Sample selection and gold standard testing for a proper group comparison

Abstract

We have read with interest the article entitled “Charcot– Marie– Tooth disease misdiagnosed as chronic inflammatory demyelinating polyradiculoneuropathy: an international multicentric retrospective study”. [1] We believe that major methodological flaws undermine the reliability of the results of this study. Indeed, the authors compared the clinical and electrophysiological features of a subgroup of patients with Charcot– Marie– Tooth disease (CMT) with controls affected by chronic inflammatory demyelinating polyradiculoneuropathy (CIDP). The patients affected by CMT underwent genetic testing on the basis of clinical judgement, as they presented with features that raised the suspicion of an inherited polyneuropathy. This represents a selection bias, [2] since CMT cases are not representative of the source population, and cases and controls are differently selected. Moreover, as is clearly stated in the methods, no genetic investigations were performed in the control group, making it impossible to exclude CMT in these patients. This constitutes a classification bias [2] and no inference can be made about different clinical features in the two groups. Thus, the only reliable result of the cited study [1] is represented by the clinical description of the 35 CMT patients who previously received a CIDP diagnosis, although a remote possibility of the cooccurrence of CMT and CIDP cannot be ruled out.