BACKGROUND Group B streptococcus (GBS) is a leading cause of neonatal sepsis in many industrialized countries, but reports from the developing world infrequently identify this pathogen among newborns with sepsis. Studies of GBS colonization among women living in developing countries were reviewed to determine whether lower colonization rates might account for these findings. METHODS Literature was reviewed with the use of Medline Express (1980 to 1996) and Abstracts on Tropical Agriculture and Rural Development in the Tropics (1975 to 1995). The methods of each report were considered adequate if specimens were collected from the vagina and if selective broth media were used. RESULTS Thirty-four studies reported results of cultures from 7730 women; overall colonization was 12.7%. Among only those studies in which methods were adequate, 17.8% (675 of 3801) women were identified as colonized. Studies with adequate methods found significantly higher colonization rates (relative risk, 2.3; 95% confidence interval, 2.0 to 2.6) than those using inadequate methods. When analysis was restricted to reports with adequate methods, the prevalence of colonization by region was as follows: Middle East/North Africa, 22%; Asia/Pacific, 19%; Sub-Saharan Africa, 19%; India/Pakistan, 12%; and Americas, 14%. CONCLUSIONS Although there is significant geographic variation in the proportion of women colonized with GBS, the range of colonization reported from developing countries is similar to that identified in populations studied in the United States. Specimen collection and microbiologic methods are important factors in identification of women colonized with GBS.