Objective: To describe survival to intensive care unit (ICU) discharge and 6-month survival in a large cohort of pediatric oncology patients with severe sepsis.
Design: Retrospective analysis.
Setting: The ICU of a single pediatric oncology center.
Patients: Patients with cancer admitted to the ICU of St. Jude Children's Research Hospital between January 1, 1990, and December 31, 2002, who met the following criteria: 1) severe sepsis by ACCP/SCCM (American College of Chest Physicians/Society of Critical Care Medicine) Consensus Conference criteria and 2) receipt of fluid boluses of > or =30 mL/kg to correct hypoperfusion or receipt of a dopamine infusion of >5 microg.kg.min for inotropic support.
Interventions: None.
Measurements and main results: Data evaluated were demographic variables, oncologic diagnosis and time from diagnosis to ICU admission, Pediatric Risk of Mortality III score and absolute neutrophil count at admission, use of inotropes or pressors, use of mechanical ventilation, maximum organ system failure score, blood culture results, survival to ICU discharge, and 6-month survival. We identified 446 ICU admissions of 359 eligible patients. Overall ICU mortality was 76 of 446 (17%): 40 of 132 (30%) in post-bone marrow transplant (BMT) admissions and 36 of 314 (12%) in non-BMT admissions (p < .0001). In the 106 admissions requiring both mechanical ventilation and inotropic support, ICU mortality was 68 of 106 (64%). Regarding individual patients, 6-month survival was 170 of 248 (69%) among non-BMT patients vs. 43 of 111 (39%) for BMT patients (p < .001). When the 38 patients who survived to ICU discharge after requiring both mechanical ventilation and inotropic/vasopressor support are considered, 27 (71%) were alive 6 months after ICU discharge (22 of 27 [81%] non-BMT vs. 5 of 27 BMT [19%; p < .001]). ICU mortality varied by causative pathogen, from 63% for fungal sepsis (12 of 19) to 9% (5 of 53) for Gram-negative sepsis. Logistic regression analysis of factors significantly associated with ICU mortality in admissions requiring both mechanical ventilation and inotropic support identified four variables: BMT (odds ratio, 2.9; 95% confidence interval, 1.1-7.4; p = .03); fungal sepsis (odds ratio, 10.7; 95% confidence interval, 1.2-94.4; p = .03); use of multiple inotropes (odds ratio, 4.1; 95% confidence interval, 1.4-11.8; p = .01); and Pediatric Risk of Mortality III score (odds ratio, 1.1; 95% confidence interval, 1.0-1.2; p = .04).
Conclusions: In a large series of pediatric oncology patients with severe sepsis, ICU mortality was only 17% overall, although mortality remained quite high in the higher acuity patients. Mortality among the higher acuity patients was significantly associated with only a small number of variables. The number of patients alive at 6 months and the encouraging ICU survival rate further justifies the use of aggressive ICU interventions in this population.