Search

In Reply.— Soller and Stander, on the basis of an unpublished analysis of our data, raise the issue of two potential biases: protopathic and product confusion bias. The existence of protopathic bias, or bias caused by the disease preceding the risk factor, depends on how the onset of RS is defined. We defined the onset of RS as the first day of severe vomiting. In the second year of our study, aspirin was classified as a positive exposure for cases only if ingested before the first day of severe vomiting. Thus, this bias regarding the temporal association of aspirin and RS should not be a problem if the onset of vomiting properly identifies the onset of RS. Even if clinical RS can precede severe vomiting, a retrospective determination of the exact onset on the basis of a complex of nonspecific symptoms, eg, headache, fever, or sore throat, symptoms all compatible

Ninety-seven Reye's syndrome (RS) cases in Ohio children with onsets from December 1978 through March 1980 were studied for medication use during their pre-RS illness. They were matched with 156 control subjects for age, race, sex, geographic location, time, and type of illness. Only the use of aspirin was reported by significantly more cases (97%, 94/97) than controls (71%, 110/156) during the pre-RS matched illness. Using a multiple logistic model to control for the presence of fever, headache, and sore throat statistically, the difference in aspirin use remained significant. Conversely, fewer cases (16%) took medications containing acetaminophen than controls (33%). In 87% of the cases receiving aspirin, their maximum daily dosage did not exceed recommended levels, but their doses were higher than those of controls receiving aspirin. No relationship was found between dosage and stage of RS encephalopathy. ( JAMA 1982;248:687-691)

In Reply.— The Ohio Department of Health received only two reports of Reye's syndrome cases from the Cleveland Clinic during the study period (December 1978 through March 1980). The first of the two cases reported to us was not included in the study because our nurse coordinator could not get timely access to the child's record to verify fulfillment of the case definition. This child was reported by Dr Orlowski to have received aspirin. The other patient who was included in our study did not have a history of aspirin ingestion. Thirty-three hospitals reported cases to the Ohio Department of Health. With the exception of Dr Orlowski, there is no evidence of similar reporting problems at the other 32 hospitals. The vast majority of cases (85% ) of the 227 cases were reported from the state's six major pediatric referral centers. In Cleveland, this was Rainbow Babies and Children's Hospital (affiliated