Clinical characteristics and outcomes of patients with acute myelogenous leukemia admitted to intensive care: a case-control study.

BACKGROUND: There is limited epidemiologic data on patients with acute myelogenous (myeloid) leukemia (AML) requiring life-sustaining therapies in the intensive care unit (ICU). Our objectives were to describe the clinical characteristics and outcomes in critically ill AML patients. METHODS: This was a retrospective case-control study. Cases were defined as adult patients with a primary diagnosis of AML admitted to ICU at the University of Alberta Hospital between January 1st 2002 and June 30th 2008. Each case was matched by age, sex, and illness severity (ICU only) to two control groups: hospitalized AML controls, and non-AML ICU controls. Data were extracted on demographics, course of hospitalization, and clinical outcomes. RESULTS: In total, 45 AML patients with available data were admitted to ICU. Mean (SD) age was 54.8 (13.1) years and 28.9% were female. Primary diagnoses were sepsis (32.6%) and respiratory failure (37.3%). Mean (SD) APACHE II score was 30.3 (10.3), SOFA score 12.6 (4.0) with 62.2% receiving mechanical ventilation, 55.6% vasoactive therapy, and 26.7% renal replacement therapy. Crude in-hospital, 90-day and 1-year mortality was 44.4%, 51.1% and 71.1%, respectively. AML cases had significantly higher adjusted-hazards of death (HR 2.23; 95% CI, 1.38-3.60, p = 0.001) compared to both non-AML ICU controls (HR 1.69; 95% CI, 1.11-2.58, p = 0.02) and hospitalized AML controls (OR 1.0, reference variable). Factors associated with ICU mortality by univariate analysis included older age, AML subtype, higher baseline SOFA score, no change or an increase in early SOFA score, shock, vasoactive therapy and mechanical ventilation. Active chemotherapy in ICU was associated with lower mortality. CONCLUSIONS: AML patients may represent a minority of all critically ill admissions; however, are not uncommonly supported in ICU. These AML patients are characterized by high illness severity, multi-organ dysfunction, and high treatment intensity and have a higher risk of death when compared with matched hospitalized AML or non-AML ICU controls. The absence of early improvement in organ failure may be a useful predictor for mortality for AML patients admitted to ICU.