Coronary Artery Calcium Progression Is Associated with Cardiovascular Events Among Asymptomatic Individuals: From the North Texas Primary Care Practice-based Research Network (NorTex-PBRN)

Study Population

The North Texas Healthy Heart (NTHH) study is a prospective cohort study involving a convenience sample of 571 subjects recruited from 12 sites participating in the North Texas Primary Care Practice-based Research Network (NorTex) from April 2006 to May 2008. NorTex is a collaborative network of primary care clinics serving low-income, underrepresented populations of the Dallas/Fort Worth, Texas, metropolitan area. The 12 family medicine/internal medicine clinic sites that participated in the NTHH study included 4 academic community-based clinics, 3 community health centers, 4 solo-practitioner private practices, and 1 federally qualified health center. Participants were eligible for the study if they were aged >44 years; self-identified as non-Hispanic white, non-Hispanic African American, or Hispanic/Latino; and had no history of self-reported CVD (coronary artery disease, peripheral arterial disease, history of myocardial infarction or stroke, or congestive heart failure), renal failure, or liver failure. All participants were consecutively screened for eligibility either on site or via phone from a centralized NorTex research office located within the Department of Family Medicine at the University of North Texas Health Science Center. Initial contact was made with 1062 individuals, of whom 860 met eligibility criteria. Of those who were eligible, 670 individuals were invited and 571 agreed to participate (85% recruitment rate). A total of 364 participants completed the follow-up annual telephone calls and an in-person study visit at 2 years. The analysis presented here includes the 311 participants who completed multislice computed tomography (MSCT) of the chest at baseline and 2 years. This represents a 58% follow-up rate over 2 years. All study procedures were approved by the University of North Texas Health Science Center and John Peter Smith Health Network institutional review boards.

Study Procedures

All participants who consented to participate underwent an initial 1-hour private, in-person interview. Women without a history of hysterectomy underwent a urine pregnancy test; none of these were positive. Participants then completed weight, height, waist-to-hip circumference, and blood pressure measures at both visits. Height and weight measurements were used to calculate a body mass index (BMI) for each subject using the Quetelet equation (kilograms divided by square meters).¹⁸ Automated Welch Allyn sphygmomanometers were used to measure heart rate and systolic and diastolic blood pressures in each arm using a appropriately sized cuff. The measures were taken after the participant had been seated quietly for 5 minutes with both feet flat on the floor and the back comfortably supported. Mean heart rate and systolic and diastolic blood pressure were calculated for each subject based on 2 separate measures. These procedures were repeated at the 2-year visit.

Demographic and Health Behavior Measures

The NTHH study used standardized questions from the Behavioral Risk Factor Surveillance System to collect select demographic and health behavior information. Age (years) was registered as a continuous variable. Race/ethnicity was self-reported and categorized as non-Hispanic white, non-Hispanic African American, Hispanic, and other.

Physiologic and Clinical Measures

The presence of coronary calcium was measured using 16-slice MSCT, and Agatston units¹⁹ were calculated. The MSCT took images every 3 mm from the carina to the base of the heart, and double inspiration was used to minimize motion artifacts. Electrocardiography correlated heart rate with the images produced by Vitrea software (Vital Images (Minnetonka, MN) was used to quantify calcium. Total time in the scanner was 10 to 15 minutes, which was open with no contrast administered. CAC quantification was reviewed and interpreted by a radiologist from Radiology Associates at the Center for Diagnostic Imaging at the University of North Texas Health Science Center; this radiologist was blinded to participant characteristics. These procedures were repeated at the 2-year visit, and the differences from baseline CAC measures were calculated for each participant. Because the resulting values were skewed and transformation attempts failed to normalize the variable, CAC differences were categorized as increase in CAC (CAC difference >0) or no change or improvement in CAC (CAC difference ≤0). In addition, each participant was assessed yearly on whether they had suffered a myocardial infarction, stroke, transient ischemic attack, new-onset heart failure, peripheral arterial disease, or any other new heart condition. A composite CVD event was coded as having none or any of these events. CAC and CVD composite outcomes functioned as our dependent variables.

Other clinical factors included hypertension, diabetes, and lipid status. After patients had fasted for 8 hours, blood was collected for serum chemistries and analyzed using a commercial laboratory. Hypertension was considered present if blood pressure was ≥140/90 mmHg, the subject reported being diagnosed with hypertension, or the subject was taking antihypertensive medications. Diabetes was considered present if the fasting glucose concentration was ≥126 mg/dL, the subject reported being previously diagnosed with diabetes, or the subject was taking any diabetic medication. Dyslipidemia was considered to be present if, based on Adult Treatment Panel III guidelines, the participant had an LDL ≥160 mg/dL, the participant reported being previously diagnosed with high cholesterol, or the participant was taking a lipid-lowering medication. Because the study was conducted from 2006 to 2008, Adult Treatment Panel III guidelines were used; these define LDL concentrations ≥160 mg/dL as high or very high.

Primary independent variables included the presence of hypertension, dyslipidemia, diabetes, or any combination of these chronic conditions and baseline values of LDL, non-HDL, HDL, triglycerides, systolic and diastolic blood pressures, fasting glucose, and BMI. We contemplated assessing the mean of or difference in physiologic measures between clinic visits, but variability at the time of measurement limited the reliability and clinical significance of such measures.

Sample Size and Statistical Analyses

A retrospective sample size calculation for cohort study designs was conducted based on data from the Multi-Ethnic Study of Atherosclerosis (MESA) published by Budoff et al.²⁰ The MESA study found that just more than 50% of its analytic cohort (n = 6778) had CAC progression with a median follow-up of 7.6 years. Using a relative risk of 1.3 in the likelihood of developing coronary heart disease from progressive CAC changes, as found in the MESA study, we would achieve a desired power of 80% with a sample size of 167. These analyses are, in fact, based on a sample of 311 participants, ensuring sufficient power was achieved.

All statistical analyses were performed using SPSS version 21.0 (IBM, Chicago, IL). Descriptive statistics are provided for all variables. Counts and frequencies are provided for categorical data, and means with standard deviations (SDs) are provided for continuous variables. Independent sample t tests and χ² analyses were performed to test for differences in independent variables between participants with CAC progression. Differences between CAC progression and a composite CVD event were also assessed. Logistic regression was performed to assess associations between independent variables, CAC progression, and composite CVD events. Unadjusted and adjusted odds ratios and 95% confidence intervals were calculated. Statistical significance was assessed at the α = 0.05 level.

The multiple logistic regression model assessed for multicollinearity using tolerance and variation inflation factor, with age and sex variables in the final models. No collinear relationships were identified.