Duration of antibiotic therapy for bacteraemia in immune compromised hosts—a post-hoc subgroup analysis of the Bacteremia Antibiotic Length Actually Needed For Clinical Effectiveness (BALANCE) randomized clinical trial

OBJECTIVES: The Bacteremia Antibiotic Length Actually Needed For Clinical Effectiveness (BALANCE) randomized clinical trial demonstrated 7 days to be noninferior to 14 days of antibiotics for bacteraemia. Uncertainty remains as to whether shortened treatment durations are sufficient for immune compromised patients. METHODS: /L) and transplantation but allowed enrolment of other forms of immune compromise; this post-hoc subgroup analysis focused on these immune compromised patients. The primary outcome was 90-day mortality. The primary analysis was in the intention-to-treat population with sensitivity analyses among the modified intention-to-treat population limited to those who survived ≥7 days. RESULTS: Among BALANCE patients, 1037 (28.7%) were considered immune compromised; 530 were randomized to receive 7 days of antibiotics and 507 to receive 14 days. The most common types of immune compromise included solid malignancies (75.4%), leukaemia/lymphoma (9.7%), chemotherapy (15.4%), corticosteroid use (4.7%), and HIV/AIDS (2.3%). Despite being formally excluded, a small number of neutropenic patients were enrolled (2.5%). Patients received a median (interquartile range) of 8 (7-11 days) versus 14 (14-15 days) days of antibiotics in the two arms. All-cause 90-day mortality in the intention-to-treat population was 106/528 (20.1%) in the 7 days arm and 120/507 (23.7%) in the 14 days arm (risk difference, -3.6%; 95% CI, -8.6-1.4; p = 0.16). In the modified intention-to-treat analysis, 90-day mortality was 104/526 (19.8%) in the 7 days arm and 117/504 (23.2%) in the 14 days arm (risk difference, -3.4%; 95% CI -8.5-1.6, p = 0.18). CONCLUSIONS: This post-hoc analysis of the immune compromised subgroup of the BALANCE randomized controlled trial suggests that the shorter treatment duration was sufficient for bacteraemia in these patients.