Study objectives: The survival rate of patients with a hematologic malignancy requiring mechanical ventilation (MV) in the ICU has improved over the last few decades. The objective of this study was to identify the factors affecting the in-hospital mortality of these particular patients, and to assess whether the use of noninvasive positive pressure ventilation (NPPV) was protective in our study population.
Design: We retrospectively collected variables in 166 consecutive patients with hematologic malignancies who had acute respiratory failure (ARF) requiring MV, and identified factors obtained within 24 h of ICU admission affecting in-hospital mortality in univariate and multivariate stepwise logistic regression analyses. The effect of NPPV on mortality was assessed using a pair-wise matched exposed-unexposed analysis.
Results: The mean simplified acute physiology score (SAPS) II was 58.9. The in-hospital mortality rate was 71%. In a multivariate logistic regression analysis, the in-hospital mortality rate was predicted by increasing severity of illness, as measured by SAPS II (odds ratio [OR] per point of increase, 1.07; 95% confidence interval [CI], 1.04 to 1.11) and a diagnosis of acute myelogenous leukemia (OR, 2.73; 95% CI, 1.05 to 7.11). Female sex (OR, 0.36; 95% CI, 0.16 to 0.82), endotracheal intubation (ETI) within 24 h of ICU admission (OR, 0.29; 95% CI, 0.11 to 0.78), and recent bacteremia (defined as blood cultures positive for bacteria < 48h before or < 24h after ICU admission) [OR, 0.22; 95% CI, 0.08 to 0.61] were associated with a lower mortality rate. Twenty-seven patients who received NPPV were matched for SAPS II (+/- 3) with 52 patients who required immediate ETI on a 1:2 basis. The crude in-hospital mortality rate was 65.4% in both groups.
Conclusion: Although the in-hospital mortality rate in hematologic patients who develop ARF remains high, the reluctance to intubate and start treatment with invasive MV in this population is unjustified, especially when bacteremia has precipitated ICU admission.