Microbiology and Mortality of Pediatric Febrile Neutropenia in El Salvador Sumit Gupta, MD,* Miguel Bonilla, MD,w Mario Gamero, MD,w Soad L. Fuentes, MD,w Miguela Caniza, MD,zy and Lillian Sung, MD, PhD* Background: Febrile neutropenia (FN) and infection-related mortality are major problems for children with cancer in low- income countries. Identifying predictors for adverse outcome of FN in low-income countries permits targeted interventions. We describe the nature and predictors of microbiologically documen- ted infection (MDI) and mortality of FN in children with cancer in El Salvador. Methods: We examined Salvadoran pediatric oncology patients admitted with FN over a 1-year period. Data were collected pros- pectively. Demographic, treatment, and admission-related vari- ables were examined as predictors of outcomes. Results: Hundred six FN episodes among 85 patients were in- cluded. Twenty-three of 106 episodes (22%) were microbiologically documented; 13 of106 episodes (12%) resulted in death. Gram- positive and gram-negative organisms were isolated in 14 of 23 and 11 of 23 specimens; polymicrobial infections were common (11 of 23 episodes of MDI). Older age decreased the MDI risk [odds ratio (OR) per year=0.87, 95% confidence interval (CI), 0.75-0.99; P=0.04] while increasing number of days since the last chemother- apy increased the risk (OR=1.03 per day, 95% CI, 1.01-1.04; P=0.002). Pneumonia diagnosed either clinically (OR=6.6, 95% CI, 1.8-30.0; P=0.005) or radiographically (OR=5.5, 95% CI, 1.7-18.1; P=0.005) was the only predictor of mortality. Conclusions: In El Salvador, polymicrobial infections were common. Pneumonia at admission identified children with FN at high risk of death; these children may benefit from targeted interventions. Key Words: cancer, developing countries, fever, neutropenia, pneumonia (J Pediatr Hematol Oncol 2011;33:276–280) C ure rates of childhood cancer have steadily increased over the past half century in high-income countries (HICs), reaching overall cure rates of more than 80% across all diagnoses. 1 Children with cancer who live in low-income countries (LICs), however, experience cure rates significantly below those evident in HICs. This survival gap between LICs and HICs may have multiple underlying fac- tors, including delayed presentation and diagnosis leading to advanced stage, different underlying biology, and higher rates of treatment toxicity. 2 Infection-related mortality in particular is an important contributor to both treatment- related mortality and overall mortality. 3 In HICs, the mortality rate during episodes of febrile neutropenia (FN) in children varies between <1% to 3%. 4–7 Recently, there has been significant interest in identifying groups of children with FN who are at low risk of developing serious complications. 4,5,8–10 Most of this interest has come from HICs, with the goal of reducing the intensity of care for low-risk patients through strategies such as oral antibiotics, early discharge, and outpatient management. Only a few studies have examined risk stratification of FN in LICs, and found contradictory results. For example, although magnitude of fever and monocyte count were predictive of adverse outcomes in some studies, this finding was not universal. 11–13 In addition, these studies have been conducted in more developed LICs such as Chile and Turkey. Further research is required to better understand the epidemiology and microbiology of FN in a broad range of LICs. A deeper understanding of the characteristics and predictors of adverse outcomes may allow earlier identifica- tion and management of children at highest risk. This is of particular importance in resource-constrained settings. The objectives of this study were, therefore, to describe micro- biology and outcomes of FN among children with cancer in El Salvador, an LIC, and to determine predictors of adverse outcomes. MATERIALS AND METHODS Study Population and Data Source The study sample included all episodes of FN among pediatric oncology patients aged 1 to 16 years of age who were admitted to the Benjamin Bloom National Children’s Hospital, San Salvador, El Salvador, between December 25, 2000 and December 31, 2001. This hospital is the only pediatric oncology referral center in the country, allowing for population-based analyses. No episodes were excluded. Fever was defined as an oral temperature Z38.01C or equivalent documented either at home or on presentation. Neutropenia was defined as an absolute neutrophil count (ANC) of <500  10 9 /L or an ANC Z500  10 9 /L with either the child looking clinically unwell or the ANC expected to drop further in the next few days. Copyright r 2011 by Lippincott Williams & Wilkins Received for publication July 14, 2010; accepted January 12, 2011. From the *Division of Haematology/Oncology and Program in Child Health Evaluative Sciences, The Hospital for Sick Children, University Avenue, Toronto, Ontario, Canada; wPediatric Onco- logy, Benjamin Bloom National Children’s Hospital, San Salvador, El Salvador; zDepartment of Infectious Diseases; and yInter- national Outreach Program, St Jude Children’s Research Hospital, Memphis, TN. Lillian Sung was supported by a New Investigator Award from the Canadian Institutes of Health Research. Conflict of Interest: None. Reprints: Lillian Sung, MD, PhD, Division of Haematology/Oncology, The Hospital for Sick Children, 555 University Avenue, Toronto, Ontario M5G 1 8 (e-mail: Lillian.sung@sickkids.ca). ORIGINAL ARTICLE 276 | www.jpho-online.com J Pediatr Hematol Oncol  Volume 33, Number 4, May 2011