TO THE EDITOR: I was confused and disappointed by the conflicting data in two different articles in the February 1, 2007, issue of American Family Physician regarding wrist splints for patients with carpal tunnel syndrome. In the article, “Braces and Splints for Musculoskeletal Conditions,”¹ Drs. Gravlee and Van Durme cite two studies showing that neutral splints relieved symptoms more than cock-up splints (20 degrees of extension) at two weeks,² and that full-time splint wear improved symptoms and functional deficits at six weeks compared with nighttime-only wear.³ In the Clinical Evidence Concise, “Carpal Tunnel Syndrome,”⁴ Dr. Ashworth comes to opposite conclusions about the same studies: no significant difference in symptoms between neutral angle and 20-degree extension wrist splinting after two weeks, and no significant difference in symptoms between full-time and nighttime-only neutral angle wrist splinting at six weeks.⁴

As a busy, full-time, practicing family physician, I rely on the AFP editors to provide me with concise, practical information to help me care for my patients. Please help me clarify the conflicting data cited.

in reply: Dr. Van Zandt raises some important questions that highlight the differences between systematic and narrative reviews. Clinical Evidence Concise reviews (published by the British Medical Journal) are systematic by design and involve a priori guidelines that attempt to add rigor to the review process. This includes a thorough review of each identified randomized controlled trial (RCT) and even reanalysis of data if necessary. Narrative reviews such as the article by Drs. Gravlee and Van Durme,¹ although readable and informative, are not constrained by the same standards.

Only one RCT compares different types of splinting (neutral versus 20 degrees of extension),² and one RCT compares different splinting regimens (full-day wear versus nighttime-only wear) in patients with carpal tunnel syndrome. Burke and colleagues examined the neutral angle versus 20-degree extension splints in 90 hands (59 patients) at two weeks and at two months.³ The main outcome was whether the patient's overall symptoms had improved “not at all,” “a little,” “a lot,” or “completely.” Because of low numbers, the “not at all” and the “a little” groups were collapsed and the “a lot” and “completely” groups were collapsed; this analysis showed neutral splints were significantly ‘superior’, as Drs. Gravlee and Van Durme stated. However, there was essentially no one (1 out of 90 wrists) in the “completely” improved symptoms group; therefore, this is a comparison of those who said their symptoms improved “a lot” versus those who said they improved “a little” plus those who said “not at all.”

I would argue strongly that the most logical way of comparing the data would be to divide patients into “those who improved” and “those who did not,” because what does “a little” or “a lot” mean to a patient with carpal tunnel syndrome? Accordingly, a reanalysis of the data showed no difference between neutral and extension splints in terms of those who improved (40 out of 45 hands) versus those who did not (38 out of 45 hands).

Drs. Gravlee and Van Durme state: “Another randomized clinical trial, comparing symptoms and functional deficits in nighttime versus full-time splint wear, found the most significant improvements at six-week follow-up in the group instructed to wear the splints full-time.”³ This description perhaps is misleading because the authors of this study clearly state that there was no difference in patient symptoms and function between nighttime and full-time wearer groups.³ What the authors did find was a weak, statistically significant difference between the two groups in terms of slight improvements in motor distal latency and for sensory distal latency. However, the statistical methods used to obtain these results are questionable, and more rigorous methods suggest that there is no true statistical difference between nighttime and full-time wear in the data presented in that study.³

in reply: We appreciate Dr. Van Zandt's request for clarification regarding the use of wrist splints for carpal tunnel syndrome, specifically whether the splint should be in the neutral or cock-up position and whether it should be worn full-time or only at night.

The discrepancies between our recommendations¹ and the Clinical Evidence Concise by Dr. Ashworth² are based on differing interpretations of two randomized clinical trials. First, as we originally reported, one study concluded that neutral splints relieved symptoms more than did cock-up splits (20 degrees of extension).³ in particular, 17 of the 45 wrists (37.8 percent) placed in neutral splints experienced “a lot” or “complete” relief, compared with only seven of the 45 wrists (15.6 percent) placed in extension (risk ratio = 2.4, 95% confidence interval 1.1 to 5.3).³

Ashworth's conclusion that there are no significant differences between neutral and extension splints is based on his reanalysis² of the same data. In the original study, Burke and colleagues compared patients who reported “a lot” or “complete” relief to those who reported “a little” or “no” relief. In his reanalysis, Ashworth chose instead to compare patients who reported no relief to those who reported any relief at all. This strategy masked the differences between patients who experienced “a little” versus “a lot” or “complete” relief, and it reduced the difference between neutral and extension splints to nonsignificance.² We suggest that Ashworth's alternate grouping may not be as helpful to the practicing physician. When making recommendations to patients, we believe it would be more intuitive to weigh the odds of achieving “a lot” or “complete” relief, as opposed to “little” or “no” relief. Thus, we believe that Burke and colleagues' data support the use of neutral splints.

Second, our recommendation for full-time use of splints is based on a small study.⁴ Ashworth correctly points out that the higher level of symptom relief in the full-time versus nighttime group is not statistically significant. However, the study did show statistically significant improvements in motor distal latency and sensory distal latency in the full-time group.⁴ This finding is remarkable, given the low statistical power, with subsample sizes of only 13 patients for the nighttime group and 11 patients for the full-time group. Nevertheless, one could argue that the recommendation for full-time use of wrist splints should be classified as evidence level C (consensus, disease-oriented evidence, usual practice, expert opinion, or case series) using the SORT criteria, because the statistically significant data on full-time wear come from physiologic rather than patient-oriented evidence.

Dr. Van Zandt's question underscores the need for more studies with larger samples to clarify the evidence regarding splinting for carpal tunnel syndrome.