Increasing aminoterminal-pro-B-type natriuretic peptide precedes the development of arterial hypertension: the multiethnic study of atherosclerosis

Abstract

Background: Longitudinal associations between the aminoterminal pro-B-type natriuretic peptide (NT-proBNP) and incident hypertension are lacking.

Methods: We tested associations between baseline NT-proBNP (bNT-proBNP) and change in NT-proBNP (ΔNT-proBNP) (visit 3 NT-proBNP - bNT-proBNP, 3.2 years apart) with incident hypertension (SBP ≥ 140 and/or DBP ≥90 mmHg or taking antihypertensive medications). Incident hypertension was evaluated in 5596 individuals in the Multi-Ethnic Study of Atherosclerosis without hypertension at baseline (53% women, age range 45-84 years without overt cardiovascular disease) and follow-up for 9.5 years and in a subgroup (1550) who had bNT-proBNP less than 100 pg/ml and no hypertension at visit 3. Incident hypertension was regressed (proportional hazards) on quintiles of bNT-proBNP (range) (reference <19.2, 19.3-40.8, 40.9-70.9, 71-135.2, and >135.5) and also on ΔNT-proBNP categories (reference <-10, -10 to 10, >10 to 50, and >50 pg/ml). Hazard ratios were adjusted for age, race, sex, education, diabetes, obesity, left ventricle mass/height, SBP and DBP, interleukin-6, salt intake, estimated glomerular filtration rate, and exercise.

Results: Compared with the reference category, hazard ratios (95% confidence interval) for incident hypertension compared with the first quintile of bNT-proBNP were 1.47 (1.13-1.93), 1.57 (1.18-2.09), 1.52 (1.12-2.06), and 2.36 (1.62-3.41). Hazard ratios for incident hypertension by categories of ΔNT-proBNP from 3.2 to 9.5 years follow-up were 0.98 (0.62-1.56), 1.13 (0.72-1.79), and 1.82 (1.07-3.12).

Conclusion: The development of hypertension tended to be preceded by elevated levels of bNT-proBNP or a substantial positive ΔNT-proBNP.

Figure 1

Cross-sectional analysis at baseline. Dip in NT-proBNP levels observed in pre-hypertensive BP categories. Values are back transformed geometric means ± 95% CI. Model 1 adjusted for age, race and gender. Model 2 = model 1 + highest level of education, presence of diabetes, presence of obesity (BMI > 30 kg/m²), left ventricular end diastolic volume, heart rate categories < or > then 60 beats per minute, left ventricular mass/height, add table salt, estimated glomerular filtration rate, IL-6, and total intentional exercise in MET-min/week. * indicates that NT-proBNP is lower in category c than in categories a, b, d and e for both model 1 and 2, p < 0.02.

Figure 2

NT-proBNP values as a function of left ventricular diastolic wall distensibility. NT-proBNP values are back transformed geometric means ± 95% CI. Quadratic model adjusted for age, race and sex, p < 0.0001.

Figure 3

Hazard ratios for incident hypertension according to quintiles of NT-proBNP concentrations. Prospective analysis using baseline NT-proBNP and any incident high blood pressure at exams 2–5 (9.7 yr followup), N=2925. Hazard ratios ± 95% CI. Models 1 and 2 covariates are described in Methods and in the footnote to Table 2. Bar represents the hazard ratio reference line. p values for trend for linear associations regressed on NT-proBNP as a continuous variable for model 1 < 0.0001 and model 2 = 0.0007. Quadratic associations were not significant.