The Paris Prospective Study is a long-term investigation of coronary heart disease (CHD) risk factors in a large population of working men. The baseline cohort.
Insulin and Cardiovascular Disease
Annick M. Fontbonne, MD Eveline M. Eschwege, MD
Paris Prospective Study
The Paris Prospective Study is a long-term investigation of coronary heart disease (CHD) risk factors in a large population of working men. The baseline cohort included 7028 men, 6093 who had a 75-g oral glucose tolerance test with measurement of plasma insulin and glucose levels (0 and 2 h) and 125 who were known noninsulin-treated diabetic patients. After a mean follow-up of 11 yr, 126 deaths ascribed to CHD were reported. Major independent predictors of CHD death were blood pressure, smoking, plasma cholesterol level, and fasting and 2-h postload plasma insulin level. Impairment of glucose tolerance, including overt diabetes, did not rank as an independent predictor when other baseline variables were accounted for. In the subset of the baseline cohort who presented with impaired glucose tolerance or diabetes (n = 943), 26 died from CHD during the follow-up. The strongest independent predictor of subsequent CHD death in this subsample with abnormal glucose tolerance was plasma triglyceride level. In view of the accumulating evidence that hyperinsulinemia and hypertriglyceridemia generally occur in the same type of subjects, in relation to insulin resistance and central obesity, the epidemiological findings of the Paris Prospective Study and of other investigations support the hypothesis that a constellation of mild metabolic abnormalities may play a deleterious role with regard to cardiovascular disease risk. Diabetes Care 14:461-69, 1991
P
rospective community-based epidemiological studies have led to the identification of major risk factors of cardiovascular disease, a target public health issue in industrialized countries. Although their intimate mechanisms remain unclear, the responsibility of high plasma cholesterol levels, high blood pressure, and smoking is unquestioned (1).
DIABETES CARE, VOL. 14, NO. 6, JUNE 1991
The case for diabetes, glucose intolerance, obesity, and high plasma triglyceride levels is much more debated, because prospective studies have never yielded clear-cut results concerning their predictive power toward cardiovascular disease (2-6). However, clinical experience teaches that diabetic patients, and maybe obese subjects to a lesser degree, are more prone to cardiovascular complications than lean nondiabetic individuals (7-9). Furthermore, these are often the same subjects who are overweight and present with mild glucose intolerance, slightly elevated blood pressure, and lipidemic disturbances (10-12). If any of these abnormalities cannot separately determine a significant increase in cardiovascular risk, their clustering may still interact to precipitate cardiovascular complications. The fact that this cluster of risk factors is a common clinical finding has given rise to the hypothesis that it was not constituted at random but was the multifaceted symptom of a basic abnormality, probably rooted in overweight and insulin resistance (13,14). This hypothesis is not new; decades ago, central obesity, i.e., preferentially abdominal fat deposits, was identified as a frequent concomitant of non-insulin-dependent (type II) diabetes and cardiovascular disease (15,16). However, it has received new input from the results of prospective studies, which identified either central obesity (17-22) or hyperinsulinemia (23-25), a direct marker of insulin resistance, as independent significant predictors of coronary heart disease (CHD) incidence and mortality.
From the National Institute of Health and Medical Research, Villejuif, France. Address correspondence and reprint requests to Dr. Annick Fontbonne, INSERM U21, 16 Avenue Paul Vaillant Couturier, F94807 Villejuif Cedex, France.
461
INSULIN AND CORONARY HEART DISEASE
RESEARCH DESIGN AND METHODS The Paris Prospective Study (PPS) is a long-term investigation of cardiovascular disease risk factors, which was initiated in the late 1960s on a large cohort of middleaged men. The first annual follow-up session (19681973) was attended by 7434 male employees of the Paris Civil Service, aged 43-54 yr. It included a cardiovascular examination, biological measurements, and a medical history questionnaire. At the time of this examination, 270 men with diagnosed cardiovascular disease were excluded from the follow-up. Blood pressure was measured by standard cuff, with the subjects in the sitting position. Body mass index (BMI) was calculated as the ratio of weight to squared height (kg/m2). Cigarette consumption was recorded, and an overnight fasting venous blood sample was taken for measurement of serum cholesterol (26), triglyceride (27), plasma glucose (28), and insulin levels (29). A 75-g oral glucose tolerance test (OGTT) was given to 6903 subjects with no known diabetes and to 99 subjects with known diabetes, and plasma glucose and insulin levels were measured 2 h after the load. All deaths up to 1 January 1983 were counted (mean follow-up 11 yr). Deaths were systematically reported by the different administrative departments of the Paris Civil Service. Complementary inquiries to families, practitioners, or hospitals were organized to obtain information about the circumstances and causes of death. A panel of physicians then had to ascertain the nature of the disease as the underlying cause of death. Coding was performed according to the International Classification of Diseases, 8th revision. All the analyses reported herein are based on the following cohort or a subset of it: 6903 men who had an OGTT at the first follow-up examination and 125 men who attended the first follow-up examination and were
already known to have diabetes at the time but with no insulin treatment (only 10 men had insulin-treated diabetes at the time of this examination). In this cohort, CHD death was identified by the following codes: 410.0-414.9 was myocardial infarction; 795.0 was sudden death; 782.0-782.9, 427.0, 427.1, and 519.1 was heart failure in a broad extent. Statistical analyses for comparing levels of the various measurements performed at entry (1st follow-up examination) between the subjects who subsequently died from CHD and those who died from another cause or were still alive by 1 January 1983 were based on standard parametric tests. CHD mortality rates by levels of a given risk factor were computed with a denominator of person-years and compared with the Cox model (30). They were not age adjusted, given the narrow range of age of the population. Multivariate analyses were based on the Cox model, computed with either BMDP (statistical software P2L) or SAS (Cary, NC) statistical packages. Age was systematically entered in the regression equations. Variables with markedly skewed distributions were log transformed before computing. Subjects with missing values for a computed variable were excluded from the analysis.
INSULIN AS INDEPENDENT PREDICTOR OF CHD Results of PPS. At the 11 -yr follow-up, 651 men of the initial cohort had died, 126 from a CHD death (29 myocardial infarctions and 70 sudden deaths). Compared with the others, the 126 who died from CHD had significantly higher entry levels of blood pressure, BMI, plasma cholesterol and triglyceride, fasting glycemia, and insulinemia, and they were more often smokers (Table 1). Those who were not known to have diabetes had higher postload glycemia and insulinemia than their counterparts who did not die from CHD. When all sub-
TABLE 1 Entry variables in all subjects by subsequent death from coronary heart disease (CHD) CHD death
n Age (yr) Cigarettes (n/day) Body mass index (kg/m2) Systolic blood pressure (mmHg) Cholesterol (mM) Triglyceride (mM) Insulin (pM) Fasting 2-h postload Glucose (mM) Fasting 2-h postload
Yes
No
P
126 49.0 (48.6-49.4) 13 (0-15) 27 (26.2-27.3) 157(151-162) 6.0 (5.8-6.2) 1.43 (1.30-1.57)
6902 48.5 (48.4-48.6) 9 (0-9) 26 (25.9-26.1) 144(143.5-144.5) 5.63 (5.61-5.66) 1.24 (1.23-1.26)
