Central Adipose Distribution Is Related to Coronary Atherosclerosis

Transcription

1 327 Central Adipose Distribution Is Related to Coronary Atherosclerosis Corleen J. Thompson, Jacqueline E. Ryu, Timothy E. Craven, Frederic R. Kahl, and John R. Crouse III A "male" distribution of adipose tissue in women (excess of fat in the abdomen compared with that in the hips; i.e., elevated waist/hip ratio) has been related to symptomatic cardiovascular disease. An elevated waist/hip ratio has also been related to symptomatic cardiovascular and cerebrovascular diseases in men, as well as to risk factors for these diseases and various metabolic conditions. To determine whether adipose distribution was related to coronary atherosclerosis, we performed a case control study in patients with angiographically documented coronary atherosclerosis (cases) and in angiographically normal hospital and neighborhood. The data show that distribution of adiposity as assessed by waist/hip ratio is significantly related to coronary atherosclerosis in both females and males. Waist/hip ratio is significantly greater in female cases compared with either control group; in males, waist/hip ratio is significantly greater in cases compared with asymptomatic neighborhood but not compared with patients with normal coronary arteries. These results persist after control for age, plasma concentrations of lipids and lipoproteins, body mass index, history of hypertension, history of diabetes, and smoking status. The connection between the male adipose distribution in females and coronary atherosclerosis partly explains the greater likelihood of symptomatic cardiovascular disease in them. with excess deposition of fat in the abdominal region are also likely to experience increased risk. {Arteriosclerosis and Thrombosis 1991;ll: ) A'male" or "android" distribution of adipose tissue was first noted to be related to cardiovascular disease and its risk factors by Vague in The observed relation to risk factors has been confirmed in several subsequent studies that include connections to glucose and diabetes, 2 " 19 lipids and lipoproteins, 3-7 ' blood pressure, and cigarette smoking In several instances, the male adipose distribution has been related to development of diabetes 2 ' ; in addition, in most of these studies, the relation of the male fat distribution pattern to risk factors was independent of obesity per se Although these observa- Frora the Department of Epidemiology (CJ.T.), University of North Carolina, Chapel Hill, and the Departments of Comparative Medicine (J.E.R.), Public Health Sciences (T.E.C.), and Medicine (F.R.K., J.R.C.), Bowman Gray School of Medicine, Winston- Salem, N.C. Supported by Specialized Center of Research in Arteriosclerosis grant HL and Preventive Cardiology Academic award HL to J.R.C. from the National Heart, Lung, and Blood Institute. Address for correspondence: Ms. Corleen J. Thompson, MSPH, Department of Epidemiology, University of North Carolina, CB #7400, Chapel Hill, NC Received February 20, 1990; revision accepted October 24, tions are of considerable interest per se, importance was added to studies of adipose distribution by the discovery of a relation to cardiovascular disease prevalence and incidence in the Health Examination Survey 4 and in studies in Gothenburg, Framingham, 26 Paris, 15 and Honolulu. 27 Several excellent reviews of these relations have been published. 28 " 31 Many of the factors linked to adipose distribution are risk factors for both symptomatic cardiovascular disease and coronary or cerebrovascular atherosclerosis; however, previous studies have not addressed the potential for an independent relation of adipose distribution to coronary atherosclerosis. To determine whether the male pattern of adipose distribution was independently related to coronary atherosclerosis, we compared adipose distribution in a group of patients who had coronary atherosclerosis (patients hospitalized for coronary angiography) with that in both coronary disease-free and in a group of asymptomatic accessed from the neighborhoods of the patients and chosen to be representative of the population from which the cases arose. Methods A description of hospital patient accession and methods has been reported previously. 32 Briefly, patients

2 328 Arteriosclerosis and Thrombosis Vol 11, No 2 March/April 1991 referred for coronary angiography were screened for eligibility during the evening before catheterization. Patients who had any of the following were excluded from the study: 1) history of angioplasty, coronary artery bypass graft surgery, or myocardial infarction within the past 6 weeks; 2) use of certain drugs known to influence plasma lipid levels, for example, lipidlowering agents, corticosteroids, or heparin; and 3) renal or hepatic disease, or alcohol abuse. Those who met eligibility criteria and agreed to participate gave their signed informed consent. At that time, one of several interviewers specially trained in research methods administered a questionnaire relating to risk factors for coronary artery disease, performed a brief physical examination including certain anthropometric measurements, and provided the patient with additional questionnaires to ascertain family history and psychosocial characteristics. On the morning of catheterization, blood was drawn for determination of plasma lipids and lipoproteins, fasting blood glucose, hemoglobin, and other chemistries. After results of the angiogram became available, those patients with either obstructive coronary artery disease (^50% obstruction of at least one coronary artery) or normal coronary arteries (no luminal irregularities) were eligible for selection into one of eight cells based on disease status, sex, and age (> 50 years or ^50 years), according to a schema designed to achieve approximately equal numbers in each cell. We failed to maintain the goal of accessing approximately equal numbers in each of the eight cells, as fewer elderly male hospital and younger male cases were identified and agreed to participate compared with younger male cases and older male hospital. Beginning in 1985, a third group of subjects was recruited: neighborhood matched on the basis of age (±5 years), sex, and race to hospital patients with obstructive coronary artery disease who had been accessed before The method of identifying these has been described 33 and included use of trained surveyors to canvass the neighborhoods where cases resided to find suitable matches, that is, those without a history or symptoms of heart disease. At a later date, physical examination (including anthropometry), phlebotomy, electrocardiogram, and questionnaire administration were performed by a research nurse either in the home of the neighborhood control or in an outpatient clinic. Because of the nonconcurrent accession of neighborhood, hospital cases, and hospital and the inclusion of waist and hip circumference measurements after 1985, neighborhood and hospital cases reported here are not necessarily pahmatched, resulting in different sample sizes. After 1985, waist and hip measurements were included with anthropometric determinations. Each anthropometric determination was measured according to a protocol that employed a scale provided by the project for weight and a vinyl tape measure for anthropometric measurements. "Waist" was denned as the minimal circumference between the rib cage and the umbilicus for lean volunteers and as the maximal circumference for obese volunteers, and "hip" was defined as the maximal circumference between the superior border of the iliac crest and the thigh region. Adipose distribution was expressed as the ratio of the waist circumference to the hip circumference (waist/hip ratio), and relative weight was expressed as body mass index (wt/ht 2 ). Statistical analyses included simple descriptive statistics such as means and SDs for continuous variables, or frequency counts and percentages for dichotomous variables. Statistical analyses were performed to test two null hypotheses generally stated as 1) coronary artery disease patients are the same as patients with normal coronary arteries and 2) coronary artery disease patients are the same as neighborhood. Bonferroni adjustment of the a-level (two-tailed a=0.025) was used with unpaired t tests and x 1 tests to guarantee the overall significance level of Associations between waist/hip ratio, body mass index, and each factor of interest were measured univariately by Spearman rank correlation coefficients. Analysis of covariance was used to adjust waist/hip ratio and body mass index for various continuous factors before comparing adjusted means across subgroups. All tests were performed separately on males and females, and all analyses were limited to white patients and without a history of valvular heart disease. Analyses are based on unmatched samples, as cases were not matched to hospital and only 23 (15%) of the neighborhood are pair matched to cases with waist and hip circumference measurements. Results Levels of variables for volunteers with obstructive coronary artery disease (cases), normal coronary artery hospital, and neighborhood are shown in Table 1 for continuous variables and Table 2 for dichotomous variables. Eight individuals with body mass index greater than 37 (two female hospital, two female neighborhood, two male cases, one male hospital control, and one male neighborhood control) have been eliminated from these analyses because their values were greater than three SDs from the sex-specific means and they proved to be outlying data points. 34 Waist/hip ratio varied significantly for both males and females depending on group membership. In females, mean waist/hip ratio in cases was higher than in either hospital or neighborhood ; mean waist/hip ratios were similar in hospital and neighborhood. In males, mean waist/hip ratios were higher in cases than in neighborhood but were similar for cases and hospital. For both females and males, mean body mass index was greater for hospital than cases, which in turn was greater than that for neighborhood, but none of the comparisons was statistically significant. The higher mean waist/hip ratio in cases could have been attributable to an association between waist/hip ratio and some other risk factor for coro-

3 Thompson et al Adiposity and Atherosclerosis 329 TABLE 1. Characteristics of and by Group Membership for Continuous Variables Factor 71 Age (yr) Waist/hip ratio Body mass index (kg/m 2 ) Total cholesterol Ln triglycerides HDL cholesterol LDL cholesterol Pack-years smoking ± ± ± ± ± ± ±0.07* 25.88±4.94 2O8.80± ± ±12.99* ±30.89* 6.18±12.60* * * * * ± Ln, natural logarithm; HDL, high density lipoprotein; LDL, low density lipoprotein. Values are mean±sd. 'Significantly different from cases, unpaired / test (/><0.025) ± ± ± ± ± J5± ±9.66* ± ± * 5.09± ± ±35.67* 37.98± ±10.02* 0.92±0.06* 26.44± ±33.17* 4.80 ±0.52* ±9.91* ±30.40* 23.63±25.11' nary atherosclerosis. To test for an association with any of these risk factors, we calculated Spearman rank correlations for waist/hip ratio and body mass index with age, lipid and lipoprotein levels, packyears of smoking, family history of coronary heart disease, and history of diabetes mellitus and hypertension as displayed in Table 3. For males and females, significant correlations with waist/hip ratio were positive except for high density lipoprotein (HDL) cholesterol. Correlations with body mass index were positive except for age and HDL cholesterol (males) and pack-years of smoking (females). These significant correlations led us to retest the relation of waist/hip ratio to atherosclerosis after adjusting waist/hip ratio for age and, separately, for body mass index, total cholesterol, natural logarithm of triglyceride values, HDL cholesterol, low density lipoprotein (LDL) cholesterol, pack-years of smoking, history of high blood pressure, and history of diabetes mellitus. A strikingly consistent pattern emerged. With few exceptions, the adjusted TABLE 2. Factor mean waist/hip ratio in female cases was significantly higher than the adjusted mean waist/hip ratio in either hospital or neighborhood ; adjusted mean waist/hip ratio in male cases was significantly higher than the adjusted mean waist/ hip ratio in neighborhood but not higher than in hospital. Age-adjusted mean waist/ hip ratios shown in Table 4 for the female and male groups are nearly identical to those obtained when adjusting for age and, separately, for body mass index, total cholesterol, ln triglycerides, HDL cholesterol, LDL cholesterol, and pack-years of smoking. Body mass index was unrelated to group membership after comparable covariate adjustment in either males or females (data not shown). Discussion An extensive literature has developed that links adipose distribution to risk factors for and symptoms of cardiovascular disease. The risk factors most con- Characteristics of and by Group Membership for Dichotomous Variables n History of high blood pressure History of diabetes mellitus Family history ofchd Ever smoke * CHD, coronary heart disease. Values are percentages. Significantly different from cases, x 2 test (pso.025) * 7* * 9* 16* 68

4 330 Arteriosclerosis and Thrombosis Vol 11, No 2 March/April 1991 TABLE 3. Spearman Correlations Between Select Variables With Waist/Hip Ratio and Body Mass Index Waist/hip ratio Body mass index Variable n Body mass index Age Total cholesterol Ln trigfyceride HDL cholesterol LDL cholesterol Pack-years smoking History of high blood pressure History of diabetes mellitus Family history of CHD Ln, natural logarithm; coronary heart disease. Values are correlation (0.0002) 0.20 (0.02) (0.006) 0.18 (0.04) (0.01) 0.21 (0.001) (0.004) 0.19 (0.004) 0.17 (0.01) 0.30 (0.0002) (0.02) 0.25 (0.003) (0.01) (0.03) HDL, high density lipoprotein; LDL, low density lipoprotein; NS, not significant; CHD, coefficients, with p values in parentheses. sistently associated with the distribution of adipose tissue are described in Table 5 and include plasma concentrations of HDL cholesterol and triglycerides, blood pressure, and various metabolic and clinical correlates of glucose metabolism (elevated fasting glucose or insulin, increased area under the glucose or insulin curve after oral glucose tolerance test, diabetes history). Cigarette smoking has also been correlated with adipose distribution. In a study of males, Shimokata et al 24 found a dose-response effect between waist/hip ratio and number of cigarettes smoked, and a positive relation was also noted in the Diagnostisch Onderzoek Mammacarcnoom Project. 9 Investigators have used various approaches TABLE 4. Adjusted Mean Waist/Hip Ratio for and by Group Membership Adjusted variables Age Age and history of blood pressure No Yes Age and history of diabetes raellitus No Yes * * 0.87 Significantly different from cases (pso.025). 0.83* * 0.84* 0.80* for defining a pattern of adipose distribution that best discriminates patients at risk. Relative skinfold thicknesses measured from the upper body/lower body or from the peripheral/central adipose depots were first used as indexes of a male or android fat pattern Subsequently, it has appeared that the ratio of the circumferences of the body at the waist and hips (waist/hip ratio) is equally well correlated with various risk factors. 11 Several such indexes have been calculated using the minimum waist, waist at the umbilicus, hips at the iliac crest, maximal hips, and thighs as referent circumferences. Of these, the ratio used in this study, the circumference of the waist midway between the lower rib margin and the

5 Thompson et al Adiposity and Atherosclerosis 331 TABLE 5. Previous Studies of Adipose Distribution in Humans First author Ref Date Population n, M/F Method Lipids Cross-sectional studies Vague Men and women SF, und Butler Men and women 2,749/2,986 SF, ss Albrink Men 419 SF, ss&tc Haffner Men and women 293/441 WHR, u:m (SF,ss:tc) Gillum Men and women 3,027/2^79 WHR Haffner Men and women 338/468 WHR, u:m Folsom Men and women 2,256/2,735 WHR, m:m Stefanick Men 73 WHR, b:m Shimokawa Men 1,122 WHR, m:m Kalkhoff Obese women 110 WHR NS Hartz Obese women 21,065 WHR Tonkelaar Women 452 WHR, m:m Landin Obese postmen women 145 WHR, u:m Evans Premen women 80 WHR, m:m WHR,b:m Seidell Premen women 452 (SF,ss:tc;WTR) Fujioka Men and women 15/31 CT, ia:sc Sparrow Men 41 CT, ia CT, ia Peiris Obese premen women 48 (WHR,m:m;SF,ss:tc) Desprfis Premen women 52 CT, ia Cohort studies Butler Men and women 2,749/2,986 SF, ss Stokes Men and women 1,934/2,420 SF, ss Ducimetiere Men 6,718 SF, mult Donahue Men 7,692 SF, ss Feskens Men 878 SF, ss Larsson Men 792 WHR, u:i Ohlson Men 792 WHR,u:i Lapidus Women 1,462 WHR, m:m Lundgren Women 1,462 WHR, m:m Blood pressure Relation to Diabetes +/+ (i) +/+ (i) + + (g, 0 + +/++ (g) Smoking Cardiovascular disease Neg + + For relation, += associated in univariate analysis; + + =associated in multivariate analysis after control for body mass index and/or other risk factors; +/+=related in both males and females; NS, not significant; Neg=negative relation observed. For method, SF, imd; ss; ss:tc; mult=skinfolds, index of masculine differentiation (skinfold/muscle average); subscapular skinfold; subscapulantriceps skinfold; multiple weighted sum of several sites; WHR, u:m; m:m; b:m; u:i=waist/hip ratio for the waist at umbilicus/hip at maximum; waist at minimum circumference/hip at maximum; waist between lower rib margin and iliac crest/hip at maximum; waist at umbilicus/hip at iliac crest; WTR, waist/thigh ratio; CT, ia:sc; ia, computed tomographic measurement of intra-abdominahsubcutaneous fat at the umbilicus or of intra-abdominal fat alone. For lipids, generally positive association with trigrycerides and negative association with high density lipoprotein cholesterol. For diabetes, adipose distribution related to glucose (g) or insulin (i) or, if not indicated, to clinically evident diabetes. No association of waist/hip ratio and trigh/cerides after control for insulin; no association of waist/hip ratio and high density lipoprotein cholesterol in males after control for body mass index; waist/hip ratio was associated with systolic blood pressure in males and females after control for body mass index or insulin. Differences in waist/hip ratio noted between premenopausal and postmenopausal women, between estrogen-using and non-estrogen-using postmenopausal women, and in relation of waist/hip ratio to age, hypertension, and diabetes in premenopausal versus postmenopausal women. Computed tomographic measurement of intra-abdominal fat correlated with risk factors better than did skinfolds or waist/hip ratio. Ref, reference number, postmen, postmenopausal; premen, premenopausal. iliac crest relative to that at the maximal hip circumference, has considerable precedent ' More recently, computed tomographic scanning has been used to identity intra-abdominal and subcutaneous fat at the midabdomen, and this method holds great promise We also found the waist/hip ratio to correlate with several metabolic and clinical variables including age, cholesterol, triglycerides, HDL cholesterol, cigarette smoking, high blood pressure, and history of diabetes. In general, correlations were stronger in females than in males, in agreement with previous studies

6 332 Arteriosclerosis and Thrombosis Vol 11, No 2 March/April 1991 Adipose distribution has also been independently associated with development of symptomatic cardiovascular disease (myocardial infarction, angina pectoris, stroke, and death) in several studies It has previously been unclear, however, whether the relation to symptomatic cardiovascular disease reflects an accompanying association to atherosclerosis, and furthermore, whether such an association is independent of relations to other risk factors. Several investigators have urged use of atherosclerosis as an outcome variable (as opposed to myocardial infarction or death from cardiovascular disease ). This approach has certain advantages such as specificity of the outcome variable and relative certainty that the control group lacks the disease that affects the cases. The latter may be particularly important when dealing with an elderly population in whom the condition of interest may be very prevalent. On the other hand, a study of patients undergoing angiography is not without its own pitfalls: an important one is the possible nonrepresentativeness of the control group. For this reason, use of more than one control group for angiography case-control studies has been advocated. 37 We have previously documented the effort involved in collecting such an asymptomatic neighborhood control group. 32 In addition, we have outlined similarities and differences between the hospital and community regarding traditional risk factors for cardiovascular disease. 38 Briefly, the hospital and neighborhood control groups are very similar for all risk factors studied except cigarette smoking (more current smokers and fewer quitters in the neighborhood ), plasma concentrations of triglycerides and HDL cholesterol (higher triglycerides and lower HDL cholesterol in the hospital ; associated with widespread /3-blocker use), and type A behavior profile (hospital have higher type A profiles than neighborhood ). In general, the coronary angiography model appears to reflect many aspects of population-based studies and may have certain advantages over populationbased studies for exploratory analyses. In the present study, waist/hip ratios in female coronary disease cases differed from both control groups after adjustment for various other risk factors. Male cases differed from neighborhood only, thus prompting more caution in interpretation of the results in males. The lack of waist/hip ratio differences between male cases and hospital may be due to several factors. The number of males who were found to have angiographically normal coronary arteries is small. There may also be selection bias in the group of males studied who have no atherosclerotic lesions in their coronary arteries; the reasons for referral are varied, and personal traits may increase the likelihood of hospitalization of a select group of males for cardiac catheterization. Additionally, the true difference in mean waist/hip ratios for males with and without coronary atherosclerosis may be smaller than that for females. Each of these might lead to our failure to find an apparent association between waist/hip ratio and atherosclerosis in males. Of interest, a sex differential in the relation of adipose distribution to symptomatic disease may also be inferred from male and female results in the Gothenburg study Although we found a statistically significant correlation between waist/hip ratio and body mass index, the relation of waist/hip ratio to atherosclerosis persisted after adjustment for body mass index and age. This is in agreement with published reports that indicate that various anthropometric indexes measure different phenomena " 19 Although body mass index was significantly correlated with several risk factors for coronary atherosclerosis, adjusting for these separately with age led to no significant differences in body mass index for those with atherosclerosis and those with normal coronary arteries or neighborhood. Hence, body mass index was not an independent risk factor for coronary atherosclerosis in this sample, once other known risk factors were taken into account. In addition to the various traditional risk factors measured in this study, waist/hip ratio has been linked to insulin level, a factor gaining importance for its association with cardiovascular disease. 39 The possibility exists that the relation of adipose distribution to disease shown in this article might disappear in a multivariate model that included plasma insulin concentrations. Whether a defect in insulin homeostasis, adipose distribution, or a related third factor (e.g., biologically available sex steroids ) is primary in such a relation is unclear at present. The connection of adipose distribution to coronary atherosclerosis, however, while not excluding the possibility of a link to clinical events through older (e.g., thrombotic) mechanisms, does provide a rationale for a link between symptomatic cardiovascular disease and body habitus. Acknowledgments The authors acknowledge the help of Edna Mitchell, Pat Mueller, and Nancy Murphy for volunteer accession. Mike Hines, Robert Brunner, Monte Hunter, Steve Mishkind, John Bent, Polly Ross, and Susan Epstein also helped in volunteer accession. Barbara McGill and Ellen Burleson provided support in the laboratory. Mary Lib Slate assisted with data entry. We thank Amy Harmon and Misty Agientas for manuscript preparation. References 1. Vague J: The degree of masculine differentiation of obesities: A factor determining predisposition to diabetes, atherosclerosis, gout, and uric calculous disease. Am J Clin Nutr 1956;4: Butler WJ, Ostrander LD Jr, Carman WJ, Lamphiear DE: Diabetes mellitus in Tecumseh, Michigan: Prevalence, incidence and associated conditions. Am J Epidemiol 1982;116: Haffner SM, Stern MP, Hazuda HP, Pugh J, Patterson JK: Do upper-body and centralized adiposity measure different aspects of regional body-fat distribution? Relationship to non-insulin-dependent diabetes mellitus, lipids, and lipoproteins. Diabetes 1987;36;43-51

7 4. Gillum RF: The association of body fat distribution with hypertension, hypertensive heart disease, coronary heart disease, diabetes and cardiovascular risk factors in men and women aged years. / Chronic Dis 1987;40: Haffner SM, Fong D, Hazuda HP, Pugh JA, Patterson JK: Hyperinsulinemia, upper body adiposity, and cardiovascular risk factors in non-diabetics. Metabolism 1988^7: Folsom AR, Burke GL, Ballew C, Jacobs DR, Haskell WL, Donahue RP, Liu K, Hilner JE: Relation of body fatness and its distribution to cardiovascular risk factors in young blacks and whites: The role of insulin. Am J Epidemiol 1989;130: Kalkhoff RK, Hartz AH, Rupley D, Kissebah AH, Kelber S: Relationship of body fat distribution to blood pressure, carbohydrate tolerance, and plasma lipids in healthy obese women. / Lab Clin Med 1983;102: Hartz AJ, Rupley DC, Rimm AA: The association of girth measurements with disease in 32,856 women. Am J Epidemiol 1984;119:71-8O 9. Tonkelaar ID, Seidell JC, Van Noord PAH, Baanders-Van Halewijn EA, Jacobus JH, Bruning PF: Factors influencing waist/hip ratio in randomly selected pre- and post-menopausal women in the DOM-project (preliminary results). Int J Obes 1989;13: Landin K, Krotkiewski M, Smith U: Importance of obesity for the metabolic abnormalities associated with an abdominal fat distribution. Metabolism 1989;38: Evans DJ, Hoffmann RG, Kalkhoff RK, Kissebah AH: Relationship of body fat topography to insulin sensitivity and metabolic profiles in premenopausal women. Metabolism 1984; 33: Fujioka S, Matsuzawa Y, Tokunaga K, Tarui S: Contribution of intra-abdominal fat accumulation to the impairment of glucose and lipid metabolism in human obesity. Metabolism 1987;36: Sparrow AD, Borkan GA, Gerzof SG, Wisnewski C, Silbert CK: Relationship of fat distribution to glucose tolerance: Results of computed tomography in male participants of the Normative Aging Study. Diabetes 1986^5: Peiris AN, Hennes MI, Evans DJ, Wilson CR, Lee MB, Kissebah AH: Relationship of anthropometric measurements of body fat distribution to metabolic profile in premenopausal women. Acta Med Scand Suppl 1988;723: Ducimetiere P, Richard J, Cambien F: The pattern of subcutaneous fat distribution in middle-aged men and the risk of coronary heart disease: The Paris Prospective Study. IntJ Obes 1986;10: Feskens EJM, Kromhout D: Cardiovascular risk factors and the 25-year incidence of diabetes mellitus in midde aged men. Am J Epidemiol 1989;13O: Ohlson LO, Larsson B, Svardsudd K, Welin L, Eriksson H, Wilhelmsen L, Bjdrntorp P, Tibblin G: The influence of body fat distribution on the incidence of diabetes mellitus: 13.5 years of follow-up of the participants in the study of men born in Diabetes 1985;34: Lapidus L, Bengtsson C, Larsson B, Pennert K, Rybo E, Sjostrom L: Distribution of adipose tissue and risk of cardiovascular disease and death: A 12 year follow up of participants in the population study of women in Gothenburg, Sweden. Br Med] 1984;289: Lundgren H, Bengtsson C, Blohme G, Lapidus L, Sjostrflm L: Adiposity and adipose tissue distribution in relation to incidence of diabetes in women: Results from a prospective population study in Gothenburg, Sweden. Int J Obes 1989;13: Albrink MJ, Meigs JW: The relationship between serum triglycerides and skinfold thickness in obese subjects. Am J Clin Nutr 1964;15: Stefanick ML, Williams PT, Krauss RM, Terry RB, Vranizan KM, Wood PD: Relationships of plasma estradiol, testosterone, and sex hormone-binding globulin with lipoproteins, Thompson et al Adiposity and Atherosclerosis 333 apolipoproteins, and high density lipoprotein subtractions in men. / Clin EndocrincA Metab 1987;64: Seidell JC, Cigolini M, Charzewska J, Ellsinger BM, Contaldo F: Regional obesity and serum lipids in European women born in 1948: A multicenter study. Acta Med Scand Suppl 1988;723: Despres J-P, Moorjani S, Ferland M, Tremblay A, Lupien PJ, Nadeau A, Pinault S, Thfiriault G, Bouchard C: Adipose tissue distribution and plasma lipoprotein levels in obese women: Importance of intra-abdominal fat. Arteriosclerosis 1989;9: Shimokawa H, Muller DC, Andres R: Studies in the distribution of body fat: HI. Effects of cigarette smoking. JAMA 1989;261: Larsson B, Svardsudd K, Welin L, Wilhelmsen L, Bj6rntorp P, Tibblin G: Abdominal adipose tissue distribution, obesity, and risk of cardiovascular disease and death: 13 year follow-up of participants in the study of men bom in Br Med J 1984;288: Stokes J m, Garrison RJ, Kannel WB: The independent contributions of various indices of obesity to the 22-year incidence of coronary heart disease: The Framingham Heart Study, in Vague J, BjSrntorp P, Guy-Grand B, Rebuff6-Scrive M, Vague P (eds): Metabolic Complications of Human Obesities. New York/Amsterdam, Elsevier Science Publishing Co, Inc, 1985, pp Donahue RP, Abbott RD, Bloom E, Reed DM, Yano K: Central obesity and coronary heart disease in men. Lancet 1987;l: Kissebah AH, Peiris AN, Evans DJ: Mechanisms associating body fat distribution to glucose intolerance and diabetes mellitus: Window with a view. Acta Med Scand Suppl 1988;723: Larsson B: Regional obesity as a health hazard in men: Prospective studies. Acta Med Scand Suppl 1988;723: BjSrntorp P: The associations between obesity, adipose tissue distribution and disease. Acta Med Scand Suppl 1988;723: Stern MP, Haffner SM: Body fat distribution and hyperinsulinemia as risk factors for diabetes and cardiovascular disease. Arteriosclerosis 1986;6: Crouse JR, Toole JF, McKinney WM, Digman MB, Howard G, Kahl FR, McMahan MR, Harpold GH: Risk factors for extracranial carotid artery atherosclerosis. Stroke 1987;18: Ryu JE, Thompson CJ, Crouse JR: Selection of neighborhood for a study of coronary artery disease. Am J Epidemiol 1989;129: Neter J, Wasserman W, Kutner MH: Applied Linear Statistical Models, ed 2. Homewood, 111, RD Irwin, 1985, pp Seidell JC, Bjorntorp P, Sjostrom L, Sannerstedt R, Krotkiewski M, Kvist H: Regional distribution of muscle and fat mass in men: New insights into the risk of abdominal obesity using computed tomography. Int J Obes 1989;13: Kuller LH: Epidemiology of cardiovascular diseases: Current perspectives. Am J Epidemiol 1976;104: Pearson TA: Coronary angiography in the study of the epidemiology of coronary artery disease. Epidemiol Rev 1984;6: Tell GS, Ryu JE, Thompson CJ, Kahl FR, Crouse JR: Control selection bias in case-control studies of coronary atherosclerosis (abstract). Circulation 1988;(suppl II):II Reaven GM: Banting lecture, 1988: Role of insulin resistance in human disease. Diabetes 1988;37: Haffner SM, Katz MS, Stern MP, Dunn JF: Association of decreased sex hormone binding globulin and cardiovascular risk factors. Arteriosclerosis 1989;9: KEY WORDS obesity coronary atherosclerosis adipose distribution risk factors