Cardiovascular Outcomes in the Irbesartan Diabetic Nephropathy Trial of Patients with Type 2 Diabetes and Overt Nephropathy

Context Previously published results of this randomized, double-blind trial showed that high-risk patients with type 2 diabetic nephropathy had better renal protection if they were treated with irbesartan rather than amlodipine in addition to conventional antihypertensive therapy. Contribution These detailed analyses showed no differences in overall cardiovascular outcomes between patients given irbesartan or amlodipine. Fewer patients given irbesartan had heart failure and fewer patients given amlodipine had heart attacks. Cautions The trial had limited power to detect important differences between groups in mortality or strokes, and most patients received several antihypertensive agents. The Editors Patients with diabetes have an increased risk for cardiovascular complications and death (1). Studies that analyzed the effects of inhibition of the reninangiotensin system on the risk for cardiovascular complications included a substantial number of patients with diabetes (2-5) or were done exclusively in patients with diabetes (6-8). The meta-analysis of these studies (9), the analysis of the diabetic cohorts in the Heart Outcomes Prevention Evaluation (HOPE) study (2), and the Losartan Intervention for Endpoint Reduction in Hypertension (LIFE) trial (5) demonstrated that angiotensin-converting enzyme (ACE) inhibitors (2, 9) and angiotensin-receptor blockers (5) had a statistically significant advantage over placebo or alternative agents in decreasing the risk for several cardiovascular events. These studies randomly assigned few patients with renal involvement and overt proteinuria. Overt proteinuria occurred in fewer than 20% of the 470 patients in the Appropriate Blood Pressure Control in Diabetes (ABCD) trial (6), and only 11% of the 1195 patients in the LIFE trial (5). The Captopril Prevention Project (CAPP) (3) and the Swedish Trial in Old Patients with Hypertension-2 (STOP Hypertension-2) (4) did not state the number of patients with diabetes and overt proteinuria. There were no such patients in the Fosinopril versus Amlodipine Cardiovascular Events Trial (FACET) (7), and patients with dipstick-positive albuminuria were excluded from the HOPE trial (2). Since proteinuria is an independent risk factor for cardiovascular disease (10, 11), the data obtained in the aforementioned trials cannot be extrapolated to patients with type 2 diabetes and overt nephropathy. Trials performed in such patients have reported a blood pressureindependent effect of two different angiotensin-receptor blocker agents to protect against nephropathy (12, 13) without a change in all-cause mortality. Apart from studies in heart failure, few cardiovascular data exist for receptor blockers compared with either placebo or calcium-channel blockers. We report on the analysis of the cardiovascular end points that were monitored as secondary end points in the Irbesartan Diabetic Nephropathy Trial (IDNT) (12) and assess whether an angiotensin II receptor blocker or a calcium-channel blocker alters the risk for cardiovascular events beyond those observed by blood pressure reduction alone without such agents. Methods Patients The IDNT was a randomized, double-blind study on the effect of treatment with irbesartan or amlodipine compared with placebo in patients with type 2 diabetic nephropathy. The protocol of this study has been published (12, 14). Entry criteria required that patients be between 30 and 70 years of age and have type 2 diabetes mellitus and overt nephropathy, as evidenced by current treatment for hypertension or by a protein excretion rate of 900 mg/d or greater, serum creatinine level of 89 mol/L (1.0 mg/dL) to 266 mol/L (3.0 mg/dL) in women or of 106 mol/L (1.2 mg/dL) to 266 mol/L (3.0 mg/dL) in men, and baseline seated blood pressure greater than 135/85 mm Hg. The institutional review boards of each center approved the protocol. All patients gave written informed consent. Treatment and Randomization Patients were randomly assigned centrally by computer to receive treatment with irbesartan, 300 mg/d (Avapro, Bristol-Myers Squibb, Princeton, New Jersey); amlodipine, 10 mg/d (Norvasc, Pfizer, New York); or matched placebo. To minimize any center effect, randomization was blocked by center. All patients had blood pressure controlled to the same blood pressure goal of less than 135/85 mm Hg by using antihypertensive agents other than ACE inhibitors, angiotensin II receptor blocking agents, or calcium-channel blockers. For the analysis of cardiovascular end points, patients were followed to initiation of treatment for end-stage renal failure (dialysis or renal transplantation), reaching a serum creatinine level of 530.4 mol/L (6.0 mg/dL) or higher, death, or administrative censoring in December 2000. Outcomes We prospectively established cardiovascular outcomes, defined in the Appendix Table. Appendix Table. Classification for Fatal and Nonfatal Cardiovascular Events Ascertainment of Cardiovascular Events Information about hospitalizations and adverse events were screened at Bristol-Myers Squibb, Princeton, New Jersey, by trained, blinded clinical research associates to identify potential cardiovascular events. Investigators used study forms to report and characterize all cardiovascular outcomes. For all potential events, records, including laboratory values, electrocardiograms, and radiographic reports were obtained for clarification. Since myocardial infarctions may go unrecognized, a central electrocardiogram reading center was established at Brigham and Women's Hospital, Boston, Massachusetts, where two cardiologists reviewed every electrocardiogram. Electrocardiography was performed at baseline, 6 months, 12 months, and annually thereafter. A total of 5698 electrocardiograms were reviewed at the center. When a new Q-wave infarction was found, the cardiologists asked whether a clinical myocardial infarction was reported. Even when myocardial infarctions were not clinically reported, these Q-wave infarctions were adjudicated as myocardial infarctions. Adjudication of Cardiovascular Events Investigators at each center reported cardiovascular events, defined in the Appendix Table. The information on all potential events was referred to one member of the Outcomes Confirmation and Classification Committee (Appendix). If the committee member agreed with the judgment of the center investigator, their combined judgment was accepted. If the center investigator and the committee member differed, the case material was reviewed by the membership of the committee, whose decision was accepted. Deaths were adjudicated by a Mortality Committee (Appendix). Each death was reviewed by two members of the committee and presented to the membership, whose decision was accepted as final. Statistical Analysis For graphical presentation (Figure) and overall testing for statistically significant differences among the three treatment groups, time to the first occurrence of either a specific cardiovascular outcome or one of the composite outcomes was analyzed by product-limit survival curves and the log-rank test (15). We used proportional hazards modeling to determine hazard ratios. For the cardiovascular death outcome, which could occur only once, we used the standard proportional hazards model (16), with treatment assignment as the only independent covariate. For other cardiovascular outcomes, which could occur more than once, we used the AndersonGill formulation of the proportional hazards model (17), in which patients are considered at risk for the first event from randomization to the first event, at risk for the second event from the day following the first event to the second event, and so forth, permitting use of all the data. In accordance with the method of Lee and colleagues (18), we used a robust variance estimate that accounts for the possibility of correlation of risk for several events within a patient. We believed that occurrence of a first event of a given type increases the likelihood of a subsequent similar event. Therefore, both treatment assignment and a time-dependent covariate indicating whether the event was the first of its type or a subsequent event were included in these analyses. The time-dependent covariate was statistically significant in each case, confirming the above assumption. There was no statistically significant interaction between treatment and the time-dependent covariatethe effects of treatment assignment were similar for first and subsequent eventsand inclusion of the time-dependent covariate did not change either the estimates of the treatment effect or their statistical significances. Figure. Time to first cardiovascular composite event as a function of treatment assignment. P Data management and computations were done by using SAS software for Windows, version 8 (SAS Institute, Inc., Cary, North Carolina), or S-Plus for Windows, version 6.0 (Insightful Corp., Seattle, Washington). Statistical tests were two sided. A P value of 0.05 or less, unadjusted for the multiple comparisons, was considered statistically significant. Role of the Funding Sources The funding sources were involved in the data collection but not in the analysis or interpretation or the decision to submit the manuscript for publication. Results The baseline characteristics of the three groups are shown in Table 1. A flow diagram of the study is shown in the Appendix Figure. Table 1. Baseline Characteristics Appendix Figure. Flow diagram for the Irbesartan Diabetic Nephropathy Trial. Clinical Management During the study, the blood pressure decreased from the baseline values to 140/77 mm Hg in the irbesartan group, 141/77 mm Hg in the amlodipine group, and 144/80 mm Hg in the placebo group. Blood pressure in the two active treatment groups did not differ; values in both groups were statistically significantly lower than in the placebo group (P = 0.001). The distribution of nonstudy drugs used to achieve the target blood pressure was similar i