#53: High Burden of Serious Bacterial Infections in African Children Treated for Cancer

Background Infectious complications in children treated for cancer contribute to their morbidity and mortality. There is a paucity of studies on the incidence, microbiological etiology, risk factors, and outcome of serious bacterial infections in African children treated for cancer. Aim The aim of the study was to delineate the epidemiology of infectious morbidity and mortality in South African children with cancer. Methods This prospective, single-center, longitudinal-cohort study enrolled children one-19 years old hospitalized for cancer treatment at the Paediatric Oncology Unit, Chris Hani Baragwanath Academic Hospital, Soweto, South Africa. Children were investigated for infection as part of the standard of care. Results In total, 169 children were enrolled, 82 with hematological malignancy (HM), 87 with a solid tumor (ST), median age was 68.5 months and 10.7% were living with HIV. The incidence (per 100 child-years) of septic episodes (SE) and microbiologically confirmed SE (MSCE) was 101 (138 vs. 70, P < 0.001) and 70.9 (99.1 vs. 47.3; P < 0.001), respectively; higher in children with HM than ST. The incidence of MCSE in children with high-risk HM (137.7) was 4.32-fold greater compared with those with medium-risk HM (30.3; P < 0.001). Children with metastatic ST had a higher incidence (84.4) of MSCE than those with localized ST (33.6; aOR: 2.52; P < 0.001). The presence of an indwelling catheter was 3-fold (P < 0.001) more likely to be associated with MCSE compared with those without. There was no association for age group, nutritional status or HIV-status, and incidence of MCSE. The incidence of gram-positive (GPB) and gram-negative (GNB) SEs was 48.5 and 37.6, respectively, and higher in children with an HM. The most commonly identified GPB were Coagulase-negative Staphylococci, Streptococcus viridans and Enterococcus faecium; while the most common GNB were Escherichia coli, Acinetobacter baumannii, and Pseudomonas species. The median CRP was higher in children with MSCE compared with those with culture-negative SE (CNSE) (116.5 vs. 92; P < 0.001) in both HM (132.5 vs. 117; P < 0.001) and ST (87.5 vs. 46; P < 0.001). The procalcitonin was higher in those with MSCE compared with those with CNSE (2.30 vs. 1.40; P < 0.001) in both HM (2.95 vs. 1.60; P = 0.002) and ST (2.10 vs. 1.20; P < 0.001). The case fatality risk was 40.4%; 80% was attributed to sepsis. Of these, 35 (72.92%) had HM and 34 of the 35 (97.14%) had HR-HM. Children with HM had an overall sepsis CFR of 42.68%. Four (30.77%) of the 13 sepsis-related deaths in STs had metastatic disease and 8 (16.67%) of the total number of sepsis-related deaths were in children living with HIV. There was no association between malnutrition or HIV-positivity and death. The odds of dying from sepsis were higher in children with profound (aOR 3.96; P = 0.004) and prolonged (aOR 3.71; P = 0.011) neutropenia. Pneumonia (58.85% vs. 29.23%; aOR 2.38; P = 0.025) and tuberculosis (70.83% vs. 34.91%; aOR 4.3; P = 0.005) were independently associated with a higher CFR. Conclusion The current study emphasizes the high burden of sepsis in African children treated for cancer, and especially HM, and highlights the association of tuberculosis and pneumonia as independent predictors of death in children with cancer.