LETTERS TO THE EDITOR DISTRIBUTION OF FIBRILLATION POTENTIALS IN RADICULOPATHIES The report by Dillingham et al. 1 questions the general electromyographic (EMG) wisdom that fibrillation poten- tials develop and resolve earliest in proximal muscles in a damaged root distribution. The data presented do not al- low the authors to make such a judgment. Apparently, patients were selected for entry into the study based on EMG features of cervical radiculopathy requiring denerva- tion in at least two muscles in the diagnosed root distribu- tion, but we are not informed of the criteria by which specific root level diagnoses were made. 2 A number of patients were entered into the study based on the presence of paraspinal fibrillation only (‘‘indeterminate levels’’). Some patients demonstrated multilevel paraspinal fibrilla- tion potentials and evidence of bilateral cervical radicu- lopathies (‘‘multiple radiculopathies’’), further confusing the issue of the timing of the onset of the radiculopathy, and the segmental source of the fibrillation potentials. Five of the patients included in the analysis were said to have fibrillation potentials in a cervical root distribution 5 years after onset of the symptoms, adding doubt to the cause of the paraspinal fibrillation and the timing of onset. Clinical criteria used for the inclusion of patients with cervical ra- diculopathy were not reviewed in the methods section and there was no information about the underlying structural causes of the radiculopathies or their anatomic correlation with the root levels diagnosed by EMG. 2 Regardless of the quality of the statistical analysis, the lack of precision in the data on which the statistics were based prevents any con- clusions from being drawn. Kerry H. Levin, MD Department of Neurology, Cleveland Clinic Foundation, Cleveland, Ohio 44195, USA 1. Dillingham TR, Pezzin LE, Lauder TD. Cervical paraspinal muscle abnormalities and symptom duration: a multivariate analysis. Muscle Nerve 1998;21:640–642. 2. Levin KH, Maggiano HJ, Wilbourn AJ. Cervical radiculopa- thies: comparison of surgical and EMG localization of single- root lesions. Neurology 1996;46:1022–1025. Reply We appreciate the comments by Dr. Levin on our study examining the relationship between symptom duration and the probability of finding denervation potentials in upper limb muscles among patients with cervical radicu- lopathy. 1 Dr. Levin raises questions about our ability to reach conclusions based on our sample selection criteria and the lack of information about the underlying cause of the radiculopathy. Although these are important issues for other investigations, we believe that Dr. Levin’s concerns are not relevant given the focus of our study. The proposition that paraspinal and other limb muscles demonstrate a defined time course of denervation and reinnervation presupposes that the radiculopathy can be electrodiagnostically confirmed. There are cases in which a radiculopathy is suggested either clinically or by imaging studies, yet electromyographic (EMG) muscle as- sessment produces normal results. Conversely, an EMG study may be suggestive of a radiculopathy even in the absence of structural findings. Because our purpose was not to evaluate structural causes of radiculopathy but rather to examine the relationship between denervation and symptom duration, the inclusion criteria of electrodi- agnostically confirmed cervical radiculopathies was most appropriate. As stated in our article, root level classifica- tion was based upon the minimum number of root levels necessary to account for all abnormal muscles. Dr. Levin’s contention that a lack of precision in the data prevents any conclusions from being drawn is incor- rect. The large number of cases studied, capturing a broad time spectrum, ensured adequate power to detect statisti- cally significant relationships. As Dr. Levin pointed out, 5 patients reported symptom durations of many years. Sim- ply discarding these patients from the analyses would mis- represent the sample. The absence of a significant associa- tion between symptom duration and the probability of denervation potentials for any muscle examined suggests that this relationship does not exist. Of note, similar nonsignificant findings were found in our analyses of lumbosacral radiculopathies. 2 As discussed Muscle Nerve 22: 287–292 CCC 0148-639X/99/020287-06 © 1999 John Wiley & Sons, Inc. Letters to the Editor MUSCLE & NERVE February 1999 287