Data derived from a cohort study of 191 men, seen 5 years apart, were used to investigate the involvement in allergic as well as in nonallergic upper airway disease (UAD) of two risk factors, immediate hypersensitivity and tobacco smoking, the roles of which have been well established in lower airway disease. At both surveys, UAD and smoking habits were assessed by an extended version of the BMRC/ECSC questionnaire. UAD consisted of usual or chronic rhinitis, seasonal allergic rhinitis and perceived nasal hyperresponsiveness (PNHR) to tobacco smoke, cold air or exercise. Immediate hypersensitivity was determined either in vivo (skin prick test, SPT, positivity) or in vitro (total IgE). UAD prevalence and smoking habits did not vary significantly over 5 years. On the contrary, SPT positivity increased significantly between the two surveys. At both surveys, SPT positivity for common aeroallergens (grass pollens overall) was significantly related to seasonal allergic rhinitis but not to usual or chronic rhinitis and to PNHR. Similarly, the total IgE level was increased in seasonal allergic rhinitis, but never significantly. Current smoking was always a habit significantly more frequent in men reporting chronic rhinitis (odds ratios of 5.3 and 4.9 in 1985 and 1990, respectively). The relationship was of the dose-response type: the more the subjects smoked, the more they reported chronic rhinitis. In contrast, seasonal allergic rhinitis was more associated with exsmoking, either in 1985 or in 1990. These results were confirmed longitudinally and after exclusion of asthmatics. Further investigations are needed to support the hypothesis raised by our data according to which immediate hypersensitivity, as assessed by SPT positivity for common aeroallergens, and tobacco smoking might intervene alternatively in UAD, probably because of the 'healthy smoker effect'.