Explaining patterns in the sex ratio (male/female) of cardiovascular disease (CVD) mortality would improve understanding of mortality transitions under modernisation. Little research has examined secular trends in this ratio across populations, taking age and cohort into account. We examine cohort effects in the ratios of CVD mortality (including ischaemic heart disease and cerebrovascular disease) among 4 East Asian populations that vary in the timing of their modernisation, and assess the effect of smoking on these patterns in comparison with Western populations. The sequential method for log-linear models is applied to analyse age, period and cohort effects for sex ratios. Age and cohort effects are fitted first, with population as offset; period effects are fitted in a second model using the fitted values from the first model as the offset. Lung cancer mortality serves as a proxy for smoking. Increases in sex ratios of CVD mortality began in earlier cohorts in Western than in East Asian populations. Once begun, increases were more rapid in East Asia. The cohort effect for the sex ratio of CVD mortality differs from that for lung cancer mortality. Trends in sex ratios of CVD mortality by cohort are similar before and after adjustment for lung cancer mortality in East Asia; the increasing trend across 1900-1945 cohorts is maintained in Western populations after adjustment. The sex ratio of CVD mortality has increased across successive cohorts living in increasingly modernised environments. There is scant evidence that this increase is attributable to changing sex-specific rates of smoking.