Background: Obesity is accompanied with low-grade inflammation and leucocytosis and increases the risk of venous thromboembolism. Associations with platelet count, however, are unclear because several studies have reported positive associations only in women. Associations with body shape are also unclear, because waist and hip circumferences reflect overall body size, as well as body shape, and are correlated strongly positively with body mass index (BMI).

Methods: We evaluated body shape with the allometric body shape index (ABSI) and hip index (HI), which reflect waist and hip size among individuals with the same weight and height and are uncorrelated with BMI. We examined the associations of BMI, ABSI, and HI with platelet count, mean platelet volume (MPV), and platelet distribution width (PDW) in multivariable linear regression models for 125,435 UK Biobank women and 114,760 men. We compared men with women, post-menopausal with pre-menopausal women, and older (≥52 years) with younger (<52 years) men.

Results: BMI was associated positively with platelet count in women, more strongly in pre-menopausal than in post-menopausal, and weakly positively in younger men but strongly inversely in older men. Associations of BMI with platelet count were shifted towards the inverse direction for daily alcohol consumption and current smoking, resulting in weaker positive associations in women and stronger inverse associations in men, compared to alcohol≤3 times/month and never smoking. BMI was associated inversely with MPV and PDW in pre-menopausal women but positively in post-menopausal women and in men. ABSI was associated positively with platelet count, similarly in women and men, while HI was associated weakly inversely only in women. ABSI was associated inversely and HI positively with MPV but not with PDW and only in women. Platelet count was correlated inversely with platelet size and positively with leucocyte counts, most strongly with neutrophils.

Conclusions: Competing factors determine the associations of BMI with platelet count. Factors with sexually-dimorphic action (likely thrombopoietin, inflammatory cytokines, or cortisol), contribute to a positive association, more prominently in women than in men, while age-dependent factors (likely related to liver damage and fibrosis), contribute to an inverse association, more prominently in men than in women.