Coronary artery calcification in black women and white women

Coronary artery calcification in black women and white women Charanjit Khurana, MD,a Christina G. Rosenbaum, MD,a Barbara V. Howard, PhD,c Lucile L. Adams-Campbell, PhD,d Robert C. Detrano, MD, PhD,e Afifa Klouj, MBA,b and Judith Hsia, MDa Washington, DC, and Torrance, Calif

Background Coronary calcification is a potent independent predictor of coronary risk. Sex-specific risk categories based on calcium scores have been established, but ethnic differences in coronary calcification have been little studied. This prospective cohort study compares coronary calcification, assessed by computed tomography, in postmenopausal black women and white women.

Methods and Results Computed tomographic scans were performed on 128 black women and 733 white women without known coronary artery disease (mean age 63 ⫾ 8 years). Although coronary risk factors were more prevalent among black women (P ⬍ .0001), total calcium scores were similar to those in white women. By use of the Framingham algorithm, higher calcium scores were associated with higher 10-year risk of myocardial infarction or coronary death. In multiple regression analysis, age was independently associated with higher calcium scores in both ethnic groups (P ⫽ .002 for black women, P ⬍ .0001 for white women). Diabetes mellitus and not exercising at least 3 times per week were independently associated with higher calcium scores in white women but not black women. Educational level, body mass index, current hormone replacement therapy, hysterectomy, dietary fat consumption, family history of premature coronary disease, hypertension, self-reported high cholesterol, and current smoking were not independently associated with coronary calcium score in black women, white women, or the combined cohort; neither was ethnicity an independent predictor of coronary calcification. Conclusions Despite higher dietary fat consumption, higher body mass index, and greater prevalence of hypertension, diabetes, and smoking, black women had coronary calcium scores similar to those of white women. Ethnicity was not an independent predictor of coronary calcification. (Am Heart J 2003;145:724-9.)

See related Editorial on page 579.

Coronary calcification is a marker for atherosclerosis.1 Computed tomography (CT) is a quick, noninvasive technique for quantification of coronary calcium and, in turn, coronary artery disease risk.2-5 A recent report suggests that CT may have advantages over other established tests such as treadmill testing or nuclear imaging studies for the detection of angiographic coronary artery disease.6 CT may be especially valuable in women, who are less likely than men to report classic symptoms of ischemic heart disease.7

From the aDivision of Cardiology and eBiostatistics Center Medical Center Unit, The George Washington University, bMedStar Research Institute, and cHoward University Cancer Center, Washington, DC, and dHarbor-UCLA Medical Center, Torrance, Calif. Supported by NIH contracts NO1-WH-4-2108 and WH-4-2123. Submitted October 22, 2001; accepted June 6, 2002. Reprint requests: Judith Hsia, MD, 2150 Pennsylvania Ave NW, #4-414, Washington, DC 20037. E-mail: [email protected] Copyright 2003, Mosby, Inc. All rights reserved. 0002-8703/2003/$30.00 ⫹ 0 doi:10.1067/mhj.2003.99

Coronary calcification has been studied predominantly in white men and women. In a subset (283 subjects, 51 black women) of a multiethnic cohort of high-risk, asymptomatic subjects studied with CT for coronary calcium, black women had lower coronary artery calcium scores than white women.8 However, that subset included only 6 black women. In a subsequent analysis of the entire cohort, coronary calcium scores in blacks appeared to confer a comparatively higher risk of coronary events than in whites with similar Framingham risk profiles; however, the group included only 13 black women.9 The age-adjusted coronary heart disease (CHD) mortality rate is lower among black women than men (3.2 vs 4.6 per 1000 person-years). Similarly, among whites, CHD mortality rate is 2.1 for women compared with 5.2 per 1000 person-years for men.10 Coronary calcium scores mimic this relation in that women’s scores are approximately one-half those found in men.11 Differences in the extent, predictive value and determinants of coronary calcification in black women compared with white women remain uncertain. This analysis compares coronary artery calcification in black women

American Heart Journal Volume 145, Number 4

and white women enrolled at two Women’s Health Initiative clinics.

Methods Patient population The Women’s Health Initiative Observational Study (OS) is a cohort of 93,717 postmenopausal women, enrolled at 40 sites, which ultimately will provide 9 to 11 years of follow-up for clinical outcomes including cardiovascular events.12 Women were recruited through a media campaign, direct unsolicited mailings, public notices, and community and medical sources. Potential participants contacted their local clinical center for eligibility screening, often by telephone. Women meeting initial eligibility requirements were invited to a screening visit that included an interview, physical examination, and blood sample collection. OS inclusion criteria were broad to increase the generalizeability of findings: age (50-79 years), postmenopausal status, likely residence in one of the study areas for at least 3 years, and provision of written informed consent. Exclusion criteria included report of a health risk with predicted survival of ⬍3 years, staff judgment of inability to follow the protocol requirements, and active current participation in any randomized trial. Participants at The George Washington University and Howard University/Medstar clinics (n ⫽ 4435), recruited between February 1995 and December 1998, were invited for CT; women who joined this ancillary study provided informed consent in a form approved by the respective institutional review boards. Participants provided data on a wide range of health variables including dietary habits, medical history, and anthropometric measures. Questionnaire measures assessed self-reported hypertension, diabetes mellitus (excluding gestational diabetes), current smoking, high cholesterol requiring pills, hormone replacement therapy, hysterectomy status, and family history of premature coronary disease (father with myocardial infarction at ⱕ55 years or mother with myocardial infarction at ⱕ65). The number of risk factors was determined for each woman by assigning 1 point for each of the following: hypertension, diabetes, high cholesterol, family history of premature coronary disease, and current smoking. Dietary fat consumption was assessed with the use of a food frequency questionnaire based on instruments used in the Women’s Health Trial Vanguard and full-scale studies,13,14 the Working Well Trial,15 and the Women’s Health Trial Feasibility Study in Minority Populations. Nutrient estimates from the food frequency questionnaire are similar to those from shortterm dietary recall and from 4-day food records.16 Women reporting moderate or strenuous exercise for 20 minutes or longer at least 3 times per week were defined as exercising at least 3 times weekly. Educational level was assessed by asking “What is the highest grade in school you finished?” with 11 possible responses ranging from “Didn’t go to school” to “Doctoral degree (PhD, MD, JD, and so forth)”. For multivariate analysis, the education variable was collapsed into 3 categories, with masters or doctoral degree as the referent.

Khurana et al 725

Framingham CHD score Fasting lipid profiles were offered to women at The George Washington clinic as part of this ancillary study. Consequently, Framingham CHD risk scores,17 which incorporate risk associated with cholesterol, smoking, diabetes, age, and blood pressure, could be calculated for 480 predominantly white women. For risk score determination, blood pressure and lipid categorization was performed without regard to medication use.

CT image acquisition and analysis CT was performed on 916 women. Those with prior CHD, defined as myocardial infarction or coronary revascularization, were excluded from this analysis, as were those who reported their ethnic origin as other than black or white. The remaining 861 women constitute the cohort for this analysis. Images were acquired with the use of an Imatron C-150 scanner. Thirty contiguous 3-mm slides (100 ms per slice) were acquired during a single breath hold beginning 1 cm caudad to the carina. Each level was triggered by electrocardiography in end-diastole (80% of R-R interval). Images were obtained with a 30-cm2 field of view (pixel size 0.586 mm). Images were analyzed by the Agatson method.18 Extent of calcification was categorized as described by Hoff et al.19

Statistical analysis Descriptive statistics such as frequencies, percentages, and means and standard deviations were used to describe the study population and to explore the relations between coronary calcium score and several explanatory variables. Group comparisons were made by the t test, ␹2 and, where appropriate, the Cochran-Mantel-Haenszel means test. The relation between coronary calcium scores and Framingham risk scores was evaluated by the Spearman rank correlation. Separate proportional odds models were developed to describe factors associated with coronary calcium score in black women and white women. In a third model, both ethnic groups were combined and ethnicity was included as an independent variable. Calcium scores in these analyses were categorized as 0 to 10, 11 to 100, 101 to 399, or ⱖ400; independent continuous variables were included as linear covariates and categorical variables as indicator levels. Analyses were carried out by means of the SAS System for Windows, version 6.12 (SAS Institute, Cary, NC).

Results The study population included 128 black women and 733 white women (Table I). The cardiovascular risk profile was less favorable among black women, among whom body mass index was higher and hypertension, current smoking, and diabetes mellitus more prevalent. Overall, black women had a significantly higher number of conventional coronary risk factors than did white women (P ⬍ .0001). One or no risk factor was reported by 13% of black women and by 52% of white women; ⱖ3 risk factors were reported by 55% of black women and by 17% of white women.

American Heart Journal April 2003

726 Khurana et al

Table I. Patient characteristics

Table III. Framingham risk in relation to calcium score

Black No. Age (y) BMI (kg/m2) Percent of total calories from fat Hypertension (%) Diabetes mellitus (%) High cholesterol (%) Current smoking (%) Family history premature CHD (%) Current hormone replacement therapy (%) Hysterectomy (%) Exercise 3 times weekly (%)

White

P

128 62.0 ⫾ 6.4 28.3 ⫾ 5.5 27.3 ⫾ 8.1

733 62.8 ⫾ 7.6 25.5 ⫾ 5.6 26.7 ⫾ 7.7

.3 ⬍.0001 .5

53 (41) 7 (5) 21 (16) 11 (9) 16 (13)

143 (20) 10 (1) 98 (13) 32 (4) 87 (12)

⬍.0001 .002 .4 .04 .8

46 (37)

446 (62)

⬍.0001

71 (55) 80 (62)

236 (32) 401 (55)

⬍.0001 .1

Continuous variables, mean ⫾ SD. BMI, Body mass index.

Table II. Severity of coronary calcification in black and white women

Age (y) 50-59

60-69

ⱖ70

Coronary calcium score (Agatston units)

Black, no. (%)

White, no. (%)

0-10 11-100 101-399 ⱖ400 0-10 11-100 101-399 ⱖ400 0-10 11-100 101-399 ⱖ400

44 (85) 7 (13) 0 (0) 1 (2) 41 (66) 9 (15) 8 (13) 4 (6) 3 (21) 3 (21) 7 (50) 1 (7)

271 (81) 48 (14) 13 (4) 3 (1) 163 (68) 48 (18) 20 (8) 15 (6) 62 (39) 41 (26) 29 (18) 26 (16)

P .5

.6

.3

Duration of antihypertensive treatment was similar in the 2 groups: 5.9 ⫾ 6.4 years (mean ⫾ SD) for black women and 5.1 ⫾ 5.7 years for white women. Of women reporting hypertension, 76% were taking antihypertensive medication; the number of antihypertensive medications used was similar among black women and white women (1.3 ⫾ 0.5 vs 1.3 ⫾ 0.6, respectively). Onset of diabetes before age 30 years was reported by 2 white women, age 30 to 39 years by 2 black women and 3 white women, age 40 to 49 years by 2 black women and 4 white women, and age ⱖ50 by 3 black women and 1 white woman. Black women were more likely to be current smokers and reported more years of lifetime smoking: ⬍10 years for 73 of 128 black women (56%) and 502 of 733 white women (69%), 10 to 29 years for 27% and 22%, respectively, and ⱖ30 years for 16% and 10%, respectively (P ⫽ .02).

Calcium score

No.

Framingham points

10-year CHD risk

0-10 11-100 101-399 ⱖ400

327 81 41 28

4.9 ⫾ 3.0 5.6 ⫾ 3.1* 7.3 ⫾ 2.6* 6.7 ⫾ 3.0†

5% 6% 7% 7%

*P ⬍ .05 versus calcium score 0-10. †P ⫽ .002 versus calcium score 0-10.

Women in both ethnic groups were highly educated, although white women reported more years of education (P ⫽ .0003 across the 11 response categories). A college degree or education beyond college was reported by 54% and 75% of black women and white women, respectively. Doctoral degrees were reported by 5% of black women and 11% of white women. Black women were more likely to have undergone hysterectomy (55% vs 32%, P ⬍ .0001) but less likely to report current hormone replacement therapy (37% vs 62%, P ⬍ .0001). Age at hysterectomy was similar for the 2 groups, with 27% having hysterectomy before age 40 years, 50% from age 40 to 49 years, and 21% at age ⱖ50 years. Black women and white women reported bilateral oophorectomy with similar frequency (38% vs 44%), although 32% either did not indicate or reported uncertainty about their oophorectomy status. Most women taking hormone replacement therapy were taking conjugated estrogens either alone (53% of black women and 30% of white women) or with a progestin (24% and 49%, respectively). Use of other oral estrogens was reported by 12% of black women and 10% of white women; use of other estrogens with a progestin was reported by no black women but by 11% of white hormone users. Estrogen patch and cream use was infrequent (1% and 2% of women, respectively). Coronary calcium scores were similar in black women and white women (Table II). Framingham risk scores for the subsample of 477 women with measured fasting lipid profiles increased with severity of coronary calcification (Table III). Overall, 10-year risk of myocardial infarction or coronary death in this group of healthy volunteers was low, averaging 5%. The Framingham score was positively correlated with the total calcium score (P ⬍ .0001). Characteristics associated with coronary calcification in black women and white women were evaluated in separate multiple regression models (Table IV). Age was independently associated with calcium score in both ethnic groups. Self-reported diabetes mellitus was associated with higher coronary calcium scores in white women and in the combined cohort. Failure to

American Heart Journal Volume 145, Number 4

Khurana et al 727

Table IV. Determinants of coronary calcium score in black and white women Black

Age Education Less than a college degree College degree Masters or doctoral degree Hypertension Diabetes mellitus High cholesterol Current smoking Family history premature CHD Number of coronary risk factors Current hormone replacement therapy Hysterectomy Percent of total calories from fat Exercise ⬍3 times weekly BMI Ethnicity

White

All women

OR (95% CI)

P

OR (95% CI)

P

OR (95% CI)

P

1.13 (1.05-1.22)

.002 .5

1.11 (1.08-1.13)

⬍.0001 .5

1.11 (1.08-1.13)

⬍.0001 .5

4.89 (1.38-17.25) 1.62 (0.60-4.38) 1.00 0.82 (0.14-4.74) 0.83 (0.09-7.82) 1.68 (0.56-5.10) 0.29 (0.04-2.22) 0.73 (0.11-4.71) 2.60 (0.65-10.40) 1.14 (0.44-2.96) 1.19 (0.48-2.96) 0.96 (0.92-1.03) 1.13 (0.47-3.20) 1.00 (0.98-1.02)

.8 .9 .4 .2 .9 .2 .8 .7 .3 .7 .7

exercise at least 3 times weekly predicted higher calcium scores but only in white women. Hormone replacement therapy was not associated with coronary calcium score, whether defined as “current hormone replacement therapy” or “ever used hormone replacement therapy” (data not shown). In regression analysis of the entire cohort of black women and white women, ethnicity was not independently associated with total coronary calcium score (odds ratio 0.83, 95% CI 0.51-1.35, P ⫽ .45).

Discussion Despite a higher prevalence of coronary risk factors, coronary calcium scores in black women were similar to those of white women. Age and diabetes mellitus were the only independent predictors of coronary calcium score in the combined cohort despite inclusion of a number of variables beyond conventional coronary risk factors, such as educational level, dietary fat consumption, exercise, and body mass index, which have not been studied in prior analyses.1,5 Women in general and minority women in particular have been poorly represented in previous CT reports. The ethnic composition was not defined in 2 prior large studies1,11 of asymptomatic patients undergoing CT. The study by Doherty et al9 did specify ethnicity but included only 13 black women. Sex-specific ranges of coronary calcium scores for coronary risk stratification have been established, but the appropriateness of these ranges for nonwhite populations has been little studied.

0.82 (0.45-1.47) 0.80 (0.55-1.16) 1.00 1.38 (0.90-2.12) 11.59 (3.21-41.87) 1.13 (0.72-1.78) 1.53 (0.71-3.28) 1.40 (0.81-2.42) 1.05 (0.86-1.29) 0.92 (0.64-1.31) 0.95 (0.66-1.37) 1.01 (0.99-1.04) 1.48 (1.04-2.10) 1.00 (1.00-1.01)

.1 .0002 .6 .3 .2 .6 .6 .8 .2 .03 .1

1.18 (0.71-1.96) 0.86 (0.61-1.21) 1.00 1.48 (0.99-2.21) 4.27 (1.58-11.58) 1.28 (0.85-1.92) 1.24 (0.63-2.45) 1.31 (0.79-2.18) 1.07 (0.88-1.31) 0.94 (0.67-1.31) 0.99 (0.71-1.34) 1.01 (0.99-1.03) 1.36 (0.98-1.88) 1.00 (0.99-1.01) 0.83 (0.51-1.35)

.1 .004 .2 .5 .3 .5 .7 .9 .3 .1 .9 .5

This group of predominantly healthy volunteers would be expected to have lower coronary risk than the general population or than cohorts reported in other CT studies. For example, prevalence of hypertension in the general population ranges from 47% to 78% in 50- to 79-year-old black women and from 38% to 68% in white women.20 In contrast, 47% and 25% of black women and white women, respectively, in this analysis had hypertension. Similarly, 10% of 65- to 74year-old women in the general population report having diabetes mellitus21 compared with 2% in this cohort. By applying the Framingham algorithm to the 480 women in this analysis who had fasting lipid profiles available, the average 10-year risk of myocardial infarction, angina, or coronary death was considerably lower than the risk for 60- to 64-year-old women in the Framingham study (5.2% vs 12%).17 Previous CT studies have included individuals referred or self-referred for screening because of their perceived high risk for CHD. For example, in the St Francis cohort,1 which is 71% male, prevalence of diabetes mellitus (5%), current smoking (10%), hypercholesterolemia, (42%) and family history of premature CHD (44%) are all higher than in this report. In the South Bay Heart Watch cohort, which is 88% male, prevalence of diabetes mellitus (18%), current smoking (21%), and family history (45%) were also higher.9 Although age and diabetes mellitus were associated with coronary calcium score both in this analysis and in prior reports from predominantly male cohorts, some differences in characteristics associated with coronary

American Heart Journal April 2003

728 Khurana et al

calcification are apparent in this ethnically diverse group of women. For example, hypertension, smoking, hypercholesterolemia, and family history of premature CHD were not associated with coronary calcium score in this analysis. Possible explanations for the similarity of coronary calcium scores despite a higher prevalence of risk factors among black women include ethnic differences in predictors of coronary risk, risk factor intensity, estrogen exposure, and limitations of coronary calcium as a marker of coronary risk. Reported risk factors for CHD in black women had been similar to those in white women.22-24 The predictive accuracy of the Framingham algorithm is similar in black women and white women,25 suggesting that risk factors of similar intensity convey similar risk in the 2 ethnic groups. Conspicuous differences in risk factor intensity were not apparent between black women and white women in this cohort. Neither current nor past hormone replacement therapy predicted coronary calcium score, suggesting that differences in estrogen exposure are unlikely to account for the dissociation between coronary calcification and number of risk factors. The utility of calcium score as a predictor of clinical coronary events has not been confirmed in women; systematic outcomes ascertainment in this cohort until 2005 will ultimately permit comparison of the predictive value of coronary calcium score for coronary events in black women and white women. Strengths of this analysis include the inclusion of substantial numbers of women and nonwhites and inclusion of socioeconomic, dietary, and exercise variables, as well as physical measures. The absence of clinical outcomes is a weakness, which will be remedied as follow-up continues for identification and confirmation of coronary death, nonfatal myocardial infarction, overnight hospitalization for angina, and coronary revascularization. Despite a higher prevalence of coronary risk factors, coronary calcium scores of black women were similar to those of white women in this analysis, results similar to those reported for young adults.26 Blacks have a disproportionately high cardiovascular mortality rate in relation to their risk characteristics.10 Thus, the relation between coronary calcification, traditional coronary risk factors, and coronary events in black compared with white populations bears examination.

References 1. Arad Y, Spadaro LA, Goodman K. Prediction of coronary events with electron beam computed tomography. J Am Coll Cardiol 2000;36:1253-60. 2. Rumberger JA, Behrenbeck T, Breen JF. Coronary calcification by electron beam computed tomography and obstructive coronary artery disease: a model for costs and effectiveness of diagnosis as

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17. 18.

19.

20.

compared with conventional cardiac testing methods. J Am Coll Cardiol 1999;33:453-62. Detrano R, Hsiai T, Wang S. Prognostic value of coronary calcification and angiographic stenoses in patients undergoing coronary angiography. J Am Coll Cardiol 1996;27:285-90. Detrano R, Schwendener C, Doherty T. Coronary calcium results predict coronary heart disease deaths in high risk asymptomatic adults [abstract]. J Am Coll Cardiol 1997;29:128A. Kennedy J, Shavelle R, Wang S. Coronary calcium and standard risk factors in symptomatic patients referred for coronary angiography. Am Heart J 1998;135:696-702. Shavelle DM, Budoff MJ, LaMont DH. Exercise testing and electron beam computed tomography in the evaluation of coronary artery disease. J Am Coll Cardiol 2000;36:32-8. Kennedy JW, Killip R, Fisher LD. The clinical spectrum of coronary artery disease and its surgical and medical management, 19741979: the Coronary Artery Surgery Study. Circulation 1982;66(3 Suppl):16. Doherty T, Tang W, Dascalos S. Ethnic origin and serum levels of 1␣,25-dihydroxyvitamin D3 are independent predictors of coronary calcium mass measured by electron beam computed tomography. Circulation 1997;96:1477-81. Doherty TM, Tang W, Detrano RC. Racial differences in the significance of coronary calcium in asymptomatic black and white subjects with coronary risk factors. J Am Coll Cardiol 1999;34:78794. Keil JE, Sutherland SE, Knapp RG. Mortality rates and risk factors for coronary disease in black as compared with white men and women. N Engl J Med 1993;329:73-8. Janowitz WR, Agatston AS, Kaplan G. Differences in prevalence and extent of coronary artery calcium detected by ultrafast computed tomography in asymptomatic men and women. Am J Cardiol 1993;72:247-54. Women’s Health Initiative Study Group. Design of the Women’s Health Initiative Clinical Trial and Observational Study. Control Clin Trials 1998;19:61-109. Henderson MM, Kushi LH, Thompson DJ. Feasibility of a randomized trial of a low-fat diet for the prevention of breast cancer: dietary compliance in the Women’s Health Trial Vanguard Study. Prev Med 1990;19:115-33. White E, Shattuck AL, Kristal AR. Maintenance of a low-fat diet: follow-up of the Women’s Health Trial. Cancer Epidemiol Biomarkers Prev 1992;1:315-23. Kristal AR, Patterson RE, Glanz K. Psychosocial correlates of healthful diets and intention to improve diet: baseline results from the Working Well Study. Prev Med 1995;24:221-8. Patterson RE, Kristal AR, Tinker LF. Measurement characteristics of the Women’s Health Initiative Food Frequency Questionnaire. Ann Epidemiol 1999;9:178-87. Wilson PWF, D’Agostino RB, Levy D. Prediction of coronary heart disease using risk factor categories. Circulation 1998;97:1837-47. Agatston A, Janowitz W, Hildner F. Quantification of coronary artery calcium using ultrafast computed tomography. J Am Coll Cardiol 1990;15:827-32. Hoff JA, Chomka EV, Krainik AJ. Age and gender distributions of coronary artery calcium detected by electron beam tomography in 35,246 adults. Am J Cardiol 2001;87:1335-9. Burt VL, Whelton P, Roccella EJ. Prevalence of hypertension in the US adult population: results from the Third National Health and Nutrition Examination Survey, 1988-1991. Hypertension 1995; 25:305-13.

American Heart Journal Volume 145, Number 4

21. Kenny SJ, Aubert RE, Geiss LS. Prevalence and incidence of noninsulin-dependent diabetes. In: National Diabetes Data Group. Diabetes in America. (NIH Publication No. 95-1468) Bethesda (Md); NIDDKD, 47-68, 1995. 22. Rosenberg L, Palmer JR, Rao RS. Risk factors for coronary heart disease in African American women. Am J Epidemiol 1999;150: 904-9. 23. Bensen JT, Li R, Hutchinson RG. Family history of coronary heart disease and pre-clinical carotid artery atherosclerosis in AfricanAmericans and Whites: the ARIC study: Atherosclerosis Risk in Communities. Genet Epidemiol 1999;16:165-78.

Khurana et al 729

24. Folsom AR, Szklo M, Stevens J, et al. A prospective study of coronary heart disease in relation to fasting insulin, glucose and diabetes: the Atherosclerosis Risk in Communities (ARIC) Study. Diabetes Care 1997;20:935-42. 25. D’Agostino RB Sr, Grundy S, Sullivan LM. Validation of the Framingham coronary heart disease prediction scores: results of a multiple ethnic groups investigation. JAMA 2001;286:180-7. 26. Bild DE, Folsom AR, Lowe LP. Prevalence and correlates of coronary calcification in black and white young adults: the Coronary Artery Risk Development in Young Adults (CARDIA) Study. Arterioscler Thromb Vasc Biol 2001;21:852-7.