Our objective was to evaluate influenza vaccine effectiveness during an influenza epidemic by means of a matched case-control study. The study was performed by 35 general practitioners who collected specimens for influenza virus testing from 610 patients who consulted for infectious syndrome: 168 (28%) were influenza-positive. Two designs were used for selecting controls to take into account the high incidence-rate of influenza-like illness and the various possible protective effects of the vaccine. A first disease-free control matched for age and sex was selected during the same week as the case. A second control matched for age and sex was selected at the end of the epidemic period, irrespective of disease history during the epidemic period. Upper and lower bounds of vaccine effectiveness can be derived from these case-control designs. After adjustment for chronic conditions and exposure to an index case, analysis of the matched-pairs whose case was influenza-positive showed, with the first group of controls, an influenza vaccine effectiveness of 68% (95% CI, 10% to 88%) and, in the second group, 53% (95% CI, -19% to 82%). Among the pairs whose case was negative for influenza, vaccine effectiveness was, respectively, 31% (95% CI, -17% to 59%) and 12% (95% CI, -47% to 47%). Vaccine effectiveness was highest for the H3N2 subtype whose vaccine strain was identical to that of the wild-type strain. The results suggest that influenza vaccine is effective in the field in preventing influenza morbidity.