Impact of Major Cardiovascular Disease Risk Factors, Particularly in Combination, on 22-Year Mortality in Women and Men FREE

CARDIOVASCULAR diseases (CVDs), primarily coronary heart disease (CHD), remain leading causes of death in both men and women in the United States.¹ While much of the knowledge about CVD risk factors comes from studies in men, some prospective studies have included women.^2- 14 Despite limitations in some studies due to inclusion of small numbers of women, few years of follow-up, or limited numbers of clinical events, findings generally indicate that major modifiable risk factors for CVD in men (high blood pressure, high blood cholesterol level, and smoking) are also important in women. Nevertheless, the relative importance of CVD risk factor status and appropriate approaches to prevention of CVD in women continue to be debated.^15- 17 A less intensive approach to CVD prevention in women than in men has been advocated by some,^6,15,18- 19 but this view and the argument for it have been challenged.^{16- 17,20- 21}

Guidelines from the National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults²² and the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure (JNC VI)²³ give recommendations for CVD risk assessment and treatment in both women and men. Both sets of guidelines recommend that the presence or absence of multiple CVD risk factors should be considered in the evaluation and care of women and men. Smoking cessation policy also emphasizes the value for both women and men of controlling this major risk factor. The basic approach of such guidelines is that treatment intensity should match severity of CVD risk.²⁴ Although several studies have reported CVD risks in women,^2- 14 none have presented risk factor data based on current NCEP and JNC VI guidelines, and neither have they published data on the impact of combinations of major risk factors. The debate proceeds in this context, especially on CVD risk in women.^15- 17

The Chicago Heart Association Detection Project in Industry is one of the largest and longest prospective studies providing CVD mortality data for women and men. The goals of this report are (1) to use NCEP and JNC VI guidelines to classify baseline CVD risk factor status of middle-aged women and men and to estimate the prevalence of adverse risk factor status; (2) to present the impact of baseline risk factor status—for each major risk factor considered separately, and particularly for combinations of these factors—on CHD, CVD, and all-cause mortality in women compared with men, based on average follow-up of 22 years; and (3) to thereby provide data relevant to the discussion of single and multiple CVD risk factors in women.

The methods of the Chicago Heart Association Detection Project in Industry have been described elsewhere.^25- 26 Briefly, 39573 men and women aged 18 years and older were screened between November 1967 and January 1973. All employees at 84 Chicago-area cooperating companies and institutions were invited and encouraged to participate; from a labor force of more than 75000 people, the response rate was 55%. The study received institutional review board approval at periodic reviews.

Screenings were done by 2 trained and standardized 4-person field teams. Risk factors for CVD ascertained at baseline screening included age, gender, race, level of education, blood pressure, total serum cholesterol level, smoking status, height and weight used to calculate body mass index (BMI), resting electrocardiographic (ECG) findings, and medical history of and drug treatment for hypertension or diabetes. A single casual supine blood pressure measurement was obtained by trained staff using a standard mercury sphygmomanometer. Standardized, high-quality methods were used for total serum cholesterol level determination.²⁶ Smoking status was ascertained via a self-administered questionnnaire. Criteria of the National Cooperative Pooling Project and the Hypertension Detection and Follow-up Program²⁷ were used to code ECG abnormalities as major or minor.

Follow-up information is available for all but 0.2% of participants, in whom vital status could not be determined; they were excluded from the analyses here. Also excluded were those with missing data on risk factors at baseline (<1.0%). Vital status through 1992 was ascertained by periodic follow-up before 1979, both locally and through the Social Security Administration, and by the National Death Index since 1979. Death certificates were obtained for known decedents and were coded for underlying cause of death according to the International Classification of Diseases (Adapted for Use in the US), Eighth Revision (ICDA-8) .²⁸ Coronary heart disease deaths were those assigned ICD codes 410.0 through 414.9, and CVD deaths were those assigned ICD codes 400.0 through 445.9, which include heart failure and stroke.

Men and women aged 40 to 64 years at baseline were included in these analyses; persons with baseline ECG evidence of a previous myocardial infarction (0.5%) were excluded based on coding described above. Included were 8686 women and 10503 men. After average follow-up of 22 years, there were 405 and 1105 CHD deaths, 608 and 1483 CVD deaths, and 1577 and 3150 deaths caused by all causes among women and men, respectively.

Distributions of individual CVD risk factors were assessed in women and men. Total serum cholesterol levels were classified according to NCEP guidelines: less than 5.20 mmol/L (<200 mg/dL), 5.20 to 6.19 mmol/L (200-239 mg/dL), and 6.20 mmol/L or more (≥240 mg/dL) designate desirable, borderline high, and high blood cholesterol levels, respectively.²² Blood pressure was categorized by JNC VI criteria²³: optimal (systolic blood pressure [SBP] <120 mm Hg and diastolic blood pressure [DBP] <80 mm Hg), normal not optimal (SBP 120-129 mm Hg and DBP <85 mm Hg or SBP <130 mm Hg and DBP 80-84 mm Hg), high normal (SBP 130-139 mm Hg and DBP <90 mm Hg or SBP <140 mm Hg and DBP 85-89 mm Hg), and high (SBP ≥140 mm Hg or DBP ≥90 mm Hg or taking antihypertensive medication). Cigarette smoking status was classified as never, former, or current.

For women and men, age- and race-adjusted mortality rates per 10000 person-years of observation were calculated for CHD, CVD, and all-cause mortality by direct standardization. Person-years of exposure were calculated as time from date of baseline examination to time of death, to time of loss to follow-up, or to December 31, 1992. Cox proportional hazards analysis was used to determine gender-specific multivariate-adjusted relative risks (RRs) and corresponding 95% confidence intervals (CIs) of CHD, CVD, and all-cause mortality for strata of each major risk factor, adjusted for other major risk factors, excluding the risk factor of interest, and for potential confounders (age [in years], race, education [in years], BMI [calculated as weight in kilograms divided by the square of the height in meters {kg/m²}], BMI², any ECG abnormality, and diabetes). Because of small numbers of nonwhites, including blacks, Hispanics, and other nonwhite races, race was classified simply as white or nonwhite (all races reported other than white); ECG abnormalities were classified as noted earlier; and diabetes was categorized by medical history (yes or no). An alternative approach based on the Cox multivariate proportional hazards model and continuous values for the major risk factors (SBP, serum cholesterol level, and number of cigarettes per day) was also used. Findings for each major risk factor considered separately are henceforth referred to as findings for "individual" risk factors.

Analyses were repeated to determine mortality rates and RRs according to baseline absence or presence of any 1, 2, or all 3 major risk factors. Absence of all major risk factors (SBP <130 mm Hg and DBP <85 mm Hg and no antihypertensive drug treatment, serum cholesterol level <5.20 mmol/L [<240 mg/dL], and no current smoking) was categorized as none. Absolute excess risk (ie, attributable risk) was calculated as the difference between the mortality rate associated with the presence of a single or multiple risk factors compared with the absence of all risk factors. To assess the relationship of any 2 or all 3 risk factors to mortality, terms for each combination were replaced by a single summary term. The residual stratum of nonsmoking persons with high normal blood pressure or a borderline high serum cholesterol level was designated borderline only. Potential interactions between gender and CVD risk factor status were tested in models with women and men combined. Henceforth, the presence of only 1 risk factor is referred to as a "single" risk factor, and more than 1 risk factor is termed a "combination."

Table 1 gives distributions of individual major CVD risk factors and other baseline characteristics for women and men. According to JNC VI criteria, high blood pressure was the most common major risk factor, present in 53% of women and 64% of men. Relatively small proportions of hypertensive patients were taking medication at the baseline screenings. Based on NCEP criteria, 29.7% of women and 21.6% of men had high blood cholesterol levels. More than one third of both women and men were current smokers.

Risk factor status according to the prevalence of any 1, 2, or all 3 major CVD risk factors is shown in Table 2. Approximately 80% of both women and men had at least 1 of the 3 major CVD risk factors at baseline. More than one third of women and men had at least 2 major risk factors, and 5.7% of women and 6.1% of men had all 3 risk factors. For women and men with risk factors, prevalence of 1 risk factor only was much less frequent than prevalence of combinations; eg, for 7% of women and 3.6% of men, high blood cholesterol concentration was the sole major risk factor, whereas high blood cholesterol concentration plus 1 or 2 other factors was prevalent in 22.6% of women and 18% of men.

Table 3 presents 22-year age- and race-adjusted rates and multivariate-adjusted RRs of CHD mortality for men and women, associated with the presence of individual major risk factors. Among women and men, there was a positive association between high baseline level of each risk factor considered separately and CHD mortality. Presence of high blood pressure, a high blood cholesterol level, or current smoking was independently associated with significantly increased risk of CHD mortality compared with the desirable level of each risk factor, after adjustment for the other 2 major risk factors and other potential confounders. Current smoking was associated with a significantly greater RR in women compared with men. Findings from the analysis using continuous values for the major risk factors showed significant positive independent associations between SBP, serum cholesterol level, and number of cigarettes smoked and CHD mortality in both women and men.

Results for CVD mortality (data not shown) closely resembled those for CHD mortality. Findings were also generally similar for all-cause mortality (Table 4). Among both women and men, high blood pressure and current smoking were each significantly related to increased risk of all-cause mortality. High blood cholesterol level was positively associated with increased risk of all-cause mortality among women and men, but for women the RR was not statistically significant.

The results presented above do not match the manner in which clinicians are advised to use risk factor data in medical practice; that is, in clinical practice, the recommendation is that risk factors be considered together. Therefore, our major aim in these analyses was to assess impact of risk factor combinations on mortality. Table 5 shows adjusted rates and RRs of CHD mortality associated with the presence of only 1, 2, or all 3 major risk factors as currently defined in medical practice. Among women and men, the CHD mortality rate increased significantly with the number of major risk factors, from 7 and 13 per 10000 person-years of observation for women and men, respectively, with no baseline risk factors to 33 and 72 per 10000 person-years of observation for women and men, respectively, with combinations of high blood pressure, high blood cholesterol levels, and current smoking. Risk of CHD mortality in women and men significantly increased with the presence of any 2 or all 3 risk factors compared with the absence of all 3 risk factors, and there were no gender differences in RRs. With combinations of baseline risk factors, RR of CHD mortality increased markedly; women and men who had all 3 risk factors at baseline had the highest RRs (for women, RR=8.92, 95% CI=3.58-22.26; for men, RR=6.73, 95% CI=3.70-12.25).

Results for CVD mortality (data not shown) were similar to those for CHD mortality. Relative risk associated with any 2 or all 3 risk factors was similar for CVD mortality in women and men, respectively: RR=4.54 (95% CI, 2.33-8.84) and RR=4.12 (95% CI, 2.56-6.37). Findings for all-cause mortality (Table 6) were similar to those for mortality caused by CHD and CVD; eg, with any 2 or all 3 major risk factors, for women and men, respectively, RR=2.34 (95% CI, 1.73-3.15) and RR=3.20 (95% CI, 2.47-4.14). Absolute excess risks for women and men, respectively, were 83.1 and 158.3 per 10000 person-years. For both women and men, all 4 combinations of the 3 major risk factors were associated with similar significant increases in RR (range, 1.64-3.07 for women and 2.11-3.88 for men) and sizable increases in absolute excess risk (range, 42.0-128.9 per 10000 person-years for women and 87.4-201.1 per 10000 person-years for men).

Current medical practice guidelines^22- 23 emphasize the importance of the concomitant assessment of major CVD risk factors in both women and men and of matching preventive treatment intensity to severity of CVD risk. Few studies have used current clinical cut points to evaluate actual risks associated with individual risk factors in women and men, and none have assessed impact of combinations of the major risk factors. This was the purpose of this study. It is the first prospective population study to use nationally recognized clinical cut points^22- 23 to describe CVD risk factor status and to compare the impact of combinations of risk factors on mortality in middle-aged women and men. Its large sample size, lengthy follow-up, similar numbers of women and men, and sizable numbers of fatal events permitted meaningful gender comparisons of rates and risks of mortality. Findings show that major CVD risk factors (high blood pressure, high blood cholesterol level, and current smoking), first considered individually, are positively associated with risks of CHD, CVD, and all-cause mortality in women and men. Furthermore, they demonstrate that the presence of multiple CVD risk factors places women at considerably higher relative and absolute excess risks of CHD, CVD, and all-cause mortality compared with women without any risk factors, similar to men. These findings show that the major CVD risk factors known to be important in men are also important in women. These findings also support current NCEP²² and JNC VI²³ approaches to risk factor assessment and management in women and men.

Several studies have assessed risk factors for CVD incidence or mortality in both women and men; however, most have not used clinical cut points to assess risk,^{4- 5,7- 8,10,12- 14,29} none compared the impact in women and men, and none evaluated the impact of combinations of major risk factors. In 1 study⁶ that compared the impact of individual major risk factors on CHD mortality, RRs were slightly larger in women than in men; attributable and absolute risks were lower. The Framingham Study³⁰ showed increased probability of CVD associated with greater numbers of risk factors for women and men, but relative impact of multiple risk factors was not compared. One study³¹ examined the predictive validity of current clinical guidelines for risk assessment in the Lipid Research Clinics cohort, but utility was not examined separately for women.

As for the individual risk factors considered separately, high blood pressure as defined by JNC VI cut points (SBP >140 mm Hg or DBP >90 mm Hg or taking antihypertensive medication) and current smoking were each significant risk factors for CHD, CVD, and all-cause mortality in both women and men in the present study. These findings for women are consistent with reviews of several long-term studies^32- 34 that documented the relationship of elevated blood pressure and smoking to excess mortality. High blood cholesterol level as defined by NCEP criteria (>6.21 mmol/L [>240 mg/dL]) was a significant risk factor for CHD and CVD mortality in both women and men in the present study. This finding for women is concordant with results from the Nurses' Health Study¹¹ and an overview of 14 studies by Manolio et al.³⁵ In a recently reported large Dutch study,¹³ hypercholesterolemia was also significantly related to mortality risk for women and men who were followed up for 12 years. In our analyses, high blood cholesterol level was significantly associated with increased all-cause mortality among men but not women. These findings are similar to those reported in an overview of 11 studies of women,³⁶ whereas the recent Dutch report¹³ found significantly increased risk for all-cause mortality in women and men in the highest compared with the lowest quintile of cholesterol levels.

Our findings also show that women with none of the major CVD risk factors (high blood pressure, high blood cholesterol level, and current smoking) are at low risk of CHD, CVD, and all-cause mortality. Only 6.6% of women and 4.8% of men aged 40 to 64 years had none of the major risk factors at baseline. On the other hand, more than one third of women, as well as men, had combinations of these risk factors, placing them at high risk of death from CHD, CVD, and all causes.

The latest data on US adults from the Third National Health and Nutrition Examination Survey (1988-1991)³⁷ show that the prevalence of high blood pressure (categorized by the same cut points used in our study) was 18% among all US women. Data from the Second National Health and Nutrition Examination Survey³⁸ show that, among adult women, prevalence rates of high blood cholesterol level and smoking were 20% and 33%, respectively. Two or more major CVD risk factors were present in 10% of white and 15% of black women³⁴ and in 33% of women aged 65 years and older.³⁹ Thus, the prevalence of major CVD risk factors remains substantial.

Our results also show that these risk factors tend to occur in combination, not in isolation. For example, as noted, 33.9% of women and 37.6% of men had more than 1 risk factor at baseline. Therefore, risk assessment based on multiple CVD risk factors is appropriate to help guide decisions about preventive interventions in both women and men. This conclusion is in accord with current NCEP and JNC VI recommendations that emphasize that measurement of multiple major CVD risk factors should be used to guide type and intensity of cholesterol-lowering and antihypertensive drug treatments.^22- 23

There are limitations to the present study. First, CVD risk factor status was not determined after baseline. Because the prevalence of high blood pressure and high blood cholesterol level typically increases with aging,^10,39- 40 lack of data on follow-up risk factor status most likely biased our findings conservatively. High blood pressure^14,32 and smoking^10,14 are important risk factors for CVD in older women and men, and hypercholesterolemia has been associated with increased CHD risk in older persons.³⁵ Second, risk relationships based on a single measurement are underestimates because of regression dilution bias.⁴¹ This would tend to reduce the strength of risk factor associations. Thus, had multiple risk factor measurements been available, RRs would likely have been even higher than those observed here. Third, only total serum cholesterol concentration was measured. Information on other lipid risk factors, eg, high-density lipoprotein cholesterol level, is not available. Fourth, information on menopausal status or use of hormonal agents was unavailable for women; therefore, impact of risk factors according to hormonal status cannot be addressed. Despite the limitations, the data from the present study clearly demonstrate the adverse effects of unfavorable levels of major CVD risk factors—especially when present in combination—on risks of CHD, CVD, and all-cause mortality in both middle-aged women and men.

Current NCEP²² and JNC VI²³ clinical practice guidelines emphasize the value of using CVD risk factor status when a clinician decides whether to offer preventive treatment and how intensively such treatment should be undertaken. The data presented here demonstrate the applicability of current clinical practice guidelines for CVD risk in both women and men. In particular, these data show marked increases in RRs and in absolute excess (attributable) risks in women who have any 2 or all 3 of the major CVD risk factors (high blood pressure, high blood cholesterol level, and cigarette smoking). Although absolute mortality rates were lower in women than in men, absolute excess risks in women were substantial when 2 or more CVD risk factors were present. The data also show low rates of CHD, CVD, and all-cause mortality, especially in women but also in men, in the small percentages with none of the major CVD risk factors. The appropriate preventive emphasis seems consistent with current public health recommendations—by lifestyle means—to avoid development of risk factors in the first place, especially combinations of risk factors.

Accepted for publication February 19, 1998.

This study was supported by the Chicago Health Research Foundation, Chicago, Ill; Illinois Regional Medical Program, Springfield; Chicago Heart Association; American Heart Association, Dallas, Tex; Illinois Heart Association, Springfield; and grant R01 HL21010 from the National Heart, Lung, and Blood Institute, Bethesda, Md.

Presented in part at the 36th Annual Conference on Cardiovascular Epidemiology and Prevention, San Francisco, Calif, March 13, 1996.

We thank the Chicago companies and organizations, and their officers, staffs, and employees, whose volunteer efforts made the Chicago Heart Association Detection Project in Industry possible. This study would not have been possible without the efforts of numerous other people. The extensive list of acknowledgments has been previously published.⁴²

Reprints: Philip Greenland, MD, Department of Preventive Medicine, Northwestern University Medical School, 680 N Lake Shore Dr, Suite 1102, Chicago, IL 60611 (e-mail: p-greenland@nwu.edu).