Abstract
OBJECTIVES: To evaluate the predictive value of fasting plasma glucose (FPG) for in-hospital mortality in patients with acute myocardial infarction (AMI) with different glucose metabolism status. METHODS: We selected 5,308 participants with AMI from the prospective, nationwide, multicenter CAMI registry, of which 2,081 were diabetic and 3,227 were nondiabetic. Patients were divided into high FPG and low FPG groups according to the optimal cutoff values of FPG to predict in-hospital mortality for diabetic and nondiabetic cohorts, respectively. The primary endpoint was in-hospital mortality. RESULTS: Overall, 94 diabetic patients (4.5%) and 131 nondiabetic patients (4.1%) died during hospitalization, and the optimal FPG thresholds for predicting in-hospital death of the two cohorts were 13.2 mmol/L and 6.4 mmol/L, respectively. Compared with individuals who had low FPG, those with high FPG were significantly associated with higher in-hospital mortality in diabetic cohort (10.1% vs. 2.8%; odds ratio [OR] = 3.862, 95% confidence interval [CI]: 2.542-5.869) and nondiabetic cohort (7.4% vs. 1.7%; HR = 4.542, 95%CI: 3.041-6.782). After adjusting the potential confounders, this significant association was not changed. Furthermore, FPG as a continuous variable was positively associated with in-hospital mortality in single-variable and multivariable models regardless of diabetic status. Adding FPG to the original model showed a significant improvement in C-statistic and net reclassification in diabetic and nondiabetic cohorts. CONCLUSIONS: This large-scale registry indicated that there is a strong positive association between FPG and in-hospital mortality in AMI patients with and without diabetes. FPG might be useful to stratify patients with AMI.