Cardiovascular risk assessment in the metabolic syndrome: results from the Prospective Cardiovascular Munster (PROCAM) Study

Transcription

1 (28) 32, S11 S16 & 28 Nature Publishing Group All rights reserved 37-6/8 $3. ORIGINAL ARTICLE Cardiovascular risk assessment in the metabolic syndrome: results from the Prospective Cardiovascular Munster (PROCAM) Study G Assmann 1,2, H Schulte 1 and U Seedorf 1 1 Leibniz-Institute of Arteriosclerosis Research, University of Münster, Münster, Germany and 2 Institute of Clinical Chemistry and Laboratory Medicine, University of Münster, Münster, Germany Objectives: We aimed (1) to construct a modified PROCAM risk algorithm, which incorporates BMI/waist circumference in a model for predicting coronary events; (2) to evaluate how accurate this and the previously established PROCAM risk algorithm predict coronary risk in individuals with metabolic syndrome. Design: Prospective Cardiovascular Münster (PROCAM) Study, a prospective study of men and women at work in the northwest of Germany. Subjects: A total of 7134 men aged 3 6 years at study entry. Measurements: On the basis of 44 major coronary events (defined as nonfatal MI and coronary deaths), which occurred within 1 years of follow-up, a modified PROCAM risk algorithm was constructed by incorporating BMI/waist circumference as fixed variable in a Cox proportional hazards model for predicting coronary events. The metabolic syndrome was defined according to the latest recommendations proposed by the NCEP-ATP III Panel. Results: Men who were classified as having the metabolic syndrome (n ¼ 232, prevalence: 32.6%) were 2.9-fold more likely to experience a major coronary event within 1 years of follow-up than men not having the metabolic syndrome. In men with metabolic syndrome, the observed major coronary event rate of 9.6% corresponded well with their estimated global risk according to the modified BMI-based PROCAM risk algorithm (1.2%). Comparative calculations performed with the previously published fully adjusted PROCAM algorithm yielded very similar results. Conclusion: Both PROCAM algorithms provide very accurate means to ascertain coronary risk in male patients with metabolic syndrome. (28) 32, S11 S16; doi:1.138/ijo Keywords: epidemiology; body mass index; metabolic syndrome; coronary heart disease; myocardial infarction Introduction In western societies like the European Union or the United States, roughly every fifth adult man and every fifth adult woman is affected by cardiovascular disease. 1 Despite considerable progress over the past 1 years with respect to treatment of coronary heart disease, its prevention remains a major public health challenge. Benefits that came with the widespread introduction of effective cholesterol lowering drugs (that is, statins) and a better awareness concerning a healthy lifestyle are counteracted by the current epidemic of overweight and obesity in many parts of the world. 2 Excess Correspondence: Dr G Assmann, Leibniz-Institute of Arteriosclerosis Research, Domagkstrasse 3, University of Münster, Münster 48149, Germany. assmann@uni-muenster.de body weight is a well-established component of the metabolic syndrome, which is an expanding health threat due to its tight association with diabetes, hypertension, dyslipidaemia and atherosclerosis. 3,4 The metabolic syndrome accounts for 6 7% of all-cause mortality in the United States, and it may be expected that life expectancy will plateau or even decline in the United States and other western societies within the first half of this century as a consequence of rising incidence rates of obesity-related morbidity. Two widely used clinical definitions of the metabolic syndrome have been proposed. The latest version of the definition developed by the NCEP-ATP III Panel 6 relies on the presence of at least three out of the following five simple criteria: (1) increased waist circumference with population-specific cutoff values, (2) increased levels of

2 S12 fasting triglycerides or treatment for hypertriglyceridaemia, (3) low high-density lipoprotein (HDL) cholesterol levels or treatment for this condition, (4) elevated blood pressure or antihypertensive treatment and () elevated blood glucose or treatment with a hypoglycaemic agent. The definition by the International Diabetes Federation (IDF) requires the presence of central obesity plus any two of the following four criteria: (1) elevated fasting triglycerides, (2) reduced HDL cholesterol, (3) hypertension and (4) raised plasma glucose or previous diagnosis of type 2 diabetes. 7 Irrespective of which definition is used to diagnose the metabolic syndrome, its presence is associated with increased risk for cardiovascular disease. 8 Despres and Lemieux 9 have recently introduced a concept that suggests that the presence of the metabolic syndrome should be considered separately from traditional risk factors. On the other hand, this concept has been questioned because it may solely depend on differences in the evaluation of cardiovascular risk by different risk scores. To provide further insights into this debate, we constructed a modified risk algorithm, which incorporates body mass index (BMI)/waist circumference as fixed variable in a Cox proportional hazards model. In addition, we evaluated this new risk algorithm with respect to its ability to predict coronary risk in men with and without the metabolic syndrome. Methods Study design and recruitment of participants Recruitment to the PROCAM Study 1,11 was started in The basic study design has been described elsewhere. 12 We certify that all applicable Institutional and Governmental Regulations concerning the ethical use of human volunteers were followed during this research. The sample of 7134 men (3 6 years of age) used for the current evaluation represented a subcohort of the PROCAM Study recruited before The basic characteristics of the study population are shown in Table 1. In all, 44 major coronary events occurred within 1 years of follow-up, enough to allow statistically valid longitudinal analysis. A total of 29 participants died from other causes, and 6 nonfatal cases of stroke occurred. Table 1 Characteristics of the study population. Outcome in men with follow-up of 1 years Age at study entry, years 3 6 Men with 1 years follow-up 78 Men with missing data or TG levels 44 mg per 1 ml 41 Men studied 7134 Nonfatal MI and CHD deaths 44 Cerebrovascular events 6 All deaths except for CHD or CVD 29 Deaths from malignant neoplasms 16 Abbreviations: CHD, coronary heart disease; CVD, cardiovascular disease; MI, myocardial infarction; TG, triglyceride. Risk algorithms The PROCAM risk algorithm employed in this study was from the previously published Cox proportional hazards model. 12 The algorithm incorporates the following eight independent variables: age, low-density lipoprotein (LDL) cholesterol, smoking, HDL cholesterol, systolic blood pressure, family history of premature myocardial infarction, diabetes mellitus and triglycerides. The modified PROCAM risk algorithm was constructed by incorporating BMI as fixed variable in a Cox proportional hazards model along with 6 other clinical and laboratory variables measured in the PROCAM Study. 13 Definition of the metabolic syndrome The metabolic syndrome was defined according to the recommendations proposed by the NCEP-ATP III Panel. 6 BMI data were converted into waist circumference values according to the following equation: BMI (kg/m 2 ) ¼.319 waist (cm) 3.3. The equation resulted from linear regression analysis between BMI and waist circumference (r ¼.87) obtained from studying a representative subcohort of the PROCAM Study consisting of 4896 men aged 18 6 years. A BMI of 29.2 kg/m 2 corresponded to a waist circumference value of 12 cm and was used as cutoff value in the diagnosis of the metabolic syndrome according to the NCEP-ATP III definition (Table 2). Results The strong correlation that existed between BMI and waist circumference (r ¼.87) in a representative subsample of the PROCAM cohort consisting of 4896 men suggested that BMI and waist circumference provided similar information (Table 2). We therefore developed the modified PROCAM risk algorithm by incorporating BMI as fixed variable in a multilogistic regression model along with 6 other clinical and laboratory variables measured in the PROCAM study. Table 2 Conversion of BMI data into waist circumference values in men according to the equation BMI (kg m 2 ) ¼.319 waist (cm) 3.3 (for details see Methods section) BMI (kg/m 2 ) Waist circumference (cm) a 12 a Abbreviation: BMI, body mass index. a Waist circumference cutoff value for men according to the NCEP-ATP III definition of the metabolic syndrome.

3 Table 3 Independent variables used in construction of the Cox proportional hazards model with BMI as fixed variable Coefficient of Cox model P Hazard ratio 9% CI S13 Age, years.1 o LDL cholesterol, mg per 1 ml.12 o Smoking.77 o HDL cholesterol, mg per 1 ml.28 o Pulse pressure, mm Hg.14 o Diabetes mellitus.8 o Family history of MI Body Mass Index, kg/m Abbreviations: BMI, body mass index; CI, confidence interval; HDL, high-density lipoprotein; LDL, low-density lipoprotein; MI, myocardial infarction. Table 4 Variables used in construction of the classical fully adjusted Cox proportional hazards model Coefficient of Cox model P Hazard ratio 9% CI Age, years.13 o LDL cholesterol, mg per 1 ml.13 o Smoking.68 o HDL cholesterol, mg per 1 ml.32 o Systolic blood pressure, mm Hg.1 o Family history of MI Diabetes mellitus Triglycerides, mg per 1 ml, logarithmically transformed Abbreviations: CI, confidence interval; HDL, high-density lipoprotein; LDL, low-density lipoprotein; MI, myocardial infarction. Ranking of Risk Factors R MIs (%) in 1 Years I II III IV V Quintiles of Cox Proportional Hazards Model 1. Age 2. LDL cholesterol 3. Smoking 4. Diabetes/ Glucose 12 mg per 1 ml. Pulse pressure 6. HDL cholesterol 7. Waist Circumference (BMI) 8. Family history of MI Figure 1 Body mass index (BMI)/waist circumference-based Cox proportional hazards model for men. Observed incidence rates of major coronary events within 1 years of follow-up are shown according to quintiles of estimated global risk calculated with the BMI-based Cox proportional hazards model for men (left panel). The ranks of BMI/waist circumference and the seven independent risk factors according to their correlation coefficients are shown in the right panel. The evaluation is based on 44 major coronary events, which occurred in 1 years of follow-up in 7134 men aged 3 6 years. MI, myocardial infarction. The following seven variables were found to be independently predictive of event risk and were used along with BMI to construct the risk algorithm based on a Cox proportional hazards model: age, LDL cholesterol, smoking, HDL cholesterol, pulse pressure, family history of premature myocardial infarction and diabetes mellitus (Table 3). The variables and their coefficients used in the construction of the classical Cox proportional hazards model are shown for comparison in Table 4. The observed cases of major coronary events within 1 years of follow-up in each of the five quintiles of estimated global risk calculated with the BMI-based Cox proportional hazards model for men are shown in Figure 1. As is evident, individuals in the uppermost quintile of the global risk distribution were almost -fold more likely to be affected by a major coronary event compared with those in the lowest quintile. We also compared the performance of the BMI-based modified model with the previously established traditional PROCAM algorithm using receiver-operating characteristics (ROC) curve analysis (Figure 2). Although the area under the ROC curve obtained by means of the conventional Cox function of 82.1% was marginally larger than the area under the ROC curve obtained with the modified Cox model (82.%), the difference was not significant, indicating that the ability of both models to predict the relative risk of an acute coronary event was essentially equivalent.

4 S14 sensitivity specificity Figure 2 Diagnostic performance of the modified body mass index (BMI)/ waist circumference-based PROCAM Cox model. Receiver-operating characteristics (ROC) curves (black line: BMI-based modified PROCAM Cox model; grey line: conventional fully adjusted PROCAM Cox model). Note that both curves show almost complete overlap and essentially identical areas under the ROC curves (BMI-modified Cox model: 82.%; conventional Cox model: 82.1%). Incidence of coronary events in 1 years (%) without with metabolic syndrome 44 coronary events in 1 years in 7134 men aged 3 to 6 years To evaluate the new BMI-based PROCAM risk algorithm with respect to its ability to predict coronary events in individuals with metabolic syndrome, we divided the study population into two groups with or without the metabolic syndrome at study entry. As shown in Figure 3, 232 men (32.6%) fulfilled the criteria of the metabolic syndrome. In all, 224 major coronary events occurred in this group within 1 years of follow-up (incidence rate: 9.6%). In the 489 men not classified as having metabolic syndrome, 18 major coronary events were observed (incidence rate: 3.7%). Thus, men with metabolic syndrome were 2.9-fold more likely to experience a major coronary event within 1 years of followup than men not having the metabolic syndrome. Most importantly, the mean estimated risk based on the new observed expected (PROCAM) Figure 3 Observed (grey bars) and expected (hatched bars) coronary events in individuals with and without metabolic syndrome. The study population (7134 men aged 3 6 years) was divided into two groups with (232 men, 32.6%) or without (489 men, 67.4%) the metabolic syndrome at study entry based on the definition proposed by the NCEP-ATP III Panel. A total of 18 major coronary events were observed within 1 years of follow-up in men without metabolic syndrome (left grey bar, incidence rate: 3.7%); 224 major coronary events were observed in men with metabolic syndrome (right grey bar, incidence rate: 9.6%). The expected rates of major coronary events calculated with the body mass index-based modified PROCAM Cox model are indicated by hatched bars on the right. modified PROCAM risk algorithm was 1.2% in the group with metabolic syndrome and 4.6% in the group without metabolic syndrome (Figure 3), and thus closely resembled the observed event rates. As shown in Figure 4, the observed major coronary event incidence rates corresponded well with the expected rates according to the modified PROCAM risk algorithm over the entire global risk distribution range in both groups. Discussion Individuals with metabolic syndrome are characterized by the presence of a cluster of metabolic abnormalities, including central obesity, dyslipidaemia (elevated triglycerides, low HDL cholesterol and often also high LDL cholesterol), elevated blood pressure, and insulin resistance or glucose intolerance. Additional frequently observed anomalies include a prothrombotic state (that is, high fibrinogen or plasminogen activator inhibitor-1 in the blood) and/or a proinflammatory state (that is, elevated C-reactive protein in the blood). In practice, patients represent with a continuum of anomalies ranging from moderate overweight linked to moderately decreased HDL and mildly elevated blood pressure to full expression of the syndrome with presence of all its anomalies in a very significant manner It is conceivable that coronary risk depends importantly on this clinical variability and interaction with other coronary risk factors, which can best be accounted for by a validated risk algorithm that incorporates central obesity as one of the most significant features of the metabolic syndrome. Increased waist circumference is a measure for central obesity. We here show that waist circumference is strongly correlated with BMI in the PROCAM Study population. Therefore, we considered BMI and waist circumference as essentially exchangeable variables and used BMI as a predefined fixed variable in constructing the new PROCAM algorithm for assessing coronary event risk in the metabolic syndrome. In comparison with the traditional fully adjusted PROCAM algorithm, pulse pressure came out as independent variable in the new modified BMI-based model with a similar coefficient and hazard ratio as systolic blood pressure in the traditional PROCAM algorithm. The fixed variable BMI provided similar information as triglycerides did in the conventional model. Minor differences with respect to the coefficients of all other risk factors (age, HDL and LDL cholesterol, diabetes, smoking status and family history of premature coronary heart disease) were not significant. As shown by ROC curve analysis, the diagnostic performance of the BMI-based modified model was essentially equal compared with the previously published fully adjusted PROCAM algorithm. 12 We show that men with metabolic syndrome according to the NCEP-ATP III definition were 2.9-fold more likely to

5 MI incidence (% in 1 years) without metabolic syndrome MI incidences in 489 men, aged 3 6 years <1 1 2 >2 MI incidence (% in 1 years) Prevalence (%) Global risk (% in 1 years) 4.7 with metabolic syndrome <1 1 2 >2 224 MI incidences in 232 men, aged 3 6 years S1 observed expected Figure 4 Major coronary event incidence rates and calculated risk. The observed (grey bars) and expected (hatched bars) coronary events in individuals without (left panel) and with (right panel) metabolic syndrome stratified according to low (o1% in 1 years), moderate (1 2% in 1 years) and high (42% in 1 years) global risk are shown. Note that observed major coronary event incidence rates correspond well with the expected rates according to the modified body mass indexbased PROCAM risk algorithm over the entire global risk distribution range in both groups. MI, myocardial infarction. experience a major coronary event compared with those not having metabolic syndrome. This result corresponded well with the 2.22-fold elevated relative risk calculated with the new modified algorithm. Comparative calculations performed with the conventional PROCAM algorithm, which includes fasting triglycerides and systolic blood pressure as variables, yielded very similar results (data not shown). Thus, our data are in line with results from previous studies showing that presence of the metabolic syndrome leads to an approximately twofold increase in relative risk for cardiovascular disease. 19 On the basis of pathophysiological considerations, Despres and Lemieux 9 have recently introduced a concept of cardiometabolic risk. They suggested that the presence of the metabolic syndrome should be considered separately from traditional risk factors in global risk assessment. On the basis of the observation that risk calculations performed with the Framingham algorithm tended to significantly underestimate the risk of myocardial infarction in individuals with the metabolic syndrome, it was argued that the syndrome may per se add excess risk to the global coronary heart disease risk estimates provided by established risk algorithms. Specifically, they proposed that the calculated risk scores should be doubled in individuals with the diagnosis of metabolic syndrome. 9 The results of our study argue against this concept and instead suggest that both PROCAM algorithms accurately predict risk of myocardial infarction in men with the metabolic syndrome. The origin of these discrepant observations is at present unclear. It should be noted, however, that the PROCAM algorithms incorporate different variables compared with the Framingham algorithm. Notably, triglycerides, BMI/waist circumference and family history of myocardial infarction, which are components of the PROCAM algorithms, are not considered by the Framingham algorithm. Another difference is that the Framingham algorithm incorporates total cholesterol, whereas the PROCAM algorithms incorporate LDL cholesterol. Thus, it is conceivable that the set of variables that is incorporated in the PROCAM algorithms may represent an improvement in risk ascertainment, especially in individuals with the metabolic syndrome. Conflict of interest US has received consultancy fees from MSD and lecture fees from Pfizer. The other authors declared no financial interests. References 1 Leal J, Luengo-Fernandez R, Gray A, Petersen S, Rayner M. Economic burden of cardiovascular diseases in the enlarged European Union. Eur Heart J 26; 27: Flegal KM, Carroll MD, Ogden CL, Johnson CL. Prevalence and trends in obesity among US adults, JAMA 22; 288: Ford ES. Risks for all-cause mortality, cardiovascular disease, and diabetes associated with the metabolic syndrome: a summary of the evidence. Diabetes Care 2; 28: Eckel RH, Grundy SM, Zimmet PZ. The metabolic syndrome. Lancet 2; 36: Olshansky SJ, Passaro DJ, Hershow RC, Layden J, Carnes BA, Brody J et al. A potential decline in life expectancy in the United States in the 21st century. N Engl J Med 2; 32: Grundy SM, Cleeman JI, Daniels SR, Donato KA, Eckel RH, Franklin BA et al. Diagnosis and management of the metabolic syndrome: an American Heart Association/National Heart, Lung, and Blood Institute Scientific Statement. Circulation 2; 112: Alberti KG, Zimmet P, Shaw J. The metabolic syndromefa new worldwide definition. Lancet 2; 366:

6 S16 8 Assmann G, Guerra R, Fox G, Cullen P, Schulte H, Willett D et al. Harmonizing the definition of the metabolic syndrome: comparison of the criteria of the Adult Treatment Panel III and the International Diabetes Federation in United States American and European populations. Am J Cardiol 27; 99: Despres JP, Lemieux I. Abdominal obesity and metabolic syndrome. Nature 26; 444: Assmann G, Cullen P, Schulte H. The Munster Heart Study (PROCAM): results of follow-up at 8 years. Eur Heart J 1998; 19: A2 A Cullen P, Schulte H, Assmann G. The Münster Heart Study (PROCAM): total mortality in middle-aged men is increased at low total and LDL cholesterol concentrations in smokers but not in nonsmokers. Circulation 1997; 96: Assmann G, Cullen P, Schulte H. Simple scoring scheme for calculating the risk of acute coronary events based on the 1-year follow-up of the prospective cardiovascular Munster (PROCAM) study. Circulation 22; 1: Assmann G, Schulte H. Results and conclusions of the Prospective Cardiovascular Münster (PROCAM) Study. In: Assmann G (ed). Lipid Metabolism Disorders and Coronary Heart Disease. MMV Medizin Verlag: München, 1993, pp Anderson PJ, Critchley JA, Chan JC, Cockram CS, Lee ZS, Thomas GN et al. Factor analysis of the metabolic syndrome: obesity vs insulin resistance as the central abnormality. Int J Obes 21; 2: Carr DB, Utzschneider KM, Hull RL, Kodama K, Retzlaff BM, Brunzell JD et al. Intra-abdominal fat is a major determinant of the National Cholesterol Education Program Adult Treatment Panel III criteria for the metabolic syndrome. Diabetes 24; 3: Nakamura T, Tokunga K, Shimomura I, Nishida M, Yoshida S, Kotani K et al. Contribution of visceral fat accumulation to the development of coronary artery disease in non-obese men. Atherosclerosis 1994; 17: Bonora E, Kiechl S, Willeit J, Oberhollenzer F, Egger G, Targher G et al. Prevalence of insulin resistance in metabolic disorders: the Bruneck Study. Diabetes 1998; 47: Nesto RW. The relation of insulin resistance syndromes to risk of cardiovascular disease. Rev Cardiovasc Med 23; 4: S11 S Galassi A, Reynolds K, He J. Metabolic syndrome and risk of cardiovascular disease: a meta-analysis. Am J Med 26; 119: