Rationale and tutorial for analysing and reporting sex differences in cardiovascular associations

Abstract

Cardiovascular disease (CVD) is the leading cause of death in women and men. Yet biological and social factors differ between the sexes, while the importance of CVD in women may be underestimated due to the higher age-specific rates in men and the historical bias towards the male model of CVD. Consequently, sex differences in risk factor associations with CVD occur, but these are not always recognised. This article argues that sex disaggregation should be the norm in CVD research, for both humanitarian and clinical reasons. A tutorial on how to design and analyse sex comparisons is provided, including ways of reducing bias and increasing efficiency. This is presented both in the context of analysing individual participant data from a single study and a meta-analysis of sex-specific summary data. Worked examples are provided for both types of research. Fifteen key recommendations are included, which should be considered when undertaking sex comparisons of CVD associations. Paramount among these is the need to estimate sex differences, as ratios of relative risks or differences in risk differences, rather than merely test them for statistical significance. Conversely, when there is no evidence of statistical or clinical significance of a sex difference, the conclusions from the research should not be sex-specific.