Summary: | Abstract Background The prognostic value of in-hospital hemoglobin drop in non-overt bleeding patients with acute myocardial infarction (AMI) admitted to the intensive care unit (ICU) remains insufficiently investigated. Methods A retrospective analysis was performed based on the Medical Information Mart for Intensive Care (MIMIC)-IV database. 2,334 ICU-admitted non-overt bleeders diagnosed with AMI were included. In-hospital hemoglobin values (baseline value on admission and nadir value during hospitalization) were available. Hemoglobin drop was defined as a positive difference between admission and in-hospital nadir hemoglobin. The primary endpoint was 180-day all-cause mortality. The time-dependent Cox proportional hazard models were structured to analyze the connection between hemoglobin drop and mortality. Results 2,063 patients (88.39%) experienced hemoglobin drop during hospitalization. We categorized patients based on the degree of hemoglobin drop: no hemoglobin drop (n = 271), minimal hemoglobin drop (< 3 g/dl; n = 1661), minor hemoglobin drop (≥ 3 g/dl & < 5 g/dl, n = 284) and major hemoglobin drop (≥ 5 g/dl; n = 118). Minor (adjusted hazard ratio [HR] = 12.68; 95% confidence interval [CI]: 5.13–31.33; P < 0.001) and major (adjusted HR = 13.87; 95% CI: 4.50-42.76; P < 0.001) hemoglobin drops were independently associated with increased 180-day mortality. After adjusting the baseline hemoglobin level, a robust nonlinear relationship was observed in the association between hemoglobin drop and 180-day mortality, with 1.34 g/dl as the lowest value (HR = 1.04; 95% CI: 1.00-1.08). Conclusion In non-overt bleeding ICU-admitted patients with AMI, in-hospital hemoglobin drop is independently associated with higher 180-day all-cause mortality.
|