Framingham Heart Study 100K Project: genome-wide associations for blood pressure and arterial stiffness

The Framingham Heart Study began in 1948 when 5209 men and women from Framingham, Mass, who were between 28 and 62 years of age were recruited to participate in an observational study [12]. Subjects underwent a medical history, physician-administered physical examination including blood pressure measurement, laboratory tests, and electrocardiography. Examinations have been repeated every 2 years. In 1971, 5124 offspring and spouses of offspring of original participants were recruited into the Framingham Offspring Cohort [13]. The offspring cohort was reexamined approximately every 4 years, except for an 8 year interval between their initial and second visit. All subjects gave written informed consent before each clinic visit, and the examination protocol was approved by the Institutional Review Board at Boston Medical Center (Boston, Mass).

At each clinic visit, the examining physician measured the systolic and diastolic BP in the left arm using a mercury column sphygmomanometer. BP was measured twice at each visit, with the exception of the first Offspring Cohort clinic visit, when it was measured once in about half the participants. Systolic and diastolic pressures were determined by the first and fifth Korotkoff sounds, respectively, and the two BP measurements were averaged to derive the systolic and diastolic pressures for that examination.

Examination cycles for the two cohorts were overlaid temporally as follows [offspring cohort/original cohort (earliest - latest year)]: examination 1/examination 12 (1971–1975), examination 2/examination 16 (1979–1983), examination 3/examination 18 (1983–1987), examination 4/examination 20 (1986–1991), examination 5/examination 22 (1990–1995), examination 6/examination 24 (1995–1998) and examination 7/examination 26 (1998–2001). Referring to offspring cycle numbers, the six BP phenotypes analyzed for this investigation were residuals for SBP and DBP at Examination 1, at Examination 7, and average of residuals from available Examinations 1 to 7. BP was imputed for treated observations as previously described [6]. No adjustment was made for untreated observations, which constituted the vast majority of BP values. Systolic and diastolic BP phenotypes were analyzed independently. Residuals were obtained from cohort- and examination-specific regression models accounting for sex, age and BMI; for DBP, age-squared was added. For inclusion in long-term BP analyses, each participant had to have BP measured on at least three examinations over a period of 12 years or more.

Arterial tonometry for assessment of arterial stiffness was conducted on Offspring Cohort participants attending their 7th clinic examination. Five primary tonometry phenotypes were analyzed: carotid-femoral and carotid-brachial pulse wave velocity, forward and reflected pressure wave amplitude, and mean arterial pressure. Tonometry was performed in the supine position after 5 minutes of rest. Arterial tonometry with simultaneous ECG recording was obtained from brachial, radial, femoral and carotid arteries using a commercially available tonometer (SPT-301, Millar Instruments, Houston, TX). Carotid-brachial, carotid-radial and carotid-femoral PWV were calculated as previously described [14]. Mean arterial pressure was calculated from the planimetered brachial arterial tracing after calibration to the brachial blood pressure, which was obtained by an oscillometric device. Forward pressure wave amplitude was defined as the difference between pressure at the waveform foot and pressure at the first systolic inflection point or peak of the carotid pressure waveform; reflected pressure wave amplitude was defined as the difference between the central systolic pressure and the pressure at the forward wave peak. Sex-specific regressions were conducted for each tonometry phenotype with the following covariates: age, age², height, weight, to generate sex-specific residuals.

Details of the genotyping methods are available in the Executive Summary [15]. Briefly, 112990 autosomal SNPs on the Affymetrix 100K chip were genotyped in the Boston University School of Medicine Genetics Laboratory on the Framingham Heart Study family plate set. SNPs were excluded for the following reasons: minor allele frequency <10% (n = 38062); call rate <80% (n = 2346); Hardy Weinberg equilibrium p value < 0.001 (n = 1595), leaving 70,987 SNPs available for analysis.

Standardized multivariable-adjusted blood pressure and tonometry residuals were generated as described above. Table 1 lists the covariates used for each phenotype. As described in the Executive Summary [15], we conducted association testing using family based association testing (FBAT), and generalized estimating equations (GEE) applied to the additive genetic effects model. In secondary analyses that used the GEE general genetic effects model, which is more sensitive to recessive genetic effects, to be more conservative, we limited analyses to two phenotypes: long-term SBP and long-term DBP, and we limited eligible SNPs to those with a minor allele frequency >= 0.20 and Hardy-Weinberg equilibrium p value >= 0.05. The software package Merlin [16] was used to compute exact identity by descent linkage probabilities for allele sharing, and linkage analysis by variance component method was carried out SOLAR using 11,200 SNPs and STRs. Heritability was estimated using variance-components methods (SOLAR). For BP, 2155 study participants were used for examination 1 SBP and DBP, 1479 for examination 7, and 2009 for long-term average; 770 individuals were used in heritability analysis of arterial stiffness phenotypes.

GEE and FBAT additive genetic effect models were run for SNPs in or near 6 genes in the renin-angiotensin-aldosterone pathways. These genes were selected a priori because of a substantial body of literature implicating them in hypertension and altered vascular properties. All SNPs from 200 Kb proximal to the start and extending to 200 kb of the terminus of each gene were included in analysis providing the minor allele frequency was >= 0.1, the genotype call rate was 0.8, and the Hardy-Weinberg equilibrium p value was >= 0.001.