Insulin-Resistant Prediabetic Subjects Have More Atherogenic Risk Factors Than Insulin-Sensitive Prediabetic Subjects

Insulin-Resistant Prediabetic Subjects Have More Atherogenic Risk Factors Than Insulin-Sensitive Prediabetic Subjects Implications for Preventing Coronary Heart Disease During the Prediabetic State Steven M. Haffner, MD; Leena Mykkänen, MD; Andreas Festa, MD; James P. Burke, PhD; Michael P. Stern, MD Background Subjects who convert to type 2 diabetes mellitus have increased cardiovascular risk factors relative to nonconverters. However, it is not known whether these atherogenic changes in the prediabetic state are predominantly due to insulin resistance, decreased insulin secretion, or both. Methods and Results We examined this issue in the 7-year follow-up of the San Antonio Heart Study, in which 195 of 1734 subjects converted to type 2 diabetes. At baseline, converters had significantly higher body mass index, waist circumference, triglyceride concentration, and blood pressure and lower HDL cholesterol than nonconverters. Atherogenic changes in converters were markedly attenuated (and no longer significant) after adjustment for the homeostasis model assessment of insulin resistance (HOMA IR, a surrogate for insulin resistance); in contrast, the differences in risk factors between converters and nonconverters increased after adjustment for the ratio of early insulin increment to early glucose increment ( I 30-0 / G 30-0 ) during an oral glucose tolerance test (a surrogate for insulin secretion). We also compared converters who had a predominant insulin resistance (high HOMA IR and high I 30-0 / G 30-0 )(n 56) and converters who had a predominant decrease in insulin secretion (low HOMA IR and low I 30-0 / G 30-0 )(n 31) with nonconverters (n 1539). Only the converters who were insulin resistant had higher blood pressure and triglyceride levels and lower HDL cholesterol levels than nonconverters. Conclusions Our data suggest that atherogenic changes in the prediabetic state are mainly seen in insulin-resistant subjects and that strategies to prevent type 2 diabetes might focus on insulin-sensitizing interventions rather than interventions that increase insulin secretion because of potential effects on cardiovascular risk. (Circulation. 2000;101:975-980.) Key Words: insulin diabetes mellitus lipids cholesterol blood pressure Type 2 diabetes mellitus is associated with a marked increase in coronary heart disease (CHD). 1 3 The relationship between glycemia and CHD in type 2 diabetes has been controversial, with some studies showing strong associations, 4 some weak associations, 5 and others no associations. 6 The recently published United Kingdom Prospective Diabetes Study (UKPDS) found a greater benefit of glycemic control on microvascular events than for CHD or stroke. 7 One explanation for the relatively weak effect of glycemia on CHD in type 2 diabetes might be the existence of a highly atherogenic state before the onset of diabetes. 8 Increased risk factors for CHD before the onset of type 2 diabetes have been shown in several populations, including Israelis, 9 elderly American subjects, 10 elderly Finnish subjects, 11 Mexican American subjects, 12 and Pima Indians. 13 The causes of increased atherogenicity of the prediabetic state are not fully understood. Both insulin resistance (measured directly 14 or through surrogates such as fasting insulin 15 17 ) and decreased insulin secretion 18 21 predict the development of type 2 diabetes. It is not known whether the increased atherogenicity of the prediabetic state is primarily due to increased insulin resistance or decreased insulin secretion, although increased resistance may be likely given the amount of information on the insulin-resistance syndrome. 22,23 In this report, we examine whether insulin resistance or decreased insulin secretion is responsible for the atherogenic prediabetic state. In particular, we were interested in whether cardiovascular risk factors were similar in prediabetic subjects who had a predominant insulin-secretory defect (normal Received July 12, 1999; revision received September 17, 1999; accepted October 1, 1999. From the Division of Clinical Epidemiology, Department of Medicine, University of Texas Health Science Center at San Antonio. Reprint requests to Steven M. Haffner, MD, Division of Clinical Epidemiology, Department of Medicine, University of Texas Health Science Center at San Antonio, 7703 Floyd Curl Drive, San Antonio, TX 78284-7873. E-mail haffner@uthscsa.edu 2000 American Heart Association, Inc. Circulation is available at http://www.circulationaha.org 975

976 Circulation March 7, 2000 insulin sensitivity but low insulin-secretory response) as opposed to subjects with a predominant insulin-resistance defect (insulin resistant but with good insulin-secretory response). This issue is important because recently, a number of projects to prevent type 2 diabetes have been undertaken with methods that either improve insulin sensitivity (Diabetes Prevention Project [DPP: metformin, intensive lifestyle 24 ] and STOP-NIDDM [acarbose] 25 ) or increase insulin secretion (NANSY [sulfonylurea {Amaryl}]; Arne Melander, Sweden, oral communication, 1999). If atherogenic changes in the prediabetic state are limited to subjects with insulin resistance, the use of insulin-sensitizing agents to prevent diabetes could have a beneficial effect on CHD. We used data from the San Antonio Heart Study, in which we have previously shown that both high fasting insulin (a surrogate for insulin resistance) and a decreased ratio of insulin increment (over the first 30 minutes) to glucose increment (over the first 30 minutes) during an oral glucose tolerance test ( I 30-0 / G 30-0 ) (a surrogate for insulin secretion) predict the development of type 2 diabetes in Mexican Americans. 20 Methods The San Antonio Heart Study is a population-based study of diabetes and cardiovascular disease in Mexican American and non-hispanic whites. From 1984 to 1988, we randomly selected households from low-income (barrio), middle-income (transitional), and high-income (suburban) census tracts in San Antonio, Tex. 26 All men and nonpregnant women aged 25 to 64 years who resided in the randomly sampled households were eligible to participate. Mexican Americans were defined as individuals whose ancestry derived from a Mexican national origin. Detailed descriptions of this study have been published previously. 26 This study was approved by the Institutional Review Board of the University of Texas Health Science Center at San Antonio. All subjects gave informed consent. Beginning in October 1991, we began a 7-year follow-up of the cohort. 20 The results in this report are based on risk factors for the development of type 2 diabetes. Subjects with diabetes at the baseline examination were excluded from this report. At the baseline and follow-up visits, blood specimens were obtained after a 12- to 14-hour fast for determination of plasma glucose, serum insulin, and serum lipids and lipoproteins. Methods for determination of lipids and lipoproteins and glucose have been described previously. 20 We measured serum insulin with a solidphase radioimmunoassay (Diagnostic Products Corporation) that shows a relatively high degree of cross-reactivity with proinsulin ( 70% to 100%). 26 A 75-g oral glucose load (Orangedex; Custom Laboratories) was administered, and blood specimens were obtained 30 minutes, 1 hour, and 2 hours later for plasma glucose and serum insulin concentrations. At the follow-up examination, post glucoseload specimens were obtained only at the 2-hour time point. Diabetes was diagnosed according to World Health Organization (WHO) criteria. 27 Subjects who did not meet WHO plasma glucose criteria but who were undergoing treatment with oral antidiabetic agents or insulin were considered to have diabetes. In this report, we use the homeostasis model of insulin resistance (HOMA IR) as a measure of insulin resistance 28 30 and I 30-0 / G 31 30-0 as a measure of insulin secretion (early secretory response to an oral glucose load). The formula for the HOMA IR model 28 follows: HOMA IR Fasting insulin ( U/mL) fasting glucose (mmol/l) 22.5 The correlation of HOMA IR and fasting insulin in nondiabetic subjects is 0.98. Anthropometric measurements (height, weight, and waist and hip circumferences) were made after participants had removed their shoes and upper garments and donned an examination gown. Body mass index (BMI) was calculated as weight (in kilograms) divided by height (in meters squared). Waist circumference was chosen as a measure of central adiposity. The systolic (first phase) and diastolic (fifth phase) blood pressures were measured to the nearest even digit by use of a randomzero sphygmomanometer (Hawksley-Gelman). Three readings were recorded for each individual, and the average of the second and third readings was defined as the patient s blood pressure. Statistical analyses included ANCOVA performed with SAS statistic software. Two-way ANCOVA was done initially with conversion to diabetes and ethnicity (Mexican American versus non-hispanic whites as the grouping variable). The P value for these interaction terms (ethnicity times conversion status) were all 0.100. Because there was no evidence of different effect of conversion status by ethnicity on variables of interest (ie, triglycerides or blood pressure), we pooled the ethnic groups with control for ethnicity to increase statistical power and to simplify the analysis. One-way ANCOVA was done with conversion to diabetes as the main effect (Tables 1 and 2). Additional analysis was done with 2-way ANCOVA among the converters to diabetes by dividing subjects by their insulinresistance or insulin-secretion status at baseline (HOMA IR above and below median of 3.0 and insulin secretion [ I 30-0 / G 30-0 in pmol/mmol] above and below median of 155.6 pmol/l) (Table 3). The median was based on the overall nondiabetic population at baseline. Finally, 1-way ANCOVAs (with pairwise contrasts) were done with conversion to diabetes as the dependent variable to compare subjects with predominant insulin resistance (above median for both HOMA IR and I 30-0 / G 30-0 ) and subjects with a predominant insulin-secretory defect (below median for both fasting insulin and I 30-0 / G 30-0 ) with subjects who did not convert to type 2 diabetes (Figure 2). Triglyceride, fasting insulin, HOMA IR, and I 30-0 / G 30-0 were transformed to improve the skewness and kurtosis of their distribution for statistical testing. These variables were both back-transformed for presentation in the tables. All probability values are 2-sided. Results Table 1 shows the baseline characteristics of subjects by conversion to diabetes, adjusted for age, sex, and ethnicity. Fasting insulin and HOMA IR were higher and I 30-0 / G 30-0 was lower in converters to diabetes than in subjects who remained nondiabetic. Subjects who converted to diabetes had greater obesity and an unfavorable body fat distribution, higher blood pressure, higher prevalence of hypertension, higher glucose levels, higher triglyceride levels, and lower HDL cholesterol than subjects who did not convert to diabetes. Total and LDL cholesterol levels and smoking status were similar in converters and nonconverters. As shown in Table 2 (model A), after further adjustment for fasting glucose and waist circumference, converters continued to have higher blood pressure and higher triglyceride levels and lower HDL cholesterol levels than subjects who did not convert to diabetes. Table 2 shows the effects of additional adjustment for HOMA IR (a surrogate for insulin resistance) (model B) versus the effect of adjustment for I 30-0 / G 30-0 (model C), a surrogate for insulin secretion. Adjustment for HOMA IR attenuated the differences between converters and nonconverters, making them no longer statistically significant. In contrast, after adjustment for insulin secretion ( I 30-0 / G 30-0 ), the differences between converters and nonconverters to type 2 diabetes remained statistically significant. We next categorized the subjects simultaneously by insulin resistance (above and below the median for HOMA IR in the overall nondiabetic population at baseline) and insulin secretion (above and below the median for I 30-0 / G 30-0 ). The

Haffner et al Prediabetic Subjects and Atherogenic Risk 977 TABLE 1. Clinical Characteristics at Baseline (Mean SEM) According to Conversion Status at Follow-Up (Adjusted for Age, Sex, and Ethnicity) Conversion Status at Follow-Up Nondiabetic (n 1539) Type 2 Diabetes (n 195) P Age,* y 42.8 0.3 48.0 0.8 0.001 Male 43% 39% 0.31 Mexican American 64% 81% 0.0001 Smokers 10.9% 11.4% 0.812 BMI, kg/m 2 26.9 0.1 30.3 0.4 0.001 WHR ( 100) 87.9 91.8 0.5 0.001 Waist, mm 887 3 988 9 0.001 Fasting glucose, mmol 4.7 0.01 5.3 0.04 0.001 Fasting insulin, pmol/l 72 2 130 6 0.001 HOMA IR 2.6 0.23 5.2 0.08 0.001 I 30-0 / G 30-0, pmol/mmol 272 10 130 29 0.001 LDL cholesterol, mmol/l 3.23 0.02 3.18 0.07 0.305 HDL cholesterol, mmol/l 1.25 0.02 1.06 0.02 0.001 Triglyceride, mmol/l 1.52 0.03 2.17 0.08 0.001 HBP 9.9% 25.0% 0.001 SBP, mm Hg 117.6 0.4 123.3 1.0 0.001 DBP, mm Hg 71.6 0.2 74.7 0.7 0.005 WHR indicates waist-hip ratio; HBP, prevalence of hypertension (DBP 95 mm Hg and/or presence of blood pressure medications); SBP, systolic blood pressure; and DBP, diastolic blood pressure. *Not adjusted for age; not adjusted for sex; not adjusted for ethnicity. incidence of type 2 diabetes by insulin resistance and secretion categories is shown in Figure 1. As expected, subjects with the highest rate of developing type 2 diabetes had both insulin resistance and decreased insulin secretion (31.8% in 7 years), and the lowest rate was in subjects who were insulin sensitive with good secretory capacity (1.0%). Subjects with insulin resistance but good insulin secretion had a higher conversion rate than subjects with low insulin secretion who were insulin sensitive (11.0% versus 6.2%). These results are similar to those presented for a smaller cohort of Mexican Americans only. 20 We also characterized the distribution of insulin resistance and secretory effects of converters to diabetes and insulin secretion. Fifty-four percent of converters had both an insulin secretory defect and were insulin resistant compared with 1.5% of converters who were insulin sensitive with good secretion at baseline. The subjects who were predominantly insulin resistant with good insulin secretion at baseline comprised 28.7% of all converters to type 2 diabetes compared with 15.9% of subjects who had low insulin secretion but were insulin sensitive (predominantly insulin sensitive). Table 3 shows anthropometric and cardiovascular risk factors by insulin resistance and secretion categories. Insulin resistance was associated with higher BMI, greater waist circumference, and higher blood pressure and triglyceride levels and lower HDL cholesterol levels. Insulin secretion was not related to anthropometric or cardiovascular risk factors. Fasting and 2-hour glucose levels were similar in each group. Additional adjustment for BMI or waist circumference did not appreciably change these results (data not shown). Figure 2 compares the triglyceride and HDL levels and systolic blood pressure in converters to type 2 diabetes with predominant insulin resistance (high HOMA IR and high I 30-0 / G 30-0 ), converters with a predominant insulin-secretory defect (low I 30-0 / G 30-0 and low HOMA IR), and nonconverters to type 2 diabetes. Among converters to diabetes, the only subjects with adverse cardiovascular risk factors (high systolic blood pressure and triglyceride levels and low HDL cholesterol levels) were converters to diabetes with high IR and I 30-0 / G 30-0 (insulin-resistant subjects). Subjects who converted to diabetes but had predominant insulin resistance had a BMI 3 to 4 kg/m 2 higher than subjects who did not convert to diabetes or who converted to diabetes but had a predominant insulin-secretory defect. Adjustment for differences in BMI somewhat attenuated the differences in lipoproteins or blood pressure ( 30%), but converters to diabetes who had predominant insulin resistance continued to have significantly more atherogenic risk factors than the other 2 groups (P 0.01). Discussion We have confirmed that prediabetic subjects have increased cardiovascular risk factors at baseline relative to subjects who do not convert to type 2 diabetes. These results extend previous results in earlier reports in a variety of ethnic groups, 9 13 including reports on smaller groups of Mexican Americans in the San Antonio Heart Study. 12 The increased atherogenicity of the prediabetic state was due only in part to differences in overall adiposity and upper-body adiposity between the converters and nonconverters to type 2 diabetes (Table 1).

978 Circulation March 7, 2000 TABLE 2. Clinical Characteristics of Subjects at Baseline According to Conversion Status at Follow-Up, Adjusted for Either Insulin Resistance or Insulin Sensitivity Conversion Status at Follow-Up Nondiabetic (n 1539) Type 2 Diabetes (n 195) Difference (DM Non-DM) P A. Adjusted for age, sex, ethnicity, fasting glucose, and waist circumference HDL cholesterol, mmol/l 1.25 0.01 1.14 0.02 0.11 0.001 LDL cholesterol, mmol/l 3.23 0.02 3.09 0.07 0.14 0.459 Triglyceride, mmol/l 1.53 0.03 2.02 0.07 0.49 0.001 SBP, mm Hg 117.8 0.35 120.7 1.0 2.9 0.016 DBP, mm Hg 71.5 0.2 73.2 0.7 2.2 0.042 Fasting insulin, pmol/l 74 2 108 6 34.0 0.001 HOMA IR 2.8 0.1 3.9 0.2 1.1 0.001 I 30-0 / G 30-0, pmol/mmol 227 11 119 30 108 0.001 B. Adjusted for age, sex, ethnicity, fasting glucose, waist circumference, and HOMA IR (surrogate for insulin resistance) HDL cholesterol, mmol/l 1.21 0.01 1.18 0.02 0.03 0.352 LDL cholesterol, mmol/l 3.24 0.01 3.09 0.07 0.15 0.552 Triglyceride, mmol/l 1.69 0.03 1.82 0.08 0.13 0.101 SBP, mm Hg 119.9 2.6 122.1 0.9 2.2 0.271 DBP, mm Hg 72.2 0.7 72.2 0.23 0.0 0.900 C. Adjusted for age, sex, ethnicity, fasting glucose, waist circumference, and I 30-0 / G 30-0 (surrogate for insulin secretion) HDL cholesterol, mmol/l 1.26 0.01 1.11 0.02 0.15 0.001 LDL cholesterol, mmol/l 3.25 0.01 3.07 0.07 0.18 0.373 Triglyceride, mmol/l 1.44 0.03 2.10 0.08 0.66 0.001 SBP, mm Hg 117.3 0.4 121.7 1.0 4.4 0.008 DBP, mm Hg 71.0 0.2 73.4 0.7 2.4 0.024 Values are mean SEM. DM indicates diabetes mellitus; SBP, systolic blood pressure; and DBP, diastolic blood pressure. More interesting is whether the atherogenic differences in prediabetic subjects are due to increased insulin resistance, decreased insulin secretion, or both. To address this issue, we used 2 different approaches: that of statistical adjustment (Tables 1 and 2) and that of stratification (Table 3 and Figure 2). After adjustment for HOMA IR, the differences between converters and nonconverters were attenuated and were no longer statistically significant. In contrast, after adjustment for I 30-0 / G 30-0 (a surrogate for decreased secretory response that has been shown to be a significant predictor of type 2 diabetes in this cohort 20 ), the differences between converters and nonconverters actually widened (Table 2). Figure 1. Seven-year incidence of type 2 diabetes by baseline status for insulin resistance (HOMA IR surrogate for insulin resistance) and insulin secretion ( I 30-0 / G 30-0 surrogate for insulin secretion). Subjects who converted to diabetes but were insulin resistant had significantly higher triglyceride levels, systolic blood pressure, and diastolic blood pressure and lower HDL cholesterol levels than subjects who converted to diabetes but who were insulin sensitive. The 2 groups of converters to type 2 diabetes (predominantly insulin-resistant versus insulinsensitive converters) had similar values for both fasting and 2-hour glucose levels, but the insulin-resistant converters were more obese (Table 3). After additional adjustment for BMI (data not shown), insulin-resistant converters to type 2 diabetes still had worse cardiovascular risk factors than insulin-sensitive converters. Interestingly, insulin-sensitive converters to type 2 diabetes had triglyceride and HDL cholesterol levels and systolic blood pressure similar to those of subjects who remained nondiabetic at baseline. We have thus identified different subgroups of converters to type 2 diabetes with markedly different patterns of cardiovascular risk factors. The implication is that the subjects who were insulin resistant and converted to diabetes would have more cardiovascular disease than the insulin-sensitive subjects who converted to diabetes. We should point out that the differences in lipids (triglyceride 0.9 mmol and HDL cholesterol 0.24 mmol) and systolic blood pressure (6.5 mm Hg) (Table 3) are actually larger than the differences between diabetic (n 303) and nondiabetic (n 2564) subjects in the San Antonio Heart Study (for diabetic versus nondiabetic subjects, respectively: triglyceride 2.3 versus 1.6 mmol/l, a 0.7 mmol/l difference; HDL

Haffner et al Prediabetic Subjects and Atherogenic Risk 979 TABLE 3. Clinical Characteristics of Subjects at Baseline Who Converted to Type 2 Diabetes by Whether They Had Predominantly Insulin Resistance, Decreased Insulin Secretion, or a Mixed Picture (Adjusted for Age, Sex, and Ethnicity) High HOMA IR ( Insulin Resistant ) Low HOMA IR ( Insulin Sensitive ) P * HOMA IR I 30-0 / G 30-0 I 30-0 / G 30-0 (insulin secretion status) High ( good ) Low ( bad ) High ( good ) Low ( bad ) n 56 105 3 31 Fasting insulin, pmol/l 236 15 128 11 50 41 39 19 0.001 0.092 HOMA IR 6.5 2.6 6.0 1.8 1.4 0.9 1.9 2.6 0.001 0.191 I 30-0 / G 30-0, pmol/mmol 281 12 99 9 239 33 78 16 0.125 0.001 Waist circumference, mm 1011 23 901 17 964 65 875 30 0.001 0.293 BMI, kg/m 2 32.3 0.7 30.7 0.6 25.0 2.0 28.7 1.0 0.001 0.344 HDL cholesterol, mmol/l 0.97 0.04 1.07 0.03 1.17 0.12 1.22 0.06 0.005 0.305 LDL cholesterol, mmol/l 2.96 0.13 3.26.10 2.95 0.6 3.34 0.17 0.608 0.465 Triglyceride, mmol/l 2.80 0.48 2.43 0.15 2.38 0.56 1.91 0.25 0.019 0.112 SBP, mm Hg 128.8 2.4 126.2 1.8 114.4 6.6 122.0 3.1 0.009 0.351 DBP, mm Hg 78.5 1.3 76.5 1.0 71.4 3.9 70.9 1.8 0.001 0.825 Fasting glucose, mmol 5.38 0.08 5.31 0.06 5.28 0.10 5.41 0.33 0.866 0.855 2-h Glucose, mmol 7.85 0.24 7.81 0.17 7.70 0.33 7.39 0.69 0.772 0.812 %IGT 54.0% 54.6% 50% 53.5% 0.920 0.940 Values are mean SEM unless otherwise indicated. SBP indicates systolic blood pressure; DBP, diastolic blood pressure; and IGT, impaired glucose tolerance. *P value by 2-way ANCOVA. cholesterol 1.12 versus 1.24 mmol/l, a 0.12 mmol/l difference; and systolic blood pressure 124.9 versus 118.8 mm Hg, or a difference of 6.1 mm Hg). Our results may have important implications for the prevention of diabetes. Currently, a number of clinical trials on insulin-sensitizing agents are under way (DPP [metformin, intensive lifestyle 24 ] and STOPNIDDM [acarbose] 25 ). Similarly, there are prevention trials involving insulin secretagogues (NANSY [sulfonylurea {Amaryl}]). If our results are correct, they suggest that the use of a sulfonylurea to prevent diabetes might increase the risk of CHD (or at least prove less beneficial) than the use of insulin-sensitizing agents. The effects of different modalities for the prevention of diabetes on CHD will be particularly informative. However, improvement in glycemic control in diabetic subjects by sulfonylurea has led to reduction in insulin resistance in diabetic subjects, 32 Figure 2. Levels of cardiovascular risk factors by HOMA IR (fasting insulin), insulin secretion ( I 30-0 / G 30-0 ), and conversion status. suggesting that the differential between insulin sensitizers and insulin secretagogues with respect to cardiovascular risk factors could be overestimated in epidemiological studies such as the present report. However, whether insulin secretagogues would improve insulin sensitivity in nondiabetic subjects at high risk of diabetes (impaired glucose tolerance), which are the focus of the current report, is not known. In this study, we have a number of limitations. First, we have not directly measured insulin resistance or insulin sensitivity. Few studies have compared fasting insulin versus insulin resistance as a predictor of type 2 diabetes. Lillioja et al 14 showed that insulin resistance (as determined by hyperinsulinemic euglycemic clamp) was a better predictor than was fasting insulin, although both were strong predictors (hazard ratios of 30 and 15, respectively). Fasting insulin and I 30-0 / G 30-0 have been correlated with more definitive methods for assessing insulin secretion and resistance. 28 31 In the Mexico City Diabetes Study, HOMA IR was a slightly better predictor of the incidence of type 2 diabetes than were fasting insulin levels. 33 It is likely that more precise measurements of these variables would decrease misclassification and perhaps strengthen the present results. In conclusion, we have shown that prediabetic subjects have an atherogenic pattern of cardiovascular risk factors, and these changes are predominantly observed in prediabetic subjects with increased HOMA IR and fasting insulin ( insulin resistance ) at baseline. Insulin-sensitive converters to diabetes have a pattern of cardiovascular risk factors similar to nonconverters to diabetes. The most important possible implication of these findings is that different methods of preventing diabetes may have different effects on CHD, which is the most common cause of death in diabetic subjects. 1 3

980 Circulation March 7, 2000 Acknowledgments This work was supported by grants RO1HL24799 and R37HL36820 from the National Heart, Lung, and Blood Institute. Dr Burke was supported by the American Diabetes Association s mentor-based postdoctoral fellowship program. References 1. Kannel WB, McGee DL. Diabetes and glucose tolerance as risk factors for cardiovascular disease: the Framingham Study. Diabetes Care. 1979; 2:120 126. 2. Wingard DL, Barrett-Connor E. Heart disease and diabetes. In: Harris MI, ed. Diabetes in America. 2nd ed. Bethesda, Md: National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases, National Diabetes Data Group; 1995:429 448. 3. Haffner SM, Lehto S, Rönnemaa T, Pyörälä K, Laakso M. Mortality from coronary heart disease in subjects with type 2 diabetes and in nondiabetic subjects with and without prior myocardial infarction. N Engl J Med. 1998;339:229 234. 4. Kuusisto J, Mykkänen L, Pyörälä K, Laakso M. NIDDM and its metabolic control predict coronary heart disease in elderly subjects. Diabetes. 1994;43:960 967. 5. Klein R. Kelly West Lecture 1994: hyperglycemia and microvascular and macrovascular disease in diabetes. Diabetes Care. 1995;18:258 268. 6. West KM, Ahuja MMS, Bennett PH, Czyzyk A, DeAcosta OM, Fuller JH, Grab B, Bragauskas V, Jarrett RJ, Kosaka K. The role of circulating glucose and triglyceride concentrations and their interactions with other risk factors as determinants of arterial disease in nine diabetic population samples from the WHO Multinational Study. Diabetes Care. 1983; 6:361 369. 7. UKPDS Group. Intensive blood-glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes (UKPDS 33). Lancet. 1998;352:837 853. 8. Haffner SM. Management of dyslipidemia in adults with diabetes. Diabetes Care. 1998;21:160 178. 9. Medalie JH, Papier DM, Goldbourt U, Herman JB. Major factors in the development of diabetes mellitus in 10 000 men. Arch Intern Med. 1975; 135:811 817. 10. McPhillips JB, Barrett-Connor E, Wingard DL. Cardiovascular disease risk factors prior to the diagnosis of impaired glucose tolerance and non-insulin dependent diabetes mellitus in a community of older adults. Am J Epidemiol. 1990;131:443 453. 11. Mykkänen L, Kuusisto J, Pyörälä K, Laakso M. Cardiovascular disease risk factors as predictors of type II (non-insulin-dependent) diabetes mellitus in elderly subjects. Diabetologia. 1993;36:553 559. 12. Haffner SM, Stern MP, Hazuda HP, Mitchell BD, Patterson JK. Cardiovascular risk factors in confirmed prediabetic individuals: does the clock for coronary heart disease start ticking before the onset of clinical diabetes? JAMA. 1990;263:2893 2898. 13. Fagot-Campagna A, Narayan KM, Hanson RL, Imprematore G, Howard BV, Nelson RG, Pettitt DJ, Knowler WC. Plasma lipoproteins and incidence of non-insulin-dependent diabetes mellitus in Pima Indians: protective effect of HDL cholesterol in women. Atherosclerosis. 1997; 128:113 119. 14. Lillioja S, Mott DM, Spraul M, Ferraro R, Foley JE, Ravussin E, Knowler WC, Bennett PH, Bogardus C. Insulin resistance and insulin secretory dysfunction as precursors of non-insulin-dependent diabetes mellitus: prospective studies of Pima Indians. N Engl J Med. 1993;329:1988 1992. 15. Haffner SM, Stern MP, Mitchell BD, Hazuda HP, Patterson JK. Incidence of type II diabetes in Mexican Americans predicted by fasting insulin and glucose levels, obesity, and body fat distribution. Diabetes. 1990;39: 283 288. 16. Bergstrom RW, Newell-Morris LL, Leonetti DL, Shuman WP, Wahl PW, Fujimoto WY. Association of elevated fasting C-peptide levels and increased intra-abdominal fat distribution with development of NIDDM in Japanese American men. Diabetes. 1990;39:104 111. 17. Sicree RA, Zimmet PZ, King HOM, Coventry JS. Plasma insulin response among Nauruans: prediction of deterioration in glucose tolerance over 6 years. Diabetes. 1987;36:179 186. 18. Kadowaki T, Miyake Y, Hagura R, Akanuma Y, Kajinuma H, Kuzuya N, Takaku F, Kosaka K. Risk factors for worsening to diabetes in subjects with impaired glucose tolerance. Diabetologia. 1984;26:44 49. 19. Chen KW, Boyko EJ, Bergstrom RW, Leonetti DL, Newell-Morris L, Wahl PW, Fujimoto WY. Earlier appearance of impaired insulin secretion than of visceral adiposity in the pathogenesis of NIDDM: 5 year follow-up of initially non-diabetic Japanese American men. Diabetes Care. 1995;18:747 753. 20. Haffner SM, Miettinen H, Gaskill SP, Stern MP. Decreased insulin secretion and increased insulin resistance are independently related to the 7-year risk of non-insulin dependent diabetes mellitus. Diabetes. 1995; 44:1386 1391. 21. Lundgren H, Bengtsson C, Blohme G, Lapidus L, Waldenstrom J. Fasting serum insulin concentrations and early insulin response as risk determinants for developing diabetes. Diabet Med. 1990;7:407 413. 22. Reaven GM. Role of insulin resistance in human disease. Diabetes. 1988;37:1596 1607. 23. Haffner SM, Miettinen H. Insulin resistance implications for type II diabetes mellitus and coronary heart disease. Am J Med. 1997;193: 152 162. 24. The Diabetes Prevention Program Research Group. The Diabetes Prevention Program: design and methods for a clinical trial in the prevention of type 2 diabetes. Diabetes Care. 1999;22:623 634. 25. Chiasson JL, Gomis R, Hanefeld M, Josse RG, Karasik A, Laakso M, the STOP-NIDDM Trial Research Group. The STOP-NIDDM trial: an international study on the efficacy of an -glucosidase inhibitor to prevent type 2 diabetes in a population with impaired glucose tolerance: rationale, design, and preliminary screening data. Diabetes Care. 1998;21: 1720 1725. 26. Haffner SM, Stern MP, Hazuda HP, Pugh J, Patterson J. Hyperinsulinemia in a population at high risk for non-insulin dependent diabetes mellitus. N Engl J Med. 1986;315:220 224. 27. World Health Organization Study Group on Diabetes Mellitus. Diabetes Mellitus: Report of a WHO Study Group. Geneva, Switzerland: World Health Organization (Technical Report Series, No. 727); 1985:94 98. 28. Matthews DR, Hosker JP, Rudenski AS, Naylor GA, Treacher DF, Turner RL. Homeostasis model assessment: insulin resistance and -cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia. 1985;28:412 419. 29. Laakso M. How good a marker is insulin level for insulin resistance? Am J Epidemiol. 1993;137:959 965. 30. Saad MF, Anderson RL, Laws A, Watanabe RM, Kades WW, Chen YD, Sands RE, Pei D, Savage PJ, Bergman RN. A comparison between the minimal model and the glucose clamp in the assessment of insulin sensitivity across the spectrum of glucose tolerance: Insulin Resistance Atherosclerosis Study. Diabetes. 1994;43:1114 1121. 31. Philips DI, Clark PM, Hales CN, Osmond C. Understanding oral glucose tolerance: comparison of glucose and insulin measurements during the oral glucose tolerance test with specific measurements of insulin resistance and secretion. Diabet Med. 1994;11:286 292. 32. Groop LC. Sulfonylureas in NIDDM. Diabetes Care. 1992;15:737 754. 33. Haffner SM, Miettenen H, Stern MP. The homeostasis model in the San Antonio Heart Study. Diabetes Care. 1997;20:1087 1092.