Cardiovascular disease (CVD) is the leading cause of morbidity and mortality worldwide.1 Guidelines exist for the primary prevention of CVD and their aim is to establish recommendations to reduce risk,2 but there are considerable differences between men and women.3 According to data from the United States for the period 2013 to 2016, premenopausal women had lower rates of CVD than men of the same age, whereas postmenopausal women had higher rates.4 The difference was even greater for women in early menopause.5 Differences in health care are also a cause for concern. Because CVD presents differently in women, it is often diagnosed and treated later than in men, increasing the risk of poor outcomes.6

Many CV risk prediction models exist for the general population.7 While most do not account for the differential characteristics of women, a number of models in the United States have included female-specific risk factors.8,9

It has been shown that women with diabetes have a 58% greater risk of coronary heart disease and a 13% greater risk of all-cause mortality than men with the same condition.10 Similarly, hyperlipidemia appears to be associated with an increased CVD risk in postmenopausal women.11 No clear associations, however, have been identified for other classic risk factors such as smoking and alcohol consumption in this population.12 Obesity13 and sedentary behavior may increase the risk of CVD and hospitalization for heart failure in women after menopause.14

The aim of this study was to determine the influence of lifestyle habits and chronic disease on increased CV risk in a representative sample of women of menopausal age from a national health survey in Spain and to use our findings to build an easy-to-apply risk scale.

We performed a population-based retrospective observational cohort study using data from the 2011 National Spanish Health Survey (ENSE11) conducted by the National Statistics Institute (INE) between July 2011 and June 2012. This survey is representative of all residents aged over 18 years in Spain and the sample was obtained using a complex stratified 3-stage design. The study was conducted in accordance with the principles of the Declaration of Helsinki and was approved by the Office of Responsible Research at Universidad Miguel Hernández in Elche, Alicante.

We included women aged 50 years and older and excluded those with missing data for any of the study variables. Information was collected on sociodemographic variables, the population characteristics that best defined the lifestyle habits of the women included, self-reported health status and medical history at the time of the survey, and mortality data for a 6-year follow-up period (2011-2017). Mortality data were provided by the INE, which used probabilistic record linkage to match data from the ENSE11 to causes of death from the national death registry. We distinguished between mortality due to CVD (deaths due to diseases of the circulatory system [codes 100-199 in the 10th revision of the International Classification of Diseases]) and mortality due to other causes.

The following exposure variables were extracted from the ENSE11:

• •

  Sociodemographic variables: age, place of residence (autonomous community), social class based on occupation of reference person,15 country of birth, civil status, level of education, and net monthly household income.

• •

  Lifestyle variables: smoking, exposure to environmental smoke, alcohol consumption, hours of daily sleep, type of physical activity in main daily activity, frequency of physical activity in leisure time, dental hygiene, and consumption of fruit, vegetables, pulses, dairy products, sweet products, and fast food.

• •

  Health status: body mass index (BMI), use of a hearing aid, chronic or long-term diseases, current or past history of hypertension, myocardial infarction, other heart diseases, varicose veins, osteoarthritis, chronic neck pain, chronic lower back pain, allergies, asthma, chronic obstructive pulmonary disease, diabetes mellitus, stomach ulcer, urinary incontinence, high cholesterol, cataracts, chronic skin diseases, constipation, cirrhosis, migraine, hemorrhoids, osteoporosis, thyroid problems, stroke, malignant tumor, chronic depression, chronic anxiety, permanent injury due to an accident, or mental health.

• •

  Self-reported health and health-related quality of life: self-perceived health in the past 12 months, self-rated health on the 5-level EQ-5D (EQ-5D-5L) visual analog scale (VAS), and limitations in activities of daily living due to health problems in the past 6 months or longer.

• •

  Use of health services: admission to hospital in the past 12 months, admission to a day hospital in the past 12 months, emergency care in the past 12 months, primary care visit in the past month, specialist consultation in the past month, physical therapy in the past year, psychology consultation in the past year, radiography in the past year, computed tomography in the past year, ultrasound in the past year, magnetic resonance imaging in the past year, and vaccination in the last flu vaccination campaign.

The items within each category are shown in the methods section of the supplementary data.

Of the 21 007 individuals surveyed in the ENSE11, 6223 were women older than 50 years; 270 (4.3%) were excluded due to missing values, leaving 5953 for analysis. Their mean age was 66.4±11.4 years (range, 50-103 years). The number and percentage of women in each of the lifestyle and chronic CVD-related disease categories are shown in table 1, together with cumulative rates for CV mortality and mortality due to other causes. The estimated rate calculated for the 6 years of follow-up was 4% (n=239) for CV mortality and 7% (n=419) for mortality due to other causes. Mean age at death was 83.5±8.6 years for women who died of CVD and 78.6±10.8 years for those who died of another cause. Cumulative CV mortality and mortality due to other causes according to sociodemographic characteristics, other chronic diseases, self-rated health and quality of life, and use of health care services are shown in table 1 of the supplementary data. The mean EQ-5D-5L VAS score for self-rated health was 67.5±20.9 (50.3±21.9) for women who died of CVD and 52.2±23.0 for those who died of another cause.

The results of the multivariate Cox model for CV mortality adjusted for competing causes of death and all the study variables analyzed are shown in table 2. The model was built using a random sample of 4204 women (71%) and tested in the remaining 1749 (29%). The HRs for CV mortality are shown for each item within the predictor categories. Sitting during most of the day vs walking around and performing tasks requiring exertion in the person's main daily activity was the strongest predictor of CV mortality (HR=2.757). This was followed by active smoking (HR=1.816) vs having never smoked, sleeping for more than 9hours a day (HR=1.809) vs 7 to 9hours, eating vegetables less than once a week (HR=1.758) vs every day, and having been admitted to hospital in the past year (HR=1.700). The only chronic diseases that independently predicted CV mortality were diabetes mellitus (HR=1.522) and high cholesterol. BMI did not have independent predictive power in the presence of the other factors. As seen in the last row of table 2, there was a significant interaction between age and mortality due to other causes, such that the HR represents the difference in the effect of age on CV mortality and mortality due to other causes. In brief, the risk of mortality due to other causes was 0.955 times that of CV mortality. In other words, age contributed to a slightly increased risk of CV mortality compared with mortality due to other causes. The model fit the data well (LRT=1072.6; P<.001), met the proportional hazards assumption (P=.257), and showed high predictive capacity in the validation sample, with a C index of 0.8407 (95%CI, 0.8025-0.8789).

The results of the multivariate model were used to assign a score to each risk factor and build a CV risk scale for women based on their lifestyle habits and the presence of chronic diseases (table 3). BMI and high cholesterol levels were not included as they did not have independent predictive power in the multivariate analysis. The sum of scores for each predictor indicates the probability of CV mortality over 6 years (table 4). For example, an 85-year-old woman (5 points) who scored 80 on the EQ-5D-5L VAS (0 points), smoked (1 point), slept an average of 8hours a day (0 points), spent most of her day sitting (2 points), ate vegetables 3 times a week (0 points), did not have diabetes (0 points), and had been admitted to hospital once in the past year (1 point) would have a 33% risk of dying of CVD in the next 6 years (total score, 9 points). The curve showing the risk of CV mortality according to total score is shown in figure 1.

Figure 1.

Cumulative risk curve according to risk prediction scale scores.

(0.09MB).

We have shown that lifestyle habits are closely linked to CV mortality in women of menopausal age. Vegetable consumption, physical activity, smoking, and number of hours spent sleeping carried as much or more weight than chronic diseases or classic CV risk factors. We used our findings to design a simple 6-year CV risk scale based on self-reported data that had high predictive capacity.

CV risk in women of menopausal age is significantly influenced by lifestyle habits. Factors such as vegetable consumption and physical activity are very important in this population. Other factors, however, such as high cholesterol and obesity, which are strong predictors in men, do not appear to have predictive value in postmenopausal women.

No funding was received for this study.

J.A. Quesada, V. Bertomeu-González, J.M. Ruiz-Nodar, A. López-Pineda, and F. Sánchez-Ferrer contributed to the study design. J.A. Quesada requested the data and performed the statistical analyses. J.A. Quesada, V. Bertomeu-González, J.M. Ruiz-Nodar, A. López-Pineda, and F. Sánchez-Ferrer contributed to the interpretation of data. J.A. Quesada, V. Bertomeu-González, J.M. Ruiz-Nodar, A. López-Pineda, and F. Sánchez-Ferrer contributed to writing the manuscript. All the authors approved the final version of this manuscript.

None.