Blood pressure and cardiovascular risk in relation to birth weight and urinary sodium: an individual-participant meta-analysis of European family-based population studies

Abstract

Background: Although the relation of salt intake with blood pressure (BP) is linear, it is U-shaped for mortality and cardiovascular disease (CVD). This individual-participant meta-analysis explored whether the relation of hypertension, death or CVD with 24-h urinary sodium excretion (UVNA) or sodium-to-potassium (UNAK) ratio was modified by birth weight.

Methods: Families were randomly enrolled in the Flemish Study on Genes, Environment and Health Outcomes (1985-2004) and the European Project on Genes in Hypertension (1999-2001). Categories of birth weight, UVNA and UNAK (≤2500, >2500-4000, >4000 g; <2.3, 2.3-4.6 and >4.6 g; and <1, 1-2, >2, respectively) were coded using deviation-from-mean coding and analyzed by Kaplan-Meier survival functions and linear and Cox regression.

Results: The study population was subdivided into the Outcome ( n = 1945), Hypertension ( n = 1460) and Blood Pressure cohorts ( n = 1039) to analyze the incidence of mortality and cardiovascular endpoints, hypertension and BP changes as function of UVNA changes. The prevalence of low/medium/high birth weight in the Outcome cohort was 5.8/84.5/9.7%. Over 16.7 years (median), rates were 4.9, 8 and 27.1% for mortality, CVD and hypertension, respectively, but were not associated with birth weight. Multivariable-adjusted hazard ratios were not significant for any endpoint in any of the birth weight, UVNA and UNAK strata. Adult body weight tracked with birth weight ( P < 0.0001). The partial r in the low-birth-weight group associating changes from baseline to follow-up in UVNA and SBP was 0.68 ( P = 0.023) but not significant in other birth weight groups.

Conclusion: This study did not substantiate its prior hypothesis but showed tracking of adult with birth weight and suggest that low birth weight increases salt sensitivity.

Graphical abstract

FIGURE 1

Flowchart for participants in the study. Outcome cohort and Hypertension cohort refer to participants used to study the incidence of mortality and morbidity of cardiovascular endpoints and the incidence of hypertension, respectively. The relation of the changes from baseline to follow-up in blood pressure with the changes in the 24-h urinary sodium excretion was investigated in the Blood Pressure cohort. See Methods for the definition of inaccurate urine collections.

FIGURE 2

Boxplots showing the distributions of birth weight (a) and adult body weight (b) at baseline in the low (≤2500 g; yellow), medium (>2500–4000 g; blue) and high (>4000 g; red) birth weight categories. The central line, the upper and lower lines, and the upper and lower caps represent the median, interquartile range, and the 5th–95th percentile interval. The arithmetic means and extreme measurements are represented by circles inside the box and outside the whiskers, respectively. The arithmetic means and standard deviations are given above/beneath the whiskers. The number of the participants in each category is given along the horizontal axis. Significance of the difference with the medium group: ^‡P ≤ 0.001.

FIGURE 3

Kaplan–Meier survival function estimates for the incidence of total mortality, all cardiovascular endpoints and hypertension. The yellow, blue and red lines in panels a, b and c represent the low, medium and high categories of birth weight (≤2500, >2500–4000 and >4000 g). The yellow and red lines in panels d, e and f represent the low, medium and high categories of 24-h urinary sodium (<2.3, 2.3–4.6, >4.6 g per day. The results for all-cause mortality and all cardiovascular endpoints were obtained in the Outcome cohort, and the results for hypertension in the Hypertension cohort. P values indicate the significance of log-rank test.

FIGURE 4

Relation between the changes from baseline to follow-up in SBP and the changes in 24-h urinary sodium excretion in the low-birth-weight (a), medium-birth-weight (b) and (c) high-birth-weight categories (≤2500, >2500–4000, >4000 g). n and r refer to the number of participants in each group and to the partial correlation coefficients with adjustments applied for recruitment (FLEMENGHO vs. EPOGH), sex, and the baseline variables age, heart rate, the serum total-to-HDL cholesterol ratio, blood glucose, educational attainment (elementary school vs. secondary and high education) and the glomerular filtration rate estimated from serum creatinine. To illustrate the fit of the linear model, the average changes in SBP and in 24-h urinary sodium were computed by quantiles and plotted over the regression lines given with 95% confidence interval. Categories were defined by the median, quintiles and tertiles in the low-birth-weight, medium-birth-weight and high-birth-weight groups, respectively. The results were obtained in the Blood Pressure cohort (see Fig. 1).