Despite the use of antibiotic prophylaxis, bacterial infections remain the leading cause of treatment-related morbidity in children with acute myeloid leukemia (AML). The objectives of this study were to: 1) evaluate the cumulative incidence of blood stream infection (BSI) during treatment according to the last three consecutive protocols of the Dutch Childhood Oncology Group (DCOG); 2) determine the most common bacterial pathogens; and 3) determine risk factors.


A retrospective chart review was performed on 246 patients (DCOG treatment protocols ANLL-97/AML-12 n=119, AML-15 n=60, DB-AML01 n=67) diagnosed with de novo AML (APL and Down syndrome excluded) between 1998-2013. All patients received prophylaxis with co-trimoxazole. From AML-15 onwards, patients also received prophylactic fluoroquinolones. BSIs were defined as a positive blood culture confirmed by the hospital's microbiologist. The cumulative incidence of BSIs was determined using competing risk analysis with death, switch to another protocol, or stem cell transplantation as competing events. Age, sex, BMI, white blood cell count (WBC) at diagnosis, cytogenetics, and treatment protocol were evaluated as risk factors.


Median age was 6.4 years [range 17.4], 58% were male. The cumulative incidence of BSIs was 78.2%. The most common pathogens were Staphylococcus epidermidis (15.6%), Streptococcus mitis (14.9%) and other Staphylococci (11.6%). ANLL-97/AML-12 (78%) and AML-15 (72%) showed a lower BSI incidence than DB-AML01 (84%; hazard ratio(HR)s 0.69 [0.50-0.97], p=0.033 and 0.59 [0.40-0.89], p=0.012, respectively). WBC >20x109/l was associated with a lower risk of infections (HR 0.72 [0.54-0.96], p=0.026). The other risk factors were not statistically significant.


The most recent DCOG protocol was associated with the highest incidence of BSIs, probably as a result of the high intensity of this protocol. Furthermore, the use of prophylactic antibiotics with gram-negative coverage may have resulted in a relatively higher incidence in gram-positive BSI. The high incidence stresses the urge to improve anti-infective supportive care measures.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.