Cardiovascular Disease Epidemiology and Risk Factors in Primary Care

INTRODUCTION

Although mortality from cardiovascular diseases (CVD) has shown a tendency to decrease in developed countries, including Spain,¹ CVD nevertheless remains the main cause of death, especially ischemic heart disease and cerebrovascular disease (CD).^1-3 The origin of these diseases is multifactorial, but cardiovascular risk factors (CVRF) represent some of the most important causes. Of importance among CVRF are age and sex, as unmodifiable CVRF, and smoking, high blood pressure, hypercholesterolemia, and diabetes mellitus (DM) as modifiable CVRF.⁴ Recently, other CVRF, such as hypertriglyceridemia,⁵ have also been shown to play an important role. The impact of the different CVRF on the onset of CVD, however, differs among countries with similar levels of development.^6,7 This can be seen in the Mediterranean basin, for instance, where rates of ischemic heart disease are much lower than in countries in the north of Europe, which have a similar, or even better, CVRF profile.^6,7 This phenomenon has also been confirmed in Spain.⁸

Study of the epidemiology of CVRF in our setting suffers from several limitations. Firstly, few studies have been undertaken in persons older than 65 years of age, which is the fastest growing segment of the population and which has the highest number of CVD. Secondly, recent demographic, cultural, and lifestyle changes are modifying the prevalence of these CVRF, a clear example of which is the pattern of smoking according to sex. Thirdly, CVRF are often not stratified by age and sex, and only a few studies have used sampling techniques. Finally, up-to-date studies are required which make use of the new diagnostic criteria, as in DM. Similar limitations surround the epidemiology of CVD, although added to these are the low number of studies of patients with CD and, especially, peripheral artery disease. Accordingly, the purpose of this study was to examine the epidemiology of the main CVRF and CVD using a cross-sectional descriptive design in an urban health care center.

PATIENTS AND METHODS

Study Type and Patient Selection

We undertook a descriptive, cross-sectional study in an urban health care center located in a suburb of Barcelona, Spain. This suburb has a reasonably heterogeneous population regarding social and economic status, a high rate of patient visits and a total of 35 275 clinical histories (CH). The patients were selected by simple random sampling of the CH files and were included if they were older than 15 years of age and had at least 5 visits recorded in the CH, with at least 3 over the previous 2 years. If the CH failed to fulfil these conditions, then the next consecutive CH was chosen.

Data Collection

Data were collected according to a standardized protocol which included age (in years) and sex as non-modifiable CVRF, as well as various modifiable CVRF and CVD. A pilot study was carried out with 20 cases in order to detect any errors in design, to train the 12 physicians who were to review the CH and to verify the degree of concordance between the survey takers and a reference pattern (consensus review). The data were collected between June 1998 and May 1999. To reduce the proportion of lack of responses we contacted the physicians responsible for those patients whose CH failed to include all CVRF and telephoned the patients whose CH failed to record details about smoking habits (a minimum of three calls at different times of day and at least one call on a Saturday or Sunday).

Modifiable Cardiovascular Risk Factors

1. Smoking. This variable was coded dichotomously as either smoker or nonsmoker/ex-smoker (more than 12 months without smoking) if the smoking habit during the previous 2 years was reflected in the CH.⁹

2. High blood pressure. Patients were classified as having hypertension, if they took antihypertensive drugs or they had 3 blood pressure recordings ≥140/90 mm Hg at 3 consecutive visits, or as not having hypertension, taking into consideration their blood pressure recordings over the previous four years if they were aged 15-40 years and the previous 2 years if they were aged over 40 years.¹⁰ The last blood pressure reading was also recorded.

3. Hypercholesterolemia. This was coded as a dichotomous variable (yes/no) depending on a cut-off point of 250 mg/dL on at least 2 occasions, which were considered valid if they were recorded in the CH during the previous 6 years.¹⁰ The last value in the CH was also recorded.

4. Hypertriglyceridemia. This was coded according to a cut-off point of 200 mg/dL (2 measurements, considered valid if they were included in the CH over the previous 6 years).¹⁰ The last value in the CH was also recorded.

5. Diabetes mellitus. The criteria used were those of the American Diabetes Association and the World Health Organization, adopted¹¹ by the Spanish Society of Family and Community Medicine: characteristic clinical features plus a random blood glucose reading >200 mg/dL, 2 baseline plasma blood glucose readings ≥126 mg/dL or a 2-hour oral glucose tolerance test ≥200 mg/dL. Persons already diagnosed with diabetes or who were receiving treatment with insulin or oral antidiabetic agents were also included. Baseline blood glucose levels were only taken into account for the previous 3 years¹¹ and the last value in the CH was also recorded.

Cardiovascular Diseases

1. Ischemic heart disease: clinical diagnosis, confirmed from the hospital discharge report, of acute myocardial infarction (AMI) obtained by enzyme measurements, or angina diagnosed by a compatible clinical picture plus electrocardiogram, exercise stress test, scintigraphy, or coronary angiography.

2. Cerebrovascular disease: clinical diagnosis, confirmed from the hospital discharge report, of transient ischemic attack (with computerized tomography [CT] or normal magnetic resonance imaging [MRI]) or established stroke confirmed by CT or MRI.

3. Peripheral artery disease: explicit clinical diagnosis in the CH or hospital reports of intermittent claudication, gangrene or ischemic ulcers, or diagnostic arteriography or Doppler ultrasound.

Statistical Analysis

Statistical analysis was done with SPSS. The concordance of the pilot study was studied by kappa statistics (values >0.80 were considered to represent good concordance). The sample size needed to estimate the prevalence of CVRF and CVD was calculated using the GRANMO program,¹² with a 2-sided alpha of 5%, a precision of 2% and the maximum possible uncertainty (P=.5). The prevalence of the CVRF and CVD studied was stratified by sex and age into 10-year groups, except above the age of 75 years, which was considered as a single category. For the quantitative variables blood pressure, cholesterol, triglycerides, and baseline glycemia the results were stratified in age groups of 25-74 years and 35-64 years as the mean and standard deviation (SD), to facilitate comparison with other studies. Proportions were compared by χ² test, means by the Student t test for independent data or by the corresponding nonparametric test. An alpha <.05 was used as the level of statistical significance in all cases.

RESULTS

The concordance study carried out among those taking the survey produced values >0.80 for all the variables. The study included a total of 2248 patients, with a mean age of 49.1±18.9 years and 53.5% women. Table 1 shows the distribution by age and sex in 10-year strata.

The type of modifiable CVRF and the non-response rate are shown in Table 2. Most patients (57.9%) had at least 1 modifiable CVRF. A total of 803 patients (35.7%) had 1 modifiable CVRF, 338 (15.0%) had 2, 115 (5.2%) had 3, and 46 (2.0%) had more than 3 risk factors.

Table 3 shows the distribution by sex. Men had a significantly higher proportion of modifiable CVRF (P<.05), except for high blood pressure, for which no significant differences were observed.

In the 25-74 year age group, the mean systolic and diastolic BP were 128.4±18.7 and 78.4±10.7 mm Hg, respectively; the mean level of cholesterol was 212.9±41.3 mg/dL, of triglycerides 127.4±78.7 mg/dL and glycemia 105.6±38.6 mg/dL. In the 35-64 year age group, the mean systolic and diastolic BP were 127.5±17.2 and 79.4±10.5 mm Hg, respectively, the mean level of cholesterol was 217.5±39.3 mg/dL, of triglycerides 128.9±81.5 mg/dL, and glycemia 103.8±36.0 mg/dL.

Table 4 shows the distribution of the modifiable CVRF according to age and sex. Of note are the following aspects: a) the important proportion of women younger than 45 years of age who smoked; b) the increase in the prevalence of hypertension in both sexes with age, such that most of the older patients had hypertension; c) the progressive increase in the prevalence of hypercholesterolemia with age, although it tended to fall with effect from the age of 74 years; d) the similar pattern of hypertriglyceridemia to that of cholesterol in the women, although the rising curve was not so evident in the men, with a high prevalence among those 35-44 years of age, and e) the high prevalence of DM (15.8%), which also rose with age, with almost 30% of persons older than 65 years of age having diabetes, although in the men aged 35-45 years the prevalence of diabetes was already 8.3%.

The proportions in the 25-74 year age group were as follows: smoking: 36.1% (46.2% in men and 26.6% in women), high blood pressure: 32.2% (32.2% in men and 32.1% in women), hypercholesterolemia: 22.5% (21.0% in men and 23.8% in women), hypertriglyceridemia: 13.8% (19.1% in men and 9.2% in women) and diabetes mellitus: 15.5% (18.3% in men and 13.2% in women). The proportions in the 35-64 year age group were smoking: 38.5% (49.1% in men and 28.3% in women), high blood pressure: 28.5% (29.1% in men and 27.5% in women), hypercholesterolemia: 22.3% (22.7% in men and 21.9% in women), hypertriglyceridemia: 13.8% (21.7% in men and 6.9% in women) and diabetes mellitus: 12.6% (15.5% in men and 10.2% in women). A total of 224 patients (10.0%) had CVD. Ischemic heart disease was present in 123 patients (5.5%), CD in 84 (3.7%), and peripheral artery disease in 55 (2.4%). Most of the patients had just one recorded CVD (192 cases, i.e., 8.5%), whereas 28 patients (1.3%) had 2 CVD and 4 had 3 CVD (0.2%).

Table 5 shows the distribution of the CVD according to sex. The men had a higher proportion (12.2%) of these events than the women (8%), the difference being statistically significant (P<.05) except for CD. Finally, Table 6 shows the CVD stratified by age groups and sex. Of note was the low proportion of patients younger than 55 years with a CVD, especially in the women, and the progressive increase with age. Age is a very relevant CVRF, with almost one in 3 patients older than 74 years of age having at least 1 CVD. The men had a higher risk of having CVD in almost all the age groups, and most cases of ischemic heart disease involved patients older than 65 years. This tendency was more marked in CD (especially among those older than 74 years). Peripheral artery disease mainly affected men older than 55 years of age.

DISCUSSION

The results of this study show a high prevalence of modifiable CVRF in our area. The prevalence was greater in men and tended to increase with age. Of special note was the situation in patients with DM. Although reducing the cut-off point to 126 mg/dL might slightly increase the diagnostic prevalence of the disease,¹³ the prevalence of this CVRF may be increasing as a result of the increasing overall life expectancy and the rise in obesity and unhealthy lifestyles. The low prevalence of CVD in men <55 years and women <60 years was notable.

We shall now comment on our results in comparison with other studies undertaken in our area using some sort of sampling technique among adult persons, irrespective of sex, and which include patients up to at least 74 years of age.

With regard to smoking, men smoked more than women and the number of smokers tended to decrease with age, similarly to the trend reported in other studies.^14-23 However, the difference between sexes was much less in persons younger than 45 years of age, as can be seen in Table 4. In our study the men smoked less and the women more than the corresponding data reported in the National Health Survey,¹⁴ which confirms that the patterns of smoking habits are changing,^14-23 with a reduction in men and an increase in women.

With regard to high blood pressure, the results are similar (Table 4) to those of other studies,^15,18,23-29 although the older studies show a lower prevalence.²⁶ The prevalence of hypertension may also be increasing, as a consequence of the aging of the population and the increase in unhealthy lifestyles.

Similar considerations can be made regarding studies of dyslipidemia,^{15,17,18,23,25,30-32} although in this case comparisons are more difficult owing to the variation in cut-off points between the different studies. The greater consumption of alcohol, especially among young and middle-aged persons, may account for the different distribution according to age and sex (Table 4).

Special consideration should be given to DM, which is the CVRF whose prevalence has increased most in recent years. Studies of DM undertaken a decade ago^33,34 showed a prevalence of around 6%. More recent studies,^{13,15,17,35-39} however, generally show a higher prevalence, varying from 6.7% to 18.3%, although the different diagnostic criteria used and the inclusion criteria by age hinder comparison of the results. The aging of the population, unhealthy lifestyles and obesity are all raising the prevalence of DM⁴⁰; as a consequence, this phenomenon may be classified as the epidemic of the 21st century.

Fewer epidemiologic studies concerning the prevalence of CVD have been undertaken in our area. Most studies of ischemic heart disease have been done from hospital records. Nevertheless, the REGICOR⁴¹ group examined the incidence of AMI and the PANES⁴² study, which analyzed the prevalence of angina in different autonomous regions of Spain, reported a prevalence of 7.5% in the population aged 45-74 years. A recent study⁴³ suggested that the demand for health care due to ischemic heart disease is tending to increase. Regarding CD, most studies have also been undertaken in a hospital setting. Caicoya et al⁴⁴ studied the incidence of CD in Asturias and López Pousa et al⁴⁵ studied the prevalence in persons older than 64 years of age, giving a figure (4%) lower than that seen in our results, although their study was undertaken in a rural setting. A large hospital series of patients with peripheral artery disease⁴⁶ coincides with the results of our study in the high proportion of men (80%) and older age. The use of mainly clinical diagnostic criteria may have led to a lower prevalence of peripheral artery disease than expected, especially in older persons.

One of the limitations of the study may involve a selection bias. The large population covered by the health center (the number of CH is greater than the population census) and selection by random sampling tend to guarantee their absence. Information bias may also have occurred. However, in an attempt to avoid this we undertook a pilot study to train the survey takers and confirm the inter-observer concordance, with excellent results on the kappa index. Likewise, strict criteria were used for the cut-off points of the quantitative variables, the period during which the information was considered valid and the definition of the variables. Moreover, we attempted to minimize loss of information by contact with the physicians and by telephone contact with the patients in the case of smoking habits. The main problem in cross-sectional descriptive studies is usually related to lack of responses, which in the present study was increased owing to the strict criteria used. The high prevalence of some risk factors, such as smoking or diabetes, may thus be an overestimate.

CONCLUSIONS

The results of this study show a high prevalence of modifiable CVRF, which was greater in men and in older patients. Of note was the low prevalence of CVD in men and women younger than 55 and 60 years of age, respectively. The increase in life expectancy might possibly raise the prevalence of CVD, which is still the main cause of death in Spain.^1-3 These diseases are closely related with CVRF, which have an accumulative, additive and progressive effect on the incidence of CVD. An overall strategy is required to prevent CVD by means of the prevention, control and treatment of CVRF, in order to reduce the incidence of CVD.

See Editorial on Pages 338-40

---

Correspondence: Dr. J.M. Baena Díez.
Sant Joan, 123, esc. 2, 3.o 2.a. 08150 Parets del Vallès. Barcelona. España.
E-mail: [e-mail]

Received July 7, 2004.
Accepted for publication January 20, 2005.