Changes in the association of lifecourse body-mass index trajectories with adult blood pressure across two cohorts in Britain: an observational study

Abstract

BackgroundRecent increases in the prevalence of obesity and more rapid gains in body-mass index (BMI) have been noted for all ages. The effects on blood pressure (BP) of secular changes in BMI or lifetime BMI trajectories are little. We investigated associations between lifecourse BMI trajectories and adult BP across two generations. MethodsUsing the 1946 (n∼5300) and 1958 (n∼17 000) British birth cohorts, we fitted joint multivariate response models to repeated BMI measures (at ages 7, 11, 16, 20, 26, 36, 43, and 50 years in the 1946 cohort; and at 7, 11, 16, 23, 33, and 45 years in the 1958 cohort) and adult BP (measured at 43 years for the 1946 cohort [centred at 45 years] and at 45 years for the 1958 cohort). To characterise distinct slopes for childhood and adulthood BMI gain, we used linear spline models with random coefficients for BMI measures. Although there was loss to follow-up with age (n=3035 at 53 years for the 1946 cohort; n=9377 at 45 years for the 1958 cohort), most participants (4787 and 16 820, respectively) had at least one BMI or BP measure and were included in the joint models. Different cuff sizes and BP devices were used in the two cohorts, potentially introducing bias; thus, we standardised measurements to the mercury sphygmomanometer readings. FindingsMean systolic BP (SBP) decreased from the older to younger cohort by 2·8 mm Hg in female but not male participants; mean diastolic BP (DBP) decreased by 3·2–3·3 mm Hg (both sexes). Geometric mean adult BMI (45 years) was higher in the younger than the older cohort (27·7 kg/m2 [95% CI 27·6–27·8] vs 25·9 kg/m2 [25·7–26·2] in males and 26·6 kg/m2 [26·5–26·8] vs 25·2 kg/m2 [25·0–25·3] in females). Adult SBP was associated with BMI onwards from 11 years in the younger cohort, and late teens (female) and late 20s (male) in the older cohort. Associations of SBP with adult BMI were stronger in the 1958 than 1946 cohort: the correlation coefficient (r) was 0·27 (95% CI 0·24–0·29) versus 0·09 (0·04–0·13) for males and 0·29 (0·26–0·32) versus 0·08 (0·03–0·12) for females, suggesting that mean SBP decreased more among those with lower than with higher BMI values between cohorts. The slope of BMI trajectory was steeper in the younger cohort from early adulthood and its association with adult BP was stronger. For males, the childhood slope for BMI gain was associated with adult SBP only in the 1958 cohort (r=0·21, 95% CI 0·17–0·24); the association for adult BMI slope was stronger in the younger cohort (r=0·28, 0·25–0·33 vs 0·13, 0·06–0·20). For females, childhood BMI slope was associated with SBP for both cohorts (r=0·19, 0·12–0·26 for the 1958 cohort and 0·11, 0·02–0·21 for the 1946 cohort); adult slope was associated with SBP only in the younger cohort (r=0·34, 0·31–0·37). Associations were similar for DBP. InterpretationBP did not increase between two generations of contemporary adults born 12 years apart despite higher adult BMI levels in the younger cohort. A stronger association between BMI trajectory and BP in mid-adulthood in the younger cohort suggests that BMI-related effects may have been offset by improvements in other factors linked to BP; for example, in detection and treatment of hypertension or decreasing prevalence of smoking. However, our study is limited to two cohorts. We need to understand the explanations for the strengthening BMI and BP association and its implications in the context of the obesity epidemic. FundingLL was funded by a Medical Research Council (MRC) Career Development Award in Biostatistics (grant G0601941). The MRC has funded the 1946 cohort since 1962 (grants U1200632239 and U12309272) and also the 45-year survey of the 1958 cohort (grant G0000934). This work was undertaken at Great Ormond Street Hospital and University College London Institute of Child Health who received a proportion of funding from the Department of Health's National Institute for Health Research Biomedical Research Centres funding scheme. The Centre for Paediatric Epidemiology and Biostatistics also benefits from funding support provided by the MRC in its capacity as the MRC Centre of Epidemiology for Child Health (grant G0400546).