In phase I of a 2-phase study, 56 evaluable children (0.8 to 5 years) with lobar or segmental pneumonia received intravenous or intramuscular ceftriaxone 50 mg/kg/day for 2 days followed by oral cefetamet pivoxil 20 mg/kg/day in 2 divided doses to complete 7 days of treatment. All patients achieved a clinical cure. In phase II, a randomised open multicentre study, 62 children with pneumonia received an identical regimen to phase I (arm A), and 59 children received ceftriaxone 50 mg/kg/day for 1 day followed by 6 days' treatment with cefetamet pivoxil 20 mg/kg/day (arm B). Patients from phase I and arm A were combined giving a total of 118 evaluable patients in arm A. At the end of treatment, 100% of patients in arm A and 96% in arm B achieved a clinical cure; cure was maintained in 99 and 98% of patients, respectively. Two (4%) patients in arm B failed therapy; in both cases, factors other than treatment failure may have accounted for the poor response. 11 and 12% of patients in treatment arms A and B, respectively, experienced adverse events; gastrointestinal events (nausea and/or vomiting) were reported in 9 and 8% of patients, respectively. In conclusion, 1 or 2 days' treatment with parenteral ceftriaxone before switching to oral cefetamet pivoxil was safe and effective in the treatment of childhood pneumonia. Therefore, parenteral-oral switch is a feasible treatment option in the treatment of serious paediatric community-acquired pneumonia.