ABSTRACT
BACKGROUND: In 2010, after regulatory review of rosiglitazone licensing, the US Food and Drug Administration (FDA) requested a reevaluation of cardiovascular end points in the RECORD trial. METHODS: Automated screening of the original clinical trial database and manual case report form review were performed to identify all potential cardiovascular and noncardiovascular deaths, and nonfatal myocardial infarction (MI) and stroke events. Search techniques were used to find participants lost to follow-up, and sites were queried for additional source documents. Suspected events underwent blinded adjudication using both original RECORD end point definitions and new FDA end point definitions, before analysis by the Duke Clinical Research Institute. RESULTS: The reevaluation effort included an additional 328 person-years of follow-up. Automated screening identified 396 suspected deaths, 2,052 suspected MIs, and 468 suspected strokes. Manual review of documents by Duke Clinical Research Institute clinical events classification (CEC) coordinators identified an additional 31 suspected deaths, 49 suspected MIs, and 28 suspected strokes. There were 127 CEC queries issued requesting additional information on suspected deaths; 43 were closed with no site response, 61 were closed with a response that no additional data were available, and additional data were received for 23. Seventy CEC queries were issued requesting additional information for suspected MI and stroke events; 31 were closed with no site response, 20 were closed with a response that no additional data were available, and 19 resulted in additional data. CONCLUSIONS: Comprehensive procedures were used for rigorous event reascertainment and readjudication in a previously completed open-label, global clinical trial. These procedures used in this unique situation were consistent with other common approaches in the field, were enhanced to address the FDA concerns about the original RECORD trial results, and could be considered by clinical trialists designing event readjudication protocols for drug development programs that have been completed.
Subject(s)
Hypoglycemic Agents/adverse effects , Myocardial Infarction/chemically induced , Randomized Controlled Trials as Topic , Research Design , Stroke/chemically induced , Thiazolidinediones/adverse effects , Follow-Up Studies , Humans , Mortality , Myocardial Infarction/epidemiology , Rosiglitazone , Stroke/epidemiology , United States , United States Food and Drug AdministrationABSTRACT
BACKGROUND: The US Food and Drug Administration (FDA) required a reevaluation of cardiovascular (CV) outcomes in the RECORD trial. This provided an opportunity to assess the implications of event adjudication by 2 groups and quantify the differences as well as to use new FDA end point definitions in development. METHODS: Original data were used to systematically identify all potential deaths, myocardial infarctions (MIs), and strokes. Site investigators were approached for additional source documents and information about participants lost to follow-up. Suspected events were adjudicated using standard procedures, and the results were compared with the original trial outcomes. RESULTS: Follow-up for mortality was 25,833 person-years, including an additional 328 person-years identified during the reevaluation effort. A total of 184 CV or unknown-cause deaths (88 rosiglitazone, 96 metformin/sulfonylurea), 128 participants with an MI (68 rosiglitazone, 60 metformin/sulfonylurea), and 113 participants with a stroke (50 rosiglitazone, 63 metformin/sulfonylurea) were included. The hazard ratio (HR) for rosiglitazone versus metformin/sulfonylurea for the end point of CV (or unknown cause) death, MI, or stroke was 0.95 (95% CI 0.78-1.17) compared with 0.93 (95% CI 0.74-1.15) for the original RECORD results. Treatment comparisons for MI (HR 1.13, 95% CI 0.80-1.59) and mortality (HR 0.86, 95% CI 0.68-1.08) were also the same compared with the original RECORD results. Sensitivity analyses were also consistent with the original RECORD results. Analyses using the FDA definitions showed similar results. CONCLUSIONS: Only a modest number of additional person-years of follow-up were ascertained from this reevaluation of CV end points in RECORD. Observed HRs and CIs from these analyses using the original RECORD or new FDA end point definitions showed similar treatment effects of rosiglitazone compared with the original RECORD results.