Chin Yuan Hui Andrew MBBS, FRCSEd, Lim Keng Hua MBBS, Poh Beow Kiong MBBS, MMed (Surg), Koh Dennis MBBS
Department of Hand Surgery, SGH

* Presented at the International Conference on Evidence-based Medicine, Singapore on 18 to 19 October 2003.

Background. Although carpal tunnel syndrome (CTS) is the most common entrapment neuropathy, there is still no consensus on its treatment. While splinting is still a popular treatment for patients with mild symptoms, surgery is often preferred in very symptomatic cases. The objective of this systematic review is to compare the efficacy of surgery with that of splinting in the treatment of CTS.

Methods. Four independent reviewers searched the electronic databases, MEDLINE, EMBASE, OVID and the Cochrane Neuromuscular Disease Group Registry for relevant prospective randomised controlled trials. The quality and eligibility of the trials were assessed independently and data were extracted. Wherever possible, data were pooled and analysed. Only 2 studies fulfilled our criteria for inclusion. Both studies were not blinded. The patient population comprised patients with carpal tunnel syndrome who are randomised to either surgery or splinting arms. One was a single centre study comprising 22 female patients, whereas the other was a multi-centre trial with a total of 176 patients.

Results. When independently analysed, the results, in terms of clinical improvement in both studies, favoured surgery as the treatment of choice for CTS. In each study, the results were statistically significant. However, when results of both studies were pooled, the differences in efficacy were not statistically significant.

Conclusion. The outcome of this review is inconclusive. More well-designed randomised controlled trials comparing surgery and splinting are required in order to draw meaningful conclusions.

Keywords: carpal tunnel syndrome, randomised controlled trial, splinting, surgery

Carpal tunnel syndrome (CTS) is the common entrapment neuropathy encountered in clinical practice.^1,2 It has a prevalence of 9.2% in females and 0.6% in males.³ Its symptoms include tingling, numbness, loss of sensation in the hand, weakness and loss of muscle function of the thumb. The carpal tunnel is a fibro-osseous confined space formed by the roof, which is that of the transverse carpal ligament and the floor comprising the carpal bones. CTS is due to the compression or entrapment of the median nerve where it passes under the transverse ligament in the wrist. This is often due to the presence of thickened tendon sheath or a space-occupying lesion within the tunnel. As a result, there is a pressure build-up in the carpal tunnel, resulting in compromised median nerve function.

CTS is a clinical diagnosis. There is currently no universally accepted clinical and laboratory diagnostic criteria. Neither is there any universally accepted treatment of choice. Treatments range from conservative management like splinting to open surgical procedure to decompress the tunnel. In very symptomatic patients, surgery is widely preferred over conservative therapy.^4,5 Splinting remains one of the most common conservative treatment modalities. Other conservative treatment options include corticosteroid injections, Vitamin B6, diuretics, anti-inflammatory drugs and acupuncture.⁶

The objective of this paper was to compare the efficacy of surgery with splinting in the treatment of CTS.

Subjects diagnosed with CTS irrespective of the diagnostic criteria used, aetiology of the syndrome, associated pathology, gender and age.

Patients who had undergone surgical release (open method) or splinting as the only form of conservative treatment.

Primary outcome measured was clinical improvement at 3, 6 and 12 months of follow-up. Clinical improvement was defined as complete or significant relief of symptoms, by at least 50% of the baseline.⁷ The other outcome measure studied was that of requirement of secondary surgery after primary treatment was instituted i.e. either surgery or splinting.

Only randomised controlled trials comparing surgery with splinting were included.

A search was carried out using the MEDLINE, OVID, EMBASE (all up to September 2003) and The Cochrane Neuromuscular Disease Group Trials register for randomised controlled trials. The keywords used were "Carpal Tunnel Syndrome", "median nerve entrapment", and "splinting" and "surgery". Bibliographies and references in all relevant papers were also searched. Using the above-mentioned keywords, 40 articles were selected for further review. Based on the selection criteria, 2 randomised controlled trials were found and appraised in detail. This process was done by 4 independent reviewers.

Two randomised controlled trials were selected for detailed appraisal (Table 1). One single centre study (Garland 1964) comprised of 22 females between 35 to 63 of age with symptoms of CTS ranging from 1 month to 20 years, and distal motor latency of the median nerve greater than 4.5ms.⁸ The patients were randomised equally to both surgery and splinting by a secretary using a random list. The patients in the surgical group underwent open carpal tunnel release while the other group had splinting to their hand, wrist and arm with plaster-of-Paris for 1 month. Results of clinical and electrophysiological studies were reported at 1 year follow-up.

Table 1. Characteristics of included trials.

The second study is a multi-centre trial involving 176 patients (143 females, 33 males) from 13 hospitals, who were more than 18 years of age, with clinical and electrophysiological confirmation of idiopathic CTS.⁹ Patients with secondary causes of CTS like diabetes mellitus or previous distal wrist fracture were excluded. Randomisation was performed using a random number table. Eighty-seven patients were allocated to the surgical group while 89 patients were allocated to the splinting group. Fourteen patients in the surgical group did not receive surgery as assigned, while 13 in the splinting group did not receive splinting according to protocol as assigned. Results of clinical improvement at 3, 6, 12 and 18 months follow-up were published. Electrophysiological studies were performed at 12 months follow-up.

Both papers were randomised but no blinding was performed. Allocation concealment was performed in the Gerritsen study but was not mentioned in the Garland study. Both studies published their intention-to-treat (ITT) analysis.

Statistical analysis was performed using the Review Manager 4.2.1 (RevMan) statistical package.¹⁰ Proportional and absolute risk reductions were calculated for each outcome. Test for heterogeneity was performed using standard chi-square test. Due to the high level of heterogeneity between the 2 studies, relative risks were calculated based on the random effects model.

In Gerritsen study, clinical improvement was considered at 3, 6 and 12 months (Fig. 1). Sixty-two (71%) out of 87 patients in the surgery group and 42 (51.6%) out of 89 patients in the splinting group had clinical success at 3 months follow-up. The results favoured surgery as a method of treatment (with a relative risk of 1.38 and confidence interval of 1.08 to 1.75). In the same trial, 72 (82.7%) out of 87 patients in the surgery group and 57 (64%) out of 89 patients in the splinting group had clinical success at 6 months follow-up. These results, too, favoured surgery (with a relative risk of 1.29 and confidence interval of 1.08 to 1.55).

Fig. 1. Forest plot of outcome variable — clinical improvement.

In Garland study, all 11 patients randomised to surgery were completely relieved from symptoms of CTS at 1 year follow-up (Fig. 1). However, only 2 patients in the splinting group had improvement at 1 year follow-up. This result strongly favoured surgery with a relative risk of 6.81 (confidence interval of 1.57 to 19.27). In Gerritsen’s study, 67 (77%) out of 87 patients in the surgery group and 60 (67.4%) out of 89 patients in the splinting group had clinical success at 1 year follow-up. The results also favoured surgery as a method of treatment (relative risk of 1.14 and confidence interval of 0.95 to 1.37). When results from both studies were pooled for meta-analysis, surgery remained as the preferred treatment option (relative risk of 2.24 and confidence interval of 0.44 to 11.39).

A test for heterogeneity was performed using chi square which showed a significant level of heterogeneity with a chi square value of 6.72 and df=1 (P=0.01).

ITT analysis was also done for both studies (Fig. 2). In Gerritsen’s trial, 62 (80%) out of 78 patients in the surgery group and 46 (54%) out of 86 patients in the splinting group had clinical success at 3 months follow-up. The results favoured surgery as a method of treatment (relative risk of 1.49 and confidence interval of 1.18 to 1.86). At 6 months follow-up, 72 (94%) out of 77 patients in the surgery group and 57 (68%) out of 84 patients in the splinting group had clinical success. These results also favoured surgery (relative risk of 1.38 and confidence interval of 1.18 to 1.61). At 1 year follow-up, Gerritsen’s study showed that 67 (92%) out of 73 patients in the surgery group and 60 (72%) out of 83 patients in the splinting group had clinical success. These results, too, favoured surgery as a method of treatment (with a relative risk of 1.27 and confidence interval of 1.09 to 1.47). When results from both studies were pooled for meta-analysis, surgery remained as the preferred treatment option (relative risk of 2.37 and confidence interval of 0.48 to 11.66). Test for heterogeneity again showed that there was a significant level of heterogeneity with a chi square value of 6.52 and df=1 (P=0.01).

Fig. 2. Forest plot of outcome variable — clinical improvement (Intention to Treat analysis).

The 2 studies were analysed in relation to other outcome measures, i.e. the need to undergo secondary surgery after primary treatment (Fig. 3). In the Garland trial, 8 out of 11 patients in the splinting group required surgery during their follow-up whereas none in the surgical arm needed a repeat operation. The result favoured the surgical arm (relative risk of 0.06, 95% confidence interval of 0.00 to 0.91). In the other series by Gerritsen, 1 out of 87 patients in the surgical group underwent repeat surgery while 35 out of 89 patients in the splinting group required surgery (relative risk of 0.003, 95% confidence interval of 0.00 to 0.21). The data from the 2 studies were pooled and the results indicated that a significant number of patients in the splinting group required surgery whereas the risk of repeat surgery in the surgical arm was low (relative risk of 0.04, 95% confidence interval of 0.41 to 0.98).

Fig. 3. Forest plot comparing the need for secondary surgery after primary treatment between the surgery and splinting groups.

The 2 trials show that there is a favourable outcome in patients with CTS undergoing surgery as compared to splinting. When analysed separately and independently, the difference was statistically significant. However, when the data were pooled (as in the clinical outcome at 1 year follow-up), the results were no longer statistically significant. This was attributed to the fact that the heterogeneity between the 2 studies were statistically significant. The Garland trial has a relatively small sample size compared to Gerritsen trial. Since there were no baseline clinical and electrophysiological data in the Garland study, similarities in patient profile between the 2 different arms could not be ascertained. There is likelihood that there were more patients with severe symptoms found in one treatment arm compared to the other. Even within the study, there was also heterogeneity between the patients selected e.g. symptoms ranging from as short as 1 month to as long as 20 years. Although a review by Verdugo indicated that after pooling the data from the 2 studies, clinical improvement at 1 year follow-up favoured surgery; we noted that the fixed effects model was used instead of the random effects model that was applied in our review.¹¹ This resulted in the difference between the 2 reviews. We feel that the random effects model should be applied in the review because of the heterogeneity of the studies.

Significant clinical improvement at 3 and 6 months follow-up was reported in the study with better methodological quality (Gerritsen). Although both treatment arms reported good clinical outcome and success rate, a large number of patients in the splinting group had to undergo surgery during the follow-up period.

Detection bias cannot be ruled out in both studies as both were not blinded. Since the randomisation process was unclear in the Garland study, selection bias is highly probable.

Based on the 2 studies, surgery seems to be more efficacious compared to splinting in the treatment of CTS. But analysis of the results is still not conclusive. There is still a need for well designed clinical trials to provide a more conclusive answer to the efficacy of surgery in CTS. There should also be stratification of patients with CTS according to the severity, duration of symptoms, occupation and age before analysis so as to provide a more meaningful review and conclusion.

The authors would like to thank Dr Miny Samuel, Clinical Trials and Epidemiology Research Unit, Singapore for her kind assistance in statistical analysis and advice.

1. Martyn CN, Hughes RAH. Epidemiology of peripheral neuropathy. J Neurol Neurosurg Psychiatry 1997; 62:310-8.
2. Steward JD. Compression and entrapment neuropathies. In: Dyck PJ, Thomas PK, Griffin JW, Low PA, Poduslo JF, editors. Peripheral Neuropathy, 3rd Edition, Vol 2. Philadelphia: Saunders, 1993;961-79.
3. de Krom MC, Knipschild PG, Kester AD, Thijs CT, Boekkooi PF, Spaan F. Carpal tunnel syndrome : prevalence in the general population. J Clin Epidemiol 1992; 45:373-6.
4. Gerritsen AA, Uitdehaag BM, van Geldere D, Scholten RJ, de Vet HC, Bouter LM. Systematic review of randomized clinical trials of surgical treatment for carpal tunnel syndrome. Br J Surg 2001; 88:1285-95.
5. American Academy of Neurology. Practice parameter for carpal tunnel syndrome (summary statement). Neurology 1993; 43:2406-9.
6. Shaw Wilgis EF. Treatment options for Carpal Tunnel Syndrome. JAMA. 2002; 288:1281-2.
7. Verdugo RJ, Ochoa JL. "Sympathetically maintained pain" I. Phentolamine block questions the concept. Neurology 1994; 44:1003-10.
8. Garland H, Langworth EP, Tarverner D, Clark JM. Surgical treatment for the carpal tunnel syndrome. Lancet 1964; 1:1129-30.
9. Gerritsen AA, de Vet HC, Scholten RJ, Bertelsmann FW, de Krom MC, Bouter LM. Splinting vs surgery in the treatment of carpal tunnel syndrome: a randomized controlled trial. JAMA 2002; 288:1245-51.
10. Review manager (RevMan) [computer program]. Version 4.2.1 for Windows. Oxford, England: The Cochrane Collaboration, 2003.
11. Verdugo RJ, Salinas RS, Castillo J, Cea JG. Surgical versus non-surgical treatment for carpal tunnel syndrome. The Cochrane Database of Systematic Reviews, The Cochrane Library 2003; 3:1-14.