- Background: Studies on short-course treatment of childhood pneumonia in the developed world are lacking. We compared clinical and laboratory outcomes of a 3-day or a 5-day to a 10-day treatment in young children with community-acquired alveolar pneumonia. Methods: A double-blind, randomized, placebo-controlled trial was conducted in 2 stages: (1) 3 days versus 10 days and (2) 5 days versus 10 days. Amoxicillin (80 mg/kg/d; divided into 3 daily doses) was used for all arms. Case definition was: age 6–59 months; radiologically confirmed community-acquired alveolar pneumonia; temperature ≥38.5°C; peripheral white blood cell count ≥15,000/mm3; status permitting outpatient treatment. Scheduled visits were on days 4–5 and 10–14 with daily telephone interviews for 30 days. Treatment failure was defined by the need for a rescue treatment or hospitalization. Secondary outcomes were: duration of fever and symptoms and white blood cell and C-reactive protein responses. Results: During Stage 1, 4/10 (40%) and 0/12 (0%) evaluable patients failed in the 3-day and 10-day arms, respectively (P = 0.16). Therefore, the 3-day arm was replaced by a 5-day arm (Stage 2). No failures occurred in the 5-day (n = 56) and 10-day (n = 59) arms. Overall, 4/10 (0%), 0/56 and 0/42 (0%) children failed in the 3-day, 5-day and 10-day arms, respectively (P < 0.001, 3-day versus 5-day or 10-day). Secondary outcomes were similar in the 5-day and 10-day arms. Conclusion: In 6- to 59-month-old outpatients with community-acquired alveolar pneumonia, a 5-day course with high-dose oral amoxicillin was not inferior to a 10-day course. The 3-day regimen may be associated with an unacceptable failure rate.