Oral step-down, optimal drug, and total duration of antibiotic treatment in African children hospitalised with severe community-acquired pneumonia (PediCAP): a factorial randomised controlled trial

BACKGROUND: WHO recommends 5 days of intravenous antibiotics for children hospitalised with severe community-acquired pneumonia (CAP). We aimed to investigate the safety of a step-down to different oral antibiotics and the shortest effective duration. METHODS: PediCAP was an open-label, parallel group, 2 × 5 factorial randomised controlled trial of children aged 2 months to 6 years hospitalised with severe CAP without complicating factors in 13 hospitals across five sub-Saharan African countries. Children weighing 3-30 kg and with point-of-care C-reactive protein of more than 10 mg/L were randomly assigned (5:5:1) to either a step-down from intravenous antibiotics to oral amoxicillin (250 mg) or amoxicillin-clavulanate (co-amoxiclav; 200 mg amoxicillin to 28·5 mg clavulanate) via dispersible tablets twice daily (superiority comparison) for five total (intravenous plus oral) durations (4-8 days; non-inferiority comparison), or a 5-day, fixed-duration, intravenous-only treatment (non-inferiority comparison). Children were stepped down when clinically improved and able to take oral antibiotics. Clinicians could change antibiotics if clinically indicated. The primary outcome was the proportion of readmission or death at day 28 (non-inferiority margin vs intravenous only: +10%). The study was registered with the ISRCTN registry (ISRCTN63115131) and is complete. FINDINGS: Between Dec 7, 2020, and Aug 14, 2023, 2248 children were screened for eligibility and 1101 were randomly allocated (480 [43·6%] girls and 621 [56·4%] boys). The primary outcome was available in 1055 (95·8%) children. For children in the step-down groups, the mean total antibiotic exposure was 6·0 days (SD 3·5) in the 4-day group, 6·8 days (4·2) in the 5-day group, 7·4 (3·6) in the 6-day group, 8·3 days (3·4) in the 7-day group, and 9·2 days (2·9) in the 8-day group (p<0·0001), with 140 (68·6%) of 204, 151 (76·3%) of 198, 167 (85·2%) of 196, 179 (89·5%) of 200, and 186 (91·6%) of 203, respectively, stepping down within their randomised duration (p<0·0001). Primary outcomes occurred in 33 (6·9%) of 475 in the co-amoxiclav group, 27 (5·6%) of 484 in the amoxicillin group, and six (6·3%) of 96 in the intravenous-only group, with both oral step-down strategies non-inferior to the intravenous-only strategy (upper 95% CIs of 6·0 for co-amoxiclav and 5·7 for amoxicillin) and no evidence of superiority for co-amoxiclav over amoxicillin (adjusted risk difference 1·3% [95% CI -1·8 to 4·4]; p=0·40). Primary outcomes occurred in eight (4·1%) of 194 in the 4-day group, ten (5·3%) of 190 in the 5-day group, 16 (8·5%) of 188 in the 6-day group, 16 (8·3%) of 193 in the 7-day group, and ten (5·2%) of 194 in the 8-day group; all durations were non-inferior to the 8-day group (all upper 95% CIs ≤6·0%). There was no evidence of consistent differences in adverse events for oral antibiotic or duration comparisons. For antibiotic-related or serious adverse events, there was one (1·0%) in the intravenous-only group and 12 (2·5%) in the amoxicillin group (adjusted risk difference -2·3% [95% CI -4·2 to -0·3]; p=0·025), and 16 (3·4%) in the co-amoxiclav group (+0·8% [-1·2 to 2·8]; p=0·42); rates increased with randomised duration (slope estimate +0·9% [0·1 to 1·7]; p=0·032). INTERPRETATION: For sub-Saharan African children hospitalised with severe CAP without complicating factors, a strategy of stepping down upon clinical improvement to oral amoxicillin after initial intravenous antibiotics, with a total treatment duration of 4-5 days, is non-inferior to WHO-recommended 5-day intravenous treatment. FUNDING: The Second European and Developing Countries Clinical Trials Partnership (EDCTP2).