Objective:To assess whether the difference in risk of cardiovascular mortality between urban and rural areas of Serbia could be explained by differences in the use of cardiovascular medication. Methods: The Serbian cohorts of the Seven Countries Study, Velika Krsna (VK), Zrenjanin (ZR) and Belgrade (BG), were enrolled in 1962-1964 and were followed up for 25 years. The survivors of these cohorts were re-examined in 1987, 1988 and 1989, respectively. This second examination of elderly men aged 65 to 84 years included a questionnaire about current use of cardiovascular medication, risk factors and diseases and a physical examination. All subjects were followed until death or the predefined censor date (10 years after baseline). The Cox proportional hazards model was used to calculate the risk of cardiovascular mortality in the rural cohorts compared to the urban cohort and to adjust for confounding. Main outcome measure: Cardiovascular death. Results: A total of 227 men from VK, 184 men from ZR and 287 men from BG were followed for a mean duration of 7.4 years and was complete for all subjects. After exclusion of 13 subjects with missing medication data, the incidences of cardiovascular mortality in VK, ZR, and BG were 60, 74, and 26 per 1000 person-years, respectively. The prevalence of cardiovascular medication use was 38% in VK, 52% in ZR, and 59% in BG. The greatest difference in use of specific medication was observed for betablockers (0% in VK and ZR, 13% in BG). After adjustment for cardiovascular risk factors, diseases and age, the relative risks (RRs) of cardiovascular mortality were 2.12 [95% CI: 1.44¿3.12], and 2.27 [95% CI: 1.56¿3.30] in VK, and ZR compared to BG. Additional adjustment for the use of cardiovascular medication increased these RRs to 2.40 [95% CI: 1.61¿3.60] and 2.55 [95% CI: 1.72¿3.78], respectively. Conclusion:The variation in cardiovascular medication use could not explain the excess risk of mortality in the rural Serbian cohorts compared to urban Belgrade.