Early intravenous corticosteroid and 28-day mortality in critically ill adults with an ICD-coded aspiration syndrome: a retrospective MIMIC-IV cohort study

Systemic corticosteroid therapy is not established for aspiration-related respiratory illness. Because ICD-9 code 507.0 and ICD-10 code J69.0 capture a heterogeneous syndrome that may include chemical pneumonitis, bacterial pneumonia, or overlapping processes, the association between early corticosteroid treatment and clinical outcomes remains uncertain. We analyzed linked Medical Information Mart for Intensive Care IV (MIMIC-IV) version 3.1 and MIMIC-IV-Note version 2.2 data. Adults with International Classification of Diseases, Ninth Revision (ICD-9) code 507.0 or Tenth Revision (ICD-10) code J69.0 were eligible. The operational index time (t0) was the earliest high-confidence radiology-note timestamp when available; otherwise, admission time was used with discharge-note support or ICD evidence alone. Only the first eligible hospitalization and first intensive care unit (ICU) stay were retained. The primary analysis was conditional on survival and observability through t0 plus 24 h. The exposure was a prescription-recorded intravenous systemic corticosteroid order starting within this 24-hour window. Follow-up began at the 24-hour landmark. The primary outcome was death through day 28. The Cox model adjusted for demographics, Acute Physiology Score III (APS III), major comorbidities, immunosuppression history, and recorded pre-t0 oral or enteral systemic corticosteroid exposure. Among 5,270 patients, 413 (7.8%) had an early intravenous corticosteroid order and 1,426 died by day 28. Early intravenous corticosteroid ordering was associated with a higher adjusted mortality hazard in the primary landmark model (hazard ratio [HR], 1.40; 95% confidence interval [CI], 1.18–1.66; p < 0.001). The landmark propensity-score-matched estimate was HR 1.45 (95% CI, 1.15–1.84). The adjusted HR through day 90 was 1.29 (95% CI, 1.11–1.49). Early intravenous corticosteroid ordering was also associated with fewer index-ICU-free days (adjusted mean difference, −1.38; 95% CI, −2.45 to −0.31) and hospital-free days ( − 1.20; 95% CI, −2.06 to −0.33). In this ICD-coded aspiration-syndrome cohort, early intravenous corticosteroid ordering was associated with higher short-term mortality in the primary landmark analysis. Material attenuation in a stricter index-time restriction and the possibility of phenotype, timing, and indication bias limit causal interpretation. These findings are hypothesis-generating and do not establish harm from corticosteroids. Not applicable.