Abstract
Acute myocardial infarction (AMI) substantially fuels the worldwide escalation in both morbidity and mortality. The hemoglobin glycation index (HGI) is linked to a range of undesirable outcomes, but its relationship with short-term outcomes in AMI patients has not been explored. This study analyzed data from 1008 first-time ICU AMI patients in the MIMIC-IV 3.1 database. To calculate the HGI, a linear regression equation was developed based on fasting glucose (FPG) and glycosylated hemoglobin (HbA1c), and patients were classified into four quartile groups. The main outcome of interest was 28-day ICU mortality, with the secondary outcome being 28-day in-hospital mortality. Kaplan-Meier survival analysis revealed that the Q1 group (low HGI) exhibited significantly higher mortality rates compared to the other groups. In a well-adjusted Cox proportional hazards model, low HGI was drastically linked with 28-day ICU mortality and 28-day in-hospital mortality. Restricted cubic spline (RCS) analysis revealed a U-shaped association between HGI and outcome events, mainly characterized by a correlation between low HGI and poor outcomes. Subgroup studies revealed that the association between HGI and endpoints was constant across subgroups. Machine learning models, including Boruta and SHAP, confirmed HGI's predictive value for short-term adverse outcomes. This shows that HGI could be a useful indicator of short-term mortality in AMI patients.