The Epidemiology of Hearing Loss Study (EHLS) began in 1993 in a population-based cohort developed for the Beaver Dam Eye Study (BDES). A private census conducted in 1987-88 identified residents of the city or township of Beaver Dam 43 to 84 years of age (n=5,924) who were then invited to participate in the BDES (n=4,926) (Klein, 1994
). Participants alive as of March 1, 1993 were eligible for the EHLS. Of these, 3,753 people (82.6%) participated in the baseline examination (1993-1995), 2,800 people participated in the 5-year follow-up (EHLS2 1998-2000), and 2,395 people participated in the 10-year follow-up examination (EHLS3 2003-2005). Data from the EHLS and concurrent BDES examinations were included in these analyses of the 2,042 EHLS2 participants with IMT measurements. These studies were approved by the University of Wisconsin-Madison Internal Review Board.
Educational status obtained at the baseline BDES was categorized as < 12 years, 12 years, and > 12 years of school completed. At the baseline EHLS, the type of work done the longest was coded by 1980 census classifications; homemakers were coded with service occupations. Household annual incomes from 1998-2000 (EHLS2), were used. At EHLS3, participants were asked: “When you were age 13, how many people lived in your home?” and “When you were age 13, how many bedrooms were in your home?” Childhood household density was defined as the number of people living in the home divided by the number of bedrooms in the home, then split at the median. Participants were also asked if at age 13, their parents rented or owned their home.
High resolution B-mode carotid artery ultrasound images (Biosound AU4, Biosound Esaote, Indianapolis, IN USA) were obtained at EHLS2 by a certified sonographer using a modification of the Atherosclerosis Risk In Communities (ARIC) study protocol (Bond, 1991
). The areas of focus were the 1 cm of the distal common carotid artery (CCA) closest to the bifurcation, the bifurcation, and 1 cm of the proximal internal carotid artery (ICA) closest to the flow divider. IMT and plaque measurements were made by certified graders using a custom program interfaced with ImagePro software (Image Pro Plus version 4.1. Media Cybernetics, Silver Spring, MD USA), and a modified ARIC protocol (Riley, 1991
). Mean IMT was defined as the mean of the near and far walls of the CCA, the bifurcation and the ICA on both the left and right sides. Mean inter-grader difference in IMT was 0.03 mm. Plaque was determined by evaluating change in wall shape, change in wall texture, and wall thickness (greater than or equal to 1.5 mm). Plaque was considered present if one of these was present with acoustic shadowing, or two were observed in an area without acoustic shadowing. The number of sites (left and right CCA, ICA, and bifurcation) with plaque was categorized: 0, 1-3 and 4-6 sites. Inter-grader reliability for plaque was excellent; kappas averaged 0.76 and percent agreement averaged 90%. IMT measures were available for 2,042 (73%) participants. Ultrasounds were not obtained for participants unable to come to the central site, unable to lay flat, or when the sonographer was unavailable.
Cardiovascular risk factors were measured at the same study phase as IMT measurements. History of CVD was a self-reported physician-diagnosed stroke, myocardial infarction (MI), or angina. Family history of CVD was positive if either parent had had a stroke or MI. Blood pressure was measured using the Hypertension Detection and Follow-up Protocol (HDFP, 1976
). Hypertension was defined as systolic blood pressure ≥140 mmHg, diastolic blood pressure ≥90 mmHg, or physician-diagnosed hypertension and current use of hypertension medication. Diabetes status was defined as self-report of physician-diagnosed and treated diabetes or elevated glycated hemoglobin levels using age- and sex- adjusted norms (Klein, 1992
). Body mass index (BMI) was calculated as weight in kilograms divided by height in meters squared, with obesity defined as BMI ≥ 30 kg/m2
. Total and HDL cholesterol were measured using reflectance spectrophotometry. Pack-years were calculated for smokers. History of heavy drinking meant having consumed four or more alcoholic beverages daily at any point in life.
All analyses were performed with SAS version 9.1 (SAS Institute, Cary NC). Participants with and without IMT measurements were compared using chi-square, Cochran-Mantel-Haenszel, t-test and linear regression procedures. Multivariable least squares regression was used to test for associations between SES indicators and IMT. Ordinal logistic regression (the cumulative odds model) was used to estimate odds ratios and examine the associations between SES indicators and plaque score. Model 1 adjusted for age and sex, and model 2 adjusted for age, sex and traditional CVD risk factors (family history of CVD, prevalent CVD, diabetes status, hypertension, HDL and non-HDL cholesterol, pack-years of smoking (non-smokers coded as 0), and history of heavy drinking) in order to determine if SES associations were independent of these known risk factors. Analyses were repeated in a subgroup of participants without CVD, to examine associations in the healthiest participants. Based on the results of the separate childhood and adult SES analysis, an accumulation of risk model (Ben-Schlomo, 2002
; Roswall, 2002
) was built using the combined “exposures” of childhood household density and education as they were the most consistent childhood and adult SES indicators respectively in their associations with carotid IMT and plaque. An ordinal variable with six categories was created according to the possible groupings of 2 levels of density split at the median value (≤1.5 people/bedroom, >1.5 people/bedroom), and 3 levels of education (<12 years, 12 years, >12 years) with the group having low density and >12 years of education as the referent group.