Local corticosteroid injection versus placebo for carpal tunnel syndrome

Carpal tunnel syndrome (CTS) is a very common clinical syndrome manifested by signs and symptoms of irritation of the median nerve at the carpal tunnel in the wrist. Direct and indirect costs of CTS are substantial, with estimated costs of two billion US dollars for CTS surgery in the USA in 1995 alone. Local corticosteroid injection has been used as a non-surgical treatment for CTS many years, but its effectiveness is still debated.

To evaluate the benefits and harms of corticosteroids injected in or around the carpal tunnel for the treatment of carpal tunnel syndrome compared to no treatment or a placebo injection.

We used standard, extensive Cochrane search Methods. The searches were 7 June 2020 and 26 May 2022.

We included randomised controlled trials (RCTs) or quasi-randomised trials of adults with CTS that included at least one comparison group of local injection of corticosteroid (LCI) into the wrist and one group that received a placebo or no treatment.

We used standard Cochrane methods. Our primary outcome was 1. improvement in symptoms at up to three months of follow-up. Our secondary outcomes were 2. functional improvement, 3. improvement in symptoms at greater than three months of follow-up, 4. improvement in neurophysiological parameters, 5. improvement in imaging parameters, 6. requirement for carpal tunnel surgery, 7. improvement in quality of life and 8.

We used GRADE to assess the certainty of evidence for each outcome.

We included 14 trials with 994 participants/hands with CTS. Only nine studies (639 participants/hands) had useable data quantitatively and in general, these studies were at low risk of bias except for one quite high-risk study. The trials were conducted in hospital-based clinics across North America, Europe, Asia and the Middle East. All trials used participant-reported outcome measures for symptoms, function and quality of life. There is probably an improvement in symptoms measured at up to three months of follow-up favouring LCI (standardised mean difference (SMD) -0.77, 95% confidence interval (CI) -0.94 to -0.59; 8 RCTs, 579 participants; moderate-certainty evidence). Up to six months this was still evident favouring LCI (SMD -0.58, 95% CI -0.89 to -0.28; 4 RCTs, 234 participants/hands; moderate-certainty evidence). There is probably an improvement in function measured at up to three months favouring LCI (SMD -0.62, 95% CI -0.87 to -0.38; 7 RCTs, 499 participants; moderate-certainty evidence). We are uncertain if there is a difference in median nerve DML at up to three months of follow-up (mean difference (MD) -0.37 ms, 95% CI -0.75 to 0.02; 6 RCTs, 359 participants/hands; very low-certainty evidence). The requirement for surgery probably reduces slightly in the LCI group at one year (risk ratio 0.84, 95% CI 0.72 to 0.98; 1 RCT, 111 participants, moderate-certainty evidence). Quality of life, measured at up to three months of follow-up using the Short-Form 6 Dimensions questionnaire (scale from 0.29 to 1.0; higher is better) probably improved slightly in the LCI group (MD 0.07, 95% CI 0.02 to 0.12; 1 RCT, 111 participants; moderate-certainty evidence). Adverse events were uncommon (low-certainty evidence). One study reported 2/364 injections resulted in severe pain which resolved over "several weeks" and 1/364 injections caused a "sympathetic reaction" with a cool, pale hand that completely resolved in 20 minutes. One study (111 participants) reported no serious adverse events, but 65% of LCI-injected and 16% of the placebo-injected participants experienced mild-to-moderate pain lasting less than two weeks. About 9% of participants experienced localised swelling lasting less than two weeks. Four studies (229 participants) reported that they experienced no adverse events in their studies. Three studies (220 participants) did not specifically report adverse events.

Local corticosteroid injection is effective for the treatment of mild and moderate CTS with benefits lasting up to six months and a reduced need for surgery up to 12 months. Where serious adverse events were reported, they were rare.