Infection is a significant cause of morbidity and mortality in children receiving treatment for AML. Pediatric cancer centers vary widely in use of antibacterial and antifungal prophylaxis. Here we report on the incidence of bacteremia, fungal infections and associated outcomes of 76 children diagnosed with AML at our institution over a 12 year period. Further, we analyze the outcomes of patients diagnosed after 2010 when new policies were implemented requiring cefepime prophylaxis and inpatient hospitalization during count recovery.
We retrospectively reviewed the medical charts of 76 children with AML treated at Children's Hospital of Alabama between January 2000 and March 2012. We examined in detail 101 episodes of bacteremia and 66 episodes of concern for fungal infection (defined as an episode of febrile illness where the patient received antifungal medication at treatment doses rather than only prophylaxis).
Fifty-two patients (68%) had one or more episodes of bacteremia from time of diagnosis to an endpoint of bone marrow transplant, death, or end of therapy. There were a total of 101 episodes of bacteremia (68 gram-positive, 26 gram-negative, and 7 mixed gram-positive and gram-negative). Strep Viridans was the most common organism isolated, accounting for 33% of positive bacterial cultures. Eighty-six (85%) positive cultures occurred while the patient was severely neutropenic (ANC<200). Only nine (9%) positive cultures occurred in patients recovered from severe neutropenia. There was a 33% risk of PICU stay per episode of bacteremia. There was a 10% (10/101) risk of death per episode of bacteremia, and 13% (10/76) of patients in our cohort died of bacterial infection. Twenty-seven patients (36%) had 2 or more hospitalizations with bacteremia, and these patients had a disproportionate fraction of infection related complications. They accounted for 88% (29/33) of PICU stays and 80% (8/10) of patient deaths resulting from bacteremia. Seven of 76 (9%) AML patients received cefepime prophylaxis and inpatient hospitalization while severely neutropenic (ANC <200). Fifty seven percent (4/7) of these patients had one bacteremic episode which was similar to the 69% (48/69) rate of bacteremia in patients not receiving prophylaxis (p=0.67). However, there were no PICU stays or deaths associated with positive blood cultures in patients receiving cefepime prophylaxis. Comparing patients with bacteremia on cefepime prophylaxis (n=4) to patients with bacteremia not on a prophylaxis regimen (n=48), there were significantly less PICU stays in the patients receiving cefepime prophylaxis (p=0.014).
All but one of the 76 patients in our cohort received antifungal prophylaxis (fluconazole 70%, voriconazole 11%, micafungin 5%, other/combination 14%). Fifty eight percent (44/76) of patients had at least one episode of clinical concern for fungal infection where they received an antifungal drug at treatment doses, and overall there were 66 episodes of fungal concern. Positive fungal cultures (from various sources including blood, lung biopsy, BAL, urine, trachea, skin, wound and stool cultures) were documented in 16% (12/76) of patients. Seven cultures were positive for Candida species (4 bloodstream, 1 wound, 1 stool, and 1 patient positive in both urine and trachea), 4 for Aspergillus (all lung biopsies), and 1 for Geotrichum (skin culture). Eleven episodes of positive fungal culture occurred during neutropenia with only the one skin culture occurring at ANC above 500. CT scans of the chest were used regularly to evaluate for fungal disease in 83% (55/66) of episodes of concern. Three (8%) chest CTs showed cavitary lesions highly suspicious for fungal disease, and 10 (27%) chest CTs had nodules. There was only one death attributed to fungal infection (Candida fungemia).
In conclusion, AML patients have a high risk of infection. Bacteremia occurred in 68% of our cohort and resulted in a 13% mortality rate. However, patients receiving cefepime prophylaxis had significant decrease in PICU stays associated with bacteremia and did not account for any infection related deaths. Nearly all patients in our cohort received antifungal prophylaxis, and while 16% of patients had culture proven fungal disease, only one death was attributed to fungal infection. These results emphasize the importance of supportive care practices that prevent infection and monitor patients during times of increased risk.
No relevant conflicts of interest to declare.
Asterisk with author names denotes non-ASH members.