Background
Although idiopathic median neuropathy at the carpal tunnel (IMNCT) is objective and verifiable, distinction of normal and abnormal nerves is imprecise and probabilistic. The associated symptoms and signs (carpal tunnel syndrome [CTS]) vary, particularly for nonsevere (mild and moderate) median neuropathy. Discordance between diagnosis of mild or moderate median neuropathy at the carpal tunnel using symptoms and signs and diagnosis based on objective tests is a measure of the potential for overdiagnosis and overtreatment.
Question/purpose
What is the difference in the estimated prevalence of mild-to-moderate IMNCT using nonsevere signs and symptoms compared with the estimated prevalence using electrodiagnostic studies (EDS) and ultrasound (US)?
Methods
We used data from an existing cross-sectional data registry. To create this registry, between January 2014 and January 2019, we considered all new adult English-speaking people who had an EDS that included the median nerve or people with a diagnosis of CTS who did not have surgery yet. A small and unrecorded number of people declined participation. The cross-sectional area of the median nerve at the distal wrist crease using US in people who already had EDS was measured. People with a diagnosis of CTS underwent both EDS and US. The six signs and symptoms of Carpal Tunnel Syndrome 6 (CTS-6, a validated tool to estimate the probability of IMNCT using ratings of symptoms and signs of CTS) were recorded. This resulted in a registry of 185 participants; we excluded 75 people for obvious, severe IMNCT (defined as nonrecordable nerve conduction velocity, thenar atrophy, or greater than 5 mm 2-point discrimination). Three of the 110 qualifying patients had missing information on ethnicity or race, but we accounted for this in our final analysis. Without a reference standard, as is the case with IMNCT, latent class analysis (LCA) can be used to establish the probability that an individual has specific pathophysiologic findings. LCA is a statistical method that identifies sets of characteristics that tend to group together. This technique has been used, for example, in diagnosing true scaphoid fractures among suspected fractures based on a combination of demographic, injury, examination, and radiologic variables. The prevalence of mild-to-moderate IMNCT was estimated in two LCAs using four signs and symptoms characteristic of mild-to-moderate IMNCT, as well as EDS and US measures of median neuropathy.
Results
The estimated prevalence of mild-to-moderate IMNCT based on signs and symptoms was 73% (95% CI 62% to 81%), while the estimated prevalence using EDS and US measurements was 51% (95% CI 37% to 65%).
Conclusion
The notable discordance of 22% between the estimated prevalence of mild-to-moderate IMNCT using signs and symptoms and prevalence based on EDS and US criteria, and the overlapping CIs of the probability estimations, indicate considerable uncertainty and a corresponding notable potential for underdiagnosis or overdiagnosis. When signs and symptoms suggest mild-to-moderate median neuropathy and surgery is being considered, patients and clinicians might consider additional testing, such as EDS or US, to increase the probability of actual median neuropathy that can benefit from surgery. We might benefit from a more accurate and reliable diagnostic strategy or tool for mild-to-moderate IMNCT; this might be the focus of a future study.
Level of Evidence
Level III, diagnostic study.