The Effect of Acupuncture on Hand and Wrist Pain Intensity, Functional Status, and Quality of Life in Adults: A Systematic Review

Abstract

Objective: This systematic review examined the effects of acupuncture on hand-and-wrist pain intensity, functional status, quality of life, and incidence of adverse effects in adults.

Methods: Searches of 6 databases and previous reviews for randomized controlled trials (RCTs) were performed. Each outcome was analyzed for participant conditions, interventions, controls, and follow-up times determined a priori. Active controls were excluded. Follow-up periods were based on Cochrane 5.1.0 guidelines. The results were tabulated and described narratively.

Results: In the 10 included RCTs (622 participants), 6 had a low risk of bias. For cryotherapy-induced pain, 1 trial showed significant pain reduction post treatment. For rheumatoid arthritis, 1 trial shown significant pain reduction and function improvements post treatment and short-term. For carpal tunnel syndrome, 1 trial showed significant pain reduction and functional improvements intermediate-term, while 3 trials suggested no significant difference. For tenosynovitis, 1 trial showed significant pain reduction and function improvements short-term. For poststroke impairments, 1 trial showed significant function improvements post treatment and at short-term, while another trial suggested no significant difference. No significant improvements were noted for trapezio-metacarpal joint osteoarthritis. In 2 trials, adverse effects occurred in patients with carpal tunnel syndrome; yet acupuncture appeared to be relatively safe.

Conclusions: Acupuncture may be effective and safe for short-term pain reduction and functional improvement in hand-and-wrist conditions. Clinicians should interpret the results with caution due to small sample sizes and clinical heterogeneity. Future research is warranted.

Introduction

Pain is a common symptom resulting from many hand- or wrist-related disorders and injuries. Pain can be caused by physiologic disorders such as rheumatoid arthritis¹ and trigger finger,² as well as by injuries such as fractures.³ Occasionally, infections may also lead to pain.^4,5 Due to its wide variety of causes, hand-and-wrist pain can manifest in individuals who have various demographic characteristics with a range of intensities and durations,⁶ in addition to having a substantial prevalence in some populations.^7–9

A 2019 systematic review of 32 cross-sectional studies suggested that the median prevalence of wrist pain was 4.2% within the general population, and 24% in individuals who participate in demanding occupations or sports.⁸ The pain, along with its underlying disorder, often result in a decrease in the functions of the hand and wrist and may also interfere severely with one's daily activities and quality of life (QoL). Carpal tunnel syndrome, for instance, may result in severe pain and decreased grip strength, along with a reduction in health-related QoL.¹⁰ Ultimately, this may lead to substantial opportunity costs due to lost productivity time.¹¹

Currently, many conventional treatments are used to relieve hand-and-wrist pain and improve function, including, but not limited to, splints, exercises, pain medications, and steroid injections.^12–15 Due to the potential drawbacks of conventional treatments, such as poor user experience when splinting¹⁶ and adverse effects of steroid injections,¹⁷ some patients have turned to alternative treatments, such as acupuncture and its variants, to address the musculoskeletal pain and functional problems.¹⁸

Acupuncture is a practice that originated from China and has also been historically present in Japan and Korea.¹⁹ Modern acupuncture involves insertion of fine, flexible needles into specific acupoints by trained acupuncturists or practitioners and may involve other stimulations such as electricity.²⁰ Dry needling acupuncture have also been developed to treat specific conditions with potential efficacy.²¹ Previous studies have demonstrated that acupuncture may produce its analgesic effects by stimulating release of hormones and neurotransmitters such as serotonin,²² dopamine,²³ and endorphins.²⁴ In addition, acupuncture may locally reduce the levels of inflammatory mediators, such as tumor necrosis factor-α.²⁵

Within the last 2 decades (in 2005, 2008, and 2011), the clinical efficacy and safety of acupuncture for a variety of musculoskeletal conditions, such as plantar fasciitis and lower-back pain, have been examined in many trials and reviews.^26–29 However, to the current authors' knowledge, there has yet to be a systematic review that examines the effect and safety of acupuncture on the hand and wrist in adult populations. Moreover, randomized controlled trials (RCTs) that examine related topics typically have small sample sizes with a mix of outcome-reporting methodologies, which often leads to insufficient power and a lack of generalizability of the results.³⁰

Due to the aforementioned reasons, the current systematic review was conducted to summarize the findings of relevant RCTs, and to assess, with greater confidence and applicability, the efficacy of acupuncture for pain reduction, functional status improvements, and QoL improvements, as well as the incidence of adverse effects.³¹

Methods

The guidelines from the 2021 online version of the Cochrane Handbook for Systematic Reviews of Interventions³² were followed to conduct this systematic review.

Search Strategy

With the assistance of a health science librarian who has extensive experience with Cochrane systematic reviews, systematic searches were made in the following English-language databases from inception to April 8, 2021 for relevant trials: (1) Medical Literature Analysis and Retrieval System Online (MEDLINE^®); (2) Excerpta Medica Database (EMBASE); (3) Allied and Complementary Medicine (AMED); (4) Cochrane Central Register of Controlled Trials (CENTRAL); (5) Cumulative Index of Nursing and Allied Health Literature (CINAHL); and (6) Physiotherapy Evidence Database (PEDro). The search strategy is shown in supplementary Tables S1 to S6 (online only). Hand searches were also performed in the reference sections of previous reviews of relevant RCTs. Trials of all languages were included or screening.

Types of Studies

Published parallel or crossover RCTs in full-text or abstract form were included when sufficient information for analyses was obtained. Any quasi-RCTs and clinical controlled trials were excluded.

Eligibility Criteria

Participants

Participants in the studies that were analyzed had to meet the following criteria to be included in the review: (1) Adult (age 18 or older) and (2) presented with any type of hand- or wrist-related condition.

Trials with child or adolescent participants (younger than 18) were excluded from the review. Trials that did not report the age range of the participants and were not able to be judged reasonably to have only included adult participants were excluded from the review.

Types of interventions

Trials must have used acupuncture with penetrating needles with a methodology that punctures the skin. Stimulation might or might not have been applied to the needles. Types of stimulation included, but were not limited to, manual, electric, and heat. Trials that used interventions without skin penetration (e.g., transcutaneous electrical nerve stimulation [TENS]) were excluded from this review.

Types of controls

Trials must have compared the intervention with sham acupuncture, inactive treatment, wait-list, or no treatment. Trials without control groups or trials that compared interventions to active treatments, such as night-splinting, were excluded from this review.

Outcomes and Measures

Outcomes were to be measured using clinically validated or custom-made scales and questionnaires, or clinically validated indicators. The primary outcomes of interest of this systematic review were: (1) pain-intensity change; (2) functional-status change; and (3) QoL. Scales or indicators that reported on 2 or more outcomes of interest simultaneously (e.g., global symptom scale being an indicator for both pain intensity and functional status) were not included. The secondary outcome of interest was incidence of adverse effects.

Follow-Up Timeperiod Definitions

The 2019 Cochrane 5.1.0 guidelines (in print) for follow-up timeperiods were followed.³³ A follow-up timeperiod was defined as the time between the completion of the last intervention session and the measurement of the outcome. Post treatment was defined as up to 1 day for follow-up. Short-term was defined as more than 1 day to 3 months. Intermediate-term was defined as between more than 3 months and 1 year. Long-term was defined as longer than 1 year. The latest outcome was used to incorporate time-based effects if multiple outcomes from the same trial fell within a defined timeperiod.³³

Study Selection

Two authors (F.Z. and C.Y.W.) performed title and abstract screening using Rayyan³⁴ independently and in duplicate based on the eligibility criteria. Relevant abstracts were entered into the full-text screening phase. Two authors (F.Z. and J.D.) performed full-text screening independently and in duplicate using an Excel spreadsheet. Conflicts during both phases were resolved by recruiting a third author (N.B.) for arbitration. Attempts were made to obtain the full-texts from the first authors or the corresponding authors if the full-text were not retrievable.

Data Extraction

Two authors (F.Z. and N.B.) performed data extraction independently and in duplicate using an Excel spreadsheet. Disagreements were resolved through discussion and consensus. Attempts were made to contact the first author and the corresponding author for any missing or unpublished data. These requests were followed-up after 1 month if no responses occurred.

Risk of Bias Assessment

Two authors (F.Z. and N.B.) assessed the RCTs for methodological quality independently using prepiloted forms based on the Cochrane Back Review Group risk of bias assessment tool (12 criteria).³⁵ Each item was rated to have either a low, unclear, or high risk of bias. The overall risk of bias was rated as either low or high risk of bias. Conflicts were resolved by recruiting a third author (K.T.) for arbitration. No predefined cutoff score was used for the risk of bias assessment.

Qualitative Synthesis

The findings of the included trials were tabulated and narratively described. Due to clinical heterogeneity, it was decided to analyze the outcomes in subgroups of participant conditions, interventions, controls, and follow-up durations a priori. This ensured clinical homogeneity and proper analysis of the results.³⁶

Quality of Evidence Assessment

Outcomes were assessed for overall quality of evidence using the Cochrane Grades of Recommendation, Assessment, Development and Evaluation (GRADE) approach.³⁷ the GRADE approach was used as recommended in the Cochrane Collaboration's reference, The GRADE Handbook.³⁸ In addition, high-quality evidence was reserved for conclusions that were unlikely to make incorrect inferences—high-quality evidence should have consistent results from multiple trials with low risks of bias. Therefore, inconsistencies were downgraded by 1 level for outcomes with only 1 entry.

Results

Study Selection

The systematic database searches identified 2330 articles (Fig. 1). There were 14 reviews^39–52 from which 284 additional articles were retrieved. After discarding the duplicates, 1731 articles were screened. After both screening phases, a total of 10 RCTs^53–62 were included.

FIG. 1.

Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow diagram. AMED, Allied and Complementary Medicine; CENTRAL, Cochrane Central Register of Controlled Trials; CINAHL, Cumulative Index of Nursing and Allied Health Literature; EMBASE, Excerpta Medica Database; MEDLINE,^® Medical Literature Analysis and Retrieval System Online; PEDro, Physiotherapy Evidence Database.

Included Studies

See Table S7 (online only) for characteristics of the included studies.

General characteristics

All 10^53–62 trials were parallel RCTs. Six trials^{53,57,59–62} were single-center trials, 3 trials^55,56,58 were multicenter trials, and 1 trial⁵⁴ did not report the number of centers involved. The trials were performed in Canada,⁵³ the United Kingdom,⁵⁶ the United States,^57,58,61,62 Hong Kong,⁵⁵ Iran,^54,60 and Portugal.⁵⁹

Study participants

A total of 622 participants were included. One trial⁵³ included 30 participants with cryotherapy-induced pain, 1 trial⁵⁶ included 12 participants with trapezio-metacarpal joint osteoarthritis (OA), 1 trial⁵⁹ had 105 participants with hand rheumatoid arthritis (RA), 4 trials^55,57,58,62 included a total of 360 participants with carpal tunnel syndrome, 1 trial⁵⁴ reported 58 participants with tenosynovitis, and 2 trials^60,61 reported 57 participants with upper-extremity impairment due to previous strokes.

Interventions and controls

Three RCTs^56,59,62 used verum acupuncture with manual stimulation. Three trials^53,55,57 used acupuncture with electrical stimulation. Two trials^58,61 used both manual and electroacupuncture (EA). Two trials^54,60 reported the use of dry needling acupuncture.

Six RCTs^{54–56,60–62} had 2 groups, 2 trials^57,58 had 2 treatment groups with 1 control group, and 2 trials^53,59 had 1 treatment group with 2 control groups.

Five trials^{57,58,60–62} included a nonpenetrating sham-acupuncture control group, 2 trials^54,55 included a wait-list or no-treatment control group, and 1 trial⁵⁶ included an inactive control group. Two trials^53,59 included both an off-point sham acupuncture control group and a wait-list or no-treatment control group.

Reported outcomes

Two RCTs^53,57 reported on pain intensity alone and 3 trials^58,60,62 reported outcomes on functional-status changes alone. One trial⁶¹ reported both functional status changes and QoL changes. Four trials^54–56,59 reported both pain intensity and functional-status changes. Only 2 trials^55,62 reported the incidence of adverse effects.

With regard to follow-up timeperiods, 2 trials^53,57 reported their outcomes post treatment, and 2 trials^54,61 reported their outcomes for short-term follow-ups. Five trials^{56,58–60,62} reported their outcomes both post treatment and for short-term follow-ups. One trial⁵⁵ reported its outcomes for both short- and intermediate-term follow-ups. No trial reported an outcome for long-term follow-up.

Risk of Bias of Included Studies

The results of the risk of bias evaluations is shown in Figures 2 and 3.

FIG. 2.

Risk of bias traffic-light plot.

FIG. 3.

Risk of bias summary graph.

The studies avoided or had similar co-interventions and had similar timing for outcome assessments. The majority of trials were rated to have a low risk of bias for random-sequence generation (80%), allocation concealment (70%), patient blinding (70%), outcome-assessor blinding (80%), dropout rate (80%), participants analyzed in groups that they were allocated to (60%), selective reporting (90%), and similarity of baseline characteristics (80%). However, the majority of trials were rated to have an unclear risk for compliance (70%) due to inadequate descriptions. In addition, the majority of trials were rated to have a high risk for care-provider blinding (90%) due to the nature of acupuncture interventions. Overall, 6 trials were rated to have a low risk of bias and 4 trials were rated to have a high risk of bias.

Participants with Cryotherapy-Induced Pain

One trial reported on pain intensity⁵³ (Table 1) in patients with cryotherapy-induced pain.

Table 1.

Acupuncture for Participants with Cryotherapy-Induced pain

Study	Comparator	Outcome	Follow-up	Acupuncture Group	Control Group	Between-group P-value	Comment
Anderson et al. (1974)53	Off-point sham	Pain (0–11 scalea)	Post treatment	Mean = 5.9	Mean = 7.8	P < 0.01	Measurement of distribution not reported.
	Wait-list or no treatment	Pain (0–11 scalea)	Post treatment	Mean = 5.9	Mean = 8.5	P < 0.001

^a Lower score indicates less pain.

In 1974, Anderson et al.⁵³ (n = 30 analyzed/30 randomized) included an EA group (n = 10), an off-point sham EA group (n = 10), and a no-treatment group (n = 10). Immediately post treatment, there was a statistically significant between-group difference in pain intensity score that favored the EA group, compared to both control groups. However, the GRADE assessments indicated a very low quality of evidence for both outcomes. Table S8 (online only) shows all GRADE assessments).

Participants with Trapezio-Metacarpal Joint OA

One trial reported on pain intensity and functional status⁵⁶ (Table 2) in patients with trapezio-metacarpal join RA.

Table 2.

Outcomes for Participants with Trapezio-Metacarpal Joint Osteoarthritis

Study	Comparator	Outcome	Follow-up	Acupuncture group	Control group	Between-group P-value	Comment
Dickens et al. (1989)56	Inactive treatment	Pain (VASa)	Post treatment	Change in median (95% CI) = −69.22 (−76.17, −32.77)	Change in median (95% CI) = −73 (−8.01, 0)	NR	Only reported change from baseline. Did not report absolute final values.
			Short-term (2 weeks)	Change in median (95% CI) = −76.1 (−90.54, −28.37)	Change in median (95% CI) = −20 (−90.16, −2.39)	P = 1
		Functional statusb	Post treatment	Change in median (95% CI) = 1 (0, 6)	Change in median (95% CI) = 2 (0, 7)	P > 0.05
			Short-term (2 weeks)	Change in median (95% CI) = 2 (1, 7)	Change in median (95% CI) = 1 (0, 7)	P > 0.05

^a Lower score indicates less pain.

^b Higher score indicates better function.

VAS, visual analogue scale; CI, confidence interval; NR, not reported.

In 1989, Dickens et al.⁵⁶ (n = 12/13) included a verum acupuncture group (n = 7) and a mock TENS group (n = 5). Immediately post treatment, the control group was favored for pain intensity visual analogue scale (VAS) scores and the P-value was not reported. Functional status scores did not differ significantly. In the short-term follow-up, neither outcome differed significantly between 2 two groups. The GRADE assessments indicated a very low quality of evidence for all outcomes.

Participants with RA

One trial reported on pain intensity and functional status⁵⁹ in patients with RA (Table 3).

Table 3.

Participants with Hand Rheumatoid Arthritis

Study	Comparator	Outcome	Follow-up	Acupuncture group	Control group	Between-group P-value	Comment
Seca et al. (2019)59	Off-point sham	Pain (VASa)	Post treatment	Change in median (IQR) = −4 (−6, −3) (P < 0.001)	Change in median (IQR) = −1 (−2, −1) (P < 0.001)	P < 0.001	Skewed data. Only reported median and IQR.
			Short-term (4 wks)	Median (IQR) = 1 (0, 2)	Median (IQR) = 4 (3, 6)	P < 0.001
		Functional status (hand grasp strengthb [kgf])	Post treatment	Change in median (IQR) = 6.2 (4.65, 8.8) (P < 0.001)	Change in median (IQR) = −0.05 (−1.55, 1.95) (P = 0.808)	P < 0.001
			Short-term (4 wks)	Median (IQR) = 21.83 (15.53, 24.83)	Median (IQR) = 11.5 (9.1, 20.2)	P < 0.001
		Functional status (HAQc)	Post treatment	Change in median (IQR) = 0.25 (−0.5, −0.125) (P < 0.001)	Change in median (IQR) = 0 (−0.125, 0.125) (P = 0.597)	P < 0.001
			Short-term (4 wks)	Median (IQR) = 1 (0.75, 1.25)	Median (IQR) = 1.25 (0.5, 1.5)	P = 0.057
	Wait-list or no treatment	Pain (VASa)	Post treatment	Change in median (IQR) = −4 (−6, −3) (P < 0.001)	Change in median (IQR) = 0 (0, 2) (P-value NR)	P < 0.001
			Short-term (4 wks)	Median (IQR) = 1 (0, 2)	Median (IQR) = 5 (4, 7)	P < 0.001
		Functional status (hand grasp strengthb [kgf])	Post treatment	Change in median (IQR) = 6.2 (4.65, 8.8) (P < 0.001)	Change in median (IQR) = −0.295 (−1.67, 0.63) (P = 0.029)	P < 0.001
			Short-term (4 wks)	Median (IQR) = 21.83 (15.53, 24.83)	Median (IQR) = 11.126 (7.94, 15.25)	P < 0.001
		Functional status (HAQc)	Post treatment	Change in median (IQR) = −0.25 (−0.5, −0.125) (P < 0.001)	Change in median (IQR) = 0.125 (0, 0.25) (P = 0.012)	P < 0.001
			Short-term (4 weeks)	Median (IQR) = 1 (0.75, 1.25)	Median (IQR) = 1.25 (1, 1.625)	P = 0.057

^a Lower score indicates less pain.

^b Higher value indicates better function.

^c Lower score indicates better function.

VAS, visual analogue scale; IQR, interquartile range; wks, weeks; NR, not reported; kgf, kilogram force; HAQ, Health Assessment Questionnaire.

In 2019, Seca et al.⁵⁹ (n = 102/105) included a verum acupuncture group (n = 34), an off-point sham group (n = 33), and a wait-list group (n = 35). Compared to off-point sham, verum acupuncture was significantly favored both post treatment and at a short-term follow-up for all outcomes and measures, including pain intensity VAS score, hand-grasp strength, and the Disability Index of the Health Assessment Questionnaire (HAQ) score. The GRADE assessments indicated moderate quality for these outcomes.

Compared to the wait-list group, verum acupuncture was significantly favored post treatment and at a short-term follow-up for pain intensity VAS score and hand-grasp strength. Verum acupuncture was only significantly favored for HAQ post treatment, and the trial did not reach statistical significance for the outcome at the short-term follow-up. The GRADE assessments indicated moderate quality for short-term pain and post treatment functional status and low quality of evidence for post treatment pain and short-term functional status.

Participants with Carpal Tunnel Syndrome

Four trials reported on pain intensity and functional status in patients with carpal tunnel syndrome.^55,57,58,62 Among these trials, 2 reported on the incidence of adverse effects^55,62 (Table 4).

Table 4.

Participants with Carpal Tunnel Syndrome

Study	Comparator	Outcome	Follow-up	Acupuncture group	Control group	Between-group P-value	Comments
Chung et al. (2016)55	Wait-list or no treatment	Pain (NRSb)	Short-term (5 wks)	Change in mean (95% CI) = −0.68 (−1.18, −0.19)	Change in mean (95% CI) = −0.55 (−1.11, 0.02)	P = 0.5	Only reported change from baseline. Did not report absolute final values.
			Intermediate-term (17 wks)	Change in mean (95% CI) = −1.22 (−1.79, −0.65)	Change in mean (95% CI) = −0.61 (−1.22, 0.00)	P = 0.03
		Functional status (BCTQ FFSc)	Short-term (5 wks)	Change in mean (95% CI) = −0.01 (−0.12, 0.11)	Change in mean (95% CI) = 0.06 (−0.07, 0.18)	P = 0.3
			Intermediate-term (17 wks)	Change in mean (95% CI) = −0.16 (−0.28, −0.04)	Change in mean (95% CI) = 0.02 (−0.13, 0.17)	P = 0.01
		Adverse event incidence	Intermediate-term (17 wks)	n = 11	n = 12	P = NR	—
Maeda et al. (2013)57	Non-penetrating sham	Pain (VASb)	Post treatment	Change in mean (SD) = −1.2 (1.8)	Change in mean (SD) = −0.5 (1.4)	P > 0.1	Only reported change from baseline. Did not report absolute final values.
Maeda et al. (2017)58	Non-Penetrating sham	Functional status (BCTQ FFSc [%])	Post treatment	Change in mean (SD) = −26.2 (19.1)	Change in mean (SD) = −18.2 (33.6)	P = 0.23	Only reported change from baseline. Did not report absolute final values.
			Short-term (3 mos)	Change in mean (SD) = −16.8 (32.4)	Change in mean (SD) = −2.3 (33.7)	P = 0.18
Yao et al. (2012)62	Non-penetrating sham	Functional status (key pincha [lb])	Short-term (3 mos)	Mean (SD) = 14.0 (5.4)	Mean (SD) = 13.8 (4.9)	P = 0.87	—
		Functional status (tip pincha [lb])	Short-term (3 mos)	Mean (SD) = 10.9 (4.6)	Mean (SD) = 10.5 (4.0)	P = 0.75
		Adverse event incidence	Short-term (3 mos)	n = 0 (serious events only)	n = 0 (serious events only)	P = NR

^a Higher value indicates better function.

^b Lower score indicates less pain.

^c Lower score indicates better function.

NRS, numeric rating scale; wks, weeks; CI, confidence interval; BCTQ, Boston Carpal Tunnel Questionnaire; FFS, Functional Status Scale; NR, not reported; VAS, visual analogue scale; SD, standard deviation; mos, months.

In 2016, Chung et al.⁵⁵ (n = 181/181) included an EA group (n = 90) and a wait-list group (n = 91). EA was significantly favored for pain intensity numeric rating scale (NRS) score and the Boston Carpal Tunnel Questionnaire (BCTQ) Functional Status Scale (FFS) score at an intermediate term follow-up, while no significant differences in either measures were observed in a short-term follow-up. The trial reported 11 and 12 incidents of adverse effects by the end of the intermediate term follow-up for the EA group and the control group, respectively. The GRADE assessments indicated a moderate quality of evidence for BCTQ FFS and incidence of adverse effects during intermediate-term and low quality of evidence for the other outcomes.

In 2013, Maeda et al.⁵⁷ (n = 41/41) included an EA group (n = 22) and a nonpenetrating sham group (n = 19). Immediately post treatment, EA resulted in a greater reduction in pain intensity VAS score, compared to nonpenetrating sham. However, the difference in the VAS scores did not reach statistical significance. The GRADE assessment indicated a very low quality of evidence for this outcome.

In 2017, Maeda et al.⁵⁸ (n = 43/51 post treatment, 38/51 short-term) included an EA group (n = 22, 21) and a nonpenetrating sham group (n = 21, 17). EA was favored for BTCQ FFS score post treatment and at a short-term follow-up. However, between-group differences did not reach statistical significance for either timeperiod. The GRADE assessments indicated a very low quality of evidence for the outcome post treatment, and a low quality for the short-term follow-up.

In 2012, Yao et al.⁶² (n = 41/41) included a verum acupuncture group (n = 20) and a nonpenetrating sham group (n = 21). At a short-term follow-up, key pinch and tip pinch were relatively similar between the groups and statistical significance was not reached. The trial reported no incidents of serious adverse effects in either group at the follow-up timeperiod. The GRADE assessments indicated a low quality of evidence for functional status and a very low quality of evidence for incidence of adverse events.

Participants with Tenosynovitis

One trial reported on pain intensity and functional status in patients with tenosynovitis⁵⁴ (Table 5).

Table 5.

Participants with Tenosynovitis

Study	Comparator	Outcome	Follow-up	Acupuncture group	Control group	Between-group P-value	Comment
Azizian et al. (2019)54	Wait-list or no treatment	Pain (VASa)	Short-term (1 wk)	Mean (SD) = 1.79 (0.54)	Mean (SD) = 3.59 (0.9)	P < 0.05	—
		Functional status (pinch gripb [kg])	Short-term (1 wk)	Mean (SD) = 5.4 (1.45)	Mean (SD) = 4.63 (1.5)	P < 0.05

^a Lower score indicates less pain.

^b Higher value indicates better function.

VAS, visual analogue scale; wk, week; SD, standard deviation; kg, kilogram.

In 2019, Azizian et al.⁵⁴ (n = 58/58) included a dry needling acupuncture group (n = 29) and a no-treatment group (n = 29). At a short-term follow-up, dry needling acupuncture was significantly favored for pain intensity VAS score and pinch grip. The GRADE assessments indicated a low quality of evidence for both outcomes.

Participants with Poststroke Impairment

One trial reported on functional status and 1 trial reported on functional status and quality of life in patients with poststroke impairment^60,61 (see Table 6).

Table 6.

Participants with Poststroke Upper-Extremity Impairment

Study	Comparator	Outcome	Follow-up	Acupuncture group	Control group	Least-squares mean difference between groups (95% CI)a	Between-group P-value	Comments
Tavakol et al. (2021)60	Non-penetrating sham	Functional status (MMAS)	Post treatment	Mean (SD) = 0.92 (0.90)	Mean (SD) = 2.00 (0.85)	—	P = 0.012	—
			Short-term (1 mo)	Mean (SD) = 1.42 (0.90)	Mean (SD) = 2.00 (0.85)	—
		Functional status (box-and-block testb)	Post treatment	Mean (SD) = 7.00 (9.42)	Mean (SD) = 3.34 (2.74)	—	P = 0.249
			Short-term (1 mo)	Mean (SD) = 6.84 (9.54)	Mean (SD) = 3.25 (2.77)	—
		Functional status (passive ROMb)	Post treatment	Mean (SD) = 93.25 (3.98)	Mean (SD) = 64.08 (22.23)	—	P = 0.04
			Short-term (1 mo)	Mean (SD) = 94.08 (4.09)	Mean (SD) = 64.18 (20.98)	—
		Functional status (active ROMb)	Post treatment	Mean (SD) = 26.16 (27.59)	Mean (SD) = 14.50 (16.12)		P = 0.249
			Short-term (1 month)	Mean (SD) = 25.836 (27.59)	Mean (SD) = 14.75 (16.12)
Wayne et al. (2005)61	Non-penetrating sham	Functional status (ROM wrist, sagittal planeb)	Short-term (12 wks)	—	—	18.20 (−17.1, 53.5)	P = 0.29	Did not report absolute values.
		Functional status (ROM wrist, frontal planeb)	Short-term (12 wks)	—	—	10.47 (−2.6, 23.5)	P = 0.11
		Functional status (ROM thumbb)	Short-term (12 wks)	—	—	1.15 (−8.1, 10.4)	P = 0.79
		Functional status (ROM digitsb)	Short-term (12 wks)	—	—	6.91 (−3.2, 17.0)	P = 0.17
		Functional status (3 jaw chuckb)	Short-term (12 wks)	—	—	–2.22 (−6.2, 1.7)	P = 0.25
		QoL (NHPc)	Short-term (12 wks)	—	—	–1.27 (−7.5, 4.9)	P = 0.68

^a Least-squares mean of the acupuncture group minus the least-squares mean of the control treatment group.

^b Higher value indicates better function.

^c Higher score indicates better QoL.

^d Lower score indicates better function.

CI, confidence interval; MMAS, modified Modified Ashworth Scale; SD, standard deviation; mo, month; ROM, range of motion; wks, weeks; QoL, quality of life; NHP, Nottingham Health Profile.

In 2021, Tavakol et al.⁶⁰ (n = 24/24) included a dry needling acupuncture group (n = 12) and a sham dry needling acupuncture group (n = 12). Dry needling acupuncture was favored immediately post treatment and at a short-term follow-up for all measures, including modified Modified Ashworth Scale (MMAS), the box-and-block test, passive range of motion (ROM), and active ROM. However, only the differences in MMAS and passive ROM results reached statistical significance. The GRADE assessments indicated a low quality of evidence for all outcomes.

In 2005, Wayne et al.⁶¹ (n = 33/33) included a combined manual and EA group (n = 16) and a nonpenetrating sham acupuncture group (n = 17). At a short-term follow-up, mixed results were observed for both functional status and QoL changes. Acupuncture resulted in greater ROM improvements compared to the control condition, while that condition resulted in greater improvements in 3 jaw chuck grasping and the Nottingham Health Profile. However, the study did not reach statistical significance in any of the reported outcomes. The GRADE assessments indicated a low quality of evidence for all aforementioned outcomes.

Discussion

Summary of the Evidence

The results of this systematic review suggested that acupuncture and related interventions may be efficacious for various hand-and-wrist conditions with regard to pain relief and functional-status improvement across multiple timeperiods. Evidence suggested that the acupuncture groups were generally favored for cryotherapy-induced pain, RA, tenosynovitis, and poststroke impairment—either immediately post treatment or in the short-term, or in both follow-up periods. For carpal tunnel syndrome, acupuncture was favored over the intermediate-term as opposed to the shorter timeperiods. However, convincing evidence was not obtained showing that acupuncture resulted in significant improvements in patients with trapezio-metacarpal joint OA.

With regard to the effect on QoL, it was not possible to establish that acupuncture was significantly better due to a lack of reporting on this outcome for most of the conditions. No evidence was obtained on the effect of acupuncture in long-term follow-ups. It must also be noted that, for all the outcomes when the point estimate favored the control group, none reached statistical significance.

In terms of the safety of acupuncture, 2 of the 10 trials reported adverse effects in patients with carpal tunnel syndrome. Yet, based on these results, verum and EA appeared to be relatively safe for those patients, compared with nonpenetrating sham and wait-list, respectively.

The risk of bias assessments indicated that the majority of the trials had an overall low risk of bias. Most, if not all, of the trials were rated to be of low bias in the randomization process, outcome-assessor blinding, dropout rate, selective reporting, co-intervention, and the timing of outcome assessment. However, most trials failed to blind the acupuncturists—as was expected due to the nature of acupuncture interventions. In addition, the majority of the trials did not report patient compliance adequately. As such, the reported treatment effects may be exaggerated in the case that the acupuncturists were subconsciously biased toward the treatment group or the patients in the treatment group may have sought other interventions such as splinting.

Finally, the GRADE framework suggested that the majority of evidence was of low (11/28, 39%) or very low (9/28, 32%) quality and no evidence was rated to be of high quality. The primary reasons for downgrading the quality were imprecision and inconsistency. However, despite having lower quality for most outcomes, the intermediate-term effect for carpal tunnel syndrome on functional status and the majority of the evidence for RA on both pain and function were of moderate quality. Therefore, these findings may be interpreted with greater confidence.

Strengths and Limitations

This study had several strengths.

First of all, similar to previous reviews,^28,36,63 it was decided to stratify the subgroups based on participant conditions, interventions, controls, and timeperiods a priori, as these are major sources of clinical heterogeneity. Stratification would ensure clinical homogeneity and the applicability of the findings under each category.

Second, the majority of the included RCTs were rated to be of low risk of bias based on the assessments, which indicated adequate internal validity for those trials.

Finally, it was decided prospectively to exclude trials comparing acupuncture to active controls. Focusing on comparing acupuncture to inactive controls, sham acupuncture, wait-list, or no treatment enabled the effects of acupuncture to be isolated between groups. This enabled the direct interpretation of the effect of acupuncture without the need to consider the effect of the active control.

However, the current study also had several limitations that must be noted.

First, the sample size for most of the hand-and-wrist conditions was very small and only consisted of 1 trial. Due to the aforementioned reason and methodological heterogeneity in other cases, it was not possible to perform meta-analyses to pool the treatment effects quantitatively. In addition, it was not possible to pool the results of all trials together, disregarding the presence of clinical heterogeneity, as this would have resulted in misleading pooled treatment effects that might not have been applicable in all circumstances.⁶⁴ The inability to meta-analyze the treatment effects was also a limitation in previous systematic reviews that examined the effect of acupuncture on other musculoskeletal conditions.^28,29,36

Second, no outcome was rated to be of high quality according to GRADE and the quality of most outcomes was downgraded due to imprecision and inconsistency. This was mainly due to the lack of power in the included trials, which led to insignificant and inconsistent conclusions.

Finally, it was not possible to obtain enough evidence, due to a lack of reporting, to establish a confident conclusion on the effects of acupuncture on QoL, on the longer follow-up timeperiods, and safety.

Implications

To the current authors' knowledge, this study is the most comprehensive review of acupuncture interventions on hand-and-wrist pain intensity, functional status, and QoL to date. All available RCTs that investigated the effect of acupuncture on any hand-and-wrist related conditions were examined. The results indicated that acupuncture interventions are generally clinically efficacious for hand-and-wrist conditions with regard to pain relief and functional improvement immediately post treatment and in the short-term. Previous clinical trials and systematic reviews on other musculoskeletal conditions found similar benefits post treatment and in the short-term.^{29,36,65–67} Due to the consistent findings on the benefits of acupuncture for musculoskeletal conditions in the shorter terms, acupuncture may be a viable alternative therapy for hand-and-wrist conditions.

However, currently, the present authors caution clinicians who seek to use acupuncture therapies for hand-and-wrist improvements over the intermediate and long-term, as well as for QoL improvements in general, given that sufficient evidence was not obtained to establish confidently that acupuncture alone would result in significant benefits. Therefore, the use of conventional therapies in combination with acupuncture in these situations may be warranted.⁶⁸

Additionally, future research should seek to examine the efficacy of acupuncture in longer follow-up terms, and if its use would result in clinically significant QoL improvements. With regard to the safety of acupuncture, verum and EA were relatively safe in patients who had carpal tunnel syndrome. This finding should be considered with caution as few patients with carpal-tunnel syndrome were included in the 2 trials that reported on this outcome, and that the finding may not be applicable to other conditions.

Despite limited evidence in the current study, however, numerous revies in the past have suggested that adverse effects relating to the use of acupuncture are often transient and serious events are rare.^69–72 Therefore, use of acupuncture for hand-and-wrist conditions is not discouraged by the current authors. Nevertheless, clinicians should monitor any potential adverse effects closely in patients, and future research should continue to examine and report on the safety with regard to acupuncture.

With regard to the methodology of RCTs, few trials were adequate in terms of reporting patient compliance. Researchers should report on patient compliance and attempt to avoid related issues, such as deviation or contamination, that may be a threat to the internal validity of a trial. As a general recommendation, future RCTs on acupuncture should follow proper reporting guidelines, such as the Revised STandards for Reporting Interventions in Clinical Trials of Acupuncture (STRICTA).⁷³ In addition, numerous trials have obtained inconclusive results due to small sample sizes. As such, researchers should estimate the required sample size that would allow sufficient power to reach statistical significance a priori. This may reduce the likelihood of accepting the null hypothesis.⁷⁴

Conclusions

Evidence suggests that acupuncture may be beneficial for hand-and-wrist pain and function for individuals with various conditions post treatment and in the short-term. QoL improvements require further investigation. Limited evidence suggests that acupuncture is relatively safe in patients with carpal tunnel syndrome. Clinicians should be cautious when interpreting these results as few samples were included for most conditions, and clinical heterogeneity may affect the applicability of the findings. Most outcomes were of low or very low quality of evidence according to GRADE. Future research with standardized reporting methodology and sufficient sample size may be warranted to validate the current findings further.

Authors' Contributions

The authors were responsible for the following: Project supervision: Dr. Trinh. Conception and design of the study methodology: Dr. Trinh; Zhou; and Belski. Hand searching relevant reviews: Zhou and Wong. Title and abstract screening: Zhou; Wong; and Belski (arbitration). Full-text retrieval: Zhou and Deng. Full text screening: Zhou; Deng; and Belski (arbitration). Data extraction: Zhou and Belski. Data analysis: Zhou. Risk of bias assessments: Zhou; Belski; and Trinh (arbitration). GRADE assessment: Zhou; Drafting the article: Zhou. Critical editing of the article: Dr. Trinh; Deng; and Wong.

All of authors reviewed and approved the final and submitted version of the article and agreed to be held accountable for all aspects of the work.

Author Disclosure Statement

The lead author, Dr. Kien Trinh, personally funded this review.

Funding Information

This review was personally funded by the lead author, Dr. Kien Trinh.