Although a coronary artery calcium (CAC) score of 0 is associated with a very low 10-year risk for cardiac events, this risk is nonzero. Subjects with a family history of coronary heart disease (CHD) has been associated with more subclinical atherosclerosis than subjects without a family history of CHD. The purpose of this study was to assess the significance of a family history for CHD in subjects with a CAC score of 0. The Multi-Ethnic Study of Atherosclerosis cohort includes 6,814 participants free of clinical cardiovascular disease (CVD) at baseline. Positive family history was defined as reporting a parent, sibling, or child who had a heart attack. Time to incident CHD or CVD event was modeled using the multivariable Cox regression; 3,185 subjects were identified from the original Multi-Ethnic Study of Atherosclerosis cohort as having a baseline CAC score of 0 (mean age 58 years, 37% men). Over a median follow-up of 10 years, 101 participants (3.2%) had CVD events and 56 (1.8%) had CHD events. In age- and gender-adjusted analyses, a family history of CHD was associated with an ∼70% increase in CVD (hazard ratio 1.73, 95% confidence interval 1.17 to 2.56) and CHD (hazard ratio 1.72, 95% confidence interval 1.01 to 2.91) events. CVD events remained significant after further adjustment for ethnicity, risk factors, and baseline medication use. In conclusion, asymptomatic subjects with a 0 CAC score and a positive family history of CHD are at increased risk for CVD and CHD events compared with those without a family history of CHD, although absolute event rates remain low.
Highlights
- •
A family history of coronary heart disease is a potent risk factor.
- •
We investigated its significance in those with a coronary artery calcium score of 0.
- •
A 70% increase in cardiovascular events was noted in those with a family history.
- •
Family history is a potent risk factor, even in those with coronary artery calcium scores of 0.
Although the presence and extent of coronary artery calcium (CAC) is directly proportional to subsequent cardiovascular events, a 0 CAC score is associated with a very low cardiac event rate of about 1% over a 10-year period. However, data from clinical trials and observational studies have documented the presence of noncalcified plaque by cardiac computed tomographic angiography with varying extent and severity in subjects with a 0 CAC score, with a prevalence ranging from 4% to 24% depending on the population studied and symptom status. Subclinical noncalcified atherosclerotic plaque is more prevalent in men and those with diabetes, active smoking, a family history of premature coronary heart disease (CHD), or those presenting with chest pain symptoms. These data suggest that although the absolute 10-year risk is low for a cardiovascular event in asymptomatic patients with a CAC score of 0, there may be subsets in whom risk is higher and for whom the overall predicted risk may be a misattribution. Data addressing this question in asymptomatic patients with a CAC score of 0 and a positive family history of CHD have not been published to the best of our knowledge. We, therefore, investigated the CHD and cardiovascular disease (CVD) event rates in a diverse population of subjects with a 0 CAC score at baseline comparing those with versus without a family history of CHD. We hypothesized that after adjustment for standard risk factors, the observed event rates would be higher in the group with a positive family history.
Methods
The Multi-Ethnic Study of Atherosclerosis (MESA) is a longitudinal, population-based study of 6,814 men and women, initially free of clinical CVD, aged 45 to 84 years at baseline, and recruited from 6 field centers: Baltimore, Maryland; Chicago, Illinois; Forsyth County, North Carolina; Los Angeles, California; New York, New York; and St. Paul, Minnesota. Specific racial/ethnic groups enrolled included white, black, Hispanic, and Chinese. Approximately 50% of the participants enrolled were women. Details of the MESA recruitment strategy are published elsewhere. The baseline visit took place from July 2000 to September 2002. MESA was approved by institutional review boards at each site, and all participants gave written informed consent. The design of the study has been described in detail previously.
Information was obtained at the baseline examination in 2000 to 2002 regarding age, gender, ethnicity, and medical history by questionnaires. At this examination, participants also reported the presence or absence of a family history of CHD. The family history of CHD was obtained by asking participants whether any parent, sibling, or child had experienced a heart attack. The age of the family member at the time of their heart attack was not obtained during the baseline examination and, therefore, was not available for analysis. Current smoking was defined as having smoked in the last 30 days, whereas former smoker was defined as a subject who is not currently smoking but had smoked ≥ 100 cigarettes in his or her lifetime. Diabetes was defined as a fasting glucose ≥ 126 mg/dl or on hypoglycemic medication. Use of antihypertensive and other medications was based on clinic staff entry of prescribed medications. Blood pressure at rest was measured 3 times in the seated position using a Dinamap model Pro 100 automated oscillometric sphygmomanometer (Critikon, Tampa, Florida), and the average of the second and third readings was recorded. Hypertension was defined as a systolic blood pressure ≥140 mm Hg, diastolic blood pressure ≥90 mm Hg, or use of medications together with a self-reported diagnosis of high blood pressure. Total and high-density lipoprotein cholesterol and triglyceride levels were measured from blood samples obtained after a 12-hour fast. Low-density lipoprotein cholesterol was calculated with the Friedewald equation. The Framingham Risk Score for men and women was calculated on the basis of age, total cholesterol, high-density lipoprotein cholesterol, current smoking status, high blood pressure, and the use of antihypertensive medications.
CAC was assessed by chest computed tomography using either a cardiac-gated electron-beam computed tomography scanner (Chicago, Los Angeles, and New York field centers) or a multidetector computed tomography scanner system (Baltimore, Forsyth County, and St. Paul field centers). Certified technologists scanned all participants twice over phantoms of known physical calcium concentration. A radiologist or cardiologist read all computed tomography scans at a central reading center (Los Angeles Biomedical Research Institute at Harbor—UCLA in Torrence, California). We used the average Agatston score for the 2 scans in all analyses. Carr et al have reported the details of the MESA computed tomography scanning and interpretation methods.
At the time of analysis, the cohort had been followed for incident CVD and CHD events for a median follow-up of 10 years. At intervals of 9 to 12 months, a telephone interviewer contacted each participant to enquire about interim hospital admissions, cardiovascular outpatient diagnoses, and deaths. Trained personnel abstracted medical records suggesting possible cardiovascular events. Two physicians independently classified the events and assigned incidence dates. If, after review and adjudication, disagreements persisted, a full mortality and morbidity review committee made the final classification. For the purposes of this study, we used all CVD and CHD events as the end point. Specifically, the end point included myocardial infarction, CHD death, resuscitated cardiac arrest, angina, stroke, stroke death, or other CVD death for total CVD events and myocardial infarction, CHD death, and resuscitated cardiac arrest and angina for total CHD events.
There were 3,416 participants with a CAC score of 0 at baseline. Of these, 183 were missing data for family history, 17 were removed because they lacked follow-up time and an additional 31 participants were removed because of missing covariate information. Values from the baseline examination were used for the covariates. Only the first CVD or CHD event for each participant was considered. The annual incident rates of first CVD and CHD events, for subjects with and without family history, were estimated with Poisson rate models. The family history groups were compared with rate ratios. Robust standard error was used. Kaplan-Meier graphs compared the cumulative probability distributions over time between family history groups. Differences between the curves were tested using nonparametric log-rank and Peto-Peto tests. Cox proportional hazards models were used to estimate the effect of family history on time to incident CVD or CHD, adjusted for other covariates. Controlling covariates included age, gender, race, Framingham Risk Score, and use of baseline medications (statins and aspirin). The proportional hazards assumption was tested with Schoenfeld residuals and time-interacting covariates. The number needed to treat (NNT) for statin and aspirin was calculated separately using the subset of participants who were not on the medication at examination 1, using a 10-year time interval and relative risk estimates for statins and aspirin from the Cochrane meta-analyses. Because some participants were started on statins or aspirin at a later date, the risk ratios were adjusted to compensate for the fraction of the exposure time on the medications. The 10-year NNT was then rescaled to 5 years. All statistical analyses were performed using STATA 12.0 (StataCorp 2011, College Station, Texas) and SAS 9.3 for Windows (SAS Institute Inc., Cary, North Carolina).
Results
At the time of analysis, a total of 3,185 subjects had a baseline CAC score of 0 and complete data, thus comprising the study cohort for this analysis (mean age 58 years, 37% men). There were 1,185 (37%) subjects with a self-reported family history of CHD in a first-degree relative. Table 1 shows the baseline characteristics according to the presence or absence of a family history of CHD. The 2 groups had a similar median age, comparable frequencies of most CVD risk factors, and comparable Framingham Risk Scores. However, the group with a positive family history of CHD had a lower percentage of men and differed in ethnicity (more Whites and less Chinese Americans). Despite no differences in the frequency of dyslipidemia or Framingham Risk Scores, baseline use of statins and aspirin were more common in those with a family history of CHD.
| Variable | Baseline CAC Score = 0 | p-Value | |
|---|---|---|---|
| FamHx + (n = 1185) | FamHx − (n = 2000) | ||
| Mean age (years) | 58 | 58 | 0.27 |
| Men | 370 (31%) | 799 (40%) | <0.0001 |
| Ethnicity | <0.0001 | ||
| White | 477 (40%) | 584 (29%) | |
| Chinese | 68 (6%) | 303 (15%) | |
| African-American | 373 (31%) | 609 (30%) | |
| Hispanic | 267 (23%) | 504 (25%) | |
| Dyslipidemia ∗ | 336 (28%) | 515 (26%) | 0.11 |
| Smoker | 0.09 | ||
| Never | 637 (54%) | 1142 (57%) | |
| Former | 372 (31%) | 608 (30%) | |
| Current | 176 (15%) | 250 (13%) | |
| Hypertension † | 452 (38%) | 658 (33%) | 0.003 |
| Diabetes mellitus | 102 (9%) | 182 (9%) | 0.64 |
| Impaired fasting glucose | 144 (12%) | 230 (12%) | 0.58 |
| Mean Framingham Risk Score | 6.1% | 6.2% | 0.84 |
| Baseline statin use | 139 (12%) | 174 (9%) | 0.005 |
| Baseline aspirin use | 256 (22%) | 309 (16%) | 0.000 |
∗ Dyslipidemia—based on adult treatment panel (ATP) III definitions of total and/or low-density cholesterol.
† Hypertension—systolic blood pressure >140 mm Hg, diastolic blood pressure > 90 mm Hg, or use of medications together with a self-reported diagnosis of high blood pressure.
Over the median follow-up period of 10 years, 101 (3.2%) participants had CVD events and 56 (1.8%) had CHD events. In subjects with a positive family history for CHD, 51 of 1,185 (4.3%) experienced CVD events (vs 2.5% for negative family history), whereas 28 of 1,185 (2.4%) had CHD events (vs 1.4% for negative family history). The annual incident event rate in the group with a positive family history for CHD compared with those with negative family history was 0.44% versus 0.26% for CVD events (rate ratio 1.70, p = 0.007) and 0.24% versus 0.14% (rate ratio 1.67, p = 0.056) for CHD events. Figures 1 and 2 illustrate the cumulative probabilities for CVD and CHD events in the subjects with a positive family history versus negative family history. There was a statistically significant difference in the cumulative probability for CVD events, whereas CHD events were marginally nonsignificant.
Following adjustment for age and gender, a family history of CHD was significantly associated with both CVD (hazard ratio [HR] 1.73, 95% confidence interval [CI] 1.17 to 2.56) and CHD (HR 1.60, 95% CI 1.08 to 2.38) events. After additional adjustment for ethnicity, Framingham Risk Score, and baseline use of aspirin or statin, a family history of CHD was only significantly associated with CVD events (HR 1.72, 95% CI 1.01 to 2.91).
Of the 3,185 subjects in this analysis, only 3,167 had complete data involving aspirin use. A separate analysis involving this smaller cohort revealed no important differences in outcomes. Because of the limited number of events in the CAC score = 0 cohort, the Framingham Risk Score was chosen as a surrogate covariate. A separate analysis involving the major individual risk factors revealed no important differences ( Table 2 ). A model involving smoking, hypertension, diabetes, diabetes by time interaction, statin, and aspirin medication revealed no significant change in the hazard for CVD events (HR 1.54, 95% CI 1.04 to 2.30). Because of the limited number of CHD events, a similar model involving the individual risk factors and medications was not possible. Hypertensive medications were also added as a controlling covariate with no important changes in the HRs for CVD and CHD events. The new cardiovascular risk score was also substituted for the Framingham Risk Score in the fully adjusted model with no important differences in events noted (HR 1.58, 95% CI 1.06 to 2.34, for CVD and HR 1.54, 95% CI 0.90 to 2.62, for CHD). The potential for interaction between the effect of family history and time on the hazards for CVD and CHD events was also investigated with no significant interaction noted (p = 0.83 for CVD and p = 0.62 for CHD). Gender and ethnic differences were also investigated, with no important differences in either CVD or CHD events for gender. For ethnicity, Caucasians (relative to Chinese-Americans) were found to have a slightly higher hazard for CVD events, which was marginally significant at a p = 0.04.
