Abstract
Background Overweight and obesity are positively correlated with increased risk of morbidity and mortality.
Aim To evaluate whether obesity may be considered an independent cardiovascular risk factor in patients of ages from 35 to 74 years followed-up for 10 years.
Design of study Observational, longitudinal retrospective study.
Setting Primary care practices in Badajoz (Spain).
Method A cohort of 899 patients (mean 55.7 years; 58.2% female) without evidence of cardiovascular disease was studied.
Results A total of 33.5% of the population were obese (body mass index ≥30 kg/m2). Patients meeting the obesity criteria were more commonly female (36.6%) and were older, had higher mean values of blood pressure and triglycerides, higher percentages of diabetes, and higher coronary risk using either the original Framingham or the Framingham function calibrated for the Spanish population (Framingham-REGICOR). During the follow-up period, the rates of cardiovascular events and death in patients with obesity tended to be higher: 16.3% versus 11.7%, P = 0.056 and 4.7% versus 2.2%, P<0.05, respectively. In the final model of the logistic regression multivariate analysis, the significant predictors of cardiovascular events in patients with obesity were age, sex (male), diastolic blood pressure, diabetes, and smoking. The highest odds ratio corresponded to smoking (odds ratio 2.03; 95% confidence interval = 1.22 to 3.38).
Conclusion Obesity may not be considered an independent cardiovascular risk factor in patients aged from 35 to 74 years followed-up for 10 years.
INTRODUCTION
Obesity is a serious and prevalent disorder. Overweight and obesity are the most common nutritional disorders in the US, affecting the majority of adults in the country. Given a normal body mass index (BMI) ranging from 18.5 to 24.9 kg/m2, 34% of the adult population is overweight (BMI, 25–29.9 kg/m2), and another 27% is obese (BMI ≥30 kg/m2).1 More recent data confirm the epidemic of obesity in most racial and ethnic groups.2 In Spain it is estimated that the prevalence of obesity among people aged 25–60 years is 14% and that 8.5% of all deaths are due to obesity.3,4 In addition to their negative effects on health and quality of life, obesity and associated comorbidities may have a considerable impact on healthcare expenditures.5
Several studies have proved that overweight and obesity are positively correlated with increased risk of morbidity and mortality.6–16 Also a number of studies have documented the association between obesity and cardiovascular disease risk factors and with markers of subclinical cardiovascular disease.17–19 A recent study indicates that adverse effects of obesity (BMI ≥30 kg/m2) on blood pressure and cholesterol levels could account for about 45% of the total increase of risk of coronary heart disease due to obesity.20 Therefore, part of the increased risk of coronary heart disease in obese individuals is independent of the effects on blood pressure and cholesterol. This implies that even under the theoretical scenario where optimal treatment is available against hypertension and hypercholesterolaemia in overweight persons, they would still have an elevated risk of coronary heart disease. Whatever the mechanisms involved, there is general acceptance that obesity is a risk factor for cardiovascular and kidney diseases.21,22 However, the utility of BMI as a cardiovascular risk factor has been questioned.23
The aim of this study was to determine the risk of developing a cardiovascular event associated with the presence of obesity in patients aged 35–74 years during a 10-year follow-up period, controlling the estimated effect for potential confounding factors.
METHOD
The design of the present work was an observational longitudinal study of a retrospective cohort of 1011 patients aged 35 to 74 years (9% of the total population of this age group attending a primary healthcare centre), without pre-existing cardiovascular disease, and followed over 10 years. Of this cohort, 899 patients were included in the study. The remaining 112 patients (11% of the cohort) were excluded by the absence of height data. There were no differences between the excluded and the included patients concerning the estimated coronary risk or the rate of cardiovascular events. The criterion for inclusion in the study cohort was that their medical history included, before 1 January 1995, the variables necessary for the purpose of this study: weight, height, and the laboratory variables needed to estimate: (1) the glomerular filtration rate (GFR) by applying the Cockcroft–Gault formula,24 corrected for body surface area,25 and the simplified Modification of Diet in Renal Disease (MDRD) Study formula;26 and (2) coronary risk with the original Framingham function27 and the Framingham function calibrated for the Spanish population (Framingham-REGICOR).28 Additional data recorded included glycaemia, triglycerides, total cholesterol, low-density lipoprotein (LDL) cholesterol, and medication with hypolipidaemic or antihypertensive drugs. The cardiovascular events investigated were those included in the calculation of total coronary risk (angina and myocardial infarction, fatal and non-fatal) and fatal cardiovascular disease (cardiac death of coronary and non-coronary origin, death of cerebrovascular origin, and deaths from other cardiovascular causes). Acceptance as an event of cardiovascular origin required confirmation of the diagnosis by specialists in the field or by the pertinent tests in the referral hospital (such as, thallium stress tests and coronary angiography). Similarly, the acceptance of a cardiovascular cause for death required confirmation in the hospital archives, inquiry in the civil registry office to review the death certificate, and contact with relatives by telephone for confirmation of the event.29
How this fits in
The increased relative risk of cardiovascular disease and diabetes for overweight and obese individuals compared to those of normal weight is well documented. The role of obesity as an independent cardiovascular risk factor at middle and old age is conflicting. The results of this study do not confirm that a high body mass index (BMI) or obesity (BMI ≥30 kg/m2) behave as independent predictors of cardiovascular events. The cardiovascular risk associated with a high BMI could be influenced by the development of diabetes, hypertension, dyslipidaemia, or other risk factors associated with obesity.
Statistical analysis
The statistics used as representative of the sample for the univariate descriptive analysis were the mean and standard deviation for normal distributions, the median and quartile 1-to-3 range for non-normal distributions, and the observed frequencies and proportions for categorical variables. The normal distribution of the variables was verified by normality plots.
In the bivariate analyses, a t-test for independent samples was used for normally-distributed quantitative variables, a non-parametric Mann–Whitney U test for non-normal variables, and a χ2 test or Fisher's exact test for categorical variables. Due to the high number of statistical tests performed, the minimum significance level was taken to be P<0.01.
Multivariate analysis was performed using a binary logistic regression model, and included all the variables that were clinically significant or showed a P-value <0.10 in the bivariate analyses. Backward regression analysis was used to select the best predictor variables, and the odds ratio (OR) was used as a measure of risk, with a 95% confidence interval (CI).
The data were processed and analysed using the software packages SPSS for Windows (version 14.0) and Epi Info (version 6.04).
RESULTS
Table 1 lists the relevant baseline characteristics of the study subjects. The mean age was 55.7 years, 58.2% were female, 33.5% of the patients were obese (BMI ≥30 kg/m2), 23.8% had diabetes, and 40.5% used antihypertensive drugs. Obesity was more frequent in female patients, and the proportion of smokers was higher in males (43.9%, Table 1). During the 10-year follow-up period there was a higher percentage of coronary and cardiovascular events in male patients (Table 1).
Table 1 General characteristics of the patients in the study.
Obese patients showed higher systolic blood pressure (SBP) (148.2 versus 139.0 mmHg, P<0.001), higher diastolic blood pressure (DBP) (89.3 versus 83.8 mmHg, P<0.001), and higher proportion of diabetes (33.6% versus 18.9%, P<0.001) and cardiovascular deaths (4.7% versus 2.2%, P<0.05). Moreover, obese patients had lower mean values of HDL-cholesterol (48.7 versus 52.7 mg/dl, P<0.001) and higher mean values of triglycerides (133.5 versus 110.0 mg/dl, P<0.001).
Obese males showed higher values of triglycerides (150.5 versus 122.5 mg/dl, P<0.001) and smoking (47.5% versus 6.5%, P<0.001), and a higher incidence of coronary events during the 10 years of follow-up (17.2% versus 9.2%, P<0.05), when compared to obese females.
Patients who developed cardiovascular events during the follow-up period (Table 2) were older (61.2 versus 54.9 years, P<0.001), predominantly male (55.5% versus 42.1%, P<0.01), with higher values of SBP, DBP, and BMI, and higher percentages of diabetes and smoking, and lower values of HDL-cholesterol and GFR estimated by the Cockroft–Gault formula.
Table 2 Characteristics of the patients with and without cardiovascular events
The dependent variable in the logistic regression multivariate analysis was the presence of cardiovascular events, and obesity (BMI ≥30 kg/m2) was the main independent variable. The initial model introduced all the variables that could act as confounding factors identified in the bivariate analysis (Table 2) and the interactions of the main independent variable with each of them. In the final model (Table 3), the significant predictors of cardiovascular events were age, sex (male), DBP, diabetes, and smoking. The highest OR corresponded to smoking (2.03; 95% CI = 1.22 to 3.38). Similar results were obtained when the BMI was taken as a quantitative variable in the regression model. The highest ORs corresponded also to smoking (2.02; 95% CI = 1.20 to 3.41, P<0.01), sex (male) (1.74; 95% CI = 1.10 to 2.76, P<0.05), diabetes (1.57; 95% CI = 1.03 to 2.49, P<0.05) and age (1.09; 95% CI = 1.05 to 1.10, P<0.001).
Table 3 Predictor variables of cardiovascular events in the logistic regression multivariate analysis after categorisation of BMI as ≥ or <30 kg/m2.
DISCUSSION
Summary of main findings
This observational study investigated the relationship between obesity and cardiovascular events and, specifically, the extent to which this relationship is mediated by adverse effects of obesity on other cardiovascular risk factors. The study reveals that obese patients showed a higher percentage of cardiovascular events and deaths during a follow-up period of 10 years. However, obesity (BMI ≥30 kg/m2) does not behave as an independent predictor of cardiovascular events after adjusting for several confounding factors. The study also confirmed an increased presence of cardiovascular risk factors in the obese population. This aggregation of cardiovascular risk factors explains the increased coronary risk using the original Framingham and the calibrated Framingham-REGICOR charts and could justify the higher number of cardiovascular events and deaths during follow-up in patients with BMI ≥30 kg/m2. In the final model of the logistic regression multivariate analysis, the significant predictors of cardiovascular events were age, sex (male), DBP, diabetes, and smoking, and the highest OR corresponded to smoking. These findings raise the question, still unresolved, of whether obesity is an independent cardiovascular risk factor and, therefore, can be used as an additional modifiable risk factor in risk stratification schemes like the Framingham equation, or rather, diseases associated with obesity (diabetes, hypertension, lower values of HDL-cholesterol, or higher values of triglycerides) are responsible for the largest number of cardiovascular events in obese individuals.
Strengths and limitations of the study
The strengths of this study are the large number of participants attending a primary healthcare centre, the detailed information on multiple baseline covariates including age, sex, smoking, glycaemia, triglycerides, total cholesterol, LDL-cholesterol, use of lipid-lowering and antihypertensive drugs, estimation of GFR and coronary risk, and different cutoff values of BMI (data not shown); however, it does also have limitations.
The population was not randomly selected, but corresponded to patients who had been attended to at the centre and had a clinical history available that included the information necessary for this study. Reverse causation owing to pre-existing chronic disease and inadequate control for smoking status can distort the true relation between body weight and the risk of death, as chronic illness and smoking are associated with both decreased BMI and increased risk of death. Statistical adjustment for smoking status does not fully address the problem. Restriction of analyses to patients who have never smoked is a powerful tool for addressing this potential bias, but the relatively small number of non-smokers in the study cohort makes this approach difficult (in any case, no differences were found in the percentages of various categories of cardiovascular events among non-smoker patients with BMI <30 kg/m2 and ≥30 kg/m2).
This study was also limited because there were no values of waist circumference available and it is not known if those values were associated with an increased cardiovascular risk in the study cohort, as has been demonstrated recently.30,31 Moreover, there were no data for a statistical adjustment for other potential confounding factors such as physical activity and diet. The risk of coronary heart disease associated with elevated BMI is considerably reduced by increased physical activity levels,32 but, in the Nurses' Health Study, adjustment for diet had virtually no impact on the association between BMI and risk of coronary heart disease.33 Finally, a follow-up period of 10 years is perhaps small, if one considers that the full effect of obesity on cardiovascular mortality may begin after 15 years or more.34
Comparison with existing literature
The study results coincide with another study that showed that the association between overweight and death from atherosclerotic cardiovascular causes is attenuated to statistically non-significant levels after adjustment for blood pressure, cholesterol level, and blood glucose level.35 Furthermore, findings of reduced or no association of increased BMI with subsequent mortality in older subjects have been reported in several longitudinal studies.36–39 In contrast, other studies point to an association between BMI and coronary heart disease, stroke, and cardiovascular disease mortality among white and black individuals.14 There are also reports of increased mortality associated with being underweight or obese, especially with higher levels of obesity, while overweight and mildly obese patients do not have a significantly higher risk.40,41
These results as a whole, and also the present study, support the hypothesis that BMI alone does not appear to be adequate in overall risk assessment, and measures of fat distribution are necessary to provide a more comprehensive assessment of morbidity and mortality risk.42 An increased risk of cardiovascular morbidity and mortality has been reported in patients with increased waist-to-hip ratio and waist and hip circumferences, even with a normal BMI,30,31,43 suggesting that waist circumference and not BMI explains obesity-related health risk.44,45 Nevertheless, other studies have emphasised that higher levels of adiposity, however measured, confer increased risk of incident cardiovascular disease and these data show that BMI should be the single measure of obesity.21,46
Implications for future research
There is a great deal about the relationship between obesity, mortality, and disability that is still not understood, and, given the cost, the difficulties, and the burden associated with treating obesity, there is an overwhelming need for research that addresses these questions.32 Determining the best method for quantifying adiposity is important because this may reveal insights into the mechanisms by which obesity is linked to cardiovascular diseases.21
In summary, the results of this study do not confirm that a high BMI or obesity (BMI ≥30 kg/m2) behave as independent predictors of cardiovascular events, and suggest that the cardiovascular risk associated with a high BMI could be influenced by the development of diabetes, hypertension, dyslipidaemia, or other risk factors associated with obesity. Therefore, obesity alone would not justify the intensification of pharmacological cardiovascular prevention, but its association with the mentioned diseases, in addition to quality-of-life aspects, make the reduction of BMI advisable. Finally, because of the high prevalence of obesity and the expected future increases, it is essential to gain precise insight into the consequences of obesity for health and into the metabolic mechanisms that link the two.
Notes
Funding body
This work was supported by redIAPP (Innovation and Integration of Prevention and Health Promotion in Primary Care), thematic cooperative research network G03/170 and by a grant from the Program for Promotion of Research in Primary Care, of the Instituto de Salud Carlos III. The second and fourth authors of the article also received a predoctoral scholarship from the Spanish Society of Family and Community Medicine.
Competing interests
The authors have stated that there are none.
- Received August 15, 2009.
- Revision received November 17, 2009.
- Accepted February 23, 2010.
- © British Journal of General Practice, 2010.
