Gram-negative bacilli (GNB) cause many cases of pneumonia in Indonesia. We investigated nasopharyngeal carriage of GNB in Semarang, Indonesia. Klebsiella pneumoniae carriage in adults (15%) was higher than in children (7%) (P â«Ű⏠0.004), while that of other GNB was comparable. Poor food and water hygiene are determinants of carriage of these bacteria.W orldwide, Streptococcus pneumoniae is the most common cause of community-acquired pneumonia (CAP). However, in low-to-middle-income countries, Gram-negative bacilli (GNB) are reported as frequent causes of CAP (1). In a recent study on CAP in Indonesia, Klebsiella pneumoniae was the most common etiological agent (our unpublished observation). Since nasopharyngeal colonization may precede pneumonia, as for S. pneumoniae (2), we investigated nasopharyngeal carriage of K. pneumoniae and other GNB among children and adults compared to that of S. pneumoniae in Indonesia. Subjects were healthy children (6 to 60 months old) and adults (45 to 70 years old) living in the city of Semarang on the island of Java in Indonesia, without respiratory symptoms and without antibiotic consumption within the last 3 days. Subjects were recruited from all 16 districts of Semarang by cluster random sampling from February to April 2010.Nasopharyngeal swabs were obtained and transported in Amies (COPAN, Italy) and then were inoculated on 5% sheep blood agar with gentamicin (5 mg/liter) and on MacConkey agar. The plates were incubated at 35°C in 5% CO 2. Identification and antimicrobial susceptibility tests for GNB were performed using Vitek 2 (bioMĂ©rieux, Marcy l'Etoile, France). Identification of S. pneumoniae was performed using the optochin test (Oxoid, Basingstoke, United Kingdom). Control strains were included, and CLSI guidelines were applied.A questionnaire was developed to identify determinants of carriage of these bacteria. Data on demography, water supply (for drinking, preparing food, and bathing), food hygiene and house sanitation (crowding, flies, and exposure to smoke from cigarettes and mosquito coils), were collected. Water hygiene was considered poor when water other than tap water or bottled water was used. Food hygiene was considered poor if food was bought from street vendors. Crowding was considered if the ratio of bedroom space per family member was Ïœ4 m 2 (3). Univariate analysis was done with chi-square or Fisher's exact test when appropriate, followed by logistic regression using SPSS 17 (SPSS Inc., Chicago, IL, USA). Confidence interval (CI) was calculated at the 95% level (CI 95 ), and P values of Ïœ0.05 were considered significant. Written informed consent was obtained from the subjects or their caregivers. The study was approved by the Ethical Committee of the Faculty of Medicine, Diponegoro University.A total of 253 adults and 243 children participated in the study; approximately 25% resided in the suburbs, and the majority were living under hygienically restricted circumstances regarding their food and water supplies. Crowding and exposure to smoke were common...