The aim of our study was to explore factors associated with the intensity of signs and symptoms of flu-like illness, on adult patients seeking medical advice for flu. We analysed 554 index patients (56% influenza positive) included in a prospective household contact study during winter 1999-2000. The intensity of illness was calculated from 19 symptoms recorded by the GP during the visit (Z-score-GP) and from 13 symptoms reported by the patients at the first day of follow-up (Z-score-Pat). The two Z scores were highly correlated (intraclass correlation of 0.43). Multivariate regression identified the following predictors of Z-score-GP: influenza-positive diagnosis (+0.26, p=0.002) and gender female (+0.19, p=0.03). The predictors of Z-score-Pat were: influenza-positive diagnosis (+0.39, p=0.001), current smoker (+0.20, p=0.04), gender female (+0.20, p=0.02) and influenza vaccination (-0.39, p=0.003). In influenza-positive patients, influenza vaccine was highly linked with a decrease in intensity of symptoms (Z-score-GP: -0.38, p=0.04; Z-score-Pat: -0.49, p=0.009). This last result was not found in influenza-negative patients. The study identifies tobacco use and influenza-positive diagnosis as predictors of increased intensity of influenza-like illness. Interestingly, influenza vaccine decreased the intensity of symptoms in influenza-positive, but not in influenza-negative patients.