TYPE 2 DIABETES MELLITUS (T2DM) is a heterogeneous

Transcription

1 X/04/$15.00/0 The Journal of Clinical Endocrinology & Metabolism 89(6): Printed in U.S.A. Copyright 2004 by The Endocrine Society doi: /jc Pharmacological Treatment of Insulin Resistance at Two Different Stages in the Evolution of Type 2 Diabetes: Impact on Glucose Tolerance and -Cell Function ANNY H. XIANG, RUTH K. PETERS, SIRI L. KJOS, JOSE GOICO, CESAR OCHOA, AURA MARROQUIN, SYLVIA TAN, HOWARD N. HODIS, STANLEY P. AZEN, AND THOMAS A. BUCHANAN Departments of Preventive Medicine (A.H.X., R.K.P., S.T., H.N.H., S.P.A.), Obstetrics and Gynecology (S.L.K., T.A.B.), and Medicine (J.G., C.O., A.M., H.N.H., T.A.B.), University of Southern California Keck School of Medicine, Los Angeles, California The purpose of this study was to compare the impact of treating insulin resistance with a thiazolidinedione drug before vs. at the onset of diabetes on glucose levels and -cell function. Nondiabetic Hispanic women of Mexican or Central American descent with prior gestational diabetes mellitus (GDM) were randomized to troglitazone (early intervention), 400 mg/d, or placebo (later intervention). Women who developed diabetes were placed on open-label troglitazone. Glucose tolerance, insulin resistance, and -cell function were measured at randomization, at the diagnosis of diabetes, and 8 months post trial to determine the long-term impact of the two treatment strategies on glucose levels and -cell function. During a mean follow-up of 4.3 yr between baseline and posttrial tests, glucose tolerance (oral glucose tolerance test glucose area, P 0.04) and insulin resistance (MINMOD S I, P 0.02) worsened more in women randomized to late intervention (n 69) than to early intervention (n 57). Insulin TYPE 2 DIABETES MELLITUS (T2DM) is a heterogeneous disorder in which pancreatic -cell dysfunction that is usually progressive occurs on a background of insulin resistance (1). We found that successful treatment of insulin resistance in young, nondiabetic Hispanic women with recent GDM lowered endogenous insulin requirements and stabilized -cell function, thereby preventing T2DM for 4.5 yr (2). That observation suggests strongly that insulin resistance causes or worsens the -cell dysfunction that characterizes GDM and subsequent T2DM in Hispanic patients (3 6). Our prior observations on stabilization of -cell function (2) were made in women who had either normal or impaired glucose tolerance at the initiation of treatment. The present analysis was conducted to compare that early treatment strategy to one in which treatment is withheld until the onset of T2DM. Abbreviations: AIRg, Acute insulin response to iv glucose; DI, disposition index; GDM, gestational diabetes mellitus; ivgtt, iv glucose tolerance test; Kg, glucose disappearance; ogtt, oral glucose tolerance test; S I, insulin sensitivity ; T2DM, type 2 diabetes mellitus; TRIPOD, Troglitazone in Prevention of Diabetes (study); UKPDS, United Kingdom Prospective Diabetes Study. JCEM is published monthly by The Endocrine Society ( endo-society.org), the foremost professional society serving the endocrine community. secretion (acute insulin response in the iv glucose tolerance test, P 0.09) and -cell compensation for insulin resistance (disposition index, P 0.07) also tended to worsen more in the late intervention group. Among women in the late intervention group who developed diabetes, oral glucose tolerance test glucose area (P ) and -cell function (P < 0.04) deteriorated significantly during development of diabetes on placebo and then did not change significantly (P > 0.50) during treatment with troglitazone and posttreatment washout. In high-risk Hispanic women, amelioration of insulin resistance can stabilize glycemia at the time diabetes develops. These findings highlight the role of insulin resistance in the genesis of progressive -cell dysfunction during the evolution of type 2 diabetes. (J Clin Endocrinol Metab 89: , 2004) Subjects and Methods Subjects and design The Troglitazone in Prevention of Diabetes (TRIPOD) study was designed to address the impact of troglitazone treatment on pancreatic -cell function and glucose levels in Hispanic women with prior GDM before the development of diabetes and also at the onset of diabetes. Descriptions of the design of the study and the results of the primary blinded trial (treatment before diabetes) have been published previously (2, 7). Briefly, between August 1995 and May 1998, 266 nondiabetic women of Mexican or Central American descent who were at least 18 yr old, with a history of GDM in the prior 4 yr and a total glucose area on a 75-g oral glucose tolerance test (ogtt) above the median for women with GDM (8), were randomly assigned to receive placebo or 400 mg of troglitazone daily in a double-blind fashion. Fasting plasma glucose was measured every 3 months, and 75-g ogtts were performed annually to test for diabetes (1). Women continued on their assigned medication until they developed diabetes, dropped out of the trial, or reached the end of the trial, when final on-trial testing for diabetes was performed. Women who developed diabetes while on blinded treatment before the end of the trial were offered treatment with troglitazone until they dropped out, reached the end of the trial, or developed a hemoglobin A 1C concentration, measured every 3 months from the diagnosis of diabetes onward, that was more than 7%. Thus, women initially assigned to troglitazone were given active drug before they developed diabetes (early intervention strategy). Women initially assigned to placebo were given active drug if and when they were first found to have diabetes during frequent testing for the disease (late intervention strategy). The study was designed to end in August Troglitazone was withdrawn from clinical use in March Subjects active in the trial at that time were asked to return for a final on-trial ogtt and then for a posttrial 2846

2 Xiang et al. Treatment of Insulin Resistance J Clin Endocrinol Metab, June 2004, 89(6): ogtt 8 months after stopping study medications. Frequently sampled ivgtts were performed at randomization, three months later, 3 months after entering the open label phase of the study for patients who did so, and 8 months post trial. All participants gave written informed consent for participation in the IRB-approved study. IRB approval and informed consent were updated during the study as information about potentially serious heptotoxicity of troglitazone became available. No serious hepatotoxicity was observed during the TRIPOD study. A summary of the frequency of reversible transaminase elevations appears in the description of the blinded trial (2). Clinical testing protocols ogtts and ivgtts were initiated between h, after an 8- to 12-h overnight fast. For ogtts, subjects drank 75 g dextrose. Venous blood was sampled from an indwelling catheter before and 30, 60, 90, and 120 min after the dextrose ingestion. For ivgtts, dextrose (300 mg/kg body weight) was injected into an antecubital vein. Tolbutamide (125 mg/m 2 body surface area) was injected 20 min later. Twenty-two arterialized venous blood samples were drawn and placed on ice before and up to 240 min after the dextrose injection. Plasma was separated within 20 min and stored at 80 C. Laboratory methods Glucose was measured by glucose oxidase (Beckman Glucose Analyzer II, Beckman Instruments, Brea, CA). Insulin was measured by a RIA (Linco Research, St. Charles, MO) that provided less than 0.2% cross-reactivity with proinsulin. Data analysis Whole-body insulin sensitivity (S I ) was calculated from ivgtts using the Bergman minimal model (9). Glucose disappearance (Kg) during ivgtts was calculated as 100 times the fractional Kg rate between 10 and 40 min after the glucose injection. Areas under glucose and insulin curves were calculated using the trapezoid rule. The acute insulin response to iv glucose (AIRg; the incremental insulin area between 0 10 min after the glucose injection) was used as a sensitive measure of -cell well-being (10), reflecting a combination of -cell mass (11) and function. (12). The product of AIRg and S I (the disposition index, DI) was used as a measure of the ability of -cells to compensate for insulin resistance (13 15). Baseline characteristics were compared between groups using twogroup t tests for continuous variables and 2 or Fisher s exact tests for categorical variables. For nonnormally distributed variables, such as insulin, insulin area, S I, and first-phase insulin response, natural logtransformation was used before t tests. Square root transformation was used for the DI. Means and sds are presented in the original measurement scales. Changes in variables over time were assessed as actual changes, percent changes, and rates of change, because subjects had different lengths of follow-up. These analyses led to similar conclusions, and only absolute changes are presented. Absolute changes were normally distributed and analyzed using parametric t tests. Data are presented as means sd in tables and text and as means se in figures. All statistical tests were two-sided, and statistical significance was set at P 0.05 for presentation of results. Blinded trial Results As reported previously (2), 236 women returned for at least one follow-up visit, 122 randomized to placebo (late intervention) and 114 randomized to troglitazone (early intervention). Annual dropout rates were 13.4% and 16.3%, respectively (P 0.44). Pill compliance rates, based on pill counts at each follow-up visit, were 87 10% and 85 16%, respectively (P 0.30). Average annual incidence rates of diabetes during the blinded trial were 12.1% and 5.4%, respectively (P 0.009). Comparison of early vs. later intervention strategies One hundred twenty-six women completed the trial and returned for posttrial testing, providing the full set of data required for inclusion in this report. Of the 110 women who ended participation before the final posttrial tests, three developed HbA 1c more than 7% during open-label treatment with troglitazone and were referred for management of their diabetes (two in the late and one in the early intervention group), and one (in the late intervention group) was pregnant at the time of posttrial testing. The remaining 106 women were lost to follow-up during the trial or posttrial period. At baseline, none of the following variables differed significantly between women who completed all required testing and women who ended their participation before posttrial testing in the cohort overall or within groups assigned to late intervention or early intervention: age, body mass index, ogtt glucose and insulin areas, ivgtt S I, AIRg, and DI (all P 0.18). Of the 126 women who completed the trial and posttrial testing, 69 had been randomized to late intervention, and 57 had been randomized to early intervention. Baseline characteristics were similar in the two groups (Table 1). Posttrial testing was completed at medians of 7.9 and 7.6 months after study medications were stopped in late and early intervention groups, respectively. These posttrial intervals corresponded to means of 4.3 and 4.4 yr after randomization in each group (Table 2). Glucose and insulin values during ogtts rose significantly between baseline and posttrial testing in the late intervention group but remained stable in the early intervention group (Fig. 1). Changes from baseline for other variables appear in Table 2. Weight increased significantly in both groups and slightly, but not significantly, more in the early intervention group. ogtt glucose area rose significantly, and S I fell significantly in the late intervention group, whereas neither variable changed significantly in the early intervention group. As a result, there were significant differences between groups in the changes in these two variables. AIRg, DI, and Kg fell significantly in the late intervention group. These measures fell less, and not significantly, in the early intervention group. Differences between groups in changes in AIRg, DI, and Kg were of borderline statistical significance (P ). Thus, assignment to the early intervention strategy was associated with significantly less worsening of glucose tolerance and insulin resistance and somewhat greater preservation of -cell function, compared with assignment to the late intervention strategy. Fifty-five of the 57 women in the early intervention group had ivgtts 3 months after starting blinded troglitazone. Thirteen of them had a change in S I that was less than min 1 / U/ml ( min 1 /pmol/liter) at 3-months (nonresponders, as defined in the description of the TRIPOD blinded trial (2). The other 42 women had a greater increase in S I (troglitazone responders (2). The ogtt glucose area was essentially unchanged in the 4.4 yr between baseline and posttrial testing in responders [ mg/dl min ( mmol/liter min), P 0.75] but rose significantly during the 4.3 yr between tests in nonresponders [ mg/dl min ( mmol/liter min), P 0.002]. The DI during the same interval was unchanged in the responders (1 845, P 0.99) but fell in the nonresponders ( , P 0.07). Thus, women who failed to respond to the primary insulin-sensitizing effect of troglitazone accounted for the small rise in glucose levels and small loss of -cell function that occurred in the early intervention group. Impact of troglitazone in women who developed diabetes Of the 126 women who completed the trial and posttrial testing, 23 in the late intervention group and 13 in the early intervention group developed diabetes and went on to open label treatment with troglitazone during the trial. Characteristics at study entry, at the diagnosis of diabetes, and at posttrial testing for these 36 women appear in Table 3 and Fig. 2. In the late intervention group (n 23), the average time from enrollment to development of diabetes was yr. The average duration of treatment with troglitazone plus posttrial washout was yr. Weight increased during both periods, although somewhat less during active treatment plus washout. Glucose areas during ogtts rose during the development of diabetes and then remained essentially stable during troglitazone treatment and washout. S I fell progressively across the two periods but not significantly in either period. AIRg fell in the presence of worsening insulin resistance during the development of diabetes and then increased slightly in the presence of additional insulin resistance during troglitazone treatment and washout. As a result, the DI fell during the development of diabetes, then stabilized during tro-

3 2848 J Clin Endocrinol Metab, June 2004, 89(6): Xiang et al. Treatment of Insulin Resistance TABLE 1. Baseline characteristics of women who completed the trial and returned for posttrial testing a Variable Intervention strategy Late (n 69) Early (n 57) P-value b Age (yr) 34.3 (6.6) 35.4 (6.5) 0.34 Body mass index (kg/m 2 ) 29.6 (4.6) 30.1 (5.3) 0.59 Waist:hip circumference ratio 0.86 (0.05) 0.85 (0.07) 0.33 ogtt Fasting glucose (mg/dl) 98.8 (9.6) 99.3 (10.0) h glucose (24.1) (23.8) 0.40 Total glucose area (mg/dl min 10 3 ) c 18.8 (2.0) 19.1 (2.2) 0.38 Impaired glucose tolerance d 78% 75% 0.83 Fasting insulin ( U/ml) 14.1 (6.5) 14.5 (7.4) 0.72 Total insulin area ( U/ml min) c 8791 (5166) 7975 (4018) 0.33 ivgtt e S I (min 1 / U/ml 10 4 ) f 2.02 (1.42) 2.17 (1.30) 0.53 AIRg ( U/ml min) g 392 (334) 329 (265) 0.25 DI (S I AIRg) h 873 (663) 804 (626) 0.55 Kg (min 1 100) i 1.49 (0.40) 1.42 (0.40) 0.39 To convert values for glucose to millimoles per liter, multiply by To convert values for insulin to picomoles per liter, multiply by 6.0. a Fraction of women in treatment group or mean (SD) in original scale except insulin, insulin area, S I and AIRg (geometric means), and DI (squared means after square root transformation). b By two-group t test for means, 2 or Fisher s exact test for proportions. c Calculated by trapezoid method using data from entire duration of test. d Plasma glucose mg/dl 2 h after 75-g oral glucose load. e ivgtt, as described in Subjects and Methods. f Calculated by the Bergman minimal model (9). g Incremental insulin area between 0 and 10 min after the glucose injection. h A measure of -cell compensation for insulin resistance. i 100 Fractional Kg rate min after glucose injection. TABLE 2. Changes between baseline and posttrial tests a Variable Late (n 69) Intervention strategy Within-group Early (n 57) Within-group P-value b P-value b Between-group P-value c Age (yr since randomization) 4.3 (0.8) 4.4 (0.7) 0.77 Weight (kg) 3.0 (5.4) (4.1) ogtt glucose area (mg/dl min 10 3 ) 1.8 (4.4) (3.5) S I (min 1 / U/ml 10 4 ) d 0.60 (1.69) (1.66) AIRg ( U/ml min) d 197 (491) (362) DI (S I AIRg) d 344 (642) (788) Kg (min 1 100) d 0.18 (0.37) (0.41) To convert values for glucose to millimoles per liter, multiply by To convert values for insulin to picomoles per liter, multiply by 6.0. a Mean (SD) in original measurement scales, without transformation. b By paired t test comparing baseline and posttrial values within groups. c By two groups t test comparing changes from baseline between groups. d As defined in Table 1. glitazone treatment and washout. In the early intervention group (n 13), the average times from enrollment to development of diabetes ( yr) and the duration of treatment with open-label troglitazone plus posttrial washout ( yr) were similar to the analogous intervals in the late intervention group. Weight increased significantly during development of diabetes on blinded medication and then decreased very slightly during open-label treatment and washout. Glucose areas from ogtts rose during both periods, but the rise was statistically significant only during development of diabetes on blinded medication. S I did not change significantly during either period or consistently across periods. AIRg and the DI fell during both periods, but the reductions within periods were not statistically significant in this small subgroup of women. Eighteen of the 23 women who participated in the late intervention by receiving open-label troglitazone after developing diabetes had an IVGTT 3 months after starting open-label drug. Twelve were responders to the insulin-sensitizing effect of troglitazone, and six were nonresponders, as defined above (2). The responders had better preservation of -cell function than the nonresponders. Specifically, in the 2.2 yr of troglitazone treatment and washout, the DI rose in the responders and fell in the nonresponders (P 0.04 between groups). These changes in -cell function were associated with changes in ogtt glucose areas in the responders [fall of mg/dl min ( mmol/liter min)] and nonresponders [rise of mg/dl min ( mmol/liter min)] that were directionally opposite but not significantly different (P 0.58). Discussion We report two main observations from the early vs. late intervention component of the TRIPOD study. First, early treatment designed to ameliorate insulin resistance before the diagnosis of diabetes was superior to later intervention at the initial diagnosis of diabetes regarding stabilization of glucose levels and insulin resistance and tended to be superior for preservation of pancreatic -cell function. Second,

4 Xiang et al. Treatment of Insulin Resistance J Clin Endocrinol Metab, June 2004, 89(6): FIG. 1. Plasma glucose and insulin concentrations during 75-g ogtts at baseline (closed symbols) and posttrial testing (open symbols) in women who participated in the entire trial and returned for posttrial testing after randomization to late intervention strategy (left panels, n 69) or early intervention strategy (right panels, n 57), as defined in Subjects and Methods. *,P 0.05; **, P 0.01 vs. baseline in the same treatment strategy. To convert values for glucose to millimoles per liter, multiply by To convert values for insulin to picomoles per liter, multiply by 6.0. the loss of -cell function that characterized progression to diabetes during placebo treatment was arrested when treatment with troglitazone was initiated at the onset of diabetes. Stabilization of -cell function was limited to women whose S I improved when they received troglitazone. These findings reveal that -cell function can be preserved by treatment of insulin resistance not only before the onset of diabetes, as we reported previously (2), but also when diabetes first develops. Delaying treatment until diabetes develops is accompanied by a rise in glucose levels and a loss of -cell function that cannot be reversed by slightly more than 2 years of troglitazone treatment. Whether the additional hyperglycemia and loss of -cell function are important to long-term health remains to be determined. The present findings compliment our prior analysis of the blinded component of the TRIPOD trial (2). That analysis revealed a significant reduction in the incidence of diabetes, along with preservation of -cell function in high-risk nondiabetic women who were randomized to troglitazone compared with placebo. The present results indicate that a similar effect can be achieved at a somewhat later stage, soon after glucose concentrations reach the level of diabetes. Our findings contrast with the common observation that T2DM is a progressive disease in patients who present with more advanced diabetes. For example, in the United Kingdom Prospective Diabetes Study (UKPDS), patients who presented with diabetic symptoms and complications (16) manifested rising glycemia (17 19) and falling -cell function (19) during treatment with diet, sulfonylurea drugs, metformin, and/or insulin. The precise reason for the different patterns of -cell function between the UKPDS and the present study is not known. One possibility is that thiazolidinedione drugs, which were not used in the UKPDS, have a unique effect to stabilize -cell function that was not provided by the treatments that were used in the UKPDS. In our analysis of the blinded component of the TRIPOD study, we found no evidence for a direct effect of troglitazone on -cells. Rather, protection from diabetes appeared to be mediated by a reduction in endogenous insulin requirements ( -cell rest) that was achieved, to varying degrees, by amelioration of insulin resistance (2). Given their large effect to ameliorate insulin resistance, thiazolidinediones may be more effective at mediating -cell rest than are treatments that were used in the UKPDS. Alternatively, delaying treatment of insulin resistance until patients have developed clinically apparent T2DM, as was done in the UKPDS and is common in clinical practice, could be inferior to earlier treatment, to the extent that reduced -cell mass at later stages (20) makes it progressively difficult to rest individual -cells. Finally, the duration of observation after development of diabetes in the TABLE 3. Characteristics at baseline, diagnosis of diabetes, and posttrial testing in women who developed diabetes during blinded treatment, switched to open-label troglitazone treatment, and returned for posttrial testing a Variable Baseline Diagnosis of diabetes Post trial P b P c Late intervention group (n 23) Age (yr) 34.4 (6.4) 36.4 (6.8) 38.7 (6.6) Weight (kg) 71.2 (13.7) 73.6 (13.0) 75.1 (14.8) OGTT glucose area (mg/dl min 10 3 ) d 20.3 (2.3) 23.2 (2.6) 23.1 (4.2) S I (min 1 / U/ml 10 4 ) d 1.63 (0.88) 1.48 (1.00) 1.23 (1.21) AIRg ( U/ml min) d 245 (236) 89 (244) 120 (145) DI (S I AIRg) d 517 (572) 335 (478) 318 (528) Early intervention group (n 13) Age (yr) 36.5 (4.7) 38.5 (4.9) 40.8 (4.7) Weight (kg) 74.3 (10.4) 78.9 (11.3) 78.2 (11.1) OGTT glucose area (mg/dl min 10 3 ) d 21.1 (1.8) 23.4 (2.7) 24.9 (4.2) S I (min 1 / U/ml 10 4 ) d 2.02 (0.95) 2.46 (1.73) 2.25 (1.30) AIRg ( U/ml min) d 182 (195) 125 (134) 80 (128) DI (S I AIRg) d 468 (511) 361 (377) 241 (464) To convert values for glucose to millimoles per liter, multiply by To convert values for insulin to picomoles per liter, multiply by 6.0. a Mean (SD) in original scale except for insulin, insulin area, S I and AIRg (geometric means), and DI (squared mean after square root transformation). b For change from baseline to development of diabetes (paired t test). c For change from diagnosis of diabetes to posttrial test (paired t test). d As defined in Table 1.

5 2850 J Clin Endocrinol Metab, June 2004, 89(6): Xiang et al. Treatment of Insulin Resistance FIG. 2. DI, a measure of -cell compensation for insulin resistance, in women from late (left panel, n 23) and early (right panel, n 13) intervention groups who developed diabetes on blinded medication (blinded placebo, blinded trog.) and then participated in open-label troglitazone treatment (open trog.) and posttrial testing (post trial). P-values are for paired t tests between successive time points. present study was relatively short (approximately 2 yr), whereas -cell deterioration in the UKPDS was observed over a longer period of time. Whether the stabilization of -cell function that we observed with troglitazone treatment at the onset of diabetes can be sustained for prolonged periods of time remains to be determined. Women who developed diabetes while taking blinded troglitazone did not appear to benefit from additional treatment with open-label troglitazone. Their glucose levels rose and their -cell function fell during development of diabetes on blinded drug and continued to do so when they were switched to open-label drug. These results are not surprising, because knowledge of treatment rather than the treatment itself was changed when subjects switched from blinded to open-label troglitazone. Nonetheless, the pattern of -cell function displayed by those women (Fig. 2, right panel) is important in that it demonstrates that continued loss of -cell function is the natural progression after diabetes develops in the absence of effective treatment. Against that background, the stabilization in glucose levels and -cell function that occurred in the late intervention group when they were placed on troglitazone at the onset of diabetes can be attributed with confidence to the effect of the drug. In the TRIPOD blinded trial (2), women who failed to increase S I when placed on troglitazone did benefit from the drug. The same general pattern was observed in the present analysis, although sample sizes and statistical power were lower than in the blinded phase of the study. Nonetheless, we observed that women in the early intervention group who failed to respond to the insulin-sensitizing effects of troglitazone lost -cell function, whereas responders had stable function for more than 4 yr. Similarly, nonresponders in the late intervention group continued to lose -cell function after being placed on troglitazone, whereas responders had a small increase in function 2.4 yr later. These findings, combined with the more robust analysis from the blinded phase of TRIPOD, provide strong evidence for the importance of treating insulin resistance to preserve -cell function. The present report indicates that the effect can be achieved at the time diabetes first develops. We focused this report on women who had measures of glucose tolerance, insulin resistance, and -cell function at baseline and after the posttrial washout period. This approach allowed us to determine changes in relevant variables in the absence of acute drug effects. It also resulted in the exclusion of data from a relatively large fraction of women who ended their participation before posttrial testing. The exclusions did not appear to be systematically biased regarding crucial metabolic parameters, because those parameters were similar at baseline in women who dropped out before posttrial tests and women who provided posttrial data for this report. Thus, our approach provided a valid assessment of the chronic effects of treatment on the biology of T2DM. The clinical relevance of our findings is limited by two facts. First, troglitazone is no longer available for clinical use. As detailed in our publication of the results of the blinded component of the TRIPOD study (2), the protective effect of troglitazone appeared to be mediated by a reduction in endogenous insulin requirements mediated by amelioration of insulin resistance. Thus, other methods for reducing insulin resistance, including weight loss and treatment with other insulin-sensitizing compounds, should have analogous beneficial effects. Such effects remain to be proven but could have contributed to protection from diabetes observed in the lifestyle intervention arms of the Diabetes Prevention Program (21) and the Diabetes Prevention Study (22), as well as the protection observed in the metformin arm of the former study (21). The second limitation is the tight ethnic focus of the TRIPOD study. All subjects were Hispanic American of Mexican or Central American ancestry. Whether amelioration of insulin resistance in other ethnic groups can preserve -cell function is unknown, although the relatively uniform effects of exercise and weight loss to reduce the incidence of diabetes in different ethnic groups in the Diabetes Prevention Program (21) suggests a generalized beneficial effect. In summary, amelioration of insulin resistance with troglitazone in Hispanic women with previous GDM stabilized glycemia and -cell function before the onset of diabetes and at the time diabetes was first detected by frequent testing. The cost of waiting until diabetes developed was some loss of glucose tolerance and -cell function that could not be recovered by 2 yr of treatment with troglitazone. When considered in light of the usually progressive course of hyperglycemia and declining -cell function in T2DM (17 19), our findings support a focus on prevention and early treatment of the disease through amelioration of insulin resistance to preserve pancreatic -cell function. Acknowledgments The authors thank Susie Nakao, Carmen Martinez, and the staff of the General Clinical Research Center for assistance with metabolic studies; the TRIPOD Data Safety Monitoring Committee for their guidance in the

6 Xiang et al. Treatment of Insulin Resistance J Clin Endocrinol Metab, June 2004, 89(6): conduct of the study; and Lilit Zeberians, Mike Salce, and Jay Sisson for performance of assays. Received November 24, Accepted March 9, Address all correspondence and requests for reprints to: Thomas A. Buchanan, M.D., Room 6602 GNH, 1200 North State Street, Los Angeles, California buchanan@usc.edu. This work was supported by Research Grant PD , Parke- Davis Pharmaceutical Research; M01-RR-43 from the General Clinical Research Branch, National Center for Research Resources, National Institutes of Health (NIH); Clinical Research Award from the American Diabetes Association; and R01-DK from the National Institute of Diabetes and Digestive and Kidney Diseases, NIH. References 1. American Diabetes Association 1997 Report of the Expert Committee on the Diagnosis and Classification of Diabetes Mellitus. Diabetes Care 20: Buchanan TA, Xiang AH, Peters RK, Kjos SL, Marroquin A, Goico J, Ochoa C, Tan S, Berkowitz K, Hodis HN, Azen SP 2002 Preservation of pancreatic B-cell function and prevention of type 2 diabetes by pharmacological treatment of insulin resistance in high-risk Hispanic women. Diabetes 51: Buchanan TA, Metzger BE, Freinkel N, Bergman RN 1990 Insulin sensitivity and B-cell responsiveness to glucose during late pregnancy in lean and moderately obese women with normal glucose tolerance or mild gestational diabetes. Am J Obstet Gynecol 162: Xiang AH, Peters RK, Trigo E, Kjos SL, Lee WP, Buchanan TA 1999 Multiple metabolic defects during late pregnancy in women at high risk for type 2 diabetes mellitus. Diabetes 48: Buchanan TA, Xiang AH, Kjos SL, Trigo E, Lee WP, Peters RK 1999 Antepartum predictors of the development of type 2 diabetes in Latino women months after pregnancies complicated by gestational diabetes. Diabetes 48: Buchanan TA, Xiang A, Kjos SL 1998 Gestational diabetes mellitus: antepartum metabolic characteristics that predict postpartum glucose intolerance and type 2 diabetes. Diabetes 47: Azen SP, Peters R, Berkowitz K, Kjos S, Xiang A, Buchanan TA 1998 TRIPOD: a randomized placebo-controlled trial of troglitazone in women with prior gestational diabetes mellitus. Control Clin Trials 19: Kjos SL, Peters RK, Xiang A, Henry OA, Montoro MN, Buchanan TA 1995 Predicting future diabetes in Latino women with gestational diabetes: utility of early postpartum glucose tolerance testing. Diabetes 44: Pacini G, Bergman RN 1986 MINMOD, a computer program to calculate insulin sensitivity and pancreatic responsivity from the frequently sampled intravenous glucose tolerance test. Comput Methods Programs Biomed 23: Brunzell JD, Robertson RP, Lerner RL, Hazzard WR, Ensinck JW, Bierman EL, Porte Jr D 1976 Relationships between fasting plasma glucose levels and insulin secretion during intravenous glucose tolerance tests. J Clin Endocrinol Metab 42: Kjems LL, Kirby BM, Welsh EM, Veldhuis JD, Straume M, McIntyre SS, Yang D, Lefebvre P, Butler PC 2001 Decrease in -cell mass leads to impaired pulsatile insulin secretion, reduced postprandial hepatic insulin clearance, and relative hyperglucagonemia in the minipig. Diabetes 50: Vague P, Moulin J-P 1982 The defective glucose sensitivity of the B-cell in noninsulin dependent diabetes: improvement after twenty hours of normoglycemia. Metabolism 31: Bergman RN 1989 Toward physiological understanding of glucose tolerance. Minimal model approach. Diabetes 38: Kahn SE, Prigeon RL, McCulloch DK, Boyko EJ, Bergman RN, Schwartz MW, Neifing JL, Ward WK, Beard JC, Palmer JP, Porte Jr D 1993 Quantification of the relationship between insulin sensitivity and B-cell function in human subjects: evidence for a hyperbolic function. Diabetes 42: Buchanan TA, Xiang AH, Peters RK, 2000 Response of pancreatic B-cells to improved insulin sensitivity in women at high risk for type 2 diabetes. Diabetes 49: Colagiuri S, Cull CA, Holman RR, UKPDS Study Group 2003 Are lower fasting plasma glucose levels at diagnosis of type 2 diabetes associated with improved outcomes? Diabetes Care 25: UKPDS Study Group 1998 Intensive blood glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes. Lancet 352: UKPDS Study Group 1998 Effect of intensive blood glucose control with metformin on complications in overweight patients with type 2 diabetes. Lancet 352: UKPDS Study Group 1995 Overview of 6 years therapy of type II diabetes: a progressive disease. Diabetes 44: Butler AE, Janson J, Bonner-Weir S, Ritzel R, Rizza RA, Butler PC 2003 B-cell deficit and increased B-cell apoptosis in humans with type 2 diabetes. Diabetes 52: Diabetes Prevention Program Research Group 2002 Reduction in the incidence of type 2 diabetes with lifestyle intervention or metformin. N Engl J Med 346: Tuomilehto J, Lindstrom J, Eriksson JG, Valle TT, Hamalainen H, Ilanne- Parikka P, Keinanen-Kiukaanniemi S, Laakso M, Louheranta A, Rastas M, Salminen V, Uusitupa M; Finnish Diabetes Prevention Study Group 2001 Prevention of type 2 diabetes mellitus by changes in lifestyle among subjects with impaired glucose tolerance. N Engl J Med 344: JCEM is published monthly by The Endocrine Society ( the foremost professional society serving the endocrine community.