Abstract
The World Health Organization (WHO) recommends intravenous (IV) ampicillin and gentamicin as first-line therapy to treat severe pneumonia in children under five years of age. Ampicillin needs to be administered at a six-hourly interval, which requires frequent nursing intervention and bed occupancy for 5-7 days, limiting its utility in resource-poor settings. We compared the efficacy of IV amoxicillin over IV ampicillin, which is a potential alternative drug in treating severe pneumonia in children between 2-59 months. We conducted an unblinded, randomized, controlled, non-inferiority trial in the Dhaka hospital of icddr,b from 1 January 2018 to 31 October 2019. Children from 2-59 months of age presenting with WHO defined severe pneumonia with respiratory danger signs were randomly assigned 1:1 to either 50 mg/kg ampicillin or 40 mg/kg amoxicillin per day with 7.5 mg/kg gentamicin. The primary outcome was treatment failure as per the standard definition of persistence of danger sign(s) of severe pneumonia beyond 48 h or deterioration within 24 h of therapy initiation. The secondary outcomes were: (i) time required for resolution of danger signs since enrolment, (ii) length of hospital stay, (iii) death during hospitalization, and (iv) rate of nosocomial infections. Among 308 enrolled participants, baseline characteristics were similar among the two groups. Sixty-two (20%) children ended up with treatment failure, 21 (14%) in amoxicillin, and 41 (27%) in ampicillin arm, which is statistically significant (relative risk [RR] 0.51, 95% CI 0.32-0.82; p = 0.004). We reported 14 deaths for serious adverse events, 4 (3%) and 10 (6%) among amoxicillin and ampicillin arm, respectively. IV amoxicillin and IV gentamicin combination is not inferior to combined IV ampicillin and IV gentamicin in treating severe pneumonia in under-five children in Bangladesh. Considering the less frequent dosing and more compliance, IV amoxicillin is a better choice for treating children with severe pneumonia in resource-limited settings.