Carpal tunnel syndrome is a common cause of hand pain, sensory disturbance, and weakness affecting daily activities and quality of life. It is a frequent reason for medical consultation and up to 40% of patients are managed exclusively in primary care.
Treatment goals are to relieve symptoms, improve function, and prevent disease progression to nerve damage. The two treatments that constitute standard care for most patients with first-time carpal tunnel syndrome are night splinting and local steroid injections. In a large-sample, US national database study (2009–13), 71% of patients with carpal tunnel syndrome were treated with immediate surgery without previous steroid injection. Evidence regarding effectiveness of night splinting in carpal tunnel syndrome is weak, optimal treatment duration is unclear, and the benefit is generally small.
Efficacy of night splinting has not been assessed in placebo-controlled trials, and efficacy of local steroid injection is supported by strong evidence of efficacy 8 weeks after injection. However, the long-term benefit is still being debated.
In The Lancet, Linda S Chesterton and colleagues report the results of a randomised trial done in 25 UK primary care centres, which compared a single steroid injection (20 mg methylprednisolone acetate) with 6 weeks' night splinting in patients with mild or moderate carpal tunnel syndrome. Among the 212 patients who completed the questionnaire, the overall Boston Carpal Tunnel Questionnaire score (combined symptom severity and functional status scores, 1–5 scale) at 6 weeks was significantly better in the injection group than in the night splint group by an adjusted mean difference of −0·32 (95% CI −0·48 to −0·16), corresponding to an effect size of 0·41. At 6 weeks, mean symptom severity score improvement was 0·84 in the injection and 0·48 in the splint group. Between-group differences did not persist at 6 months. It is important to highlight the participants' characteristics: 199 (85%) had a first-time carpal tunnel syndrome diagnosis and all had symptoms for at least 6 weeks.
The clinical diagnosis of mild or moderate carpal tunnel syndrome was made by different examiners (general practitioners, physiotherapists, or occupational therapists) at 25 centres. However, nerve conduction tests were not used, which is a limitation because they reflect diagnostic accuracy and disease severity. Nonetheless, the trial is pragmatic and treatment is often given on the basis of clinical diagnosis in practice.
Severe carpal tunnel syndrome (constant numbness, sensory loss, and thenar atrophy) was excluded and self-reported symptoms were measured with the symptom severity scale, a standard scale completed by all patients. Notably, mean symptom severity score before treatment was greater than 2·9, a value similar to preoperative scores in UK studies.
Thus, although the trial was done in primary care, the patients are largely similar to those in hospital setting in terms of self-reported symptom severity and duration, making the results relevant beyond primary care.
The statistically significant score difference of −0·32 in favour of steroid injection could be argued to be of small clinical importance. In a randomised trial of steroid injection for carpal tunnel syndrome, mean symptom severity score improved after placebo injection by 0·30 at 10 weeks.
Injections are believed to have a larger placebo effect than non-invasive therapies. In Chesterton and colleagues' study, the only score improvements that are likely to be considered clinically important were seen in patients allocated to their preferred intervention and patients who preferred injection (score difference >0·50), whereas patients not allocated to preferred intervention or had splint or no preference improved by less than 0·25. The magnitude and durability of the treatment-specific benefits were moderate. A mean symptom severity score greater than 2·1 at 6 weeks and 6 months suggests many patients still have symptoms. At 6 months, sleep-disturbing symptoms were reported by almost a third of the patients, and about a third of the patients in the steroid group and a fourth in the night splinting group had, or were awaiting, surgery.
In this trial, adherence to night splinting was self-reported at 6 weeks. Because adherence is exclusive to the splinting group, low adherence might reduce treatment effect. A common steroid injection technique was used; different techniques exist but have no strong evidence for different outcomes.
However, injections were given not only by doctors but also by physiotherapists or occupational therapists, which is relevant to the conclusion that steroid injection is more cost-effective. Access to splints is easier whereas steroid injections are often prescribed and given by a doctor; costs would increase depending on the extent to which patients are referred for injections.
A small treatment benefit in favour of steroid injection should be sufficient to consider injection as a first-line treatment, provided that it can be given safely in primary care without referral and the patient accepts possible transient post-injection pain. This trial might justify a change in clinical practice in that patients with mild or moderate carpal tunnel syndrome can choose a single steroid injection in primary care instead of night splinting. A policy of initial treatment with steroid injection and considering surgery in case of inadequate improvement or recurrence of symptoms is reasonable and supported by evidence.