Hertzel C Gerstein1, Guillaume Paré2, Matthew J McQueen2, Heinz Haenel2, Shun Fu Lee2, Janice Pogue2, Aldo P Maggioni2, Salim Yusuf2, Sibylle Hess2. 1. From Population Health Research Institute (H.C.G., G.P., M.J.M., S.F.L., J.P., S.Y.) and Thrombosis and Atherosclerosis Research Institute (G.P.), Hamilton Health Sciences and McMaster University, Ontario, Canada; Sanofi Aventis Deutschland GmbH R&D Division Diabetes (H.H., S.H.), Frankfurt, Germany; and ANMCO Research Centre (A.P.M.), Florence, Italy. gerstein@mcmaster.ca. 2. From Population Health Research Institute (H.C.G., G.P., M.J.M., S.F.L., J.P., S.Y.) and Thrombosis and Atherosclerosis Research Institute (G.P.), Hamilton Health Sciences and McMaster University, Ontario, Canada; Sanofi Aventis Deutschland GmbH R&D Division Diabetes (H.H., S.H.), Frankfurt, Germany; and ANMCO Research Centre (A.P.M.), Florence, Italy.
Abstract
BACKGROUND: Serum biomarkers may identify people at risk for cardiovascular (CV) outcomes. Biobanked serum samples from 8494 participants with dysglycemia in the completed Outcome Reduction With Initial Glargine Intervention trial were assayed for 284 biomarkers to identify those that could identify people at risk for a CV outcome or death when added to clinical measurements. METHODS AND RESULTS: A multiplex analysis measured a panel of cardiometabolic biomarkers in 1 mL of stored frozen serum from every participant who provided biobanked blood. After eliminating undetectable or unanalyzable biomarkers, 8401 participants who each had a set of 237 biomarkers were analyzed. Forward-selection Cox regression models were used to identify biomarkers that were each independent determinants of 3 different incident outcomes: (1) the composite of myocardial infarction, stroke, or CV death; (2) these plus heart failure hospitalization or revascularization; and (3) all-cause death. When added to clinical variables, 10 biomarkers were independent determinants of the 1405 CV composite outcomes observed during follow-up; 9 biomarkers (including 8 of these 10) were independent determinants of the 2435 expanded composite outcomes; and 15 (including the 10 CV composite biomarkers) were independent determinants of the 1340 deaths. Adjusted C statistics increased from 0.64 for the clinical variables to 0.71 and 0.68 for the 2 CV composite outcomes, respectively, with the greatest increase to 0.75 for death (P<0.001 for the change). CONCLUSIONS: A systematic hypothesis-free approach identified combinations of up to 15 cardiometabolic biomarkers as independent determinants of CV outcomes or death in people with dysglycemia. CLINICAL TRIAL REGISTRATION: URL: http://www.clinicaltrials.gov. Unique identifier: NCT00069784.
RCT Entities:
BACKGROUND: Serum biomarkers may identify people at risk for cardiovascular (CV) outcomes. Biobanked serum samples from 8494 participants with dysglycemia in the completed Outcome Reduction With Initial Glargine Intervention trial were assayed for 284 biomarkers to identify those that could identify people at risk for a CV outcome or death when added to clinical measurements. METHODS AND RESULTS: A multiplex analysis measured a panel of cardiometabolic biomarkers in 1 mL of stored frozen serum from every participant who provided biobanked blood. After eliminating undetectable or unanalyzable biomarkers, 8401 participants who each had a set of 237 biomarkers were analyzed. Forward-selection Cox regression models were used to identify biomarkers that were each independent determinants of 3 different incident outcomes: (1) the composite of myocardial infarction, stroke, or CV death; (2) these plus heart failure hospitalization or revascularization; and (3) all-cause death. When added to clinical variables, 10 biomarkers were independent determinants of the 1405 CV composite outcomes observed during follow-up; 9 biomarkers (including 8 of these 10) were independent determinants of the 2435 expanded composite outcomes; and 15 (including the 10 CV composite biomarkers) were independent determinants of the 1340 deaths. Adjusted C statistics increased from 0.64 for the clinical variables to 0.71 and 0.68 for the 2 CV composite outcomes, respectively, with the greatest increase to 0.75 for death (P<0.001 for the change). CONCLUSIONS: A systematic hypothesis-free approach identified combinations of up to 15 cardiometabolic biomarkers as independent determinants of CV outcomes or death in people with dysglycemia. CLINICAL TRIAL REGISTRATION: URL: http://www.clinicaltrials.gov. Unique identifier: NCT00069784.
Authors: Jennifer Sjaarda; Hertzel C Gerstein; Salim Yusuf; Darin Treleaven; Michael Walsh; Johannes F E Mann; Sibylle Hess; Guillaume Paré Journal: J Am Soc Nephrol Date: 2018-03-06 Impact factor: 10.121