Vincent Bataille1, Jean Ferrières1,2, Nicolas Danchin3, Etienne Puymirat3, Marianne Zeller4, Tabassome Simon5,6,7, Didier Carrié1. 1. 1 Department of Cardiology B, Toulouse-Rangueil University Hospital, Toulouse University School of Medicine, France. 2. 2 Department of Epidemiology, Health Economics and Public Health, UMR 1027 INSERM- University of Toulouse III, Toulouse University School of Medicine, France. 3. 3 APHP, Hôpital Européen Georges Pompidou, Paris, France. 4. 4 Laboratoire de Physiopathologie et Pharmacologie Cardiométaboliques, UMR INSERM 866, UFR Sciences de Santé, Dijon, France. 5. 5 APHP, Department of Pharmacology, URCEST-CRB-CRCEST-Hôpital Saint Antoine, Paris, France. 6. 6 Université Pierre et Marie Curie, Paris, France. 7. 7 INSERM, U-1148, CHU Bichat, Paris, France.
Abstract
BACKGROUND: Merits of insulin use for diabetes treatment in patients with advanced atherosclerosis are debated. This observational study conducted in diabetic patients after an acute myocardial infarction aimed to assess whether insulin prescription at discharge (IPD) was related to all-cause mortality during follow-up. METHODS: Subjects were diabetic patients admitted in intensive- or coronary-care units for acute myocardial infarction (consecutively recruited in 223 centres in France) and discharged alive from the hospital, with or without an IPD. Vital status after five years was obtained and the relationship between insulin prescription at discharge and survival was studied. RESULTS: Overall, 1221 diabetic patients were discharged alive and 38% had an IPD. Factors independently related to IPD were female gender, hospitalization in a public hospital, duration of diabetes, HbA1c level, smoking, peripheral artery disease, history of coronary heart disease and Killip class. After adjustment, IPD was independently related to all-cause mortality after five years of follow-up (adjusted hazard ratio = 1.72 (1.42-2.09), p<0.001). This increased mortality in subjects with IPD was also observed in propensity matched analyses, when subjects actually treated or actually not treated with insulin at discharge were compared in two groups matched on their computed probability of having had insulin prescribed. CONCLUSIONS: Insulin was preferably prescribed in seriously affected patients, regarding diabetes and cardiovascular risk. However, insulin prescription at discharge was associated with increased all-cause mortality after extensive adjustments for confounders. These results suggest possible intrinsic harmful effects of insulin in high-risk diabetic patients after myocardial infarction.
BACKGROUND: Merits of insulin use for diabetes treatment in patients with advanced atherosclerosis are debated. This observational study conducted in diabetic patients after an acute myocardial infarction aimed to assess whether insulin prescription at discharge (IPD) was related to all-cause mortality during follow-up. METHODS: Subjects were diabetic patients admitted in intensive- or coronary-care units for acute myocardial infarction (consecutively recruited in 223 centres in France) and discharged alive from the hospital, with or without an IPD. Vital status after five years was obtained and the relationship between insulin prescription at discharge and survival was studied. RESULTS: Overall, 1221 diabetic patients were discharged alive and 38% had an IPD. Factors independently related to IPD were female gender, hospitalization in a public hospital, duration of diabetes, HbA1c level, smoking, peripheral artery disease, history of coronary heart disease and Killip class. After adjustment, IPD was independently related to all-cause mortality after five years of follow-up (adjusted hazard ratio = 1.72 (1.42-2.09), p<0.001). This increased mortality in subjects with IPD was also observed in propensity matched analyses, when subjects actually treated or actually not treated with insulin at discharge were compared in two groups matched on their computed probability of having had insulin prescribed. CONCLUSIONS: Insulin was preferably prescribed in seriously affected patients, regarding diabetes and cardiovascular risk. However, insulin prescription at discharge was associated with increased all-cause mortality after extensive adjustments for confounders. These results suggest possible intrinsic harmful effects of insulin in high-risk diabetic patients after myocardial infarction.