Secondary Logo

Journal Logo

Original Studies

Cardiovascular disease (CVD) risk scores, age, or years since menopause to predict cardiovascular disease in the Women's Health Initiative

Wild, Robert A. MD, MPH, PhD1; Hovey, Kathleen M. MS2; Andrews, Christopher PhD3; Robinson, Jennifer G. MD, MPH4; Kaunitz, Andrew M. MD5; Manson, JoAnn E. MD, DrPH6,7; Crandall, Carolyn J. MD, MS8; Paragallo, Rachel MD, MS9; Shufelt, Chrisandra MD, MS10; Merz, C. Noel Bairey MD10

Author Information
doi: 10.1097/GME.0000000000001753

Abstract

Midlife is a time of increasing cardiovascular disease (CVD) risk for women, attributable to chronological and biological aging.1 A greater burden of CVD risk factors, including high blood pressure, overweight/obesity, diabetes, smoking, and dyslipidemia, is associated with greater subclinical vascular disease in midlife women.2,3 Due to the elevation of CVD risk attributable to menopausal hormone therapy (HT) among older women at high underlying risk of CVD,4 various strategies have been proposed to estimate risk with HT use when HT is being considered for menopausal symptoms. First, older age and years since menopause have been used to estimate CVD risk.4,5 Second, management based on vascular, rather than chronological age, has been proposed.6 Third, Bray et al7 found that higher low-density lipoprotein cholesterol (LDL-C), lower high-density lipoprotein cholesterol (HDL-C), and higher LDL/HDL-C ratios (≥ 2.5) at baseline were associated with higher HT-related CVD risks. Metabolic syndrome at trial baseline was associated with greater HT-related CVD risk in the Women's Health Initiative (WHI).8 Previous studies have not evaluated whether established CVD risk scoring systems predict CVD outcomes in the WHI randomized controlled trial (RCT) cohorts. Our objective was to compare calculated CVD risk scores to age and to years since menopause for prediction of CVD events in the HT clinical trial cohorts of the WHI. We used the validated Framingham Risk Score and the American Heart Association atherosclerotic cardiovascular disease (ASCVD) risk assessment tools. We did not include the Reynold's Risk Score because all the required parameters were not collected in the WHI.

METHODS

Details of the WHI hormone therapy trials and outcome adjudication have been published.9-11 Briefly, 27,347 postmenopausal women ages 50 to 79 were enrolled from 1993 to 1998 at 40 US clinical centers. Women with a uterus (n = 16,608) were randomized to daily oral conjugated equine estrogen (CEE) −0.625 mg plus medroxyprogesterone acetate (MPA) −2.5 mg or placebo; 10,739 women with prior hysterectomy were randomized to daily oral CEE-0.625 mg (E) or placebo. The primary efficacy and safety outcomes were coronary heart disease and invasive breast cancer. Institutional review board approval was received at each clinical center. All participants provided written informed consent. Race and ethnicity were self-reported. The intervention phase of the trial was stopped early because of increased risks for breast cancer and overall risks exceeded benefits in the CEE + MPA arm after 5.7 years (interquartile range [IQR], 4.8-6.5 years) and because of an increased risk of stroke in the CEE-alone arm after 7.2 years (IQR, 6.4-8.1 years). Post intervention follow-up to determine the incidence of CVD events continued in 81.1% of surviving participants who provided additional written informed consent in 2010. For each trial, our primary interests were the CVD outcomes until 2005, when the main study ended (baseline to end of study a median of 8.2 years), and the cumulative CVD rates (baseline to end 17.6 years). The current analysis is through February 2017. Outcomes were assessed by self-reported health history every 6 months during the main study and annually thereafter, CVD events were adjudicated.

The following demographic lifestyle and health-related information was collected at baseline: age, race/ethnicity, years since menopause, prior HT, vasomotor symptoms (hot flashes and night sweats), body mass index (BMI), measured systolic and diastolic blood pressure, self-reported treated hypertension (or measured blood pressure ≥ 140/90), smoking history, self-reported high cholesterol requiring treatment, treated diabetes, statin and aspirin use, history of prior myocardial infarction (MI), angina, revascularization, stroke, pulmonary embolism or deep vein thrombosis, fracture by age 55, family history of breast cancer, education, and income status.

We compared two established CVD risk scores in common use in The North America versus chronological age and years since menopause at entry to the observed follow-up absolute CVD events. We determined rates and their hazard ratios up to the end of the main study and up to the end of the cumulative follow-up.

The Framingham D ’Agostino CVD risk score (FHS)12 allows BMI (available for the entire cohort n = 26,773) to be substituted for HDL-C. This score is a sex-specific and well-validated index developed in the Framingham study cohort. It estimates 10-year risk of a CVD event, defined as coronary heart disease, coronary death, myocardial infarction, coronary insufficiency or angina, cerebrovascular events (ischemic or hemorrhagic stroke, transient ischemic attack, peripheral arterial disease, or heart failure). Variables included are age, systolic blood pressure, antihypertensive medication use, smoking, diabetes, HDL-C (or BMI).12 This scoring system can also provide an estimate of vascular age compared to chronological age when no risk factors are present. This is the extrapolated age a person would be if she had no other CVD risk factors adding to the score beyond her chronological age alone. It provides an indication of persons who likely have accelerated aging of their vascular system.12 This score was also calculated using HDL-C in the subset for whom HDL was measured (n = 14,018) (Fig. 1).

FIG. 1
FIG. 1:
Flow diagram for study analyses. ASCVD, atherosclerotic cardiovascular disease; HT, hormone therapy; WHI, Women's Health Initiative. The numbers don’t add up to the total because women can meet more than one of the listed exclusion criteria.

The American College of Cardiology/American Heart Association (ACC/AHA) atherosclerotic cardiovascular disease (ASCVD) risk score (Pooled cohort equations)13 is easily available free of charge from the uniform resource locator: https://tools.acc.org/ascvd-risk-estimator-plus/#!/calculate/estimate. The score is sex-specific and well validated among women age 40-75 years, to predict 10-year CVD risk defined as nonfatal or fatal myocardial infarction or stroke. The ASCVD risk score variables include age, race, total cholesterol, HDL-C, systolic blood pressure, antihypertensive medication use, diabetes, and smoking status.13 This score was used in the subset of women (N = 13,599) who had all of these variables measured in the current analysis. Women with values outside of the range for the score were excluded.

Statistical analysis

Baseline characteristics were summarized by quintiles (Q) of Framingham CVD and ASCVD risk scores. CVD rates are reported as events per 1,000 person-years. Cox models were used to estimate hazard ratios separately for quintiles of each risk score, chronologic age, and years since menopause using the lowest quintile as the reference category. The associations between CVD events and five baseline measures of risk were measured with Uno's concordance statistic, a C-statistic that measures the predictive ability of a model in the presence of right censoring.14 Comparison of concordance was restricted to either women with complete data on all measures (CVD-BMI) risk score, chronologic age, and years since menopause) or women in a cohort with lipid measurements (ASCVD and Framingham Health Study HDL). Sensitivity analyses were performed excluding women who reported history of CVD events (MI or Stroke). All analyses were performed using SAS software (v9.4, SAS Institute, Cary, NC).

RESULTS

Baseline characteristics of the women stratified by FHS risk quintiles are shown in Table 1 (see Table 1 Supplemental Digital Content 1, http://links.lww.com/MENO/A715, which provides baseline characteristics by quintile of ASCVD score). The mean age at baseline was 63.4 (7.2) years and 80.6% were White. Women with higher predicted CVD risk scores at baseline were older, had a greater number of years since menopause onset, had fewer vasomotor symptoms, had higher BMIs, were more likely to be current smokers, more likely to be treated for diabetes, hypertension, or high cholesterol, and more likely to have prior bilateral salpingo-oophorectomy. Those with higher CVD predicted risk scores were also more likely to be using statins or aspirin and to have had a prior myocardial infarction, stroke, or revascularization procedure.

TABLE 1 - Baseline characteristics by quintiles of Framingham D’Agostino score (BMI)
D’Agostino score
Q1 (5-9) Q2 (10-12) Q3 (13-15) Q4 (16-17) Q5 (18-33)
Quintile (range) 8 11 14 16 20
Median 4,840 5,657 6,636 3,754 5,886
N N (%) N (%) N (%) N (%) N (%)
Age
 Mean (SD) 57.1 (4.9) 61 (6.4) 63.7 (6.2) 66.3 (6.1) 68.7 (6.1)
 50-59 3,348 (69.2) 2,555 (45.2) 1,739 (26.2) 526 (14.0) 488 (8.3)
 60-69 1,492 (30.8) 2,505 (44.3) 3,748 (56.5) 2,058 (54.8) 2,311 (39.3)
 ≥70 0 (0.0) 597 (10.6) 1,149 (17.3) 1,170 (31.2) 3,087 (52.4)
Race/ethnicity
 American Indian or Alaskan Native 18 (0.4) 23 (0.4) 29 (0.4) 19 (0.5) 33 (0.6)
 Asian or Pacific Islander 111 (2.3) 113 (2.0) 118 (1.8) 55 (1.5) 120 (2.0)
 Black 311 (6.4) 477 (8.4) 645 (9.7) 412 (11.0) 817 (13.9)
 Hispanic/Latino 391 (8.1) 361 (6.4) 340 (5.1) 156 (4.2) 230 (3.9)
 White 3,948 (81.6) 4,597 (81.3) 5,423 (81.7) 3,073 (81.9) 4,581 (77.8)
 Unknown 61 (1.3) 86 (1.5) 81 (1.2) 39 (1.0) 105 (1.8)
Years since menopause
 <10 2,515 (58.4) 1,938 (38.5) 1,528 (25.9) 506 (15.1) 510 (9.8)
 10- < 20 1,382 (32.1) 1,959 (38.9) 2,404 (40.8) 1,334 (39.8) 1,721 (32.9)
 ≥20 408 (9.5) 1,140 (22.6) 1,966 (33.3) 1,512 (45.1) 2,998 (57.3)
Hormone use
 Never used 3,059 (63.2) 3,639 (64.3) 4,322 (65.2) 2,469 (65.8) 3,956 (67.2)
 Past user 1,126 (23.3) 1,414 (25.0) 1,772 (26.7) 997 (26.6) 1,631 (27.7)
 Current user 652 (13.5) 603 (10.7) 537 (8.1) 286 (7.6) 298 (5.1)
Vasomotor symptoms
 None 2,351 (49.0) 3,089 (55.1) 3,977 (60.5) 2,436 (65.5) 3,883 (66.8)
 Mild 1,554 (32.4) 1,597 (28.5) 1,736 (26.4) 824 (22.2) 1,275 (21.9)
 Severe 892 (18.6) 917 (16.4) 861 (13.1) 458 (12.3) 658 (11.3)
BMI (kg/m2)
 Mean (SD) 26.5 (5.2) 28.3 (5.7) 29.6 (6.1) 30.3 (6.1) 30.7 (6.1)
 <25 2,264 (46.8) 1,752 (31.0) 1,524 (23.0) 681 (18.1) 929 (15.8)
 25- < 30 1,708 (35.3) 2,192 (38.7) 2,266 (34.1) 1,297 (34.5) 1,939 (32.9)
 ≥30 868 (17.9) 1,713 (30.3) 2,846 (42.9) 1,776 (47.3) 3,018 (51.3)
Blood pressure mean (SD)
 Systolic (mm Hg) 110.4 (8.2) 119.3 (10.3) 128.6 (11.2) 136.5 (12.5) 147.8 (15.6)
 Diastolic (mm Hg) 71.3 (7.2) 74.2 (8.3) 76.9 (8.9) 78.1 (9.2) 79.3 (9.8)
Smoking
 Never Smoked 2,589 (53.5) 2,851 (50.4) 3,307 (49.8) 1,839 (49.0) 2,894 (49.2)
 Past smoker 2,142 (44.3) 2,280 (40.3) 2,594 (39.1) 1,442 (38.4) 2,036 (34.6)
 Current Smoker 109 (2.3) 526 (9.3) 735 (11.1) 473 (12.6) 956 (16.2)
Bilateral oophorectomy
 No 4,271 (89.9) 4,745 (86.0) 5,418 (84.2) 2,942 (81.5) 4,541 (80.8)
 Yes 481 (10.1) 774 (14.0) 1,015 (15.8) 669 (18.5) 1,079 (19.2)
 Missing 88 138 203 143 266
Treated diabetes
 No 4,840 (100.0) 5,622 (99.4) 6,527 (98.4) 3,590 (95.6) 4,682 (79.5)
 Yes 0 (0.0) 35 (0.6) 109 (1.6) 164 (4.4) 1,204 (20.5)
Hypertension
 No 4,021 (92.4) 3,965 (76.3) 3,291 (52.6) 944 (25.7) 351 (6.0)
 Yes 330 (7.6) 1,235 (23.8) 2,969 (47.4) 2,723 (74.3) 5,499 (94.0)
High cholesterol requiring pills ever
 No 3,998 (93.5) 4,514 (89.1) 5,178 (86.9) 2,879 (84.5) 4,248 (78.8)
 Yes 276 (6.5) 553 (10.9) 778 (13.1) 528 (15.5) 1,143 (21.2)
Statin use
 No 4,717 (97.5) 5,364 (94.8) 6,210 (93.6) 3,440 (91.6) 5,155 (87.6)
 Yes 123 (2.5) 293 (5.2) 426 (6.4) 314 (8.4) 731 (12.4)
Aspirin use (≥80 mg)
 No 4,222 (87.2) 4,710 (83.3) 5,367 (80.9) 2,887 (76.9) 4,249 (72.2)
 Yes 618 (12.8) 947 (16.7) 1,269 (19.1) 867 (23.1) 1,637 (27.8)
MI
 No 4,819 (99.6) 5,586 (98.7) 6,512 (98.1) 3,638 (96.9) 5,604 (95.2)
 Yes 21 (0.4) 71 (1.3) 124 (1.9) 116 (3.1) 282 (4.8)
Angina
 No 4,758 (98.6) 5,456 (96.7) 6,300 (95.4) 3,498 (93.8) 5,224 (89.5)
 Yes 69 (1.4) 184 (3.3) 305 (4.6) 230 (6.2) 610 (10.5)
Revascularization
 No 4,793 (99.8) 5,552 (99.1) 6,487 (98.8) 3,631 (97.9) 5,548 (96.1)
 Yes 8 (0.2) 48 (0.9) 78 (1.2) 77 (2.1) 227 (3.9)
Stroke
 No 4,827 (99.7) 5,622 (99.4) 6,577 (99.1) 3,724 (99.2) 5,729 (97.3)
 Yes 13 (0.3) 35 (0.6) 59 (0.9) 30 (0.8) 157 (2.7)
PE or DVT
 No 4,793 (99.0) 5,593 (98.9) 6,565 (98.9) 3,706 (98.7) 5,811 (98.7)
 Yes 47 (1.0) 64 (1.1) 71 (1.1) 48 (1.3) 75 (1.3)
Fracture age 55
 No 2,327 (53.8) 3,402 (69.1) 4,310 (76.4) 2,479 (77.7) 3,909 (78.1)
 Yes 234 (5.4) 578 (11.7) 874 (15.5) 614 (19.2) 1,020 (20.4)
 age <55 1,762 (40.8) 946 (19.2) 461 (8.2) 99 (3.1) 76 (1.5)
Family history of breast cancer
 No 1,372 (65.3) 1,650 (64.6) 1,947 (65.8) 1,041 (64.3) 1,636 (63.8)
 Yes 730 (34.7) 903 (35.4) 1,010 (34.2) 579 (35.7) 928 (36.2)
Education
 High school or less 1,028 (21.4) 1,507 (26.8) 1,960 (29.8) 1,129 (30.2) 2,002 (34.2)
 Some college or vocational training 1,854 (38.7) 2,275 (40.5) 2,687 (40.8) 1,587 (42.5) 2,417 (41.3)
 College grad or more 1,913 (39.9) 1,831 (32.6) 1,939 (29.4) 1,018 (27.3) 1,439 (24.6)
Income
 <50,000 2,666 (58.1) 3,623 (67.9) 4,601 (73.7) 2,739 (77.4) 4,600 (82.7)
 ≥50,000 1,922 (41.9) 1,714 (32.1) 1,646 (26.3) 800 (22.6) 961 (17.3)
HT study arms
 E + P intervention 1,761 (36.4) 1,789 (31.6) 2,097 (31.6) 1,107 (29.5) 1,599 (27.2)
 E + P control 1,581 (32.7) 1,784 (31.5) 1,953 (29.4) 1,071 (28.5) 1,530 (26.0)
 E-alone intervention 716 (14.8) 1,041 (18.4) 1,305 (19.7) 786 (20.9) 1,348 (22.9)
 E-alone control 782 (16.2) 1,043 (18.4) 1,281 (19.3) 790 (21.0) 1,409 (23.9)
Self-reported hypertension or medication for hypertension or systolic blood pressure ≥140 mm Hg or diastolic blood pressure ≥90 mm Hg.Category counts may not equal total because of missing data.BMI, body mass index; DVT, deep vein thrombosis; E+P, estrogen plus progesterone; E-alone, estrogen alone; HT, hormone therapy; MI, myocardial infarction; PE, pulmonary embolism.

Tables 2 and 3 display the absolute CVD event rates and the hazard ratios for events at the end of the main study and at cumulative follow-up, by quintiles defined by the Framingham BMI, HDL, and ASCVD risk scores, and age and years since menopause. There were higher absolute rates and hazard ratios (relative to Q1) for each successive higher quintile of the predictors calculated at trial onset to the end of the main study and to the end of cumulative follow-up. Overall, the hazard ratios calculated from baseline data using the three risk scores were higher than age or years since menopause at entry (baseline) in the RCT. Framingham (BMI) quintile 5 versus 1 had the highest hazard ratio and event rate at both follow-up time points. Of the two risk scores measured on the subsample, the hazard ratios and event rates for Framingham (HDL) scoring were higher than ASCVD scoring. A few women who entered into the WHI RCT had had a prior CVD event. For this reason, we conducted a sensitivity analysis limited to people without prior CVD events (See tables, Supplemental Digital content 2a and 2b, http://links.lww.com/MENO/A716, which illustrates that results are quite similar at both follow-up time points when those with prior CVD events are not included). The risk increases by increasing baseline quintile risk just as it does when the complete cohorts were evaluated.

TABLE 2 - Absolute CVD event rates and hazard ratios for events through end of the main study (2005) by quintile of three CVD risk scores, age, and years since menopause.
Quintile 1 Quintile 2 Quintile 3 Quintile 4 Quintile 5
Framingham (BMI)
 Quintile range 5-9 10-12 13-15 16-17 18-33
N 4,840 5,657 6,636 3,754 5,886
 Event (rate) 79 (1.9) 181 (3.9) 392 (7.4) 296 (10.1) 881 (20.2)
 HR (95% CI) 1 (reference) 2.03 (1.56-2.65) 3.87 (3.04-4.93) 5.31 (4.15-6.81) 10.8 (8.56-13.6)
Framingham (HDL)
 Quintile range 2-10 11-12 13-15 16-17 18-30
N 2,961 2,225 3,792 2,137 2,903
 Event (rate) 82 (3.4) 130 (7.3) 307 (10.4) 251 (15.5) 556 (26.8)
 HR (95% CI) 1 (reference) 2.20 (1.67-2.90) 3.12 (2.44-3.98) 4.67 (3.64-5.99) 8.12 (6.44-10.2)
ASCVD
 Quintile range 0-5 5-8 8-12 12-18 18-146
N 2,719 2,720 2,720 2,720 2,720
 Event (rate) 94 (4.1) 179 (8.3) 230 (10.8) 289 (14.1) 486 (25.3)
 HR (95% CI) 1 (reference) 2.02 (1.57-2.59) 2.64 (2.08-3.36) 3.47 (2.75-4.38) 6.27 (5.03-7.82)
Age
 Quintile range 50-56 57-61 62-65 66-70 71-81
N 5,545 5,600 4,838 5,808 4,982
 Event (rate) 152 (3.2) 238 (5.2) 310 (8.1) 500 (11.1) 629 (17.0)
 HR (95% CI) 1 (reference) 1.63 (1.33-2.00) 2.55 (2.10-3.10) 3.51 (2.93-4.21) 5.45 (4.56-6.51)
Years since Menopause
 Quintile range 0-6 7-12 13-17 18-23 24-55
N 4,734 4,994 4,429 4,734 4,930
 Event (rate) 144 (3.7) 216 (5.4) 291 (8.5) 394 (10.8) 561 (15.1)
 HR (95% CI) 1 (reference) 1.48 (1.20-1.83) 2.35 (1.92-2.87) 3.00 (2.48-3.63) 4.23 (3.52-5.09)
Framingham (HDL) and ASCVD computed on subsample. Rate: events per 1,000 person-years.ASCVD, atherosclerotic cardiovascular disease; BMI, body mass index; CVD, cardiovascular disease; HDL, high-density lipoprotein.

TABLE 3 - Absolute CVD event rates and hazard ratio for events through the end of follow-up by (2017) quintile of three CVD risk scores, age, and years since menopause.
Quintile 1 Quintile 2 Quintile 3 Quintile 4 Quintile 5
Framingham (BMI)
 Quintile range 5-9 10-12 13-15 16-17 18-33
N 4,840 5,657 6,636 3,754 5,886
 Event (rate) 254 (3.2) 557 (6.4) 982 (10.1) 712 (13.8) 1,719 (24.5)
 HR (95% CI) 1 (reference) 2.07 (1.79-2.40) 3.37 (2.93-3.86) 4.71 (4.08-5.43) 8.86 (7.76-10.1)
Framingham (HDL)
 Quintile range 2-10 11-12 13-15 16-17 18-30
N 2,961 2,225 3,792 2,137 2,903
 Event (rate) 284 (5.8) 342 (10.1) 736 (13.2) 554 (18.9) 1,014 (28.7)
 HR (95% CI) 1 (reference) 1.76 (1.50-2.06) 2.33 (2.04-2.68) 3.42 (2.96-3.95) 5.34 (4.68-6.09)
ASCVD
 Quintile range 0-5 5-8 8-12 12-18 18-146
N 2,719 2,720 2,720 2,720 2,720
 Event (rate) 218 (4.9) 436 (10.3) 543 (13.4) 690 (18.4) 954 (29.9)
 HR (95% CI) 1 (reference) 2.16 (1.83-2.54) 2.87 (2.45-3.36) 4.06 (3.49-4.73) 6.93 (5.97-8.03)
Age
 Quintile range 50-56 57-61 62-65 66-70 71-81
N 5,545 5,600 4,838 5,808 4,982
 Event (rate) 378 (4.3) 587 (6.8) 733 (10.4) 1,148 (14.2) 1,378 (23.2)
 HR (95% CI) 1 (reference) 1.62 (1.42-1.84) 2.53 (2.23-2.86) 3.55 (3.16-3.99) 6.23 (5.56-6.99)
Years since Menopause
 Quintile range 0-6 7-12 13-17 18-23 24-55
N 4,734 4,994 4,429 4,734 4,930
 Event (rate) 350 (4.6) 585 (7.7) 712 (11.3) 897 (13.8) 1,197 (19.5)
 HR (95% CI) 1 (reference) 1.69 (1.48-1.93) 2.57 (2.26-2.92) 3.21 (2.83-3.63) 4.68 (4.16-5.28)
Framingham (HDL) and ASCVD computed on subsample. Rate: events per 1,000 person-years.ASCVD, atherosclerotic cardiovascular disease; BMI, body mass index; CVD, cardiovascular disease; HDL, high density lipoprotein.

Table 4 displays the Uno's concordance statistics and standard errors (SE) comparing years since menopause, chronological age, and the Framingham BMI, HDL, and ASCVD risk scores. In the full sample, the Framingham (BMI) risk score had the highest C-statistic at the end of the main study (0.711) and at the end of follow-up (0.689). In the subsample, the Framingham (HDL) risk score had the highest C-statistic (0.684) at the end of the main study and the ASCVD risk score had the highest C-statistic (0.677) at the end of follow-up.

TABLE 4 - Uno concordance statistic and standard error (SE) for assessing CVD prediction by five single-variable Cox models: age, years since menopause, and Framingham score using BMI in the full sample (n = 23,821); and additionally, the Framingham score using HDL and ACC/AHA ASCVD score in the subset (n = 12,164).
CVD events main study follow-up (2005) CVD events through end of follow-up (2017)
Full sample Subsample Full sample Subsample
Events 1,606 1,134 3,741 2,547
Source Estimate SE Estimate SE Estimate SE Estimate SE
Years since menopause 0.622 0.013 0.587a 0.010 0.650a 0.006 0.605 0.007
Age 0.650 0.011 0.590a 0.008 0.662a 0.007 0.628 0.006
Framingham score (BMI) 0.711 0.010 0.673b 0.008 0.689 0.008 0.665a 0.005
Framingham score (HDL) 0.684c 0.008 0.665a 0.006
ACC/AHA ASCVD score 0.673bc 0.008 0.677 0.006
Estimates reported using data from the main study only (left) and from all available follow-up (right). Estimates with like letters are not statistically significantly different.ACC/AHA, American College of Cardiology/The American Heart Association; ASCVD, atherosclerotic cardiovascular disease; BMI, body mass index; CVD, cardiovascular disease; HDL, high density lipoprotein.

Figure 2A and B displays ROC curves for the full sample after 10 years of follow-up and for the subsample contrasting age, years since menopause and CVD risk scores.

FIG. 2
FIG. 2:
ROC curves estimated at 10 years. CVD events through end of follow-up. (A) Full Sample (n = 23,821). (B) Subsample (n = 12,164). CVD, cardiovascular disease; ROC, receiver operating characteristic.

DISCUSSION

While age and years since menopause have traditionally guided menopause management of HT decisions, we found that CVD risk score assessment to quantify risk factors better informed a person's risk for the actual CVD event. Older age and more years since menopause were associated with more CVD events; however, established, validated risk scores, the best for this cohort being the Framingham CVD score, accomplished better risk prediction for an actual event than either chronological age or years since menopause at entry into the WHI.

The ASCVD risk score office tool provides a 10-year risk estimate,13 and is endorsed by ACC/AHA/National Lipid Association (NLA) cholesterol guidelines.15 The Framingham CVD score is validated for 10-year risk assessment also.12 It has the advantage of providing an estimate of heart age/vascular age compared with current chronological age. The utility of these scores is to better predict women at advanced vascular age at potentially greater risk for a CVD event. Elevated vascular age suggests higher coronary plaque burden, indicating greater susceptibility to HT-enhanced plaque rupture or erosion and acute coronary events.6 Relating this score to a patient emphasizes vascular rather than chronological age for improved CVD risk prediction. Experienced clinicians are typically aware that some of their patients are younger from a cardiovascular or biologic perspective than their chronologic age, others are older. Diagnostic and therapeutic decisions are substantially influenced by biologic age6 and vascular age is more accurate for the prediction of longer-term CVD event risk. Other work demonstrates that the concept of “vascular age” to inform risk is easy for patients to understand.16

Most of the women who went on to suffer a CVD event had scores high enough at baseline to qualify for statin use currently (2018 ACC/AHA/NLA Guidelines of CVD 10-y absolute event rate of ≥7.5%).15 Many women in our WHI cohorts used aspirin for primary prevention. This is now recommended predominantly only for women at high CVD risk (ASCVD ≥ 20%) or who have diabetes with a 10-year CVD risk ≥ 10%15 or for secondary prevention. This might have influenced prediction estimation with each of the predictors we compared.

Current guidelines endorse score calculation of CVD risk annually in women ≥ 40 years of age. While assessment is easy online, via smartphone, tablet, or computer, all scoring systems have some potential shortcomings.17,18 Scoring is better validated for middle-age patients. Traditional CVD risk factors (such as dyslipidemia) are more strongly associated with premature CVD. The management of dyslipidemia and hypertension has improved since the WHI began.15,16,19 Many younger individuals (especially those with elevated LDL-C) are at low risk in the short term, yet are at high lifetime CVD risk.20,21 Risk score calculation for 30-year, lifetime, “vascular age,” and “cardiovascular age” risk are also available17,22 and useful for clinical care and preventive counseling. Using CVD risk scoring applies the concept that the vascular system of some persons is older than their chronological age. Office CVD risk scoring can be supplemented with other assessments such as coronary calcium to further refine a risk estimate (enhance or modify to reclassify as higher or lower risk). Scores may help with a concept of red-light versus green light versus yellow light for cardiovascular risk. Each of the scoring systems includes age as an important component of determining risk.

Our analysis has some important strengths and limitations. The WHI hormone therapy trial is a large, well-characterized cohort and most CVD risk factors were available at baseline for assessment. The cohort was followed for a relatively long time with high retention rates. Outcomes were adjudicated. Continuation of HT after the active phase was optional after the WHI ended. Only oral HT at fixed doses was used in the RCT. Changes in exercise habits and/or improvements in diets and/or medication for CVD risk were not evaluated and this could affect prediction. None of the scores, as currently constructed, takes into consideration a prior history of complications of pregnancy and/or premature ovarian insufficiency as risk factors for chronic disease. Albeit CVD is one of the most important risks to consider, the information provided needs to be in consideration of other risks and benefits when HT is considered.

CONCLUSIONS

At entry, CVD risk scores better identified women at higher risk of a CVD event than age or years since menopause alone in the WHI. This suggests that emphasis on vascular/biological age beyond chronological age or years since menopause is important for CVD risk prevention and is important to inform HT risk/benefit discussions.

REFERENCES

1. Barinas-Mitchell E, Duan C, Brooks M, et al. Cardiovascular disease risk factor burden during the menopause transition and late midlife subclinical vascular disease: does race/ethnicity matter? J Am Heart Assoc 2020; 9:e013876.
2. Matthews KA, El Khoudary SR, Brooks MM, et al. Lipid changes around the final menstrual period predict carotid subclinical disease in postmenopausal women. Stroke 2017; 48:70–76.
3. Derby CA, Crawford SL, Pasternak RC, Sowers M, Sternfeld B, Matthews KA. Lipid changes during the menopause transition in relation to age and weight: the Study of Women's Health Across the Nation. Am J Epidemiol 2009; 169:1352–1361.
4. Manson JE, Chlebowski RT, Stefanick ML, et al. Menopausal hormone therapy and health outcomes during the intervention and extended poststopping phases of the Women's Health Initiative randomized trials. JAMA 2013; 310:1353–1368.
5. Chester RC, Kling JM, Manson JE. What the Women's Health Initiative has taught us about menopausal hormone therapy. Clin Cardiol 2018; 41:247–252.
6. Santen RJ. Use of cardiovascular age for assessing risks and benefits of menopausal hormone therapy. Menopause 2017; 24:589–595.
7. Bray PF, Larson JC, Lacroix AZ, et al. Usefulness of baseline lipids and C-reactive protein in women receiving menopausal hormone therapy as predictors of treatment-related coronary events. Am J Cardiol 2008; 101:1599–1605.
8. Wild RA, Wu C, Curb JD, et al. Coronary heart disease events in the Women's Health Initiative hormone trials: effect modification by metabolic syndrome: a nested case-control study within the Women's Health Initiative randomized clinical trials. Menopause 2013; 20:254–260.
9. Rossouw JE, Anderson GL, Prentice RL, et al. Risks and benefits of estrogen plus progestin in healthy postmenopausal women: principal results From the Women's Health Initiative randomized controlled trial. JAMA 2002; 288:321–333.
10. Anderson GL, Limacher M, Assaf AR, et al. Effects of conjugated equine estrogen in postmenopausal women with hysterectomy: the Women's Health Initiative randomized controlled trial. JAMA 2004; 291:1701–1712.
11. Design of the Women's Health Initiative clinical trial and observational study. The Women's Health Initiative Study Group. Control Clin Trials 1998; 19:61–109.
12. D’Agostino RB Sr, Vasan RS, Pencina MJ, et al. General cardiovascular risk profile for use in primary care: the Framingham Heart Study. Circulation 2008; 117:743–753.
13. Stone NJ, Robinson JG, Lichtenstein AH, et al. 2013 ACC/AHA guideline on the treatment of blood cholesterol to reduce atherosclerotic cardiovascular risk in adults: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines. J Am Coll Cardiol 2014; 63 (25 pt B):2889–2934.
14. Uno H, Cai T, Pencina MJ, D’Agostino RB, Wei LJ. On the C-statistics for evaluating overall adequacy of risk prediction procedures with censored survival data. Stat Med 2011; 30:1105–1117.
15. Grundy SM, Stone NJ, Bailey AL, et al. 2018 AHA/ACC/AACVPR/AAPA/ABC/ACPM/ADA/AGS/APhA/ASPC/NLA/PCNA Guideline on the Management of Blood Cholesterol. Circulation 2018; Cir0000000000000625.
16. Soureti A, Hurling R, Murray P, van Mechelen W, Cobain M. Evaluation of a cardiovascular disease risk assessment tool for the promotion of healthier lifestyles. Eur J Cardiovasc Prev Rehabil 2010; 17:519–523.
17. Jaspers NEM, Blaha MJ, Matsushita K, et al. Prediction of individualized lifetime benefit from cholesterol lowering, blood pressure lowering, antithrombotic therapy, and smoking cessation in apparently healthy people. Eur Heart J 2020; 41:1190–1199.
18. Ridker PM, Cook NR. Comparing cardiovascular risk prediction scores. Ann Intern Med 2015; 162:313–314.
19. Cooney MT, Dudina AL, Graham IM. Value and limitations of existing scores for the assessment of cardiovascular risk: a review for clinicians. J Am Coll Cardiol 2009; 54:1209–1227.
20. Grover SA, Paquet S, Levinton C, Coupal L, Zowall H. Estimating the benefits of modifying risk factors of cardiovascular disease: a comparison of primary vs secondary prevention. Arch Intern Med 1998; 158:655–662.
21. Thanassoulis G, Williams K, Altobelli KK, Pencina MJ, Cannon CP, Sniderman AD. Individualized statin benefit for determining statin eligibility in the primary prevention of cardiovascular disease. Circulation 2016; 133:1574–1581.
22. Catapano AL, Reiner Z, De Backer G, et al. ESC/EAS Guidelines for the management of dyslipidaemias The Task Force for the management of dyslipidaemias of the European Society of Cardiology (ESC) and the European Atherosclerosis Society (EAS). Atherosclerosis 2011; 217:3–46.
Keywords:

CVD; Hormone; Risk; Scores; Therapy

Supplemental Digital Content

© 2021 by The North American Menopause Society