This study deals with regional differences in annual prevalence of 44 declared diseases or clinical disorders in 1989 in GAZEL cohort participants from "Electricité de France-Gaz de France" company. The studied sample consisted of 20,325 people, from 35 to 50 years old, living in France, who answered to a mail questionnaire in 1989. The collected data from one year were analysed. An overall "regional effect" was searched about every listed disease in the questionnaire. For this purpose, a logistic model, controlling for age and sex, was used. For comparing prevalence of a disease in one region with average prevalence, an odds ratio (OR) was calculated: it was defined here as the ratio of disease odd in the region to the average French odd. The regional effect was studied in greater detail for five diseases, making allowance for extra factors of adjustment (tobacco and alcohol consumption, Quetelet index, salary category, number of declared diseases). The regional effect remained significative for four diseases: hypertension with OR from 1.26 to 1.68 in northern regions, and from 0.63 to 0.75 in southeastern ones; osteoarthrosis with OR lower than 1 in western regions (0.60 to 0.79), and larger in the South (1.24 to 1.46); hyperthyroidism, goiter with OR elevated in Midi-Pyrénées (2.91) and Lorraine (1.82) [corrected]; renal stones with OR high in the South (1.85 in Provence Riviera). These results can be compared to known observations about geographical differences in mortality. They also suggest persistence of hazardous areas for some diseases.