Phil B.FontanarosaMD, Deputy EditorIndividualAuthorStephen J.LurieMD, PhD, Fishbein FellowIndividualAuthor
To the Editor: Several factors cause biases
in the CTS prevalence estimates reported by Dr Atroshi and colleagues1 (at least 1 upward and multiple downward biases);
consequently, the accuracy of these estimates is questionable.
The authors used a median-ulnar sensory latency difference of 0.8 milliseconds
or longer derived from normal values reported by Kimura2
as their electrophysiological criterion of abnormality. However, the practice
parameter for electrodiagnostic studies3
cited by the authors in their choice of criterion requires that reference
values be obtained with either concomitant studies of a reference population
or with previous studies of a reference population in the same laboratory.
The Kimura data used by the authors were obtained from a different population
in a different laboratory. In fact, Atroshi et al did study an appropriate
reference population, that is, asymptomatic respondents to their survey. However,
instead of using this reference population to derive the criterion of abnormality
to be applied to their test subjects, they applied the Kimura criterion to
the asymptomatic responders, yielding an 18.4% rate of supposed subclinical
median neuropathy. If the authors had instead appropriately derived their
criterion of abnormality from this reference population, their criterion of
abnormality would have been significantly higher, yielding lower estimates
for the prevalence of CTS. (A criterion of abnormality applied to the reference
population from which it was derived is expected to yield 2.5% "abnormal"
results for normally distributed populations.)
Zucker RS. Clinical Diagnosis of Carpal Tunnel Syndrome. JAMA. 2000;283(8):1000-1003. doi:10.1001/jama.283.8.999