In a previous study, we found fecal colonization with multiresistant Escherichia coUl exhibiting high-level trimethoprim resistance in 19% of diapered children attending six day-care centers in Houston, Tex. To examine the potential risk factors associated with this finding, we conducted cross-sectional studies among 203 children attending 12 day-care centers, 51 children attending a well-child clinic (controls), and 64 medical students. The prevalence of fecal colonization with trimethoprim-resistant E. coli among children attending day-care centers (30%) was higher (P < 0.001) than among control children (6%) or medical students (8%). The prevalence of colonization among the children attending the 12 centers ranged from 0 to 59% and was correlated with the number of diapered children enrolled (r = 0.73; P < 0.01). In a case control study among the day-care center children, significant risk factors were an age of less than 12 months and attendance at a center with an enrollment of over 40 diapered children (odds ratios of 2.2 and 3.5, respectively); ethnicity, duration of attendance, and prior antibiotic administration were not associated with colonization. Plasmid analysis of 60 of the day-care center strains revealed 22 profiles, each of which was unique to a given day-care center. Transmission and carriage of trimethoprim-resistant strains for as long as 6 months was documented in one center studied on three occasions. Given the documented transmission of enteric pathogens among diapered children attending day-care centers and their spread into family members, it is likely that day-care centers are an important community reservoir of plasmid-associated antibiotic-resistant E. coli.Resistance to antimicrobial agents, an increasing problem worldwide, is particularly common in certain settings, including tertiary care hospitals, animal feedlots, and developing countries (10). In a study conducted in 1983, trimethoprim resistance (Tmp) was found in 40 to 50% of over 2,500 human clinical isolates of Escherichia coli in Chile and Thailand but in only 4 to 8% of concurrent clinical isolates of E. coli in the United States (11). Since most E. coli infections result from endogenous fecal E. coli, one might hypothesize that the frequency of Tmpr among fecal E. coli would parallel that among clinical isolates in the same population. Levy et al. (7) found 7.5% of healthy medical students in Boston in 1987 to be colonized with Tmpt E. coli, a frequency similar to that among U.S. clinical E. coli isolates (11).We conducted a pilot study of antimicrobial agent-resistant fecal flora among children attending day-care centers (DCC) in Houston, Tex., in 1986 (17) and reported that 19% of 79 diapered children in seven DCC were colonized with Tmpr E. coli, all of which were multiresistant. This frequency of Tmpr was higher than that previously reported among strains colonizing healthy U.S. adults (7) or adults receiving trimethoprim (TMP) (3,14,19,20) Moreover, the overall Tmpr E. coli colonization rate of 19% was due primarily...