Bloodstream infections in paediatric cancer patients. Prospective comparative study in 2 university hospitals.

BACKGROUND Preventive approaches (including those related to care of long term central venous catheters, CVADs) and the incidence of bloodstream infections (BSI) in 2 German university affiliated paediatric oncology units. PATIENTS AND METHODS Non-interventional prospective observational study using the Oncoped surveillance module. Center A included 85 patients in 31 months and Center B 84 patients in 21 months. The populations did not differ in terms of age, gender, malignancy and disease status (first illness vs. relapse). Center A used ports (46 %) and 2 different Broviac catheters (54 %), in Center B nearly all patients with a CVAD had Broviacs (96 %). 30 BSI (24 patients) were diagnosed in Centre A and 28 BSI (22 patients) in Center B. Patients with relapsed malignancy experienced more BSI (51.4 % vs. 20.9 %; p = 0.001). Incidence rates were significantly lower in Center A (3.47 vs. 7.93 BSI/1000 CVAD days; p = 0.037). Poisson regression analysis revealed a significant lower incidence density (BSI/100 inpatient days) for all BSI in Center A (RR 0.47 CI95 0.27-0.81, p = 0.006). Overall, 52 % of all pathogens detected in blood cultures in Center A were Gram-positive (57 % in Center B) and 48 % Gram-negative (43 in Center B). One ALL patient without a CVAD died due to overwhelming sepsis caused by an ESBL-producing E. cloacae isolate. CONCLUSION Paediatric cancer treatment centers differ substantially in regard to management of CVADs and in other preventive strategies. The most important use of local surveillance data is longitudinal internal assessment in close cooperation with microbiology and hospital hygiene experts.