Background: With documented rise in bacterial resistance in vitro to co-trimoxazole, there is a need to document in vivo effectiveness of the drug in treatment of non-severe pneumonia in the community setting.

Setting: The study will be conducted in 18 rural primary health center (PHC) in 9 districts in India near Lucknow, Nagpur, New Delhi, Mumbai, Chennai, Trivandrum, Vellore, Chandigarh and Bhopal.

Design: The unit of randomization will be the PHC. The recruited children will be followed up on days 4, 6 and 13-15 to assess primary and secondary outcome measures.

Hypothesis: The primary hypothesis is that the clinical failure rate with use of either oral amoxycillin for three days or five days co-trimoxazole is similar.

Intervention: Oral Co-trimoxazole (8mg/kg/day trimithoprim) twice a day for five days vs. oral amoxycillin (20 mg/kg/day) thrice a day for three days.

Main objective: To compare the proportion of children aged 2 to 59 months presenting with non-severe pneumonia with or without wheeze, who do achieve clinical cure on day 6 on treatment with 5 days oral co-trimoxazole and 3 day oral amoxycillin, respectively.

Main outcomes measures: Clinical cure on day 6; clinical failure between day 1-6; clinical relapse between day 7-15.

Inclusion criteria: Children aged 2-59 months with non-severe pneumonia based on WHO criteria of respiratory rate above the age specific cut-off with or without wheeze, accessible to follow up, whose guardians give written informed consent.

Exclusion criteria: Children with severe disease, other infectious disease requiring antibiotic treatment, documented use of antibiotic for the last 48 hours, three or more episodes of wheezing in a year, diagnosed asthmatics, known immunodeficiency disease, acute or chronic organic disease, history of hospitalization within last 15 days, measles within last one month, drug allergy, refusal to give consent, prior enrollment in the study.

Sample size: Has been calculated to test the null hypothesis. There will be 1100 children in each arm. Thus each PHC is required to recruit a minimum of 122 cases within one year.

Policy relevance: The findings of the study can effect the ARI management policy. If the current study proves that three day amoxycillin is clinically as effective as five day co-trimoxazole and with the well documented evidence of rising resistance to co-trimoxazole, short course amoxycillin may become the first line treatment for non-severe pneumonia Time line: 18 months. Pilot in March 2003, enrollments from April 2003, Interim analysis Nov. 2003, DMC meet Dec. 2003, Data cleaning June 2004, DMC meet July 2004, Manuscript writing Aug.2004.