Infection-associated transitions to critical care–level interventions in adult cancer hospitalizations: A National Inpatient Sample analysis.

Abstract

e23214 Background: Infectious diagnoses and complications during cancer hospitalizations are commonly evaluated using static outcomes such as mortality or organ-specific failure. Whether infection precipitates discrete shifts from routine inpatient management to critical care–level interventions remains poorly defined. This study examined the infection-associated transition to intensive care in hospitalized adults with cancer. Methods: A serial cross-sectional analysis of adult hospitalizations with a principal diagnosis of malignancy in the 2018–2022 Healthcare Cost and Utilization Project National Inpatient Sample was conducted. We identified infection using validated any-diagnosis ICD-10-CM phenotypes for sepsis, infection-associated shock, or infection-associated acute organ dysfunction. The primary outcome was a composite escalation-anchored transition to critical care defined by initiation of mechanical ventilation, dialysis-requiring acute kidney injury, or shock during hospitalization. Secondary outcomes included in-hospital mortality, length of stay, and hospitalization cost derived using cost-to-charge ratios. National estimates were generated using survey-weighted analyses accounting for discharge weights, hospital-level clustering, and stratification, with multivariable adjustment for demographics, cancer lineage, APR-DRG severity, and hospital characteristics. Results: Among an estimated 4.81 million cancer hospitalizations nationally, 12.9% were complicated by infection. Critical care transitions occurred in 25.0% of infected admissions compared with 0.7% of non-infected hospitalizations. Infected admissions demonstrated substantially higher rates of mechanical ventilation (17.0% vs 0.5%), dialysis-requiring acute kidney injury (2.9% vs 0.3%), and shock (10.8% vs 0%). In-hospital mortality increased markedly following escalation, from 3.2% among admissions without transition to 34.4% among those requiring critical care–level intervention. Transitions to critical care were associated with significantly longer length of stay (16.5 vs 6.4 days) and higher mean hospitalization cost ($84,849 vs $27,964). After multivariable adjustment, infection remained independently associated with transition to critical care (adjusted OR 12.51; 95% CI 11.48–13.63). Conclusions: Infection in hospitalized adults with cancer is associated with discrete threshold crossings to critical care–level interventions rather than functioning solely as a static diagnostic category. Framing infection as a driver of inpatient escalation highlights opportunities for earlier risk identification, triage prioritization, and proactive escalation planning in high-risk oncology admissions.