I read with interest the superb review/AAEM minimonograph number 26, ''The Electrodiagnosis of Carpal Tunnel Syndrome'' by Stevens. I was puzzled by his statement, ''It is not wise, however, to do multiple different sensitive NCSs in the same patient, because of the risk of a type I error (normal patient is mistakenly called abnormal) increases with each additional technique used.''
The same is true of the clinical evaluation. Does the author advocate the performance of a cursory history and physical examination because this would also decrease the probability of a type I error? The skilled clinician appreciates the difference between statistical and clinical significance. Just as it is not wise to make a diagnosis of amyotrophic lateral sclerosis on the basis of a few scattered fasciculations in the presence of an otherwise normal exam, so too the thoughtful electromyographer does not make a diagnosis of carpal tunnel syndrome on the basis of a minimal isolated abnormality on serial, 1st digit, 4th digit, or palmar study in a patient whose clinical presentation is not strongly suggestive of carpal tunnel syndrome.
While it is not absolutely necessary to perform a wide battery of sensitive examinations in a patient with electrophysiologically obvious carpal tunnel syndrome, the more sensitive parameters become invaluable in the minority of patients whose symptomatology does not respond to surgery. In patients whose symptomatology does not respond, the comparison of multiple sensitive postoperative parameters with the preoperative values enables one to document with great certainty the quality of the electrophysiologic result. This is often not the case when only single motor and sensory determinations, are performed preoperatively. Since response to surgery cannot be predicted, and these examinations take only a few extra minutes for the experienced examiner to perform, I believe that the more sensitive examinations should be done in all patients undergoing surgery, regardless of how obvious their abnormalities are using conventional parameters.