Mortality and Resource Utilization in Children with De Novo Acute Myeloid Leukemia Treated with Chemotherapy and Gemtuzumab Ozogamicin in the United States

The role of Gemtuzumab ozogamicin (GO) for acute myeloid leukemia (AML) remains controversial. GO was removed from the U.S. market in 2010 due to concerns of increased induction mortality in adults. Other studies have shown a survival benefit without increased treatment related mortality. Moreover, no data are available on the resources required to deliver GO based chemotherapy. Since pediatric data are limited, we evaluated in-hospital mortality and resource utilization in pediatric AML patients treated with GO and standard chemotherapy.

We used the Pediatric Information Health System (PHIS) to establish a cohort of children < 19 years old treated for de novo AML with GO and standard cytarabine, daunorubicin, and etoposide (ADE) induction. Cohort assembly was validated by local chart review and used ICD-9 diagnosis codes and manual review of chemotherapy. Case fatality was determined after induction (defined from the start of therapy to the initiation course 3), at 6 months and at 12 months. Resource utilization was determined for each patient based on daily billing data. Each resource variable was dichotomized (exposure or no exposure) for each inpatient day and then summarized during each study period to determine resource utilization days per 1,000 hospital days.

In total, 253 children who had billing data for GO during the first course of ADE induction were identified. Median age was 9.6 years; a slight male predominance was observed (54%) and most patients were white (69%). In-hospital case-fatality rates were 2.4% during induction, 6.7% at 6 months, and 13.0% at 12 months from start of therapy. PHIS billing data demonstrated that patients received opioids almost on one in four hospital days, that during induction period 12% of patients received vasopressors on at least two consecutive days, and 12% needed assisted ventilation. Mean inpatient stay and resource utilization rates are presented in Table 1.

In-hospital mortality rates at the three time points were low and concordant with published data on pediatric AML trials using an ADE induction (Gibson, BJH 2011) and ADE Induction + GO (Cooper, Cancer 2012) and lower than trials using intensively timed DCTER regimens (Woods, Blood 2001; Lange, Blood 2008). Resource utilization data demonstrated an extensive use of resources needed to manage infections (blood cultures, imaging, antimicrobials). While infections are the leading cause of non-relapse morbidity and mortality in pediatric AML, such extensive use of resources has not been previously quantified. In addition, PHIS billing data describe toxicities such as pain (opioid use), hypotension (vasopressor support), and respiratory failure (assisted ventilation) at rates higher than those previously reported in clinical trials. In conclusion, the in-hospital mortality of children treated with GO at PHIS centers appears comparable to previously published studies of ADE and ADE + GO. The resource utilization data provide a more comprehensive description of resources needed to treat pediatric AML than previously reported. In addition, the resource utilization data suggest that toxicities reported on clinical trials may underestimate the resources needed to administer AML induction therapy safely.

No relevant conflicts of interest to declare.