Racial Differences in the Association of Serum 25-Hydroxyvitamin D Concentration With Coronary Heart Disease Events

Study Population

The Multi-Ethnic Study of Atherosclerosis (MESA) is a multicenter, community-based prospective cohort study of clinical and subclinical cardiovascular disease.²¹ Each center’s institutional review board approved the study, and all participants provided informed consent. From 2000 to 2002, MESA enrolled 6814 adults aged 45 to 84 years from 6 field centers (New York and Bronx counties, New York; Baltimore and Baltimore County, Maryland; Forsyth County, North Carolina; Chicago, Illinois; St Paul, Minnesota; and Los Angeles, California). By design, MESA recruited a study population that was 38% white, 28% black, 22% Hispanic, and 12% Chinese. MESA excluded individuals who had prevalent clinical cardiovascular disease, defined as myocardial infarction, angina, stroke, transient ischemic attack, heart failure, atrial fibrillation, use of nitroglycerin, prior angioplasty, coronary artery bypass graft surgery, valve replacement, pacemaker or defibrillator implant, or any surgery on the heart or arteries.

We restricted our study population to 6476 participants for whom we measured serum concentrations of 25(OH)D at the baseline MESA examination. We excluded 6 participants with serum 25(OH)D concentration suggestive of high-dose vitamin D supplementation (>100 ng/mL; to convert to nanomole per liter, multiply by 2.496) and 34 participants without follow-up data, resulting in a final sample size of 6436 (94% of all MESA participants).

Measurement of Serum 25(OH)D Concentration

Sera samples were collected at the baseline MESA exam in 2000–2002, after an overnight fast. Samples were stored at −80°C until they were shipped to the University of Washington for analysis in 2011–2012. Total 25(OH)D (sum of 25-hydroxyvitamins D₂ and D₃) was measured using high-performance liquid chromatography– tandem mass spectrometry. Calibration was confirmed with National Institute of Standards and Technology’s standard reference material 972.²² The interassay coefficient of variation was 4.4% at 10.4 ng/mL with a lower limit of detection of 2.0 ng/mL for 25(OH)D₃ and a coefficient of variation of 4.4% at 9.4 ng/mL with a lower limit of detection of 0.5 ng/mL for 25(OH)D₂. 25-hydroxyvitamin D is stable during long-term storage at −80°C.²³

CHD Events

At intervals of 9 to 12 months, a telephone interviewer contacted each participant to inquire about interim hospital admissions, cardiovascular outpatient diagnoses, and deaths. Self-reported diagnoses were verified via death certificate and medical record review, as previously described.²⁴ Cardiovascular disease outcomes were adjudicated by the MESA events committee, which included cardiologists, physician epidemiologists, and neurologists.²¹ Incident CHD, used as the primary outcome for this study and other MESA studies,²⁵ was assessed until May 2012 and defined as the first occurrence of any 1 of the following events: myocardial infarction, definite or probable angina, resuscitated cardiac arrest, or CHD death. Definite angina was defined as symptoms of typical chest pain or atypical symptoms, followed by 1 or more additional criteria, including coronary artery bypass graft surgery or other revascularization procedure, 70% or greater obstruction on coronary angiography, or evidence of ischemia by stress tests or by resting electrocardiogram. Probable angina required symptoms of typical chest pain or atypical symptoms, a physician’s diagnosis of angina, and medical treatment for it. As a secondary outcome, we used a more restrictive definition of CHD that included only incident myocardial infarction, CHD death, and resuscitated cardiac arrest, which we labeled hard CHD events.²⁶

Covariates

All covariates were ascertained at the baseline MESA examination, concurrent with 25(OH)D measurements. Participants completed self-administered questionnaires, interviewer-administered standardized interviews, and extensive in-person examinations yielding demographic and lifestyle characteristics, medical history, anthropometric measurements, and laboratory data. Race/ethnicity was characterized on the basis of participants’ responses to questions modeled from the year 2000 US Census. General health was self-reported on a questionnaire as excellent, very good, good, fair, or poor. Leisure-time physical activity was estimated as the total amount of intentional exercise performed in a usual week and measured in metabolic equivalent task–minutes. Participants were asked to report the average frequency of consumption of specific food items over the previous year using a 120-item food frequency questionnaire. Dietary vitamin D intake was estimated by multiplying the frequency and serving size for each food consumed by the vitamin D content of that food (Nutritional Data Systems for Research). Data regarding vitamin D supplement use were not available. Body mass index (BMI) was calculated as weight in kilograms divided by height in meters squared. Diabetes was defined as a fasting blood glucose concentration of 126 mg/dL (to convert to mmol/L, multiply by 0.0555) or more or the use of insulin or oral hypoglycemic medications. Blood pressure was ascertained as the mean of the last 2 of 3 seated measurements. Total and high-density lipoprotein cholesterol were measured using the cholesterol oxidase method. Low-density lipoprotein cholesterol levels were calculated using the Friedewald equation. Intact serum parathyroid hormone was quantified using a 2-site immunoassay on a clinical analyzer.²⁷ Glomerular filtration rate was estimated from serum creatinine using the Chronic Kidney Disease Epidemiology Collaboration equation.²⁸ Chronic kidney disease was defined as a glomerular filtration rate less than 60 mL/min/1.73 m².

Statistical Analyses

Mean annual serum 25(OH)D concentration was estimated for each participant using a cosinor model previously developed and validated in MESA.²⁰ The cosinor model assumes that 25(OH)D concentration follows a sine wave pattern across months of measurement. Using the observed mean amplitude of 25(OH)D excursion in the MESA population, each participant’s mean annual 25(OH)D concentration was estimated from the participant’s single baseline 25(OH)D measurement and the month during which it was obtained. Race is not included in the cosinor model because its inclusion did not improve prediction performance. Annualized 25(OH)D concentration was evaluated using clinically relevant categories that have been previously published: 30 ng/mL or more (reference group), 20–29 ng/mL, and less than 20 ng/mL.^18,29,30 Because the distribution of serum 25(OH)D concentration and perhaps the functional form of its association with CHD risk was expected to vary by race/ethnicity group, we also evaluated annualized 25(OH)D using race-specific quintiles.

Participants were considered at risk of incident CHD events from the date of their baseline MESA examination until the first occurrence of the composite outcome or until their data were censored due to death from non-CHD cause (n = 217), loss to follow-up, or the end of available follow-up (May 2012). Functional forms of the race-specific associations of 25(OH)D with CHD risk were examined graphically using adjusted penalized smoothing splines with evenly spaced knots, among the inner 99% of annualized 25(OH)D concentrations.³¹ We used the Cox proportional hazards regression model to estimate the relative hazard of CHD, adjusting for covariates selected prior to analyses based on known or suspected biologic relationships. The first model included demographic data, including age, race/ethnicity, sex, and study site. The second model added likely potential confounding variables, including BMI (continuous variable), smoking status (never, former, current), educational attainment (high school, some college, completed college), total gross family income in the past 12 months, self-reported health status (poor, fair, good, very good, excellent), nutritional vitamin D intake (log-transformed continuous variable), and leisure-time physical activity level (log-transformed continuous variable). The third model added variables that may possibly confound or mediate the associations of interest, including diabetes status, chronic kidney disease status, systolic blood pressure, serum concentrations of high-density lipoprotein, low-density lipoprotein, triglycerides, parathyroid hormone, serum C-reactive protein (log-transformed), and the use of antihypertensive and lipid-lowering medications.

Approximately 5% or less of the study participants were missing data on education, income, physical activity, smoking status, serum C-reactive protein, systolic blood pressure, antihypertensive medication use, or lipid concentrations. For the regression analyses, these participants’ values were multiply imputed using chained equations.³² The multiple analyses over the imputations were combined using Rubin rules to account for the variability in the imputation procedure.³³

Subgroup-specific hazard ratios (HRs) were calculated via linear combination of regression coefficients for main effect and cross-product terms. Interactions were tested by the Wald test of a product term for 25(OH)D concentration and race/ethnicity categories. Analyses completely stratified by race/ethnicity yielded similar results.

All analyses were conducted with Stata 11.2 (StataCorp). The nominal level of significance was defined as P < .05 (2-sided).

Baseline Characteristics

At baseline, the mean age was 62 years and 53% of participants were women. Mean (SD) serum 25(OH)D concentration was 25.5 (10.6) ng/mL and varied substantially by race/ethnicity: 30.1 (10.6) ng/mL for white participants (n = 2501), 26.7(8.3) for Chinese participants (n = 784), 19.2 (9.0) for black participants (n = 1750), and 24.6 (9.4) for Hispanic participants (n = 1401) (P < .001). Participants with lower serum 25(OH)D concentrations were less likely to be white, were more likely to smoke or have diabetes or chronic kidney disease, and were characterized by larger body mass, higher serum parathyroid hormone concentration, and lower physical activity levels (Table 1). Race/ethnicity was associated with differences in baseline characteristics, particularly for the level of education completed, gross family income, self-reported health, prevalence of diabetes, and amount of intentional exercise (eTable 1 in Supplement).

CHD Events

During amedian follow-up of8.5 years (interquartile range, 7.6–8.6 years), there were 361 occurrences of the composite clinical end point (incidence rate: 7.38 events per 1000 person-years). The qualifying event was myocardial infarction for 139 articipants, CHD death for 46 participants, resuscitated cardiac arrest for 13 participants, and angina for 163 participants (Table 2).

25(OH)D and CHD Events

Unadjusted incidence rates and adjusted risks of incident CHD were higher with lower baseline serum 25(OH)D concentration (Table 3). However, we observed significant heterogeneity in the association of 25(OH)D with CHD risk by race/ethnicity, with P values for interaction < .05 across race/ethnicity groups for each model (Table 3). Lower serum 25 (OH)D concentration was associated with significantly higher risks of CHD among white participants (fully adjusted hazard ratio [HR], 1.26 [95%CI, 1.06–1.49] per 10 ng/mL decrement in 25(OH)D concentration) and Chinese participants (HR, 1.67 [95% CI, 1.07–2.61]). However, there was no evidence of association among black participants (HR, 0.93 [95% CI, 0.73–1.20]) or Hispanic participants (HR, 1.01 [95% CI, 0.77–1.33]). Evaluation of the association of 25(OH)D with CHD risk using splines and race/ethnicity-specific 25(OH)D quintiles similarly revealed that significant associations were limited to white and Chinese participants (Figure and Table 4).

Sensitivity Analyses

For the more restricted hard CHD outcome (n = 216 events), associations of low serum 25(OH)D concentration were of stronger magnitude for white and Chinese participants and were null for black and Hispanic participants (eTable 2 in Supplement). Similar results were observed when mean annual 25(OH)D concentration was replaced with untransformed 25(OH)D concentration and models were adjusted for season of blood draw, when participants with incident CHD events during the first 12 months of follow-up were excluded, and in analyses restricted to participants with good, very good, or excellent self-reported health statuses (eTables 3, 4, and 5 in Supplement).