Association between the stress hyperglycemia ratio and 1-year all-cause mortality in patients with secondary pulmonary embolism: a retrospective cohort study using the MIMIC-IV database

Background The stress hyperglycemia ratio (SHR) reflects relative hyperglycemia during acute physiological stress. Its prognostic value in pulmonary embolism (PE), particularly in patients with PE recorded as a nonprincipal diagnosis during a hospitalization involving intensive care unit admi.ssion (operationally defined as secondary PE), remains unclear. Using the MIMIC-IV database, we investigated the association between SHR and 1-year all-cause mortality in patients with secondary PE. Methods This retrospective cohort study included patients with PE from MIMIC-IV version 2.2. Patients with PE recorded as the principal diagnosis for the hospitalization were excluded; therefore, the cohort comprised patients with operationally defined secondary PE. Patients were categorized into SHR quartiles. The primary outcome was 1-year all-cause mortality. Associations were assessed using Cox proportional hazards models and restricted cubic spline analyses. Covariates were selected a priori based on prior evidence, clinical relevance, and biological plausibility. Multicollinearity was assessed using variance inflation factors. Subgroup analyses were exploratory, with Bonferroni-corrected interaction P values reported. Results A total of 784 patients were included; 49.49% were male. In-hospital and 1-year mortality rates were 4.21% and 20.41%, respectively. In the fully adjusted Cox model, the highest SHR quartile was associated with higher 1-year mortality than the lowest quartile (hazard ratio [HR] = 1.49, 95% confidence interval [CI]: 1.01–2.24, P = 0.049), with a significant trend ( P = 0.045). Restricted cubic spline analyses showed a visually J-shaped trend in the association between SHR and 1-year all-cause mortality; however, the tests for non-linearity were not statistically significant (P for non-linearity = 0.623). The apparent curve pattern should therefore be interpreted cautiously. An exploratory interaction with insurance type was observed before but not after Bonferroni correction. Conclusions Higher SHR was associated with increased 1-year all-cause mortality in patients with secondary PE, suggesting potential prognostic value for longer-term risk assessment. Although the restricted cubic spline curve showed a visually J-shaped trend, the non-linearity test was not statistically significant; therefore, this apparent pattern requires confirmation in future studies. Clinical trial number Not applicable.