Changes in Intensive Care Unit Admission Rates, Organ Support, and Mortality in Patients with Acute Myeloid Leukemia: A Danish Nationwide Cohort Study

Abstract

Introduction The mortality in non-promyelocytic acute myeloid leukemia (AML) has decreased as a result of improved supportive care, allowing for more aggressive treatment options, including increased use of allogeneic stem cell transplantation. To investigate effects on ICU admission and mortality, we evaluated trends in ICU admission rates, use of organ support, 1-year mortality overall and after ICU admission in all Danish AML patients. Methods The study included all adult Danish AML patients registered in the Danish Acute Leukemia Registry diagnosed between 2005 and 2016 who were treated with a curative intent. Information on first ICU admission was obtained from the Danish Intensive Care Database. We estimated the rate of ICU admitted AML patients within 3 years of diagnosis and examined the proportion of ICU patients receiving organ supportive treatments including invasive and non-invasive mechanical ventilation, dialysis, and vasopressors. The trends were examined using linear regression analysis and presented as predicted yearly percent point difference (pp/year) with 95% confidence interval (95%CI). Results A total of 1378 patients treated with remission-induction chemotherapy were included. On average, 27.9% of patients were admitted to an ICU within 3 years of diagnosis. The proportion of ICU admitted patients did not change over time (pp/year: 0.1, 95%CI: -0.5 to 0.8), nor did the available Simplified Acute Physiology Score (SAPS) II, which is used to describe disease severity and predict mortality risk in ICUs (predicted yearly SAPS II score difference: -0.47, 95%CI: -1.80 to 0.86). Over the 11 year study period, the use of mechanical ventilation dropped from 66.7% to 40.6% (pp/year: -2.0, 95%CI: -3.5 to -0.5), with a parallel increase in the use of non-invasive ventilation from 20.0% to 50.0% (pp/year: 2.6, 95%CI: 1.1 to 4.0). The use of dialysis decreased within the same time period (pp/year: -1.7, 95%CI: -3.0 to -0.4), whereas the use of vasopressors did not change significantly (decreased from 60.0% to 56.3%, pp/year: -1.2, 95%CI: -2.8 to 0.3). Within the study period, the overall 1-year mortality decreased from 36.0% in 2005 to 28.8% in 2016 (pp/year: -1.0, 95%CI: -1.7 to -0.25). Similarly, in the subgroup of ICU patients (n=370), the mortality rate decreased from 80.0% to 65.6% (pp/year: -1.4, 95%CI: -2.8 to -0.07). Conclusion In a Danish nationwide AML study, the overall mortality, as well as in the subset of patients admitted to the ICU have decreased substantially since 2005. This is in contrast to a recent study (Halpern et al, JAMA, 2017), which failed to demonstrate survival improvements in American ICU-admitted AML patients from 2004 to 2012. The reduced mortality in Danish AML patients following ICU admission could not be explained by a change in ICU admission rate or disease severity, as measured by SAPS II. However, whether the observed changes in ICU treatment patterns contribute to the improvement in survival (e.g. increasing use of non-invasive ventilation over mechanical ventilation) or reflect a change in treatment strategies versus a change in treatment needs, remain to be further investigated. Disclosures No relevant conflicts of interest to declare.