Ultrasound improves motor distal latency on patients with carpal tunnel syndrome: systematic review and meta-analysis

Abstract

INTRODUCTION

Is the application of ultrasound effective on pain, the severity of the symptoms, physical function, strength, and neurophysiological parameters of the median nerve conduction in patients with carpal tunnel syndrome?

EVIDENCE ACQUISITION

A systematic review and meta-analysis of randomized controlled trials was performed by using a structured search strategy in Scopus, CINAHL, Web of Science and PEDro databases. All the primary studies included samples with carpal tunnel syndrome treated by: ultrasound versus no treatment, therapeutic ultrasound versus sham ultrasound, ultrasound and usual care versus usual care, or ultrasound and other intervention versus the same intervention. The outcomes measures registered were pain, severity of symptoms, function, strength, and neurophysiological parameters (motor distal latency and sensory distal latency) of the median nerve. Methodological quality was evaluated by PEdro Scale.

EVIDENCE SYNTHESIS

Ten clinical trials met the inclusion criteria for the systematic review. Eight trials were meta-analyzed, which included a total of 2069 patients with carpal tunnel syndrome. The methodological quality of the included studies ranged among limited (5 trials), moderate (3 trials), and high (2 trials). In one of the electrophysiological parameters (motor distal latency), a significant difference between groups after the use of ultrasound was observed (MD=-0.10; fixed 95% CI=-0.20, -0.01; P=0.04). No significant differences between groups were observed at post-treatment for pain (P=0.29), severity of symptoms (P=0.99), function (P=0.54), strength (P=0.27) and for the rest of the electrophysiological parameters evaluated (P>0.05).

CONCLUSIONS

The use of ultrasound on patients with carpal tunnel syndrome seems to improve motor distal latency. This finding implies a partial improvement at the neurophysiological level, representing a reduction in the grade of clinical severity. Additional clinical trials with a high methodological quality are needed to investigate the doses at which ultrasound are most effective.

Introduction

Carpal tunnel syndrome (CTS) refers to the set of symptoms and signs associated with the compression of the median nerve at the wrist. This syndrome is the most common peripheral neuropathy. The symptoms of this syndrome are usually hypoesthesia or paresthesia in the median nerve innervation area of the hand, and weakness or paralysis in the abductor pollicis brevis or opponens pollicis. The data about the prevalence and incidence of CTS vary in literature. In workers, the percentage of electrophysiologically confirmed CTS is higher than in the general population. Particularly, the prevalence of CTS among manufacturing and meat-packing workers ranges between 5% and 21%, while the percentage in general populations ranges between 1% and 5%.¹ There are occupational factors that have been associated with this syndrome, including experimenting physical vibrations for working time, mechanical stress on the palm of the hand, and forceful repetitive hand motions.² CTS derived from work is one of the most disabling and costly upper extremity disorders, resulting in lost work days and consequently, represents a major cause of workers compensation cost.³

The diagnosis of CTS is based on the clinical examination by assessing pain, paresthesia, Tinel and Phalen signs, function and weakness.⁴ The diagnosis and the grade of CTS is usually established by a damage of the sensory and motor fibers and, occasionally, vegetative fibers of the median nerve.^5-8 The approaches to assess the median nerve function often involve sensory and motor examination.⁹ The most common sensory parameters in the CTS diagnosis are the sensory distal latency (SDL), sensory conduction velocity and amplitude of the sensory nerve action potential.⁷ In addition, the diagnosis of CTS can be also made by using ultrasound.¹⁰

In terms of sensory techniques of the median nerve, one of the most frequently used is those which explore the fibers that run from the index finger to the wrist and elbow. This exploration is developed antidromically or orthodromically, depending on whether the potential is recorded on the finger (antidromic) or on the wrist (orthodromic) and whether it is stimulated at the finger (orthodromic) or at the wrist (antidromic).^{5, 7, 11} The most important motor parameters in the CTS screening are: 1) compound muscle action potential amplitude; 2) motor distal latency (MDL) of compound muscle action potential; and 3) the motor conduction velocity. The motor conduction velocity is usually measured in the branch for the thenar muscles of the median nerve, recording the potential at the abductor pollicis brevis muscle and stimulating at the wrist and elbow.^{7, 11, 12}

Currently, the usual treatment for this syndrome is surgical; however, this intervention can generate complications, as a high incidence of associated complications has been attributed to this treatment.¹³ Recent publications have shown favorable results from long-term conservative treatment compared to the surgical approach;^{14, 15} nevertheless, some of these techniques have shown heterogeneity in their results.^16-22 A combination of education, night splinting and home exercises appears to reduce the need for surgery in 21% of patients remaining on the waiting list for surgery.²³ Conservative treatment of chronic median nerve compression aims to reduce pressure in the carpal tunnel, improve blood flow through the tunnel, and restore nerve excursion.²⁴ On the other hand, splinting,²³ gliding exercises,²⁵ neurodynamics²⁶ or laser²⁷ have also been shown to be effective conservative therapies. However, ultrasound (US) is the technique most commonly used in rehabilitation area for the treatment of CTS.^28-34

US is a conservative electrophysical agent that requires little time with easy use and with no severe adverse effects reported. Some studies have shown improvements after its application on patients of this syndrome.^{33, 35, 36} There are also other studies that conclude that its use is not effective or question its effectiveness,^{37, 38} and various systematic reviews and meta-analysis have indicated this^{15, 18} for cases other than CTS. For this reasons, it is necessary to perform a meta-analysis to show the existing evidence on the efficacy of the US application on clinically relevant variables such as pain, severity of symptoms, functionality, strength and conductivity of the median nerve in patients with CTS.³⁹ In line, synthesized information is required to confirm the evidence that exists in relation to this therapy. This knowledge will facilitate decision making process in the clinical setting about the application of US in patients with CTS.

Therefore, the main aim of the present systematic review and meta-analysis was to evaluate the efficacy of therapeutic US on pain, severity of symptoms, function and electrophysiological parameters, in comparison with no therapy or sham therapy. The research question is: is the application of US effective on pain, the severity of the symptoms, physical function, strength, and neurophysiological parameters of the median nerve conduction in patients with carpal tunnel syndrome?

Evidence acquisition

Design

This study has followed the Preferred Reporting Items for Systematic Reviews and MetaAnalyses (PRISMA) guidelines. The systematic review and meta-analysis was also performed following general methods for Cochrane reviews, and prospectively registred in PROSPERO database (CRD 42020216366).

Identification and selection of studies

The database search was performed by two independent investigators in the following databases: Scopus, CINAHL, Web of Science, PUBMED and PEDro database. The following MESH terms or keywords were used: Ultrasound, Carpal, Tunnel Syndrome, Physical Therapy. The search strategy used was (ultrasound OR “physical therapy”) AND (carpal OR tunnel OR syndrome). Firstly, we perform a search for meta-analysis, systematic or narrative revisions on the same topic, consulting in Cochrane Plus, Cochrane Library, the Proquest Platform and Google Scholar. There were 4 revisions about physical therapy’s conservative treatment using US on CTS.^{15, 18, 27, 40} One of the reviews compares the effect of US with low level laser therapy.²⁷ Two reviews include various techniques related to physical therapy such as ultrasound, but excluded relevant articles.^{15, 40} Another review performs an extensive systematic review on the specific use of US,¹⁸ but it was published in 2013 and new clinical trials have been conducted since then. Therefore, no review with the same characteristics of the present study was identified. Secondly, the titles and abstracts were independently assessed against the eligibility criteria. Thirdly, two reviewers (APM and JMPM) independently extracted data from the full text of the included studies using a data extraction form (codebook).

The inclusion criteria were:

• design: randomized controlled trials, clinical trials;

• participants: adults with a diagnosis of STC and no other serious comorbid conditions of the hand or wrist;

• intervention: being treated with therapeutic ultrasound;

• outcome measures: pain, severity of symptoms, function, strength, MDL and SDL;

• comparison: non-exposed control group, US versus sham, US and conventional treatment, and other therapy versus conventional treatment or the other therapy alone.

Assessment of characteristics of studies

Methodological quality and risk of bias

The Physiotherapy Evidence Database (PEDro) Scale, an 11-item scale, was used to evaluate and ascertain the methodological quality of the trials, as they are randomized controlled trials (RCTs). This instrument was utilized due to its good validation and reliability indices.^{41, 42} Almost all items (with exception of item 1, which is of relevance to external validity) were awarded 1 point, with the total score ranging from 0 to 10 points. It is important to note that a high score indicates high quality. With the purpose of reducing inter-examiner bias, the studies were evaluated by two independent investigators (APM and MCGR) to ascertain the level of quality. The reviewers then proceeded to settle any discrepancies in the scoring process by consensus.

The PEDro Scale was also used to evaluate the internal validity criteria.⁴³ With the use of this method, a quantitative analysis of the methodological quality of a study can be performed. Internal Validity Score (IVS) was calculated using 7 items related to internal validity (2, 3, and 5 to 9) by adding the positive scores of each of the 7 items together to calculate the IVS. The result of this calculation was then classified into 3 categories: high methodological quality (from 6 to 7 points), moderate quality (from 4 to 5 points), and limited quality (from 0 to 3 points).

The risk of bias was measured by RoB2, the revised version of the Cochrane tool recommended when randomized trials are included.⁴⁴

Quality of evidence

As recommended by Cochrane Handbook for Systematic Reviews of Interventions, the quality of evidence must be measure by using GRADE tool. GRADE approach provides a specific definition of the quality of evidence to summarize the findings of a systematic review.⁴⁴

Participants

The participants included in the original studies have the following characteristics: a) patients with clinical and electrophysiological evidence of mild or moderate idiopathic CTS;^45-49 b) over 18 years old; c) unilateral⁴⁵ or bilateral⁵⁰ wrist involvement.

Intervention

Therapeutic ultrasound.

Outcome measures

The main outcome measures registered for this study are: pain, severity of symptoms, function and median nerve conduction. All of the instruments included in the primary studies have been validated tools for CTS patients. Pain was evaluated by Visual Analogue Scale (VAS).⁵¹ Severity of symptoms was evaluated using the Boston carpal tunnel Questionnaire and the Symptom Severity Subscale (SSS). Function was evaluated by Boston carpal tunnel Questionnaire, Functional Status Subscale (FSS) and Health Assessment Questionnaire.⁵² Hand grip and finger pinch were used for assessing strength. According to the recommendation of the American Association of Electrodiagnostic Medicine guidelines, the electrophysiological assessment of Median Nerve involved parameters used in routinary electroneuromyography studies, such as MDL, motor conduction velocity, amplitude of compound muscle action potential, SDL, sensory conduction velocity and amplitude of sensory nerve action potential.⁷ Finally, the main outcomes studied for median nerve conduction were MDL and SDL.

Data extraction and analysis

Codebook, a form of data extraction, was elaborated to register the effect sizes (Cohen’s d) or essential statistical information required for its calculation, from each study. When data was found to be insufficient in the published trials, the authors were contacted via available means of communication to request the collection of missing data to enable the calculation of the size effects. Pilot testing of the fulfillment of the codebook was implemented on several trials and modified according to the arising need. The reviewers divided their tasks as the first (APM) independently extracted the data while the second (JMPM) confirmed the accuracy of this extracted data. The data extraction from each study was based on outcome data (measurements and timing of assessment), sample characteristics data from each group (therapeutic US group vs control group) such as sample size, age and gender of participants, as well as a description of interventions (type, duration, frequency, setting) and trial results, as shown in Table I. The reviewers revised this data again upon any flagged discrepancies. A two-stage process has been followed to conduct the systematic review and meta-analysis.

Table I. —PEDro Score and IVS of included trials.^{37, 38, 45-50, 53, 54}.

Author (year of publication)	1	2	3	4	5	6	7	8	9	10	11	Total PEDro Score	IVS
Ebenbichler (1998) 48	x	x	x	x	x	x	x			x	x	8/10	5/7 Moderate
Oztas (1998) 65		x			x					x	x	4/10	2/7 Limited
Baysal (2006) 51	x	x	x	x			x			x	x	6/10	3/7 Limited
Dincer (2009) 49	x	x	x	x			x			x	x	6/10	3/7 Limited
Yildiz (2011) 50	x	x	x	x	x		x	x	x	x	x	9/10	6/7 Hight
Duymaz (2012) 47	x	x		x						x	x	4/10	1/7 Limited
Armagan (2014) 46	x	x	x	x	x	x				x	x	7/10	4/7 Moderate
Lazovic (2018) 64	x	x	x	x	x	x				x	x	7/10	4/7 Moderate
Çatalbaş (2018)38		x		x			x	x		x	x	6/10	3/7 Limited
Jothi (2019) 38	x	x	x	x	x	x	x	x		x	x	9/10	6/7 Hight

X: yes; PEDro: Physiotherapy Evidence Database; IVS: Internal Validity Score.

In the first stage, a summary statistic has been calculated for each study, describing the observed intervention effect. This summary was performed for the mean differences (Cohen’s d) from pre-post intervention and pre-follow-up. Cohen’s d was calculated when the studies included descriptive data (sample size, pre and post-treatment and from pre-follow-up mean and standard deviation values) in the intervention group which was compared to the control group for the main groups of variables.⁵⁵ When insufficient information was obtained from primary sources, a conversion of the available data was performed to find Cohen’s d.

In the second stage, a summary (pooled) intervention effect estimate was calculated as a weighted average of the intervention effects estimated in the individual studies, with the use of Revman 5.4 software.⁵⁵ A fixed-effect model was used with the inverse variance method by including the means and standard deviations at post-intervention. The meta-analyses for the different groups of outcomes were performed by the calculation of mean differences (MD) for pain, symptom severity, motor distal latency and sensory distal latency, and the standardized mean differences (SMD) for function, and 95% confidence intervals. The strength of the SMD was determined using Cohen’s d criteria.⁵⁶ Statistical heterogeneity was assessed using the I² statistic. This heterogeneity responds to the variability in the intervention effects being evaluated in the different studies. The most common data collection time was chosen where outcome data were measured at different time points between the trials.

Evidence synthesis

Flow of studies through the review

The present review includes the published clinical trials evaluating the effectiveness of US on CTS, with the variables of pain intensity, severity of symptoms, function, strength, MDL and SDL. Two reviewers applied the inclusion and exclusion criteria. The initial search, which was conducted by analyzing the title, identified 5509 potentially relevant papers. After eliminating duplicate papers and reading the abstracts of the relevant papers, 371 trials were obtained. From these results, 28 items were revised, of which 10 items met the inclusion criteria to be included in the present systematic review. The flow chart used to select the primary studies, according to the PRISMA guideline, is shown in Figure 1.⁵⁷

Figure 1.

—PRISMA flow diagram. From: Moher et al.⁵⁷

Two studies were excluded because they were published in a language other than those established in the inclusion criteria,^{58, 59} while two others were excluded due to the lack of full text.^{60, 61} Despite the use of different strategies such as contacting the authors, the full texts could not be retrieved.^{47, 48} A number of other studies were discarded because they were superiority studies^{28, 29, 31-33, 35, 36, 62, 63} or they included drug therapy.^{20, 49, 64}

In the meta-analysis of the present study, 8 clinical trials were finally included, with a total sample of 2069 wrists with CTS. These clinical trials included data on pain, severity of symptoms, function, strength evaluation, and the neurophysiological parameters of MDL and SD. A total sample of 259 wrists included a pain evaluation, 307 wrists the severity of symptoms, 305 wrists function, 172 wrists strength, 373the MDL and 305 SDL.

Characteristics of studies

Methodological quality and risk of bias

Table I^{37, 38, 45-50, 53, 54} provides details about the PEDro Scale of each trial and IVS scores. It can be seen that these studies^{37, 49} scored 9/10 on the PEDro Scale. The study⁴⁷ received a score of 8/10, like the studies of^{45, 53} which received a score of 7/10, while the studies of^{38, 48, 50} a score of 6/10 and the studies of^{46, 54} received a score of 4/10. Five studies showed limited methodological quality, three others showed moderate quality, while two of the included studies showed high quality. The risk of bias is shown in Figure 2, 3.^{37, 38, 45-50, 53, 54}

Figure 2.

—Risk of bias graph: review authors’ judgements about each risk of bias item presented as percentages across all included studies.

Figure 3.

—Risk of bias summary: review authors’ judgements about each risk of bias item for each included study.^{37, 38, 45-50, 53, 54}

Participants

The characteristics of the clinical trials included in the systematic review and meta-analysis are in Supplementary Digital Material 1, Supplementary Table I.^{37, 38, 45-50, 53, 54}

Interventions

Amongst the 8 meta-analyzed studies, only one looked at therapeutic US versus placebo,⁴⁷ while the other 7 looked at conventional therapy alone versus the same conventional therapy but with therapeutic US.^{37, 38, 45, 48-50, 53, 54} However, two other articles that met the inclusion criteria could not be used for the meta-analysis due to lack of data.^{46, 47} This information is detailed in Supplementary Table I and Supplementary Digital Material 2, Supplementary Table II.^{37, 38, 45-50, 53, 54}

Outcome measures

The outcome measures used for each primary study are listed in Supplementary Table I.^{37, 38, 45-50, 53, 54}

Effect of US on CTS

Effect of US on pain

There were no differences between the therapeutic US group and the control group for pain (MD=-0.30; fixed 95% CI=-0.86, 0.25; P=0.29) (Figure 4).^{38, 45, 49, 50, 54} For the effect of the US in the short term, the magnitude of the effect varied between 0.835 and 1.878. In relation to the pre-post-treatment differences of the therapeutic US group, an immediate effect size using Cohen’s d varied in a range from 0.108 to 1.625.

Figure 4.

—Forest plot pain FE. Forest plot and meta-analysis of ultrasound treatment compared with control treatment for the outcome pain intensity.^{38, 45, 49, 50, 54}

Effect of US on severity of symptoms

With respect to the severity of symptoms registered by the SSS, no differences were observed between the therapeutic US group and the control group (MD=-0.00; fixed 95% CI= -0.21, 0.21; P=0.99) (Figure 5).^{37, 38, 45, 49, 50}

Figure 5.

—Forest Plot SSS FE. Forest plot and meta-analysis of ultrasound therapy compared with control treatment illustrating the improvement in self-reported severity of symptoms.^{37, 38, 45, 49, 50}

While in regards to the individual improvement of the therapeutic US group, the studies immediately reported an effect size, with a Cohen’s d, that varies in a range from 0.437 to 2.797. For short-term improvement, an effect size between 0.823 and 2.116 was found. One of the studies had an effect size of 2.307 in relation to the individual improvement of the therapeutic US group in the long term (Supplementary Table II).^{37, 38, 45-50, 53, 54}

Effect of US on function

There were no observed differences between the therapeutic US group and control group in regards to function registered by the FSS (SMD=0.07; fixed 95% CI=-0.16, 0.31; P=0.54) (Figure 6).^{38, 45, 46, 48-50}

Figure 6.

—Forest Plot FSS FE. Forest plot and meta-analysis of ultrasound therapy compared with control treatment illustrating the improvement in self-reported function.^{38, 45, 46, 48-50}

An immediate effect size that varied in a range from 0.219 to 1.563 was derived in relation to the individual improvement of the therapeutic US group, in the short term, one study had an effect size that varied in a range between 0.110 and 1.419 (Supplementary Table II).^{37, 38, 45-50, 53, 54}

Effect of US on strength

Regarding the strength variant, no differences were observed between the therapeutic US group and control group (MD=0.62; fixed 95% CI=-0.48, 1.72; P=0.27) (Figure 7).^{38, 50}

Figure 7.

—Forest plot strength FE. Forest plot and meta-analysis of ultrasound therapy compared with control treatment illustrating the improvement in self-reported pinch and grip strength.^{38, 50}

An immediate effect size for the independent improvement of the therapeutic US group, varying in a range from 0.187 to 0.392, was obtained using Cohen’s d. The effect of US in the short term had a magnitude of the effect that varied between 0.302 and 0.759 (Supplementary Table II).^{37, 38, 45-50, 53, 54}

Effect of US on MDL

In the case of MDL, a significant difference was observed between the therapeutic US group and the control group, this being in favor of the therapeutic US group (MD=-0.10; fixed 95% CI=-0.20, 0.01; P=0.04) (Figure 8).^{38, 45, 48-50, 53, 54}

Figure 8.

—Forest Plot MDL FE. Forest plot and meta-analysis of ultrasound treatment compared with control treatment for the outcome motor distal latency.^{38, 45, 48-50, 53, 54}

The therapeutic US group effect size varies between 0.000 y 0.246 y and for a short term between 0.142 and 0.270 (Supplementary Table II).^{37, 38, 45-50, 53, 54}

Effect of US on SDL

In respect to the SDL, the therapeutic US and control groups do not show statistically significant differences in post-treatment improvement (MD=0.02; fixed 95% CI=-0.10, 0.13); P=0.77) (Figure 9).^{38, 45, 49, 50, 54}

Figure 9.

—Forest Plot SDL FE. Forest plot and meta-analysis of ultrasound treatment compared with control treatment for the outcome sensory distal latency.^{38, 45, 49, 50, 54}

Specifically, the therapeutic US group immediately obtained a Cohen’s d ranging between 0 and 0.654; in the short term, and the effect size ranged between 0.362 and 0.711 (Supplementary Table II).^{37, 38, 45-50, 53, 54}

Discussion

The main purpose of this systematic review and meta-analysis has been to evaluate the available scientific evidence about the effectiveness of the use of US in patients with CTS. The studies involved in this meta-analysis showed a methodological quality that varied between limited and high. The clinical relevance of the present review is based on the changes recorded for motor distal latency after the application of the therapeutic US. These changes suggest a partial improvement and a reduction of the degree of neurophysiological severity of the CTS. All the randomized controlled trials had a moderate-high methodological quality and the sample populations from the primary studies were similar within the systematic review (idiopathic CTS). The quality of evidence for MDL and SDL is moderate due to the indirectness of evidence in terms of differences in interventions (applicability).⁴⁴ Interventions may have been delivered differently since the calibration in the application of the therapeutic US is unknown. On the other hand, the application of the neurophysiological measures in some primary studies was unrecorded; therefore the implementation of this outcome measure could be heterogeneous. Similarly, the quality of evidence for pain, severity of symptoms, function, and strength is moderate due to possible imprecision related to the sample size of the studies.⁴⁴

In relation to the differences between groups in the studies which were meta-analyzed, the therapeutic US group showed a lower Median Nerve Motor Distal Latency after the application of US compared to a control group. The difference between both groups had a mean effect size of 0.09 milliseconds. The immediate improvement recorded after the application of US was of an effect size that varied between negligible and null. However, in the short term, an improvement that ranged from negligible to small was observed. Taking into account the different neurophysiological severity scales of CTS, the interpretation of a control neurophysiological study in which a decrease in MDL values is detected within normality in the set parameters (in comparison to a previous pathological study but with persistence of slowing down in the sensitive studies) probably generating a partial improvement of the degree of severity of the CTS. This interpretation is due to the fact that the different scales, although with disparity in degrees, classify with less severity only sensory slowdowns of median nerve conduction. They also show, with mixed intermediate severity, sensitive and motor slowdowns, and with greater severity, the absence of action potential, especially motors.^{9, 65-68} On the other hand, the effect of US on musculoskeletal structures can vary depending on a multitude of factors such as if the pathology is chronic or acute, the depth of the tissues, the acoustic windows, tissue structure, among many other factors.⁶⁹

Regarding the rest of the electrophysiological parameters recorded in the primary studies, the difference found for the SDL and the sensory conduction speed (index finger-wrist) between the groups was not significant. Although this finding could be considered a discrepancy with that found for motor nerve conduction of the median nerve, it can be explained by the following reasons: 1) there is a relationship between extreme values of MDL and the absence of the sensory nerve action potential, but this relationship is often seen in severe degrees of CTS; 2) this relationship can be altered by rare anatomical defects that affect the recurrent branch of the median nerve of the thenar muscles, with the exclusive involvement of sensory conduction being more common as an early and less severe finding.⁷⁰ The improvement of these parameters can occur after more than a year post injury, so a longer duration of the treatments of the meta-analyzed studies could realize significant findings also in sensory conduction.^{11, 12, 65, 67, 70}

No significant differences were observed between groups for pain, severity of symptoms, function, or strength. It should be taken into account that the primary symptoms such as numbness, tingling and nocturnal symptoms are generally considered more specific to the nerve injury, and therefore tend to have a greater correlation with the electrophysiological parameters. While the secondary symptoms such as pain, weakness or clumsiness are typical associated with soft tissue and other musculoskeletal injuries.^{11, 12, 71, 72}

Specifically, in relation to pain levels, the therapeutic US group did not show a significant improvement greater than the control group, after the application of US. Their effect size varied from null and large. However, in the short term, a long improvement was observed for both groups. The results found may be due to the fact that biologically there is an increase in cellular production or their metabolism is modulated,⁷³ tissue regeneration,⁷⁴ or it could be due to how they inhibit inflammation.⁷⁵ Hence, it is possible that these changes at the molecular and cellular level reduce pain and inflammation and contribute to the structural improvement of damaged tissues.

In relation to the severity of the symptoms and the degree of function of the upper limb, the therapeutic US group did not show a significant difference higher than the control group. However, for severity of symptoms in the therapeutic US group, a small-large improvement was observed immediately and a large improvement in the short term.

With regard to function, a similar improvement was observed after the intervention, and a negligible improvement in the long term and short term. These results can be explained by the same physiological mechanisms on which US acts. As there are lower levels of pain, it is likely that there is in turn a decrease in the symptoms that accompany this pathology and patients are able to better perform normal activities.⁷⁶

The grip and pinch strength sample showed no significant difference between the therapeutic US and control groups after the intervention. For the improvement recorded in the therapeutic US group, the effect ranged from negligible to small immediately, then from negligible to moderate, in the short term.

Very low heterogeneity has been observed in the research methods, reaching a value of 0% for some variables studied. No adverse effects have been observed with this technique as in no case have serious effects been reported with the application of US. The methods and variables recorded in the primary studies have been evaluated using validated and reliable instruments.

In regards to the main critical points, several aspects have been identified that should be taken into account when interpreting the evidence found. First of all, it was not possible to include all the studies that rigorously analyzed the effect of US due to language^{58, 59} or as a result of incomplete text.^{60, 61} Two studies were excluded due to lack of data in the full text^{47, 48} or because the articles were not accessible in full text. In order to maintain the greatest possible stringency in this synthetic review and meta-analysis, we only included US as therapy. We have not included clinical trials in which other types of therapies were associated, nor have superiority trials been taken into account in which the action of US versus another therapeutic technique was evaluated. More recent studies that had not been taken into account in previous systematic reviews have been included in this review, some of them with high methodological quality, and the results obtained have been meta-analyzed. On that note, the present review also aims to serve as a relevant contribution to different areas such as clinical, research or even political. All articles related to sonophoresis have been excluded, although US is used for the application of this technique, as a pharmacological compound is also associated. It must also be considered that the long-term effects (at least 12 months after the end of the intervention) have not been evaluated in most of the included studies. Furthermore, other electrophysiological parameters could not be used, due to lack of data to compare them (motor conduction speed or the motor and sensory amplitude of potentials), as well as, due to methodological differences in the performance of the tests that made it impossible to homogenize the data without committing biases. An example is assessing the sensory conduction velocity of the Median nerve in the palm-wrist stretch, through which sensory and motor fibers pass. This being an assessment of mixed nerve conduction and not purely sensitive.^{12, 77}

As regards to the implications for practice, the included studies reflect the parameters of intensity, duration and frequency of US that make it possible to easily replicate it in clinical practice. However, given the data obtained, it is necessary to delve into the dose of its application, either the frequency, the intensity or the time. Also, the cost-benefit of these therapies should be evaluated in future studies in order to know if it is profitable to perform them in clinical practice. On the other hand, it would be interesting to carry out superiority studies where the mixed effect of US is evaluated along with other emerging therapies for pain control, such as the treatment of chronic pain from neuropedagogy, cognitive behavioral therapies, motor graded imagery or including desensitization maneuvers of the central nervous system.¹⁴ Other neurophysiological parameters that have not been developed in the meta-analyzed studies could reinforce the finding of the reduction in MDL. A needle EMG examination of the APB muscle is recommended and the number of motor units should also be estimated.^{78, 79} On the other hand, in the meta-analyzed studies, the somatic innervation of the median nerve (motor and sensory) has been recorded, but the autonomic efferents and afferents responsible for the innervation of the sweat glands and for collecting pain or temperature has not been included. These fibers are more susceptible to anoxia than compression and have also been the subject of neurophysiological study in CTS using nerve conduction techniques such as the sympathetic cutaneous response, or evoked potential techniques, in this case, by laser.^{8, 80, 81}

Conclusions

In conclusion, US has shown a superior post-treatment improvement in motor distal latency compared to the control groups used in the primary studies carried out so far. This could lead to a reduction in the degree of severity in the neurophysiological scales of the CTS. However, the application of US does not appear to have a therapeutic effect on pain, severity of symptoms, strength, function, sensory conduction velocity and sensory distal latency.

Supplementary Digital Material 1

Supplementary Table I

Characteristics of the clinical trials incluidos en la revision sistemática y meta-analysis.^{37, 38, 45-50, 53, 54}

Supplementary Digital Material 2

Supplementary Table II

Pre, postintervention, follow-up values and size effects for pain, function and strength.^{37, 38, 45-50, 53, 54}