The contribution of smoking to differences in cardiovascular disease incidence between men and women across six ethnic groups in Amsterdam, the Netherlands: The HELIUS study

Abstract

It is unclear to what extent differences in cardiovascular disease (CVD) risk between men and women are explained by differences in smoking, and whether this contribution to risk is consistent across ethnic groups. In this prospective study, we determined the contribution of smoking to differences in CVD incidence between men and women, also in various ethnic groups. We linked baseline data of 18,058 participants of six ethnic groups from the HELIUS study (Amsterdam, the Netherlands) to CVD incidence data, based on hospital admission and death records from Statistics Netherlands (2013-2019). The contribution of smoking to CVD incidence, as estimated by the population attributable fraction, was higher in men than in women, overall (24.1% versus 15.6%) and across most ethnic groups. Among Dutch participants, however, the contribution of smoking was higher among women (21.0%) than men (16.2%). Using Cox regression analyses, we observed that differences in smoking prevalence explained 22.0% of the overall lower hazard for CVD in women compared to men. Smoking contributed minimally to the lower hazards for CVD in women among participants of Dutch (0%), Ghanaian (4.9%) and Moroccan origin (0%), but explained 28.6% and 48.6% of the lower hazards in women in South-Asian Surinamese and African Surinamese groups, respectively. While smoking prevention and cessation may lead to lower CVD incidence in most groups of men and women, it may not substantially reduce disparities in CVD risk between men and women in most ethnic groups.

Keywords: Cardiovascular disease; Ethnicity; HELIUS study; Sex and gender differences; Smoking.

Fig. 1

Age-standardized CVD incidence rates split for the proportions related to smoking and not related to smoking, by sex, overall and per ethnic group.

Fig. 2

Forest plot of adjusted hazard ratios (HRs) for CVD incidence of women compared to men and relative change in HR for sex after additional adjustment for smoking, overall and per ethnic group. ^aProportional hazards assumption violated. Estimates should thus be treated with caution.