Diabetes and stress hyperglycemia associated with myocardial infarctions at an urban municipal hospital: prevalence and effect on mortality

Abstract

Municipal hospitals in large cities provide care for patients from immigrant and mixed ethnic communities that are at high risk for diabetes. Both diabetes and stress hyperglycemia increase the risk of adverse outcome after myocardial infarctions, and the impact of stress hyperglycemia on the outcome of myocardial infarctions in this particular setting has not been previously studied. We therefore undertook a retrospective cohort study to determine the prevalence of diabetes and stress hyperglycemia in patients presenting to a university-affiliated Bronx municipal hospital with myocardial infarction, and the relationship of these conditions to the extent of coronary disease and mortality. We obtained data on 106 consecutive patients from July 1998 to April 1999 with a diagnosis-related group diagnosis of either myocardial infarction or acute coronary syndrome, in which myocardial infarction was confirmed by serum enzymes or characteristic electrocardiographic changes. Patients were followed until March 30, 2001. Measurements of clinical parameters and results of catheterization were obtained for all patients. Death rates were determined by laboratory database, direct patient contact, or data from National Death Index. Eighty percent of the cohort had either a diagnosis of diabetes (n = 45, 42% of cohort) or evidence of stress hyperglycemia (defined as serum glucose greater than 126 mg/dL at the time of admission without prior diagnosis of diabetes, n = 40, 38%). In-hospital mortality for patients with diabetes, stress hyperglycemia, or normal glucose was 20%, 15%, and 14%, respectively. Eighty-three percent of the cohort received beta blockers, and 61% of hospital survivors had catheterization. Left main or triple vessel disease was common in both patients with diabetes (52%) and patients with stress hyperglycemia (32%). Mortality at follow up (maximum follow up 3 years; mean follow up 19.6 months) was much higher in patients with either diabetes (42%) or stress hyperglycemia (52%) than normal subjects (24%). Kaplan-Meier analysis of the difference in mortality between patients with high glucose on admission and normal subjects was borderline significant (P = 0.06). Multivariate regression demonstrated that age (P = 0.020), increase in admission serum creatinine (P = 0.001), and reduction in either ejection fraction (P = 0.016) or admission systolic blood pressure (P = 0.005) were significant predictors of mortality. Glycemic status and sex were not independently associated with death after controlling for these other factors. These results show that the prevalence of both diabetes and stress hyperglycemia on presentation with myocardial infarction is strikingly high in this immigrant, mixed ethnic, urban population. Patients with diabetes and stress hyperglycemia had advanced disease on presentation and much higher mortality at 2 to 3 years than those with normal blood glucose. The mortality difference is the result of older age and more advanced disease rather than hyperglycemia per se.