Age‐related suppression effect of current body weight on the association between birthweight and blood pressure: The Bogalusa heart study

Background It has been controversial whether to adjust for current body weight while examining the relationship between birthweight and blood pressure (BP). Objective The present study aims to partition the total effect of birthweight on BP into its direct effect and indirect effect through current body mass index (BMI). Methods The study cohort consisted of 6251 participants who had birthweight information with 13 443 observations on BMI and BP in pre-adolescence (4-11 years), adolescence (12-19 years), young adulthood (20-30 years) and mid-adulthood (31-58 years). General third-variable models were used to distinguish the mediation and suppression effects of current BMI on the birthweight-BP association at different ages. Results The total effect of birthweight on systolic BP measured as standardized regression coefficient (β) without current BMI included in the model was 0.003 (P = .810) in pre-adolescents, -0.032 (P = .029) in adolescents, -0.066 (P = .002) in young adults and -0.051 (P = .023) in midlife adults. With additional adjustment for BMI, the direct effect of birthweight on systolic BP was strengthened to β = -0.066 (P = .013), β = -0.058 (P = .014), β = -0.094 (P = .020), β = -0.066 (P = .023); the suppression effects of BMI were calculated at 0.070, 0.027, 0.028 and 0.015 in the respective age groups. The decreasing trend of suppression effects with increasing age mimicked the trends of birthweight-BMI and BMI-BP correlations. Conclusions Current body weight has a suppression effect, not a mediation effect, on the birthweight-BP association, with pre-adolescents having the greatest suppression effect. The suppression effect is predominantly determined by birthweight-BMI and BMI-BP correlations.