Prognostic Factors for Mortality in ICU Patients With Healthcare-Associated Infections: A Five-Year Institutional Analysis

TL;DR

This study identifies age, cancer, and pneumonia as key factors increasing death risk in ICU patients with healthcare infections, emphasizing the need for better infection control in cancer ICUs.

Contribution

The study identifies independent prognostic factors for ICU mortality in oncology patients with healthcare-associated infections using a five-year institutional analysis.

Findings

ICU mortality was 81.1% among patients with healthcare-associated infections.

Pneumonia was the most common infection type and independently associated with higher mortality.

Older age and presence of malignancy were significant predictors of ICU mortality.

Abstract

Background

Healthcare-associated infections (HAIs) represent a major challenge in intensive care units (ICUs), leading to increased morbidity, prolonged hospitalization, and high mortality. Critically ill oncology patients are particularly vulnerable because of immunosuppression and frequent exposure to invasive devices.

Objective

This study aims to describe the epidemiology and microbiological profile of HAIs among oncology ICU patients over a five-year period and to identify independent risk factors for ICU mortality using multivariable analysis.

Methods

We conducted a single-center retrospective cohort study in the anesthesiology ICU of a tertiary oncology hospital between January 1, 2020, and December 31, 2024. Among 1,766 ICU admissions, 259 adult patients who developed at least one HAI (343 episodes in total) and met the inclusion criteria were analyzed. Data on demographics, comorbidities, infection characteristics, microbiology, and outcomes were retrieved from the hospital infection control surveillance system and medical records. Comparative analyses between survivors and non‑survivors were performed, and a parsimonious multivariable logistic regression model was constructed to identify independent predictors of ICU mortality. Model performance was assessed by the area under the receiver operating characteristic (ROC) curve (AUC) and the Hosmer-Lemeshow test.

Results

The median ICU length of stay (LOS) was 29 days (interquartile range (IQR): 17-49), and ICU mortality was 81.1% (n = 210). Pneumonia (including ventilator‑associated and hospital‑acquired pneumonia) was the most common HAI (45.1%, n = 117), followed by central line‑associated bloodstream infection (CLABSI) (35.1%, n = 91) and catheter‑associated urinary tract infection (CAUTI) (11.6%, n = 30). Klebsiella spp. (37.6%, n = 129) and Acinetobacter baumannii (23.9%, n = 82) were the predominant pathogens. In the final multivariable model, higher age (adjusted odds ratio (aOR): 1.03 per year, 95% confidence interval (CI): 1.01-1.05, p = 0.014), presence of malignancy (aOR: 3.01, 95% CI: 1.49-6.10, p = 0.002), and pneumonia compared with other infection sites (aOR: 2.89, 95% CI: 1.41-5.94, p = 0.004) were independently associated with increased ICU mortality. The model showed acceptable discrimination (AUC: 0.741) and good calibration (Hosmer-Lemeshow, p = 0.363). In a sensitivity analysis restricted to ventilator‑associated pneumonia cases only, pneumonia remained independently associated with ICU mortality (aOR: 2.71, p = 0.007).

Conclusion

Among ICU patients with HAIs, mortality was very high. Older age, underlying malignancy, and pneumonia were the main independent prognostic factors associated with ICU mortality, particularly in the context of prolonged device exposure and highly virulent Gram‑negative pathogens. These findings highlight the need for intensified, site‑specific infection prevention strategies and strengthened antimicrobial stewardship in oncology ICUs, while recognizing that residual confounding related to illness severity and antimicrobial resistance patterns cannot be fully excluded.

Figures