Blood viscosity and cardiovascular disease

Abstract

Background Men have higher rates of ischemic cardiovascular disease (CVD) than women which contributes to shorter lives in men than women. In adults, hematocrit (Hct) is higher in men than women with the difference emerging at puberty; androgens raise Hct. At the same time the effect of some CVD risk factors, such as HDL-cholesterol and fasting glucose, on CVD is becoming increasingly unclear and attention is returning to all elements of Virchow’s triad, i.e., hypercoagulability as well as the more heavily studied endothelial function and blood flow. Hct affects blood viscosity, and has been considered as a CVD risk factor, but observational studies are inconsistent, possibly due to confounding by poor health status resulting in both lower Hct and higher CVD. Randomized controlled trials (RCTs) have largely targeted raising Hct and have not been definitive. This thesis took advantage of a large cohort in south China and publically available data to identify potentially modifiable determinants of Hct and using an observational analysis, a Mendelian randomization (MR) study and a systematic review and meta-analysis of RCTs, where available, assessed the causal effect of Hct and the fertility related anti-coagulant protein C on CVD and/or its sub-types. Methods In a large study (Guangzhou Biobank Cohort Study) from Southern China, I used adjusted linear regression to assess the associations of components of height with Hct, an environment-wide association study (EWAS) to search agnostically for drivers of Hct, and an adjusted Cox proportional hazards regression to assess the associations of Hct with all cause and cause-specific mortality overall and stratified by sex and health status, because of possible confounding by health status. I used MR studies of publically available summary genetic data to assess whether people with genetically higher Hct or protein C were more prone to ischemic heart disease (IHD). Finally, I conducted a systematic review and meta-analysis of RCTs assessing the effect of increasing Hct on CVD and its subtypes. Results Sitting height and height were positively associated with Hct but leg length was not. Vitamin A, serum calcium, serum magnesium and alcohol use were robustly associated with higher Hct while physical activity was robustly associated with lower Hct. The associations of Hct with IHD mortality varied by health status, with an inverse relation in the unhealthier group and potentially a positive association in the healthier group. In the MR studies, genetically higher Hct was not clearly associated with IHD, but genetically higher protein C was associated with lower IHD. In the systematic review and meta-analysis of RCTs, raising Hct increased stroke, thrombotic events, CVD and mortality, but not IHD. Little evidence of different effects by type of intervention, health status or baseline Hct was found.

Conclusion My study suggests that Hct and modifiable drivers of Hct might be targets of intervention to prevent stroke, thrombosis and CVD, but not IHD, instead protein C may be a target to prevent IHD. Moreover, Hct could partially underlie higher CVD rates in men than women, possibly operating via androgens.