Trachoma, an ocular infection caused by Chlamydia trachomatis, is the second leading cause of blindness worldwide. The blinding sequelae, which occur in middle age, are felt to be the result of numerous or lengthy episodes of severe inflammatory trachoma in childhood. Risk factors for constant, severe trachoma were identified in a group of children enrolled in a longitudinal study in Kongwa, Tanzania, where villages were randomized in a clinical trial of mass treatment and a behavior modification campaign. In 1989, each of 1,417 randomly selected children had photographs taken of an upper eyelid for determination of their trachoma status. The photographs were graded by a reader who was masked as to the village and date of each photograph. Risk factor data on the family's socioeconomic status, distance to water, and hygiene practices were obtained at baseline. Follow-up examinations occurred 2, 6, and 12 months from baseline. Data from all four time points were available for 82% of the children enrolled. Overall, 10% of the children had constant, severe trachoma, defined as severe trachoma at three or four examinations. The odds ratio for severe trachoma was 1.9 for female children (95% confidence interval 1.3–2.7). Familial cattle ownership and having one or more siblings with trachoma at baseline were also significantly related to the odds of having severe trachoma. Children with a sustainably clean face had lower odds (odds ratio = 0.4, 95% confidence interval 0.3–0.7). A subgroup of 10% of children in these hyperendemic communities always seemed to have severe trachoma, despite enrollment in a mass treatment campaign. Improved face-washing plus antibiotic treatment may decrease the likelihood that these children will be at risk for blinding complications in adulthood. Am J Epidemiol 1996;143:73–8