Abstract
OBJECTIVE: To investigate the extent to which (1) secular changes in mid-adulthood WC are independent of BMI and (2) secular changes in low HDL-C are dependent on WC in each sex. METHODS: The sample comprised 19,406 adults (aged 43-47 years) from three birth cohort studies with BMI and WC measured in 1990, 2003, or 2018; 13,239 participants additionally had HDL-C measured in 2003 or 2018. Quantile regression was used to model differences between 1990-2003 and 2003-2018 in (1) BMI and WC internal Z-scores and (2) WC in cm before and after adjustment for BMI. Binary logistic regression was used to model differences between 2003 and 2018 in low HDL-C, before and after adjustment for BMI or WC. RESULTS: Secular increases in BMI and WC were larger between 1990 and 2003 than 2003 and 2018 and at the upper ends of the distributions. At the 85th quantile, effect sizes were larger for WC than BMI Z-scores in females but not males. Adjustment for BMI attenuated estimates of secular increases in WC in cm more in males than females. Odds ratios for low HDL-C in 2018 compared to 2003 were 1.73 (95% CI 1.32, 2.28) in males and 1.34 (1.01, 1.78) in females. Adjustment for WC did not substantially change the estimate in males but attenuated the estimate for females to 1.09 (0.81, 1.47). CONCLUSIONS: In women much more so than in men, secular increases in mid-adulthood WC appear to have occurred independently of BMI and largely explain the observed rise in low HDL-C prevalence between 2003 and 2018.