This article is only available in the PDF format. Download the PDF to view the article, as well as its associated figures and tables.
As we pointed out in our discussion, we agree with Drs Vogt and Brondum that the selection of controls in our investigation was suboptimal and that selection bias might have been introduced.As noted in our article, most control subject specimens were obtained during a two-week period in March 1982, when an active effort was made to complete this investigation. In the two institutions in which specimens were sought, essentially all subjects from whom a serum specimen was available and who could be matched to a case were used. Twenty of the 24 childhood subjects were hospitalized children with a variety of illnesses. Three of these children had proved or suspected neoplastic disease (including two with leukemia). These children might not have responded normally to antigenic challenge, thus introducing a possible bias against detecting recent cytomegalovirus (CMV) or herpes simplex virus (HSV) infection; alternatively, they might have been
Kaplan JE, Stewart J, Schoenberger LB, Greenspan JR, Bomgaars M, Bart RD, Wiebe R. Guillain-Barré Syndrome and Bell's Palsy in Hawaii in 1981-Reply. JAMA. 1984;252(10):1282. doi:10.1001/jama.1984.03350100016017