Surgical versus non‐surgical treatment for carpal tunnel syndrome

Vieda Lusa; Teemu V Karjalainen; Markus Pääkkönen; Tuomas Jaakko Rajamäki; Kati Jaatinen

doi:10.1002/14651858.CD001552.pub3

. 2024 Jan 8;2024(1):CD001552. doi: 10.1002/14651858.CD001552.pub3

Surgical versus non‐surgical treatment for carpal tunnel syndrome

Vieda Lusa ¹, Teemu V Karjalainen ^2,^3,^✉, Markus Pääkkönen ⁴, Tuomas Jaakko Rajamäki ⁵, Kati Jaatinen ⁶

Editor: Cochrane Neuromuscular Group

PMCID: PMC10772978 PMID: 38189479

Abstract

Background

Carpal tunnel syndrome (CTS) is a compression neuropathy of the median nerve at the wrist. Surgery is considered when symptoms persist despite the use of non‐surgical treatments. It is unclear whether surgery produces a better outcome than non‐surgical therapy. This is an update of a Cochrane review published in 2008.

Objectives

To assess the evidence regarding the benefits and harms of carpal tunnel release compared with non‐surgical treatment in the short (< 3 months) and long (> 3 months) term.

Search methods

In this update, we included studies from the previous version of this review and searched the Cochrane Neuromuscular Specialised Register, CENTRAL, Embase, MEDLINE, ClinicalTrials.gov and WHO ICTRP until 18 November 2022. We also checked the reference lists of included studies and relevant systematic reviews for studies.

Selection criteria

We included randomised controlled trials comparing any surgical technique with any non‐surgical therapies for CTS.

Data collection and analysis

We used the standard methodological procedures expected by Cochrane.

Main results

The 14 included studies randomised 1231 participants (1293 wrists). Eighty‐four per cent of participants were women. The mean age ranged from 32 to 53 years, and the mean duration of symptoms from 31 weeks to 3.5 years. Trial sizes varied from 22 to 176 participants.

The studies compared surgery with: splinting, corticosteroid injection, splinting and corticosteroid injection, platelet‐rich plasma injection, manual therapy, multimodal non‐operative treatment, unspecified medical treatment and hand support, and surgery and corticosteroid injection with corticosteroid injection alone.

Since surgery is generally used for its long‐term effects, this abstract presents only long‐term results for surgery versus splinting and surgery versus corticosteroid injection.

1) Surgery compared to splinting in the long term (> 3 months)

Surgery probably results in a higher rate of clinical improvement (risk ratio (RR) 2.10, 95% confidence interval (CI) 1.04 to 4.24; 3 studies, 210 participants; moderate‐certainty evidence).

Surgery probably does not provide clinically important benefit in symptoms or hand function compared with splinting (moderate‐certainty evidence). The mean Boston Carpal Tunnel Questionnaire (BCTQ) Symptom Severity Scale (scale 1 to 5; higher is worse; minimal clinically important difference (MCID) = 1) was 1.54 with splint and 0.26 points better with surgery (95% CI 0.52 better to 0.01 worse; 2 studies, 195 participants). The mean BCTQ Functional Status Scale (scale 1 to 5; higher is worse; MCID 0.7) was 1.75 with splint and 0.36 points better with surgery (95% CI 0.62 better to 0.09 better; 2 studies, 195 participants). None of the studies reported pain. Surgery may not provide better health‐related quality of life compared with splinting (low‐certainty evidence). The mean EQ‐5D index (scale 0 to 1; higher is better; MCID 0.074) was 0.81 with splinting and 0.04 points better with surgery (95% CI 0.0 to 0.08 better; 1 study, 167 participants). We are uncertain about the risk of adverse effects (very low‐certainty evidence). Adverse effects were reported amongst 60 of 98 participants (61%) in the surgery group and 46 of 112 participants (41%) in the splinting group (RR 2.11, 95% CI 0.37 to 12.12; 2 studies, 210 participants).

Surgery probably reduces the risk of further surgery; 41 of 93 participants (44%) were referred to surgery in the splinting group and 0 of 83 participants (0%) repeated surgery in the surgery group (RR 0.03, 95% CI 0.00 to 0.21; 2 studies, 176 participants). This corresponds to a number needed to treat for an additional beneficial outcome (NNTB) of 2 (95% CI 1 to 9).

2) Surgery compared to corticosteroid injection in the long term (> 3 months)

We are uncertain if clinical improvement or symptom relief differs between surgery and corticosteroid injection (very low‐certainty evidence). The RR for clinical improvement was 1.23 (95% CI 0.73 to 2.06; 3 studies, 187 participants). For symptoms, the standardised mean difference (SMD) was ‐0.60 (95% CI ‐1.88 to 0.69; 2 studies, 118 participants). This translates to 0.4 points better (95% CI from 1.3 better to 0.5 worse) on the BCTQ Symptom Severity Scale. Hand function or pain probably do not differ between surgery and corticosteroid injection (moderate‐certainty evidence). For function, the SMD was ‐0.12 (95% CI ‐0.80 to 0.56; 2 studies, 191 participants) translating to 0.10 points better (95% CI 0.66 better to 0.46 worse) on the BCTQ Functional Status Scale with surgery. Pain (0 to 100 scale) was 8 points with corticosteroid injection and 6 points better (95% CI 10.45 better to 1.55 better; 1 study, 123 participants) with surgery. We found no data to estimate the difference in health‐related quality of life (very low‐certainty evidence).

We are uncertain about the risk of adverse effects and further surgery (very low‐certainty evidence). Adverse effects were reported amongst 3 of 45 participants (7%) in the surgery group and 2 of 45 participants (4%) in the corticosteroid injection group (RR 1.49, 95% CI 0.25 to 8.70; 2 studies, 90 participants). In one study, 12 of 83 participants (15%) needed surgery in the corticosteroid group, and 7 of 80 participants (9%) needed repeated surgery in the surgery group (RR 0.61, 95% CI 0.25 to 1.46; 1 study, 163 participants).

Authors' conclusions

Currently, the efficacy of surgery in people with CTS is unclear. It is also unclear if the results can be applied to people who are not satisfied after trying various non‐surgical options. Future studies should preferably blind participants from treatment allocation and randomise people who are dissatisfied after being treated non‐surgically.

The decision for a patient to opt for surgery should balance the small benefits and potential risks of surgery. Patients with severe symptoms, a high preference for clinical improvement and reluctance to adhere to non‐surgical options, and who do not consider potential surgical risks and morbidity a burden, may choose surgery. On the other hand, those who have tolerable symptoms, who have not tried non‐surgical options and who want to avoid surgery‐related morbidity can start with non‐surgical options and have surgery only if necessary. We are uncertain if the risk of adverse effects differs between surgery and non‐surgical treatments. The severity of adverse effects may also be different.

Keywords: Adult, Female, Humans, Male, Middle Aged, Adrenal Cortex Hormones, Adrenal Cortex Hormones/therapeutic use, Carpal Tunnel Syndrome, Carpal Tunnel Syndrome/surgery, Drug-Related Side Effects and Adverse Reactions, Occupational Therapy, Pain, Quality of Life

Plain language summary

Surgical versus non‐surgical treatment for carpal tunnel syndrome

Surgery or non‐surgical treatment: which works better for carpal tunnel syndrome?

Key messages

Surgery probably results in a higher rate of clinical improvement than splinting after a follow‐up of 6 to 12 months. The evidence on which of the two treatments may have less harmful effects is uncertain.

The evidence is uncertain if clinical improvement or the rate of harmful effects differ between surgery and corticosteroid injection after a follow‐up of 6 to 12 months.

Generally, we lack confidence about the efficacy of surgery in people with carpal tunnel syndrome because we did not find studies comparing surgery with placebo surgery or no treatment. Future studies should address this evidence gap.

What is carpal tunnel syndrome?

Carpal tunnel syndrome is a condition in which the median nerve at the wrist is compressed, causing numbness, tingling in the thumb, index and middle finger, and pain. In severe cases, skin sensation can be permanently diminished and compression may cause muscle wasting at the base of the thumb.

How is carpal tunnel syndrome treated?

Usually, non‐surgical treatments such as splints, corticosteroid injections, exercises and manual therapy are offered as first‐line treatments. Surgery is considered for people with persisting symptoms and sometimes as the primary treatment for people with severe symptoms.

What did we want to find out?

If surgery or non‐surgical treatment is more beneficial and less harmful for treating carpal tunnel syndrome.

What did we do?

We searched for studies that compared surgery with 1) no treatment or placebo treatment, or 2) any non‐surgical treatment. We compared the rate of clinical improvement, symptoms, hand function, pain and health‐related quality of life, as well as harmful effects and need for further surgery. We collected and analysed data according to Cochrane methods.

What did we find?

We found 14 studies randomising 1231 people: mean age ranged between 32 and 53 years; 84% female; representing nine countries from Asia, Europe and North America; symptom duration 31 weeks to 3.5 years with a varying degree of severity. Surgery was compared with 1) splinting, 2) corticosteroid injection, 3) splinting and corticosteroid injection, 4) platelet‐rich plasma injection, 5) manual therapy (three 30‐minute treatment sessions including desensitisation manoeuvres, once per week), 6) multimodal non‐operative treatment (combination of six visits of hand therapy, activity modification, nonsteroidal anti‐inflammatory drugs, splinting, followed by ultrasound treatment if needed), 7) unspecified medical treatment and hand support. Additionally, one study compared corticosteroid injection with surgery plus corticosteroid injection. Two studies reported including people who had been unresponsive to at least 2 weeks of non‐surgical treatment. Study sizes varied from 22 to 176 people. Ten studies measured outcomes at long‐term follow‐up (over 3 months, usually 6 or 12 months).

We did not find studies comparing surgery with placebo surgery or no treatment.

Key results

Because surgery is often used for its long‐term effects, this summary focusses on long‐term follow‐up (6 to 12 months).

Surgery compared to splinting

Surgery probably results in a higher rate (twice) of clinical improvement compared with splinting. However, surgery may not be more beneficial than splinting for improving symptoms or hand function, or general health‐related quality of life.

We are uncertain if the risk of harmful effects differs between surgery and splinting. Some people in the splinting group had surgery‐related harms because they had undergone surgery before the outcomes were measured. However, surgery probably reduces the need for further surgery.

Surgery compared to corticosteroid injection

We are uncertain if clinical improvement, symptom relief, risk of harmful effects and the need for further surgery differ between surgery and corticosteroid injection. Hand function or pain probably do not differ considerably.

Other comparisons

Surgery is probably slightly more beneficial than multimodal non‐operative treatment for improving symptoms, but may not provide benefits for other outcomes.

Surgery probably results in a higher rate of clinical improvement than manual therapy (1.6 times) but may not provide benefits for other outcomes.

We are uncertain if surgery provides more benefits compared with a combination of splint and corticosteroid injection.

What are the limitations of the evidence?

The most important limitation is that there is no existing evidence to determine whether surgery is better than no treatment or if surgery is more effective than continuing non‐surgical treatments for people who have not improved with non‐surgical interventions. Additionally, it is unclear if surgery could provide more long‐lasting effects after several years' follow‐up.

Our confidence in the evidence regarding the differences between surgery and non‐surgical treatment was mostly affected by the fact that people in all the studies were aware of which treatment they received. This may affect how people report the outcomes and result in biased estimations of treatment effects. Furthermore, in two large studies, up to 40% of people, who were randomised to non‐surgical treatment, opted for surgery during the study and had already undergone it by the time long‐term outcomes were measured. If these people had not had surgery, their outcomes might have been worse, with the resulting benefits of surgery being underestimated.

How up‐to‐date is this evidence?

The evidence is up‐to‐date until November 2022.

Summary of findings

Summary of findings 1. Surgery compared to splint for carpal tunnel syndrome in the long term.

Surgery compared to splint for carpal tunnel syndrome
Patient or population: people with carpal tunnel syndrome  Setting: outpatient clinics in the Netherlands and Turkey, and hospital in the UK Intervention: surgery  Comparison: splint
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect (95% CI)	Number of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with splint	Risk with surgery
Clinical improvement ‐ Long‐term improvement: > 3 months	Study population		RR 2.10 (1.04 to 4.24)	210 (3 RCTs)	⊕⊕⊕⊝ Moderate^a	Surgery probably results in higher clinical improvement at long‐term follow‐up. Absolute risk difference 50% better (12% better to 89% better) with surgery. NNTB 2 (95% CI 2 to 9).
	603 per 1000	925 per 1000 (628 to 1000)
Symptoms (Boston CTS questionnaire) ‐ Long‐term improvement: > 3 months Scale: 1 to 5, higher is worse	The mean symptoms was 1.54 points (data from 1 study used)	MD 0.26 better (0.52 better to 0.00 better or worse)	‐	195 (2 RCTs)	⊕⊕⊕⊝ Moderate^a	Surgery probably does not provide important symptom relief compared with splinting. Absolute difference 6.5% better (13% better to 0% better/worse) with surgery.
Function (Boston CTS questionnaire) ‐ Long‐term improvement: > 3 months Scale: 1 to 5, higher is worse	The mean function was 1.75 points (data from 1 study used)	MD 0.36 better (0.62 better to 0.09 better)	‐	195 (2 RCTs)	⊕⊕⊕⊝ Moderate^a	Surgery probably provides little (clinically unimportant) benefit compared with splinting at long‐term. Absolute difference 9% better (2.25% better to 15.5% better) with surgery.
Pain ‐ Long‐term improvement: > 3 months	‐	‐	‐	‐	‐	No data. We are uncertain about the effect.
Health‐related quality of life (EQ‐5D) ‐ Long‐term improvement: > 3 months  Scale: 0 to 1, higher is better	The mean health related quality of life (EQ‐5D) was 0.81 points	MD 0.04 points better (0 to 0.08 better)	‐	167 (1 RCT)	⊕⊕⊝⊝ Low^a,b	Surgery may not provide clinically important benefits compared with splinting. Absolute difference 4% better (95% CI 0% different to 8% better) with surgery.
Adverse effects	Study population		RR 2.11 (0.37 to 12.12)	210 (2 RCTs)	⊕⊝⊝⊝ Very low^a,c	We are uncertain if surgery is associated with more or fewer adverse effects compared with splinting. Absolute risk difference 16% higher risk (95% CI 4% to 28% higher) with surgery.
	411 per 1000	867 per 1000 (152 to 1000)
Need for surgery or secondary surgery during follow‐up ‐ Long‐term improvement: > 3 months	Study population		RR 0.03 (0.00 to 0.21)	176 (2 RCTs)	⊕⊕⊕⊝ Moderate^a	Surgery probably results in fewer reoperations compared with risk of surgery with splinting. Absolute risk difference 63% lower (114% lower to 12% lower) with surgery. NNTB 2 (95% CI 1 to 9).
	441 per 1000	13 per 1000 (0 to 93)
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

Open in a new tab

^aDowngraded once for high overall risk of bias in the included studies (mainly due to the lack of blinding). ^bDowngraded once for imprecision as the 95% confidence interval overlapped with the MCID value. ^cDowngraded twice for serious imprecision as 95% confidence intervals included substantial effect in both directions.

Abbreviations: CI: confidence interval; CTS: carpal tunnel syndrome; EQ‐5D: instrument for measuring health‐related quality of life; MCID: minimal clinically important difference; MD: mean difference; NNTB: number needed to treat for an additional beneficial outcome; RCTs: randomised controlled trials; RR: risk ratio.

Summary of findings 2. Surgery compared to corticosteroid injection for carpal tunnel syndrome in the long term.

Surgery compared to corticosteroid injection for carpal tunnel syndrome
Patient or population: people with carpal tunnel syndrome  Setting: hospitals in Hong Kong, Iran, Pakistan and Spain Intervention: surgery  Comparison: corticosteroid injection
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect (95% CI)	Number of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with corticosteroid injection	Risk with surgery
Clinical improvement ‐ Long‐term improvement: > 3 months	Study population		RR 1.23 (0.73 to 2.06)	187 (3 RCTs)	⊕⊝⊝⊝ Very low^a,b,c	We are uncertain if clinical improvement differs between surgery and corticosteroid injection at long‐term follow‐up. Absolute difference 15% better (95% CI 23% worse to 54% better) with surgery.
	615 per 1000	757 per 1000 (449 to 1000)
Symptoms ‐ Long‐term improvement: > 3 months	‐	SMD 0.6 lower (1.88 lower to 0.69 higher)	‐	118 (2 RCTs)	⊕⊝⊝⊝ Very low^a,d,e	We are uncertain if symptom severity differs between surgery and corticosteroid injection at long‐term follow‐up. SMD translates to 0.4 points better in BCTQ symptom severity score (95% CI from 1.3 better to 0.5 worse) with surgery compared with corticosteroid injection (using baseline SD of 0.70 from Jarvik 2009). Absolute difference 10% better (95% CI 32.5% better to 12.5% worse) with surgery.
Function ‐ Long‐term improvement: > 3 months	‐	SMD 0.12 lower (0.8 lower to 0.56 higher)	‐	191 (2 RCTs)	⊕⊕⊕⊝ Moderate^a	Surgery probably does not improve function compared with corticosteroid injection at long‐term follow‐up. SMD translates to 0.10 points better in BCTQ functional status score (95% CI from 0.66 better to 0.46 worse) with surgery compared with corticosteroid injection (using baseline SD of 0.82 from Jarvik 2009). Absolute difference 2.5% better (95% CI 16.5% better to 11.5% worse) with surgery.
Pain (VAS) ‐ Long‐term improvement: > 3 months Scale: 0 to 100, higher is worse	The mean pain was 8 points	MD 6 points better (10.45 better to 1.55 better)	‐	123 (1 RCT)	⊕⊕⊕⊝ Moderate^a	Surgery probably provides a small but clinically unimportant, benefit on pain compared with corticosteroid injection. Absolute difference 6% better (95% CI 1.6% to 10.5% better) with surgery.
Health‐related quality of life ‐ Long‐term improvement: > 3 months	‐	‐	‐	‐	‐	No data. We are uncertain about the effect.
Adverse effects	Study population		RR 1.49 (0.25 to 8.70)	90 (2 RCTs)	⊕⊝⊝⊝ Very low^a,c	We are uncertain if surgery is associated with more or fewer adverse effects compared with corticosteroid injection. Absolute risk difference 2% higher risk (95% CI 7% lower to 11% higher) with surgery.^f
	44 per 1000	66 per 1000 (11 to 387)
Need for surgery or secondary surgery during follow‐up ‐ Long‐term improvement: > 3 months	Study population		RR 0.61 (0.25 to 1.46)	163 (1 RCT)	⊕⊝⊝⊝ Very low^a,c	We are uncertain whether surgery or corticosteroid injection results in higher risk of subsequent surgery at long‐term follow‐up. Absolute risk difference 6% lower risk (95% CI 15% lower to 4% higher) with surgery.
	145 per 1000	88 per 1000 (36 to 211)
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

Open in a new tab

^aDowngraded once for high overall risk of bias in the included studies (mainly due to the lack of blinding). ^bDowngraded once for inconsistency (substantial unexplained heterogeneity). ^cDowngraded two levels for very serious imprecision; 95% confidence intervals included substantial effect in both directions. ^dDowngraded once for imprecision as the 95% confidence interval overlapped with the MCID value. ^eDowngraded once for inconsistency. ^fOne study reported that no adverse effects or complications occurred during the trial, with 0 of 83 participants in the corticosteroid group and 0 of 80 participants (0%) in the surgery group (Ly Pen 2005). Due to zero events, this study did not contribute to the meta‐analysis.

Abbreviations: BCTQ: Boston Carpal Tunnel Syndrome Questionnaire; CI: confidence interval; CTS: carpal tunnel syndrome; MCID: minimal clinically important difference; MD: mean difference; RCTs: randomised controlled trials; RR: risk ratio; SMD: standardised mean difference; VAS: visual analogue scale.

Background

Description of the condition

Carpal tunnel syndrome (CTS) is a cluster of symptoms and clinical findings resulting from entrapment of the median nerve at the wrist level, where it passes under the transverse carpal ligament through the carpal tunnel.

The most common symptoms are tingling, numbness and pain within the median nerve distribution (particularly the thumb, index and middle fingers), worsening typically at night and sometimes during hand use. Pain may radiate proximally to the forearm or shoulder. In severe cases, the skin sensation can be permanently diminished and axonal injury results in atrophy (wasting) of the thenar muscles (muscles at the base of the thumb) and loss of dexterity of the hand (Padua 2016).

The carpal tunnel is a short, confined space within which pressure may rise. Altered location, shape or mobility of the median nerve and tendons in the carpal tunnel can also contribute to the increased pressure on the median nerve (Liong 2013). Association with metabolic syndrome and diabetes suggests that reduced perfusion due to vascular changes may play a key role in the development of CTS (Shiri 2011). Initially, the symptoms are caused by ischaemia in the median nerve. This results in injury to the myelin sheath and eventually can lead to axonal injury.

CTS is the most common entrapment neuropathy (pathology in which a nerve becomes compressed or entrapped between two other structures in the body) (Martyn 1997; Stewart 1993). A study in the Netherlands found a prevalence of 9.2% in the female population (5.8% undetected CTS and 3.4% diagnosed CTS) and 0.6% in the male population (de Krom 1992). In Sweden, nearly 3% of the general population had clinically and electrophysiologically‐confirmed CTS (Atroshi 1999). It has an important economic impact, affecting active people and may present as a work‐related disorder (Rossignol 1997), leading to compensation claims (Leigh 1998).

CTS is a clinical diagnosis based upon history and clinical examination. There are no universally accepted diagnostic clinical and laboratory criteria. However, it is agreed that certain abnormalities in the function of the nerve support the diagnosis. The most frequently used measured parameters are distal motor and sensory latencies and sensory conduction velocity across the carpal tunnel (Stevens 1997). Other techniques have been used, such as comparisons between the distal sensory or motor latencies stimulating the ulnar and median nerve (Felsenthal 1977), or between the radial and median nerves (Carroll 1987). The 'inching technique' allows precise localisation of the site of entrapment (Kimura 1979), but its clinical relevance is under debate (Geiringer 1998). Ultrasound has also been used for diagnostic purposes (Erickson 2021).

There is no universally accepted therapy for CTS (Rosenbaum 1993). A wide range of non‐surgical treatments are offered initially, which vary widely around the world, within individual countries, and even within hospitals. Typically, most patients are first treated non‐surgically with splints, corticosteroids, various exercises or manual therapy, but the evidence of the efficacy of non‐operative treatments is inconsistent, and people can also improve without any treatment (Burton 2016; Erickson 2021).

Between 23% and 80% of people trying non‐surgical treatments are not satisfied with the outcome and many (up to 60% to 90%) eventually progress to surgery because their symptoms persist despite non‐surgical management (Burton 2016; Hofer 2021). However, the high rate of surgery may arise from its perceived benefits, rather than its greater effectiveness. Predictors of poor outcomes with non‐surgical treatment are long symptom duration, a positive Phalen's test, and thenar muscle wasting, more than one non‐surgical intervention and a high initial symptom burden. On the other hand, short symptom duration (less than one year) and less severe night‐time symptoms predict success with non‐surgical management (Erickson 2021).

Description of the intervention

In surgical treatment (commonly referred to as carpal tunnel release), the surgeon divides the transverse carpal ligament through an open incision (of 2 cm to 4 cm), mini‐open incision (of 1 cm to 2 cm) or endoscopically (a less than 1 cm incision), after which the skin is closed. The procedure can be performed under local, regional or general anaesthesia. The operation normally takes from 10 to 20 minutes. Symptoms arising from ischaemic changes (e.g. temporary numbness or tingling) can vanish immediately after release, but recovery of the nerve may take months.

How the intervention might work

In surgical treatment, division of the ligament opens the carpal tunnel, and this decreases the pressure inside it (Murata 2007; Schuind 2002).

The mechanism of action for non‐surgical treatments varies between particular treatments. Pressure reduction is also considered to be the mechanism of action for splinting, because a straight position of the wrist ensures the lowest pressure in the carpal tunnel (Gelberman 1984). Steroid injections may provide anti‐fibrotic and anti‐inflammatory effects in the subsynovial connective tissue in the carpal tunnel (Yamanaka 2021). It is also believed to have an anti‐oedematous effect. Platelet‐rich plasma injection is hypothesised to have an anti‐fibrotic effect on the nerve repair site, but empiric data about the mechanism is very limited (Karjalainen 2022a). As for manual therapy, the theoretical model suggests that including neurodynamic exercises (exercises that aim to change the mechanical characteristics around the nerve) could improve the gliding of the median nerve within the carpal tunnel, reducing stress, and also improve the nerve's ability to tolerate mechanical loading (Bialosky 2009).

Why it is important to do this review

This is an update of a previous Cochrane review that aimed to assess the benefits and harms of carpal tunnel release compared with non‐operative treatment (Verdugo 2008). It concluded that surgical treatment of CTS relieves symptoms significantly better than splinting, but it is uncertain if surgery is better compared with corticosteroid injection. Carpal tunnel release is one of the most common surgical procedures, yet its effect compared with no surgery is unclear. Since the last version of the review, several new trials have been published comparing surgery with various non‐operative treatments, making an update timely.

Objectives

To assess the evidence regarding the benefits and harms of carpal tunnel release compared with non‐surgical treatment in the short (up to 3 months) and long (over 3 months) term.

Methods

Criteria for considering studies for this review

Types of studies

We included all published and unpublished randomised controlled trials (RCTs) without language restrictions, regardless of publication status, and regardless of whether they contained outcomes of interest.

Types of participants

We included all participants diagnosed with carpal tunnel syndrome (CTS) (as defined by the authors) irrespective of the diagnostic criteria used, aetiology of the syndrome, associated pathology, gender and age. If studies with participants with other compression neuropathies had been included in the study, we planned to include only patients with CTS if possible and exclude the study if the data were not obtainable.

Due to the lack of agreement regarding criteria for diagnosis of CTS, we considered all studies of symptomatic patients including a control group, regardless of the diagnostic criteria applied. We planned subgroup analysis of trials using the American Academy of Neurology (AAN) practice parameter for the diagnosis of CTS (AAN 1993), if data had been available.

Types of interventions

We included all surgical techniques and all non‐surgical treatments, no treatment or placebo surgery. Because there is no universally accepted surgical technique for the treatment of this condition, we included all procedures, such as open or endoscopic section of the transverse carpal ligament.

Types of outcome measures

For this update, we focused on patient‐reported outcomes.

We did not consider electrodiagnostic outcomes in this update as their clinical relevance is unclear (Schrijver 2005; see also Differences between protocol and review).

Primary outcomes

Clinical improvement at:

the short term (up to 3 months of follow‐up); and
the long term (over 3 months of follow‐up).

Improvement was considered relevant if it implied significant relief of symptoms or pain, or improvement in function from the baseline, as defined by the study authors.

Secondary outcomes

1. Symptoms (prioritising Boston Carpal Tunnel Questionnaire, BCTQ) at:

the short term (up to 3 months of follow‐up);
the long term (over 3 months of follow‐up).

2. Function (prioritising BCTQ) at:

the short term (up to 3 months of follow‐up);
the long term (over 3 months of follow‐up).

3. Pain (measured by a visual analogue scale, VAS, or numerical rating scale, NRS, or any other scale) at:

the short term (up to 3 months of follow‐up);
the long term (over 3 months of follow‐up).

4. Health‐related quality of life at:

the short term (up to 3 months of follow‐up);
the long term (over 3 months of follow‐up).

5. Adverse effects (as defined by the authors, including formation of a painful neuroma (disorganised growth of nerve cells) of the palmar cutaneous branch of the median nerve, tender or hypertrophic scar, section of the motor branch, subluxation ('bow stringing') of flexor tendons, wound infection and complex regional pain syndrome) at the final time point of the study.

6. Need for surgery (in participants treated non‐surgically), or secondary surgery (in those treated surgically) during follow‐up at:

the short term (up to 3 months of follow‐up);
the long term (over 3 months of follow‐up).

We considered the long‐term time point as the primary time point of this review.

Search methods for identification of studies

Electronic searches

On 18 November 2022, the Information Specialist searched the following resources:

Cochrane Neuromuscular's Specialised Register via Cochrane Register of Studies‐Web (until search date; Appendix 1);
Cochrane Central Register of Controlled Trials (CENTRAL) via the Cochrane Library (until search date; Appendix 2);
MEDLINE via Ovid SP (1946 to 17 November 2022; Appendix 3);
Embase via Ovid SP (1974 to 2022 Week 45; Appendix 4);
ClinicalTrials.gov (until search date; Appendix 5);
World Health Organization International Clinical Trials Registry Platform (WHO ICTRP) (until search date; Appendix 6).

There were no limitations to the searches. For details on the search in the previous version of this review, see Appendix 7.

Searching other resources

We checked the reference lists of the included studies and relevant systematic reviews.

Data collection and analysis

The review authors followed the recommended strategies for data collection and analysis as documented in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2020).

Selection of studies

Four review authors (KJ, VL, MP and TR) working in pairs independently screened the titles and abstracts to identify potentially eligible studies conforming to the predetermined inclusion criteria (study is randomised controlled trial; study investigates surgery versus non‐surgical treatment, no treatment or placebo surgery for carpal tunnel syndrome). If a title or abstract suggested that the trial was eligible for inclusion, or we could not be certain, we obtained the full text of the paper, and two review authors independently assessed it. We resolved discrepancies through discussion or adjudication by a third author (TK). We collated several references related to the same study during the full‐text screening phase.

Data extraction and management

Four authors working in pairs using a piloted structured sheet independently extracted data. Any disagreements were settled by discussion. The authors extracted trial characteristics and review outcomes at all the provided time points.

We extracted:

trial characteristics: study design, country, sample size, primary analysis, source of funding, and trial registration status;
number of participants randomised, number analysed at each time point, inclusion/exclusion criteria, participant key characteristics (age, sex, duration of symptoms, outcomes at baseline);
intervention characteristics for the treatment groups, and use of co‐interventions;
outcomes at each time point.

We always prioritised values from intention‐to‐treat analysis over values from 'per protocol' or 'as‐treated' analyses. If the trial authors reported both final values and change from baseline values for the same outcome, we extracted the final values. If trial authors reported several outcomes for one outcome domain (e.g. function), we chose the outcome based on a priori hierarchies (see Types of outcome measures). If outcomes had different scales or directions, we transformed or reversed the scales to match the most commonly used instrument scale before pooling.

For adverse effects, we excluded pain during or immediately after injection or surgery, since we considered this to be an inherent part of the intervention and not an adverse effect. However, if the authors reported long‐term or disproportionate pain (over 3 days for injections, or if the participant needed to seek care for postoperative pain), we considered that an adverse effect.

Assessment of risk of bias in included studies

Four review authors (KJ, VL, MP and TR) working in pairs independently assessed the risk of bias for the outcomes of this review using the risk of bias (RoB) tool version 2 (Sterne 2019). We assessed all outcomes and assessed the risk of bias separately for each outcome at long‐term and short‐term follow‐up. We resolved discrepancies through discussion or adjudication by a third author (TK). We used the Excel tool to implement the RoB 2 tool (available at www.riskofbias.info, version as of 22 August 2019). Future updates may focus on assessing the risk of bias for the outcomes included in the summary of findings tables only.

Our primary interest was the effect of the assignment to the intervention (the 'intention‐to‐treat effect'). The Excel tool's algorithm defined the risk of bias as either 'low', 'some concerns' or 'high' in five domains: 1) randomisation process; 2) deviation from intended interventions, including evaluation if co‐interventions were balanced between the study groups; 3) missing outcome data; 4) measurement of the outcome; and 5) selection of the reported result. The risk of bias for a study outcome was considered to be at high risk if any domain was at high risk of bias or more than one domain was judged as having 'some concerns'. When only one domain was judged as having 'some concerns', the overall risk of bias for a study outcome was 'some concerns'.

In the domain of deviations from the intended interventions, we rated 'some concerns' when a high rate of participants who were allocated to non‐surgical treatment had surgery before the outcome measurement, even if the authors used the intention‐to‐treat principle. Although switching to surgery often occurs in usual care when people are not satisfied with non‐surgical treatment, we deemed that the effect does not reflect the difference between surgery and non‐surgical treatment, but rather the difference between having immediate surgery or having surgery when the non‐surgical treatment fails to improve the symptoms.

In the domain of missing outcome data, we considered over 20% loss in any of the study groups as a reason for an assessment of high risk, and 10% to 20% loss as raising 'some concerns', particularly if the loss was not balanced.

Measures of treatment effect

We performed statistical analysis using Cochrane Review Manager (RevMan) software (Review Manager 2020). We used the risk ratio (RR) and 95% confidence interval (CI) for dichotomous outcomes. When we found a statistically significant difference, we also calculated the number needed to treat for an additional beneficial outcome (NNTB) or the number needed to treat for an additional harmful outcome (NNTH) from surgery as 1/risk difference. NNTB/NNTH were reported in whole numbers, rounded up.

We expressed the difference as mean difference (MD) with 95% CI for continuous outcomes. If we could not pool mean values because of several outcome measures used, we pooled the results as a standardised mean difference (SMD) and then back‐transformed the SMD to the most commonly used outcome measure (multiplying the SMD and its 95% CI by the standard deviation (SD) from the most representative study).

Unit of analysis issues

The unit of analysis was the participant wherever possible. If a study provided data only at the wrist level, we used the wrist as a unit, and noted this in the Characteristics of included studies table of the study.

In the case of multi‐arm studies, we compared all the non‐surgical arms with surgical arms in separate analysis, to avoid double‐counting the same participants in the total number of the analysed participants.

Dealing with missing data

When data were missing from trial reports, we contacted the authors. When the SD was not reported, we calculated it based on standard errors, 95% CIs, the reported P values, interquartile range (as IQR/1.35) or range (range/4) (Higgins 2020). If we found no relevant data to calculate or estimate the SD, we imputed the value from another study using the same outcome (using control group baseline SD from the most representative study). If we had to calculate or impute the SD, we noted this in the Characteristics of included studies tables.

Assessment of heterogeneity

We assessed clinical diversity by comparing participant characteristics and used interventions to assess if pooling the studies in a meta‐analysis was appropriate.

We assessed statistical heterogeneity by visual inspection and using the I² statistic. We interpreted the I² statistic using the following guidelines:

0% to 40% might not be important;
30% to 60% may represent moderate heterogeneity;
50% to 90% may represent substantial heterogeneity;
75% to 100% represents considerable heterogeneity (Deeks 2021).

When substantial to considerable heterogeneity was present, we explored the reasons for it and, if no plausible reasons existed, we downgraded the certainty of evidence for inconsistency and noted it in the Effects of interventions section and in Table 1 and Table 2. When heterogeneity could be explained (e.g. one study explained most heterogeneity and measured the outcomes at different time points or with different instruments), we noted this and did not downgrade for inconsistency. When one study drove the heterogeneity, we reported the difference between the groups, primarily including the study and, secondarily, excluding it.

Assessment of reporting biases

We planned to use funnel plots whenever 10 or more studies were included in meta‐analyses, but did not do so as we did not have a sufficient number of studies in any analysis.

Data synthesis

We defined the following review questions, in people with CTS.

Is surgery (carpal tunnel release) more effective than placebo surgery or no treatment?
Is surgery more effective than splinting?
Is surgery more effective than corticosteroid injection?
Is surgery more effective than splinting and corticosteroid injection together?
Is surgery more effective than platelet‐rich plasma injection?
Is surgery more effective than manual therapy?
Is surgery more effective than multimodal non‐operative treatment?
Is surgery more effective than unspecified medical treatment and hand support together?
Is surgery combined with corticosteroid injection more effective than corticosteroid injection alone?

We considered surgery versus placebo surgery or no treatment to be the primary comparison but, since we identified no studies for this comparison, we considered surgery versus splinting and surgery versus corticosteroid injection as the primary comparisons. Surgery could be followed by postoperative physical therapy or rehabilitation or an exercise programme.

We pooled data from included trials with similar characteristics to provide estimates of benefits and harms at short term follow‐up (up to three months) and long‐term follow‐up (over three months). We pooled outcomes using the random‐effects model as a default. The primary analysis included all eligible studies.

We used minimal clinically important difference (MCID) values to assess the clinical importance of differences in patient‐reported outcomes: a 1‐point difference as a MCID for the BCTQ Symptom Severity Scale and 0.7 points for the Functional Status Scale (Kim 2013); 0.074 points difference as a MCID for the EQ‐5D (Walters 2005); 5 points difference for the global symptom score (MCID value unknown, thus we used 10% from the scale); 1.5 points for the VAS from 0 to 10 (Hao 2019), 11 points for the Numerical Rating Scale, 15 points difference for the VAS from 0 to 100 (MCID value not defined; we used 15% from these scales); and 2.5 points difference for SF‐36 Physical Component Score and Mental Component Score (Teitsma 2017).

Subgroup analysis and investigation of heterogeneity

We did not perform subgroup analyses. We planned subgroup analysis based on the severity of CTS (mild, moderate, severe, based on the American Academy of Neurology (AAN) practice parameter for the diagnosis of CTS (AAN 1993)), if data had been available.

Sensitivity analysis

We planned sensitivity analyses for the outcomes at low risk of bias compared to those at high or unclear risk of bias in all domains, and low risk versus high or unclear risk of bias in the randomisation process for the primary outcome (relevant clinical improvement). Since no studies were at low risk of bias in all domains (mostly due to the lack of blinding in all studies), we could only conduct sensitivity analysis for low versus high or unclear risk in the randomisation process.

Summary of findings and assessment of the certainty of the evidence

For the primary comparisons (surgery compared with splinting, and surgery compared with corticosteroid injection), we present the outcomes (clinical improvement, symptoms, function, pain, health‐related quality of life, adverse effects, need for surgery) at long‐term follow‐up in the summary of findings tables (Table 1; Table 2) and at short‐term follow‐up in the additional tables (Table 3; Table 4). The tables summarise the treatment effects and certainty of evidence defined using the GRADE approach (Schunemann 2017). We used GRADEpro software to prepare the summary of findings tables (GRADEpro GDT 2021).

1. Surgery compared to splint for carpal tunnel syndrome in the short term.

Surgery compared to splint for carpal tunnel syndrome in the short term
Patient or population: people with carpal tunnel syndrome  Setting: outpatient clinics in the Netherlands and Turkey Intervention: surgery  Comparison: splint
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect (95% CI)	Number of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with splint	Risk with surgery
Clinical improvement ‐ Short‐term improvement: < 3 months	Study population		RR 1.06 (0.48 to 2.34)	198 (2 RCTs)	⊕⊝⊝⊝ Very low^a,b	It is uncertain if surgery or splint results in higher clinical improvement at short‐term follow‐up. Absolute risk difference was 4% better (45% worse to 52% better) with surgery.
	569 per 1000	603 per 1000 (273 to 1000)
Symptoms (Boston CTS questionnaire) ‐ Short‐term improvement: < 3 months Scale from: 1 to 5, higher is worse	The mean symptoms was 1.39 points (data from 1 study used)	MD 0.02 worse (0.83 better to 0.87 worse)	‐	198 (2 RCTs)	⊕⊕⊝⊝ Low^a,c	Surgery may not provide important symptom relief compared with splint at short‐term follow‐up. The absolute difference was 0.5% worse (20.75% better to 21.75% worse) with surgery.
Function (Boston CTS questionnaire) ‐ Short‐term improvement: < 3 months Scale: 1 to 5, higher is worse	The mean CTS function was 1.60 points (data from 1 study used)	MD 0 better/worse (0.44 better to 0.44 worse)	‐	198 (2 RCTs)	⊕⊕⊕⊝ Moderate^a	Surgery probably does not provide important benefits in hand function compared with splinting at short‐term follow‐up. The absolute difference was 0% better/worse (11% better to 11% worse) with surgery.
Pain ‐ Short‐term improvement: < 3 months	‐	‐	‐	‐	‐	No data. We are uncertain about the effect.
Health‐related quality of life ‐ Short‐term improvement: < 3 months	‐	‐	‐	‐	‐	No data. We are uncertain about the effect.
Need for surgery or secondary surgery during follow‐up ‐ Short‐term improvement: < 3 months	Study population		RR 0.08 (0.00 to 1.48)	164 (1 RCT)	⊕⊝⊝⊝ Very low^a,b	We are uncertain about the difference in risk at short‐term follow‐up. The absolute risk difference was 7% lower risk (95% CI 1% to 13% lower) with surgery.
	70 per 1000	6 per 1000 (0 to 103)
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

Open in a new tab

^aDowngraded once for high overall risk of bias in the included studies (mainly due to the lack of blinding). ^bDowngraded twice for serious imprecision as 95% confidence intervals included substantial effects in both directions. ^cDowngraded once for inconsistency.

Abbreviations: CI: confidence interval; CTS: carpal tunnel syndrome; MD: mean difference; RCTs: randomised controlled trials; RR: risk ratio.

2. Surgery compared to corticosteroid injection for carpal tunnel syndrome in the short term.

Surgery compared to corticosteroid injection for carpal tunnel syndrome in the short term
Patient or population: people with carpal tunnel syndrome  Setting: hospitals in Hong Kong, Pakistan and Spain Intervention: surgery  Comparison: corticosteroid injection
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect (95% CI)	Number of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Risk with corticosteroid injection	Risk with surgery
Clinical improvement ‐ Short‐term improvement: < 3 months	Study population		RR 0.75 (0.60 to 0.92)	185 (2 RCTs)	⊕⊕⊕⊝ Moderate^a	Surgery probably results in lower clinical improvement compared with corticosteroid injection at short‐term follow‐up. The absolute difference was 20% worse (95% CI 3% to 37% worse) with surgery, with NNTH 5 (95% CI 3 to 34).
	758 per 1000	569 per 1000 (455 to 698)
Symptoms (Global Symptom Score) ‐ Short‐term improvement: < 3 months Scale: 0 to 50, higher is worse	The mean symptom score was 16.21 points	MD 11.32 points better (21.76 better to 0.89 better)	‐	90 (2 RCTs)	⊕⊕⊝⊝ Low^a,b	The difference in symptoms may be clinically unimportant at short‐term follow‐up. The absolute difference was 22.6% better (95% CI 43.5% better to 1.78% better) with surgery.
Function (VAS functional impairment) ‐ Short‐term improvement: < 3 months Scale: 0 to 100, higher is worse	The mean functional status score was 6 points	MD 11 points worse (4.8 worse to 17.2 worse)	‐	147 (1 RCT)	⊕⊕⊝⊝ Low^a,b	Corticosteroid injection may provide clinically unimportant benefit in hand function compared with surgery. The absolute difference was 11% worse (95% CI 4.8% to 17.2% worse) with surgery.
Pain (VAS) ‐ Short‐term improvement: < 3 months Scale: 0 to 100, higher is worse	The mean pain score was 6 points	MD 9 points worse (2.79 worse to 15.21 worse)	‐	147 (1 RCT)	⊕⊕⊝⊝ Low^a,b	Corticosteroid injection may provide little benefit compared with surgery for pain relief at short‐term follow‐up. The absolute difference was 9% worse (95% CI 2.8% to 15.2% worse) with surgery.
Health‐related quality of life ‐ Short‐term improvement: < 3 months	‐	‐	‐	‐	‐	No data. We are uncertain about the effect.
Need for surgery or secondary surgery during follow‐up ‐ Short‐term improvement: < 3 months	Study population		RR 2.08 (0.19 to 22.44)	163 (1 RCT)	⊕⊝⊝⊝ Very low^a,c	We are uncertain if surgery can be associated with more or less repeated surgery than the need for surgery after corticosteroid injection in the short term. The absolute risk difference was 1% higher risk (95% CI 3% lower to 5% higher) with surgery.
	12 per 1000	25 per 1000 (2 to 270)
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

Open in a new tab

Abbreviations: CI: confidence interval; MCID: minimal clinically important difference; MD: mean difference; NNTH: number needed to harm; RCTs: randomised controlled trials; RR: risk ratio; VAS: visual analogue scale.

For binary outcomes, we present the assumed non‐surgical group risk and relative risk in the surgery group. We also calculated and noted the absolute risk difference between the surgical and non‐surgical groups.

For continuous outcomes, we reported the weighted mean value for the non‐surgical group and expressed the absolute difference as the mean difference (MD) or standardised mean difference (SMD) with 95% confidence intervals (CI). We also calculated the relative difference (relative to the scale of the measurement instrument) as the MD divided by the scale of the measure expressed as a percentage.

Two review authors (TK and VL) assessed the certainty of the evidence as high, moderate, low or very low using the five GRADE considerations to assess the body of evidence, which are as follows.

Limitations in study design or execution (risk of bias) ‐ downgrading the certainty of evidence once for high overall study risk of bias, or twice if serious overall study risk of bias.
Inconsistency in results ‐ downgrading the certainty of evidence once if unexplained heterogeneity (I²) within more than two pooled studies was present, or if the study CIs did not overlap.
Imprecision of results ‐ downgrading the certainty of evidence once if the 95% CI overlapped with the minimal clinically important difference value of the particular outcome measure, or downgrading once or twice if the 95% CIs allowed for considerable benefits for both compared treatments.
Indirectness of evidence.
Publication bias (planned if comparisons included 10 or more studies).

We reported decisions to downgrade the certainty of evidence of the findings in the Effects of interventions section, where the findings of each outcome are presented (including for those comparisons and outcomes that are not included in the summary of findings tables), as well as in the footnotes of the summary of findings tables (Table 1; Table 2; Table 3; Table 4).

We expressed the certainty of evidence using the following wording:

moderate certainty as 'probably';
low certainty as 'may';
very low certainty as 'we are uncertain'.

When the MD was smaller than our chosen value of minimal clinically important difference (MCID) and the 95% CI did not overlap zero in continuous outcomes (i.e. there was statistically significant but clinically unimportant difference between the groups), we acknowledged the benefit as a clinically unimportant benefit. This was because a wide range of MCID values have been reported in the previous literature and smaller differences compared with our choice may turn out to be important.

Results

Description of studies

See Characteristics of included studies; Characteristics of excluded studies; Characteristics of studies awaiting classification; Characteristics of ongoing studies.

Results of the search

All four trials included in the previous Cochrane review (Verdugo 2008) met the inclusion criteria for this updated review (Garland 1964; Gerritsen 2002; Hui 2005; Ly Pen 2005). One study awaiting classification in the previous Cochrane review was also completed and is included in this updated review (Ucan 2006).

The search was updated until 18 November 2022: we found 14,016 potentially relevant studies from electronic databases and seven from other sources. We screened 10,009 records (10,002 records after duplicates removed plus seven from the other sources), retrieved 63 full texts and finally included nine new studies (Awan 2015; Eltabl 2020; Fernandez‐de‐las‐Penas 2015; Fernandez‐de‐las‐Penas 2017a; Fernandez‐de‐las‐Penas 2017b; Ismatullah 2013; Jafari 2018; Jarvik 2009; Zhang 2019). We excluded 9924 records at the screening stage and 34 studies while reading the full‐text articles (26 were not RCTs, seven had inappropriate comparisons, one study was terminated) (Figure 1).

Nine studies are currently awaiting assessment for the following reasons:

unclear if the study is a randomised controlled trial (RCT) (n = 6) (Abedi 2018; Elwakil 2007; Hrkovic 2016; Linscheid 1967; Lo 2021; Shazad 2022);
trial completed but results not published (n = 3) (IRCT20200629047948N1; NCT00981565; NCT04216147).

We identified 10 ongoing studies that appeared to meet our inclusion criteria:

five studies compared surgery with corticosteroid injection (CTRI201901016881; EUCTR2013‐000873‐56‐ES; EUCTR2021‐004756‐42‐NO; NCT04014244; Palmbergen 2022);
one study compared surgery with a carpal tunnel extension device (DRKS00014585);
one study compared surgery with neural mobilisation (NCT04058041);
one study compared surgery with percutaneous electrical nerve stimulation (NCT04246216);
one study compared surgery with stretching cuff (DRKS00020570);
one study compared surgery with ultrasound‐guided median nerve hydrodissection (IRCT20141121020020N8).

Included studies

This update includes 14 RCTs, published between 1964 and 2020 (Awan 2015; Eltabl 2020; Fernandez‐de‐las‐Penas 2015; Fernandez‐de‐las‐Penas 2017a; Fernandez‐de‐las‐Penas 2017b; Garland 1964; Gerritsen 2002; Hui 2005; Ismatullah 2013; Jafari 2018; Jarvik 2009; Ly Pen 2005; Ucan 2006; Zhang 2019).

One research group published three studies with the same comparison (surgery versus manual therapy). Two were conducted in overlapping time periods (Fernandez‐de‐las‐Penas 2017a; Fernandez‐de‐las‐Penas 2017b). We contacted the authors, and they confirmed that there were no overlapping participants in these studies and consequently the studies were included as separate studies.

Participants

The studies included 1231 randomised participants, some with bilateral carpal tunnel syndrome (CTS), thus comprising 1293 wrists. Men made up 188 (16%) of participants and 1014 (84%) were women. The gender of 29 participants was not known, because three studies reported the gender distribution only for the follow‐up participants (Jafari 2018; Ucan 2006; Zhang 2019). Three studies only included women (Fernandez‐de‐las‐Penas 2015; Fernandez‐de‐las‐Penas 2017a; Fernandez‐de‐las‐Penas 2017b).

Trial size varied from 22 to 176 participants and the mean age ranged between 32 and 53 years.

The mean duration of symptoms varied from 31 weeks to 3.5 years in the 11 studies that reported it (Fernandez‐de‐las‐Penas 2015; Fernandez‐de‐las‐Penas 2017a; Fernandez‐de‐las‐Penas 2017b; Garland 1964; Gerritsen 2002; Hui 2005; Ismatullah 2013; Jarvik 2009; Ly Pen 2005; Ucan 2006; Zhang 2019). Five studies included people diagnosed as having mild‐to‐moderate CTS (Eltabl 2020; Hui 2005; Jafari 2018; Jarvik 2009; Ucan 2006), one study included people with moderate‐to‐severe CTS (Awan 2015), and any degree of CTS severity was accepted in four studies (Fernandez‐de‐las‐Penas 2015; Fernandez‐de‐las‐Penas 2017a; Fernandez‐de‐las‐Penas 2017b; Ly Pen 2005). Four studies did not report information regarding CTS severity (Garland 1964; Gerritsen 2002; Ismatullah 2013; Zhang 2019).

Two studies reported including patients who had been unresponsive to at least 2 weeks of non‐surgical treatment, including splinting (Jarvik 2009; Ly Pen 2005) (Table 5). In four studies, it was unclear if any non‐surgical treatments had been tried before enrolment (Awan 2015; Garland 1964; Hui 2005; Jafari 2018). The rest of the studies used previous interventions of interest as exclusion criteria.

3. Received treatment before inclusion in the study.

Study	Comparison	Total sample size	Interventions before inclusion	Minimum duration of symptoms for inclusion in the trial
Awan 2015	Surgery versus corticosteroid injection	116	Not reported	Not reported
Eltabl 2020	1) Surgery versus PRP injection 2) Surgery versus unspecified medical treatment and hand support	90	People were excluded if: 1) history of local corticosteroid injection in the past 3 months; 2) previous carpal tunnel release surgery; 3) NSAID consumption 2 days prior to injection.	Not reported
Fernandez‐de‐las‐Penas 2015	Surgery versus manual therapy	120	People were excluded if previous hand surgery or steroid injections treatment.	Symptoms had to have persisted for at least 12 months.
Fernandez‐de‐las‐Penas 2017a	Surgery versus manual therapy	100	People were excluded if previous surgery or steroid injections in the wrist.	Pain and paraesthesia in the median nerve distribution for at least 6 months
Fernandez‐de‐las‐Penas 2017b	Surgery versus manual therapy	100	People were excluded if previous surgery or steroid injections.	Symptoms had to have persisted for at least 12 months.
Garland 1964	Surgery versus splint	22	Not reported	Not reported
Gerritsen 2002	Surgery versus splint	176	People were excluded if previous treatment with splinting or surgery.	Not reported
Hui 2005	Surgery versus corticosteroid injection	50	Not reported	Newly diagnosed CTS of more than 3 months but less than 1 year’s duration
Ismatullah 2013	Surgery versus corticosteroid injection	40	People were excluded if recurrent cases of CTS after previous local steroid injection or previous surgery.	Longer than 3 months (patients having symptoms of less than 3 months duration were excluded)
Jafari 2018	Surgery versus corticosteroid injection	82	Not reported	Not reported
Jarvik 2009	Surgery versus multimodal non‐operative treatment	116	Participants had failed at least two consecutive weeks of non‐surgical treatment, including a trial of wrist splints, then included. People were excluded if: 1) previous carpal tunnel syndrome surgery on the study hand; 2) any wrist or hand surgery within the previous 6 months.	Not reported
Ly Pen 2005	Surgery versus corticosteroid injection	101 patients (163 wrists)	Participants were unresponsive to at least 2 weeks of NSAIDs and splinting, then included. People were excluded if: 1) previous carpal tunnel release surgery; 2) previous local injection for CTS.	Not reported
Ucan 2006	1) Surgery versus splint 2) Surgery versus splint together with corticosteroid injection	57 (available for follow‐up)	People were excluded if prior CTS treatment.	Not reported
Zhang 2019	Surgery together with corticosteroid injection versus corticosteroid injection alone	51	People were excluded if: 1) steroid injection for CTS in the preceding 6 months; 2) prior carpal tunnel decompressive surgery.	Not reported

Open in a new tab

Abbreviations: CTS: carpal tunnel syndrome; NSAID: nonsteroidal anti‐inflammatory drug; PRP: platelet‐rich plasma.

Participants in the included studies started with a moderate level of symptoms or impairment measured by the Boston Carpal Tunnel Questionnaire (BCTQ) Symptom Severity Scale and BCTQ Functional Status Scale (scale 1 to 5, higher is worse). The mean symptom severity score at baseline was 2.70 (range from 2.30 to 3.19, data available for 773 participants, 8 studies), and the mean functional status score was 2.34 (range from 1.56 to 3.05, data available for 773 participants, 8 studies).

The mean symptom severity score at the end point evaluation for all participants who received surgery was 1.51 points (range from 1 to 1.84 points, data available for 258 participants, 7 studies), but for the non‐surgically treated patients it was 1.79 points (range from 1 to 3.13 points, data available for 324 participants, 7 studies).

The mean functional status score at the end point evaluation for all participants who received surgery was also 1.51 points (range from 1.07 to 1.80, data available for 258 participants, 7 studies), and for the non‐surgically treated patients it was 1.79 points (range from 1.29 to 2.81 points, data available for 324 participants, 7 studies).

Interventions

Surgery

Participants in the surgery groups underwent carpal tunnel release using the following techniques.

Mini‐incision (Awan 2015).
Open release of transverse volar ligament (Eltabl 2020; Garland 1964; Gerritsen 2002; Ucan 2006).
Open or endoscopic decompression and release of the carpal tunnel, depending on the preferences of the surgeon and patient (Fernandez‐de‐las‐Penas 2015; Fernandez‐de‐las‐Penas 2017a; Fernandez‐de‐las‐Penas 2017b; Jarvik 2009).
Other or unspecified technique: "surgical decompression" (Hui 2005), "carpal tunnel release" (Ismatullah 2013; Jafari 2018), "limited palmar incision technique" (Ly Pen 2005), "miniscalpel‐needle release" (Zhang 2019).

Non‐surgical interventions

Three studies compared surgery with splint (Garland 1964: plaster of Paris splint for 1 month; Gerritsen 2002: custom‐made splint or a prefabricated splint at night for at least 6 weeks; Ucan 2006: standard cotton‐polyester splint at night and in the daytime whenever possible for 3 months).

Five studies compared surgery with corticosteroid injection (Awan 2015: 20 mg/mL injection of methylprednisolone; Hui 2005: 15 mg of methylprednisolone acetate; Ismatullah 2013: 40 mg of methylprednisolone; Jafari 2018: 40 mg/mL of triamcinolone acetonide; Ly Pen 2005: 20 mg/mL of paramethasone acetonide, injected once or twice).

One study compared surgery with splint (standard cotton‐polyester splint at night and in the daytime whenever possible for 3 months) and corticosteroid injection (20 mg triamcinolone acetonide and 20 mg lidocaine mixture) together (Ucan 2006).

Three studies by one author compared surgery with manual therapy, which included desensitisation manoeuvres of the central nervous system (three 30‐minute treatment sessions including desensitisation manoeuvres delivered once per week), as well as exercises as homework if necessary (the same exercises were given to the surgery group as well) (Fernandez‐de‐las‐Penas 2015; Fernandez‐de‐las‐Penas 2017a; Fernandez‐de‐las‐Penas 2017b).

One study compared surgery with multimodal non‐operative treatment, comprising nonsteroidal anti‐inflammatory drugs (ibuprofen 200 mg three times a day), six visits in a formal hand therapy, exercises, splinting, and modification for work and activity, followed by ultrasound therapy if needed (consisting of up to 12 sessions (2 to 4 per week for up to 6 weeks) of focused ultrasound at 1 Mhz, 1·0 W/cm², in pulsed mode 1:4, and 15 minutes each) (Jarvik 2009).

One study compared surgery with a single ultrasound‐guided platelet‐rich plasma injection, from 1 mL to 2 mL (Eltabl 2020), and the same study also compared surgery with unspecified medical treatment that included hand support.

One study compared surgery together with a corticosteroid injection (1.0 mL of compound betamethasone (2 mg betamethasone sodium phosphate and 5 mg betamethasone dipropionate) together with 1.0 mL of 1% lidocaine) with corticosteroid injection alone (Zhang 2019).

Co‐interventions

Six studies reported use of pain medication after treatment (Fernandez‐de‐las‐Penas 2015; Fernandez‐de‐las‐Penas 2017a; Fernandez‐de‐las‐Penas 2017b; Gerritsen 2002; Ismatullah 2013; Jarvik 2009), and seven did not provide information about postoperative pain medication (Awan 2015; Eltabl 2020; Garland 1964; Jafari 2018; Ly Pen 2005; Ucan 2006; Zhang 2019). One study reported that pain medication was not allowed during the follow‐up period (Hui 2005).

Three studies included exercises (as homework, if necessary) in addition to both main surgical and non‐surgical treatment (Fernandez‐de‐las‐Penas 2015; Fernandez‐de‐las‐Penas 2017a; Fernandez‐de‐las‐Penas 2017b).

Outcomes

1. Primary outcome: clinical improvement

Nine studies measured clinical improvement. The definition of improvement, however, varied. Awan 2015 used improvement in at least two grade scores on a visual analogue scale (VAS). Hui 2005 reported improvement by at least 50% in a global symptom score. Jafari 2018 and Ucan 2006 reported patient satisfaction on a 5‐point scale from 'completely satisfied' to 'dissatisfied'. Ly Pen 2005 reported the percentage of wrists reaching at least a 20%, 50% and 70% reduction in the VAS score for nocturnal paraesthesias, pain and functional impairment (we used "70% reduction in the VAS score for nocturnal paraesthesias" for unilateral patients and the most symptomatic wrist from bilateral patients). Garland 1964 reported the proportion of participants completely relieved of symptoms. Gerritsen 2002 reported general improvement on a 6‐point ordinal transition scale and defined success as 'completely recovered' or 'much improved'. Jarvik 2009 defined a successful outcome using three criteria: 1) improvement of 0.50 points or more or 30% improvement from baseline in Carpal Tunnel Assessment Questionnaire (CTSAQ) function; 2) improvement of 0.50 points or more or 30% improvement in CTSAQ symptom severity; and 3) a score of 0 or 1 on a 0‐ to 10‐point rating scale of hand or wrist pain interference with work or housework (we used improvement in all of these criteria). Fernandez‐de‐las‐Penas 2015 reported self‐perceived improvement using a Global Rating of Change from "a very great deal worse" to "a very great deal better" (in our analysis, the number reporting 'moderate' and 'large' improvement was used). Five studies did not measure this outcome (Eltabl 2020; Fernandez‐de‐las‐Penas 2017a; Fernandez‐de‐las‐Penas 2017b; Ismatullah 2013; Zhang 2019).

2. Symptoms

Ten studies measured symptom severity: seven used the BCTQ Symptom Severity Scale (Eltabl 2020; Fernandez‐de‐las‐Penas 2015; Fernandez‐de‐las‐Penas 2017b; Gerritsen 2002; Jafari 2018; Ucan 2006; Zhang 2019), two used the global symptom score (Hui 2005; Ismatullah 2013), and one used the CTSAQ 11‐item Symptom Severity Scale (Jarvik 2009), which is likely to be the same as the BCTQ Symptom Severity Scale. Three studies did not measure symptoms (Awan 2015; Garland 1964; Ly Pen 2005).

3. Function

Nine studies measured function: seven used the BCTQ Functional Status Scale (Eltabl 2020; Fernandez‐de‐las‐Penas 2015; Fernandez‐de‐las‐Penas 2017b; Gerritsen 2002; Jafari 2018; Ucan 2006; Zhang 2019), one used the CTSAQ 9‐item Functional Status Scale (Jarvik 2009), and one used a VAS (Ly Pen 2005). Four studies did not measure this outcome (Awan 2015; Garland 1964; Hui 2005; Ismatullah 2013).

4. Pain

Seven studies assessed pain. Four used a VAS (Awan 2015; Eltabl 2020; Jarvik 2009; Ly Pen 2005), but Awan 2015 did not report the results. Three studies used the 11‐point Numerical Rating Scale (Fernandez‐de‐las‐Penas 2015; Fernandez‐de‐las‐Penas 2017a; Gerritsen 2002). The rest of the studies did not report measuring this outcome (Garland 1964; Hui 2005; Ismatullah 2013; Jafari 2018; Ucan 2006; Zhang 2019).

5. Health‐related quality of life score (HRQoL)

Three studies reported health‐related quality of life (HRQoL) (measured using EQ‐5D in Fernandez‐de‐las‐Penas 2015 and Gerritsen 2002, and the 36‐Item Short Form Health Survey (SF‐36) in Jarvik 2009).

6. Adverse effects

Ten studies measured and reported on adverse effects (Fernandez‐de‐las‐Penas 2015; Fernandez‐de‐las‐Penas 2017a; Fernandez‐de‐las‐Penas 2017b; Gerritsen 2002; Hui 2005; Ismatullah 2013; Jarvik 2009; Ly Pen 2005; Ucan 2006; Zhang 2019).

7. Need for further surgery or secondary surgery during follow‐up

Eight studies reported on the need for further surgery (i.e. primary carpal tunnel release after non‐surgical and reoperation after surgery for any reason) (Fernandez‐de‐las‐Penas 2015; Fernandez‐de‐las‐Penas 2017a; Fernandez‐de‐las‐Penas 2017b; Garland 1964; Gerritsen 2002; Jarvik 2009; Ly Pen 2005; Zhang 2019).

Unit of analysis

The outcomes were analysed and reported at:

participant level in eight studies where each participant contributed one wrist to the study (regardless of whether there were unilateral or bilateral cases) (Fernandez‐de‐las‐Penas 2015; Fernandez‐de‐las‐Penas 2017a; Fernandez‐de‐las‐Penas 2017b; Garland 1964; Gerritsen 2002; Ismatullah 2013; Jarvik 2009; Zhang 2019);
participant level in five studies where it was unclear whether any participant had bilateral CTS and how the study dealt with bilateral CTS if present (Awan 2015; Eltabl 2020; Hui 2005; Jafari 2018; Ucan 2006);
wrist level in Ly Pen 2005 (some or all participants had bilateral CTS, where both wrists contributed to the analysis and randomisation occurred at the wrist level, and treatment assignments were made for individual wrists).

Funding

Six studies reported receiving financial support through various sources (Fernandez‐de‐las‐Penas 2015; Fernandez‐de‐las‐Penas 2017a; Fernandez‐de‐las‐Penas 2017b; Gerritsen 2002; Jarvik 2009; Zhang 2019). Eight studies did not report information related to the funding (Awan 2015; Eltabl 2020; Garland 1964; Hui 2005; Ismatullah 2013; Jafari 2018; Ly Pen 2005; Ucan 2006).

Excluded studies

In total, we excluded 34 studies after review of the full publications. Reasons for exclusion of studies are given in the Characteristics of excluded studies table. The most common reasons for exclusion were that the study:

was not an RCT;
did not compare surgical treatment to non‐surgical treatment (both groups had surgery or neither had surgery).

Risk of bias in included studies

We assessed the study outcome risk of bias using RoB 2 for all 14 included studies (totalling 72 items; 14 studies and seven outcomes in the short term and long term) (Sterne 2019).

The RoB 2 judgements along with the support for judgements for each individual outcome are presented in the Risk of bias (tables) and visualised along with the forest plots (Data and analyses). An Excel file with full risk of bias assessment is available on reasonable request from the review authors.

The overall risk of bias was high for all the outcomes, mainly due to lack of blinding of the participants in all the included studies.

Most studies used adequate methods for randomisation and allocation concealment and thus received a low risk of bias in the domain 1 'Randomisation process'. However, five studies raised some concerns because the reports did not describe the randomisation or the allocation concealment procedure sufficiently to judge whether they were adequate (Awan 2015; Eltabl 2020; Garland 1964; Ismatullah 2013; Ucan 2006).

Risk of bias in the domain 2 'Deviations from the intended interventions' was either low or raised some concerns. We had some concerns in some studies where the authors reported per‐protocol analysis instead of intention‐to‐treat analysis. Also, we rated the risk as having 'some concerns' for the studies Gerritsen 2002 and Jarvik 2009 for all the long‐term outcomes, because in these studies up to 40% of the participants in the non‐surgical treatment arms had surgery before the long‐term measurement. Other co‐interventions were also imbalanced between the study arms. However, there was no specific reason to believe these deviations happened due to the trial context, and both these studies reported using the intention‐to‐treat principle, which was of our primary interest.

Risk of bias in domain 3 'Missing outcome data' was generally low, and we had some concerns with some outcomes in five studies (Gerritsen 2002; Jafari 2018; Jarvik 2009; Ly Pen 2005; Zhang 2019). We rated the risk of bias as high in two studies (all outcomes in Ucan 2006 and some outcomes in Ly Pen 2005). In Ly Pen 2005, the loss to follow‐up was over 30% ‐ large and not balanced between the groups for the long‐term outcomes. In Ucan 2006, the authors reported outcomes for 23 participants in the non‐surgical group and for 11 participants in the surgery group, and this disproportionate loss to follow‐up was unlikely to be at random.

Risk of bias in domain 5 'Selection of the reported result' was low for studies reporting outcomes in accordance with the prospectively reported protocol or clinical trial registry record (Fernandez‐de‐las‐Penas 2015; Fernandez‐de‐las‐Penas 2017a; Fernandez‐de‐las‐Penas 2017b; Gerritsen 2002; Jarvik 2009). We had some concerns with the remaining studies for most or all of the outcomes as the study protocols were not published, or studies were not pre‐registered (Awan 2015; Eltabl 2020; Garland 1964; Hui 2005; Ismatullah 2013; Jafari 2018; Ly Pen 2005; Ucan 2006; Zhang 2019). Furthermore, we had some concerns about some outcomes that were not reported fully in accordance with the prospectively reported protocol or clinical trial registry record (Fernandez‐de‐las‐Penas 2015; Fernandez‐de‐las‐Penas 2017a; Fernandez‐de‐las‐Penas 2017b; Jarvik 2009). The risk of bias was deemed high in Eltabl 2020 for the symptoms in the long term because the authors reported the standard deviation of BCTQ as SD = 0.0 for surgery and PRP groups, which seems implausible (n = 30 in both groups). In Gerritsen 2002, the authors had planned to measure health‐related quality of life both by SF‐36 and EuroQol, but only EuroQol was reported (the reason for not providing SF‐36 results was not reported).

Effects of interventions

See: Table 1; Table 2

1. Surgery versus splint

Three studies compared surgery with splint (Garland 1964 ‐ plaster‐of‐paris splint for 1 month; Gerritsen 2002 ‐ custom‐made splint or a prefabricated splint at night for at least 6 weeks; Ucan 2006 ‐ standard cotton‐polyester splint at night and in the daytime whenever possible for 3 months).

Short‐term outcomes were reported at 3 months (Gerritsen 2002; Ucan 2006). Long‐term outcomes were reported at 6 months (Ucan 2006; Gerritsen 2002), 12 months (Garland 1964; Gerritsen 2002), and 18 months (Gerritsen 2002).

Primary outcome

1) Clinical improvement

Three studies provided data for this outcome. Clinical improvement was measured as follows.

General improvement on a 6‐point ordinal transition scale, with success defined as 'completely recovered' or 'much improved' (Gerritsen 2002).
Number of patients who were 'completely satisfied' and 'almost satisfied' according to a 5‐point satisfaction scale (Ucan 2006).
Proportion of participants with complete symptom relief; unclear how measured (Garland 1964).

Short‐term improvement

We downgraded the certainty of evidence to very low (once for the risk of bias and twice for serious imprecision, as the confidence intervals (CIs) included values compatible with large benefits for either treatment), indicating that we are uncertain if surgery or splint results in a higher rate of clinical improvement in the short term.

In the surgery group, 67 of 89 participants (75%) improved, while in the splinting group, 62 of 109 participants (57%) improved, corresponding to a risk ratio (RR) of 1.06 (95% confidence interval (CI) 0.48 to 2.34; I² = 79%; 2 studies, 198 participants) (Analysis 1.1; Table 3).

1.1 — Comparison 1: Surgery versus splint, Outcome 1: Clinical improvement at short‐term follow‐up: 3 months or less

In the sensitivity analysis (excluding studies due to 'some concerns' or 'high' risk of bias in the randomisation process), 62 of 78 participants (80%) improved with surgery, while 46 of 86 participants (54%) improved in the splinting group, corresponding to a RR of 1.49 (95% CI 1.18 to 1.86; 1 study, 164 participants).

Long‐term improvement

Moderate‐certainty evidence (downgraded due to risk of bias) indicated that surgery probably results in a higher rate of clinical improvement. We did not downgrade for inconsistency (heterogeneity I² = 79%), as it was largely driven by one study with a high rate of participants switching to surgery from the splinting group (Table 6) (Gerritsen 2002). This study found a smaller benefit for surgery compared to other studies in this analysis, but it could be due to a weakened effect, as the effect could have been larger if participants in the splinting group got only their allocated treatment. The dilution of treatment effect was supported by the 'as‐treated' analysis in this study, which showed a larger benefit for surgery.

4. Rates of switching from one study intervention to the other, and additional interventions.

Study	Comparison	Total sample size	Switched from non‐surgical to surgical (%)	Switched from surgical to non‐surgical (%)	Additional interventions in non‐surgical group	Additional interventions in surgical group
Awan 2015	Surgery versus corticosteroid injection	116	0 of 58 (0%) at 1 month	0 of 58 (0%) at 1 month	Not reported	Not reported
Eltabl 2020	1) Surgery versus PRP injection 2) Surgery versus unspecified medical treatment and hand support	90	0 of 30 (0%) in the PRP injection group at 6 months 0 of 30 (0%) in the unspecified medical treatment group at 6 months	0 of 30 (0%) at 6 months	Not reported	Not reported
Fernandez‐de‐las‐Penas 2015	Surgery versus manual therapy	120	0 of 60 (0%) at 1 month 0 of 60 (0%) at 3 months 0 of 60 (0%) at 6 months 3 of 60 (5%) at 12 months 9 of 60 (15%) at 4 years	0 of 60 (0%) at 1 month 0 of 60 (0%) at 3 months 0 of 60 (0%) at 6 months 0 of 60 (0%) at 12 months 0 of 60 (0%) at 4 years	Sporadic use of NSAIDs	Educational session for performing tendon/nerve gliding exercises at home (same as in the manual therapy group) 4 of 60 (7%) surgery for the contralateral hand at 12 months 8 of 60 (13%) had resurgery at 4 years, but unclear how many for the study hand and how many for the contralateral hand Sporadic use of NSAID
Fernandez‐de‐las‐Penas 2017a	Surgery versus manual therapy	100	0 of 50 (0%) at 1 month 0 of 50 (0%) at 3 months 0 of 50 (0%) at 6 months 1 of 50 (2%) at 12 months	0 of 50 (0%) at 1 month 0 of 50 (0%) at 3 months 0 of 50 (0%) at 6 months 0 of 50 (0%) at 12 months	2 of 50 (4%) had steroid injection at 6 months Use of NSAIDs sporadically	Educational session for performing tendon/nerve gliding exercises at home (same as in the manual therapy group) 2 of 50 (4%) had repeated surgery in the hand Use of NSAIDs sporadically
Fernandez‐de‐las‐Penas 2017b	Surgery versus manual therapy	100	0 of 50 (0%) at 1 month 0 of 50 (0%) at 3 months 1 of 50 (2%) at 6 months 3 of 50 (6%) at 12 months	0 of 50 (0%) at 1 month 0 of 50 (0%) at 3 months 0 of 50 (0%) at 6 months 0 of 50 (0%) at 12 months	Occasional use of NSAIDs	Educational session for performing cervical spine exercises at home (same as in the manual therapy group) 3 of 50 (5%) had steroid injection at 12 months Occasional use of NSAID
Garland 1964	Surgery versus splint	22	0 of 11 (0%); 1 participant excluded, but not clear if received any other treatment	0 of 11 (0%); 1 dropout – did not accept surgery, but not clear if received any other treatment	Not reported	Not reported
Gerritsen 2002	Surgery versus splint	176	Overall 35 of 89 (39%) Of participants included in analysis: 6 of 86 (7%) at 3 months 26 of 84 (31%) at 6 months 32 of 83 (39%) at 12 months 32 of 79 (41%) at 18 months	Overall 1 of 87 (1%) at 3 months had received splint Of participants included in analysis: 6 of 73 (8%) at 12 months had not undergone surgery, but not clear what kind of other treatments/no treatment they received	During the 6‐week intervention period: 1 of 89 physiotherapy 1 of 89 Mensendieck (exercise) therapy 11 of 89 (12%) had pain medication Additional treatment after the 6‐week intervention period: 52 of 89 (58%) had one or more additional treatment options: 29 of 89 had pain medication 4 of 89 had physiotherapy 1 of 89 had manual therapy 2 of 89 had Mensendieck (exercise) therapy 1 of 89 had occupational therapy 5 of 89 had one or two local steroid injections 35 of 89 (39%) had surgery	From those allocated to surgery but not receiving it: 2 of 87 had physiotherapy 2 of 87 had pain medication 1 of 87 (1%) had splint Additional treatment after the surgery: 26 of 87 (30%) had one or more additional treatment options after surgery: 25 of 87 had pain medication to relieve pain caused by operation 2 of 87 had physiotherapy 1 of 87 had occupational therapy 1 of 87 had steroid injection and surgery to relieve pain caused by complex regional pain syndrome
Hui 2005	Surgery versus corticosteroid injection	50	0 of 25 (0%) at 6 0 of 25 (0%) at 20 weeks	0 of /25 (0%) at 6 0 of 25 (0%) at 20 weeks	Not reported	Not reported
Ismatullah 2013	Surgery versus corticosteroid injection	40	0 of 20 (0%) at 12 weeks	0 of 20 (0%) at 12 weeks	Not reported	Not reported
Jafari 2018	Surgery versus corticosteroid injection	82	0 of 41 (0%) at 6 months	0 of 41 (0%) at 6 months	Not reported	Not reported
Jarvik 2009	Surgery versus multimodal non‐operative treatment	116	2 of 59 (3%) at 3 months 7 of 59 (12%) at 6 months 13 of 59 (22%) at 9 months 23 of 59 (39%) at 12 months	15 of 57 (26%) at 3 months had not undergone surgery 13 of 57 (23%) at 12 months had not undergone surgery: 2 of 57 (3.5%) continued splinting 8 of 57 (14%) had hand exercises 2 of 57 (3.5%) workplace modifications 5 of 57 (9%) used NSAIDs	Additional non‐operative treatments during the first 6 months: 17 of 59 (29%) had opioid medications 17 of 59 (29%) had vitamin B6 1 of 59 (2%) had corticosteroid injections	Non‐operative treatments during the first 6 months: 43 of 57 (75%) had hand and wrist exercises 5 of 57 (9%) had wrist splints 24 of 57 (42%) had NSAIDs 20 of 57 (35%) had opioid medications 8 of 57 (14%) had vitamin B6 6 of 57 (11%) had home or workplace modifications 1 of 57 (2%) had corticosteroid injections 1 of 57 (2%) had at least one therapeutic ultrasound treatment of the wrist
Ly Pen 2005	Surgery versus corticosteroid injection	101 patients (163 wrists)	0 of 83 (0%) 1 of 83 (1%) at 3 months were in need of surgery 12 of 83 (14.5%) at 12 months were in need of surgery 26 of 83 (31%) at 24 months were in need of surgery	0 of 80 (0%) 2 of 80 (2.5%) at 3 months were in need of another surgery 7 of 80 (9%) at 12 months were in need of another surgery 9 of 80 (11%) at 24 months were in need of another surgery	Not reported	Not reported
Ucan 2006	1) Surgery versus splint 2) Surgery versus splint and corticosteroid injection	57 (available for follow‐up)	0 of 23 (0%) at 3 months 0 of /23 (0%) at 6 months Data provided for only those who were available for follow‐up	0 of 11 (0%) at 3 months 0 of /11 (0%) at 6 months Data provided for only those who were available for follow‐up	Not reported	Not reported
Zhang 2019	Surgery together with corticosteroid injection versus corticosteroid injection alone	51	0 of 26 (0%) at 12 weeks	0 of 25 (0%) at 12 weeks	Not reported	Not reported

Open in a new tab

Abbreviations: NSAID: nonsteroidal anti‐inflammatory drug; PRP: platelet‐rich plasma.

At long‐term follow‐up, 87 of 94 participants (93%) improved with surgery compared to 70 of 116 participants (60%) with splinting, corresponding to a RR of 2.10 (95% CI 1.04 to 4.24; I² = 79%; 3 studies, 210 participants) (Analysis 1.2; Table 1). This equates to a number needed to treat for an additional beneficial outcome (NNTB) of 2 (95% CI 2 to 9). The time point used for analysis was 6 months for Ucan 2006, and 12 months for Garland 1964 and Gerritsen 2002.

1.2 — Comparison 1: Surgery versus splint, Outcome 2: Clinical improvement at long‐term follow‐up: over 3 months

In a sensitivity analysis (excluding the studies due to 'some concerns' or 'high' risk of bias in the randomisation process), the benefit of surgery attenuated, but the only remaining study was Gerritsen 2002 which, as stated above, had a high rate of participants switching from splinting to surgery. In this study, 67 of 73 participants (92%) improved with surgery, while 60 of 83 participants (72%) improved in the splinting group, corresponding to a RR of 1.27 (95% CI from 1.09 to 1.47; 1 study, 156 participants).

Secondary outcomes

1) Symptoms

Two studies reported symptoms (measured by the BCTQ Symptom Severity Scale from 1 to 5 points, higher is worse, minimal clinically important difference (MCID) value = 1 point) (Gerritsen 2002; Ucan 2006).

At short‐term follow‐up, low‐certainty evidence (downgraded for risk of bias and inconsistency) indicates that the effect on symptom relief may not differ between surgery and splint. The mean symptom severity score was 1.39 points with splint and 0.02 points worse with surgery (95% CI from 0.83 points better to 0.87 points worse; I² = 93%; 2 studies, 198 participants) (Analysis 1.3; Table 3).

1.3 — Comparison 1: Surgery versus splint, Outcome 3: Symptoms at short‐term follow‐up: 3 months or less

At long‐term follow‐up, moderate‐certainty evidence (downgraded for risk of bias) indicates that there is probably no clinically relevant difference in symptom relief between surgery and splint. The mean symptom severity score was 1.54 with splint and 0.26 points better with surgery (95% CI from 0.52 better to 0.00 better or worse; I² = 60%; 2 studies, 195 participants) (Analysis 1.4; Table 1). The time point used for analysis was 6 months from both studies (Gerritsen 2002; Ucan 2006).

1.4 — Comparison 1: Surgery versus splint, Outcome 4: Symptoms at long‐term follow‐up: over 3 months

2) Function

Two studies reported function (measured by BCTQ Functional Status Scale from 1 to 5 points, higher is worse, MCID value = 0.7 points) (Gerritsen 2002; Ucan 2006).

At short‐term follow‐up, moderate‐certainty evidence (downgraded once for risk of bias) indicates that hand function is probably comparable between surgery and splinting. The mean function was 1.60 with splint and 0.0 points different with surgery (95% CI from 0.44 better to 0.44 worse; I² = 73%; 2 studies, 198 participants) (Analysis 1.5; Table 3).

1.5 — Comparison 1: Surgery versus splint, Outcome 5: Function at short‐term follow‐up: 3 months or less

At long‐term follow‐up, moderate‐certainty evidence (downgraded once for risk of bias) indicates that there is probably no clinically relevant difference in function between surgery and splinting. At long‐term follow‐up, the mean function was 1.75 with splint and 0.36 points better with surgery (95% CI from 0.62 better to 0.09 better; I² = 57%; 2 studies, 195 participants) (Analysis 1.6; Table 1). The time point used for analysis was 6 months from both studies (Gerritsen 2002; Ucan 2006).

1.6 — Comparison 1: Surgery versus splint, Outcome 6: Function at long‐term follow‐up: over 3 months

3) Pain

Only Gerritsen 2002 measured pain, but the study did not report numerical data for pain, stating that the findings were similar to those for paraesthesia. Thus, we have no estimates for pain (very low certainty).

4) Health‐related quality of life

At short‐term follow‐up, this outcome was not reported in any of the included studies.

One study measured this outcome at long‐term follow‐up (using EQ‐5D, scale from 0 to 1 points, higher is better, MCID value = 0.074) (Gerritsen 2002).

At long‐term follow‐up (1 year), low‐certainty evidence (downgraded for risk of bias and imprecision) indicates that surgery may not provide better health‐related quality of life compared with splinting, but the confidence intervals overlapped with clinically important benefit for surgery. The mean EQ‐5D score was 0.81 in the splinting group and 0.04 points better with surgery (95% CI from 0.00 worse to 0.08 better; 1 study, 167 participants) (Analysis 1.7; Table 1). The time point used for analysis was 12 months (Gerritsen 2002).

1.7 — Comparison 1: Surgery versus splint, Outcome 7: Health‐related quality of life at long‐term follow‐up: over 3 months

5) Adverse effects

In the surgery group, adverse effects were reported for 60 of 98 participants (61%) (painful or hypertrophic scar, n = 53; stiffness of the wrist, hand, or fingers, n = 24; skin irritation, n = 19; wound haematoma, n = 10; wound infection, n = 5; severe pillar pain, n = 2; complex regional pain syndrome, n = 2; scar tenderness, n = 1) and in the splinting group, for 46 of 112 participants (41%) (painful or hypertrophic scar, n = 20; stiffness of the wrist, hand, or fingers, n = 31; skin irritation, n = 8; wound haematoma, n = 1; wound infection, n = 2; discomfort caused by pressure of the splint, n = 6; swelling of the wrist, hand, or fingers, n = 4; some of these adverse effects were likely caused by surgery as up to 40% of the participants in the splinting group had switched the treatment to surgery before the outcome was measured) (Gerritsen 2002; Ucan 2006). This corresponded to a RR of 2.11 (95% CI 0.37 to 12.12; I² = 47%; 2 studies, 210 participants) (Analysis 1.8; Table 1). We rated the certainty of evidence as very low (downgraded once for the study risk of bias and twice for very serious imprecision). The time point used for analysis was 6 months for Ucan 2006 and 18 months for Gerritsen 2002.

1.8 — Comparison 1: Surgery versus splint, Outcome 8: Adverse effects

6) Need for surgery or secondary surgery during follow‐up

At short‐term follow‐up (3 months), Gerritsen 2002 reported that 0 of 78 participants (0%) had repeated surgery in the surgery group while 6 of 86 participants (7%) were referred to surgery in the splinting group. This corresponds to a RR of 0.08 (95% CI 0.00 to 1.48; 1 study, 164 participants) (Analysis 1.9; Table 3). We downgraded the certainty of evidence to very low (once for risk of bias and twice for very serious imprecision), thus we are uncertain about the difference in risk of referral to surgery (or reoperation) at short‐term follow‐up.

1.9 — Comparison 1: Surgery versus splint, Outcome 9: Need for surgery or secondary surgery at short‐term follow‐up: 3 months or less

At long‐term follow‐up, two studies reported this outcome (Garland 1964; Gerritsen 2002). In the splinting group, 41 of 93 participants (44%) were referred to surgery, and in the surgery group, 0 of 83 participants (0%) had repeated surgery. This corresponds to a RR of 0.03 (95% CI 0.00 to 0.21; I² = 0%; 2 studies, 176 participants) (Analysis 1.10; Table 1). This equates to a NNTB of 2 (95% CI 1 to 9). We downgraded the certainty of evidence to moderate (once for the study risk of bias). The time point used for analysis was 12 months for both studies (Garland 1964; Gerritsen 2002).

1.10 — Comparison 1: Surgery versus splint, Outcome 10: Need for surgery or secondary surgery at long‐term follow‐up: over 3 months

Gerritsen 2002 reported that one patient in the surgery group developed complex regional pain syndrome and had repeat surgery to relieve the resulting pain; however, the author did not mention when it happened, therefore we did not include it in the analysis.

2. Surgery versus corticosteroid injection

Five studies compared surgery with corticosteroid injection (Awan 2015 ‐ 20 mg/mL injection of methyl prednisolone; Hui 2005 ‐ 15 mg of methylprednisolone acetate; Ismatullah 2013 ‐ 40 mg of methyl prednisolone; Jafari 2018 ‐ 40 mg/mL of triamcinolone acetonide; Ly Pen 2005 ‐ 20 mg/mL of paramethasone acetonide, injected once or twice).

At short‐term follow‐up, the outcomes were collected at 1 month by Awan 2015, at 6 weeks by Hui 2005, at 2 weeks, 4 weeks and 12 weeks by Ismatullah 2013 (we used data from 12 weeks in the analysis) and 3 months by Ly Pen 2005.

At long‐term follow‐up, the outcomes were reported at 20 weeks by Hui 2005, at 6 months by Jafari 2018 and at 6 months, 12 and 24 months by Ly Pen 2005 (we used data from 12 months in the analysis, but provided data for 24 months regarding the need for surgery too).

Primary outcome

1) Clinical improvement

Four studies provided data for this outcome. Clinical improvement was measured as follows.

Improvement in at least two grade scores on VAS (Awan 2015).
Number of participants who are 'completely satisfied' according to a 5‐point satisfaction scale (Jafari 2018).
70% reduction in the VAS score for nocturnal paraesthesias for the unilateral patients plus the most symptomatic wrist from bilateral patients (Ly Pen 2005).
Improvement by at least 50% in a global symptom score (Hui 2005).

Short‐term follow‐up

Moderate‐certainty evidence (downgraded once for risk of bias) indicates that surgery probably results in lower rates of clinical improvement compared with corticosteroid injection. In the surgery group, 54 of 94 participants (57%) improved compared to 69 of 91 participants (78%) in the corticosteroid injection group, corresponding to a RR of 0.75 (95% CI 0.60 to 0.92; I² = 5%; 2 studies, 185 participants) (Analysis 2.1; Table 4). This corresponds to a number needed to treat for an additional harmful outcome (NNTH) of 5 with surgery (95% CI 3 to 34). In the sensitivity analysis (excluding the studies with 'some concerns' about the risk of bias or high risk of bias in the randomisation process), 21 of 36 participants (58%) improved with surgery, while 29 of 33 participants (88%) improved in the corticosteroid injection group, corresponding to a RR of 0.66 (95% CI 0.49 to 0.90; 1 study, 69 participants).

2.1 — Comparison 2: Surgery versus corticosteroid injection, Outcome 1: Clinical improvement at short‐term follow‐up: 3 months or less

Long‐term follow‐up

At long‐term follow‐up, we are uncertain if the clinical improvement differs between surgery and corticosteroid injection because the certainty of evidence was very low (downgraded for risk of bias, unexplained inconsistency and imprecision). In surgery, 70 of 96 participants (73%) improved versus 56 of 91 (62%) in the corticosteroid injection group, corresponding to a RR of 1.23 (95% CI 0.73 to 2.06; I² = 83%; 3 studies, 187 participants) (Analysis 2.2; Table 2). Sensitivity analysis (excluding studies with 'some concerns' in the risk of bias or 'high' risk of bias in the randomisation process) did not change the result.

2.2 — Comparison 2: Surgery versus corticosteroid injection, Outcome 2: Clinical improvement at long‐term follow‐up: over 3 months

Secondary outcomes

1) Symptoms

Three studies provided data on symptom severity. Hui 2005 and Ismatullah 2013 measured this outcome by a global symptom score, scale from 0 to 50 points, higher is worse (MCID value unknown; we used 10%, i.e. 5 points, as a cut‐off). Jafari 2018 measured symptoms on the BCTQ Symptom Severity Scale from 1 to 5 points, higher is worse, MCID value = 1 point.

At short‐term follow‐up, the mean symptom severity score (measured by a global symptom score, 0 to 50 points, higher is worse) was 16.2 points with corticosteroid injection and 11.3 points better with surgery (95% CI from 21.76 better to 0.89 better; I² = 92%; 2 studies, 90 participants) (Analysis 2.3; Table 4). We rated the certainty of evidence as low (downgraded once for risk of bias and once for imprecision).

2.3 — Comparison 2: Surgery versus corticosteroid injection, Outcome 3: Symptoms at short‐term follow‐up: 3 months or less

We considered downgrading a third step for inconsistency (I² = 92%) at short‐term follow‐up, but the inconsistency seemed to relate to different time points for measurements (6 weeks in Hui 2005 versus 12 weeks in Ismatullah 2013), because using data from 6 weeks for both studies decreased the statistical heterogeneity. With 6‐week data, the mean symptom severity score was 10.8 with corticosteroid injection and 4.1 points better with surgery (95% CI from 7.5 better to 0.8 better; I² = 29%; 2 studies, 90 participants).

At long‐term follow‐up, we are uncertain about the effect on symptom relief between the two treatments because the certainty of evidence was very low (downgraded for risk of bias, inconsistency and imprecision). The standardised mean difference (SMD) was ‐0.60 (95% CI from ‐1.88 to 0.69; I² = 91%; 2 studies, 118 participants) (Analysis 2.4; Table 2). This translates to a 0.4‐point better BCTQ symptom severity score (95% CI from 1.3 better to 0.5 worse) with surgery compared with corticosteroid injection (using a baseline standard deviation of 0.70 from Jarvik 2009).

2.4 — Comparison 2: Surgery versus corticosteroid injection, Outcome 4: Symptoms at long‐term follow‐up: over 3 months

2) Function

Two studies reported functional status scores (measured in Jafari 2018 by the BCTQ, scale from 1 to 5 points, higher is worse, MCID value = 0.7 points; and, in Ly Pen 2005, by VAS functional impairment, scale from 0 to 100 points, higher is worse, MCID value not defined; we used 15 points).

At short‐term follow‐up, the functional status score was 6 points with corticosteroid injection and 11 points worse with surgery (95% CI from 4.8 worse to 17.2 worse; 1 study, 147 participants) (Analysis 2.5; Table 4). We rated the certainty of evidence as low (downgraded once for the risk of bias and once for imprecision).

2.5 — Comparison 2: Surgery versus corticosteroid injection, Outcome 5: Function at short‐term follow‐up: 3 months or less

At long‐term follow‐up, the certainty of evidence was moderate (downgraded once for risk of bias). The SMD was ‐0.12 (95% CI ‐0.80 to 0.56; I² = 81%; 2 studies, 191 participants) (Analysis 2.6; Table 2). This translates to a 0.10‐point better BCTQ functional status score with surgery (95% CI from 0.66 better to 0.46 worse) compared with corticosteroid injection, 95% CIs excluding a clinically relevant difference (SD of 0.82 from Jarvik 2009).

2.6 — Comparison 2: Surgery versus corticosteroid injection, Outcome 6: Function at long‐term follow‐up: over 3 months

3) Pain

One study reported diurnal pain (measured by VAS from 0 to 100 points, higher is worse, MCID value not defined; we used 15 points) (Ly Pen 2005).

At short‐term follow‐up, the pain score was 6 points with corticosteroid injection and 9 points worse with surgery (95% CI from 2.79 worse to 15.21 worse; 1 study, 147 participants) (Analysis 2.7; Table 4). We rated the evidence as low certainty (downgraded once for the risk of bias and once for imprecision).

2.7 — Comparison 2: Surgery versus corticosteroid injection, Outcome 7: Pain at short‐term follow‐up: 3 months or less

At long‐term follow‐up, moderate‐certainty evidence (downgraded once for the risk of bias) indicated that the difference in pain is probably clinically unimportant. The mean pain score was 8 with corticosteroid injection and 6 points better with surgery (95% CI from 10.45 better to 1.55 better; 1 study, 123 participants) (Analysis 2.8; Table 2).

2.8 — Comparison 2: Surgery versus corticosteroid injection, Outcome 8: Pain at long‐term follow‐up: over 3 months

4) Health‐related quality of life

None of the studies in this comparison reported this outcome.

5) Adverse effects

Adverse effects were reported in 3 of 45 participants (7%) in the surgery group (wound haematoma, complex regional pain syndrome) and in 2 of 45 participants (4%) in the corticosteroid group (cellulitis), corresponding to a RR of 1.49 (95% CI 0.25 to 8.70; I² = 0%; 2 studies, 90 participants) (Analysis 2.9; Table 2). We rated the certainty of evidence as very low (downgraded once for risk of bias, and twice for very serious imprecision), thus we are uncertain if surgery is associated with more or fewer adverse effects compared with corticosteroid injection. One study reported that no adverse effects or complications occurred during the trial, with 0 of 83 participants in the corticosteroid group and 0 of 80 participants (0%) in the surgery group (Ly Pen 2005). Due to zero events, this study did not contribute to the meta‐analysis.

2.9 — Comparison 2: Surgery versus corticosteroid injection, Outcome 9: Adverse effects

6) Need for surgery or secondary surgery during follow‐up

At short‐term follow‐up, one study reported that 1 of 83 participants (1%) was "in need of surgery" in the corticosteroid injection group and 2 of 80 participants (2.5%) were "in need of repeated surgery" in the surgery group (Ly Pen 2005). This corresponds to a RR of 2.08 (95% CI 0.19 to 22.44; 1 study, 163 participants) (Analysis 2.10; Table 4). It is unclear if the participants actually underwent surgery. We rated the certainty of evidence as very low (downgraded for risk of bias and twice for very serious imprecision), thus we are uncertain whether surgery or corticosteroid injection results in a higher risk of surgery or reoperation at short‐term follow‐up.

2.10 — Comparison 2: Surgery versus corticosteroid injection, Outcome 10: Need for surgery or secondary surgery at short‐term follow‐up: 3 months or less

At long‐term follow‐up (12 months), one study reported that 12 of 83 participants (15%) were "in need of surgery" in the corticosteroid injection group and 7 of 80 participants (9%) were "in need of repeated surgery" in the surgery group (Ly Pen 2005). This corresponds to a RR of 0.61 (95% CI 0.25 to 1.46; 1 study, 163 participants) (Analysis 2.11; Table 2). It is unclear if the participants actually underwent surgery. We rated the certainty of evidence as very low (downgraded for risk of bias and twice for very serious imprecision), thus we are uncertain whether surgery or corticosteroid injection results in a higher risk of surgery or reoperation at long‐term (12 months) follow‐up. The same study also provided data for 24‐month follow‐up. At 24 months, 26 of 83 participants (31%) were "in need of surgery" in the corticosteroid injection group and 9 of 80 participants (11%) were "in need of repeated surgery" in the surgery group (Ly Pen 2005). This corresponds to a RR of 0.36 (95% CI 0.18 to 0.72; 1 study, 163 participants). Similar to 12‐month follow‐up, it is unclear if the participants actually underwent surgery.

2.11 — Comparison 2: Surgery versus corticosteroid injection, Outcome 11: Need for surgery or secondary surgery at long‐term follow‐up: over 3 months

3. Surgery versus splint + corticosteroid injection

One study with 34 participants provided data for this comparison (Ucan 2006). Short‐term outcomes were reported at 3 months and long‐term outcomes were reported at 6 months.

Primary outcome

1) Clinical improvement

We used the number of participants reporting to be 'completely satisfied' or 'almost satisfied' as improved.

Short‐term follow‐up

We are uncertain whether surgery or splint together with corticosteroid injection results in a higher improvement rate because the certainty of evidence was very low (downgraded twice due to serious risk of bias and once for imprecision, because the study was small with a large imbalance at follow‐up). Five of 11 participants (45%) improved in the surgery group, while 23 of 23 participants (100%) improved in the splint and corticosteroid injection group, corresponding to a RR of 0.47 (95% CI 0.25 to 0.87; 1 study, 34 participants) (Analysis 3.1). This corresponds to a NNTH of 2 with surgery (95% CI 2 to 4).

3.1 — Comparison 3: Surgery versus splint + steroid injection, Outcome 1: Clinical improvement at short‐term follow‐up: 3 months or less

Long‐term follow‐up

We are uncertain whether surgery or splint together with corticosteroid injection results in a higher improvement rate because the certainty of evidence was very low (downgraded twice for serious risk of bias and once for imprecision). In the surgery group, 10 of 11 participants (91%) improved compared to 19 of 23 participants (83%) in the splint and corticosteroid injection group, corresponding to a RR of 1.10 (95% CI 0.84 to 1.43; 1 study, 34 participants) (Analysis 3.2).

3.2 — Comparison 3: Surgery versus splint + steroid injection, Outcome 2: Clinical improvement at long‐term follow‐up: over 3 months

Secondary outcomes

1) Symptoms

At short‐term follow‐up, the mean symptom severity score (measured by the BCTQ Symptom Severity Scale from 1 to 5 points, higher is worse; MCID = 1 point) was 1.41 with splint and corticosteroid injection, and 0.45 worse with surgery (95% CI from 0.07 worse to 0.83 worse; 1 study, 34 participants) (Analysis 3.3). The certainty of evidence was very low (downgraded twice for serious risk of bias and once for imprecision; only one study with 34 participants and a large imbalanced loss to follow‐up contributed to the analysis). Therefore, we are uncertain whether surgery or splint together with corticosteroid injection results in better symptom relief at short‐term follow‐up.

3.3 — Comparison 3: Surgery versus splint + steroid injection, Outcome 3: Symptoms at short‐term follow‐up: 3 months or less

At long‐term follow‐up, we are uncertain whether surgery or splint together with corticosteroid injection results in better symptom relief because the certainty of evidence was very low (downgraded twice for serious risk of bias and once for imprecision). The mean symptom severity score was 1.96 with splint and corticosteroid injection, and 0.55 better with surgery (95% CI from 0.87 better to 0.23 better; 1 study, 34 participants) (Analysis 3.4).

3.4 — Comparison 3: Surgery versus splint + steroid injection, Outcome 4: Symptoms at long‐term follow‐up: over 3 months

2) Function

At short‐term follow‐up, the mean function (measured by the BCTQ Functional Status Scale from 1 to 5 points, higher is worse; MCID = 0.7 points) was 1.32 with splint and corticosteroid injection, and 0.53 worse with surgery (95% CI from 0.13 worse to 0.93 worse; 1 study, 34 participants) (Analysis 3.5). The certainty of evidence was very low (downgraded twice for serious risk of bias and twice for imprecision). Therefore, we are uncertain whether surgery or splint together with corticosteroid injection results in better hand function at short‐term follow‐up.

3.5 — Comparison 3: Surgery versus splint + steroid injection, Outcome 5: Function at short‐term follow‐up: 3 months or less

At long‐term follow‐up, we are uncertain whether surgery or splint together with corticosteroid injection results in better hand function because the certainty of evidence was very low (downgraded twice for serious risk of bias and once for imprecision). The mean function was 1.69 with splint and corticosteroid injection, and 0.17 better (95% CI from 0.41 better to 0.07 worse; 1 study, 34 participants) with surgery (Analysis 3.6).

3.6 — Comparison 3: Surgery versus splint + steroid injection, Outcome 6: Function at long‐term follow‐up: over 3 months

3) Pain

The only study in this comparison did not report this outcome.

4) Health‐related quality of life

The only study in this comparison did not report this outcome.

5) Adverse effects

We are uncertain whether surgery or splint together with corticosteroid injection results in fewer adverse effects. Ucan 2006 reported adverse effects in 0 of 23 participants (0%) in the splint and corticosteroid injection group and 2 of 11 participants (18%) in the surgery group (complex regional pain syndrome (n = 1), scar tenderness (n = 1)). This corresponds to a RR of 10.00 (95% CI 0.52 to 192.25; 1 study, 34 participants) (Analysis 3.7). We graded the certainty of evidence as very low (downgraded twice for serious risk of bias and very serious imprecision).

3.7 — Comparison 3: Surgery versus splint + steroid injection, Outcome 7: Adverse effects

6) Need for surgery or secondary surgery during follow‐up

The only study in this comparison did not report this outcome.

4. Surgery versus PRP

One study with 60 participants provided data for this comparison at long‐term follow‐up (6 months) (Eltabl 2020).

Primary outcome

1) Clinical improvement

The study in this comparison did not report this outcome.

Secondary outcomes

1) Symptoms

This outcome was not reported at short‐term follow‐up.

At long‐term follow‐up, the mean BCTQ symptom severity score (scale from 1 to 5 points, higher is worse; MCID = 1 point) was 1 point in both groups, but the 95% CIs were not estimable because the SD was 0 in both groups (Analysis 4.1). Consequently, we performed a sensitivity analysis imputing SD from Jarvik 2009 (SD at baseline = 0.70 points). The MD was 0.0 (95% CI 0.35 better to 0.35 worse). We rated the certainty of evidence as low (downgraded once for risk of bias and imprecision, as there was only one small study).

4.1 — Comparison 4: Surgery versus PRP injection, Outcome 1: Symptoms at long‐term follow‐up: over 3 months

2) Function

This outcome was not reported at short‐term follow‐up.

At long‐term follow‐up, low‐certainty evidence (downgraded for risk of bias and imprecision) indicated that surgery may not improve hand function compared with PRP injection. The mean BCTQ functional status score (scale from 1 to 5 points, higher is worse; MCID = 0.7 points) was 1.4 with PRP and 0.33 better with surgery (95% CI from 0.35 better to 0.31 better; 1 study, 60 participants) (Analysis 4.2). Since the reported SD values deviated considerably from the other studies, we also performed sensitivity analysis imputing a SD from Jarvik 2009 (SD at baseline = 0.82 points). Using an imputed SD, the MD was 0.33 better with surgery (95% CI 0.74 better to 0.08 worse).

4.2 — Comparison 4: Surgery versus PRP injection, Outcome 2: Function at long‐term follow‐up: over 3 months

3) Pain

This outcome was not reported at short‐term follow‐up.

At long‐term follow‐up, moderate‐certainty evidence (downgraded once for risk of bias) indicates that surgery probably provides clinically unimportant benefit for pain compared with PRP at long‐term follow‐up. The mean pain score (VAS, scale from 0 to 10 points, higher is worse, MCID value = 1.5 points) was 1.97 with PRP and 0.43 better with surgery (95% CI from 0.03 better to 0.83 better; 1 study, 60 participants) (Analysis 4.3).

4.3 — Comparison 4: Surgery versus PRP injection, Outcome 3: Pain at long‐term follow‐up: over 3 months

4) Health‐related quality of life

The study in this comparison did not report this outcome.

5) Adverse effects

The study in this comparison did not report this outcome.

6) Need for surgery or secondary surgery during follow‐up

The study in this comparison did not report this outcome.

5. Surgery versus manual therapy

Three studies compared surgery with manual therapy (Fernandez‐de‐las‐Penas 2015; Fernandez‐de‐las‐Penas 2017a; Fernandez‐de‐las‐Penas 2017b). The authors confirmed that there were no overlapping participants, although recruitment periods overlapped in the two later trials (Fernandez‐de‐las‐Penas 2017a; Fernandez‐de‐las‐Penas 2017b). We used the outcomes reported at 3 months for short‐term analysis, and at 12 months for long‐term analysis.

Primary outcome

1) Clinical improvement

This outcome was not reported at short‐term follow‐up.

Long‐term follow‐up

Moderate‐certainty evidence (downgraded once for the risk of bias) indicates that surgery probably results in a higher rate of clinical improvement compared with manual therapy. In the surgery group, 52 of 56 participants (93%) improved compared to 39 of 55 participants (71%) in the manual therapy group, corresponding to a RR of 1.31 (95% CI 1.09 to 1.57; 1 study, 111 participants) (Analysis 5.1). This equates to a NNTB of 5 (95% CI 3 to 13).

5.1 — Comparison 5: Surgery versus manual therapy, Outcome 1: Clinical improvement at long‐term follow‐up: over 3 months

In addition, Fernandez‐de‐las‐Penas 2015 defined a successful outcome (post hoc) as: 1) reduction ≥ 0.7 points or 30% improvement from baseline in the BCTQ Functional Status Scale, or 2) reduction ≥ 0.7 points or 30% improvement from baseline in the BCTQ Symptom Severity Scale, or 3) decrease ≥ 2 points in the intensity of hand pain. Using this definition, 31 of 56 participants (55%) improved in the surgery group versus 32 of 55 participants (58%) in the manual therapy group, corresponding to a RR of 0.95 (95% CI 0.69 to 1.32; 1 study, 101 participants).

Secondary outcomes

1) Symptoms

At short‐term follow‐up, moderate‐certainty evidence (downgraded once for the risk of bias) indicates that symptom relief probably does not differ between surgery and manual therapy. The mean symptoms (measured by the BCTQ Symptom Severity Scale from 1 to 5 points, higher is worse, MCID value = 1 point) was 1.65 with manual therapy and 0.08 points better with surgery (95% CI from 0.28 better to 0.11 worse; I² = 41%; 2 studies, 220 participants) (Analysis 5.2).

5.2 — Comparison 5: Surgery versus manual therapy, Outcome 2: Symptoms at short‐term follow‐up: 3 months or less

At long‐term follow‐up, symptom relief probably does not differ between surgery and manual therapy (moderate‐certainty evidence, downgraded for risk of bias). The mean symptom score was 1.55 points with manual therapy and 0.09 points better with surgery (95% CI from 0.29 better to 0.10 worse; I² = 49%; 2 studies, 220 participants) (Analysis 5.3).

5.3 — Comparison 5: Surgery versus manual therapy, Outcome 3: Symptoms at long‐term follow‐up: over 3 months

2) Function

At short‐term follow‐up, moderate‐certainty evidence (downgraded once for risk of bias) indicates that surgery probably results in slightly worse hand function compared with manual therapy, but the difference is likely to be clinically unimportant. The mean function (measured by the BCTQ Functional Status Scale from 1 to 5 points, higher is worse, MCID value = 0.7 points) was 1.55 with manual therapy and 0.22 points worse with surgery (95% CI from 0.03 worse to 0.41 worse; I² = 20%; 2 studies, 220 participants) (Analysis 5.4).

5.4 — Comparison 5: Surgery versus manual therapy, Outcome 4: Function at short‐term follow‐up: 3 months or less

At long‐term follow‐up, surgery probably does not improve hand function compared with manual therapy (moderate‐certainty evidence, downgraded for risk of bias). The mean functional status score was 1.55 with manual therapy and 0.04 points better with surgery (95% CI 0.19 better to 0.11 worse; I² = 0%; 2 studies, 220 participants) (Analysis 5.5).

5.5 — Comparison 5: Surgery versus manual therapy, Outcome 5: Function at long‐term follow‐up: over 3 months

3) Pain

At short‐term follow‐up, low‐certainty evidence (downgraded for risk of bias and imprecision) indicates that pain may be worse with surgery, when compared with manual therapy. The magnitude of the difference was at the same level as the MCID value of 1.5 points on a VAS. The mean pain score (measured by 11‐point Numerical Rating Scale, 0 to 10, higher is worse) was 1.37 with manual therapy and 1.46 worse with surgery (95% CI from 0.96 worse to 1.96 worse; I² = 0%; 2 studies, 220 participants) (Analysis 5.6).

5.6 — Comparison 5: Surgery versus manual therapy, Outcome 6: Pain at short‐term follow‐up: 3 months or less

At long‐term follow‐up, moderate‐certainty evidence (downgraded once for risk of bias) indicates that there is probably no important difference in pain between surgery and manual therapy. The mean pain score was 1.29 with manual therapy and 0.05 worse with surgery (95% CI from 0.42 better to 0.53 worse; I² = 0%; 2 studies, 220 participants) (Analysis 5.7).

5.7 — Comparison 5: Surgery versus manual therapy, Outcome 7: Pain at long‐term follow‐up: over 3 months

4) Health‐related quality of life

One study measured this outcome (by EQ‐5D‐5L, scale from 0 to 1 points, higher is better, MCID value = 0.074) (Fernandez‐de‐las‐Penas 2015).

At short‐term follow‐up, moderate‐certainty evidence (downgraded for risk of bias) indicates that manual therapy probably provides clinically important benefits in health‐related quality of life compared with surgery. The mean EQ‐5D‐5L score was 0.78 in the manual therapy group and 0.10 points worse with surgery (95% CI from 0.12 worse to 0.08 worse; 1 study, 120 participants) (Analysis 5.8).

5.8 — Comparison 5: Surgery versus manual therapy, Outcome 8: Health‐related quality of life at short‐term follow‐up: 3 months or less

At long‐term follow‐up, moderate‐certainty evidence (downgraded for risk of bias) indicates that there is probably no important difference in health‐related quality of life between surgery and manual therapy. The mean EQ‐5D‐5L score was 0.93 in the manual therapy group and 0.02 points worse with surgery (95% CI from 0.04 worse to 0.00 better; 1 study, 118 participants) (Analysis 5.9).

5.9 — Comparison 5: Surgery versus manual therapy, Outcome 9: Health‐related quality of life at long‐term follow‐up: over 3 months

5) Adverse effects

Three studies reported on adverse effects, with 0 of 160 participants (0%) in the manual therapy group and 0 of 160 participants (0%) in the surgery group having adverse effects (Fernandez‐de‐las‐Penas 2015; Fernandez‐de‐las‐Penas 2017a; Fernandez‐de‐las‐Penas 2017b). Due to zero reported events, we could not estimate the risk (very low‐certainty evidence).

6) Need for surgery or secondary surgery during follow‐up

This outcome was not reported at short‐term follow‐up.

At long‐term follow‐up, 7 of 160 participants (4%) in the manual therapy group were referred for surgery and 2 of 160 participants (1.25%) in the surgery group had repeated surgery (Fernandez‐de‐las‐Penas 2015; Fernandez‐de‐las‐Penas 2017a; Fernandez‐de‐las‐Penas 2017b). This corresponds to a RR of 0.42 (95% CI 0.07 to 2.63; I² = 27%; 3 studies, 320 participants) (Analysis 5.10). The certainty of evidence was very low (downgraded for risk of bias and twice for very serious imprecision). We are thus uncertain about the effects on the need for surgery (or reoperation) between the treatments.

5.10 — Comparison 5: Surgery versus manual therapy, Outcome 10: Need for surgery or secondary surgery at long‐term follow‐up: over 3 months

6. Surgery versus multimodal non‐operative treatment

One study with 101 randomised participants provided data for this comparison at long‐term follow‐up (12 months) (Jarvik 2009). The control group received NSAIDs (ibuprofen 200 mg three times a day), six visits of formal hand therapy, exercises, splinting, and modification for work and activity, followed by ultrasound therapy if needed (consisting of up to 12 sessions (2 to 4 per week for up to 6 weeks) of focused ultrasound at 1 MHz, 1.0 W/cm², in pulsed mode 1:4, and 15 minutes each).

Primary outcome

1) Clinical improvement

Jarvik 2009 defined a successful outcome with three criteria or combination of them: 1) 0.50 points or more, or 30% improvement from baseline in CTSAQ function; 2) 0.50 points or more, or 30% improvement in CTSAQ symptom severity; and 3) a score of 0 or 1 on a 0 to 10 rating scale of hand or wrist pain interference with work or housework. We used the proportion of participants who fulfilled all these criteria.

This outcome was not reported at short‐term follow‐up.

Long‐term follow‐up

At long‐term follow‐up, low‐certainty evidence (downgraded once for the risk of bias and imprecision) indicates that clinical improvement may not differ between surgery and multimodal non‐operative treatment. In the surgery group, 22 of 49 participants (45%) improved, compared to 14 of 52 participants (27%) in the multimodal non‐operative treatment group, corresponding to a RR of 1.67 (95% CI 0.97 to 2.88; 1 study, 101 participants). The CIs excluded relevant benefit for multimodal non‐operative treatment but not for surgery (Analysis 6.1).

6.1 — Comparison 6: Surgery versus multimodal non‐operative treatment, Outcome 1: Clinical improvement at long‐term follow‐up: over 3 months

Secondary outcomes

1) Symptoms

Jarvik 2009 measured this outcome using the CTSAQ, which has the same scale as the BCTQ (11‐item Symptom Severity Scale from 1 to 5 points, higher is worse; we interpreted it using the same MCID value of 1 point as we used for the BCTQ).

This outcome was not reported at short‐term follow‐up.

At long‐term follow‐up, moderate‐certainty evidence (downgraded once for risk of bias) indicates that surgery probably provides clinically unimportant benefit compared with multimodal non‐operative treatment in symptom relief. The mean symptom severity score was 2.07 with multimodal non‐operative treatment and 0.33 better with surgery (95% CI from 0.65 better to 0.01 better; 1 study, 101 participants) (Analysis 6.2).

6.2 — Comparison 6: Surgery versus multimodal non‐operative treatment, Outcome 2: Symptoms at long‐term follow‐up: over 3 months

2) Function

Jarvik 2009 measured this outcome using the CTSAQ (9‐item Functional Status Scale from 1 to 5 points, higher is worse; we interpreted it using the same MCID value of 0.7 points as with the BCTQ Functional Status Scale).

This outcome was not reported at short‐term follow‐up.

Low‐certainty evidence (downgraded for risk of bias and imprecision) indicates that surgery may provide clinically unimportant benefit in function compared with multimodal non‐operative treatment at long‐term follow‐up. The mean function was 2.17 with multimodal non‐operative treatment and 0.43 better with surgery (95% CI from 0.77 better to 0.09 better; 1 study, 101 participants) (Analysis 6.3).

6.3 — Comparison 6: Surgery versus multimodal non‐operative treatment, Outcome 3: Function at long‐term follow‐up: over 3 months

3) Pain

This outcome was not reported at short‐term follow‐up.

At long‐term follow‐up, low‐certainty evidence (downgraded for risk of bias and imprecision) indicates that pain may not differ between surgery and multimodal non‐operative treatment. The mean pain (measured by VAS from 0 to 10 points, higher is worse; MCID value = 1.5 points) was 4.3 with multimodal non‐operative treatment and 0.80 better with surgery (95% CI from 2.03 better to 0.43 worse; 1 study, 101 participants) (Analysis 6.4).

6.4 — Comparison 6: Surgery versus multimodal non‐operative treatment, Outcome 4: Pain at long‐term follow‐up: over 3 months

4) Health‐related quality of life

At short‐term follow‐up, this outcome was not reported.

Jarvik 2009 measured this outcome with the SF‐36 Physical Component Score and Mental Component Score at long‐term follow‐up (scale from 0 to 50 points, higher is better, MCID value = 2.5 points).

At long‐term follow‐up, the health‐related quality of life may not differ between surgery and multimodal non‐operative treatment (low‐certainty evidence, downgraded for risk of bias and imprecision). The mean SF‐36 Physical Component Score was 37 in the multimodal non‐operative treatment group and 2 points better in the surgery group (95% CI from 3.10 worse to 7.10 better; 1 study, 101 participants). The mean SF‐36 Mental Component Score was 47 in the multimodal non‐operative treatment group and 2 points worse in the surgery group (95% CI from 7.85 worse to 3.85 better; 1 study, 101 participants) (Analysis 6.5).

6.5 — Comparison 6: Surgery versus multimodal non‐operative treatment, Outcome 5: Health‐related quality of life at long‐term follow‐up: over 3 months

5) Adverse effects

We could not estimate the RR because there were no reported events (very low‐certainty evidence). Jarvik 2009 reported adverse events in 0 of 49 participants (0%) in the multimodal non‐operative treatment group and 0 of 52 participants (0%) in the surgery group.

6) Need for surgery or secondary surgery during follow‐up

This outcome was not reported at short‐term follow‐up.

Moderate‐certainty evidence (downgraded once for the risk of bias) indicates that surgery probably results in less need for subsequent surgery than multimodal non‐operative treatment at long‐term follow‐up. In the surgery group, 0 of 57 participants (0%) had repeated surgery, while in the multimodal non‐operative group, 23 of 59 participants (39%) were referred to surgery. This corresponds to a RR of 0.02 (95% CI 0.00 to 0.35; 1 study, 116 participants) (Analysis 6.6).

6.6 — Comparison 6: Surgery versus multimodal non‐operative treatment, Outcome 6: Need for surgery or secondary surgery at long‐term follow‐up: over 3 months

7. Surgery versus unspecified medical treatment and hand support

One study with 60 participants compared surgery with 'conventional treatment' that included unspecified medical treatment and hand support (Eltabl 2020). The study reported outcomes at 6 months (long‐term).

Primary outcome

1) Clinical improvement

The study in this comparison did not report this outcome.

Secondary outcomes

1) Symptoms

This outcome was not reported at short‐term follow‐up.

At long‐term follow‐up, moderate‐certainty evidence (downgraded once for risk of bias), indicates that symptoms are better with surgery compared with unspecified medical treatment and hand support. The mean BCTQ Symptom Severity Scale (scale from 1 to 5 points, higher is worse, MCID value = 1 point) was 3.13 points with medical treatment and hand support and 2.1 points better with surgery (95% CI not estimable due to reported SD of 0.0 in the surgery group; 1 study, 60 participants) (Analysis 7.1). Imputing a SD from Jarvik 2009 (SD = 0.70), the MD was 2.13 points better with surgery (95% CI from 1.78 better to 2.48 better).

7.1 — Comparison 7: Surgery versus unspecified medical treatment and hand support, Outcome 1: Symptoms at long‐term follow‐up: over 3 months

2) Function

This outcome was not reported at short‐term follow‐up.

At long‐term follow‐up, moderate‐certainty evidence (downgraded once for risk of bias) indicates that surgery probably provides clinically important benefits in hand function compared with unspecified medical treatment and hand support. The mean functional status score (measured by the BCTQ Functional Status Scale from 1 to 5 points, higher is worse, MCID value = 0.7 points) was 2.81 with medical treatment and hand support and 1.74 better with surgery (95% CI from 1.91 better to 1.57 better; 1 study, 60 participants) (Analysis 7.2).

7.2 — Comparison 7: Surgery versus unspecified medical treatment and hand support, Outcome 2: Function at long‐term follow‐up: over 3 months

Due to the unusually small SD, we performed a sensitivity analysis imputing a SD from Jarvik 2009 (SD = 0.82) and, although the CIs widened, our conclusion remained unchanged. The MD was 1.74 points better with surgery (95% CI from 2.15 better to 1.33 better).

3) Pain

At short‐term follow‐up, this outcome was not reported.

At long‐term follow‐up, the mean pain score (measured by VAS, scale from 0 to 10 points, higher is worse, MCID value = 1.5 points) was 7.74 with unspecified medical treatment and hand support and 5.34 better with surgery (95% CI from 5.94 better to 4.74 better; 1 study, 60 participants) (Analysis 7.3). The certainty of evidence was moderate (downgraded once for risk of bias), indicating that surgery probably provides clinically important benefits for pain relief compared with medical treatment and hand support at long‐term follow‐up.

7.3 — Comparison 7: Surgery versus unspecified medical treatment and hand support, Outcome 3: Pain at long‐term follow‐up: over 3 months

4) Health‐related quality of life

The study in this comparison did not report this outcome.

5) Adverse effects

The study in this comparison did not report this outcome.

6) Need for surgery or secondary surgery during follow‐up

The study in this comparison did not report this outcome.

8. Surgery + corticosteroid injection versus corticosteroid injection

One study with 46 participants assessed if surgery given together with corticosteroid injection improves the outcomes over corticosteroid injection alone (Zhang 2019). We used the data reported at 12 weeks follow‐up for short‐term analysis.

Primary outcome

1) Clinical improvement

The study in this comparison did not report this outcome.

Secondary outcomes

1) Symptoms

At short‐term follow‐up, the mean symptom severity score (measured by the BCTQ Symptom Severity Scale from 1 to 5 points, higher is worse, MCID value = 1 point) was 2.06 with corticosteroid injection alone and 0.22 points better when surgery was combined with corticosteroid injection (95% CI from 0.35 better to 0.09 better; 1 study, 46 participants) (Analysis 8.1). We rated the certainty of evidence as low (downgraded once for risk of bias and once for imprecision). This means that when the carpal tunnel is released at the time of corticosteroid injection, the difference in symptom relief may be clinically unimportant compared to having the corticosteroid injection alone.

8.1 — Comparison 8: Surgery + corticosteroid injection versus corticosteroid injection, Outcome 1: Symptoms at short‐term follow‐up: 3 months or less

This outcome was not reported at long‐term follow‐up.

2) Function

At short‐term follow‐up, function (measured by the BCTQ Functional Status Scale from 1 to 5 points, higher is worse, MCID value = 0.7 points) was 2.08 with corticosteroid injection alone and 0.28 points better (95% CI from 0.46 better to 0.10 better; 1 study, 46 participants) when surgery was combined with corticosteroid injection (Analysis 8.2). We rated the certainty of evidence as low (downgraded once for the risk of bias and once for imprecision, as only one study with 46 participants contributed to the analysis).

8.2 — Comparison 8: Surgery + corticosteroid injection versus corticosteroid injection, Outcome 2: Function at short‐term follow‐up: 3 months or less

This outcome was not reported at long‐term follow‐up.

3) Pain

The study in this comparison did not report this outcome.

4) Health‐related quality of life

The study in this comparison did not report this outcome.

5) Adverse effects

Zhang 2019 reported serious adverse events in 0 of 23 participants (0%) in the corticosteroid injection group and in 0 of 23 participants (0%) in the surgery plus corticosteroid injection group at short‐term follow‐up. We could not estimate the RR due to no reported events (very low‐certainty).

6) Need for surgery or secondary surgery during follow‐up

Zhang 2019 reported that 1 of 23 participants (4%) were referred to surgery in the corticosteroid injection group, while 0 of 23 participants (0%) repeated surgery in the surgery plus corticosteroid injection group at short‐term follow‐up. This corresponded to a RR of 0.33 (95% CI 0.01 to 7.78; 1 study, 46 participants) (Analysis 8.3). We rated the certainty of evidence as very low (downgraded once for risk of bias and twice for very serious imprecision), thus we are uncertain about the effect.

8.3 — Comparison 8: Surgery + corticosteroid injection versus corticosteroid injection, Outcome 3: Need for surgery or secondary surgery at short‐term follow‐up: 3 months or less

This outcome was not reported at long‐term follow‐up.

Discussion

Summary of main results

The objective of this review was to assess the evidence regarding the benefits and harms of carpal tunnel release compared with non‐surgical treatment at the short (up to 3 months) and long (over 3 months) term. We included 14 randomised controlled trials (RCTs), which comprised 1231 participants (1293 wrists, as some participants had bilateral carpal tunnel syndrome (CTS)). Of the participants, 1014 (84%) were women and 188 (16%) were men. Trial size varied from 22 to 176 participants and the participants' mean age ranged between 32 and 53 years. The mean duration of symptoms varied from 31 weeks to 3.5 years, and the severity of CTS was commonly from mild to severe.

We are uncertain about the efficacy of surgery, since we did not identify any study comparing surgery with placebo surgery or with no treatment. The included studies compared surgery with: 1) splinting, 2) corticosteroid injection, 3) splinting and corticosteroid injection together, 4) platelet‐rich plasma (PRP) injection, 5) manual therapy (three 30‐minute treatment sessions including desensitisation manoeuvres, delivered once per week), 6) multimodal non‐operative treatment including nonsteroidal anti‐inflammatory drugs (NSAIDs), hand therapy, activity modification, splint, followed by ultrasound if needed (consisting of up to 12 sessions (2 to 4 per week for up to 6 weeks) of focused ultrasound at 1 MHz, 1.0 W/cm², in pulsed mode 1:4, and 15 minutes each), 7) unspecified medical treatment and hand support, and 8) one study combined surgery and corticosteroid injection and compared it to corticosteroid injection alone.

Due to several comparisons without an obvious hierarchy in the importance of comparisons, we present the summary of the main results by outcomes.

Clinical improvement

Short‐term follow‐up

Surgery probably results in a lower rate of clinical improvement compared with corticosteroid injection at short‐term follow‐up (2 studies, 185 participants; moderate certainty). Regarding all other comparisons, we are uncertain about the effects (splinting; corticosteroid injection and splinting together; PRP injection; manual therapy; multimodal non‐operative treatment; or unspecified medical treatment and hand support).

Long‐term follow‐up

Surgery probably results in a higher rate of clinical improvement compared with splinting (3 studies, 210 participants; moderate certainty) or compared with manual therapy (1 study, 111 participants; moderate certainty). Low‐certainty evidence from one study (101 participants) suggested that clinical improvement may not differ between surgery and multimodal non‐operative treatment, but the estimates were imprecise and nearly 40% of the participants in the non‐surgical group had had surgery by the time of the measurement (Jarvik 2009). We are uncertain about the effects regarding other comparisons (corticosteroid injection; corticosteroid injection and splinting; PRP injection; or unspecified medical treatment and hand support).

Symptoms

Short‐term follow‐up

Surgery may provide benefits in improving symptoms compared to corticosteroid injection (2 studies, 90 participants; low certainty). However, the benefit may be clinically unimportant, and this finding did not align with the other outcomes (corticosteroid injection resulted in a higher rate of clinical improvement and better function at short‐term follow‐up). Regarding the other comparisons, symptom relief is probably clinically comparable between surgery and manual therapy (2 studies, 220 participants; moderate certainty); and may be comparable between surgery and splinting (2 studies, 198 participants; low certainty). For the rest of the comparisons we are uncertain about the effects (corticosteroid injection and splint; PRP injection; multimodal non‐operative treatment; unspecified medical treatment and hand support).

Long‐term follow‐up

Surgery probably does not provide clinically important benefits in symptoms compared with splinting (2 studies, 195 participants; moderate certainty); multimodal non‐operative treatment (1 study, 101 participants; moderate certainty); or manual therapy (2 studies, 220 participants; moderate certainty). The uncertainty regarding what is a clinically important difference in symptoms limits our certainty regarding the conclusion of this outcome (see Overall completeness and applicability of evidence).

Surgery probably improves symptoms compared with unspecified medical treatment and hand support (1 study, 60 participants; moderate certainty). We are uncertain if the symptom relief differs between surgery and corticosteroid injection; corticosteroid injection and splinting; or PRP injection.

Function

Short‐term follow‐up

The hand function may be little better with corticosteroid injection compared with surgery (1 study, 147 participants; low certainty). However, the difference is likely to be clinically unimportant. There is probably no clinically important difference in function between surgery and splinting (2 studies, 198 participants; moderate certainty), or surgery and manual therapy (2 studies, 220 participants; moderate certainty). We are uncertain about the effect regarding surgery versus corticosteroid injection and splint; PRP injection; multimodal non‐operative treatment; and unspecified medical treatment and hand support.

Long‐term follow‐up

Surgery probably improves hand function compared with unspecified medical treatment and hand support (1 study, 60 participants; moderate certainty); surgery probably provides a small but likely clinically unimportant benefit compared with splinting (2 studies, 195 participants; moderate certainty); and may improve hand function compared with multimodal non‐operative treatment (1 study, 101 participants; low certainty). Function probably does not differ between surgery and 1) corticosteroid injection (2 studies, 191 participants; moderate certainty); 2) manual therapy (2 studies, 220 participants; moderate certainty); or 3) PRP injection (1 study, 60 participants; low certainty). We are uncertain about the effect of surgery compared to corticosteroid injection together with splint.

Pain

Short‐term follow‐up

Corticosteroid injection (1 study, 147 participants; low certainty), or manual therapy (2 studies, 220 participants; low certainty) may improve pain compared with surgery, but the difference is clinically unimportant. For the remaining comparisons (splinting; corticosteroid injection and splinting together; PRP injection; multimodal non‐operative treatment; or unspecified medical treatment and hand support), we are uncertain about the effect.

Long‐term follow‐up

Surgery probably provides clinically important pain relief compared with unspecified medical treatment and hand support (1 study, 60 participants; moderate certainty). Compared with corticosteroid injection, there was a small benefit for surgery, which is likely clinically unimportant (1 study, 123 participants; moderate certainty). The effect on pain probably does not differ between surgery and manual therapy (2 studies, 220 participants; moderate certainty). For the remaining comparisons (splinting; corticosteroid injection and splinting together; PRP injection; multimodal non‐operative treatment), the effect at long‐term follow‐up is unclear.

Health‐related quality of life

Short‐term follow‐up

Manual therapy may improve health‐related quality of life compared with surgery at short‐term follow‐up (1 study, 120 participants; low certainty). For the other comparisons (splinting; corticosteroid injection; corticosteroid injection and splinting; PRP injection; multimodal non‐operative treatment; or unspecified medical treatment and hand support), the effect is unclear.

Long‐term follow‐up

The effect on health‐related quality of life probably does not differ between surgery and manual therapy (1 study, 118 participants; moderate certainty), and it may not differ between surgery and multimodal non‐operative treatment (1 study, 101 participants; low certainty), or splinting (1 study, 167 participants; low certainty). For the other comparisons (corticosteroid injection; corticosteroid injection and splinting together; PRP injection; or unspecified medical treatment and hand support), the effect is unclear.

Adverse effects

Adverse effects were rare and, since all studies reported a low number of events and were at high risk of bias, we are uncertain if the surgery increases the risk of adverse effects compared with non‐surgical treatments.

Need for surgery or secondary surgery

People using splints (2 studies, 176 participants; moderate certainty) or who receive multimodal non‐operative treatment (1 study, 116 participants; moderate certainty) probably have a need for surgery more often than people with primary surgery have a need for a second surgery at 12 months follow‐up. Regarding the other comparisons (corticosteroid injection; corticosteroid injection and splinting together; PRP injection; manual therapy; or unspecified medical treatment and hand support), we are uncertain about the effects.

Overall completeness and applicability of evidence

The fundamental question of whether carpal tunnel release provides benefits compared with placebo or no treatment without unacceptable risks is unclear, since there were no studies to address this question. The comparisons with commonly used non‐operative treatments yielded inconsistent results, and we thus have limited evidence to inform us whether surgery is better than no surgery in people with CTS.

Applicability

The findings from this review can probably be generalised to typical CTS patients, but it is unclear whether our findings apply to people who have already tried several non‐surgical interventions but failed to achieve satisfactory symptom relief. Typically, people first try various non‐operative treatments or a combination of them, and are referred for surgery if non‐surgical treatments fail to provide sufficient symptom relief. Accepting this rationale for surgical treatment, a comparison of surgery and non‐surgical treatments in a population that is treatment‐naive (or that has received a very short course of non‐surgical treatment) is not optimal.

Geographically, the participants represented East Asia (Hong Kong, China), South Asia (Pakistan), Middle East (Egypt, Iran, Turkey), Southern Europe (Spain), Western Europe (the Netherlands, UK) and North America (USA). The average symptom duration in studies ranged from 31 weeks to between 3 years and 5 years before enrolment. Participants had moderate levels of symptoms at baseline, with mean scores varying between 2.30 and 3.19 (on a 1 to 5 scale of BCTQ; higher is worse). Most studies excluded participants that had coexisting conditions that can mimic CTS (such as cervical radiculopathy). Two studies reported including patients unresponsive to at least two weeks of non‐surgical treatment (including splinting), but many studies did not specify whether the participants had tried non‐operative treatments before enrolment (Table 5).

Carpal tunnel release was achieved by mini‐open, open or endoscopic techniques. Although recovery may be a little faster after endoscopic release than open release, long‐term outcomes are comparable (Vasiliadis 2014). Therefore, it is likely that the long‐term results are generalisable across all surgical techniques.

Definitions for clinical improvement (the primary outcome of this review) varied between the studies. Therefore, the absolute improvement levels may not be directly applicable to all circumstances. However, the relative risk likely reflects the effects well.

Completeness of evidence

We separated non‐operative treatment by different comparisons since it is not plausible that all non‐operative treatments have similar effects, including harms. The findings were inconsistent whenever two or more studies were included in the same comparison. The effects were also inconsistent between short‐term and long‐term follow‐up points within comparisons, which is plausible since recovery from surgery may take months.

The objective of surgery is long‐term relief from symptoms and escape from non‐operative interventions in people who have failed to improve with non‐surgical treatments. It is likely that clinicians and CTS patients deliberating over surgery, therefore, consider long‐term outcomes more relevant. High rates of surgery at long‐term follow‐up would make non‐operative treatment redundant for many patients and have an impact on clinical decision‐making. Follow‐up usually lasted for 6 months to 12 months, and that may be insufficient if most people eventually have surgery.

Repeated surgery was uncommon in this review (0% to 9% in six studies that reported this outcome), and this is supported by a registry study showing secondary surgery in 48 of 2357 patients over a 4‐year period (2%) (Schreiber 2005). This supports the hypothesis that surgery has a permanent effect on symptoms. Due to the length of follow‐up in the included studies, we do not know how large a proportion of the non‐operative groups eventually have surgery, but the highest reported rates were between 30% and 40% (Gerritsen 2002; Jarvik 2009; Ly Pen 2005), which suggests that there are people who do not improve sufficiently with non‐surgical means. This is supported by a systematic review showing that the risk of poor outcomes after non‐surgical care can be as high as 60% to 80% (Burton 2016).

Splinting is probably the simplest non‐operative intervention in people with CTS, since it requires only one visit. Its benefits may be clinically unimportant (Karjalainen 2023) and, in this review, this comparison may present the best assessment of the efficacy of surgery. Corticosteroid injection provides a transient relief of symptoms, but the effect is not maintained in the long term (Huisstede 2018), and up to 90% to 100% of participants had surgery in one trial assessing the efficacy of corticosteroids (Hofer 2021). Consistent with this, corticosteroid injection showed a higher rate of clinical improvement than surgery at short‐term follow‐up in this review, but the difference disappeared at long‐term follow‐up and, at 24 months, 31% of wrists of those injected were "in need of surgery" while 11% needed reoperation.

One small study compared surgery with a combination of splint and corticosteroid injection, theoretically a viable approach, but this study had considerable and unexplained imbalance in the number of participants (23 in the non‐surgical versus 11 in the surgery arm). The loss in surgery may depend on the outcomes as well as per‐protocol analysis. Furthermore, the estimates were imprecise, the study was not blinded and, therefore, more research is needed.

One study combined several non‐surgical treatments together (NSAID, hand therapy, activity modification, splint and ultrasound), but this comparison was confounded by a high rate of switching from non‐surgical to surgical treatment (39%) before the long‐term outcomes were measured (Table 6) (Jarvik 2009). The results from an 'as‐treated' analysis found a larger benefit for surgery, suggesting that the benefits of surgery could have been diluted (the effect would have been larger if participants in the non‐operative group received only allocated non‐surgical treatment). This study included participants who had failed to improve after at least two consecutive weeks of non‐surgical treatment, including wrist splints. Although not treatment‐naive, the participants probably had a relatively short period of non‐surgical treatment, and may not optimally represent the population that usually undergoes surgery.

One group from Spain compared surgery with manual therapy in three separate trials. The participants received three sessions of manual therapy and then continued with home‐based exercises (the same exercises were also given to the surgery group). The rates of treatment switching (from manual therapy to surgery) were relatively low (6% at 1 year and 15% at 4 years). One of these three studies continued follow‐up for up to 4 years and found no benefit for surgery, or difference in later surgery rates (15% versus 13%). The participants had had symptoms for an average of 3 years, but it is unclear whether any non‐surgical treatment options had been tried before enrolment.

Adverse effects were rarely reported, or there were no reported events and the estimates included values compatible with large differences in both directions in all comparisons. The small number of adverse effects in the included studies is inconsistent with the numbers of adverse effects reported in studies comparing open and endoscopic carpal tunnel release, where a 6% to 10% incidence was found after surgery (Vasiliadis 2014). The definition of adverse effects differed between the studies, as some included minor adverse effects, while others reported only major ones. Moreover, the adverse effects were different in the surgical and non‐surgical arms. Although this review found no evidence of a difference in risk, we can assume that the severity of adverse effects is different, with surgery causing more rare but severe adverse effects, such as deep wound infections, systemic infections or nerve injuries.

Interpretation of the BCTQ scores adds another layer of uncertainty to the conclusion of this review. A wide variety of possible MCID values have been proposed (0.16 to 1.45 for the Symptom Severity Scale, and 0.47 to 1.6 for the Functional Status Scale (De Kleermaeker 2018)). We used one point for symptoms (Kim 2013), but this corresponds to a large effect size (standardised mean difference of 1 to 2). The assumption that the MCID value may be lower than one point is supported by the finding that, although the difference in symptoms was below the MCID in comparisons of surgery to splinting and manual therapy, surgery still resulted in a higher rate of clinical improvement. If we were to assume a smaller MCID, e.g. of 10% of the scale (i.e. 0.4 points), the confidence intervals of treatment effects in many comparisons would overlap with the MCID. However, this would decrease the certainty due to imprecision but not change the conclusions given that the benefits for surgery were mostly less than 0.4 points.

Quality of the evidence

The reasons for downgrading the evidence for surgery versus splinting and surgery versus corticosteroids are reported in Table 1 and Table 2.

We did not rate any of the outcomes of interest in any comparison as having a high certainty of evidence. The main causes for downgrading the certainty were a high overall study risk of bias due to no blinding, unexplained inconsistency and imprecision of the estimates. Because we focused on patient‐reported outcomes, the lack of blinding raises concerns that outcome assessment could be influenced by the participants' knowledge of the intervention received. This could only be avoided if the trials used placebo‐surgical controls. Although recent meta‐epidemiological studies suggest that, on average, the effect of bias arising from lack of blinding may be small, particularly when no effects are found for surgery (Karjalainen 2022b), the effect of blinding is likely to be context‐specific, and it is difficult to assess how much the lack of blinding biases the results and how this affects the direction of the estimate.

Regarding surgery versus splint, the quality of evidence was downgraded to moderate for clinical improvement, symptoms and function because of study limitations, mainly lack of blinding. The estimate for health‐related quality of life was further downgraded to low because of imprecision as the confidence intervals included the minimal clinically important value. For adverse effects, the confidence intervals included large effects in both directions, and we therefore downgraded it twice for imprecision.

Regarding surgery versus corticosteroid injection, the evidence was downgraded to low or very low for all outcomes except for function (moderate certainty) because the studies were not blinded, and the estimates were imprecise and also inconsistent between studies.

The remaining comparisons suffered from the same methodological shortcomings, the certainty of evidence varying from moderate to very low. This means that, to improve our confidence in the effects of surgery versus non‐surgical treatment, we need more data from large, unbiased trials.

Potential biases in the review process

We conducted searches as expected by Cochrane methodological standards and did not use language or date restrictions. It is unlikely that we missed large‐scale, rigorously conducted published trials. There were too few trials to assess publication bias using funnel plots.

We identified nine potentially eligible studies that lacked the necessary information to decide if they should be included (Abedi 2018; Elwakil 2007; Hrkovic 2016; Linscheid 1967; Lo 2021; Shazad 2022) or they seemed to be eligible but were not published (IRCT20200629047948N1; NCT00981565; NCT04216147). These are now classified as awaiting classification. Inclusion of these trials could affect the estimates of this review in the following comparisons: surgery versus splinting (two studies awaiting classification) and surgery versus corticosteroid injection (four studies awaiting classification). There were also three new potential comparisons: surgery versus physical therapy and application of physical agents (one study); surgery versus laser therapy (one study); and surgery versus percutaneous needle electrolysis (one study).

Agreements and disagreements with other studies or reviews

It is difficult to directly compare the results of this review with others (Huisstede 2010; Klokkari 2018; Ren 2016; Shi 2011; Shi 2020), mainly due to different methodological approaches. In contrast to our review, the other reviews also included other types of studies (comparative cohort studies, or observational studies) (Klokkari 2018; Ren 2016; Shi 2011; Shi 2020). Furthermore, only one review used MCIDs (0.8 points on the symptom scale and 0.5 points on the function scale) in the interpretation of the result (Shi 2020).

Most of the previous reviews pooled all non‐surgical treatments, except Huisstede 2010 and Ren 2016, which also provided data for subgroups in the comparisons: surgery versus splinting, surgery versus corticosteroid injection and surgery versus multimodal non‐operative treatment. The findings in the subgroup analysis differed in some aspects:

Huisstede 2010 reported that clinical improvement at short‐term follow‐up was significantly better with surgery compared with splinting (RR 1.38, 95% CI 1.08 to 1.75; 1 study) at 3 months. Because we included one study more, our estimates were different and, due to inconsistency, imprecision and bias, we deemed the certainty of evidence to be very low (RR 1.06, 95% CI 0.48 to 2.34; 2 studies). Our sensitivity analysis (removing studies with 'some concerns' of risk of bias in the randomisation process) shows similar results to Huisstede 2010. Regarding the effect of surgery versus multimodal non‐operative treatment on symptom severity and function, our conclusion was different to Huisstede 2010 because we considered benefit important only if the MD exceeded the MCID value.

Ren 2016 reported subgroup analysis for clinical improvement, pooling the studies based on the non‐surgical treatment (surgery versus splinting; surgery versus corticosteroid injection; surgery versus multimodal non‐operative treatment), but included a non‐randomised study. Ren 2016 concluded that clinical improvement was not statistically significantly different between the surgery and corticosteroid groups (odds ratio (OR) 0.85, 95%CI 0.41 to 1.77; 2 studies), but our findings suggested that surgery probably results in a lower rate of clinical improvement compared with corticosteroid injection (RR 0.75, 95% CI 0.60 to 0.92; 2 studies). The different estimates are related to the definition of clinical improvement in Ly Pen 2005. Ren 2016 used "50% improvement in the VAS score for diurnal pain at two years follow‐up", whereas we used a 70% reduction in the VAS score for nocturnal paraesthesias for people with unilateral CTS plus the most symptomatic wrist from people with bilateral CTS at 1‐year follow‐up. This review also used a different summary method (RR versus OR).

Compared to the previous published version of this review, we included 10 new studies in the meta‐analysis. In contrast to the previous version, we did not pool all non‐surgical groups together. Therefore, the findings are not comparable.

Authors' conclusions

Implications for practice.

The efficacy of carpal tunnel release in people with carpal tunnel syndrome (CTS) is unclear, since we identified no studies comparing it with placebo or no treatment.

At short‐term follow‐up (up to 3 months), surgery may not provide clinically important benefits compared to various non‐operative treatments, but our conclusions are limited by risk of bias, inconsistency and often imprecision of the outcomes. However, people most likely consider surgery because it can provide long‐term improvements, with short‐term results likely to be secondary in clinical decision‐making.

At long‐term follow‐up, surgery probably provides better clinical improvement compared with splinting and manual therapy, but the benefits of surgery in symptoms and hand function seem to be small compared with non‐surgical treatment. The decision for a patient to opt for surgery should balance the small benefits and potential risks of surgery. Patients with severe symptoms, a high preference for clinical improvement and reluctance to adhere to non‐surgical options, and who do not consider potential surgical risks and morbidity a burden, may choose surgery. On the other hand, those who have tolerable symptoms, have not tried non‐surgical options and want to avoid surgery‐related morbidity can start with non‐surgical options and have surgery only if they fail to achieve a satisfactory symptom state with non‐surgical options. However, at the moment, we do not know if surgery is better than continuing non‐surgical treatments in people who are not satisfied with the outcomes of non‐surgical options.

We are uncertain if the risk of adverse effects differs between surgery and non‐surgical treatments. The severity of adverse effects may differ between surgery and non‐surgical options, with surgery potentially causing rare severe adverse effects, such as deep wound or systemic infection or nerve injury, which are not plausible risks in non‐surgical care. This may be explained to people who consider opting for surgery.

Implications for research.

A placebo‐surgery‐controlled study assessing the efficacy of carpal tunnel release would be helpful in clinical decision‐making in people considering surgery. From the clinical point of view, the most useful information would probably come from a study randomising people who are not satisfied with non‐surgical interventions, since this is the population that usually deliberates about whether to undergo surgery.

Similarly, comparisons of surgery and non‐surgical interventions should primarily recruit people who have already exhausted non‐surgical options. If, however, treatment‐naive patients are randomised to surgery versus a non‐surgical option (or a combination of them), long‐term (up to several years) follow‐up is recommended, because long‐term outcomes could impact clinical decision‐making; a high rate of conversion to surgery could influence some people to choose surgery initially, despite small benefits during the first year.

Optimally, future studies should use rigorous methods, including blinding the participants and study personnel from treatment allocation to minimise sources of bias, and plan strategies to minimise early switching to surgery to avoid bias arising from it.

In the included studies, few participants had severe CTS. This population was probably excluded in most trials because surgery is considered the primary option in this population. However, since we failed to find evidence supporting large benefits from surgery compared with non‐surgical treatments, a study randomising people at a severe stage may be appropriate.

Furthermore, more evidence is needed on the minimal clinically important difference of the Boston Carpal Tunnel Questionnaire to help to interpret the results of trials and meta‐analyses.

What's new

Date

Event

Description

8 January 2024

New citation required and conclusions have changed

Search updated to November 2022. Update completed with new author team. The conclusions have changed as the review now compares surgery with each non‐surgical intervention separately.

8 January 2024

New search has been performed

Based on an updated search, nine new studies with 815 randomised participants are included in the review (Awan 2015; Eltabl 2020; Fernandez‐de‐las‐Penas 2015; Fernandez‐de‐las‐Penas 2017a; Fernandez‐de‐las‐Penas 2017b; Ismatullah 2013; Jafari 2018; Jarvik 2009; Zhang 2019); four studies included in the previous version of this review remain included (Garland 1964; Gerritsen 2002; Hui 2005; Ly Pen 2005); one study (67 randomised participants), which was awaiting classification in the previous version of this review, is now included (Ucan 2006).
In contrast with the previous version of this review, we used a random‐effects model as default in the meta‐analyses, instead of a fixed‐effect model.
In the previous version of this review (Verdugo 2008), all studies were pooled together regardless of the comparator in one meta‐analysis. In the updated version, we performed separate meta‐analysis for each different comparator.
Furthermore, we used different data regarding:

clinical improvement from the studies of Gerritsen 2002 and Garland 1964 (in the previous version, those lost to follow‐up were included as 'not improved', but we used only reported data);
need for surgery from the studies of Garland 1964 and Ly Pen 2005 (in the previous version, patients who underwent surgery were used in analysis. We used the data for those reported as "in need of surgery" too).

We accepted clinical success or improvement as defined by the authors of the included studies.
We excluded the following secondary outcomes, which were included in the previous version of this review:

neurophysiological parameters (because their clinical relevance is unclear);
return to work at 3 months or less of follow‐up.

We included the following secondary outcomes, which were not included in the previous version of this review:

symptoms;
function;
pain (because surgery may cause chronic pain, for example painful scar);
generic health‐related quality of life.

We used the GRADE approach to assess the certainty of evidence (it was not used in the previous version of this review).
These changes were defined before the screening of the search.
Changes in the clinical implications of the findings: in the previous version of this review, the conclusion stated that surgical treatment of carpal tunnel syndrome is more effective than splinting, but further research was needed to determine its effectiveness for people with mild symptoms and whether it is better than steroid injection.
In the updated version, we have slightly changed the conclusion. Due to generally unclear benefits and harms of surgery, we emphasise the need for a balanced decision‐making process, considering the severity of the disease as well as patients' preferences.

Open in a new tab

History

Protocol first published: Issue 2, 1999 Review first published: Issue 2, 2002

Date	Event	Description
13 May 2008	Amended	Converted to new review format.

Open in a new tab

Risk of bias

Risk of bias for analysis 1.1 Clinical improvement at short‐term follow‐up: 3 months or less.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Gerritsen 2002

Low risk of bias

Quote: "By preparing a list for each hospital, the randomization was stratified by center. Permuted blocks of 4 patients were formed to ensure near‐equal distribution of patients over the 2 treatment groups in each hospital. Despite the small block sizes, the potential for unmasking was considered to be low because the neurologists did not know the method of randomization and different neurologists selected patients in each hospital. Even if a neurologist were to know the allocation scheme, the chance was high that when he/she selected a patient, this patient would not be assigned the next treatment on the list because a patient selected by another neurologist has an earlier appointment for the trial. The random sequence of the permuted blocks was generated by using random number tables. The principal investigator (A.A.M.G.), who was not involved in the selection and allocation of patients, prepared, coded, and sealed opaque envelopes containing the treatment allocation. After the baseline assessment, the next envelope was handed to the patient by the research assistant to ensure concealment of allocation."
Permuted block randomization performed, stratified by center. Allocation concealment was well described and appropriate. There were no concerns about differences between the study groups at the baseline.

Low risk of bias

Participants and care takers were not blinded. There were deviations from the intended interventions in both study groups, however we had no reason to believe they occurred due to the trial context. The analyses were performed according to the intention‐to‐treat principle.

Low risk of bias

Outcome was available for nearly all participants (less than 10% missing in each of the study groups).

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Low risk of bias

Outcome analysed in accordance with the protocol.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Ucan 2006

Some concerns

Quote: "The patients who fulfilled our inclusion criteria were asked to participate in the study. Previously prepared, randomly enumerated closed envelopes which contained the three treatment methods were given consecutively to each patient."
The randomisation and allocation concealment procedure were not sufficiently described to judge whether they were appropriate. Also, it was not clear how many participants were randomized to each of the study groups, thus we could not clearly say if there were no baseline imbalances.

Some concerns

The study did not provide information about blinding the participants and caretakers, but due to the nature of interventions it was unlikely they were blinded. There were deviations from the intended interventions (refusal of treatment), however we had no reason to believe they occurred due to the trial context. The analysis was likely done by per‐protocol principle.

High risk of bias

The authors reported outcomes for 23 participants in the non‐surgical group and for 11 participants in the surgery group, and reasons for this large disproportionate loss were unclear.

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Some concerns

The study protocol not published and no other registry records available. Results reported as intended by the methods section of the article.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome, as well as due to unexplained large imbalance between number of participants in the study groups.

Open in a new tab

Risk of bias for analysis 1.2 Clinical improvement at long‐term follow‐up: over 3 months.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Garland 1964

Some concerns

Quote: "Each patient accepted on both clinical and electromyographic grounds was then allocated to the treatment or control group by a secretary, using a previously prepared random list."
The randomisation procedure was not sufficiently described to judge whether it was appropriate. No information about the allocation sequence concealment was provided.

Some concerns

Due to the nature of interventions it is unlikely the participants and care‐takers were blinded. There were relatively small deviations from the intended interventions and intention‐to‐treat analysis were not done, but we deemed the impact on the result was likely small.

Low risk of bias

Outcome was available for nearly all participants (less than 10% missing in each of the study groups).

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Some concerns

The study protocol not published and no other registry records available. Results reported as intended by methods section of the article.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Gerritsen 2002

Low risk of bias

Some concerns

Participants and care takers were not blinded. There were deviations from the protocol in both study groups, including about 40% of the participants in the non‐operative arm had surgery before the long‐term measurement. However, we had no reason to believe it occurred due to the trial context. The analyses were performed according to the intention‐to‐treat principle.

Some concerns

There were 10 to 20 % of participants lost in the surgery group, and less than 10% in the splinting group. Reasons for missing data were relatively balanced between the study groups, except for the refusal of treatment, which might also be related to the true value of the outcome. However, we had no strong reason to believe this missingness occured due to the true value of the outcome.

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Low risk of bias

Outcome analysed in accordance with the protocol and methods section of the article.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Ucan 2006

Some concerns

High risk of bias

The authors reported outcomes for 23 participants in the non‐surgical group and for 11 participants in the surgery group, and reasons for this large disproportionate loss were unclear.

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Some concerns

The study protocol not published and no other registry records available. Results reported as intended by the methods section of the article.

High risk of bias

Open in a new tab

Risk of bias for analysis 1.3 Symptoms at short‐term follow‐up: 3 months or less.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Gerritsen 2002

Low risk of bias

Participants and care‐providers were not blinded. There were deviations from the intended interventions in both study groups, however we had no reason to believe they occurred due to the trial context. The analyses were performed according to the intention‐to‐treat principle.

Low risk of bias

Outcome was available for nearly all participants (less than 10% missing in each of the study groups).

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Low risk of bias

Outcome analysed in accordance with the protocol and methods section of the article.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Ucan 2006

Some concerns

High risk of bias

The authors reported outcomes for 23 participants in the non‐surgical group and for 11 participants in the surgery group, and reasons for this large disproportionate loss were unclear.

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Some concerns

The study protocol not published and no other registry records available. Results reported as intended by the methods section of the article.

High risk of bias

Open in a new tab

Risk of bias for analysis 1.4 Symptoms at long‐term follow‐up: over 3 months.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Gerritsen 2002

Low risk of bias

Some concerns

Participants and care takers were not blinded. There were deviations from the protocol in both study groups, including about 40% of the participants in the non‐operative arm had surgery before the long‐term measurement. However we had no reason to believe it occurred due to the trial context. The analyses were performed according to the intention‐to‐treat principle.

Some concerns

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Low risk of bias

Outcome analysed in accordance with the protocol and the methods section of the article.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Ucan 2006

Some concerns

High risk of bias

The authors reported outcomes for 23 participants in the non‐surgical group and for 11 participants in the surgery group, and the reasons for this large disproportionate loss were unclear.

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Some concerns

The study protocol not published and no other registry records available. Results reported as intended by the methods section of the article.

High risk of bias

Open in a new tab

Risk of bias for analysis 1.5 Function at short‐term follow‐up: 3 months or less.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Gerritsen 2002

Low risk of bias

Outcome was available for nearly all participants (less than 10% missing in each of the study groups).

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Low risk of bias

Outcome analysed in accordance with the protocol and methods section of the article.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Ucan 2006

Some concerns

The study did not provide information about blinding the participants and caretakers, but due to the nature of interventions it is unlikely they were blinded. There were deviations from the intended interventions (refusal of treatment), however we had no reason to believe they occurred due to the trial context. The analysis was likely not done appropriately (likely done by per‐protocol principle).

High risk of bias

The authors reported outcomes for 23 participants in the non‐surgical group and for 11 participants in the surgery group, and reasons for this large disproportionate loss were unclear.

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Some concerns

The study protocol not published and no other registry records available. Results reported as intended by the methods section of the article.

High risk of bias

Open in a new tab

Risk of bias for analysis 1.6 Function at long‐term follow‐up: over 3 months.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Gerritsen 2002

Low risk of bias

Some concerns

Participants and care takers were not blinded. There were deviations from the protocol in both study groups, including about 40% of the participants in the non‐operative arm had surgery before the long‐term measurement. However we had no reason to believe it occurred due to the trial context. The analyses were performed according to the intention‐to‐treat principle.

Some concerns

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Low risk of bias

Outcome analysed in accordance with the protocol and methods section of the article.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Ucan 2006

Some concerns

High risk of bias

The authors reported outcomes for 23 participants in the non‐surgical group and for 11 participants in the surgery group, and reasons for this large disproportionate loss were unclear.

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Some concerns

The study protocol not published and no other registry records available. Results reported as intended by the methods section of the article.

High risk of bias

Open in a new tab

Risk of bias for analysis 1.7 Health‐related quality of life at long‐term follow‐up: over 3 months.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Gerritsen 2002

Low risk of bias

Some concerns

Participants and care takers were not blinded. There were deviations from the protocol in both study groups, including about 40% of the participants in the non‐operative arm had surgery before the long‐term measurement. However we had no reason to believe it occurred due to the trial context. The analyses were performed according to the intention‐to‐treat principle.

Some concerns

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

High risk of bias

Based on the protocol HRQL should be measured both by SF‐36 and EuroQol, but only EuroQol was reported. The reason for not providing SF‐36 results not reported.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Open in a new tab

Risk of bias for analysis 1.8 Adverse effects.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Gerritsen 2002

Low risk of bias

Some concerns

Participants and care takers were not blinded. There were deviations from the protocol in both study groups, including about 40% of the participants in the non‐operative arm had surgery before the long‐term measurement. However we had no reason to believe it occurred due to the trial context. The analyses were performed according to the intention‐to‐treat principle.

Some concerns

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Low risk of bias

Outcome analysed in accordance with the protocol.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Ucan 2006

Some concerns

High risk of bias

The authors reported outcomes for 23 participants in the non‐surgical group and for 11 participants in the surgery group, and the reasons for this large disproportionate loss were unclear.

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Some concerns

The study protocol not published and no other registry records available.

High risk of bias

Open in a new tab

Risk of bias for analysis 1.9 Need for surgery or secondary surgery at short‐term follow‐up: 3 months or less.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Gerritsen 2002

Low risk of bias

Outcome was available for nearly all participants (less than 10% missing in each of the study groups).

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Low risk of bias

Outcome analysed in accordance with the protocol.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Open in a new tab

Risk of bias for analysis 1.10 Need for surgery or secondary surgery at long‐term follow‐up: over 3 months.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Garland 1964

Some concerns

Low risk of bias

Outcome was available for nearly all participants (less than 10% missing in each of the study groups).

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Some concerns

The study protocol not published and no other registry records available.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Gerritsen 2002

Low risk of bias

Some concerns

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Low risk of bias

Outcome analysed in accordance with the protocol.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Open in a new tab

Risk of bias for analysis 2.1 Clinical improvement at short‐term follow‐up: 3 months or less.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Awan 2015

Some concerns

Quote: "Patients were randomly allocated in two groups by lottery method."
The randomisation and allocation concealment procedure were not sufficiently described to judge whether they were appropriate.

Low risk of bias

The study did not provide information about blinding the participants and caretakers, but due to the nature of interventions it is unlikely they were blinded. The study did not provide any information regarding deviations from the intended interventions, but we did not have any reason to believe such deviations happened. The analysis seemed likely done appropriately.

Low risk of bias

No any loss to follow‐up reported (in both groups 58/58 participants completed the follow‐up).

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we cannot exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Some concerns

The study protocol not published and no other registry records available. Results reported as intended by the methods section of the article.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Ly Pen 2005

Low risk of bias

Quote: "Treatment assignments were randomly generated by computer in blocks of 6 cases. In patients with bilateral CTS, treatment assignments were made for individual wrists. Sealed envelopes containing the treatment assignments were provided by our biostatistics unit. Immediately after patient enrollment, the envelope containing the treatment assignment for each wrist was opened, and the specific treatment was assigned."
Block randomization performed. Allocation concealment seemed appropriate. Regarding baseline differences, there was a considerable difference between unilateral/bilateral patient proportion in each of the study groups, but we deemed it not a source of bias.

Some concerns

The study did not provide information about blinding the participants and caretakers, but due to the nature of interventions it was unlikely they were blinded. There were deviations from the intended interventions (switching treatment, refusal of treatment), however we had no reason to believe they occurred due to the trial context. The analysis was likely done by per‐protocol principle.

High risk of bias

There were over 30% missing data in each of the study subgroups (unilateral+the most symptomatic hand of bilateral wrists). Separate flow chart for this subgroup of participants with reasons for missing data was not provided.  In general, the trial excluded "treatment failures" from the analysis, i.e., those participants, who did not improve by a certain time.

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Some concerns

Trial registry record available, however it is retrospective. Results seemed to be reported as intended by the methods section of the article.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome, as well as due to large proportion of missing data.

Open in a new tab

Risk of bias for analysis 2.2 Clinical improvement at long‐term follow‐up: over 3 months.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Hui 2005

Low risk of bias

Quote: "Patients were randomly allocated using a random computer‐generated code to either surgical decompression or injection of methylprednisolone acetate. A research assistant not involved in the management of cases prepared and coded opaque envelopes containing the treatment allocation. Physicians who recruited patients did not know in advance to which group their patients would be assigned."
Randomization performed by random computer‐generated code. Allocation concealment seemed appropriate. There were no concerns about differences between the study groups at the baseline.

Low risk of bias

Due to the nature of interventions the participants and caretakers were not blinded. The study did not provide any information regarding deviations from the intended interventions, and we did not have any reason to believe such deviations happened. The analysis seemed likely done appropriately.

Low risk of bias

Quote: "No patients were lost to follow‐up."

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Some concerns

The study protocol not published and no other registry records available. The outcome was not reported in the article, but according to the methods section not planned to report either.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Jafari 2018

Low risk of bias

Quote: "The remaining 82 patients were randomly assigned into 2 study groups, 41 patients each (Figure 1). Predesigned blocks were used for randomization purposes. A computer‐generated random number list was used for random allocation of the patients. Group assignment remained concealed from investigators during the course of data gathering."
Block randomisation performed by a computer‐generated random number list. Allocation concealment seemed appropriate. There were no concerns about differences between the study groups at the baseline.

Low risk of bias

The study did not provide information about blinding the participants and caretakers, but due to the nature of interventions it is unlikely they were blinded. The study did not provide any information regarding deviations from the intended interventions, and we did not have any reason to believe such deviations happened. The analysis seemed likely done appropriately.

Some concerns

There were 15 % of participants lost in the surgery group, and 20% in the corticosteroid injection group. Reasons for the dropouts not reported, therefore, not possible to judge if the missingness was random.

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Some concerns

The registry record of the study could not be found.  Results reported only partly, but it was not clear if it happened due to the nature of the findings.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Ly Pen 2005

Low risk of bias

Some concerns

The study did not provide information about blinding the participants and caretakers, but due to the nature of interventions it was unlikely they were blinded. There were deviations from the intended interventions (switching treatment, refusal of treatment), however we had no reason to believe they occurred due to the trial context. The analysis was likely done by per‐protocol principle.

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Some concerns

Trial registry record available, however it is retrospective. Results seemed to be reported as intended by the methods section of the article.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome, as well as due to large proportion of missing data.

Open in a new tab

Risk of bias for analysis 2.3 Symptoms at short‐term follow‐up: 3 months or less.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Hui 2005

Low risk of bias

Quote: "No patients were lost to follow‐up."

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Some concerns

The study protocol not published and no other registry records available. Results seemed to be reported as intended by methods section of the article.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Ismatullah 2013

Some concerns

Quote: "Patients were randomly assigned, by lottery method, to two treatment groups: LSI group (20 patients) and CTR group (20 patients)."
The randomisation procedure was not sufficiently described to judge whether it was appropriate. No information about the allocation sequence concealment was provided.

Low risk of bias

The study did not provide information about blinding the participants and caretakers, but due to the nature of interventions it is unlikely they were blinded. The study did not provide any information regarding deviations from the intended interventions, and we did not have any reason to believe such deviations happened. The analysis seemed likely done appropriately.

Low risk of bias

No any loss to follow‐up reported (in both groups 20/20 participants completed the follow‐up).

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Some concerns

The study protocol not published and no other registry records available. According to the methods section of the article, the timepoint of the follow‐up should be 6 weeks but in the results section numbers at 4 weeks were provided. However, it did not affect our short‐term analysis.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Open in a new tab

Risk of bias for analysis 2.4 Symptoms at long‐term follow‐up: over 3 months.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Hui 2005

Low risk of bias

Quote: "No patients were lost to follow‐up."

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Some concerns

The study protocol not published and no other registry records available. Results seemed to be reported as intended by methods section of the article.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Jafari 2018

Low risk of bias

The study did not provide information about blinding the participants and caretakers, but due to the nature of interventions it is unlikely they were blinded. The study did not provide any information regarding deviations from the intended interventions, and we did not have any reason to believe such deviations happened. The analysis seemed likely done appropriately.

Some concerns

High risk of bias

The registry record of the study could not be found.  Results reported only partly, but it was not clear if it happened due to the nature of the findings.

Some concerns

The registry record of the study could not be found. According to the methods and discussion section, the outcome should be reported at six months, but in the results section the numbers for six weeks were reported. We assumed that the outcomes were measured at six months (i.e. we used the data at long‐term analysis).

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Open in a new tab

Risk of bias for analysis 2.5 Function at short‐term follow‐up: 3 months or less.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Ly Pen 2005

Low risk of bias

Some concerns

The study did not provide information about blinding the participants and caretakers, but due to the nature of interventions it was unlikely they were blinded. There were deviations from the intended interventions (switching treatment, refusal of treatment), however we had no reason to believe they occurred due to the trial context. The analysis was likely done by per‐protocol principle.

Some concerns

There were between 10 to 20 % of participants lost in the surgery group, and less than 5% in the corticosteroid injection group.
Disproportionate loss occurred due to the refusal of treatment, which might be related to the true value of the outcome. However, we had no strong reason to believe the refusal of treatment indeed occurred due to the true value of the outcome.
In general, the trial excluded "treatment failures" from the analysis, i.e., those participants, who did not improve by a certain time. However, at the short term the loss due to the "treatment failures" was small and balanced between the study groups.

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Some concerns

Trial registry record available, however it is retrospective. Results reported as intended by the methods section of the article.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Open in a new tab

Risk of bias for analysis 2.6 Function at long‐term follow‐up: over 3 months.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Jafari 2018

Low risk of bias

The study did not provide information about blinding the participants and caretakers, but due to the nature of interventions it is unlikely they were blinded. The study did not provide any information regarding deviations from the intended interventions, and we did not have any reason to believe such deviations happened. The analysis seemed likely done appropriately.

Some concerns

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Some concerns

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Ly Pen 2005

Low risk of bias

Some concerns

The study did not provide information about blinding the participants and caretakers, but due to the nature of interventions it was unlikely they were blinded. There were deviations from the intended interventions (switching treatment, refusal of treatment), however we had no reason to believe they occurred due to the trial context. The analysis was likely done by per‐protocol principle.

High risk of bias

There were over 20 % of participants lost in both study groups.
Disproportionate loss occurred due to the refusal of treatment, which might be related to the true value of the outcome. However, we had no strong reason to believe the refusal of treatment indeed occurred due to the true value of the outcome.
In general, the trial excluded "treatment failures" from the analysis, i.e., those participants, who did not improve by a certain time.

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Some concerns

Trial registry record available, however it is retrospective. Results reported as intended by the methods section of the article.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome, as well as due to large proportion of missing data.

Open in a new tab

Risk of bias for analysis 2.7 Pain at short‐term follow‐up: 3 months or less.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Ly Pen 2005

Low risk of bias

Some concerns

The study did not provide information about blinding the participants and caretakers, but due to the nature of interventions it was unlikely they were blinded. There were deviations from the intended interventions (switching treatment, refusal of treatment), however we had no reason to believe they occurred due to the trial context. The analysis was likely done by per‐protocol principle.

Some concerns

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Some concerns

Trial registry record available, however it is retrospective. Results reported as intended by the methods section of the article.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Open in a new tab

Risk of bias for analysis 2.8 Pain at long‐term follow‐up: over 3 months.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Ly Pen 2005

Low risk of bias

Some concerns

The study did not provide information about blinding the participants and caretakers, but due to the nature of interventions it was unlikely they were blinded. There were deviations from the intended interventions (switching treatment, refusal of treatment), however we had no reason to believe they occurred due to the trial context. The analysis was likely done by per‐protocol principle.

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Some concerns

Trial registry record available, however it is retrospective. Results reported as intended by the methods section of the article.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome, as well as due to large proportion of missing data.

Open in a new tab

Risk of bias for analysis 2.9 Adverse effects.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Hui 2005

Low risk of bias

Quote: "No patients were lost to follow‐up."

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Some concerns

The study protocol not published and no other registry records available.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Ismatullah 2013

Some concerns

Low risk of bias

The study did not provide information about blinding the participants and caretakers, but due to the nature of interventions it is unlikely they were blinded. The study did not provide any information regarding deviations from the intended interventions, and we did not have any reason to believe such deviations happened. The analysis seemed likely done appropriately.

Low risk of bias

No any loss to follow‐up reported (in both groups 20/20 participants completed the follow‐up).

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Some concerns

The study protocol not published and no other registry records available.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Open in a new tab

Risk of bias for analysis 2.10 Need for surgery or secondary surgery at short‐term follow‐up: 3 months or less.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Ly Pen 2005

Low risk of bias

Some concerns

There were between 10 to 20 % of participants lost in the surgery group and less than 5% in the corticosteroid group.  Disproportionate loss occurred due to the refusal of treatment, which might be related to the true value of the outcome. However, we had no strong reason to believe the refusal of treatment indeed occurred due to the true value of the outcome.

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Some concerns

Trial registry record available, however it is retrospective.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Open in a new tab

Risk of bias for analysis 2.11 Need for surgery or secondary surgery at long‐term follow‐up: over 3 months.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Ly Pen 2005

Low risk of bias

Some concerns

There were 20 % of participants lost in the surgery group and less than 10% in the corticosteroid group due to other reasons than "treatment failure" (as the "treatment failure" accounts for the outcome of interest).  Disproportionate loss occurred due to the refusal of treatment, which might be related to the true value of the outcome. However, we had no strong reason to believe the refusal of treatment indeed occurred due to the true value of the outcome.

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Some concerns

Trial registry record available, however it is retrospective.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Open in a new tab

Risk of bias for analysis 3.1 Clinical improvement at short‐term follow‐up: 3 months or less.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Ucan 2006

Some concerns

High risk of bias

The authors reported outcomes for 23 participants in the non‐surgical group and for 11 participants in the surgery group, and reasons for this large disproportionate loss were unclear.

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Some concerns

The study protocol not published and no other registry records available. Results reported as intended by the methods section of the article.

High risk of bias

Open in a new tab

Risk of bias for analysis 3.2 Clinical improvement at long‐term follow‐up: over 3 months.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Ucan 2006

Some concerns

High risk of bias

The authors reported outcomes for 23 participants in the non‐surgical group and for 11 participants in the surgery group, and reasons for this large disproportionate loss were unclear.

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Some concerns

The study protocol not published and no other registry records available. Results reported as intended by the methods section of the article.

High risk of bias

Open in a new tab

Risk of bias for analysis 3.3 Symptoms at short‐term follow‐up: 3 months or less.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Ucan 2006

Some concerns

High risk of bias

The authors reported outcomes for 23 participants in the non‐surgical group and for 11 participants in the surgery group, and reasons for this large disproportionate loss were unclear.

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Some concerns

The study protocol not published and no other registry records available. Results reported as intended by the methods section of the article.

High risk of bias

Open in a new tab

Risk of bias for analysis 3.4 Symptoms at long‐term follow‐up: over 3 months.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Ucan 2006

Some concerns

High risk of bias

The authors reported outcomes for 23 participants in the non‐surgical group and for 11 participants in the surgery group, and the reasons for this large disproportionate loss were unclear.

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Some concerns

The study protocol not published and no other registry records available. Results reported as intended by the methods section of the article.

High risk of bias

Open in a new tab

Risk of bias for analysis 3.5 Function at short‐term follow‐up: 3 months or less.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Ucan 2006

Some concerns

High risk of bias

The authors reported outcomes for 23 participants in the non‐surgical group and for 11 participants in the surgery group, and reasons for this large disproportionate loss were unclear.

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Some concerns

The study protocol not published and no other registry records available. Results reported as intended by the methods section of the article.

High risk of bias

Open in a new tab

Risk of bias for analysis 3.6 Function at long‐term follow‐up: over 3 months.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Ucan 2006

Some concerns

High risk of bias

The authors reported outcomes for 23 participants in the non‐surgical group and for 11 participants in the surgery group, and reasons for this large disproportionate loss were unclear.

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Some concerns

The study protocol not published and no other registry records available. Results reported as intended by the methods section of the article.

High risk of bias

Open in a new tab

Risk of bias for analysis 3.7 Adverse effects.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Ucan 2006

Some concerns

High risk of bias

The authors reported outcomes for 23 participants in the non‐surgical group and for 11 participants in the surgery group, and the reasons for this large disproportionate loss were unclear.

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Some concerns

The study protocol not published and no other registry records available.

High risk of bias

Open in a new tab

Risk of bias for analysis 4.1 Symptoms at long‐term follow‐up: over 3 months.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Eltabl 2020

Some concerns

Quote: "Patients of the study group were divided into three groups as they were randomized to receive PRP or medical treatment or went to surgery in an equal ratio using a computer‐generated code."
The randomisation procedure was not sufficiently described to judge whether it was appropriate. No information about the allocation sequence concealment was provided.

Low risk of bias

The study did not provide information about blinding the participants and caretakers, but due to the nature of interventions it is unlikely they were blinded. The study did not provide any information regarding deviations from the intended interventions, and we did not have any reason to believe such deviations happened. The analysis seemed likely done appropriately.

Low risk of bias

No loss to follow up reported.

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

High risk of bias

Trial registry record available, however it is retrospective.  The results for surgery and PRP group showed such standard deviation (sd=0.0), which seemed to be unlikely possible for n = 30 in both groups. As we did not receive clarification from the authors if there was any mistake in the result, the method of analysis raised concerns.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome, as well as due to concerns regarding the method of analysis.

Open in a new tab

Risk of bias for analysis 4.2 Function at long‐term follow‐up: over 3 months.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Eltabl 2020

Some concerns

Low risk of bias

The study did not provide information about blinding the participants and caretakers, but due to the nature of interventions it is unlikely they were blinded. The study did not provide any information regarding deviations from the intended interventions, and we did not have any reason to believe such deviations happened. The analysis seemed likely done appropriately.

Low risk of bias

No loss to follow up reported.

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Some concerns

Trial registry record available, however it is retrospective. Results seemed to be reported as intended by the methods section of the article.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Open in a new tab

Risk of bias for analysis 4.3 Pain at long‐term follow‐up: over 3 months.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Eltabl 2020

Some concerns

Low risk of bias

The study did not provide information about blinding the participants and caretakers, but due to the nature of interventions it is unlikely they were blinded. The study did not provide any information regarding deviations from the intended interventions, and we did not have any reason to believe such deviations happened. The analysis seemed likely done appropriately.

Low risk of bias

No loss to follow up reported.

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Some concerns

Trial registry record available, however it is retrospective. Results reported as intended by the registry record.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Open in a new tab

Risk of bias for analysis 5.1 Clinical improvement at long‐term follow‐up: over 3 months.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Fernandez‐de‐las‐Penas 2015

Low risk of bias

Quote: "Patients were randomly assigned to receive physical therapy or a surgical procedure. Concealed allocation was conducted using a computer‐generated randomized table of numbers created by a statistician who was not otherwise involved in the trial and did not participate in the analysis or interpretation of the results. Individual and sequentially numbered index cards with the random assignment were prepared. The index cards were folded and placed in sealed opaque envelopes. Another researcher opened the envelope and proceeded with allocation. Treatment allocation was revealed to the participants after collection of baseline outcomes. We blinded the clinicians who obtained the follow‐up information to the allocation. Five physical therapists provided the manual therapy protocol, and 4 surgeons conducted the surgical procedure."
Randomization performed by computer‐generated randomized table of numbers. Allocation concealment was well described and appropriate. There were no concerns about differences between the study groups at the baseline.

Some concerns

The participants were not blinded. There were small deviations from the intended interventions in both study groups, however we had no reason to believe they occurred due to the trial context. In general the study used intention‐to‐treat analysis, however, the result table for long‐term “improvement” included only participants who were available for the follow‐up. Not sure if the analysis were mITT (because the data for the lost participants were missing) or per‐protocol (excluding participants, who did not receive their assigned interventions, from the analysis).

Low risk of bias

Outcome was available for nearly all participants (less than 10% missing in each of the study groups).

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Low risk of bias

Trial registry record available. Results seemed to be reported as intended by the trial registry record and the methods section of the article.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Open in a new tab

Risk of bias for analysis 5.2 Symptoms at short‐term follow‐up: 3 months or less.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Fernandez‐de‐las‐Penas 2015

Low risk of bias

The participants were not blinded. Due to the nature of the interventions, it is unlikely that the caretakers were blinded. No deviations from the intended intervention at short‐term happened. The analyses were performed according to the intention‐to‐treat principle.

Low risk of bias

No participants were lost to follow‐up at short term.

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Low risk of bias

Trial registry record available. Results reported as intended by the trial registry record and the methods section of the article.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Fernandez‐de‐las‐Penas 2017b

Low risk of bias

Quote: "Patients were randomly assigned to receive manual therapy or surgery. Concealed allocation was conducted using a computer‐generated randomized table of numbers created prior to the start of the data collection by a researcher not involved in subject recruitment. Individual and sequentially numbered index cards with the random assignment were prepared, folded, and placed in sealed opaque envelopes. Another researcher opened the envelope and hence proceeded with treatment according to group assignment. We blinded clinicians who obtained follow‐up information to group allocation."
Randomization performed by computer‐generated randomized table of numbers. Allocation concealment was well described and appropriate. There were no concerns about differences between the study groups at the baseline.

Low risk of bias

No participants were lost to follow‐up at short term.

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Low risk of bias

Trial registry record available. Results reported as intended by the trial registry record and the methods section of the article.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Open in a new tab

Risk of bias for analysis 5.3 Symptoms at long‐term follow‐up: over 3 months.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Fernandez‐de‐las‐Penas 2015

Low risk of bias

The participants were not blinded. Due to the nature of the interventions, it is unlikely that the care takers were blinded. There were small deviations from the intended interventions in both study groups, however we had no reason to believe they occurred due to the trial context. The analyses were performed according to the intention‐to‐treat principle.

Low risk of bias

Outcome was available for nearly all participants (less than 10% missing in each of the study groups).

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Low risk of bias

Trial registry record available. Results reported as intended by the trial registry record and the methods section of the article.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Fernandez‐de‐las‐Penas 2017b

Low risk of bias

Outcome was available for nearly all participants (less than 10% missing in each of the study groups).

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Low risk of bias

Trial registry record available. Results reported as intended by the trial registry record and the methods section of the article.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Open in a new tab

Risk of bias for analysis 5.4 Function at short‐term follow‐up: 3 months or less.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Fernandez‐de‐las‐Penas 2015

Low risk of bias

No participants were lost to follow‐up at short term.

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Low risk of bias

Trial registry record available. Results reported as intended by the trial registry record and the methods section of the article.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Fernandez‐de‐las‐Penas 2017b

Low risk of bias

No participants were lost to follow‐up at short term.

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Low risk of bias

Trial registry record available. Results reported as intended by the trial registry record and the methods section of the article.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Open in a new tab

Risk of bias for analysis 5.5 Function at long‐term follow‐up: over 3 months.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Fernandez‐de‐las‐Penas 2015

Low risk of bias

Outcome was available for nearly all participants (less than 10% missing in each of the study groups).

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Low risk of bias

Trial registry record available. Results reported as intended by the trial registry record and the methods section of the article.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Fernandez‐de‐las‐Penas 2017b

Low risk of bias

Outcome was available for nearly all participants (less than 10% missing in each of the study groups).

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Low risk of bias

Trial registry record available. Results reported as intended by the trial registry record and the methods section of the article.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Open in a new tab

Risk of bias for analysis 5.6 Pain at short‐term follow‐up: 3 months or less.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Fernandez‐de‐las‐Penas 2015

Low risk of bias

No participants were lost to follow‐up at short term.

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Low risk of bias

Trial registry record available. Results reported as intended by the trial registry record and the methods section of the article.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Fernandez‐de‐las‐Penas 2017a

Low risk of bias

Quote: "Patients were randomly assigned to receive manual therapy or surgery. Concealed allocation was conducted using a computer‐generated randomized table of numbers created prior to the start of the data collection by an external researcher not involved in recruitment. Individual and sequentially numbered index cards with random assignment were prepared, folded and placed in sealed opaque envelopes. A second researcher opened the envelope and proceeded with treatment according to group assignment. We blinded clinicians who obtained follow‐up information to group allocation."
Randomization performed by computer‐generated randomized table of numbers. Allocation concealment seemed to be appropriate. There were no concerns about differences between the study groups at the baseline.

Low risk of bias

No deviations from the intended intervention at short‐term happened. The analyses were performed according to the intention‐to‐treat principle.

Low risk of bias

No participants were lost to follow‐up at short term.

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Low risk of bias

Trial registry record available. Results provided in figure only, but for all the intended timepoints.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Open in a new tab

Risk of bias for analysis 5.7 Pain at long‐term follow‐up: over 3 months.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Fernandez‐de‐las‐Penas 2015

Low risk of bias

Outcome was available for nearly all participants (less than 10% missing in each of the study groups).

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Low risk of bias

Trial registry record available. Results reported as intended by the trial registry record and the methods section of the article.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Fernandez‐de‐las‐Penas 2017a

Low risk of bias

Quote: "Patients were randomly assigned to receive manual therapy or surgery. Concealed allocation was conducted using a computer‐generated randomized table of numbers created prior to the start of the data collection by an external researcher not involved in recruitment. Individual and sequentially numbered index cards with random assignment were prepared, folded and placed in sealed opaque envelopes. A second researcher opened the envelope and proceeded with treatment according to group assignment. We blinded clinicians who obtained follow‐up information to group allocation."
Randomization performed by computer‐generated randomized table of numbers. Allocation concealment seemed to be appropriate. There were no concerns about differences between the study groups at the baseline.

Low risk of bias

Due to the nature of interventions the participants and caretakers were not blinded. There were small deviations from the intended interventions in both study groups, however we had no reason to believe they occurred due to the trial context. The analyses were performed according to the intention‐to‐treat principle.

Low risk of bias

Outcome was available for nearly all participants (less than 10% missing in each of the study groups).

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Low risk of bias

Trial registry record available. Results provided in figure only, but for all the intended timepoints.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Open in a new tab

Risk of bias for analysis 5.8 Health‐related quality of life at short‐term follow‐up: 3 months or less.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Fernandez‐de‐las‐Penas 2015

Low risk of bias

No participants were lost to follow‐up at short term.

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Low risk of bias

Trial registry record available. Results provided in figure only, but for all timepoints as planned in the trial registry record.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Open in a new tab

Risk of bias for analysis 5.9 Health‐related quality of life at long‐term follow‐up: over 3 months.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Fernandez‐de‐las‐Penas 2015

Low risk of bias

Outcome was available for nearly all participants (less than 10% missing in each of the study groups).

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Low risk of bias

Trial registry record available. Results provided in figure only, but for all timepoints as planned in the trial registry record.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Open in a new tab

Risk of bias for analysis 5.10 Need for surgery or secondary surgery at long‐term follow‐up: over 3 months.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Fernandez‐de‐las‐Penas 2015

Low risk of bias

The participants were not blinded. Due to the nature of the interventions, it is unlikely that the caretakers were blinded. Need for surgery/repeated surgery was the outcome of interest, and no other deviations from the intended interventions were reported. The analyses were performed according to the intention‐to‐treat principle.

Low risk of bias

Outcome was available for nearly all participants (less than 5% missing in one of the study groups, no loss in the other group).

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Some concerns

Trial registry record available, but the outcome not planned to be reported.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Fernandez‐de‐las‐Penas 2017a

Low risk of bias

Quote: "Patients were randomly assigned to receive manual therapy or surgery. Concealed allocation was conducted using a computer‐generated randomized table of numbers created prior to the start of the data collection by an external researcher not involved in recruitment. Individual and sequentially numbered index cards with random assignment were prepared, folded and placed in sealed opaque envelopes. A second researcher opened the envelope and proceeded with treatment according to group assignment. We blinded clinicians who obtained follow‐up information to group allocation."
Randomization performed by computer‐generated randomized table of numbers. Allocation concealment seemed to be appropriate. There were no concerns about differences between the study groups at the baseline.

Low risk of bias

Due to the nature of interventions the participants and caretakers were not blinded. Need for surgery/repeated surgery was the outcome of interest, and there were small other deviations from the intended interventions, however we had no reason to believe they occurred due to the trial context. The analyses were performed according to the intention‐to‐treat principle.

Low risk of bias

Outcome was available for nearly all participants (less than 5% missing in one of the study groups, no loss in the other group).

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Some concerns

Trial registry record available, but the outcome not planned to be reported.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Fernandez‐de‐las‐Penas 2017b

Low risk of bias

The participants were not blinded. Due to the nature of the interventions, it is unlikely that the caretakers were blinded. Need for surgery/repeated surgery was the outcome of interest, and there were small other deviations from the intended interventions, however we had no reason to believe they occurred due to the trial context. The analyses were performed according to the intention‐to‐treat principle.

Low risk of bias

Outcome was available for nearly all participants (less than 10% missing in one of the study groups, no loss in the other group).

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Some concerns

Trial registry record available, but the outcome not planned to be reported.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Open in a new tab

Risk of bias for analysis 6.1 Clinical improvement at long‐term follow‐up: over 3 months.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Subgroup 6.1.1 All criteria

Jarvik 2009

Low risk of bias

Quote: "A systems analyst with no contact with patients generated the computerised random allocation with block sizes varying between 4 and 12 (concealed from research personnel involved with recruitment), stratified by site. We placed assignments in opaque sealed envelopes opened by research assistants after patients’ baseline questionnaire assessment. We masked staff who obtained follow‐up information to patient allocation. At the beginning of each interview, the interviewer reminded the patient not to reveal his or her treatment group."
Block randomization performed. Allocation concealment was well described and appropriate. There were no concerns about differences between the study groups at the baseline.

Some concerns

Participants and care‐providers were not blinded. There were deviations from the protocol in both study groups, including about 40% of the participants in the non‐operative arm had surgery before the long‐term measurement. However, we had no reason to believe it occurred due to the trial context. The analyses were performed according to the intention‐to‐treat principle.

Some concerns

There were between 10 to 20 % of participants lost to follow‐up in both study groups. Reasons for the dropouts not reported, therefore, not possible to judge, if the missingness was random.

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Some concerns

Protocol was available but it did not provide information about what method for success measurement was used. Results reported according to the definition in the methods section.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Open in a new tab

Risk of bias for analysis 6.2 Symptoms at long‐term follow‐up: over 3 months.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Jarvik 2009

Low risk of bias

Some concerns

There were between 10 to 20 % of participants lost to follow‐up in both study groups. Reasons for the dropouts not reported, therefore, not possible to judge, if the missingness was random.

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Low risk of bias

Protocol available. Results reported as intended by the protocol and the methods section of the article.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Open in a new tab

Risk of bias for analysis 6.3 Function at long‐term follow‐up: over 3 months.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Jarvik 2009

Low risk of bias

Some concerns

There were between 10 to 20 % of participants lost to follow‐up in both study groups. Reasons for the dropouts not reported, therefore, not possible to judge, if the missingness was random.

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Low risk of bias

Protocol available. Results reported as intended by the protocol and the methods section of the article.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Open in a new tab

Risk of bias for analysis 6.4 Pain at long‐term follow‐up: over 3 months.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Jarvik 2009

Low risk of bias

Some concerns

There were between 10 to 20 % of participants lost to follow‐up in both study groups. Reasons for the dropouts not reported, therefore, not possible to judge, if the missingness was random.

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Some concerns

Protocol available, but does not provide information on what method for pain measurement used. Results reported as intended by the methods section of the article.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Open in a new tab

Risk of bias for analysis 6.5 Health‐related quality of life at long‐term follow‐up: over 3 months.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Subgroup 6.5.1 Physical component summary

Jarvik 2009

Low risk of bias

Some concerns

There were between 10 to 20 % of participants lost to follow‐up in both study groups. Reasons for the dropouts not reported, therefore, not possible to judge, if the missingness was random.

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Low risk of bias

Protocol available. Results reported as intended by the protocol and the methods section of the article.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Subgroup 6.5.2 Mental component summary

Jarvik 2009

Low risk of bias

Some concerns

There were between 10 to 20 % of participants lost to follow‐up in both study groups. Reasons for the dropouts not reported, therefore, not possible to judge, if the missingness was random.

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Low risk of bias

Protocol available. Results reported as intended by the protocol and the methods section of the article.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Open in a new tab

Risk of bias for analysis 6.6 Need for surgery or secondary surgery at long‐term follow‐up: over 3 months.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Jarvik 2009

Low risk of bias

Some concerns

There were between 10 to 20 % of participants lost to follow‐up in both study groups. Reasons for the dropouts not reported, therefore, not possible to judge, if the missingness was random.

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Low risk of bias

Outcome analysed in accordance with the protocol.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Open in a new tab

Risk of bias for analysis 7.1 Symptoms at long‐term follow‐up: over 3 months.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Eltabl 2020

Some concerns

Low risk of bias

The study did not provide information about blinding the participants and caretakers, but due to the nature of interventions it is unlikely they were blinded. The study did not provide any information regarding deviations from the intended interventions, and we did not have any reason to believe such deviations happened. The analysis seemed likely done appropriately.

Low risk of bias

No loss to follow up reported.

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome, as well as due to concerns regarding the method of analysis.

Open in a new tab

Risk of bias for analysis 7.2 Function at long‐term follow‐up: over 3 months.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Eltabl 2020

Some concerns

Low risk of bias

The study did not provide information about blinding the participants and caretakers, but due to the nature of interventions it is unlikely they were blinded. The study did not provide any information regarding deviations from the intended interventions, and we did not have any reason to believe such deviations happened. The analysis seemed likely done appropriately.

Low risk of bias

No loss to follow up reported.

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Some concerns

Trial registry record available, however it is retrospective. Results seemed to be reported as intended by the methods section of the article.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Open in a new tab

Risk of bias for analysis 7.3 Pain at long‐term follow‐up: over 3 months.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Eltabl 2020

Some concerns

Low risk of bias

The study did not provide information about blinding the participants and caretakers, but due to the nature of interventions it is unlikely they were blinded. The study did not provide any information regarding deviations from the intended interventions, and we did not have any reason to believe such deviations happened. The analysis seemed likely done appropriately.

Low risk of bias

No loss to follow up reported.

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Some concerns

Trial registry record available, however it is retrospective. Results reported as intended by the registry record.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Open in a new tab

Risk of bias for analysis 8.1 Symptoms at short‐term follow‐up: 3 months or less.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Zhang 2019

Low risk of bias

Quote: "A random number table was generated by computer and the random numbers were divided into two groups, with odd numbers into Group A and even numbers into Group B. We wrote the random number and the allocation result in sealed numbered envelopes orderly, only to open one once a patient has been recruited and consented."
Randomisation performed by computer‐generated random number table. Allocation concealment seemed appropriate. There were no concerns about differences between the study groups at the baseline.

Low risk of bias

Participants and care‐providers were not blinded. The study did not provide information about blinding the caretakers, but due to the nature of interventions it is unlikely they were blinded. The study did not provide any information regarding deviations from the intended interventions, and we did not have any reason to believe such deviations happened. The analysis seemed likely done appropriately.

Some concerns

There were 10 to 20 % of participants lost in the group receiving surgery together with steroid injection, and less than 10% in the group receiving surgery alone. Reasons for the dropouts not reported, therefore, not possible to judge, if the missingness was random.

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Low risk of bias

Trial registry record available, however it is retrospective. Results reported as intended by the methods section of the article.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Open in a new tab

Risk of bias for analysis 8.2 Function at short‐term follow‐up: 3 months or less.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Zhang 2019

Low risk of bias

Participants were not blinded. The study did not provide information about blinding the caretakers, but due to the nature of interventions it is unlikely they were blinded. The study did not provide any information regarding deviations from the intended interventions, and we did not have any reason to believe such deviations happened. The analysis seemed likely done appropriately.

Some concerns

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Low risk of bias

Trial registry record available, however it is retrospective. Results reported as intended by the methods section of the article.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Open in a new tab

Risk of bias for analysis 8.3 Need for surgery or secondary surgery at short‐term follow‐up: 3 months or less.

Study

Bias

Randomisation process

Deviations from intended interventions

Missing outcome data

Measurement of the outcome

Selection of the reported results

Overall

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Authors' judgement

Support for judgement

Zhang 2019

Low risk of bias

Some concerns

High risk of bias

Outcome was self‐reported, and participants were not blinded, thus we could not exclude that the result was influenced by participants' beliefs in intervention (e.g., surgery effect).

Some concerns

Trial registry record available, however it is retrospective.

High risk of bias

Overall risk of bias judged as high due to the lack of blinding, which could have an impact on measurement of the outcome.

Open in a new tab

Acknowledgements

We acknowledge the authors of earlier versions of this review.

We thank Ruth Brassington for her assistance in the planning and writing of this review; Farhad Shokraneh from the Cochrane Neuromuscular Disease Group for his assistance in database searches strategy; as well as Venla‐Linnea Karjalainen for checking the correctness of the data. We thank the peer reviewers for their helpful comments.

Editorial and peer reviewer contributions

Cochrane Neuromuscular supported the authors in the development of this review. The following people conducted the editorial process for this article:

Sign‐off Editor (final editorial decision): Colin Chalk, Professor, Department of Neurology & Neurosurgery, McGill University.
Managing Editor (selected peer reviewers, provided editorial guidance to authors, edited the article): Luisa Fernandez Mauleffinch, Cochrane Central Editorial Service.
Editorial Assistant (conducted editorial policy checks, collated peer reviewer comments and supported the editorial team): Leticia Rodrigues, Cochrane Central Editorial Service.
Copy Editor (copy editing and production): Anne Lethaby, Cochrane Central Production Service.
Peer reviewers (provided comments and recommended an editorial decision): Jennifer Hilgart, Cochrane (methods); Steve McDonald, Cochrane Australia (search); Pietro Caliandro, Fondazione Policlinico Universitario Agostino Gemelli IRCCS (clinical); and Janet Wale, independent consumer advocate (consumer). One additional peer reviewer provided clinical peer review but chose not to be publicly acknowledged.

Appendices

Appendix 1. Cochrane Neuromuscular's Specialised Register

1 MeSH DESCRIPTOR Carpal Tunnel Syndrome AND INREGISTER 376

2 "carpal tunnel" AND INREGISTER 595

3 ("nerve entrapment" or "nerve compression" or "entrapment neuropath*") and carpal AND INREGISTER 43

4 #1 OR #2 OR #3 595

Appendix 2. Cochrane Central Register of Controlled Trials (CENTRAL)

Date Run: 18/11/2022 22:14:34

#1 [mh "Carpal Tunnel Syndrome"] OR "Carpal Tunnel" OR (("Nerve Entrapment" OR "Nerve Compression" OR "Entrapment Neuropathy" OR "Entrapment Neuropathies") AND Carpal) with Cochrane Library publication date Between Oct 2021 and Nov 2022, in Trials 178

Appendix 3. MEDLINE search strategy

Database: Ovid MEDLINE(R) ALL <1946 to November 17, 2022>

1 ((Randomized Controlled Trial or Controlled Clinical Trial).pt. or (Randomi?ed or Placebo or Randomly or Trial or Groups).ab. or Drug Therapy.fs.) not (exp Animals/ not Humans.sh.) (4825120)

2 (clinical trial.pt. or exp clinical trial/ or (clin$ adj25 trial$).ti,ab. or ((singl$ or doubl$ or tripl$ or trebl$) adj25 (blind$ or mask$)).ti,ab. or placebos/ or placebo$.ti,ab. or random$.ti,ab. or research design/) not (exp animals/ not humans.sh.) (2121256)

3 (comparative study/ or exp evaluation studies/ or follow up studies/ or prospective studies/ or (control$ or prospectiv$ or volunteer$).ti,ab.) not (exp animals/ not humans.sh.) (6070680)

4 1 or 2 or 3 (9311989)

5 Carpal Tunnel Syndrome.mp. or Carpal Tunnel Syndrome/ or (carp$ tunn$ or tunn$ syndrom$).mp. or (nerve entrapment or nerve compression or entrapment neuropath$).mp. (27618)

6 (epineurotomy or reconstruct$ or release).mp. or SURGERY/ or surgery.mp. or SURGICAL PROCEDURES, OPERATIVE/ or surgical.mp. or (splint or splints or splinting).mp. or exp Anti‐Inflammatory Agents, Non‐Steroidal/ or non‐steroidal anti‐inflammatory.mp. or NSAID$.mp. or ((corticosteroid$ or steroid$) and injection$).mp. or diuretic$.mp. or exp DIURETICS/ (4539481)

7 4 and 5 and 6 (4785)

8 limit 7 to ed=20211002‐20221231 (240)

9 limit 7 to dt=20211002‐20221231 (225)

10 8 or 9 (345)

Appendix 4. Embase search strategy

Database: Embase <1974 to 2022 Week 45>

1 Randomized controlled trial/ or Controlled clinical study/ or randomization/ or intermethod comparison/ or double blind procedure/ or human experiment/ or (random$ or placebo or (open adj label) or ((double or single or doubly or singly) adj (blind or blinded or blindly)) or parallel group$1 or crossover or cross over or ((assign$ or match or matched or allocation) adj5 (alternate or group$1 or intervention$1 or patient$1 or subject$1 or participant$1)) or assigned or allocated or (controlled adj7 (study or design or trial)) or volunteer or volunteers).ti,ab. or (compare or compared or comparison or trial).ti. or ((evaluated or evaluate or evaluating or assessed or assess) and (compare or compared or comparing or comparison)).ab. (5965910)

2 (random$ adj sampl$ adj7 ("cross section$" or questionnaire$1 or survey$ or database$1)).ti,ab. not (comparative study/ or controlled study/ or randomi?ed controlled.ti,ab. or randomly assigned.ti,ab.) (9194)

3 Cross‐sectional study/ not (randomized controlled trial/ or controlled clinical study/ or controlled study/ or randomi?ed controlled.ti,ab. or control group$1.ti,ab.) (327035)

4 (((case adj control$) and random$) not randomi?ed controlled).ti,ab. (20421)

5 (Systematic review not (trial or study)).ti. (227538)

6 (nonrandom$ not random$).ti,ab. (18220)

7 ("Random field$" or (random cluster adj3 sampl$)).ti,ab. (4285)

8 (review.ab. and review.pt.) not trial.ti. (1040153)

9 "we searched".ab. and (review.ti. or review.pt.) (44480)

10 ("update review" or (databases adj4 searched)).ab. (54966)

11 (rat or rats or mouse or mice or swine or porcine or murine or sheep or lambs or pigs or piglets or rabbit or rabbits or cat or cats or dog or dogs or cattle or bovine or monkey or monkeys or trout or marmoset$1).ti. and animal experiment/ (1173407)

12 Animal experiment/ not (human experiment/ or human/) (2463882)

13 2 or 3 or 4 or 5 or 6 or 7 or 8 or 9 or 10 or 11 or 12 (4091434)

14 Clinical Trial/ or Multicenter Study/ or Controlled Study/ or Crossover Procedure/ or Single Blind Procedure/ or Major Clinical Study/ or PLACEBO/ or Meta Analysis/ or phase 2 clinical trial/ or phase 3 clinical trial/ or phase 4 clinical trial/ or ((clin$ adj25 trial$) or ((singl$ or doubl$ or tripl$ or trebl$) adj25 (blind$ or mask$)) or placebo$ or control$ or (meta?analys$ or systematic review$) or (cross?over or factorial or sham? or dummy) or ABAB design$).tw. (15292689)

15 1 or 14 (16934551)

16 15 not 13 (14371426)

17 Carpal Tunnel Syndrome/ or (carpal tunnel syndrome or carp$ tunn$ or tunn$ syndrom$ or nerve entrapment or nerve compression or entrapment neuropath$).mp. (37386)

18 carpal tunnel release/ or epineurotomy/ or surgical approach/ or surgical technique/ or exp Nonsteroid Antiinflammatory Agent/ or exp Diuretic Agent/ or (epineurotomy or surgery or surgical or operation or reconstruct$ or splint or splints or splinting or non‐steroid$ anti‐inflammatory or NSAID$ or ((corticosteroid$ or steroid$) and injection$) or diuretic$).mp. (6282409)

19 16 and 17 and 18 (6147)

20 limit 19 to (conference abstracts or embase) (5506)

21 limit 20 to em=202140‐202245 (480)

Appendix 5. ClinicalTrials.Gov

Advanced Search

Condition or disease: Carpal Tunnel Syndrome

Study type: Interventional Studies (Clinical Trials)

First posted on or after 04/10/2021

51 studies found

Appendix 6. WHO International Clinical Trials Registry Platform (ICTRP)

Advanced Search

Carpal Tunnel Syndrome in the Condition

Recruitment status is ALL

Date of registration is between 04/10/2021 and 18/11/2022

66 records for 66 trials found

Appendix 7. Searches in 2008

We searched Cochrane Neuromuscular's Specialised Register using 'median nerve entrapment', 'carpal tunnel syndrome' and 'entrapment neuropathy' as the search terms. We originally searched MEDLINE, Embase and the LILACS database. LILACS is a specialised database, supported by the Pan‐American Health Organisation, aiming to collect all biomedical literature published in Latin America. We updated the search of the Cochrane Neuromuscular Disease Group Trials Register (January 2008), MEDLINE (January 1966 to January 2008), Embase (January 1980 to January 2008) and LILACS (January 1982 to January 2008), which revealed two further relevant trials.

MEDLINE search strategy

1 randomized controlled trial.pt. 2 controlled clinical trial.pt. 3 randomized controlled trials/ 4 random allocation/ 5 double‐blind method/ 6 single‐blind method/ 7 or/1‐6 8 animals/ not humans/ 9 7 not 8 10 clinical trial.pt. 11 exp clinical trials/ 12 (clin$ adj25 trial$).ti,ab. 13 ((singl$ or doubl$ or tripl$ or trebl$) adj25 (blind$ or mask$)).ti,ab. 14 placebos/ 15 placebo$.ti,ab. 16 random$.ti,ab. 17 research design/ 18 or/10‐17 19 18 not 8 20 19 not 9 21 comparative study/ 22 exp evaluation studies/ 23 follow up studies/ 24 prospective studies/ 25 (control$ or prospectiv$ or volunteer$).ti,ab. 26 or/21‐25 27 26 not 8 28 27 not (9 or 20) 29 9 or 20 or 28 30 Carpal Tunnel Syndrome.mp. or Carpal Tunnel Syndrome/ 31 (carp$ tunn$ or tunn$ syndrom$).mp 32 (nerve entrapment or nerve compression or entrapment neuropath$).mp 33 or/30‐32 34 epineurotomy.mp. 35 reconstruct$.mp. 36 release.mp. 37 SURGERY/ or surgery.mp. 38 SURGICAL PROCEDURES, OPERATIVE/ or surgical.mp. 39 (splint or splints or splinting).mp. 40 exp Anti‐Inflammatory Agents, Non‐Steroidal/ or non‐steroidal anti‐inflammatory.mp. 41 NSAID$.mp. 42 ((corticosteroid$ or steroid$) and injection$).mp. [mp=title, original title, abstract, name of substance word, subject heading word] 43 diuretic$.mp. or exp DIURETICS/ 44 or/34‐43 45 33 and 44 46 29 and 45

Embase search strategy

1 Randomized Controlled Trial/ 2 Clinical Trial/ 3 Multicenter Study/ 4 Controlled Study/ 5 Crossover Procedure/ 6 Double Blind Procedure/ 7 Single Blind Procedure/ 8 exp RANDOMIZATION/ 9 Major Clinical Study/ 10 PLACEBO/ 11 Meta Analysis/ 12 phase 2 clinical trial/ or phase 3 clinical trial/ or phase 4 clinical trial/ 13 (clin$ adj25 trial$).tw. 14 ((singl$ or doubl$ or tripl$ or trebl$) adj25 (blind$ or mask$)).tw. 15 placebo$.tw. 16 random$.tw. 17 control$.tw. 18 (meta?analys$ or systematic review$).tw. 19 (cross?over or factorial or sham? or dummy).tw. 20 ABAB design$.tw. 21 or/1‐20 22 human/ 23 nonhuman/ 24 22 or 23 25 21 not 24 26 21 and 22 27 25 or 26 28 carpal tunnel syndrome.mp. or Carpal Tunnel Syndrome/ 29 (carp$ tunn$ or tunn$ syndrom$).mp. 30 (nerve entrapment or nerve compression or entrapment neuropath$).mp. 31 or/28‐30 32 epineurotomy.mp. or carpal tunnel release/ or epineurotomy/ 33 surgical approach/ or surgical technique/ 34 (surgery or surgical or operation or reconstruct$).mp 35 33 or 34 36 (splint or splints or splinting).mp. 37 exp Nonsteroid Antiinflammatory Agent/ or non‐steroid$ anti‐inflammatory.mp. 38 NSAID$.mp. 39 ((corticosteroid$ or steroid$) and injection$).mp. 40 diuretic$.mp. or exp Diuretic Agent/ 41 or/32‐40 42 31 and 41 43 27 and 42

LILACS search strategy

median nerve entrapment OR carpal tunnel syndrome OR entrapment neuropathy [Words] and ((Pt randomized controlled trial OR Pt controlled clinical trial OR Mh randomized controlled trials OR Mh random allocation OR Mh double‐blind method OR Mh single‐blind method) AND NOT (Ct animal AND NOT (Ct human and Ct animal)) OR (Pt clinical trial OR Ex E05.318.760.535$ OR (Tw clin$ AND (Tw trial$ OR Tw ensa$ OR Tw estud$ OR Tw experim$ OR Tw investiga$)) OR ((Tw singl$ OR Tw simple$ OR Tw doubl$ OR Tw doble$ OR Tw duplo$ OR Tw trebl$ OR Tw trip$) AND (Tw blind$ OR Tw cego$ OR Tw ciego$ OR Tw mask$ OR Tw mascar$)) OR Mh placebos OR Tw placebo$ OR (Tw random$ OR Tw randon$ OR Tw casual$ OR Tw acaso$ OR Tw azar OR Tw aleator$) OR Mh research design) AND NOT (Ct animal AND NOT (Ct human and Ct animal)) OR (Ct comparative study OR Ex E05.337$ OR Mh follow‐up studies OR Mh prospective studies OR Tw control$ OR Tw prospectiv$ OR Tw volunt$ OR Tw volunteer$) AND NOT (Ct animal AND NOT (Ct human and Ct animal))) [Words] and epineurotomy OR carpal tunnel release OR epineurotomy OR surg$ OR operation OR reconstruct$ OR splint$ OR non‐steroid$ anti‐inflammatory OR NSAID OR ((corticosteroid$ OR steroid$) AND injection$) OR diuretic$ [Words]

Data and analyses

Comparison 1. Surgery versus splint.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1.1 Clinical improvement at short‐term follow‐up: 3 months or less	2	198	Risk Ratio (IV, Random, 95% CI)	1.06 [0.48, 2.34]
1.2 Clinical improvement at long‐term follow‐up: over 3 months	3	210	Risk Ratio (IV, Random, 95% CI)	2.10 [1.04, 4.24]
1.3 Symptoms at short‐term follow‐up: 3 months or less	2	198	Mean Difference (IV, Random, 95% CI)	0.02 [‐0.83, 0.87]
1.4 Symptoms at long‐term follow‐up: over 3 months	2	195	Mean Difference (IV, Random, 95% CI)	‐0.26 [‐0.52, 0.00]
1.5 Function at short‐term follow‐up: 3 months or less	2	198	Mean Difference (IV, Random, 95% CI)	0.00 [‐0.44, 0.44]
1.6 Function at long‐term follow‐up: over 3 months	2	195	Mean Difference (IV, Random, 95% CI)	‐0.36 [‐0.62, ‐0.09]
1.7 Health‐related quality of life at long‐term follow‐up: over 3 months	1	167	Mean Difference (IV, Random, 95% CI)	0.04 [‐0.00, 0.08]
1.8 Adverse effects	2	210	Risk Ratio (M‐H, Random, 95% CI)	2.11 [0.37, 12.12]
1.9 Need for surgery or secondary surgery at short‐term follow‐up: 3 months or less	1	164	Risk Ratio (M‐H, Random, 95% CI)	0.08 [0.00, 1.48]
1.10 Need for surgery or secondary surgery at long‐term follow‐up: over 3 months	2	176	Risk Ratio (M‐H, Random, 95% CI)	0.03 [0.00, 0.21]

Open in a new tab

Comparison 2. Surgery versus corticosteroid injection.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
2.1 Clinical improvement at short‐term follow‐up: 3 months or less	2	185	Risk Ratio (IV, Random, 95% CI)	0.75 [0.60, 0.92]
2.2 Clinical improvement at long‐term follow‐up: over 3 months	3	187	Risk Ratio (IV, Random, 95% CI)	1.23 [0.73, 2.06]
2.3 Symptoms at short‐term follow‐up: 3 months or less	2	90	Mean Difference (IV, Random, 95% CI)	‐11.32 [‐21.76, ‐0.89]
2.4 Symptoms at long‐term follow‐up: over 3 months	2	118	Std. Mean Difference (IV, Random, 95% CI)	‐0.60 [‐1.88, 0.69]
2.5 Function at short‐term follow‐up: 3 months or less	1	147	Mean Difference (IV, Random, 95% CI)	11.00 [4.80, 17.20]
2.6 Function at long‐term follow‐up: over 3 months	2	191	Std. Mean Difference (IV, Random, 95% CI)	‐0.12 [‐0.80, 0.56]
2.7 Pain at short‐term follow‐up: 3 months or less	1	147	Mean Difference (IV, Random, 95% CI)	9.00 [2.79, 15.21]
2.8 Pain at long‐term follow‐up: over 3 months	1	123	Mean Difference (IV, Random, 95% CI)	‐6.00 [‐10.45, ‐1.55]
2.9 Adverse effects	2	90	Risk Ratio (M‐H, Random, 95% CI)	1.49 [0.25, 8.70]
2.10 Need for surgery or secondary surgery at short‐term follow‐up: 3 months or less	1	163	Risk Ratio (M‐H, Random, 95% CI)	2.08 [0.19, 22.44]
2.11 Need for surgery or secondary surgery at long‐term follow‐up: over 3 months	1	163	Risk Ratio (M‐H, Random, 95% CI)	0.61 [0.25, 1.46]

Open in a new tab

Comparison 3. Surgery versus splint + steroid injection.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
3.1 Clinical improvement at short‐term follow‐up: 3 months or less	1	34	Risk Ratio (IV, Random, 95% CI)	0.47 [0.25, 0.87]
3.2 Clinical improvement at long‐term follow‐up: over 3 months	1	34	Risk Ratio (IV, Random, 95% CI)	1.10 [0.84, 1.43]
3.3 Symptoms at short‐term follow‐up: 3 months or less	1	34	Mean Difference (IV, Random, 95% CI)	0.45 [0.07, 0.83]
3.4 Symptoms at long‐term follow‐up: over 3 months	1	34	Mean Difference (IV, Random, 95% CI)	‐0.55 [‐0.87, ‐0.23]
3.5 Function at short‐term follow‐up: 3 months or less	1	34	Mean Difference (IV, Random, 95% CI)	0.53 [0.13, 0.93]
3.6 Function at long‐term follow‐up: over 3 months	1	34	Mean Difference (IV, Random, 95% CI)	‐0.17 [‐0.41, 0.07]
3.7 Adverse effects	1	34	Risk Ratio (M‐H, Random, 95% CI)	10.00 [0.52, 192.25]

Open in a new tab

Comparison 4. Surgery versus PRP injection.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
4.1 Symptoms at long‐term follow‐up: over 3 months	1	60	Mean Difference (IV, Random, 95% CI)	0.00 [‐0.35, 0.35]
4.2 Function at long‐term follow‐up: over 3 months	1	60	Mean Difference (IV, Random, 95% CI)	‐0.33 [‐0.35, ‐0.31]
4.3 Pain at long‐term follow‐up: over 3 months	1	60	Mean Difference (IV, Random, 95% CI)	0.43 [0.03, 0.83]

Open in a new tab

Comparison 5. Surgery versus manual therapy.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
5.1 Clinical improvement at long‐term follow‐up: over 3 months	1	111	Risk Ratio (IV, Random, 95% CI)	1.31 [1.09, 1.57]
5.2 Symptoms at short‐term follow‐up: 3 months or less	2	220	Mean Difference (IV, Random, 95% CI)	‐0.08 [‐0.28, 0.11]
5.3 Symptoms at long‐term follow‐up: over 3 months	2	220	Mean Difference (IV, Random, 95% CI)	‐0.09 [‐0.29, 0.10]
5.4 Function at short‐term follow‐up: 3 months or less	2	220	Mean Difference (IV, Random, 95% CI)	0.22 [0.03, 0.41]
5.5 Function at long‐term follow‐up: over 3 months	2	220	Mean Difference (IV, Random, 95% CI)	‐0.04 [‐0.19, 0.11]
5.6 Pain at short‐term follow‐up: 3 months or less	2	220	Mean Difference (IV, Random, 95% CI)	1.46 [0.96, 1.96]
5.7 Pain at long‐term follow‐up: over 3 months	2	220	Mean Difference (IV, Random, 95% CI)	0.05 [‐0.42, 0.53]
5.8 Health‐related quality of life at short‐term follow‐up: 3 months or less	1	120	Mean Difference (IV, Random, 95% CI)	‐0.10 [‐0.12, ‐0.08]
5.9 Health‐related quality of life at long‐term follow‐up: over 3 months	1	118	Mean Difference (IV, Random, 95% CI)	‐0.02 [‐0.04, 0.00]
5.10 Need for surgery or secondary surgery at long‐term follow‐up: over 3 months	3	320	Risk Ratio (M‐H, Random, 95% CI)	0.42 [0.07, 2.63]

Open in a new tab

Comparison 6. Surgery versus multimodal non‐operative treatment.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
6.1 Clinical improvement at long‐term follow‐up: over 3 months	1		Risk Ratio (IV, Random, 95% CI)	Subtotals only
6.1.1 All criteria	1	101	Risk Ratio (IV, Random, 95% CI)	1.67 [0.97, 2.88]
6.2 Symptoms at long‐term follow‐up: over 3 months	1	101	Mean Difference (IV, Random, 95% CI)	‐0.33 [‐0.65, ‐0.01]
6.3 Function at long‐term follow‐up: over 3 months	1	101	Mean Difference (IV, Random, 95% CI)	‐0.43 [‐0.77, ‐0.09]
6.4 Pain at long‐term follow‐up: over 3 months	1	101	Mean Difference (IV, Random, 95% CI)	‐0.80 [‐2.03, 0.43]
6.5 Health‐related quality of life at long‐term follow‐up: over 3 months	1		Mean Difference (IV, Random, 95% CI)	Subtotals only
6.5.1 Physical component summary	1	101	Mean Difference (IV, Random, 95% CI)	2.00 [‐3.10, 7.10]
6.5.2 Mental component summary	1	101	Mean Difference (IV, Random, 95% CI)	‐2.00 [‐7.85, 3.85]
6.6 Need for surgery or secondary surgery at long‐term follow‐up: over 3 months	1	116	Risk Ratio (M‐H, Random, 95% CI)	0.02 [0.00, 0.35]

Open in a new tab

Comparison 7. Surgery versus unspecified medical treatment and hand support.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
7.1 Symptoms at long‐term follow‐up: over 3 months	1	60	Mean Difference (IV, Random, 95% CI)	Not estimable
7.2 Function at long‐term follow‐up: over 3 months	1	60	Mean Difference (IV, Random, 95% CI)	‐1.74 [‐1.91, ‐1.57]
7.3 Pain at long‐term follow‐up: over 3 months	1	60	Mean Difference (IV, Random, 95% CI)	‐5.34 [‐5.94, ‐4.74]

Open in a new tab

Comparison 8. Surgery + corticosteroid injection versus corticosteroid injection.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
8.1 Symptoms at short‐term follow‐up: 3 months or less	1	46	Mean Difference (IV, Random, 95% CI)	‐0.22 [‐0.35, ‐0.09]
8.2 Function at short‐term follow‐up: 3 months or less	1	46	Mean Difference (IV, Random, 95% CI)	‐0.28 [‐0.46, ‐0.10]
8.3 Need for surgery or secondary surgery at short‐term follow‐up: 3 months or less	1	46	Risk Ratio (M‐H, Random, 95% CI)	0.33 [0.01, 7.78]

Open in a new tab

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Awan 2015.

*Study characteristics*
Methods	Study design: randomised control trial Setting: Department of Orthopedics and Department of Neurosurgery, Ayub Teaching Hospital, Abbottabad, Pakistan
Participants	Details of sampling frame: Total n eligible = not reported Total n excluded pre‐randomisation = not reported Total n randomised = 116 Surgery, n randomised = 58 Local steroid injection, n randomised = 58 Surgery, n available for follow‐up = 58 Local steroid injection, n available for follow‐up = 58 Gender distribution: Surgery: 8 males, 50 females Local steroid injection: 9 males, 49 females Mean ± SD age: Surgery: 32.2 ± 5.1 years Local steroid injection: 33.4 ± 5.1 years Mean ± SD duration of symptoms: Not reported Inclusion criteria: All male and female patients of any age group: with CTS diagnosed clinically by both Phalen's test & Tinel’s sign; and having moderate (grade 2) to severe (grade 3) pain. Exclusion criteria: Diabetes mellitus (diagnosed by history and fasting blood sugar of more than 126 mg/dL) History of trauma Hypo/hyperthyroidism (diagnosed by TSH, T3 and T4 levels) Rheumatoid arthritis (diagnosed by history and positive RA factor) CTS diagnostic criteria (case definition): CTS diagnosed clinically by both Phalen's test & Tinel’s sign Pain assessment was done using VAS.
Interventions	Surgical decompression ‐ mini incision. Through a 1 cm mid palmer longitudinal incision made in the axis of the radial border of the ring finger, the median nerve was visualised through sharp dissection and the transverse carpal ligament was transacted. The skin was closed using non‐absorbable sutures. Local steroid injection ‐ 20 mg in 1 mL injection of methyl prednisolone injected 1 cm distal proximal to the distal wrist flexion crease and medial to Palmaris longus tendon at a 45‐degree angle distally.
Outcomes	Outcomes were assessed at baseline and 1 month after treatment. Pain (VAS) Effectiveness, which was assessed in terms of improvement in at least two grade scores on the VAS
Funding	Not reported
COI	Not reported
Notes

Open in a new tab

Eltabl 2020.

*Study characteristics*
Methods	Study design: randomised clinical trial Setting: outpatient clinic of Neurosurgery and Physical Medicine, Rheumatology, and Rehabilitation Department at Menoufia University Hospitals, Shebeen el kom, Egypt
Participants	Details of sampling frame: Total n eligible = not reported Total n excluded pre‐randomisation = not reported Total n randomised = 90 Surgical procedure, n randomised = 30 Platelet rich plasma (PRP) injection, n randomised = 30 Conventional treatment, n randomised = 30 Surgical procedure, n available for follow‐up = 30 Platelet rich plasma (PRP) injection, n available for follow‐up = 30 Conventional treatment, n available for follow‐up = 30 Gender distribution: Surgical procedure: 12 males, 18 females PRP injection: 11 males, 19 females Conventional treatment: 13 males, 17 females Mean ± SD age: Surgical procedure: 39.8 ± 7.39 PRP injection: 37.93 ± 7.4 Conventional treatment: 39.8 ± 7.39 Mean ± SD duration of symptoms: Not reported Inclusion criteria: Confirmed diagnosis of moderate CTS based on: History (paraesthesia or dysaesthesia and painful swelling of the hand with clumsiness due to weakness that is exacerbated by repetitive use or sleep and improved by shaking the hand) Physical examination (sensory loss and numbness in the areas of the hand, innervated by the median nerve, positive Phalen’s test, and/or Tinel’s sign) Electrophysiological studies Exclusion criteria: Pregnancy History of underlying metabolic diseases (such as diabetes mellitus, thyroid diseases, rheumatoid arthritis) History of local corticosteroid injection in the past 3 months Atrophy of thenar muscles Previous carpal tunnel release surgery Evidence of concomitant neuropathy or radiculopathy Patients with PRP contraindications, including history of malignancies, autoimmune or haematologic disorders, and NSAID consumption 2 days prior to injection, treatment with antiplatelet and anticoagulant agents, haemoglobin level under 12 g/dL, and platelet count under 150,000 in mL Severe CTS in ENMG CTS diagnostic criteria (case definition): Mild CTS was defined as sensory latency of longer than 3.6 ms with or without reduced amplitude of sensory nerve action potential. Moderate CTS was defined as sensory latency of longer than 3.6 ms plus a prolonged motor latency (4.3–6 ms) with normal motor amplitude. The study selected mild to moderate CTS patients with sensory latency of longer than 3.6 ms plus a prolonged motor latency (4.3–6 ms) with normal motor and sensory amplitude.
Interventions	Carpal tunnel release Single ultrasound guided PRP injection ‐ PRP was prepared by taking 10 mL of venous blood sample from the patient in a sterile sodium citrated tube. Then, the tubes with citrated blood were centrifuged at 3500 revolutions per minute (rpm) for 9 min to separate erythrocytes and produce about 1–2 mL of PRP for injection Medical treatment (conventional medical treatment* and hand support)
Outcomes	Outcomes were assessed at a baseline and at 6 months post‐intervention. BCTQ (Italian version) symptom severity score (1 to 5, higher is worse) BCTQ (Italian version) functional status score (1 to 5, higher is worse) Pain (VAS, 0 = no pain to 10 = agonising pain) Electrophysiological results (peak latency and the onset latency of sensory nerve action potential; distal motor latency of the compound muscle action potential)***
Funding	Not applicable
COI	The authors declared that they have no competing interests in this article.
Notes	In the study registration (NCT04235426) conventional medical treatment defined as NSAIDs, diclofenac 150 mg/day for 2 weeks and 1500 g vitamin B12 per day for 6 weeks. The authors reported SDs that considerably deviated from the other studies, therefore, we performed a sensitivity analysis imputing SD from Jarvik 2009. **Electrophysiological results not included in analysis of this review.

Open in a new tab

Fernandez‐de‐las‐Penas 2015.

*Study characteristics*
Methods	Study design: pragmatic, randomised, parallel‐group clinical trial Setting: public hospital and 2 physical therapy practices in Madrid, Spain; study conducted between February 2013 and January 2014, with 1 year follow‐up in January 2015
Participants	Details of sampling frame: Total n assessed for eligibility = 200 Total n excluded pre‐randomisation = 80 Total n randomised = 120 (120 hands)* Surgical group, n randomised = 60 (60 hands) Physical therapy group, n randomised = 60 (60 hands) Surgical group, n available for follow‐up at 1 m = 60 (60 hands) Physical therapy group, n available for follow‐up at 1 m = 60 (60 hands) Surgical group, n available for follow‐up at 3 m = 60 (60 hands) Physical therapy group, n available for follow‐up at 3 m = 60 (60 hands) Surgical group, n available for follow‐up at 6 m = 60 (60 hands) Physical therapy group, n available for follow‐up at 6 m = 58 (58 hands) Surgical group, n available for follow‐up at 12 m = 56 (56 hands) Physical therapy group, n available for follow‐up at 12 m = 55 (55 hands) Surgical group, n available for follow‐up at 4 years = 48 (48 hands) Physical therapy group, n available for follow‐up at 4 years = 49 (49 hands) Gender distribution: Surgical group: 0 males, 60 females Physical therapy group: 0 males, 60 females Mean ± SD age: Surgical group: 46 ± 9 Physical therapy group: 47 ± 10 Mean ± SD duration of symptoms: Surgical group: 3.5 ± 3.1 years of pain Physical therapy group: 3.1 ± 2.7 years of pain Inclusion criteria: Female All clinical signs: pain and paraesthesia in the median nerve distribution, increasing symptoms during the night, positive Tinel's sign, and positive Phalen's test Symptoms had to have persisted for at least 12 months Electrodiagnostic examination had to reveal deficits of sensory and motor median nerve conduction: median nerve sensory conduction velocity < 40 m/s and median nerve distal motor latency > 4.20 milliseconds. Exclusion criteria: Any sensory/motor deficit in the ulnar or radial nerves Age > 65 years Previous hand surgery or steroid injections treatment Multiple diagnoses on the upper extremity (e.g. cervical radiculopathy) Cervical, shoulder, or upper extremity trauma Any systemic disease causing CTS (e.g. diabetes mellitus, thyroid disease) Comorbid musculoskeletal medical conditions (e.g. rheumatoid arthritis or fibromyalgia) Pregnancy Presence of depressive symptoms (Beck Depression Inventory II [BDI‐II] > 8 points) Male sex CTS diagnostic criteria (case definition): Clinical signs: pain and paraesthesia in the median nerve distribution, increasing symptoms during the night, positive Tinel's sign, and positive Phalen's test Electrodiagnostic examination according to the guidelines of the American Association of Electrodiagnosis, the American Academy of Neurology, and the American Physical Medicine and Rehabilitation Academy: median nerve sensory conduction velocity < 40 m/s and median nerve distal motor latency > 4.20 milliseconds Patients were classified as having: minimal (abnormal segmental‐comparative tests only); moderate (abnormal median nerve sensory velocity conduction and distal motor latency); or severe (absence of median nerve sensory response and abnormal distal motor latency) CTS.
Interventions	Open or endoscopic decompression and release of the carpal tunnel. Surgeons referred patients for hand therapy after the operation as their usual routine if necessary. Patients allocated to this group received the same educational session for performing tendon and nerve gliding exercises as did the physical therapy group. Three treatment sessions of manual therapies including desensitisation manoeuvres of the central nervous system of 30 minutes duration, once per week. The desensitisation manoeuvres consisted of several soft tissue mobilisation and nerve/tendon gliding exercises including manual techniques directed at anatomical sites of potential entrapment of the median nerve (scalene, pectoralis minor, bicipital aponeurosis, pronator teres, transverse carpal ligament, and palmar aponeurosis). In addition, lateral glides to the cervical spine and tendon and nerve gliding interventions were also applied. The therapist alternated the combination of movement depending on the tissue resistance. Speed and amplitude of movement were adjusted such that no pain was produced during the technique. The intervention was completed over 5 to 10 minutes in two sets of 5 minutes each with 1 minute’s rest between sets. The last treatment appointment included an educational teaching session on doing the tendon and nerve gliding exercises as homework if necessary. Patients were asked not to modify any work or activity levels. None of the participants in either group reported any other intervention during the study, excluding the sporadic use of nonsteroidal anti‐inflammatory drugs.
Outcomes	Outcomes were assessed at baseline and 1, 3, 6, and 12 months after the end of therapy; and at 4 years follow‐up. Pain intensity (mean pain and the worst pain) by 11‐point NPRS (0 = no pain, 10 = maximum pain) BCTQ symptom severity score (1 to 5, higher is worse) BCTQ functional status score (1 to 5, higher is worse) Self‐perceived improvement (Global Rating of Change (GROC), from ‐7 (a very great deal worse) to +7 (a very great deal better) Successful outcome, when at least 1 of the following items was present: reduction ≥ 0.7 points or 30% improvement from baseline in BCTQ function or symptom severity subscales, or decrease ≥ 2 points in the intensity of hand pain* Side effects (an adverse event was defined as sequelae of medium‐term duration with any symptom perceived as distressing and unacceptable to the patient and requiring further treatment) Health‐related quality of life (EQ‐5D‐5L, 0‐1, higher is better)**
Funding	The study was funded by a research project grant (FIS PI11/01223) from the Health Institute Carlos III and PN I 1 D 1 I 2012‐2014, Spanish Government. The sponsor had no role in the design, collection, management, analysis, or interpretation of the data, draft, review, or approval of the manuscript or its content. The authors were responsible for the decision to submit the current manuscript for publication, and the sponsor did not participate in this decision.
COI	The authors had no conflicts of interest to declare.
Notes	For patients with bilateral symptoms, the hand with more self‐reported symptoms was assigned as the study hand; if symptoms were equivalent, the mean pain of both hands was considered. BCTQ symptom severity score and functional status score was reported as a scale from 0 to 5; however, it probably is a standard scale from 1 to 5. In methods, one of the successful outcome criteria was defined as reduction of "≥ 0.7 points or 30% improvement from baseline in BCTQ Function or Symptom Severity subscales", but in results "≥ 0.6 points or 30% improvement from baseline in BCTQ Function or Symptom Severity subscales" provided. **Data from the related article Fernandez‐de‐las‐Penas 2019 used. Data were extracted from the figure, and we assumed that the figure shows the mean and standard deviation. The author of the study confirmed that there were no overlapping participants between this study and the other studies by the same author. For improvement, we analysed n as available for follow‐up since the numbers were clearly stated in table 3. For symptoms, function and pain, we analysed the n of participants as randomised, since the authors reported using the last value of each participant when data were missing. For health‐related quality of life, we analysed the n of participants as provided in figure 2 in the article Fernandez‐de‐las‐Penas 2019. For need for surgery, we analysed the n of participants as randomised.

Open in a new tab

Fernandez‐de‐las‐Penas 2017a.

*Study characteristics*
Methods	Study design: randomised, parallel‐group, blinded, clinical trial Setting: local regional hospital in Madrid, Spain from August 2014 to February 2015
Participants	Details of sampling frame: Total n assessed for eligibility = 130 Total n excluded pre‐randomisation = 30 Total n randomised = 100 * Surgical group, n randomised = 50 Physical therapy group, n randomised = 50 Surgical group, n available for follow‐up at 3 m = 50 Physical therapy group, n available for follow‐up at 3 m = 50 Surgical group, n available for follow‐up at 6 m = 50 Physical therapy group, n available for follow‐up at 6 m = 48 Surgical group, n available for follow‐up at 9 m = 50 Physical therapy group, n available for follow‐up at 9 m = 48 Surgical group, n available for follow‐up at 12 m = 48 Physical therapy group, n available for follow‐up at 12 m = 47 Gender distribution: Surgical group: 0 males, 50 females Physical therapy group: 0 males, 50 females Mean ± SD age: Surgical group: 48 ± 9 Physical therapy group: 47 ± 10 Mean ± SD duration of symptoms: Surgical group: 3.3 ± 1.9 years of pain Physical therapy group: 3.2 ± 1.7 years of pain Inclusion criteria: Pain and paraesthesia in the median nerve distribution for at least 6 months Positive Tinel's sign Positive Phalen's test Electrodiagnostic examination had to reveal deficit of sensory and motor median nerve conduction (i.e. median nerve sensory conduction velocity < 40 m/s and median nerve distal motor latency > 4.20 ms) according to guidelines of the American Association of Electrodiagnosis, American Academy of Neurology, and the American Physical Medicine and Rehabilitation Academy Female Age < 65 years Exclusion criteria: Any sensory and/or motor deficit in either ulnar or radial nerve Age > 65 years Previous surgery or steroid injections in the wrist Multiple diagnoses on the upper extremity (e.g. co‐existing cervical radiculopathy) Cervical, shoulder, hand trauma Systemic disease causing CTS (e.g. diabetes mellitus, thyroid disease) Comorbid musculoskeletal medical conditions, e.g. rheumatoid arthritis, or fibromyalgia Pregnancy Male gender CTS diagnostic criteria (case definition): Minimal (abnormal segmental‐comparative tests only), moderate (abnormal median nerve sensory velocity conduction and distal motor latency) or severe (absence of median nerve sensory response and abnormal distal motor latency) CTS Clinical median neuropathy (pain and paraesthesia), positive Phalen's test and Tinel's sign Median nerve sensory conduction velocity < 40 m/s Median nerve distal motor latency > 4.20 ms
Interventions	Surgical intervention ‐ endoscopic decompression and release of the carpal tunnel following international guidelines. Patients allocated to this group also received the same educational sessions for performing the tendon/nerve gliding exercises as the manual therapy group with the same dosage. Three treatment sessions of manual therapies including desensitisation manoeuvres of the central nervous system of 30‐min duration, once/week. The desensitisation manoeuvres consisted of soft tissue mobilisation and nerve/tendon gliding exercises including manual techniques directed at anatomical sites of potential entrapment of the median nerve such as scalene muscles, pectoralis minor muscle, biceps brachii muscle, bicipital aponeurosis, pronator teres, wrist flexor musculature, transverse carpal ligament, palmar aponeurosis or lumbricals muscles. Finally, tendon/nerve gliding interventions of the upper extremity were also applied. The third and last treatment appointments included an educational session on performing the tendon/nerve gliding exercise as homework twice per day during the first month after discharge. Participants were asked not to modify any work or activity levels during the follow‐up period.
Outcomes	Outcomes were assessed at baseline, and 3, 6, 9 and 12 months after the intervention. Pressure pain sensitivity assessed by pressure pain thresholds Thermal pain thresholds (heat or cold pain threshold) Pain intensity by 11‐points NPRS (0 = no pain; 10 = maximum pain)** Side effects
Funding	The study was funded by a research project grant (FIS PI14/00364) from the Health Institute Carlos III and PN I+D+I 2014–2017, Spanish Government. The sponsor had no role in the design, collection, management, analysis, or interpretation of the data, draft, review, or approval of the manuscript or its content. The authors were responsible for the decision to submit the current manuscript for publication, and the sponsor did not participate in this decision.
COI	None declared
Notes	For patients with bilateral symptoms, the hand with more self‐reported symptoms was assigned as the study hand; if symptoms were equivalent, the mean pain of both hands was used. *Data were extracted from the figure. The author of the study confirmed that there were no overlapping participants between this study and the other studies by the same author. For pain, we analysed the n of participants as randomised, since the authors reported using multiple imputation method when data were missing. For need for surgery, we analysed the n of wrists as randomised.

Open in a new tab

Fernandez‐de‐las‐Penas 2017b.

*Study characteristics*
Methods	Study design: randomised parallel‐group, blinded, clinical trial Setting: local regional hospital (Madrid, Spain), between September 2014 and February 2015
Participants	Details of sampling frame: Total n assessed for eligibility = 140 Total n excluded pre‐randomisation = 40 Total n randomised = 100* Surgical group, n randomised = 50 Physical therapy group, n randomised = 50 Surgical group, n available for 1 m follow‐up = 50 Physical therapy group, n available for 1 m follow‐up = 50 Surgical group, n available for 3 m follow‐up = 50 Physical therapy group, n available for 3 m follow‐up = 50 Surgical group, n available for 6 m follow‐up = 49 Physical therapy group, n available for 6 m follow‐up = 50 Surgical group, n available for 12 m follow‐up = 47 Physical therapy group, n available for 12 m follow‐up = 47 Gender distribution: Surgical group: 0 males, 50 females Physical therapy group: 0 males, 50 females Mean ± SD age: Surgical group: 47 ± 8 Physical therapy group: 46 ± 9 Mean ± SD duration of symptoms: Surgical group: 3.1 ± 1.8 years of pain Physical therapy group: 2.8 ± 1.6 years of pain Inclusion criteria: Pain and paraesthesia in the median nerve distribution Positive Tinel's sign Positive Phalen's test Symptoms had to have persisted for at least 12 months Electrodiagnostic examination had to reveal deficits of sensory and motor median nerve conduction (i.e. median nerve sensory conduction velocity less than 40 m/s and median nerve distal motor latency greater than 4.20 milliseconds), according to guidelines of the American Association of Electrodiagnostic Medicine, American Academy of Neurology, and American Academy of Physical Medicine and Rehabilitation. Exclusion criteria: Any sensory and/or motor deficit in either the ulnar or radial nerve Age > 65 years Previous surgery or steroid injections Multiple diagnoses on the upper extremity (e.g. co‐existing cervical radiculopathy) Cervical, shoulder, hand trauma Systemic disease causing CTS (e.g. diabetes mellitus, thyroid disease) Comorbid musculoskeletal medical conditions, e.g. rheumatoid arthritis, or fibromyalgia Pregnancy Male sex CTS diagnostic criteria (case definition): Minimal (abnormal segmental‐comparative tests only), moderate (abnormal median nerve sensory velocity conduction and distal motor latency), or severe (absence of median nerve sensory response and abnormal distal motor latency) Clinical symptoms (pain and paraesthesia in median nerve), positive Phalen's test and Tinel's sign Median nerve sensory conduction velocity < 40 m/s Median nerve distal motor latency > 4.2 ms
Interventions	Endoscopic decompression and release of the carpal tunnel + educational session for performing the same cervical spine exercise programme as the physical therapy group. Three treatment sessions of manual therapies, including manoeuvres targeted to the cervical spine and to those areas anatomically related to potential entrapment of the median nerve (i.e. shoulder, elbow, forearm, wrist, and fingers), of 30 minutes’ duration, once per week. The interventions applied to the neck consisted of lateral glides applied to the cervical spine and posteroanterior pressure directed to the mid cervical spine. In addition, soft tissue interventions targeting the scalene muscles, costoclavicular space, pectoralis minor, biceps brachii muscle, bicipital aponeurosis, pronator teres, transverse carpal ligament, palmar aponeurosis, and lumbrical muscles were also applied. Finally, patients performed a cervical spine exercise programme for stretching neck muscles. The third and last treatment appointment included an educational session on how to perform the cervical exercises as homework during the follow‐up period, as needed. Patients were encouraged not to modify any work or activity levels and only perform the cervical spine exercises if they experienced increases in their symptoms during the follow‐up period.
Outcomes	Outcomes were assessed at baseline and 1, 3, 6, and 12 months after the last treatment. BCTQ functional status score (1 to 5, higher is worse) Active cervical range of motion Pinch‐tip grip force BCTQ symptom severity score (1 to 5, higher is worse) Side effects
Funding	The local human research committee (HUFA PI‐12/0023) approved the study project. The study was funded by a research project grant (FIS PI14/ 00364) from the Health Institute Carlos III (PN I+D+I 2014‐2017; Spanish Government).
COI	The authors certified that they had no affiliations with or financial involvement in any organisation or entity with a direct financial interest in the subject matter or materials discussed in the article.
Notes	*In those with bilateral symptoms, the most painful side was considered as the affected side and the less painful side as the unaffected side. The author of the study confirmed that there were no overlapping participants between this study and the other studies by the same author. For symptoms and function, we analysed the n of participants as randomised, since the authors reported using multiple imputation method when data were missing. For need for surgery, we analysed the n of wrists as randomised.

Open in a new tab

Garland 1964.

*Study characteristics*
Methods	Study design: randomised controlled trial Setting: Departments of Neurology, Electromyography and Orthopaedic Surgery, Leeds, UK
Participants	Details of sampling frame: Total n eligible = not reported Total n excluded pre‐randomisation = not reported Total n randomised = 22 hands Intervention group 1 surgical treatment, n randomised = 11 Intervention group 2 splinting, n randomised = 11 Intervention group 1 surgical treatment, n available for 12 m follow‐up = 10 Intervention group 2 splinting, n available for 12 m follow‐up = 10 Gender distribution: Total: 0 males, 22 females Mean ± SD age: Total: 46.8 (35 to 63 years) Mean ± SD duration of symptoms: 3.5 * (range from 1 month to 20 years) Inclusion criteria: Patient accepted on both clinical and electromyographic (wrist‐muscle conduction times for the median nerve of more than 4.5 msec) grounds Exclusion criteria: Not reported CTS diagnostic criteria (case definition): Conduction time for wrist muscle > 4.5 ms
Interventions	Surgical treatment, by open operation and section of the anterior carpal ligament Plaster‐of‐Paris splinting of the hand, wrist, and arm for 1 month
Outcomes	Outcomes measured clinically and electromyographically at regular intervals for up to 1 year. Relief of symptoms (not mentioned what kind of method used) Motor and sensory nerve conduction in the distal segment of the median nerve Referral to surgery
Funding	Not reported
COI	Not reported
Notes	*We assumed the provided number means years. In 16 of these participants, both arms were affected, but the worse side was selected for the trial. We analysed the number of participants as available for the follow‐up.

Open in a new tab

Gerritsen 2002.

*Study characteristics*
Methods	Study design: randomised controlled trial Setting: 13 neurological outpatient clinics in the Netherlands
Participants	Details of sampling frame: Total n assessed for eligibility = 326 Total n excluded pre‐randomisation = 150 Total n randomised = 176* Wrist splinting, n randomised = 89 Open carpal tunnel release, n randomised = 87 Wrist splinting, n available for follow‐up at 1 m = 88 Open carpal tunnel release, n available for follow‐up at 1 m = 80 Wrist splinting, n available for follow‐up at 3 m = 86 Open carpal tunnel release, n available for follow‐up at 3 m = 78 Wrist splinting, n available for follow‐up at 6 m = 84 Open carpal tunnel release, n available for follow‐up at 6 m = 77 Wrist splinting, n available for follow‐up at 12 m = 83 Open carpal tunnel release, n available for follow‐up at 12 m = 73 Wrist splinting, n available for follow‐up at 18 m = 79 Open carpal tunnel release, n available for follow‐up at 18 m = 68 Gender distribution: Wrist splinting: 12 males, 77 females Open carpal tunnel release: 21 males, 66 females Mean ± SD age: Wrist splinting: 49 ± 12 Open carpal tunnel release: 49 ± 11 Mean ± SD duration of symptoms: Wrist splinting: 52 weeks (median) (IQR 24 to 104) Open carpal tunnel release: 40 weeks (median) (IQR 16 to 104) Inclusion criteria: Pain, paraesthesia, and/or hypoesthesia in the hand in the area innervated by the median nerve Electrophysiological confirmation of the diagnosis (median nerve sensory conduction velocity of the index finger ≤ 41.9 m/s in patients < 55 years or ≤ 37.3 m/s in patients ≥ 55 years, or median nerve distal sensory latency of the index finger ≥ 3.5 ms; or median‐ulnar distal sensory latency difference of the ring finger > 0.4 ms; or median nerve distal motor latency ≥ 4.34 ms) Age of 18 years or older Ability to complete written questionnaires in Dutch Exclusion criteria: Previous treatment with splinting or surgery A history of wrist trauma (e.g. fracture) or surgery A history suggesting underlying causes of CTS (e.g. diabetes mellitus, pregnancy) Clinical signs or symptoms or electrophysiological findings suggesting conditions that could mimic CTS or interfere with its validation (e.g. cervical radiculopathy, polyneuropathy) Severe thenar muscle atrophy CTS diagnostic criteria (case definition): Median nerve sensory conduction velocity of the index finger ≤ 41.9 ms in patients < 55 years or ≤ 37.3 ms in patients ≥ 55 years or Median nerve distal sensory latency of the index finger ≥ 3.5 ms or Median‐ulnar distal sensory latency difference of the ring finger > 0.4 ms or Median nerve distal motor latency ≥ 4.34 ms
Interventions	Wrist splinting during the night for at least 6 weeks: either a custom‐made splint (made of soft cast) or a prefabricated splint (Tricodur, BSN Medical, Hamburg, Germany) that immobilised the wrist in a neutral position. The patients were instructed to wear the splint during the night for at least 6 weeks, but could wear it during the day. No other types of treatment were permitted during the intervention period except pain medication, if necessary. Open carpal tunnel release (efforts were made to make an appointment within 4 weeks after randomisation). Before surgery, no other types of treatment were permitted except pain medication. In all surgical cases, the transverse carpal ligament was released and no concomitant procedures were performed (e.g. flexor tenosynovectomy, internal neurolysis, epineurotomy). Sutures were removed after 2 weeks. Patients were instructed to perform postoperative active range‐of‐motion exercises and encouraged to use the hand as tolerated. None of the patients received a splint following the surgical procedure. No specific period off work was recommended. In the surgery group, 26 patients received 1 or more additional treatment options after surgery (pain medication; physiotherapy; occupational therapy; local corticosteroid injection and surgery to relieve pain caused by complex regional pain syndrome). In the splinting group, 74/89 patients continued to wear splints after the intervention period for a varying period of time and 52 patients received 1 or more additional treatment options after 6 weeks (pain medication, physiotherapy, manual therapy, Mensendieck (exercise) therapy, occupational therapy, local corticosteroid injections, and 35 received surgery).
Outcomes	Outcomes assessed at baseline and 1, 3, 6, 12 months and 18 months after randomisation. General improvement (6‐point ordinal transition scale, ranging from "completely recovered" to "much worse"); treatment success defined as "completely recovered" or "much improved" Number of nights waking up due to symptoms Severity of the main complaint (the complaint that patient considers to be the most important), pain, paraesthesia, or hypoesthesia at night and during the day during the past week (11‐point numerical rating scale, 0 = no symptoms and 10 = very severe symptoms)* BCTQ symptom severity score (1 to 5, higher is worse) BCTQ functional status score (1 to 5, higher is worse) Overall severity of CTS complaints (11‐point Numerical Rating Scale; 0 = no complaints, 10 = very severe complaints) Nerve conduction studies (distal sensory latency, median‐ulnar distal sensory latency difference, distal motor latency) Adverse effects Health‐related quality of life (EuroQol, expressed as utility, 0–1)** Economic evaluation**
Funding	The research for this article was funded by grant OG97‐013 from the Health Care Insurance Council of the Netherlands.
COI	Not reported in the main article. The Dutch version of the article reports no conflict of interest.
Notes	If bilateral symptoms were present, the hand with more severe symptoms (according to the patient) was treated. Randomisation was chosen as the reference point for the timing of the follow‐up assessments, and not the actual start of the treatment. Scores for pain were not presented because the findings were similar to those for paraesthesia. **Reported in a related article Korthals‐de Bos 2006, number of patients from Table 5 used. We assumed the study used EQ‐5D, even though the authors refer to EuroQol only. We analysed the number of participants as available for follow‐up, except for adverse events (for the adverse events we analysed participants as randomised, as reported in table 4).

Open in a new tab

Hui 2005.

*Study characteristics*
Methods	Study design: randomised, single‐blind, controlled trial Setting: Neurology and Rheumatology clinics at the Prince of Wales Hospital, Hong Kong
Participants	Details of sampling frame: Total n assessed for eligibility = 63 Total n excluded pre‐randomisation = 13 Total n randomised = 50 Open carpal tunnel release, n randomised = 25 Injection of steroid, n randomised = 25 Open carpal tunnel release, n available for follow‐up 6 w = 25 Injection of steroid, n available for follow‐up 6 w = 25 Open carpal tunnel release, n available for follow‐up 20 w = 25 Injection of steroid, n available for follow‐up 20 w = 25 Gender distribution: Open carpal tunnel release: 1 male, 24 females Single injection of steroid: 1 male, 24 females Mean ± SD age: Open carpal tunnel release: 50.8 ± 11.6 Single injection of steroid: 48.2 ± 6.5 Mean ± SD duration of symptoms: Newly diagnosed CTS of more than 3 months but less than 1 year’s duration Inclusion criteria: Newly diagnosed CTS of more than 3 months but less than 1 year’s duration Exclusion criteria: Patients with evidence of severe CTS as defined by the presence of thenar atrophy or unelicitable distal motor latencies Patients with ulnar, radial neuropathy, or proximal median neuropathy Coexisting disorders or conditions that may mimic CTS such as brachial plexopathy, cervical radiculopathy, tenosynovitis Contraindication to steroid use such as active affective disorder and recent peptic ulcer History of underlying disorders associated with CTS such as diabetes mellitus, wrist trauma, rheumatoid arthritis, acromegaly, hypothyroidism Pregnancy CTS diagnostic criteria (case definition): Electrophysiologic criteria for the diagnosis of CTS in our centre were as follows: median‐ulnar palmar sensory latency difference > 0.5 msec, using a distance of 8 cm between stimulation and recording sites or prolonged median nerve DML > 4 msec (mean value + 2 SD).
Interventions	Surgical decompression under local anaesthesia (1% plain lignocaine) Single injection of steroid ‐ 15 mg of methylprednisolone acetate was injected into the carpal tunnel via a 25‐gauge needle. Co‐interventions such as drugs, splinting, or injections were withheld for the duration of the study and patients were informed that alternative therapy with chiropractic manipulation or acupuncture were not permitted during the trial period.
Outcomes	Outcomes measured at baseline and at 6 and 20 weeks after intervention. GSS (0 (no symptoms) to 10 (severe) in five categories: pain, numbness, paraesthesia, weakness/clumsiness, and nocturnal awakening; the sum of the scores in each category was the GSS)* Median nerve distal motor latency (ms) Sensory nerve conduction velocity (m/s) Functional assessment in the form of grip strength measurements (JAMAR hydraulic hand dynamometer) (kg)
Funding	Not reported
COI	Not reported
Notes	*Data of improvement in GSS by at least 50% were sent by the corresponding author to the author of the previous version of this review.

Open in a new tab

Ismatullah 2013.

*Study characteristics*
Methods	Study design: prospective randomised clinical trial Setting: Khalifa Gul Nawaz Teaching Hospital and District Headquarters Teaching Hospital, Bannu, Pakistan
Participants	Details of sampling frame: Total n eligible = not reported Total n excluded pre‐randomisation = not reported Total n randomised = 40* Surgery, randomised = 20 Local steroid injection, randomised = 20 Surgery, n available for follow‐up = 20 Local steroid injection, n available for follow‐up = 20 Gender distribution: Surgery: 6 males, 14 females Local steroid injection: 5 males, 15 females Mean ± SD age: Surgery: 43.8 ± 10.98 Local steroid injection: 46.9 ± 12.33 Mean ± SD duration of symptoms: Surgery: 12.5 ± 8.76 Local steroid injection: 10.15 ± 6.75 Inclusion criteria: Diagnosed clinically on the basis of typical signs and symptoms of CTS: Intermittent pain and paraesthesia in the distribution of the median nerve in the hand at night as well as during daily activities Shaking or flicking one's hand to relieve the symptoms (Flick sign) Sensory deficit in median nerve distribution of the hand Weakness of the abductor pollicis brevis muscle and atrophy of the thenar muscles Positive Phalen's test Positive Tinel's sign Patients were included in the study irrespective of age, gender, class or ethnicity. Exclusion criteria: Peripheral polyneuropathy Cervical radiculopathy Those with a recent history of trauma or fractures of the wrist Patients having symptoms of less than 3 months duration Recurrent cases of CTS after previous local steroid injection or previous surgery CTS diagnostic criteria (case definition): Electrodiagnostic studies were used for confirming the diagnosis in a few doubtful cases of CTS.
Interventions	Carpal tunnel release ‐ the operation was done under local anaesthesia as a day‐case. Oral antibiotics and analgesics were used for 4‐5 days after surgery. Local steroid injection ‐ 40 mg of methyl‐prednisolone (Depomedrol) injection was given in the carpal tunnel. No other medicines like analgesics, B‐Complex, etc., were used.
Outcomes	Outcomes assessed at baseline and at 2 weeks, 4 weeks*** and 12 weeks. GSS (0 to 50, where 0 indicates no symptoms and 50 indicates the most severe symptoms)
Funding	Not reported
COI	Not reported
Notes	In the case of patients having bilateral CTS, only the more symptomatic hand was included in the study. Not reported if it was weeks, months or years, but seems likely to be months. **In the methodology chapter, the 2nd time point was 6 weeks but in the Results tables 4 weeks were provided.

Open in a new tab

Jafari 2018.

*Study characteristics*
Methods	Study design: randomised clinical trial Setting: Shafa Orthopaedic Hospital (Tehran, Iran)
Participants	Details of sampling frame: Total n assessed for eligibility = 128 Total n excluded pre‐randomisation = 46 Total n randomised = 82* Surgery, n randomised = 41 Local steroid injection, n randomised = 41 Surgery, n available for follow‐up = 35 Local steroid injection, n available for follow‐up = 33 Gender distribution: Surgery: 6 males, 29 females Local steroid injection: 6 males, 27 females Mean ± SD age: Surgery: 46.2 ± 7.8 Local steroid injection: 45.4 ± 8.4 Mean ± SD duration of symptoms: Not reported Inclusion criteria: Mild‐to‐moderate CTS Exclusion criteria: Those patients in whom electrophysiological findings were not in accordance with clinical findings All patients with severe CTS Patients with underlying diseases such as double crush syndrome, rheumatoid arthritis, diabetes mellitus, and thyroid disorders CTS diagnostic criteria (case definition): Evaluation of patients’ history of the symptoms Confirmed by signs of CTS, such as Tinel's sign, Phalen's test, and Durkan compression CTS severity determined by electrophysiological study: nerve conduction velocity was graded according to the previously introduced neurophysiological grading with a spectrum of grades from 0 to 6 (normal, very mild, mild, moderate, severe, very severe, and extremely severe).
Interventions	Local steroid injection: dose of 40 mg/mL of triamcinolone acetonide was injected into the carpal tunnel at the volar aspect of the wrist, just ulnar to the palmaris longus tendon, with a syringe making an angle of 60° to the horizon, and entering 10 mm into the transverse carpal ligament depth. Carpal tunnel releases: were done according to Green’s classic approach (Cohen 2017). Short arm splints were applied for 2 weeks after the operation.
Outcomes	Outcomes measured at baseline and 6 months.** BCTQ functional status score (1‐5, higher is worse) BCTQ symptom severity score (1‐5, higher is worse) Patients’ satisfaction (5‐point scale from 0 to 5, indicating completely satisfied, almost satisfied, moderately satisfied, somewhat satisfied, and dissatisfied)
Funding	Not reported
COI	Not reported
Notes	No information regarding bilateral patients *In the results section, the outcomes are reported for 6 weeks. However, in the methods and discussion and conclusions, the authors clearly stated 6 months. We tried to contact the authors but did not receive clarification, and we assumed that the outcomes were measured at 6 months (i.e. we used the data for long‐term analysis).

Open in a new tab

Jarvik 2009.

*Study characteristics*
Methods	Study design: parallel‐group randomised controlled trial Setting: four academic and three private practice centres in Washington State and the Dartmouth Hitchcock Medical Center in New Hampshire, USA
Participants	Details of sampling frame: Total n assessed for eligibility = 855 Total n excluded pre‐randomisation = 739 Total hands randomised = 116 patients (116 hands)* Surgery, randomised = 57 Non‐surgical treatment, randomised = 59 Surgery, available for 3 m follow‐up = 51 Non‐surgical treatment, available for 3 m follow‐up = 56 Surgery, available for 6 m follow‐up = 50 Non‐surgical treatment, available for 6 m follow‐up = 54 Surgery, available for 12 m follow‐up = 49 Non‐surgical treatment, available for 12 m follow‐up = 52 Gender distribution: Surgery: 29 males, 28 females Non‐surgical: 25 males, 34 females Mean ± SD age: Surgery: 50.2 ± 10.3 Non‐surgical: 51.2 ± 8.9 Mean ± SD duration of symptoms: Surgery: 3.2 years (median) (IQR 1.3 to 5.5) Non‐surgical: 3.4 years (median) (IQR 1.0–8.7) Inclusion criteria: Symptoms for at least 2 weeks in at least two digits on one hand, which include thumb, index, or 1 ring finger Classic, probable, or possible carpal tunnel syndrome on hand pain diagram (Katz 1990) Failure for at least two consecutive weeks of non‐surgical treatment, including a trial of wrist splints, where failure was defined as improvement of less than 0.75 on the Carpal Tunnel Syndrome Assessment Questionnaire (CTSAQ) Function Scale, inability to achieve a satisfactory level of work, or symptoms rated as being the same or worse In the absence of electrodiagnostic criteria (see below), night pain that woke them and a positive flick test (shaking the wrist when asked what they do to relieve symptoms) (Pryse‐Phillips 1984) Electrodiagnostic inclusion criteria (any one of the following): median motor latency at the wrist ≥ 4.4 ms; median ulnar sensory difference (14 cm digit IV to wrist) > 0.4 ms; median ulnar sensory difference (8 cm midpalm) > 0.3 ms; median radial sensory difference (10 cm thumb to wrist) > 0.5 ms; combined sensory index ≥ 1.0 ms, where combined sensory index is the sum of: median ulnar ring finger antidromic latency difference at 14 cm (ring‐diff), median radial thumb antidromic latency difference at 10 cm (thumb‐diff), and median ulnar midpalmar orthodromic latency difference at 8 cm (palm‐diff) Able to complete English‐language questionnaires and telephone interviews Plan to stay in the region for 12 months Age ≥ 18 years Willing to undergo surgery as soon as possible after randomisation Exclusion criteria: Severe carpal tunnel syndrome (the study used modification of the classification by MacKinnon 1988 to define severe cases as those with thenar muscle wasting or abnormal static 2‐point discrimination (≤ 6 mm). Furthermore, the study excluded patients with electrodiagnostic evidence of denervation, defined by either needle electromyograph or median motor amplitude ≤ 3.8 mV) Previous carpal tunnel syndrome surgery on the study hand Any wrist or hand surgery within the previous 6 months Moderate‐to‐severe arthritis involving the study hand or wrist Known tumour, mass, or deformity of the study hand or wrist History of severe trauma to the study hand or wrist Current pregnancy or lactation Evidence of diffuse peripheral neuropathy or cervical radiculopathy CTS diagnostic criteria (case definition): All participants underwent electrodiagnostic tests as part of their routine clinical assessment within 6 months before enrolment.
Interventions	Carpal tunnel surgery (open or endoscopic decompression, depending on the surgeon’s preference). Surgeons referred patients for hand therapy after surgery as their usual routine. Well‐defined, non‐surgical treatment (including hand therapy and ultrasound): nonsteroidal anti‐inflammatory drugs (ibuprofen 200 mg three times a day) and scheduled with a hand therapist for customised hand‐therapy sessions (six visits over 6 weeks). At the first visit, each participant received an educational booklet and hand exercises. Hand‐therapy sessions focused on ligament stretching, tendon gliding, and review of splint use. The therapist encouraged continued splinting at night and as tolerated during the day. The hand therapist also suggested modifications for work and activity. Patients allocated to non‐surgical therapy returned 6 weeks after randomisation and, if symptoms had improved, the study made no changes to their therapy. Those whose symptoms had not improved at six weeks were offered therapeutic ultrasound according to the protocol of Ebenbichler 1998, consisting of up to 12 sessions (2–4 per week for up to 6 weeks) of focused ultrasound at 1 Mhz, 1.0 W/cm², in pulsed mode 1:4, and 15 min each. Those whose symptoms did not improve during three months were offered surgery.
Outcomes	Outcomes evaluated at baseline and at 6 weeks, 3, 6, 9, and 12 months after randomisation. Carpal Tunnel Syndrome Assessment Questionnaire (CTSAQ), 9‐item Functional Status Scale (1‐5, higher is worse) CTSAQ symptom score, 11‐item Symptom Severity Scale (1‐5, higher is worse) Hand or wrist pain intensity and interference (0 = no pain or interference to 10 = extreme pain or interference) Effect of carpal tunnel syndrome on work and other activities, including days of work lost and days of limited activity General health‐related quality of life (Short Form‐36 (SF‐36) version 2.0)) ‐ SF‐36 PCS and MCS scores are norm‐based with a mean (SD) of 50 in a healthy population. Higher scores indicate better health‐related quality of life Successful outcome (0.50 points or > 30% improvement in CTSAQ function or CTSAQ symptom, daily work or housework interference = 0 or 1) Clinically important adverse events
Funding	NIH/NIAMS 5P60AR048093 and the Intramural Research Program of the NIH Clinical Center. The sponsor had no role in the design, analysis, and conduct of the study other than the usual NIH peer‐review process. The corresponding author was responsible for the decision to submit the manuscript for publication, and the sponsor did not participate in this decision.
COI	No conflicts of interest reported
Notes	38 patients in the surgery group and 31 in the non‐surgery group had bilateral symptoms, but only one hand for each patient was included in the study. *According to figure 1, 19 of 59 (32%) participants in the non‐surgical treatment group received surgery by 12 months but, in the text, the authors have reported 23 of 59 (39%). We used the data as reported in the text.

Open in a new tab

Ly Pen 2005.

*Study characteristics*
Methods	Study design: prospective, randomised, open, controlled clinical trial Setting: Madrid, Spain; conducted between October 1998 and May 2001
Participants	Details of sampling frame: Total n assessed for eligibility = 217 wrists (123 patients) Total n excluded pre‐randomisation = 54 wrists Total n randomised = 101 patient (163 wrists) Surgery, n randomised = 56 patients (32 unilateral + 24 bilateral = 80 wrists) Local steroid injection, n randomised = 49 patients (15 unilateral + 34 bilateral = 83 wrists) Surgery, n available for 3 m follow‐up = 67 wrists Local steroid injection, n available for 3 m follow‐up = 80 wrists Surgery, n available for 6 m follow‐up = 63 wrists Local steroid injection, n available for 6 m follow‐up = 77 wrists Surgery, n available for 12 m follow‐up = 57 wrists Local steroid injection, n available for 12 m follow‐up = 66 wrists Surgery, n available for 24 m follow‐up = 55 wrists Local steroid injection, n available for 24 m follow‐up = 48 wrists Gender distribution: Total: 8 males, 93 females Mean ± SD age: Surgery: 50.52 ± 10.87 Local steroid injection: 53.17 ± 13.93 Mean ± SD duration of symptoms: Surgery: 31.12 ± 7.27 weeks Local steroid injection: 33.25 ± 8.17 weeks Inclusion criteria: At least 18 years old Suggestive symptoms of CTS of at least 3 months’ duration Were consecutively referred by their primary care physicians to a CTS unit specifically created for this study Presumptive diagnosis of CTS Unresponsive to a course of at least 2 weeks of NSAIDs and splinting Exclusion criteria: Thenar atrophy Previous carpal tunnel release surgery Local injection for CTS Patients who were pregnant Diabetes mellitus, hypothyroidism, inflammatory arthropathy, or polyneuropathy CTS diagnostic criteria (case definition): Electrodiagnostic criteria for the diagnosis of CTS were either 1) a motor latency in the median nerve from the wrist to the base of the thumb that was > 2 SD above the normal mean (4.2 m/sec), with a difference with respect to the distal latency of the ulnar nerve that was > 2 SD above the mean (1.4 m/second); or 2) a decrease in the sensory conduction velocity from the wrist to the third finger that was <2SD from the normal mean (44 m/sec), with a latency difference with respect to the sensory potential of the ulnar nerve that was > 2 SD above the mean (0.7 m/sec). The lower normal values for the amplitude of evoked potentials of the median nerve, considered to be the mean minus 2 SD, was 3.5 mV for the motor potential and 19 mikroV for the sensory potential.
Interventions	Surgical decompression ‐ limited palmar incision technique. Local steroid injection (a standard technique), paramethasone acetonide, 20 mg in 1 mL. In the injected wrists, the protocol allowed a second injection if nocturnal paraesthesias had not disappeared completely (score of 0 on the visual analogue scale). Thirteen wrists required 1 injection, and 69 wrists required 2 injections.
Outcomes	Outcomes were assessed at baseline and at 3, 6, 12 and 24 months after treatment. Nocturnal paraesthesias and pain in the area of median nerve distribution, functional impairment (VAS, 0 = no symptoms and 100 = the most intense symptoms)* Primary end point ‐ percentage of wrists reaching at least a 20% reduction in the VAS score for nocturnal paraesthesias at 3 months of follow‐up Secondary end points ‐ the percentages of wrists with a 20% reduction in the VAS score for nocturnal paraesthesias at 6 and 12 months, a 20% response for pain and functional impairment, as well as a 50% and a 70% response in nocturnal paraesthesias, pain, and functional impairment Nerve conduction studies Adverse events
Funding	Not reported
COI	Not reported in the main article. In related articles, authors have reported no conflicts of interest.
Notes	Data from related article Andreu 2014 used. *We used data from a subgroup of wrists (n = 69) consisting of unilateral CTS and the most symptomatic wrist in patients with bilateral CTS (excluding bilateral CTS). For the need for surgery, we analysed the n of wrists as randomised.

Open in a new tab

Ucan 2006.

*Study characteristics*
Methods	Study design: randomised controlled trial Setting: Ankara Numune Education and Research Hospital Physical Medicine and Rehabilitation Outpatient Clinic, Turkey
Participants	Details of sampling frame: Total n assessed for eligibility = 160 Total n excluded pre‐randomisation = 93 Total n randomised = 67 Total n available for follow‐up = 57 Surgery, randomised = not reported* Splint, randomised = not reported* Splint + local corticosteroid injection (SI), randomised = not reported* Surgery, available for follow‐up =11 Splint, available for follow‐up = 23 Splint + SI, available for follow‐up = 23 Gender distribution: Surgery: 1 male, 10 females Splint: 1 male, 22 females Splint + SI: 2 male, 21 females Mean ± SD age: Surgery: 45.27 ± 13.19 Splint: 44.5 ± 7.24 Splint + SI: 44.46 ± 8.52 Mean ± SD duration of symptoms: Surgery: 21 ± 11 months Splint: 15.26 ± 7.19 months Splint + SI: 19.13 ± 13 months* Inclusion criteria: Idiopathic CTS Mild‐moderate CTS in nerve conduction studies (AAEM guidelines). Exclusion criteria: Patients with advanced CTS or thenar atrophy Patients with underlying metabolic disorders such as diabetes mellitus, thyroid, kidney diseases, connective tissue disorders, malignancy, distal radius fracture, and pregnancy in aetiology Conditions that could affect management response such as cervical disc herniation, fibromyalgia Prior CTS treatment. CTS diagnostic criteria (case definition): Diagnosis of CTS was confirmed with nerve conduction studies and the patients were classified as mild, moderate, or advanced CTS according to the American Association of Electrodiagnostic Medicine (AAEM) guidelines (Stevens 1997).
Interventions	Open carpal tunnel release, after which the hand was bandaged and early mobilisation of fingers was encouraged. Splinting for 3 months (splinted in neutral position with standard cotton‐polyester splint). Patients were encouraged to use the splints during night‐time and also daytime whenever possible for 3 months. Splinting for 3 months plus local steroid injection. 20 mg triamcinolone acetonide (Kenacord A©) and 20 mg lidocaine (Jetocain Simplex©) mixture was injected using a 23‐gauge needle.
Outcomes	Outcomes assessed at baseline and the 3rd and 6th month after treatment. BCTQ symptom severity score (1‐5, higher is worse) BCTQ functional status score (1‐5, higher is worse) Patient satisfaction (5‐point scale as completely satisfied, almost satisfied, moderately satisfied, somewhat satisfied, and dissatisfied) Median motor nerve proximal and distal latencies Median motor nerve velocity Compound muscle action potentials Median sensory nerve velocities Sensorial nerve action potential amplitudes
Funding	Not reported
COI	Not reported
Notes	The authors did not report how many patients were randomised in each of the study groups; however, the authors reported to us that excluded hands belonged to splinting and splinting + injection groups (plus the article suggests that 4/10 of excluded hands belonged to the surgery group). The number of female patients in the splinting + SI group was reported as 21 in the study of Ucan and 22 in the study of Yagci (although they are the same studies). We have used the numbers from the Ucan study. **Mean symptom duration for splinting + SI group was reported as 19.13 ± 13 months in the study of Ucan and 19.13 ± 11.72 in the study of Yagci (although they are the same studies). We have used the numbers from the Ucan study.

Open in a new tab

Zhang 2019.

*Study characteristics*
Methods	Study design: randomised controlled trial Setting: Department of Rehabilitation Medicine, Sun Yat‐sen Memorial Hospital, Sun Yat‐sen University, Guangzhou, Guangdong Province, China
Participants	Details of sampling frame: Total n eligible = not reported Total n excluded pre‐randomisation = not reported Total n randomised = 51 patients (51 wrists) Steroid injection (SI) + surgery, randomised = 25 patients SI, randomised = 26 patients SI + surgery, available for follow‐up = 23 patients SI, available for follow‐up = 23 patients Gender distribution: SI + surgery: 5 males, 18 females SI: 6 males, 17 females Mean ± SD age: SI + surgery: 48.7 ± 15.2 SI: 53.1 ± 14.6 Mean ± SD duration of symptoms: SI + surgery: 10.2 ± 3.5 months SI: 11.1 ± 2.8 months Inclusion criteria: Pain, numbness, or tingling in the median nerve distribution area of the hand Nocturnal worsening of the symptoms Positive Tinel's sign and/or Phalen's test, a slower median nerve conduction (sensory nerve conduction velocity ≦ 50 m/s and/or distal motor latency ≧ 4 ms) Patients with unilateral disease The desire of the participant to have either a SI or SI plus miniscalpel‐needle (MSN) release Exclusion criteria: Symptomatic CTS because of diabetes, thyroid disease, or rheumatic disease Cervical radiculopathy or other polyneuropathy Age < 18 years Pregnancy SI for CTS in the preceding 6 months History of wrist fracture Prior carpal tunnel decompressive surgery The presence of infection or skin lesion at the site of injection Patients with bilateral disease Refusal of informed consent or inability to participate in follow‐up CTS diagnostic criteria (case definition): Patients with symptoms of pain, numbness, or tingling in the median nerve distribution area of hand ‐ confirmation by physical and electrophysiological inspection.
Interventions	Steroid injection combined with MSN release: Group A was treated with US‐guided MSN release first to release the nerve entrapment. Immediately after that, a steroid injection was performed. MSN release was conducted using an inplane approach. Steroid injection (US‐guided injection was conducted using an out‐of‐plane approach). Steroid used: 1.0 mL of compound betamethasone (2 mg betamethasone sodium phosphate and 5 mg betamethasone dipropionate) together with 1.0 mL of 1% lidocaine. After treatment, in addition to proper hand movements, any other complementary or alternative treatment was not allowed during the 12‐week follow‐up. After treatment, patients in both groups were observed for 30 minutes to record any adverse reaction.
Outcomes	Outcomes assessed at baseline, 4 and 12 weeks after treatment. BCTQ symptom severity score (1‐5, higher is worse) BCTQ functional status score (1‐5, higher is worse) Compound muscle action potential Distal motor latency Sensory nerve action potential Sensory nerve conduction velocity Cross‐sectional area of the median nerve
Funding	This work was funded by National Natural Science Foundation of China (81771201, 81671088), Natural Science Foundation of Guangdong Province (2016A030311045), Guangzhou Science and Technology Project under the framework of government cooperation, and Sun Yat‐sen Clinical Research Cultivating Program.
COI	The authors have no conflicts of interest to declare.
Notes

Open in a new tab

AAEM: American Association of Electrodiagnostic Medicine BCTQ: Boston Carpal Tunnel Questionnaire BDI‐II: Beck Depression Inventory II cm: centimetre CTS: carpal tunnel syndrome CTSAQ: Carpal Tunnel Syndrome Assessment Questionnaire DML: distal motor latency ENMG: electroneuromyography EQ‐5D‐5L: EuroQol 5‐dimension 5‐level health status and quality of life measure EuroQol: EuroQol Group g: gram GROC: Global Rating of Change GSS: Global Symptom Score IQR: interquartile range m: month MCS: mental component summary mg: milligram mg/dL: milligrams per decilitre  mg/mL: milligrams per millilitre MHz: megahertz min: minute mL: millilitre mm: millimetre ms, msec, m/second: millisecond m/s: metre per second MSN: miniscalpel‐needle mikroV: mikrovolt mV: millivolt n: number NPRS: numerical pain rating scale NSAID: nonsteroidal anti‐inflammatory drug PCS: physical component summary PRP: platelet‐rich plasma RA: rheumatoid arthritis rpm: revolutions per minute SD: standard deviation SF‐36: 36‐Item Short Form Survey SI: steroid injection TSH: thyroid‐stimulating hormone T3: triiodothyronine  T4: thyroxine VAS: visual analogue scale W/cm²: watts per centimetre squared

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
Baker 2014	Study does not compare surgical versus non‐surgical treatment.
Bozkurt 2021	Surgery applied to all groups in the study.
Celik 2016	Study is not RCT.
Cha 2016	Study is not RCT.
Chang 2008	Study is not RCT.
Demirci 2002	Quasi‐randomised study (the method of randomisation was"assigning odd‐numbered patients to the LSIG group and even‐numbered ones to the OCTR group" ‐ information received from the author).
Ettema 2006	Study is not RCT.
Fernandez‐de‐las‐Penas 2016	Study is a cohort study based on Fernandez‐de‐las Penas 2015, not a RCT.
Fernandez‐de‐las‐Penas 2018	Study is not RCT.
Fernandez‐de‐las‐Penas 2019	Study is not RCT.
Ferrari 1997	Study appears to be not RCT.
Guo 2017	Study is not RCT.
Gurcay 2017	Study is not RCT.
Guvenc 2019	Study is a case‐control study, not RCT.
Harter 1993	Study is not RCT.
Hartley 1999	The study appears to have surgery in all study groups.
Jenkins 2012	Study is not RCT.
Karamese 2015	Surgery applied to all groups in the study.
Kitsis 2002	Study is not RCT.
Mason 2017	Study terminated due to problems in the recruitment process.
Miranda 2013	Study is not RCT.
NCT00694265	Study is a prospective cohort study, not RCT.
NCT03548259	Surgery applied to all groups in the study.
NCT04347746	Surgery group made of severe CTS, not randomised.
Onuma 2015	Study is a case report, not RCT.
Roitberg 2003	Not RCT (the article is a commentary).
Rubin 2018	Study is a prospective cohort study, not RCT.
Sahoo 2020	Study is not RCT.
Siciliano 2014	Surgery applied to all groups in the study, and the study is not RCT.
Siegmund‐Schultze 2017	Not RCT (the article is a commentary).
Sparapani 2006	Not randomised, as informed by the corresponding author in a personal communication.
Tudiver 2003	Study is a short review, not RCT.
Yoshii 2015	Surgery applied to all groups in the study, and the study is not RCT.
Zannakis 2020	Study is a survey, not RCT.

Open in a new tab

CTS: carpal tunnel syndrome LSIG: local steroid injection OCTR: open carpal tunnel release RCT: randomised controlled trial

Characteristics of studies awaiting classification [ordered by study ID]

Abedi 2018.

Methods	Study design: unclear (awaiting information from the authors) Setting: Orthopedic Clinic of general Hospital of Zabol, Iran
Participants	Details of sampling frame: Total n = 100 patients Gender distribution: Open surgery: 5 males, 45 females Local injections of triamcinolone acetonide: 2 males, 48 females Mean ± SD age: Open surgery: 43.52 ± 8.6 Local injections of triamcinolone acetonide: 41.06 ± 10 Duration of symptoms: Not reported Inclusion criteria: All patients with clinical symptoms of CTS including paraesthesia, pain, motor impairment and exacerbation of symptoms at night, who had been referred to Orthopaedic Clinic of general Hospital of Zabol. Patients with moderate and mild pain. Exclusion criteria: Underlying metabolic diseases like diabetes, thyroid diseases, mellitus, rheumatoid arthritis, pregnancy, and evidence of associated radiculopathy or neuropathy along with the patient’s intention to abandon the study. Patients with severe pain.
Interventions	Open surgery Local injections of triamcinolone acetonide (2 cc of triamcinolone (Ampul 40 mg) was mixed with 1 cc of lidocaine 2%)
Outcomes	Outcomes were assessed at the baseline and after a 6‐month follow‐up. Pain (VAS) BCTQ symptom severity score BCTQ functional status score
Notes

Open in a new tab

Elwakil 2007.

Methods	Study design: unclear (awaiting information from the authors) Setting: Egypt
Participants	Details of sampling frame: Total n = 54 (60 hands) Gender distribution: Low level laser treatment (LLLT): 25 females, 2 males Surgery: 23 females, 4 males Mean ± SD age: LLLT: 49.11 ± 7.23 years Surgery: 42.65 ± 8.05 years Mean ± SD duration of symptoms: LLLT: 28.21 ± 7.03 months Surgery: 36.17 ± 4.38 months Inclusion criteria: Not reported Exclusion criteria: Not reported
Interventions	Standard open CTR surgery: Open surgical decompression was done by an incision that gently curves from the ulnar border of the palmaris longus tendon where it crosses the distal wrist crease and extends to the midpalm. The palmer fascia was incised to expose the transverse carpal ligament, which was completely divided vertically at its midpoint. The median nerve was inspected, and neurolysis could be carried out if the nerve was found to be adherent. Finally, only the skin was closed. Low level laser treatment: Helium Neon (He–Ne) laser (632.8 nm, Level Laser M300) in continuous wave (CW) mode was used. The instrument has 25 prestored programs including a program for CTS. A minimum power of 12 mW has been used. The X–Y dimensions of the area to be treated were measured and the distance between laser head and area to be treated (height) should be accurately fixed at 30 cm. The area to be treated extends from the proximal palmer crease to the distal wrist crease and laterally from the scaphoid tuberosity to the pisiform bone. It was exposed to LLLT through a sweeping computerised scanning at an angle of 30 ± 15°. According to the prestored program for CTS, the instrument will automatically deliver 3 J/cm² at an automatically measured therapy time. Each patient received 12 treatment sessions at a rate of two sessions/week.
Outcomes	The patients were evaluated at baseline and about 6 months after the treatment. Symptomatic relief Return to light and regular duty Complications
Notes	3 patients had hypertension and 5 had diabetes mellitus. Each bilaterally affected patient (n = 6 patients) was included in one treatment group for one hand and in the other treatment group for the contralateral hand. Some patients had already received some treatment modalities: splinting during work, at night, or all over the day was attempted in 15 patients (27.78%). NSAIDs were attempted in 11 patients (20.37%). None of the investigated patients had received prior local corticosteroid injection. All patients who had undergone previous non‐operative treatment had either no improvement or minimal improvement in complaints. Moreover, two patients (3.70%) reported prior open CTR on their dominating right side. Their presentations in the present study were for the symptomatic contralateral non‐dominating left side.

Open in a new tab

Hrkovic 2016.

Methods	Study design: unclear (awaiting information from the authors) Setting: Serbia
Participants	Details of sampling frame: Total n = 80 patients Gender distribution: Not reported Mean ± SD age: Not reported Duration of symptoms: Not reported Inclusion criteria: Patients with diagnosed CTS Exclusion criteria: Patients with prior injuries or surgery on the upper limbs and diseases that can cause peripheral neuropathy
Interventions	Surgery: standard open CTR Conservative treatment: such as physical therapy – application of physical agents
Outcomes	Outcomes were assessed at baseline and 12 months after treatment. BCTQ Symptom severity score (1‐5, higher is worse)* BCTQ Functional status score (1‐5, higher is worse)* Electrophysiological studies: sensory latency of median nerve terminal (distal) motor latency of median nerve
Notes	*Reported for patients in such groups ‐ for those who had physiological terminal motor latency; for those who had increased terminal motor latency; for those who had physiological sensory latency values; and for those who had increased sensory latency values

Open in a new tab

IRCT20200629047948N1.

Methods	Study design: randomised, parallel‐assignment trial Setting: Shahid Beheshti Hospital, Iran
Participants	Details of sampling frame: 40 patients have been assigned to two groups of intervention and control. Inclusion criteria: Patients aged 20‐65 years old Patients with CTS Informed consent Exclusion criteria: Neuropathy or an old nerve injury Patients with systemic disease, such as disease, arthritis, diabetes, hypothyroidism Dissatisfaction in filling consent paper
Interventions	Open surgery Injection of Depo‐Mdrol under sonographic guidance
Outcomes	Outcomes measured before treatment and 1 and 3 months after treatment. Pain (VAS) Paraesthesia (based on patient response) Numbness (based on Boston form) Weakness (based on Boston form) Return to daily activities (based on Boston form) Sonographic improvement (median nerve cross‐sectional area on ultrasound)
Notes	It seems that the study results are not yet published, therefore the study is awaiting classification.

Open in a new tab

Linscheid 1967.

Methods	Study design: unclear (not known if RCT)
Participants	Details of sampling frame: Total n = 28 patients (49 hands)* Gender distribution: Total: 21 females, 7 males Mean ± SD duration of symptoms: Total: 2 years (range from 3 months to 20 years) Inclusion criteria: Not reported Exclusion criteria: Not reported
Interventions	Surgical decompression Conservative treatment: occasional use of wrist cock‐up splints and the use of phenoxybenzamine hydrochloride
Outcomes	Follow‐up ranged from 5 months to 9 years (average, 28 months). Relief of symptoms
Notes	*The patients had carpal‐tunnel syndrome and, in addition, either Raynaud’s phenomenon or acrocyanosis.

Open in a new tab

Lo 2021.

Methods	Study design: prospective, non‐blinded study, unclear if RCT (awaiting information from the authors) Setting: single medical centre, Taiwan
Participants	Details of sampling frame: Total n randomised = 46 patients Gender distribution: Intervention group 1 (surgery): 15 females, 8 males Intervention group 2 (injection): 15 females, 8 males Mean ± SD age: Intervention group 1 (surgery): 58.9 ± 14.4 Intervention group 2 (injection): 58.2 ± 7.06 Mean ± SD duration of symptoms: Intervention group 1 (surgery): 54.4 ± 68.8 months Intervention group 2 (injection): 28.8 ± 53.0 months Inclusion criteria: Age over 20 years Typical symptoms, including numbness in the sensory distribution of the median nerve; nocturnal, postural, or motion‐associated paraesthesia ± pain in the subject’s hand that is relieved by a flicking movement of the hand A positive Phalen's test and/or presence of Tinel's sign Persistent symptoms for more than 3 months Electrophysiologically confirmed median neuropathy at the wrist, regardless of the degree of severity Exclusion criteria: Conditions that mimic CTS, such as cervical radiculopathy, polyneuropathy, brachial plexopathy, and thoracic outlet syndrome Underlying disease such as diabetes mellitus, thyroid disease, wrist osteoarthritis, chronic renal failure on haemodialysis, autoimmune diseases, or pregnancy Recent corticosteroid injection into the carpal tunnel within six months or previous CTR Allergy to corticosteroids or local anaesthetics, and Impaired cognitive function
Interventions	Endoscopic CTR surgery 2 mL perineural injection with 1 mL of 2% lidocaine hydrochloride (xylocaine) and 1 mL of 40 mg triamcinolone acetonide
Outcomes	Outcomes were assessed 12 weeks after surgery. Changes in the morphology and mobility of the median nerve VAS BCTQ Symptom severity scale (1‐5, higher is worse) BCTQ Functional status scale (1‐5, higher is worse)
Notes

Open in a new tab

NCT00981565.

Methods	Study design: randomised, parallel assignment Setting: North‐Carelia Central Hospital, Finland
Participants	Details of sampling frame: 88 participants Inclusion criteria: Electrophysiologically proven minimal or mild CTS lasting more than 6 months Exclusion criteria: Rheumatoid arthritis Diabetes mellitus Hypothyreosis Pregnancy Wrist trauma or previous surgery Splinting or corticosteroid injection on the affected side
Interventions	Open CTR Individual night‐time splinting
Outcomes	Outcomes assessed at 1 year follow‐up. Change in Symptom Severity Score Pain VAS Change in electroneuromyography
Notes	It seems that the study results are not yet published, therefore, the study is awaiting classification.

Open in a new tab

NCT04216147.

Methods	Study design: randomised, parallel assignment Setting: Universidad de Murcia, Spain
Participants	Details of sampling frame: 140 participants Inclusion criteria: Over 18 years CTS diagnosed by electromyography (EMG) Symptoms of CTS + EMG Exclusion criteria: Difficulty expressing your feelings properly Unsurpassed fear of needles History of adverse reactions to needles Epilepsy and/or allergies to metals Existence of diffuse peripheral neuropathy or cervical radiculopathy History of potential concurrent cause of idiopathic CTS (such as diabetes, thyroid, chronic rheumatoid arthritis, renal failure with haemodialysis, pregnancy etc.)
Interventions	Percutaneous Needle Electrolysis: the application of galvanic current through an acupuncture needle Surgery for CTS
Outcomes	Outcomes assessed at baseline and after treatments: 6 weeks, 3 months, 6 months and 12 months. Changes in the BCTQ Changes in the Clinical Symptoms Carpal Tunnel Syndrome Scale Change of pain level (VAS, 0‐100) Change of Semmes Weinstein Mini monofilament kit Change of the Hand Dynamometer Changes in muscles' strength by Kendall's scale Changes in SF‐12 Questionnaire Direct and indirect health cost measures
Notes	It seems that the study results are not yet published, therefore the study is awaiting classification.

Open in a new tab

Shazad 2022.

Methods	Study design: unclear (awaiting information from the authors) Setting: King Edward Medical University/Mayo Hospital Lahore, Pakistan
Participants	Details of sampling frame: Total n = 64 Gender distribution: Not reported Mean ± SD age: Surgical decompression: 45.44 ± 8.39 Local corticosteroids: 47.22 ± 9.72 Mean ± SD duration of symptoms: Surgical decompression: 5.14 (2.28) (unit unclear) Local corticosteroids: 9.51 (5.34) (unit unclear) Inclusion criteria: Cases of CTS of at least 3 months’ duration at age 18 years and above of either gender, pregnant, confirmed by electrodiagnostic testing as having moderate‐to‐severe symptoms according to Brigham and Women's Hospital Carpal Tunnel Syndrome questionnaire Exclusion criteria: Cases with open upper limb fracture, scarring around the wrist, previous carpal tunnel decompression surgery, or local steroid injection for CTS, having polyneuropathy and inflammatory arthropathy diagnosed by history, physical examination and laboratory investigation
Interventions	Surgical decompression Local corticosteroids
Outcomes	Outcomes were assessed at baseline and at 3 months, 6 months and 1 year. Symptom severity scale Functional status scale
Notes

Open in a new tab

BCTQ: Boston Carpal Tunnel Questionnaire cc: cubic centimetre cm: centimetre CTR: carpal tunnel release CTS: carpal tunnel syndrome CW: continuous wave EMG: electromyography He‐Ne: helium neon J/cm²: joules per square centimetre LLLT: low level laser treatment mg: milligram mL: millilitre mW: milliwatt NSAIDs: nonsteroidal anti‐inflammatory drugs RCT: randomised controlled trial SD: standard deviation SF‐12: 12‐Item Short Form Survey VAS: visual analogue scale

Characteristics of ongoing studies [ordered by study ID]

CTRI201901016881.

Study name	A comparison of ultrasound guided steroid injection and endoscopic carpal tunnel release for the treatment of carpal tunnel syndrome
Methods	Study design: randomised, parallel‐group trial Setting: Department of Orthopedics, Kasturba Hospital, India
Participants	Inclusion criteria: Age from 18 to 99 years Gender: both Idiopathic CTS confirmed by USG and nerve conduction studies that did not improve with 3 weeks of conservative management (splinting and medications) Patients with moderate disease as per nerve conduction studies Exclusion criteria: Patients with other concomitant neuropathies such as cervical radiculopathies, ulnar‐radial neuropathies Patients with other associated diseases like trigger finger, Raynaud’s disease Patients with anatomical variations such as a bifid median nerve or vascular anomalies
Interventions	Endoscopic carpal tunnel release Ultrasound‐guided steroid injection
Outcomes	Outcomes measured before intervention and 24 hours post‐intervention, 10 days post‐intervention, 1 month post‐intervention, 3 months post‐intervention and 6 months post‐intervention. Time before return to work Change in BCTQ score Satisfaction with procedure in terms of change in QuickDASH score Improvement in grip strength
Starting date	15 January 2019 (date of first enrolment)
Contact information	Abhijith Anil (anil.abhijith[at]gmail.com), Junior Resident at Kasturba Hospital, Manipal, address: Department of Orthopedics, Kasturba Hospital, Manipal 576104 Udupi, Karnataka 576104, India Anil K Bhat (anilbhatortho[at]gmail.com), Professor and Head at Kasturba Hospital, Manipal, address: Department of Orthopedics Kasturba Hospital, Manipal 576104 Udupi, Karnataka 576104, India
Notes

Open in a new tab

DRKS00014585.

Study name	Conservative treatment of carpal tunnel Syndrom (CTS) with the Tuttlingen CTS‐Cuff versus surgery
Methods	Study design: controlled, non‐blinded, randomised study Setting: County Clinic Tuttlingen, Germany
Participants	Inclusion criteria: CTS diagnosed Exclusion criteria: Treatment of CTS by cortisone injection within 6 weeks before entering the study Multiple compression symptoms or previously surgically treated compression symptoms in the same hand or forearm Other nerve compression syndromes than CTS in the same side hand/forearm Acute spinal disc prolapse Cervical vertebral syndrome Below the age of 18 Paresis or paralysed upper extremities Any illness of the central nervous system Previous trauma involving the median nerve Chronic pain syndrome (e.g. CRPS etc.) Problems with the application of the cuff by the patient
Interventions	Conservative therapy using a carpal tunnel extension device, the so‐called Tuttlingen CTS‐Cuff Open decompression/neurolysis of the median nerve
Outcomes	Outcomes are measured before the start of the therapy and during the check‐up 13 weeks after the start of the therapy. Clinical improvement of CTS after 3 months (DASH Score plus clinical indication) Change of the DML reported in msec
Starting date	30 June 2018 (date of first enrolment)
Contact information	Ms Dr Caterina Poma‐Schmidt (c.poma‐schmidt[at]klinikum‐tut.de; +49 (7424) 950 4491), Gesundheitszentrum Spaichingen im Klinikum Landkreis Tuttlingen, Robert‐Koch‐Straße 31 78549 Spaichingen Germany, www.klinikum‐tut.de
Notes

Open in a new tab

DRKS00020570.

Study name	The effect of stretching the transverse carpal ligament using curpal® on clinical symptoms and various neurophysiological parameters in carpal tunnel syndrome (CTS) – a prospective, randomised trial
Methods	Study design: randomised controlled trial Setting: University Medical Center Kliniken für Neurologie und Neurochirurgie, Freiburg im Breisgau, Germany
Participants	Inclusion criteria: Gender: both, male and female Minimum age 18 years Clinically and electrophysiologically proven CTS Already unsuccessful therapy with NSAIDs, hand splints and other conservative therapies (physiotherapy, occupational therapy) over a minimum period of 4 weeks Exclusion criteria: Parallel presence of polyneuropathy Traumatic injuries in the wrist area Pre‐surgery in the wrist and nerve area Arthritis in the hand area Pregnancy Ongoing legal proceedings Contact allergies to the materials used in the stretching devices Severe secondary diseases (tumour diseases, severe organ impairment) Motor deficit beyond a fine motor disorder
Interventions	Surgery Curpal® treatment
Outcomes	Outcomes assessed after 8 weeks and 6 months. Change in BCTQ score Change in SF‐8 scores Electrophysiological and sonographic parameters
Starting date	19 March 2021 (actual study start date)
Contact information	Mr Dr Christoph Scholz (christoph.scholz[at]uniklinik‐freiburg.de; 0761‐270‐50010), Universitätsklinikum Freiburg Kliniken für Neurologie und Neurochirurgie, Breisacher Str. 64 79106 Freiburg Germany, www.uniklinik‐freiburg.de/neurochirurgie.html
Notes

Open in a new tab

EUCTR2013‐000873‐56‐ES.

Study name	Randomised, two parallel groups, open clinical trial stratified by severity to estimate the cost‐effectivity of surgical vs corticosteroid injection treatment on carpal tunnel syndrome
Methods	Study design: controlled, randomised, open, parallel‐group trial Setting: Hospital Universitario de Canarias, Spain
Participants	Inclusion criteria: Informed consent Patients between 18 and 65 who are able to answer written questionnaires Satisfying 1 or more of the following symptoms: pain, paraesthesias and/or hypoesthesia in the territory of the median nerve in the hands Nerve conduction delay in the electromyogram If the patient has bilateral clinically confirmed by electromyogram, the hand included will be chosen by the patient Exclusion criteria: Patients with known thyroid disease, rheumatic diseases, diabetes, polyneuropathy, chronic renal failure or alcoholism Patients who were treated previously with corticoid injection in the affected hand Patients with any pathology in the affected limb of CTS Patients with suggestive clinical cervicobrachalgia or nerve compression at any other levels Presence of thenar atrophy or irreversible injury to the median nerve (nerve conduction absence measured by electromyogram) Contraindication to treatment with corticosteroids (systemic fungal infection or hypersensitivity to corticosteroids or other components of the injection) Patients with any contraindication to surgery Serious psychiatric condition Evidence of uncontrolled concomitant serious illness Pregnancy
Interventions	Local steroid injection Surgery
Outcomes	Outcomes measured at baseline and at 12 months. Effectiveness (quality of life: test DASH) Direct costs Effectiveness: quality of life by the Boston test Physical examination: Dermal sensitivity (Weber test), trigger tests, hand force, wrist mobility, complications, time to return to daily activity, adverse events
Starting date	20 May 2013 (Date on which the record was first entered in the EudraCT database)
Contact information	Federico Diaz González (federico.diaz.gonzalez[at]gmail.com, 034922646792), Hospital Universitario de Canarias, Ofra s/n. La Cuesta, La Laguna 38320 Spain
Notes

Open in a new tab

EUCTR2021‐004756‐42‐NO.

Study name	A Norwegian trial comparing treatment strategies for Carpal Tunnel Syndrome (NOR‐CACTUS)
Methods	Study design: randomised, parallel‐assignment trial Setting: hospitals in Norway
Participants	Inclusion criteria: Adult (≥ 18 years of age) Patient history indicating CTS Neurophysiological examination performed within 6 months; Mild‐to‐moderate symptoms (intermittent, interfering with everyday life, and/or disturbed sleep) Diagnosis of CTS based on: Classic/probable or possible symptoms, and neurophysiological findings consistent with CTS; or, in the case of normal neurophysiological findings Classic/probable symptoms and positive physical exam findings and/or night‐time symptoms Exclusion criteria: Previous CTS surgery or corticosteroid injection in the carpal tunnel in the relevant hand Diagnosis of severe CTS, based on history and examination indicating severe CTS with constant symptoms including pain, loss of sensibility, dexterity or reduced temperature sensation, weakness of thumb abduction and opposition, or atrophy of thenar musculature. Disappearance of pain may indicate permanent sensory loss History suggesting underlying causes of CTS e.g. inflammatory wrist arthritis and/or flexor tenosynovitis Previous significant trauma or fracture, deformity or tumour in the wrist or hand in the relevant hand Presence of conditions affecting a normal nerve function, e.g. cervical disc herniation, polyneuropathy or previous nerve injury Major comorbidities, such as severe malignancies, severe or uncontrolled infections, uncontrollable hypertension, severe cardiovascular disease (NYHA class III or IV) and/or severe respiratory diseases, severe renal failure, active ulcus ventriculi, leukopenia and/or thrombocytopenia Severe psychiatric or mental disorders Local infection or wound in the affected hand/wrist Any other medical condition that according to the treating physician and/or local guidelines makes adherence to treatment protocol impossible Inadequate birth control, pregnancy, and/or breastfeeding (current at screening or planned within the duration of the study) Known hypersensitivity to triamcinolone hexacetonide (Lederspan) or any of the excipients (sorbitol, polysorbate or benzyl alcohol) Concomitant therapy with CYP3A‐inhibitors or digitalis glycosides Patients vaccinated or immunised with live virus vaccines within 2 weeks of treatment Alcohol or other substance abuse Language barriers Other factors which make adherence to study protocol impossible
Interventions	Surgical carpal tunnel release Corticosteroid injection (ultrasound‐guided injection of 1 mL/20 mg of triamcinolone hexacetonide into the carpal tunnel space close to the median nerve)
Outcomes	Outcomes assessed at 3, 6, 12, and 24 months. Primary outcomes: Successful treatment result after 12 months (BCTQ Symptom Severity Scale ≤ 1.5) Secondary outcomes: Successful treatment result after 3, 6 and 24 months (BCTQ Symptom Severity Scale ≤ 1.5) BCTQ Symptom Severity Scale BCTQ Functional Status Scale Nerve conduction studies: motor median nerve distal latency (milliseconds) Nerve conduction studies: motor median nerve proximal latency (milliseconds) Nerve conduction studies: motor median nerve amplitude (millivolts) Nerve conduction studies: motor median nerve conduction velocity (meters per second) Nerve conduction studies: motor ulnar nerve distal latency (milliseconds) Nerve conduction studies: motor ulnar nerve proximal latency (milliseconds) Nerve conduction studies: motor ulnar nerve amplitude (millivolts) Nerve conduction studies: motor ulnar nerve conduction velocity (meters per second) Nerve conduction studies: sensory median nerve latency (milliseconds) Nerve conduction studies: sensory median nerve amplitude (microvolts) Nerve conduction studies: sensory median nerve conduction velocity (meters per second) Nerve conduction studies: sensory ulnar nerve latency (milliseconds) Nerve conduction studies: sensory ulnar nerve amplitude (microvolts) Nerve conduction studies: sensory ulnar nerve conduction velocity (meters per second) Electromyography recording spontaneous activity in m.abductor pollicis brevis (Yes/No) Electromyography recording chronic neurogenic changes in m.abductor pollicis brevis (Yes/No) Scoring of motor and sensory nerve conduction studies using the Bland scoring system (Bland 2000) (from 0 = normal to 7 = not gradable) Semmes‐Weinstein monofilament test (absent/present) Grip strength (range 0‐90 kg) Grip ability test Patient‐reported assessment of treatment effect on symptoms (current state compared to before treatment), reported on a 5‐level ordinal (Likert) scale from 1 (best) to 5 (worst) Patient‐reported assessment of CTS‐related symptoms (current state), reported on a 5‐level ordinal (Likert) scale from 1 (best) to 5 (worst) Patient‐reported assessment of current CTS‐related functional impairment (current state), reported on a 5‐level ordinal (Likert) scale from 1 (best) to 5 (worst) Patient assessment of acceptability of CTS‐related symptoms and functional disability (yes or no) Ultrasound measure of proximal cross‐sectional area of the median nerve (square millimetres) Ultrasound measure of distal cross‐sectional area of the median nerve (square millimetres) Ultrasound measures of vascularity of the median nerve in the carpal tunnel Disabilities of the Arm, Shoulder, and Hand (Quick‐DASH) (11‐item patient reported outcome measure for disabilities of the arm, shoulder, and hand) Patient‐reported CTS‐related pain on a 0‐100 mm visual analogue scale (VAS) (from 0 = no pain to 100 = worst possible pain) Work Productivity and Activity Impairment Questionnaire Days of work absence since start of intervention EuroQoL 5‐dimension health‐related quality of life (EQ‐5D‐5L); Adverse events (number and nature of adverse events and serious adverse events) Successful treatment after 1 corticosteroid injection (subjects who have received 1 corticosteroid injection who have attained a successful treatment result (BCTQ Symptom Severity Scale ≤ 1.5) without surgery) Successful treatment after 2 corticosteroid injections (subjects who have received 2 corticosteroid injections who have attained a successful treatment result (BCTQ Symptom Severity Scale ≤ 1.5) without surgery) Successful treatment after secondary surgery (only applicable in the injection treatment strategy arm. Subjects who have received 1 or 2 corticosteroid injections and secondary surgery who have attained a successful treatment result (BCTQ Symptom Severity Scale ≤ 1.5)) Undergone reoperation (received a re‐operation (yes/no)) Cost of treatment Other outcome measures: Use of hospital services Use of primary care resources Work participation Total cost
Starting date	8 April 2022 (actual study start date)
Contact information	Ulf G Sundin (uffe.sundin[at]gmail.com, +4740614198), Department of Surgery and Anesthesiology, Diakonhjemmet Hospital, Oslo, Norge, Norway, 1450
Notes

Open in a new tab

IRCT20141121020020N8.

Study name	Comparison of long term effectiveness of open surgery versus ultrasound guided median nerve hydrodissection in terms of symptoms, function and electrophysiological studies in patients with Carpal Tunnel syndrome (CTS)
Methods	Study design: RCT Setting: Shahid Beheshti Hospital in Babol, Iran
Participants	Inclusion criteria: All patients with CTS referred to Shahid Beheshti Hospital in Babol during 1397‐98; age from 20 years to 65 years Exclusion criteria: Not reported
Interventions	Open surgery Ultrasound guided median nerve hydrodissection
Outcomes	Outcomes assessed at baseline and 1, 3 and 6 months. The rate of improvement in patients' symptoms (score between 11‐55 for symptoms and 8‐40 for function) The rate of improvement of patients' electrophysiological parameters
Starting date	23 August 2021 (date of first enrolment)
Contact information	Mehrafza Mir (sh.seyfi[at]yahoo.com, 009111139348), Gangh afroz Road‐Keshavarz Street 4717641367, Babol Iran
Notes

Open in a new tab

NCT04014244.

Study name	Efficacy of dextrose injections, corticosteroid injections and surgical release for treatment of carpal tunnel syndrome: a prospective, randomised, double‐blind controlled trial
Methods	Study design: prospective, randomised, double‐blind controlled trial Setting: University Medical Center Ljubljana, Slovenia
Participants	Inclusion criteria: Numbness and tingling in at least 2 out of the first 4 digits Symptoms exacerbated during the night or in the morning, by holding objects or repetitive use of the hand and/or symptoms relieved by shaking the hand or reposition of the arm Exclusion criteria: Polyneuropathy, symptoms of polyneuropathy, and all conditions causing polyneuropathy (e.g. diabetes) Multiple mononeuropathy (e.g. hereditary neuropathy with liability to pressure palsies, syndrome Lewis Sumner) Motor neurone disorders (e.g. monomelic amyotrophy, amyotrophic lateral sclerosis) Brachial plexopathy Other focal neuropathies affecting the upper limbs (e.g. other median neuropathies, ulnar neuropathies, thoracic outlet syndrome) Previous surgery or local injections for CTS
Interventions	Triamcinolone injections or 5% dextrose Surgery
Outcomes	Outcomes assessed at baseline and at 1 year follow‐up. Digital paraesthesia/dysaesthesia and wrist or hand pain on 11‐point grading scale Global assessment of treatment results BCTQ score Median nerve distal motor latency Ulnar and median antidromic sensory nerve action potentials Median nerve cross‐sectional area Palmar bowing of the flexor retinaculum
Starting date	1 March 2021 (actual study start date)
Contact information	Gregor Omejec, PT, DSc (gregor.omejec[at]kclj.si, +386 1 522 1502) Simon Podnar, MD (simon.podnar[at]kclj.si, +386 1 522 3076)
Notes

Open in a new tab

NCT04058041.

Study name	Effectiveness of neural mobilization exercises versus surgery in patients with carpal tunnel syndrome
Methods	Study design: randomised parallel‐assignment trial Setting: Universidad Rey Juan Carlos, Spain
Participants	Inclusion criteria: Tinel's sign and Phalen's test positive 4/10 on VAS Increased symptoms at night during 12 months of symptomatology Sensorial and motor dysfunction in the conduction of the median nerve Exclusion criteria: Previous surgery Wrist fractures Rheumatoid arthritis or fibromyalgia Systemic disease Pregnancy
Interventions	Surgery Neural mobilisation Surgery and neural mobilisation
Outcomes	Outcomes assessed at baseline and at 6 months. Physiological parameters Neuropathic pain (Questionnaire DN4) Pain intensity (VAS, 0‐100 mm, higher is worse) Pain expansion Pressure pain threshold Psychological factors: Anxiety (STAI‐T), depression (BDI‐II), kinesiophobia (Tampa scale), catastrophising (Pain Catastrophising scale), fear avoidance (Fear Avoidance Belief questionnaire) DASH BCTQ Grip strength without pain Pinch strength
Starting date	12 March 2018 (actual study start date)
Contact information	Josué Fernández, PhD (josue.fernandez[at]urjc.es, 34914888949) Luis Matesanz (luis.matesanzgarcia[at]gmail.com, 675042482)
Notes

Open in a new tab

NCT04246216.

Study name	Ultrasound‐guided percutaneous electrical nerve stimulation of the median nerve for carpal tunnel syndrome
Methods	Study design: randomised, parallel‐assignment trial Setting: Universidad Rey Juan Carlos, Spain
Participants	Inclusion criteria: Pain and paraesthesia in the median nerve distribution Positive Tinel's sign Positive Phalen's test Symptoms had to have persisted for at least 6 months Deficits of sensory and motor median nerve conduction according to guidelines of the American Association of Electrodiagnosis, American Academy of Neurology, and the American Physical Medicine and Rehabilitation Academy Exclusion criteria: Any sensory/motor deficit related to the ulnar or radial nerve Older than 65 years of age Previous surgical intervention, steroid injections or physical therapy intervention Multiple diagnoses of the upper extremity (e.g. cervical radiculopathy) History of neck, shoulder, or upper limb trauma (whiplash) History of any systemic disease causing CTS (e.g. diabetes mellitus, thyroid disease) History of other systemic conditions (e.g. rheumatoid arthritis, fibromyalgia) Pregnancy
Interventions	Percutaneous electrical nerve stimulation Endoscopic surgery of the carpal tunnel
Outcomes	Outcomes assessed at baseline and 1, 3, 6 and 12 months after the intervention. Changes in pain intensity (11‐point NPRS, 0‐10) BCTQ scores Self‐perceived improvement (Global Rating of Change, 15‐point scale ranging from ‐7 to +7)
Starting date	1 March 2020 (actual study start date)
Contact information	César Fernández‐de‐las‐Peñas, Alcorcon, Madrid, Spain, 28922
Notes

Open in a new tab

Palmbergen 2022.

Study name	Dutch Injection versus Surgery TRIal in patients with CTS (DISTRICTS)
Methods	Study design: multi‐centre, open‐label RCT Setting: approximately 30 Dutch hospitals
Participants	Inclusion criteria: Patients with clinically suspected CTS, which is confirmed by electrophysiological or sonographic testing and for which surgery and injection are both potential treatment options Symptoms of CTS that have been present for at least 6 weeks and treatment initiated within 6 weeks following inclusion 18 years or older at time of examination. Patients can participate using the most affected hand only in case both hands are eligible. Exclusion criteria: Previous surgery for CTS on the ipsilateral wrist An injection for CTS in the ipsilateral wrist less than 1 year ago Previous participation in the DISTRICTS Clinical or neurophysiological suggestion that the symptoms are due to another diagnosis Not able to comprehend Dutch self‐report questionnaires Pregnancy Follow‐up not possible Legally incompetent adults and No informed consent
Interventions	Surgery Corticosteroid injection If needed, these treatments can be followed by any additional treatments within the 18 months of follow‐up, such as a second injection or surgical treatment.
Outcomes	Outcomes assessed at baseline, at 6 weeks and 3, 6, 9, 12, 15 and 18 months. Proportion of participants recovered at 18 months. Recovery is defined as having no or mild symptoms; that is, a score of less than eight points on the 6‐item carpal tunnel symptoms scale (CTS‐6) Time to recovery at 18‐months follow‐up Proportion of participants recovered at all time points during the 18‐month follow‐up Symptom severity at all time points during 18‐month follow‐up Upper limb functioning at 18 months measured using QuickDASH Severity of pain in the scar/palm and pain‐related activity limitation during 18‐month follow‐up using the Palmar Pain scale Participant’s global perception of recovery at 18 months measured with a seven‐point Likert‐type item Participant’s satisfaction at 18 months measured with a seven‐point Likert‐type item Health‐related quality of life at 18 months assessed with the EuroQol 5‐level EQ‐5D (EQ‐5D‐5L) Number and type of additional treatments during 18‐month follow‐up Adverse events during 18‐month follow‐up Use of care and health‐related costs during follow‐up assessed with an adapted version of the Medical Consumption Questionnaire (iMCQ) and the Productivity Cost Questionnaire (iPCQ)
Starting date	November 2017 (recruitment of participants started)
Contact information	Dr Camiel Verhamme (c. verhamme[at]amsterdamumc.nl)
Notes

Open in a new tab

BDI‐II: Beck Depression Inventory‐II BCTQ: Boston Carpal Tunnel Questionnaire CRPS: complex regional pain syndrome CTS: carpal tunnel syndrome CTS‐6: 6‐item carpal tunnel symptoms scale DASH: Disabilities of the Arm, Shoulder and Hand DISTRICTS: short title of the trial (Dutch Injection versus Surgery TRIal in Carpal Tunnel Syndrome) DN4: Douleur Neuropathique 4 Questions EQ‐5D‐5L: EuroQol 5‐dimension 5‐level health status and quality of life measure EuroQoL: EuroQol Group iMCQ: Medical Consumption Questionnaire iPCQ: Productivity Cost Questionnaire kg: kilogram mg: milligram mL: millilitre mm: millimetre msec: millisecond NOR‐CACTUS: short title of the trial (a Norwegian trial comparing treatment strategies for Carpal Tunnel Syndrome) NSAID: nonsteroidal anti‐inflammatory drug NYHA: New York Heart Association QuickDASH: 11‐item Disabilities of the Arm, Shoulder, and Hand scale RCT: randomised controlled trial SF‐8: 8‐item Short‐Form Health Survey STAI‐T: State‐Trait Anxiety Inventory USG: ultrasound sonography VAS: visual analogue scale

Differences between protocol and review

Changes to the scope

We assessed the evidence regarding the benefits and harms of carpal tunnel release compared with non‐surgical treatment in the short (up to 3 months) and long (over 3 months) term. Due to the need for a practical application, we performed a separate meta‐analysis for each comparison (surgery versus splinting, surgery versus corticosteroid injection, etc.) at two different time points (short‐term and long‐term). In the previous version of this review (Verdugo 2008), the objective was to compare the efficacy of surgical treatment of carpal tunnel syndrome with non‐surgical treatment, pooling all studies together regardless of the comparison.

Changes to the methods

Criteria for considering studies for this review (types of outcome measures)

Primary outcome: as there is currently no standard measure of clinical improvement, in this update, we accepted any definition of clinical success or clinical improvement as long as it was based on the subjective assessment of the participants. In the previous version of this review (Verdugo 2008), the primary outcome measure was:
- relevant clinical improvement after 3 months of follow‐up (the improvement was considered relevant if it implied significant relief of pain and paraesthesiae, by at least 50% of the baseline level (Verdugo 1994)); or
- improvement of hypoaesthesia or muscle weakness resulting in improvement in quality of life and functional status.
Also, the previous version of this review included other measures of clinical improvement as the secondary outcomes:
- clinical improvement reported by authors without including its relevance to the functional status of the participant, for example, better performance with the 2‐point discrimination test;
- clinical improvement at less than 3 months of follow‐up;
- clinical improvement at 1 year of follow‐up.
Secondary outcomes:
- We excluded the following secondary outcomes, which were included in the previous version of this review (Verdugo 2008):
  - neurophysiological parameters because their clinical relevance is unclear;
  - return to work at 3 months or less of follow‐up because we focused on long‐term outcomes.
- We included the following secondary outcomes, which were not included in the previous version of this review (Verdugo 2008):
  - symptoms;
  - function (disease or upper limb specific patient‐rated outcome measure);
  - pain;
  - generic health‐related quality of life.

We included these outcomes as they are informative patient‐reported outcomes, and they are commonly used in carpal tunnel syndrome trials, as well as included in most of the Cochrane reviews dealing with carpal tunnel syndrome. Pain was included because surgery may cause chronic pain such as painful scar.

Search methods for identification of studies

Differences in search strategies are illustrated in the Appendices of this review.

Data collection and analysis

Sensitivity analysis: same as the previous version of this review, we performed sensitivity analysis excluding those with high or unclear ('some concerns') risk of bias in the randomisation process because inadequate allocation concealment is found to be associated with overestimated treatment effects. Additionally, we planned analysis excluding all studies with high or 'some concerns' in any domain but did not perform these since there were no studies with a low risk of bias in all domains.
Data collection: we used the number of participants available for analysis at the follow‐up time points both for dichotomous and continuous outcomes. The previous version of this review partly used the number of participants randomised (Verdugo 2008).
Risk of bias: we used the risk of bias (RoB) version 2 tool to assess the methodological quality of the included studies (Sterne 2019). The previous version of this review assessed the methodological quality of included studies by judging if allocation concealment, diagnostic criteria, baseline differences, patient blinding and observer blinding were appropriate (Verdugo 2008).
Data interpretation: in contrast to the previous version of this review, we assessed the certainty of the evidence as high, moderate, low or very low using the GRADE approach (Schunemann 2017). Also, we used minimal clinically important difference values to assess the clinical importance of differences in patient‐reported outcomes.

Contributions of authors

VIEDA LUSA (VL) was involved in the screening of the search results and assessing full texts; appraising the risk of bias; extracting data; performing analyses; interpreting analyses including certainty of evidence and writing of the review.

TEEMU KARJALAINEN (TK) was involved in planning the update; as arbiter in the screening of the search results; assessing full texts and the risk of bias; interpreting analyses, including the certainty of evidence, and writing of the review.

TUOMAS RAJAMAKI (TR) was involved in the screening of the search results and assessing the full texts; appraising the risk of bias; extracting data and editing the review.

KATI JAATINEN (KJ) was involved in screening of the search results and assessing the full texts; appraising the risk of bias; extracting data and editing the review.

MARKUS PAAKKONEN (MP) was involved in the screening of the search results and assessing the full texts; appraising the risk of bias; extracting data and editing the review.

Sources of support

Internal sources

No sources of support provided

External sources

Teemu Karjalainen, Finland

TK is supported by a Finnish Medical Foundation research grant.
Vieda Lusa, Finland

VL is supported by a Finnish Medical Foundation research grant.
Cochrane Neuromuscular, UK

This project was supported by the UK National Institute for Health and Care Research (NIHR) via Cochrane Infrastructure funding to Cochrane Neuromuscular until 31 March 2023. The views and opinions expressed herein are those of the authors and do not necessarily reflect those of the Evidence Synthesis Programme, NIHR, National Health Service (NHS) or the Department of Health and Social Care. Cochrane Neuromuscular is also supported by the Queen Square Centre for Neuromuscular Disease.

Declarations of interest

VL declares no conflicts of interests in this review.

TK declares working in private practice treating varying upper limb musculoskeletal problems for 6 to 9 hours per week including people with carpal tunnel syndrome.

TR declares no conflicts of interests in this review.

KJ declares working part‐time as a physiotherapist via trade name in a private medical centre (Mehiläinen Oy).

MP declares support for attending meetings and travel (KSL Martin, wrist surgery course, Berlin, Germany, October 2022, and Stryker Corporation, 2‐day course on finger endoprosthesis surgery, Amsterdam, Netherlands, October 2019), and working in private practice (as an entrepreneur consultant hand surgeon at the premises Mehiläinen hospital, Turku, 1 day per week, and occasional private practice as hand surgeon consultant at Pihlajalinna hospital, 3 to 4 hours per week).

New search for studies and content updated (conclusions changed)

References

References to studies included in this review

Awan 2015 {published data only}

Awan AS, Khan A, Afridi SA, Khan IU, Bhatti SN, Ahmed E, et al. Early response of local steroid injection versus miniincision technique in treatment of carpal tunnel syndrome. Journal of Ayub Medical College, Abbottabad : JAMC 2015;27(1):192-6. [PubMed] [Google Scholar]

Eltabl 2020 {published data only}

Eltabl MA, Saif DS, Alemam SE. Platelet-rich plasma injection versus surgical and medical treatment of mild-moderate carpal tunnel syndrome. Egyptian Journal of Neurology, Psychiatry and Neurosurgery 2020;56(1):88. [Google Scholar]
NCT04235426. Platelet rich plasma injection versus surgical and medical treatment of mild-moderate carpel tunnel syndrome. ClinicalTrials.gov/show/NCT04235426 (first posted 21 January 2020).

Fernandez‐de‐las‐Penas 2015 {published data only}

Fernández-de-Las Peñas C, Ortega-Santiago R, la Llave-Rincón AI, Martínez-Perez A, Fahandezh-Saddi Díaz H, Martínez-Martín J, et al. Manual physical therapy versus surgery for carpal tunnel syndrome: a randomized parallel-group trial. The Journal of Pain : Official Journal of the American Pain Society 2015;16(11):1087-94. [DOI: ] [DOI] [PubMed] [Google Scholar]
Fernández-de-Las-Peñas C, Arias-Buría JL, Cleland JA, Pareja JA, Plaza-Manzano G, Ortega-Santiago R. Manual therapy versus surgery for carpal tunnel syndrome: 4-year follow-up from a randomized controlled trial. Physical Therapy 2020;100(11):1987-96. [DOI: ] [DOI] [PubMed] [Google Scholar]
Fernández-de-Las-Peñas C, Ortega-Santiago R, Díaz HF, Salom-Moreno J, Cleland JA, Pareja JA, et al. Cost-effectiveness evaluation of manual physical therapy versus surgery for carpal tunnel syndrome: evidence from a randomized clinical trial. Journal of Orthopaedic & Sports Physical Therapy 2019;49(2):55-63. [DOI] [PubMed] [Google Scholar]
NCT01789645. Physical therapy in carpal tunnel syndrome. clinicaltrials.gov/show/NCT01789645 (first posted 12 February 2013).

Fernandez‐de‐las‐Penas 2017a {published data only}

Fernández-de-Las-Peñas C, Cleland J, Palacios-Ceña M, Fuensalida-Novo S, Alonso-Blanco C, Pareja JA, et al. Effectiveness of manual therapy versus surgery in pain processing due to carpal tunnel syndrome: a randomized clinical trial. European Journal of Pain (London, England) 2017;21(7):1266-76. [DOI: ] [DOI] [PubMed] [Google Scholar]
NCT02219919. Nociceptive gain processing after physical therapy in carpal tunnel syndrome. clinicaltrials.gov/show/NCT02219919 (first posted 19 August 2014).

Fernandez‐de‐las‐Penas 2017b {published data only}

Fernández-de-Las-Peñas C, Cleland J, Palacios-Ceña M, Fuensalida-Novo S, Pareja JA, Alonso-Blanco C. The effectiveness of manual therapy versus surgery on self-reported function, cervical range of motion, and pinch grip force in carpal tunnel syndrome: a randomized clinical trial. Journal of Orthopaedic and Sports Physical Therapy 2017;47(3):151-61. [DOI: ] [DOI] [PubMed] [Google Scholar]

Garland 1964 {published data only}

Garland H, Langworth EP, Taverner D, Clark JM. Surgical treatment for the carpal tunnel syndrome. Lancet 1964;I(7343):1129-30. [DOI] [PubMed] [Google Scholar]

Gerritsen 2002 {published data only}

Gerritsen AA, De Vet HC, Scholten RJ, Bertelsmann FW, De Krom MC, Bouter LM. Betere klinische effecten op carpale-tunnelsyndroom van operatie dan van spalkbehandeling; resultaten van een gerandomiseerd klinisch onderzoek. Nederlands Tijdschrift voor Geneeskunde 2002;146(45):2153-6. [Google Scholar]
Gerritsen AA, Vet HC, Scholten RJ, Bertelsmann FW, Krom MC, Bouter LM. Splinting versus surgery in the treatment of carpal tunnel syndrome. A randomized controlled trial. Journal of the American Medical Association 2002;288(10):1245-51. [DOI] [PubMed] [Google Scholar]
Gerritsen AM, Scholten RJ, Assendelft WJ, Kuiper H, Vet HC, Bouter LM. Splinting or surgery for carpal tunnel syndrome? Design of a randomized controlled trial. BMC Neurology 2001;1:8. [DOI] [PMC free article] [PubMed] [Google Scholar]
ISRCTN18853827. Splinting or surgery for carpal tunnel syndrome. isrctn.com/ISRCTN18853827 (first registered 11 February 2002).
Korthals-de Bos IB, Gerritsen AA, Tulder MW, Rutten-van Mölken MP, Adèr HJ, Vet HC, et al. Surgery is more cost-effective than splinting for carpal tunnel syndrome in the Netherlands: results of an economic evaluation alongside a randomized controlled trial. BMC Musculoskeletal Disorders 2006;16(7):86. [DOI] [PMC free article] [PubMed] [Google Scholar]
Schrijver HM, Gerritsen AA, Strijers RL, Uitdehaag BM, Scholten RJ, Vet HC, et al. Correlating nerve conduction studies and clinical outcome measures on carpal tunnel syndrome: lessons from a randomized controlled trial. Journal of Clinical Neurophysiology : Official Publication of the American Electroencephalographic Society 2005;22(3):216–21. [PubMed] [Google Scholar]

Hui 2005 {published and unpublished data}

Hui AC, Wong S, Leung CH, Tong P, Mok V, Poon D, et al. A randomized controlled trial of surgery versus steroid injection for carpal tunnel syndrome. Neurology 2005;64(12):2074-8. [PMID: 15985575] [DOI] [PubMed] [Google Scholar]

Ismatullah 2013 {published data only}

Ismatullah I. Local steroid injection or carpal tunnel release for carpal tunnel syndrome - which is more effective? Journal of Postgraduate Medical Institute 2013;27(2):194–9. [Google Scholar]

Jafari 2018 {published data only}

Irct2017011832046N. Comparing effectiveness of local steroid injection into carpal tunnel with surgery in the treatment of patients with mild to moderate carpal tunnel syndrome. https://www.irct.ir/trial/25063 (first registered 6 July 2017).
Jafari D, Shariatzadeh H, Mazhar FN, Babaei R, Mirkazemi M, Ayazi N. Comparison of the efficacy of carpal tunnel release vs. local steroid injection in the management of mild to moderate carpal tunnel syndrome: a clinical trial. Journal of Research in Orthopedic Science 2018;5(1):e58159. [Google Scholar]

Jarvik 2009 {published data only}

Jarvik JG, Comstock BA, Kliot M, Turner JA, Chan L, Heagerty PJ, et al. Surgery versus non-surgical therapy for carpal tunnel syndrome: a randomised parallel-group trial. Lancet 2009;374(9695):1074-81. [DOI] [PubMed] [Google Scholar]
Martin BI, Levenson LM, Hollingworth W, Kliot M, Heagerty PJ, Turner JA, et al. Randomized clinical trial of surgery versus conservative therapy for carpal tunnel syndrome [ISRCTN84286481]. BMC Musculoskeletal Disorders 2005;6(2):NA. [DOI] [PMC free article] [PubMed] [Google Scholar]
NCT00032227. Carpal tunnel syndrome: diagnosis and treatment trial. clinicaltrials.gov/show/NCT00032227 (first posted 13 March 2002).

Ly Pen 2005 {published data only}

Andreu JL, Ly-Pen D, Millán I, Blas G, Sánchez-Olaso A. Local injection versus surgery in carpal tunnel syndrome: neurophysiologic outcomes of a randomized clinical trial. Clinical Neurophysiology 2014;125(7):1479-84. [DOI] [PubMed] [Google Scholar]
ISRCTN26264638. Surgical decompression versus local steroid injection in carpal tunnel syndrome. isrctn.com/ISRCTN26264638 (first registered 2 February 2012).
Ly-Pen D, Andreu JL, Millán I, Blas G, Sánchez-Olaso A. Long-term outcome of local steroid injections versus surgery in carpal tunnel syndrome: observational extension of a randomized clinical trial. Hand (New York, N.Y.) 2020;Aug 6 [Epub ahead of print]:NA. [DOI: 10.1177/1558944720944263] [DOI] [PMC free article] [PubMed] [Google Scholar]
Ly-Pen D, Andréu JL, Millán I, Blas G, Sánchez-Olaso A. Comparison of surgical decompression and local steroid injection in the treatment of carpal tunnel syndrome: 2-year clinical results from a randomized trial. Rheumatology (Oxford, England) 2012;51(8):1447-54. [DOI] [PubMed] [Google Scholar]
Ly-Pen D, Andréu JL, Blas G, Sánchez-Olaso A, Millán I. Surgical decompression versus local steroid injection in carpal tunnel syndrome. A one-year, prospective, randomized, open, controlled clinical trial. Arthritis & Rheumatism 2005;52(2):612-9. [MEDLINE: ] [DOI] [PubMed] [Google Scholar]
Salinas M, Blas G, Regidor I, LyPen D, Andreu J, Sanchez-Olaso A. An electro-clinical comparison of carpal tunnel syndrome therapy. Muscle & Nerve 2003;28(S12):S178, Abstract no: 505. [Google Scholar]

Ucan 2006 {published data only}

Ucan H, Yagci I, Yilmaz L, Yagmurlu F, Keskin D, Bodur H. Comparison of splinting, splinting plus local steroid injection and open carpal tunnel release outcomes in idiopathic carpal tunnel syndrome. Rheumatology International 2006;27(1):45-51. [DOI] [PubMed] [Google Scholar]
Yağcı İ, Uçan H, Yılmaz L, Yağmurlu F, Keskin ED, Bodur H. Comparison of splinting, splinting plus local steroid injection and surgery in carpal tunnel syndrome treatment. Turkiye Fiziksel Tip ve Rehabilitasyon Dergisi 2006;52(2):55–60. [Google Scholar]

Zhang 2019 {published data only}

ChiCTR1800014530. The effectiveness of ultrasound-guided steroid injection combined with miniscalpel-needle release in the treatment of carpal tunnel syndrome: a randomized controlled study. https://www.chictr.org.cn/showprojEN.html?proj=24826 (first registered 19 January 2018). [DOI] [PMC free article] [PubMed]
Zhang S, Wang F, Ke S, Lin C, Liu C, Xin W, et al. The effectiveness of ultrasound-guided steroid injection combined with miniscalpel-needle release in the treatment of carpal tunnel syndrome vs. steroid injection alone: a randomized controlled study. BioMed Research International 2019;Feb 24 [Epub ahead of print]:NA. [DOI: 10.1155/2019/9498656] [DOI] [PMC free article] [PubMed] [Google Scholar]

References to studies excluded from this review

Baker 2014 {published data only}

Baker NA, Livengood HM. Symptom severity and conservative treatment for carpal tunnel syndrome in association with eventual carpal tunnel release. Journal of Hand Surgery 2014;39(9):1792-8. [DOI] [PubMed] [Google Scholar]

Bozkurt 2021 {published data only}

Bozkurt M, Tatar BE, Karakol P, Sezgic M, Uslu C, Gelbal C, et al. Matriderm and platelet-rich plasma combination in the treatment of recurrent carpal tunnel syndrome: a new approach. European Journal of Plastic Surgery 2021;na:na. [DOI: 10.1007/s00238-021-01863-9] [DOI] [Google Scholar]

Celik 2016 {published data only}

Celik G, Ilik MK. Effects of two different treatment techniques on the recovery parameters of moderate carpal tunnel syndrome: a six-month follow-up study. Journal of Clinical Neurophysiology : Official Publication of the American Electroencephalographic Society 2016;33(2):166-70. [DOI] [PubMed] [Google Scholar]

Cha 2016 {published data only}

Cha SM, Shin HD, Ahn JS, Beom JW, Kim DY. Differences in the postoperative outcomes according to the primary treatment options chosen by patients with carpal tunnel syndrome: conservative versus operative treatment. Annals of Plastic Surgery 2016;77(1):80-4. [DOI] [PubMed] [Google Scholar]

Chang 2008 {published data only}

Chang CW, Wang YC, Chang KF. A practical electrophysiological guide for non-surgical and surgical treatment of carpal tunnel syndrome. Journal of Hand Surgery: European Volume 2008;33(1):32-7. [DOI] [PubMed] [Google Scholar]

Demirci 2002 {published data only}

Demirci S, Kutluhan S, Koyuncuoglu HR, Kerman M, Heybeli N, Akkuş S, et al. Comparison of open carpal tunnel release and local steroid treatment outcomes in idiopathic carpal tunnel syndrome. Rheumatology International 2002;22(1):33-7. [DOI] [PubMed] [Google Scholar]

Ettema 2006 {published data only}

Ettema AM, Amadio PC, Cha SS, Harrington JR, Harris AM, Offord KP. Surgery versus conservative therapy in carpal tunnel syndrome in people aged 70 years and older. Plastic and Reconstructive Surgery 2006;118(4):947-58. [DOI] [PubMed] [Google Scholar]

Fernandez‐de‐las‐Penas 2016 {published data only}

Fernández-de-Las-Peñas C, Cleland JA, Salom-Moreno J, Palacios-Ceña M, Martínez-Perez A, Pareja JA, et al. Prediction of outcome in women with carpal tunnel syndrome who receive manual physical therapy interventions: a validation study. Journal of Orthopaedic and Sports Physical Therapy 2016;46(6):443–51. [DOI] [PubMed] [Google Scholar]

Fernandez‐de‐las‐Penas 2018 {published data only}

Fernández-de-Las-Peñas C, Ambite-Quesada S, Fahandezh-Saddi Díaz H, Paras-Bravo P, Palacios-Ceña D, Cuadrado ML. The Val158Met polymorphism of the catechol-O-methyltransference gene is not associated with long-term treatment outcomes in carpal tunnel syndrome: a randomized clinical trial. PLOS One 2018 ;13(10):e0205516. [DOI] [PMC free article] [PubMed] [Google Scholar]

Fernandez‐de‐las‐Penas 2019 {published data only}

Fernández-de-Las-Peñas C, de-la-Llave-Rincón AI, Cescon C, Barbero M, Arias-Buría JL, Falla D. Influence of clinical, psychological, and psychophysical variables on long-term treatment outcomes in carpal tunnel syndrome: evidence from a randomized clinical trial. Pain Practice : The Official Journal of World Institute of Pain 2019;19(6):644–55. [DOI] [PubMed] [Google Scholar]

Ferrari 1997 {published data only}

Ferrari FS, Della Sala L, Cozza S, Guazzi G, Belcapo L, Mariottini A, et al. High-resolution ultrasonography in the study of carpal tunnel syndrome [L'ecografia con alta risoluzione nello studio della sindrome del tunnel carpale]. La Radiologia medica 1997;93(4):336-41. [PubMed] [Google Scholar]

Guo 2017 {published data only}

Guo XY, Xiong MX, Zhao Y, He FD, Cheng XQ, Wu YY, et al. Comparison of the clinical effectiveness of ultrasound-guided corticosteroid injection with and without needle release of the transverse carpal ligament in carpal tunnel syndrome. European Neurology 2017;78(1-2):33-40. [DOI] [PubMed] [Google Scholar]

Gurcay 2017 {published data only}

Gurcay AG, Karaahmet OZ, Gurcan O, Kazanci A, Karsli PB, Umay EK, et al. Comparison of short-term clinical and electrophysiological outcomes of local steroid injection and surgical decompression in the treatment of carpal tunnel syndrome. Turkish Neurosurgery 2017;27(3):447-52. [DOI] [PubMed] [Google Scholar]

Guvenc 2019 {published data only}

Güvenç K, İlik MK, İlik F, Gönen M, Kayhan F. Examination of surgical and conservative treatment effects on depression of patients with moderate carpal tunnel syndrome. European Research Journal 2019;5(4):673–7. [Google Scholar]

Harter 1993 {published data only}

Harter BT, McKiernan JE, Kirzinger SS, Archer FW, Peters CK, Harter KC. Carpal tunnel syndrome: surgical and nonsurgical treatment. Journal of Hand Surgery 1993;18(4):734-9. [DOI] [PubMed] [Google Scholar]

Hartley 1999 {published data only}

Hartley R, Chamberlain ST, Koukou C, Bradley M, Burke Derby FD. Results of a prospective randomized trial comparing open carpal tunnel decompression using a subcutaneous tome [abstract]. Journal of Hand Surgery - British Volume 1999;24(Suppl 1):4. [Google Scholar]

Jenkins 2012 {published data only}

Jenkins PJ, Duckworth AD, Watts AC, McEachan JE. Corticosteroid injection for carpal tunnel syndrome: a 5-year survivorship analysis. Hand (New York, N.Y.) 2012;7(2):151-6. [DOI] [PMC free article] [PubMed] [Google Scholar]

Karamese 2015 {published data only}

Karamese M, Yildiran G, Kara I, Selimoglu MN, Akdag O, Tosun Z. Carpal tunnel release combined with radiofrequence therapy for faster healing and reducing the pain. Journal of Hand Surgery, European Volume 2015;41(Suppl 1):S46, Abstract no: A-0304. [Google Scholar]

Kitsis 2002 {published data only}

Kitsis CK, Savvidou O, Alam A, Cherry RJ. Carpal tunnel syndrome despite negative neurophysiological studies. Acta Orthopaedica Belgica 2002;68(2):135-40. [PubMed] [Google Scholar]

Mason 2017 {published data only}

Berwin JT, Cooper C, Mason W. Injection versus decompression for carpal tunnel syndrome (INDICATE): feasibility trial. Journal of Hand Surgery, European Volume 2020 Nov;45(9):988-90. [DOI: 10.1177/1753193420929249.] [DOI] [PubMed]
ISRCTN59894749. Steroid injection versus surgical decompression for carpal tunnel syndrome. https://doi.org/10.1186/ISRCTN59894749 (first registered 20 July 2016).
Mason W, Ryan D, Khan A, Kerr HL, Beard D, Cook J, et al. Injection versus decompression for carpal tunnel syndrome-pilot trial (INDICATE-P)-protocol for a randomised feasibility study. Pilot and Feasibility Studies 2017;3:20. [DOI] [PMC free article] [PubMed] [Google Scholar]

Miranda 2013 {published data only}

Miranda BH, Asaad K, Cerovac S. Carpal tunnel syndrome study: local corticosteroids, conversion to surgery and NHS implications. Journal of Plastic, Reconstructive & Aesthetic Surgery : JPRAS 2013;66(10):1432-3. [DOI] [PubMed] [Google Scholar]

NCT00694265 {published data only}

NCT00694265. Sonographic follow-up of patients with carpal tunnel syndrome undergoing surgical or conservative treatment: prospective cohort study. clinicaltrials.gov/ct2/show/NCT00694265 (first received 10 June 2008).

NCT03548259 {published data only}

NCT03548259. Effects of platelet-rich plasma in the surgery of carpal tunnel syndrome: a randomized controlled trial (PRP-CTS). clinicaltrials.gov/ct2/show/NCT03548259 (first received 7 June 2008).

NCT04347746 {published data only}

NCT04347746. Comparison of interventions in patients with carpal tunnel syndrome. clinicaltrials.gov/ct2/show/NCT04347746 (first received 15 April 2008).

Onuma 2015 {published data only}

Onuma K, Fujimaki H, Kenmoku T, Sukegawa K, Takano S, Uchida K, et al. Bilateral carpal tunnel syndrome due to gouty tophi: conservative and surgical treatment in different hands of the same patient. Modern Rheumatology 2015;25(2):298–302. [DOI] [PubMed] [Google Scholar]

Roitberg 2003 {published data only}

Roitberg B. A randomized trial of splinting vs. surgery for carpal tunnel syndrome. Surgical Neurology 2003;59(1):5–6. [DOI] [PubMed] [Google Scholar]

Rubin 2018 {published data only}

Rubin G, Orbach H, Rinott M, Rozen N. The effectiveness of splinting and surgery on sleep disturbance in carpal tunnel syndrome. Journal of Hand Surgery, European Volume 2018;43(3):286–9. [DOI] [PubMed] [Google Scholar]

Sahoo 2020 {published data only}

Sahoo JK, Joshi AG. Study of effect of various treatments in clinically diagnosed patients of carpal tunnel syndrome. International Journal of Research in Pharmaceutical Sciences 2020;11(3):3105–13. [Google Scholar]

Siciliano 2014 {published data only}

Siciliano G, Schirinzi E, Lucchesi C, Depaulis V, Calabrese R, Lencioni M, et al. Platelet rich plasma (PRP) as a new therapeutic strategy for severe median nerve compression in carpal tunnel syndrome. In: Journal of the Peripheral Nervous System. Vol. 19(Suppl 1). 2014:S31–S32.

Siegmund‐Schultze 2017 {published data only}

Siegmund-Schultze N. Carpal tunnel syndrome: manual therapy is over one year as effective as surgery. Deutsches Arzteblatt International 2017;114(19):A952. [Google Scholar]

Sparapani 2006 {published and unpublished data}

Sparapani FV, Aguiar PH, Zicarelli CA, Hirata MT, Tedesco-Marchese AJ, Teixeira MJ. Comparison between the clinical and surgical treatment in carpal tunnel syndrome by means of electrophysiological analysis [Comparação entre tratamentos clínico e cirúrgico da síndrome do túnel do carpo por métodos de análises eletrofisiológicas]. Jornal Brasileiro de Neurocirurgia 2006;17(2):52-5. [Google Scholar]

Tudiver 2003 {published data only}

Tudiver F, Johnson ED, O’Reilly Brown M. Does surgery for carpal tunnel syndrome improve outcomes? Journal of Family Practice 2003;52(1):70–2. [PubMed] [Google Scholar]

Yoshii 2015 {published data only}

Yoshii Y, Ishii T, Tung WL. Ultrasound assessment of the effectiveness of carpal tunnel release on median nerve deformation. Journal of Orthopaedic Research : Official Publication of the Orthopaedic Research Society 2015;33(5):726–30. [DOI] [PubMed] [Google Scholar]

Zannakis 2020 {published data only}

Zannakis M. 5 year outcomes of carpal Rx therapy to treat carpal tunnel syndrome. In: British Journal of Pain. Vol. 14. 2020:30.

References to studies awaiting assessment

Abedi 2018 {published data only}

Abedi M, Mirkazemi M, Jamebozorgi K, Padidar S, Izanloo A. Evaluating the effectiveness of surgical treatment and local steroid injection in patients with carpal tunnel syndrome. Razavi International Journal of Medicine 2018;6(3):e14338. [Google Scholar]

Elwakil 2007 {published data only}

Elwakil TF, Elazzazi A, Shokeir H. Treatment of carpal tunnel syndrome by low-level laser versus open carpal tunnel release. Lasers in Medical Science 2007;22(4):265-70. [PMID: 17334675] [DOI] [PubMed] [Google Scholar]

Hrkovic 2016 {published data only}

Hrkovic M, Lazovic M, Nikolic D, Nikcevic L, Ilic-Stojanovic O, Filipovic T. Functional status measured by Levine questionnaire in surgically and conservatively treated patients with carpal tunnel syndrome regarding nerve conduction studies. Annali Italiani Di Chirurgia 2016;87:247-51. [PubMed] [Google Scholar]

IRCT20200629047948N1 {published data only}

IRCT20200629047948N1. Comparison of the effectiveness of open surgery versus ultrasound guided median nerve dissection with injection of methyl prednisolone acetate in treatment of carpal tunnel syndrome. www.irct.ir/trial/49264 (first received 29 June 2020).

Linscheid 1967 {published data only}

Linscheid RL, Peterson LF, Juergens JL. Carpal-tunnel syndrome associated with vasospasm. Journal of Bone and Joint Surgery. American Volume 1967;49(6):1141-6. [PubMed] [Google Scholar]

Lo 2021 {published data only}

Lo IN, Hsu PC, Huang YC, Yeh CK, Yang YC, Wang JC. Dynamic ultrasound assessment of median nerve mobility changes following corticosteroid injection and carpal tunnel release in patients with carpal tunnel syndrome. Frontiers in Neurology 2021;12:710511. [DOI: 10.3389/fneur.2021.710511] [DOI] [PMC free article] [PubMed] [Google Scholar]

NCT00981565 {published data only}

NCT00981565. Operative versus conservative treatment in mild carpal tunnel syndrome, randomized prospective multicenter study. clinicaltrials.gov/ct2/show/NCT00981565 (first received 22 Sep 2009).

NCT04216147 {published data only}

NCT04216147. Percutaneous needle electrolysis versus surgery in the treatment of carpal tunnel syndrome. clinicaltrials.gov/ct2/show/NCT04216147 (first received 2 Jan 2020).

Shazad 2022 {published data only}

Shazad MM, Rizwan M, Ul Haq Khawaja MI, Napar AR, Amjad A. Outcome assessment of comparison between surgical decompression versus local corticosteroid injection in carpal tunnel syndrome. Pakistan Journal of Medical and Health Sciences 2022;16(1):177–9. [DOI: 10.53350/pjmhs22161177] [DOI] [Google Scholar]

References to ongoing studies

CTRI201901016881 {published data only}

CTRI/2019/01/016881. A comparison of ultrasound guided steroid injection and endoscopic carpal tunnel release for the treatment of carpal tunnel syndrome. ctri.nic.in/Clinicaltrials/pmaindet2.php?trialid=29922&EncHid=&userName=CTRI/2019/01/016881 (first received 3 January 2019).

DRKS00014585 {published data only}

DRKS00014585. Conservative treatment of carpal tunnel syndrom (CTS) with the Tuttlingen CTS-Cuff versus surgery. https://trialsearch.who.int/Trial2.aspx?TrialID=DRKS00014585 (first received 26 June 2018).

DRKS00020570 {published data only}

DRKS00020570. The effect of stretching the transverse carpal ligament using curpal® on clinical symptoms and various neurophysiological parameters in carpal tunnel syndrome (CTS) – a prospective, randomised trial. www.drks.de/DRKS00020570 (first received 21 December 2020).

EUCTR2013‐000873‐56‐ES {published data only}

EUCTR2013-000873-56-ES. Randomized, two paralel groups, open clinical trial stratified by severity to stimate the cost-effectivity of surgical vs corticosteroid injection treatment on carpal tunnel syndrome. www.clinicaltrialsregister.eu/ctr-search/trial/2013-000873-56/ES (first received 20 May 2013).

EUCTR2021‐004756‐42‐NO {published data only}

EUCTR2021-004756-42-NO. A Norwegian trial comparing treatment strategies for carpal tunnel syndrome. https://www.clinicaltrialsregister.eu/ctr-search/search?query=eudract_number:2021-004756-42 (start date 25 January 2022).
NCT05306548. A Norwegian trial comparing treatment strategies for carpal tunnel syndrome. https://ClinicalTrials.gov/show/NCT05306548 (first posted 1 April 2022).

IRCT20141121020020N8 {published data only}

IRCT20141121020020N8. Comparison of long term effectiveness of open surgery versus ultrasound guided median nerve hydrodissection in terms of symptoms, function and electrophysiological studies in patients with canal tunnel syndrome (CTS). https://trialsearch.who.int/Trial2.aspx?TrialID=IRCT20141121020020N8 (first registered 21 September 2021).

NCT04014244 {published data only}

NCT04014244. Efficacy of dextrose injections, corticosteroids injections and surgical release for treatment of the carpal tunnel syndrome: a prospective, randomized, double-blind controlled trial. clinicaltrials.gov/ct2/show/NCT04014244 (first received 10 July 2019).

NCT04058041 {published data only}

NCT04058041. Effectiveness of neural mobilizations exercises versus surgery in patients with carpal tunnel syndrome. clinicaltrials.gov/ct2/show/NCT04058041 (first received 15 August 2019).

NCT04246216 {published data only}

NCT04246216. Ultrasound-guided percutaneous electrical nerve stimulation of the median nerve for carpal tunnel syndrome. clinicaltrials.gov/ct2/show/NCT04246216 (first received 20 July 2021).

Palmbergen 2022 {published data only}

ISRCTN13164336. The Dutch injection versus operation trial in carpal tunnel syndrome patients. doi.org/10.1186/ISRCTN13164336 (first received 21 June 2017).
Palmbergen WA, Bie RM, Alleman TW, Verstraete E, Jellema K, Verhagen WI, et al. Dutch injection versus surgery trial in patients with carpal tunnel syndrome (DISTRICTS): protocol of a randomised controlled trial comparing two treatment strategies. BMJ Open 2022;12(4):e057641. [DOI: 10.1136/bmjopen-2021-057641] [DOI] [PMC free article] [PubMed] [Google Scholar]

Additional references

AAN 1993

American Academy of Neurology. Practice parameter for carpal tunnel syndrome (summary statement). Neurology 1993;43(11):2406-9. [MEDLINE: ] [PubMed] [Google Scholar]

Atroshi 1999

Atroshi I, Gummesson C, Johnsson R, Ornstein E, Ranstam J, Rosén I. Prevalence of carpal tunnel syndrome in a general population. JAMA 1999;282(2):153–8. [DOI: 10.1001/jama.282.2.153] [DOI] [PubMed] [Google Scholar]

Bialosky 2009

Bialosky JE, Bishop MD, Price DD, Robinson ME, George SZ. The mechanisms of manual therapy in the treatment of musculoskeletal pain: a comprehensive model. Manual Therapy 2009;14(5):531-8. [DOI: 10.1016/j.math.2008.09.001] [DOI] [PMC free article] [PubMed] [Google Scholar]

Bland 2000

Bland JD. A neurophysiological grading scale for carpal tunnel syndrome. Muscle & Nerve 2000;23(8):1280–3. [DOI: ] [DOI] [PubMed] [Google Scholar]

Burton 2016

Burton CL, Chesterton LS, Chen Y, Windt DA. Clinical course and prognostic factors in conservatively managed carpal tunnel syndrome: a systematic review. Archives of Physical Medicine & Rehabilitation 2016;97(5):836-52.e1. [DOI] [PubMed] [Google Scholar]

Carroll 1987

Carroll GJ. Comparison of median and radial nerve sensory latencies in the electrophysiological diagnosis of carpal tunnel syndrome. Electroencephalography and Clinical Neurophysiology 1987;68(2):101-6. [MEDLINE: ] [DOI] [PubMed] [Google Scholar]

Cohen 2017

Cohen MS, Hotchkiss RN, Kozin SH, Pederson WC, Wolfe SW. Green's Operative Hand Surgery. Philadelphia: Elsevier, 2017. [Google Scholar]

De Kleermaeker 2018

De Kleermaeker FG, Boogaarts HD, Meulstee J, Verhagen WI. Minimal clinically important difference for the Boston Carpal Tunnel Questionnaire: new insights and review of literature. Journal of Hand Surgery: European Volume 2018;44(3):283-9. [DOI] [PubMed] [Google Scholar]

de Krom 1992

Krom MC, Knipschild PG, Kester AD, Thijs CT, Boekkooi PF, Spaans F. Carpal tunnel syndrome: prevalence in the general population. Journal of Clinical Epidemiology 1992;45(4):373-6. [MEDLINE: ] [DOI] [PubMed] [Google Scholar]

Deeks 2021

Deeks JJ, Higgins JP, Altman DG, editor(s). Chapter 10: Analysing data and undertaking meta-analyses. In: Higgins JP, Thomas J, Chandler J, Cumpston M, Li T, Page MJ, Welch VA, editor(s). Cochrane Handbook for Systematic Reviews of Interventions Version 6.2 (updated February 2021).Cochrane, 2021. Available from www.training.cochrane.org/handbook.

Ebenbichler 1998

Ebenbichler GR, Resch KL, Nicolakis P, Wiesinger GF, Uhl F, Ghanem AH, et al. Ultrasound treatment for treating the carpal tunnel syndrome: randomised "sham" controlled trial. BMJ 1998;316(7133):731-5. [DOI: 10.1136/bmj.316.7133.731] [DOI] [PMC free article] [PubMed] [Google Scholar]

Erickson 2021

Erickson M, Lawrence M, Lucado A. The role of diagnostic ultrasound in the examination of carpal tunnel syndrome: an update and systematic review. Journal of Hand Therapy : Official Journal of the American Society of Hand Therapists 2021;S0894-1130(21):00061-2. [DOI: 10.1016/j.jht.2021.04.014] [DOI] [PubMed] [Google Scholar]

Felsenthal 1977

Felsenthal G. Median and ulnar distal motor and sensory latencies in the same normal subject. Archives of Physical Medicine and Rehabilitation 1977;58(7):297-302. [MEDLINE: ] [PubMed] [Google Scholar]

Geiringer 1998

Geiringer SR. The value of inching techniques in the diagnosis of focal nerve lesions. Inching techniques are of limited use. Muscle & Nerve 1998;21(11):1557-9. [MEDLINE: ] [DOI] [PubMed] [Google Scholar]

Gelberman 1984

Gelberman RH, Szabo RM, Mortenson MM. Carpal tunnel pressures and wrist position in patients with Colles fractures. Journal of Trauma, Injury, Infection and Critical Care 1984;24(8):747-9. [DOI] [PubMed] [Google Scholar]

GRADEpro GDT 2021 [Computer program]

GRADEpro Guideline Development Tool. McMaster University and Evidence Prime, 2021. Available from gradepro.org.

Hao 2019

Hao Q, Devji T, Zeraatkar D, Wang Y, Qasim A, Siemieniuk RA, et al. Minimal important differences for improvement in shoulder condition patient-reported outcomes: a systematic review to inform a BMJ Rapid Recommendation. BMJ Open 2019;9(2):e028777. [DOI: 10.1136/bmjopen-2018-028777] [DOI] [PMC free article] [PubMed] [Google Scholar]

Higgins 2020

Higgins JP, Li T, Deeks JJ, editor(s). Chapter 6: Choosing effect measures and computing estimates of effect. In: Higgins JP, Thomas J, Chandler J, Cumpston M, Li T, Page MJ, Welch VA, editor(s). Cochrane Handbook for Systematic Reviews of Interventions Version 6.1 (updated September 2020). Cochrane, 2020. Available from www.training.cochrane.org/handbook.

Hofer 2021

Hofer M, Ranstam J, Atroshi I. Extended follow-up of local steroid injection for carpal tunnel syndrome: a randomized clinical trial. JAMA Network Open 2021;4(10):e2130753. [DOI] [PMC free article] [PubMed] [Google Scholar]

Huisstede 2010

Huisstede BM, Randsdorp MS, Coert JH, Glerum S, Middelkoop M, Koes BW. Carpal tunnel syndrome. Part II: effectiveness of surgical treatments--a systematic review. Archives of Physical Medicine and Rehabilitation 2010;91(7):1005-24. [DOI: 10.1016/j.apmr.2010.03.023] [DOI] [PubMed] [Google Scholar]

Huisstede 2018

Huisstede BM, Randsdorp MS, den Brink J, Franke TP, Koes BW, Hoogvliet P. Effectiveness of oral pain medication and corticosteroid injections for carpal tunnel syndrome: a systematic review. Archives of Physical Medicine and Rehabilitation 2018;99(8):1609-22.e10. [DOI] [PubMed] [Google Scholar]

Karjalainen 2022a

Karjalanen T, Raatikainen S, Jaatinen K, Lusa V. Update on efficacy of conservative treatments for carpal tunnel syndrome. Journal of Clinical Medicine 2022;11(4):950. [DOI] [PMC free article] [PubMed] [Google Scholar]

Karjalainen 2022b

Karjalainen T, Heikkinen J, Busija L, Jokihaara J, Lewin AM, Naylor JM, et al. Use of placebo and nonoperative control groups in surgical trials: a systematic review and meta-analysis. JAMA Network Open 2022;5(7):e2223903. [DOI: 10.1001/jamanetworkopen.2022.23903] [DOI] [PMC free article] [PubMed] [Google Scholar]

Karjalainen 2023

Karjalainen TV, Lusa V, Page MJ, O'Connor D, Massy-Westropp N, Peters SE. Splinting for carpal tunnel syndrome. Cochrane Database of Systematic Reviews 2023, Issue 2. Art. No: CD010003. [DOI: 10.1002/14651858.CD010003.pub2] [DOI] [PMC free article] [PubMed] [Google Scholar]

Katz 1990

Katz JN, Stirrat CR. A self-administered hand diagram for the diagnosis of carpal tunnel syndrome. Journal of Hand Surgery 1990;15(2):360-3. [DOI: 10.1016/0363-5023(90)90124-a] [DOI] [PubMed] [Google Scholar]

Kim 2013

Kim JK, Jeon SH. Minimal clinically important differences in the Carpal Tunnel Questionnaire after carpal tunnel release. Journal of Hand Surgery: European Volume 2013;38(1):75-9. [DOI] [PubMed] [Google Scholar]

Kimura 1979

Kimura J. The carpal tunnel syndrome: localization of conduction abnormalities within the distal segment of the median nerve. Brain 1979;102(3):619-35. [MEDLINE: ] [DOI] [PubMed] [Google Scholar]

Klokkari 2018

Klokkari D, Mamais I. Effectiveness of surgical versus conservative treatment for carpal tunnel syndrome: a systematic review, meta-analysis and qualitative analysis. Hong Kong Physiotherapy Journal 2018 Dec;38(2):91-114. [DOI: 10.1142/S1013702518500087] [DOI] [PMC free article] [PubMed] [Google Scholar]

Leigh 1998

Leigh JP, Miller TR. Job-related diseases and occupations within a large workers-compensation data set. American Journal of Industrial Medicine 1998;33(3):197-211. [MEDLINE: ] [DOI] [PubMed] [Google Scholar]

Liong 2013

Liong K, Lahiri A, Lee S, Chia D, Biswas A, Lee HP. Mid-motion deformation of median nerve during finger flexion: a new insight into the dynamic aetiology of carpal tunnel syndrome. Hand Surgery : an International Journal devoted to Hand and Upper Limb Surgery and Related Research : Journal of the Asia-Pacific Federation of Societies for Surgery of the Hand 2013;18(2):193-202. [DOI] [PubMed] [Google Scholar]

MacKinnon 1988

MacKinnon SE. Surgery of the Peripheral Nerve. New York, NY: Thieme Medical Publishers, 1988. [Google Scholar]

Martyn 1997

Martyn CN, Hughes RA. Epidemiology of peripheral neuropathy. Journal of Neurology, Neurosurgery and Psychiatry 1997;62(4):310-8. [MEDLINE: ] [DOI] [PMC free article] [PubMed] [Google Scholar]

Murata 2007

Murata K, Yajima H, Maegawa N, Hattori K, Takakura Y. Investigation of segmental carpal tunnel pressure in patients with idiopathic carpal tunnel syndrome--is it necessary to release the distal aponeurotic portion of the flexor retinaculum in endoscopic carpal tunnel release surgery? Hand Surgery : an International Journal devoted to Hand and Upper Limb Surgery and Related Research : Journal of the Asia-Pacific Federation of Societies for Surgery of the Hand 2007;12(3):205-9. [DOI] [PubMed] [Google Scholar]

Padua 2016

Padua L, Coraci D, Erra C, Pazzaglia C, Paolasso I, Loreti C, et al. Carpal tunnel syndrome: clinical features, diagnosis, and management. Lancet Neurology 2016;15(12):1273-84. [DOI] [PubMed] [Google Scholar]

Pryse‐Phillips 1984

Pryse-Phillips WE. Validation of a diagnostic sign in carpal tunnel syndrome. Journal of Neurology, Neurosurgery, and Psychiatry 1984;47(8):870-2. [DOI: 10.1136/jnnp.47.8.870] [DOI] [PMC free article] [PubMed] [Google Scholar]

Ren 2016

Ren YM, Wang XS, Wei ZJ, Fan BY, Lin W, Zhou XH, et al. Efficacy, safety, and cost of surgical versus nonsurgical treatment for carpal tunnel syndrome: a systematic review and meta-analysis. Medicine (Baltimore) 2016;95(40):e4857. Retraction in: Medicine (Baltimore). 2017;96(16):e6778. [DOI: 10.1097/MD.0000000000004857] [DOI] [PMC free article] [PubMed] [Google Scholar] [Retracted]

Review Manager 2020 [Computer program]

Review Manager. Version 5.4. The Cochrane Collaboration, 2020.

Rosenbaum 1993

Rosenbaum RB, Ochoa JL. Carpal Tunnel Syndrome and Other Disorders of the Median Nerve. Stoneham, MA: Butterworth-Heinemann, 1993. [Google Scholar]

Rossignol 1997

Rossignol M, Stock S, Patry L, Armstrong B. Carpal tunnel syndrome: what is attributable to work? Occupational and Environmental Medicine 1997;54(7):519-23. [MEDLINE: ] [DOI] [PMC free article] [PubMed] [Google Scholar]

Schreiber 2005

Schreiber JE, Foran MP, Schreiber DJ, Wilgis EF. Common risk factors seen in secondary carpal tunnel surgery. Annals of Plastic Surgery 2005;55(3):262-5. [DOI] [PubMed] [Google Scholar]

Schrijver 2005

Schrijver HM, Gerritsen AA, Strijers RL, Uitdehaag BM, Scholten RJ, De Vet Henrica CW, et al. Correlating nerve conduction studies and clinical outcome measures on carpal tunnel syndrome: lessons from a randomized controlled trial. Journal of Clinical Neurophysiology 2005;22(3):216-21. [PubMed] [Google Scholar]

Schuind 2002

Schuind F. Canal pressures before, during, and after endoscopic release for idiopathic carpal tunnel syndrome. The Journal of Hand Surgery 2002;27(6):1019-25. [DOI] [PubMed] [Google Scholar]

Schunemann 2017

Schünemann HJ, Oxman AD, Vist GE, Higgins JP, Deeks JJ, Glasziou P, et al, on behalf of the Cochrane Applicability and Recommendations Methods Group. Chapter 12: Interpreting results and drawing conclusions. In: Higgins JP, Green S, editor(s). Cochrane Handbook for Systematic Reviews of Interventions Version 5.1.0 (updated June 2017). Cochrane Collaboration, 2017. Available from: www.cochrane-handbook.org.

Shi 2011

Shi Q, MacDermid JC. Is surgical intervention more effective than non-surgical treatment for carpal tunnel syndrome? A systematic review. Journal of Orthopaedic Surgery and Research 2011;6:17. [DOI: 10.1186/1749-799X-6-17] [DOI] [PMC free article] [PubMed] [Google Scholar]

Shi 2020

Shi Q, Bobos P, Lalone EA, Warren L, MacDermid JC. Comparison of the short-term and long-term effects of surgery and nonsurgical intervention in treating carpal tunnel syndrome: a systematic review and meta-analysis. Hand (New York, N.Y.) 2020;15(1):13-22. [DOI: 10.1177/1558944718787892] [DOI] [PMC free article] [PubMed] [Google Scholar]

Shiri 2011

Shiri R, Heliövaara M, Moilanen L, Viikari J, Liira H, Viikari-Juntura E. Associations of cardiovascular risk factors, carotid intima-media thickness and manifest atherosclerotic vascular disease with carpal tunnel syndrome. BMC Musculoskeletal Disorders 2011;12:80. [DOI] [PMC free article] [PubMed] [Google Scholar]

Sterne 2019

Sterne JA, Savović J, Page MJ, Elbers RG, Blencowe NS, Boutron I, et al. RoB 2: a revised tool for assessing risk of bias in randomised trials. BMJ 2019;366:l4898. [DOI] [PubMed] [Google Scholar]

Stevens 1997

Stevens JC. AAEM minimonograph: 26: the electrodiagnosis of carpal tunnel syndrome. American Association of Electrodiagnostic Medicine. Muscle & Nerve 1997;20(12):1477-86. [DOI] [PubMed] [Google Scholar]

Stewart 1993

Stewart JD. Compression and entrapment neuropathies. In: Dyck PJ et al, editors(s). Peripheral Neuropathy. 3rd edition. Philadelphia, PA: Saunders, 1993. [Google Scholar]

Teitsma 2017

Teitsma XM, Jacobs JW, Welsing PM, Pethö-Schramm A, Borm ME, Hendriks L, et al. Patient-reported outcomes in newly diagnosed early rheumatoid arthritis patients treated to target with a tocilizumab- or methotrexate-based strategy. Rheumatology (Oxford) 2017;56(12):2179-89. [DOI: 10.1093/rheumatology/kex319] [DOI] [PubMed] [Google Scholar]

Vasiliadis 2014

Vasiliadis HS, Georgoulas P, Shrier I, Salanti G, Scholten RJ. Endoscopic release for carpal tunnel syndrome. Cochrane Database of Systematic Reviews 2014, Issue 1. Art. No: CD008265. [DOI: 10.1002/14651858.CD008265.pub2] [DOI] [PMC free article] [PubMed] [Google Scholar]

Verdugo 1994

Verdugo RJ, Ochoa JL. "Sympathetically maintained pain" I. Phentolamine block questions the concept. Neurology 1994;44(6):1003-10. [MEDLINE: ] [DOI] [PubMed] [Google Scholar]

Walters 2005

Walters SJ, Brazier JE. Comparison of the minimally important difference for two health state utility measures: EQ-5D and SF-6D. Quality of Life Research : an International Journal of Quality of Life Aspects of Treatment, Care and Rehabilitation 2005;6:1523-32. [DOI: 10.1007/s11136-004-7713-0] [DOI] [PubMed] [Google Scholar]

Yamanaka 2021

Yamanaka Y, Tajima T, Tsujimura Y, Kosugi K, Mano Y, Zenke Y, et al. Molecular and clinical elucidation of the mechanism of action of steroids in idiopathic carpal tunnel syndrome. The Journal of Bone and Joint Surgery 2021;103(19):1777-87. [DOI: 10.2106/JBJS.20.02096] [DOI] [PubMed] [Google Scholar]

References to other published versions of this review

Verdugo 2008

Verdugo RJ, Salinas RA, Castillo JL, Cea JG. Surgical versus non-surgical treatment for carpal tunnel syndrome. Cochrane Database of Systematic Reviews 2008, Issue 4. Art. No: CD001552. [DOI: 10.1002/14651858.CD001552.pub2] [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD001552-bib-0001] Awan AS, Khan A, Afridi SA, Khan IU, Bhatti SN, Ahmed E, et al. Early response of local steroid injection versus miniincision technique in treatment of carpal tunnel syndrome. Journal of Ayub Medical College, Abbottabad : JAMC 2015;27(1):192-6. [PubMed] [Google Scholar]

[CD001552-bib-0002] Eltabl MA, Saif DS, Alemam SE. Platelet-rich plasma injection versus surgical and medical treatment of mild-moderate carpal tunnel syndrome. Egyptian Journal of Neurology, Psychiatry and Neurosurgery 2020;56(1):88. [Google Scholar]

[CD001552-bib-0003] NCT04235426. Platelet rich plasma injection versus surgical and medical treatment of mild-moderate carpel tunnel syndrome. ClinicalTrials.gov/show/NCT04235426 (first posted 21 January 2020).

[CD001552-bib-0004] Fernández-de-Las Peñas C, Ortega-Santiago R, la Llave-Rincón AI, Martínez-Perez A, Fahandezh-Saddi Díaz H, Martínez-Martín J, et al. Manual physical therapy versus surgery for carpal tunnel syndrome: a randomized parallel-group trial. The Journal of Pain : Official Journal of the American Pain Society 2015;16(11):1087-94. [DOI: ] [DOI] [PubMed] [Google Scholar]

[CD001552-bib-0005] Fernández-de-Las-Peñas C, Arias-Buría JL, Cleland JA, Pareja JA, Plaza-Manzano G, Ortega-Santiago R. Manual therapy versus surgery for carpal tunnel syndrome: 4-year follow-up from a randomized controlled trial. Physical Therapy 2020;100(11):1987-96. [DOI: ] [DOI] [PubMed] [Google Scholar]

[CD001552-bib-0006] Fernández-de-Las-Peñas C, Ortega-Santiago R, Díaz HF, Salom-Moreno J, Cleland JA, Pareja JA, et al. Cost-effectiveness evaluation of manual physical therapy versus surgery for carpal tunnel syndrome: evidence from a randomized clinical trial. Journal of Orthopaedic & Sports Physical Therapy 2019;49(2):55-63. [DOI] [PubMed] [Google Scholar]

[CD001552-bib-0007] NCT01789645. Physical therapy in carpal tunnel syndrome. clinicaltrials.gov/show/NCT01789645 (first posted 12 February 2013).

[CD001552-bib-0008] Fernández-de-Las-Peñas C, Cleland J, Palacios-Ceña M, Fuensalida-Novo S, Alonso-Blanco C, Pareja JA, et al. Effectiveness of manual therapy versus surgery in pain processing due to carpal tunnel syndrome: a randomized clinical trial. European Journal of Pain (London, England) 2017;21(7):1266-76. [DOI: ] [DOI] [PubMed] [Google Scholar]

[CD001552-bib-0009] NCT02219919. Nociceptive gain processing after physical therapy in carpal tunnel syndrome. clinicaltrials.gov/show/NCT02219919 (first posted 19 August 2014).

[CD001552-bib-0010] Fernández-de-Las-Peñas C, Cleland J, Palacios-Ceña M, Fuensalida-Novo S, Pareja JA, Alonso-Blanco C. The effectiveness of manual therapy versus surgery on self-reported function, cervical range of motion, and pinch grip force in carpal tunnel syndrome: a randomized clinical trial. Journal of Orthopaedic and Sports Physical Therapy 2017;47(3):151-61. [DOI: ] [DOI] [PubMed] [Google Scholar]

[CD001552-bib-0011] Garland H, Langworth EP, Taverner D, Clark JM. Surgical treatment for the carpal tunnel syndrome. Lancet 1964;I(7343):1129-30. [DOI] [PubMed] [Google Scholar]

[CD001552-bib-0012] Gerritsen AA, De Vet HC, Scholten RJ, Bertelsmann FW, De Krom MC, Bouter LM. Betere klinische effecten op carpale-tunnelsyndroom van operatie dan van spalkbehandeling; resultaten van een gerandomiseerd klinisch onderzoek. Nederlands Tijdschrift voor Geneeskunde 2002;146(45):2153-6. [Google Scholar]

[CD001552-bib-0013] Gerritsen AA, Vet HC, Scholten RJ, Bertelsmann FW, Krom MC, Bouter LM. Splinting versus surgery in the treatment of carpal tunnel syndrome. A randomized controlled trial. Journal of the American Medical Association 2002;288(10):1245-51. [DOI] [PubMed] [Google Scholar]

[CD001552-bib-0014] Gerritsen AM, Scholten RJ, Assendelft WJ, Kuiper H, Vet HC, Bouter LM. Splinting or surgery for carpal tunnel syndrome? Design of a randomized controlled trial. BMC Neurology 2001;1:8. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD001552-bib-0015] ISRCTN18853827. Splinting or surgery for carpal tunnel syndrome. isrctn.com/ISRCTN18853827 (first registered 11 February 2002).

[CD001552-bib-0016] Korthals-de Bos IB, Gerritsen AA, Tulder MW, Rutten-van Mölken MP, Adèr HJ, Vet HC, et al. Surgery is more cost-effective than splinting for carpal tunnel syndrome in the Netherlands: results of an economic evaluation alongside a randomized controlled trial. BMC Musculoskeletal Disorders 2006;16(7):86. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD001552-bib-0017] Schrijver HM, Gerritsen AA, Strijers RL, Uitdehaag BM, Scholten RJ, Vet HC, et al. Correlating nerve conduction studies and clinical outcome measures on carpal tunnel syndrome: lessons from a randomized controlled trial. Journal of Clinical Neurophysiology : Official Publication of the American Electroencephalographic Society 2005;22(3):216–21. [PubMed] [Google Scholar]

[CD001552-bib-0018] Hui AC, Wong S, Leung CH, Tong P, Mok V, Poon D, et al. A randomized controlled trial of surgery versus steroid injection for carpal tunnel syndrome. Neurology 2005;64(12):2074-8. [PMID: 15985575] [DOI] [PubMed] [Google Scholar]

[CD001552-bib-0019] Ismatullah I. Local steroid injection or carpal tunnel release for carpal tunnel syndrome - which is more effective? Journal of Postgraduate Medical Institute 2013;27(2):194–9. [Google Scholar]

[CD001552-bib-0020] Irct2017011832046N. Comparing effectiveness of local steroid injection into carpal tunnel with surgery in the treatment of patients with mild to moderate carpal tunnel syndrome. https://www.irct.ir/trial/25063 (first registered 6 July 2017).

[CD001552-bib-0021] Jafari D, Shariatzadeh H, Mazhar FN, Babaei R, Mirkazemi M, Ayazi N. Comparison of the efficacy of carpal tunnel release vs. local steroid injection in the management of mild to moderate carpal tunnel syndrome: a clinical trial. Journal of Research in Orthopedic Science 2018;5(1):e58159. [Google Scholar]

[CD001552-bib-0022] Jarvik JG, Comstock BA, Kliot M, Turner JA, Chan L, Heagerty PJ, et al. Surgery versus non-surgical therapy for carpal tunnel syndrome: a randomised parallel-group trial. Lancet 2009;374(9695):1074-81. [DOI] [PubMed] [Google Scholar]

[CD001552-bib-0023] Martin BI, Levenson LM, Hollingworth W, Kliot M, Heagerty PJ, Turner JA, et al. Randomized clinical trial of surgery versus conservative therapy for carpal tunnel syndrome [ISRCTN84286481]. BMC Musculoskeletal Disorders 2005;6(2):NA. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD001552-bib-0024] NCT00032227. Carpal tunnel syndrome: diagnosis and treatment trial. clinicaltrials.gov/show/NCT00032227 (first posted 13 March 2002).

[CD001552-bib-0025] Andreu JL, Ly-Pen D, Millán I, Blas G, Sánchez-Olaso A. Local injection versus surgery in carpal tunnel syndrome: neurophysiologic outcomes of a randomized clinical trial. Clinical Neurophysiology 2014;125(7):1479-84. [DOI] [PubMed] [Google Scholar]

[CD001552-bib-0026] ISRCTN26264638. Surgical decompression versus local steroid injection in carpal tunnel syndrome. isrctn.com/ISRCTN26264638 (first registered 2 February 2012).

[CD001552-bib-0027] Ly-Pen D, Andreu JL, Millán I, Blas G, Sánchez-Olaso A. Long-term outcome of local steroid injections versus surgery in carpal tunnel syndrome: observational extension of a randomized clinical trial. Hand (New York, N.Y.) 2020;Aug 6 [Epub ahead of print]:NA. [DOI: 10.1177/1558944720944263] [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD001552-bib-0028] Ly-Pen D, Andréu JL, Millán I, Blas G, Sánchez-Olaso A. Comparison of surgical decompression and local steroid injection in the treatment of carpal tunnel syndrome: 2-year clinical results from a randomized trial. Rheumatology (Oxford, England) 2012;51(8):1447-54. [DOI] [PubMed] [Google Scholar]

[CD001552-bib-0029] Ly-Pen D, Andréu JL, Blas G, Sánchez-Olaso A, Millán I. Surgical decompression versus local steroid injection in carpal tunnel syndrome. A one-year, prospective, randomized, open, controlled clinical trial. Arthritis & Rheumatism 2005;52(2):612-9. [MEDLINE: ] [DOI] [PubMed] [Google Scholar]

[CD001552-bib-0030] Salinas M, Blas G, Regidor I, LyPen D, Andreu J, Sanchez-Olaso A. An electro-clinical comparison of carpal tunnel syndrome therapy. Muscle & Nerve 2003;28(S12):S178, Abstract no: 505. [Google Scholar]

[CD001552-bib-0031] Ucan H, Yagci I, Yilmaz L, Yagmurlu F, Keskin D, Bodur H. Comparison of splinting, splinting plus local steroid injection and open carpal tunnel release outcomes in idiopathic carpal tunnel syndrome. Rheumatology International 2006;27(1):45-51. [DOI] [PubMed] [Google Scholar]

[CD001552-bib-0032] Yağcı İ, Uçan H, Yılmaz L, Yağmurlu F, Keskin ED, Bodur H. Comparison of splinting, splinting plus local steroid injection and surgery in carpal tunnel syndrome treatment. Turkiye Fiziksel Tip ve Rehabilitasyon Dergisi 2006;52(2):55–60. [Google Scholar]

[CD001552-bib-0033] ChiCTR1800014530. The effectiveness of ultrasound-guided steroid injection combined with miniscalpel-needle release in the treatment of carpal tunnel syndrome: a randomized controlled study. https://www.chictr.org.cn/showprojEN.html?proj=24826 (first registered 19 January 2018). [DOI] [PMC free article] [PubMed]

[CD001552-bib-0034] Zhang S, Wang F, Ke S, Lin C, Liu C, Xin W, et al. The effectiveness of ultrasound-guided steroid injection combined with miniscalpel-needle release in the treatment of carpal tunnel syndrome vs. steroid injection alone: a randomized controlled study. BioMed Research International 2019;Feb 24 [Epub ahead of print]:NA. [DOI: 10.1155/2019/9498656] [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD001552-bib-0035] Baker NA, Livengood HM. Symptom severity and conservative treatment for carpal tunnel syndrome in association with eventual carpal tunnel release. Journal of Hand Surgery 2014;39(9):1792-8. [DOI] [PubMed] [Google Scholar]

[CD001552-bib-0036] Bozkurt M, Tatar BE, Karakol P, Sezgic M, Uslu C, Gelbal C, et al. Matriderm and platelet-rich plasma combination in the treatment of recurrent carpal tunnel syndrome: a new approach. European Journal of Plastic Surgery 2021;na:na. [DOI: 10.1007/s00238-021-01863-9] [DOI] [Google Scholar]

[CD001552-bib-0037] Celik G, Ilik MK. Effects of two different treatment techniques on the recovery parameters of moderate carpal tunnel syndrome: a six-month follow-up study. Journal of Clinical Neurophysiology : Official Publication of the American Electroencephalographic Society 2016;33(2):166-70. [DOI] [PubMed] [Google Scholar]

[CD001552-bib-0038] Cha SM, Shin HD, Ahn JS, Beom JW, Kim DY. Differences in the postoperative outcomes according to the primary treatment options chosen by patients with carpal tunnel syndrome: conservative versus operative treatment. Annals of Plastic Surgery 2016;77(1):80-4. [DOI] [PubMed] [Google Scholar]

[CD001552-bib-0039] Chang CW, Wang YC, Chang KF. A practical electrophysiological guide for non-surgical and surgical treatment of carpal tunnel syndrome. Journal of Hand Surgery: European Volume 2008;33(1):32-7. [DOI] [PubMed] [Google Scholar]

[CD001552-bib-0040] Demirci S, Kutluhan S, Koyuncuoglu HR, Kerman M, Heybeli N, Akkuş S, et al. Comparison of open carpal tunnel release and local steroid treatment outcomes in idiopathic carpal tunnel syndrome. Rheumatology International 2002;22(1):33-7. [DOI] [PubMed] [Google Scholar]

[CD001552-bib-0041] Ettema AM, Amadio PC, Cha SS, Harrington JR, Harris AM, Offord KP. Surgery versus conservative therapy in carpal tunnel syndrome in people aged 70 years and older. Plastic and Reconstructive Surgery 2006;118(4):947-58. [DOI] [PubMed] [Google Scholar]

[CD001552-bib-0042] Fernández-de-Las-Peñas C, Cleland JA, Salom-Moreno J, Palacios-Ceña M, Martínez-Perez A, Pareja JA, et al. Prediction of outcome in women with carpal tunnel syndrome who receive manual physical therapy interventions: a validation study. Journal of Orthopaedic and Sports Physical Therapy 2016;46(6):443–51. [DOI] [PubMed] [Google Scholar]

[CD001552-bib-0043] Fernández-de-Las-Peñas C, Ambite-Quesada S, Fahandezh-Saddi Díaz H, Paras-Bravo P, Palacios-Ceña D, Cuadrado ML. The Val158Met polymorphism of the catechol-O-methyltransference gene is not associated with long-term treatment outcomes in carpal tunnel syndrome: a randomized clinical trial. PLOS One 2018 ;13(10):e0205516. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD001552-bib-0044] Fernández-de-Las-Peñas C, de-la-Llave-Rincón AI, Cescon C, Barbero M, Arias-Buría JL, Falla D. Influence of clinical, psychological, and psychophysical variables on long-term treatment outcomes in carpal tunnel syndrome: evidence from a randomized clinical trial. Pain Practice : The Official Journal of World Institute of Pain 2019;19(6):644–55. [DOI] [PubMed] [Google Scholar]

[CD001552-bib-0045] Ferrari FS, Della Sala L, Cozza S, Guazzi G, Belcapo L, Mariottini A, et al. High-resolution ultrasonography in the study of carpal tunnel syndrome [L'ecografia con alta risoluzione nello studio della sindrome del tunnel carpale]. La Radiologia medica 1997;93(4):336-41. [PubMed] [Google Scholar]

[CD001552-bib-0046] Guo XY, Xiong MX, Zhao Y, He FD, Cheng XQ, Wu YY, et al. Comparison of the clinical effectiveness of ultrasound-guided corticosteroid injection with and without needle release of the transverse carpal ligament in carpal tunnel syndrome. European Neurology 2017;78(1-2):33-40. [DOI] [PubMed] [Google Scholar]

[CD001552-bib-0047] Gurcay AG, Karaahmet OZ, Gurcan O, Kazanci A, Karsli PB, Umay EK, et al. Comparison of short-term clinical and electrophysiological outcomes of local steroid injection and surgical decompression in the treatment of carpal tunnel syndrome. Turkish Neurosurgery 2017;27(3):447-52. [DOI] [PubMed] [Google Scholar]

[CD001552-bib-0048] Güvenç K, İlik MK, İlik F, Gönen M, Kayhan F. Examination of surgical and conservative treatment effects on depression of patients with moderate carpal tunnel syndrome. European Research Journal 2019;5(4):673–7. [Google Scholar]

[CD001552-bib-0049] Harter BT, McKiernan JE, Kirzinger SS, Archer FW, Peters CK, Harter KC. Carpal tunnel syndrome: surgical and nonsurgical treatment. Journal of Hand Surgery 1993;18(4):734-9. [DOI] [PubMed] [Google Scholar]

[CD001552-bib-0050] Hartley R, Chamberlain ST, Koukou C, Bradley M, Burke Derby FD. Results of a prospective randomized trial comparing open carpal tunnel decompression using a subcutaneous tome [abstract]. Journal of Hand Surgery - British Volume 1999;24(Suppl 1):4. [Google Scholar]

[CD001552-bib-0051] Jenkins PJ, Duckworth AD, Watts AC, McEachan JE. Corticosteroid injection for carpal tunnel syndrome: a 5-year survivorship analysis. Hand (New York, N.Y.) 2012;7(2):151-6. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD001552-bib-0052] Karamese M, Yildiran G, Kara I, Selimoglu MN, Akdag O, Tosun Z. Carpal tunnel release combined with radiofrequence therapy for faster healing and reducing the pain. Journal of Hand Surgery, European Volume 2015;41(Suppl 1):S46, Abstract no: A-0304. [Google Scholar]

[CD001552-bib-0053] Kitsis CK, Savvidou O, Alam A, Cherry RJ. Carpal tunnel syndrome despite negative neurophysiological studies. Acta Orthopaedica Belgica 2002;68(2):135-40. [PubMed] [Google Scholar]

[CD001552-bib-0054] Berwin JT, Cooper C, Mason W. Injection versus decompression for carpal tunnel syndrome (INDICATE): feasibility trial. Journal of Hand Surgery, European Volume 2020 Nov;45(9):988-90. [DOI: 10.1177/1753193420929249.] [DOI] [PubMed]

[CD001552-bib-0055] ISRCTN59894749. Steroid injection versus surgical decompression for carpal tunnel syndrome. https://doi.org/10.1186/ISRCTN59894749 (first registered 20 July 2016).

[CD001552-bib-0056] Mason W, Ryan D, Khan A, Kerr HL, Beard D, Cook J, et al. Injection versus decompression for carpal tunnel syndrome-pilot trial (INDICATE-P)-protocol for a randomised feasibility study. Pilot and Feasibility Studies 2017;3:20. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD001552-bib-0057] Miranda BH, Asaad K, Cerovac S. Carpal tunnel syndrome study: local corticosteroids, conversion to surgery and NHS implications. Journal of Plastic, Reconstructive & Aesthetic Surgery : JPRAS 2013;66(10):1432-3. [DOI] [PubMed] [Google Scholar]

[CD001552-bib-0058] NCT00694265. Sonographic follow-up of patients with carpal tunnel syndrome undergoing surgical or conservative treatment: prospective cohort study. clinicaltrials.gov/ct2/show/NCT00694265 (first received 10 June 2008).

[CD001552-bib-0059] NCT03548259. Effects of platelet-rich plasma in the surgery of carpal tunnel syndrome: a randomized controlled trial (PRP-CTS). clinicaltrials.gov/ct2/show/NCT03548259 (first received 7 June 2008).

[CD001552-bib-0060] NCT04347746. Comparison of interventions in patients with carpal tunnel syndrome. clinicaltrials.gov/ct2/show/NCT04347746 (first received 15 April 2008).

[CD001552-bib-0061] Onuma K, Fujimaki H, Kenmoku T, Sukegawa K, Takano S, Uchida K, et al. Bilateral carpal tunnel syndrome due to gouty tophi: conservative and surgical treatment in different hands of the same patient. Modern Rheumatology 2015;25(2):298–302. [DOI] [PubMed] [Google Scholar]

[CD001552-bib-0062] Roitberg B. A randomized trial of splinting vs. surgery for carpal tunnel syndrome. Surgical Neurology 2003;59(1):5–6. [DOI] [PubMed] [Google Scholar]

[CD001552-bib-0063] Rubin G, Orbach H, Rinott M, Rozen N. The effectiveness of splinting and surgery on sleep disturbance in carpal tunnel syndrome. Journal of Hand Surgery, European Volume 2018;43(3):286–9. [DOI] [PubMed] [Google Scholar]

[CD001552-bib-0064] Sahoo JK, Joshi AG. Study of effect of various treatments in clinically diagnosed patients of carpal tunnel syndrome. International Journal of Research in Pharmaceutical Sciences 2020;11(3):3105–13. [Google Scholar]

[CD001552-bib-0065] Siciliano G, Schirinzi E, Lucchesi C, Depaulis V, Calabrese R, Lencioni M, et al. Platelet rich plasma (PRP) as a new therapeutic strategy for severe median nerve compression in carpal tunnel syndrome. In: Journal of the Peripheral Nervous System. Vol. 19(Suppl 1). 2014:S31–S32.

[CD001552-bib-0066] Siegmund-Schultze N. Carpal tunnel syndrome: manual therapy is over one year as effective as surgery. Deutsches Arzteblatt International 2017;114(19):A952. [Google Scholar]

[CD001552-bib-0067] Sparapani FV, Aguiar PH, Zicarelli CA, Hirata MT, Tedesco-Marchese AJ, Teixeira MJ. Comparison between the clinical and surgical treatment in carpal tunnel syndrome by means of electrophysiological analysis [Comparação entre tratamentos clínico e cirúrgico da síndrome do túnel do carpo por métodos de análises eletrofisiológicas]. Jornal Brasileiro de Neurocirurgia 2006;17(2):52-5. [Google Scholar]

[CD001552-bib-0068] Tudiver F, Johnson ED, O’Reilly Brown M. Does surgery for carpal tunnel syndrome improve outcomes? Journal of Family Practice 2003;52(1):70–2. [PubMed] [Google Scholar]

[CD001552-bib-0069] Yoshii Y, Ishii T, Tung WL. Ultrasound assessment of the effectiveness of carpal tunnel release on median nerve deformation. Journal of Orthopaedic Research : Official Publication of the Orthopaedic Research Society 2015;33(5):726–30. [DOI] [PubMed] [Google Scholar]

[CD001552-bib-0070] Zannakis M. 5 year outcomes of carpal Rx therapy to treat carpal tunnel syndrome. In: British Journal of Pain. Vol. 14. 2020:30.

[CD001552-bib-0071] Abedi M, Mirkazemi M, Jamebozorgi K, Padidar S, Izanloo A. Evaluating the effectiveness of surgical treatment and local steroid injection in patients with carpal tunnel syndrome. Razavi International Journal of Medicine 2018;6(3):e14338. [Google Scholar]

[CD001552-bib-0072] Elwakil TF, Elazzazi A, Shokeir H. Treatment of carpal tunnel syndrome by low-level laser versus open carpal tunnel release. Lasers in Medical Science 2007;22(4):265-70. [PMID: 17334675] [DOI] [PubMed] [Google Scholar]

[CD001552-bib-0073] Hrkovic M, Lazovic M, Nikolic D, Nikcevic L, Ilic-Stojanovic O, Filipovic T. Functional status measured by Levine questionnaire in surgically and conservatively treated patients with carpal tunnel syndrome regarding nerve conduction studies. Annali Italiani Di Chirurgia 2016;87:247-51. [PubMed] [Google Scholar]

[CD001552-bib-0074] IRCT20200629047948N1. Comparison of the effectiveness of open surgery versus ultrasound guided median nerve dissection with injection of methyl prednisolone acetate in treatment of carpal tunnel syndrome. www.irct.ir/trial/49264 (first received 29 June 2020).

[CD001552-bib-0075] Linscheid RL, Peterson LF, Juergens JL. Carpal-tunnel syndrome associated with vasospasm. Journal of Bone and Joint Surgery. American Volume 1967;49(6):1141-6. [PubMed] [Google Scholar]

[CD001552-bib-0076] Lo IN, Hsu PC, Huang YC, Yeh CK, Yang YC, Wang JC. Dynamic ultrasound assessment of median nerve mobility changes following corticosteroid injection and carpal tunnel release in patients with carpal tunnel syndrome. Frontiers in Neurology 2021;12:710511. [DOI: 10.3389/fneur.2021.710511] [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD001552-bib-0077] NCT00981565. Operative versus conservative treatment in mild carpal tunnel syndrome, randomized prospective multicenter study. clinicaltrials.gov/ct2/show/NCT00981565 (first received 22 Sep 2009).

[CD001552-bib-0078] NCT04216147. Percutaneous needle electrolysis versus surgery in the treatment of carpal tunnel syndrome. clinicaltrials.gov/ct2/show/NCT04216147 (first received 2 Jan 2020).

[CD001552-bib-0079] Shazad MM, Rizwan M, Ul Haq Khawaja MI, Napar AR, Amjad A. Outcome assessment of comparison between surgical decompression versus local corticosteroid injection in carpal tunnel syndrome. Pakistan Journal of Medical and Health Sciences 2022;16(1):177–9. [DOI: 10.53350/pjmhs22161177] [DOI] [Google Scholar]

[CD001552-bib-0080] CTRI/2019/01/016881. A comparison of ultrasound guided steroid injection and endoscopic carpal tunnel release for the treatment of carpal tunnel syndrome. ctri.nic.in/Clinicaltrials/pmaindet2.php?trialid=29922&EncHid=&userName=CTRI/2019/01/016881 (first received 3 January 2019).

[CD001552-bib-0081] DRKS00014585. Conservative treatment of carpal tunnel syndrom (CTS) with the Tuttlingen CTS-Cuff versus surgery. https://trialsearch.who.int/Trial2.aspx?TrialID=DRKS00014585 (first received 26 June 2018).

[CD001552-bib-0082] DRKS00020570. The effect of stretching the transverse carpal ligament using curpal® on clinical symptoms and various neurophysiological parameters in carpal tunnel syndrome (CTS) – a prospective, randomised trial. www.drks.de/DRKS00020570 (first received 21 December 2020).

[CD001552-bib-0083] EUCTR2013-000873-56-ES. Randomized, two paralel groups, open clinical trial stratified by severity to stimate the cost-effectivity of surgical vs corticosteroid injection treatment on carpal tunnel syndrome. www.clinicaltrialsregister.eu/ctr-search/trial/2013-000873-56/ES (first received 20 May 2013).

[CD001552-bib-0084] EUCTR2021-004756-42-NO. A Norwegian trial comparing treatment strategies for carpal tunnel syndrome. https://www.clinicaltrialsregister.eu/ctr-search/search?query=eudract_number:2021-004756-42 (start date 25 January 2022).

[CD001552-bib-0085] NCT05306548. A Norwegian trial comparing treatment strategies for carpal tunnel syndrome. https://ClinicalTrials.gov/show/NCT05306548 (first posted 1 April 2022).

[CD001552-bib-0086] IRCT20141121020020N8. Comparison of long term effectiveness of open surgery versus ultrasound guided median nerve hydrodissection in terms of symptoms, function and electrophysiological studies in patients with canal tunnel syndrome (CTS). https://trialsearch.who.int/Trial2.aspx?TrialID=IRCT20141121020020N8 (first registered 21 September 2021).

[CD001552-bib-0087] NCT04014244. Efficacy of dextrose injections, corticosteroids injections and surgical release for treatment of the carpal tunnel syndrome: a prospective, randomized, double-blind controlled trial. clinicaltrials.gov/ct2/show/NCT04014244 (first received 10 July 2019).

[CD001552-bib-0088] NCT04058041. Effectiveness of neural mobilizations exercises versus surgery in patients with carpal tunnel syndrome. clinicaltrials.gov/ct2/show/NCT04058041 (first received 15 August 2019).

[CD001552-bib-0089] NCT04246216. Ultrasound-guided percutaneous electrical nerve stimulation of the median nerve for carpal tunnel syndrome. clinicaltrials.gov/ct2/show/NCT04246216 (first received 20 July 2021).

[CD001552-bib-0090] ISRCTN13164336. The Dutch injection versus operation trial in carpal tunnel syndrome patients. doi.org/10.1186/ISRCTN13164336 (first received 21 June 2017).

[CD001552-bib-0091] Palmbergen WA, Bie RM, Alleman TW, Verstraete E, Jellema K, Verhagen WI, et al. Dutch injection versus surgery trial in patients with carpal tunnel syndrome (DISTRICTS): protocol of a randomised controlled trial comparing two treatment strategies. BMJ Open 2022;12(4):e057641. [DOI: 10.1136/bmjopen-2021-057641] [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD001552-bib-0092] American Academy of Neurology. Practice parameter for carpal tunnel syndrome (summary statement). Neurology 1993;43(11):2406-9. [MEDLINE: ] [PubMed] [Google Scholar]

[CD001552-bib-0093] Atroshi I, Gummesson C, Johnsson R, Ornstein E, Ranstam J, Rosén I. Prevalence of carpal tunnel syndrome in a general population. JAMA 1999;282(2):153–8. [DOI: 10.1001/jama.282.2.153] [DOI] [PubMed] [Google Scholar]

[CD001552-bib-0094] Bialosky JE, Bishop MD, Price DD, Robinson ME, George SZ. The mechanisms of manual therapy in the treatment of musculoskeletal pain: a comprehensive model. Manual Therapy 2009;14(5):531-8. [DOI: 10.1016/j.math.2008.09.001] [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD001552-bib-0095] Bland JD. A neurophysiological grading scale for carpal tunnel syndrome. Muscle & Nerve 2000;23(8):1280–3. [DOI: ] [DOI] [PubMed] [Google Scholar]

[CD001552-bib-0096] Burton CL, Chesterton LS, Chen Y, Windt DA. Clinical course and prognostic factors in conservatively managed carpal tunnel syndrome: a systematic review. Archives of Physical Medicine & Rehabilitation 2016;97(5):836-52.e1. [DOI] [PubMed] [Google Scholar]

[CD001552-bib-0097] Carroll GJ. Comparison of median and radial nerve sensory latencies in the electrophysiological diagnosis of carpal tunnel syndrome. Electroencephalography and Clinical Neurophysiology 1987;68(2):101-6. [MEDLINE: ] [DOI] [PubMed] [Google Scholar]

[CD001552-bib-0098] Cohen MS, Hotchkiss RN, Kozin SH, Pederson WC, Wolfe SW. Green's Operative Hand Surgery. Philadelphia: Elsevier, 2017. [Google Scholar]

[CD001552-bib-0099] De Kleermaeker FG, Boogaarts HD, Meulstee J, Verhagen WI. Minimal clinically important difference for the Boston Carpal Tunnel Questionnaire: new insights and review of literature. Journal of Hand Surgery: European Volume 2018;44(3):283-9. [DOI] [PubMed] [Google Scholar]

[CD001552-bib-0100] Krom MC, Knipschild PG, Kester AD, Thijs CT, Boekkooi PF, Spaans F. Carpal tunnel syndrome: prevalence in the general population. Journal of Clinical Epidemiology 1992;45(4):373-6. [MEDLINE: ] [DOI] [PubMed] [Google Scholar]

[CD001552-bib-0101] Deeks JJ, Higgins JP, Altman DG, editor(s). Chapter 10: Analysing data and undertaking meta-analyses. In: Higgins JP, Thomas J, Chandler J, Cumpston M, Li T, Page MJ, Welch VA, editor(s). Cochrane Handbook for Systematic Reviews of Interventions Version 6.2 (updated February 2021).Cochrane, 2021. Available from www.training.cochrane.org/handbook.

[CD001552-bib-0102] Ebenbichler GR, Resch KL, Nicolakis P, Wiesinger GF, Uhl F, Ghanem AH, et al. Ultrasound treatment for treating the carpal tunnel syndrome: randomised "sham" controlled trial. BMJ 1998;316(7133):731-5. [DOI: 10.1136/bmj.316.7133.731] [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD001552-bib-0103] Erickson M, Lawrence M, Lucado A. The role of diagnostic ultrasound in the examination of carpal tunnel syndrome: an update and systematic review. Journal of Hand Therapy : Official Journal of the American Society of Hand Therapists 2021;S0894-1130(21):00061-2. [DOI: 10.1016/j.jht.2021.04.014] [DOI] [PubMed] [Google Scholar]

[CD001552-bib-0104] Felsenthal G. Median and ulnar distal motor and sensory latencies in the same normal subject. Archives of Physical Medicine and Rehabilitation 1977;58(7):297-302. [MEDLINE: ] [PubMed] [Google Scholar]

[CD001552-bib-0105] Geiringer SR. The value of inching techniques in the diagnosis of focal nerve lesions. Inching techniques are of limited use. Muscle & Nerve 1998;21(11):1557-9. [MEDLINE: ] [DOI] [PubMed] [Google Scholar]

[CD001552-bib-0106] Gelberman RH, Szabo RM, Mortenson MM. Carpal tunnel pressures and wrist position in patients with Colles fractures. Journal of Trauma, Injury, Infection and Critical Care 1984;24(8):747-9. [DOI] [PubMed] [Google Scholar]

[CD001552-bib-0107] GRADEpro Guideline Development Tool. McMaster University and Evidence Prime, 2021. Available from gradepro.org.

[CD001552-bib-0108] Hao Q, Devji T, Zeraatkar D, Wang Y, Qasim A, Siemieniuk RA, et al. Minimal important differences for improvement in shoulder condition patient-reported outcomes: a systematic review to inform a BMJ Rapid Recommendation. BMJ Open 2019;9(2):e028777. [DOI: 10.1136/bmjopen-2018-028777] [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD001552-bib-0109] Higgins JP, Li T, Deeks JJ, editor(s). Chapter 6: Choosing effect measures and computing estimates of effect. In: Higgins JP, Thomas J, Chandler J, Cumpston M, Li T, Page MJ, Welch VA, editor(s). Cochrane Handbook for Systematic Reviews of Interventions Version 6.1 (updated September 2020). Cochrane, 2020. Available from www.training.cochrane.org/handbook.

[CD001552-bib-0110] Hofer M, Ranstam J, Atroshi I. Extended follow-up of local steroid injection for carpal tunnel syndrome: a randomized clinical trial. JAMA Network Open 2021;4(10):e2130753. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD001552-bib-0111] Huisstede BM, Randsdorp MS, Coert JH, Glerum S, Middelkoop M, Koes BW. Carpal tunnel syndrome. Part II: effectiveness of surgical treatments--a systematic review. Archives of Physical Medicine and Rehabilitation 2010;91(7):1005-24. [DOI: 10.1016/j.apmr.2010.03.023] [DOI] [PubMed] [Google Scholar]

[CD001552-bib-0112] Huisstede BM, Randsdorp MS, den Brink J, Franke TP, Koes BW, Hoogvliet P. Effectiveness of oral pain medication and corticosteroid injections for carpal tunnel syndrome: a systematic review. Archives of Physical Medicine and Rehabilitation 2018;99(8):1609-22.e10. [DOI] [PubMed] [Google Scholar]

[CD001552-bib-0113] Karjalanen T, Raatikainen S, Jaatinen K, Lusa V. Update on efficacy of conservative treatments for carpal tunnel syndrome. Journal of Clinical Medicine 2022;11(4):950. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD001552-bib-0114] Karjalainen T, Heikkinen J, Busija L, Jokihaara J, Lewin AM, Naylor JM, et al. Use of placebo and nonoperative control groups in surgical trials: a systematic review and meta-analysis. JAMA Network Open 2022;5(7):e2223903. [DOI: 10.1001/jamanetworkopen.2022.23903] [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD001552-bib-0115] Karjalainen TV, Lusa V, Page MJ, O'Connor D, Massy-Westropp N, Peters SE. Splinting for carpal tunnel syndrome. Cochrane Database of Systematic Reviews 2023, Issue 2. Art. No: CD010003. [DOI: 10.1002/14651858.CD010003.pub2] [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD001552-bib-0116] Katz JN, Stirrat CR. A self-administered hand diagram for the diagnosis of carpal tunnel syndrome. Journal of Hand Surgery 1990;15(2):360-3. [DOI: 10.1016/0363-5023(90)90124-a] [DOI] [PubMed] [Google Scholar]

[CD001552-bib-0117] Kim JK, Jeon SH. Minimal clinically important differences in the Carpal Tunnel Questionnaire after carpal tunnel release. Journal of Hand Surgery: European Volume 2013;38(1):75-9. [DOI] [PubMed] [Google Scholar]

[CD001552-bib-0118] Kimura J. The carpal tunnel syndrome: localization of conduction abnormalities within the distal segment of the median nerve. Brain 1979;102(3):619-35. [MEDLINE: ] [DOI] [PubMed] [Google Scholar]

[CD001552-bib-0119] Klokkari D, Mamais I. Effectiveness of surgical versus conservative treatment for carpal tunnel syndrome: a systematic review, meta-analysis and qualitative analysis. Hong Kong Physiotherapy Journal 2018 Dec;38(2):91-114. [DOI: 10.1142/S1013702518500087] [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD001552-bib-0120] Leigh JP, Miller TR. Job-related diseases and occupations within a large workers-compensation data set. American Journal of Industrial Medicine 1998;33(3):197-211. [MEDLINE: ] [DOI] [PubMed] [Google Scholar]

[CD001552-bib-0121] Liong K, Lahiri A, Lee S, Chia D, Biswas A, Lee HP. Mid-motion deformation of median nerve during finger flexion: a new insight into the dynamic aetiology of carpal tunnel syndrome. Hand Surgery : an International Journal devoted to Hand and Upper Limb Surgery and Related Research : Journal of the Asia-Pacific Federation of Societies for Surgery of the Hand 2013;18(2):193-202. [DOI] [PubMed] [Google Scholar]

[CD001552-bib-0122] MacKinnon SE. Surgery of the Peripheral Nerve. New York, NY: Thieme Medical Publishers, 1988. [Google Scholar]

[CD001552-bib-0123] Martyn CN, Hughes RA. Epidemiology of peripheral neuropathy. Journal of Neurology, Neurosurgery and Psychiatry 1997;62(4):310-8. [MEDLINE: ] [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD001552-bib-0124] Murata K, Yajima H, Maegawa N, Hattori K, Takakura Y. Investigation of segmental carpal tunnel pressure in patients with idiopathic carpal tunnel syndrome--is it necessary to release the distal aponeurotic portion of the flexor retinaculum in endoscopic carpal tunnel release surgery? Hand Surgery : an International Journal devoted to Hand and Upper Limb Surgery and Related Research : Journal of the Asia-Pacific Federation of Societies for Surgery of the Hand 2007;12(3):205-9. [DOI] [PubMed] [Google Scholar]

[CD001552-bib-0125] Padua L, Coraci D, Erra C, Pazzaglia C, Paolasso I, Loreti C, et al. Carpal tunnel syndrome: clinical features, diagnosis, and management. Lancet Neurology 2016;15(12):1273-84. [DOI] [PubMed] [Google Scholar]

[CD001552-bib-0126] Pryse-Phillips WE. Validation of a diagnostic sign in carpal tunnel syndrome. Journal of Neurology, Neurosurgery, and Psychiatry 1984;47(8):870-2. [DOI: 10.1136/jnnp.47.8.870] [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD001552-bib-0127] Ren YM, Wang XS, Wei ZJ, Fan BY, Lin W, Zhou XH, et al. Efficacy, safety, and cost of surgical versus nonsurgical treatment for carpal tunnel syndrome: a systematic review and meta-analysis. Medicine (Baltimore) 2016;95(40):e4857. Retraction in: Medicine (Baltimore). 2017;96(16):e6778. [DOI: 10.1097/MD.0000000000004857] [DOI] [PMC free article] [PubMed] [Google Scholar] [Retracted]

[CD001552-bib-0128] Review Manager. Version 5.4. The Cochrane Collaboration, 2020.

[CD001552-bib-0129] Rosenbaum RB, Ochoa JL. Carpal Tunnel Syndrome and Other Disorders of the Median Nerve. Stoneham, MA: Butterworth-Heinemann, 1993. [Google Scholar]

[CD001552-bib-0130] Rossignol M, Stock S, Patry L, Armstrong B. Carpal tunnel syndrome: what is attributable to work? Occupational and Environmental Medicine 1997;54(7):519-23. [MEDLINE: ] [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD001552-bib-0131] Schreiber JE, Foran MP, Schreiber DJ, Wilgis EF. Common risk factors seen in secondary carpal tunnel surgery. Annals of Plastic Surgery 2005;55(3):262-5. [DOI] [PubMed] [Google Scholar]

[CD001552-bib-0132] Schrijver HM, Gerritsen AA, Strijers RL, Uitdehaag BM, Scholten RJ, De Vet Henrica CW, et al. Correlating nerve conduction studies and clinical outcome measures on carpal tunnel syndrome: lessons from a randomized controlled trial. Journal of Clinical Neurophysiology 2005;22(3):216-21. [PubMed] [Google Scholar]

[CD001552-bib-0133] Schuind F. Canal pressures before, during, and after endoscopic release for idiopathic carpal tunnel syndrome. The Journal of Hand Surgery 2002;27(6):1019-25. [DOI] [PubMed] [Google Scholar]

[CD001552-bib-0134] Schünemann HJ, Oxman AD, Vist GE, Higgins JP, Deeks JJ, Glasziou P, et al, on behalf of the Cochrane Applicability and Recommendations Methods Group. Chapter 12: Interpreting results and drawing conclusions. In: Higgins JP, Green S, editor(s). Cochrane Handbook for Systematic Reviews of Interventions Version 5.1.0 (updated June 2017). Cochrane Collaboration, 2017. Available from: www.cochrane-handbook.org.

[CD001552-bib-0135] Shi Q, MacDermid JC. Is surgical intervention more effective than non-surgical treatment for carpal tunnel syndrome? A systematic review. Journal of Orthopaedic Surgery and Research 2011;6:17. [DOI: 10.1186/1749-799X-6-17] [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD001552-bib-0136] Shi Q, Bobos P, Lalone EA, Warren L, MacDermid JC. Comparison of the short-term and long-term effects of surgery and nonsurgical intervention in treating carpal tunnel syndrome: a systematic review and meta-analysis. Hand (New York, N.Y.) 2020;15(1):13-22. [DOI: 10.1177/1558944718787892] [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD001552-bib-0137] Shiri R, Heliövaara M, Moilanen L, Viikari J, Liira H, Viikari-Juntura E. Associations of cardiovascular risk factors, carotid intima-media thickness and manifest atherosclerotic vascular disease with carpal tunnel syndrome. BMC Musculoskeletal Disorders 2011;12:80. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD001552-bib-0138] Sterne JA, Savović J, Page MJ, Elbers RG, Blencowe NS, Boutron I, et al. RoB 2: a revised tool for assessing risk of bias in randomised trials. BMJ 2019;366:l4898. [DOI] [PubMed] [Google Scholar]

[CD001552-bib-0139] Stevens JC. AAEM minimonograph: 26: the electrodiagnosis of carpal tunnel syndrome. American Association of Electrodiagnostic Medicine. Muscle & Nerve 1997;20(12):1477-86. [DOI] [PubMed] [Google Scholar]

[CD001552-bib-0140] Stewart JD. Compression and entrapment neuropathies. In: Dyck PJ et al, editors(s). Peripheral Neuropathy. 3rd edition. Philadelphia, PA: Saunders, 1993. [Google Scholar]

[CD001552-bib-0141] Teitsma XM, Jacobs JW, Welsing PM, Pethö-Schramm A, Borm ME, Hendriks L, et al. Patient-reported outcomes in newly diagnosed early rheumatoid arthritis patients treated to target with a tocilizumab- or methotrexate-based strategy. Rheumatology (Oxford) 2017;56(12):2179-89. [DOI: 10.1093/rheumatology/kex319] [DOI] [PubMed] [Google Scholar]

[CD001552-bib-0142] Vasiliadis HS, Georgoulas P, Shrier I, Salanti G, Scholten RJ. Endoscopic release for carpal tunnel syndrome. Cochrane Database of Systematic Reviews 2014, Issue 1. Art. No: CD008265. [DOI: 10.1002/14651858.CD008265.pub2] [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD001552-bib-0143] Verdugo RJ, Ochoa JL. "Sympathetically maintained pain" I. Phentolamine block questions the concept. Neurology 1994;44(6):1003-10. [MEDLINE: ] [DOI] [PubMed] [Google Scholar]

[CD001552-bib-0144] Walters SJ, Brazier JE. Comparison of the minimally important difference for two health state utility measures: EQ-5D and SF-6D. Quality of Life Research : an International Journal of Quality of Life Aspects of Treatment, Care and Rehabilitation 2005;6:1523-32. [DOI: 10.1007/s11136-004-7713-0] [DOI] [PubMed] [Google Scholar]

[CD001552-bib-0145] Yamanaka Y, Tajima T, Tsujimura Y, Kosugi K, Mano Y, Zenke Y, et al. Molecular and clinical elucidation of the mechanism of action of steroids in idiopathic carpal tunnel syndrome. The Journal of Bone and Joint Surgery 2021;103(19):1777-87. [DOI: 10.2106/JBJS.20.02096] [DOI] [PubMed] [Google Scholar]

[CD001552-bib-0146] Verdugo RJ, Salinas RA, Castillo JL, Cea JG. Surgical versus non-surgical treatment for carpal tunnel syndrome. Cochrane Database of Systematic Reviews 2008, Issue 4. Art. No: CD001552. [DOI: 10.1002/14651858.CD001552.pub2] [DOI] [PMC free article] [PubMed] [Google Scholar]

PERMALINK

Surgical versus non‐surgical treatment for carpal tunnel syndrome

Vieda Lusa

Teemu V Karjalainen

Markus Pääkkönen

Tuomas Jaakko Rajamäki

Kati Jaatinen

Abstract

Background

Objectives

Search methods

Selection criteria

Data collection and analysis

Main results

Authors' conclusions

Plain language summary

Summary of findings

Summary of findings 1. Surgery compared to splint for carpal tunnel syndrome in the long term.

Summary of findings 2. Surgery compared to corticosteroid injection for carpal tunnel syndrome in the long term.

Background

Description of the condition

Description of the intervention

How the intervention might work

Why it is important to do this review

Objectives

Methods

Criteria for considering studies for this review

Types of studies

Types of participants

Types of interventions

Types of outcome measures

Primary outcomes

Secondary outcomes

Search methods for identification of studies

Electronic searches

Searching other resources

Data collection and analysis

Selection of studies

Data extraction and management

Assessment of risk of bias in included studies

Measures of treatment effect

Unit of analysis issues

Dealing with missing data

Assessment of heterogeneity

Assessment of reporting biases

Data synthesis

Subgroup analysis and investigation of heterogeneity

Sensitivity analysis

Summary of findings and assessment of the certainty of the evidence

1. Surgery compared to splint for carpal tunnel syndrome in the short term.

2. Surgery compared to corticosteroid injection for carpal tunnel syndrome in the short term.

Results

Description of studies

Results of the search

1.

Included studies

Participants

3. Received treatment before inclusion in the study.

Interventions

Surgery

Non‐surgical interventions

Co‐interventions

Outcomes

1. Primary outcome: clinical improvement

2. Symptoms

3. Function

4. Pain

5. Health‐related quality of life score (HRQoL)

6. Adverse effects

7. Need for further surgery or secondary surgery during follow‐up

Unit of analysis

Funding

Excluded studies

Risk of bias in included studies

Effects of interventions

1. Surgery versus splint

Primary outcome

1) Clinical improvement

Short‐term improvement

1.1. Analysis.