Although the original Centers for Disease Control study of the relation between A/New Jersey/8/76 (swine flu) vaccine and Guillain-Barré syndrome (polyradiculoneuritis) demonstrated a statistical association and suggested a causal relation between the two events, controversy has persisted. To reassess this association, the authors obtained medical records of all previously reported adult patients with Guillain-Barré syndrome in Michigan and Minnesota from October 1, 1976 through January 31, 1977. To identify previously unreported hospitalized cases with onset of symptoms during this period, the authors surveyed medical care facilities. A group of expert neurologists formulated diagnostic criteria for Guillain-Barré syndrome and then reviewed the clinical records in a blinded fashion. Of the 98 adult patients from the original Centers for Disease Control study eligible for consideration, three were found to have been misclassified by date of onset and were excluded. Of the remaining 95, the 28 (29%) who did not meet the diagnostic criteria were equally distributed between the vaccinated group (18 of 60, 30%) and the unvaccinated group (10 of 35, 29%). In addition to the 67 remaining cases who met the diagnostic criteria, six previously unreported cases (three of whom had been vaccinated) were found and included in this analysis. The relative risk of developing Guillain-Barré syndrome in the vaccinated population of these two states during the 6 weeks following vaccination was 7.10, comparable to the relative risk of 7.60 found in the original study. These findings suggest that there was an increased risk of developing Guillain-Barré syndrome during the 6 weeks following vaccination in adults. The excess cases of Guillain-Barré syndrome during the first 6 weeks attributed to the vaccine was 8.6 per million vaccinees in Michigan and 9.7 per million vaccinees in Minnesota. No increase in relative risk for Guillain-Barré syndrome was noted beyond 6 weeks after vaccination.