Prognostic value of serum creatinine trajectories on ICU mortality in patients with acute myocardial infarction: a longitudinal retrospective cohort study

Background Acute kidney injury (AKI) commonly complicates acute myocardial infarction (AMI) and markedly increases mortality risk. However, the prognostic value of dynamic serum creatinine trajectories in AMI patients admitted to the intensive care unit (ICU) remains insufficiently characterized. Methods This retrospective cohort study used the Medical Information Mart for Intensive Care IV (MIMIC-IV, version 3.1) database. We included adults aged ≥ 18 years admitted to the ICU with AMI who had at least four serum creatinine measurements within 96 hours of admission. Latent growth mixture modeling (LGMM) identified distinct creatinine trajectory subgroups. The primary outcomes were all-cause mortality at 30 days and 365 days after ICU admission. Cox proportional hazards regression assessed the independent association between trajectory class and mortality, adjusting for demographics, comorbidities, disease severity scores, and therapeutic interventions. Results Among 1,153 patients meeting inclusion criteria, LGMM identified three distinct creatinine trajectories: Class 1 (stable-low, n = 394, 34.2%), Class 2 (moderate-ascending, n = 436, 37.8%), and Class 3 (high-rapid rising, n = 323, 28.0%). Model fit statistics demonstrated optimal discrimination, with average posterior probabilities > 0.77 for all classes. Kaplan-Meier analysis revealed statistically significant differences in survival curves among trajectory groups (P < 0.001). In fully adjusted Cox models, Class 3 demonstrated significantly increased mortality risk compared with Class 1 at both 30 days (HR = 2.00, 95%CI: 1.18–3.40, P = 0.010) and 365 days (HR = 1.73, 95%CI: 1.07–2.79, P = 0.026). Class 2 showed elevated risk after adjustment but did not achieve statistical significance. Conclusion Serum creatinine trajectories derived by LGMM offer independent prognostic information on both short-term and long-term mortality in ICU patients with AMI. Dynamic monitoring of these creatinine patterns may improve risk stratification beyond conventional single-timepoint assessments of renal function.