Introduction: Mortality among patients with hematologic malignancies admitted to the intensive care unit (ICU) ranges from 19% to 85%, increasing to 83%–97% in those requiring invasive mechanical ventilation (IMV). However, commonly used physiological scoring systems such as the Sequential Organ Failure Assessment (SOFA) and the Acute Physiology and Chronic Health Evaluation II (APACHE II) were not developed or validated for hemato-oncologic patients, and many of their components—such as hemoglobin or platelet counts—may be chronically altered in leukemia due to the disease or its treatment. No validated mortality prediction score currently exists for adult patients with acute leukemias admitted to ICU.

Objectives: To identify prognostic factors associated with ICU mortality in adult patients with acute leukemias admitted to a tertiary care ICU. Secondary objectives included estimating ICU, 30-day, and 60-day mortality rates, and developing a tailored mortality risk score for this population.

Methods: We conducted a retrospective cohort study of adult patients (≥18 years) with acute lymphoblastic leukemia (ALL), or acute myeloid leukemia (AML) admitted to a single tertiary ICU in Mexico City from January 2015 to August 2023. Only patients with complete clinical records were included. Univariate and multivariate analyses were performed using binary logistic regression to identify factors associated with mortality. Mortality score performance was evaluated using the area under the receiver operating characteristic (ROC) curve. Model calibration was assessed using the Hosmer-Lemeshow goodness-of-fit test. Survival curves were estimated using the Kaplan-Meier method, and the log-rank test was used to compare survival between groups.

Results: A total of 143 ICU admissions were analyzed: 62.9% had ALL and 37.1% had AML; 67.8% were in induction therapy. The most frequent reasons for ICU admission were septic shock (73.9%) and acute respiratory distress syndrome (ARDS) (36.4%). ICU, 30-day, and 60-day mortality rates were 45.3%, 54.8%, and 57.9%, respectively.

On univariate analysis, significant predictors of mortality included: age >33 years, FiO₂ >35%, pH <7.35, INR >1.4, total bilirubin >1.7 mg/dL, AKI, albumin <3 g/dL, Glasgow Coma Scale <15, IMV, myocardial dysfunction, and ARDS as the cause of ICU admission.

Multivariate analysis identified three independent risk factors: IMV (OR 4.78; 95% CI 1.73–13.23, p=0.003), AKI (OR 4.56; 95% CI 1.94–10.72, p=0.001), and hypoalbuminemia (albumin <3 g/dL, OR 4.00; 95% CI 1.59–10.04, p=0.003). A 3-point mortality risk score was developed, assigning 1 point to each variable. This score demonstrated good discriminatory performance for ICU mortality (AUROC 0.796) and 30-day mortality (AUROC 0.729), outperforming SOFA (AUROC 0.678 and 0.685) and APACHE II (AUROC 0.652 and 0.642).

Thirty-day survival varied markedly by score: 80% (score 0), 76.5% (score 1), 41.1% (score 2), and 9.4% (score 3) (log-rank p<0.001). Based on these results, patients were stratified into three risk categories: low (0–1 points), intermediate (2 points), and high (3 points), with a negative predictive value of 0.898 in the low-risk group and a positive predictive value of 0.836 in the high-risk group.

Conclusions Invasive mechanical ventilation, acute kidney injury, and hypoalbuminemia were independently associated with ICU mortality in patients with acute leukemias. The proposed risk score outperformed SOFA and APACHE II in predicting ICU and 30-day mortality and may serve as a useful tool for risk stratification in this setting. While the model performed well in this ALL-enriched cohort, external validation is warranted across broader populations, to confirm its generalizability.

This content is only available as a PDF.
2025
Sign in via your Institution