Community-based treatment of serious bacterial infections in newborns and young infants: a randomized controlled trial assessing three antibiotic regimens.

BACKGROUND Sepsis in the neonatal period is a major cause of child mortality in low-income countries. Hospitalization and parenteral penicillin/ampicillin and gentamicin therapy are recommended for management. Many families, however, are unable to access hospital care, and most home-delivered newborns who develop sepsis die without receiving antibiotic therapy. Appropriate community-based therapy in such situations is undefined. We compared failure rates of 3 clinic-based antibiotic regimens in 0- to 59-day-old infants with possible serious bacterial infection whose families refused hospitalization in Karachi communities with high neonatal mortality rates>45/1000 live births. METHODS Eligible infants were randomly assigned to 7 days of: (1) procaine penicillin [50,000 units/kg once daily (OD) by intramuscular injection (IM)] and gentamicin (5 mg/kg OD IM) reference arm, (2) ceftriaxone (50 mg/kg OD IM), or (3) oral trimethoprim-sulfamethoxazole (TMP-SMX) at 10 mg/kg/day divided twice daily and gentamicin IM OD. Primary outcome was treatment failure, defined as death, deterioration in clinical condition during therapy or no improvement after 2 days. RESULTS Possible serious bacterial infection was diagnosed in 704 infants, among 5766 screened. Among 434 (61.6%) randomized to clinic-based therapy, there were 13 of 145 failures with penicillin-gentamicin, 22 of 145 with ceftriaxone and 26 of 143 with TMP-SMX-gentamicin. Treatment failure was significantly higher with TMP-SMX-gentamicin compared with penicillin-gentamicin [relative risk 2.03, 95% confidence interval: 1.09-3.79] by intention-to-treat analysis. Differences were not significant in the ceftriaxone versus penicillin-gentamicin comparison [relative risk 1.69, 95% confidence interval 0.89-3.23). By 14 days, there were 2 deaths in the penicillin-gentamicin group, 3 in the ceftriaxone group and 11 in the TMP-SMX-gentamicin group [relative risk 5.58, 95% confidence interval: 1.26-24.72 (group 3 versus 1)]. CONCLUSION When hospitalization of sick infants is unfeasible, outpatient therapy with injectable antibiotics is an effective option. Procaine penicillin-gentamicin was superior to TMP-SMX-gentamicin. Ceftriaxone is a more expensive option, and may be less effective, although this requires further research.