The objective of this study was to describe the prevalence of Escherichia coli O157 in the feces and on the hides of finishing beef cattle fed a standard diet and those fed diets supplemented with direct-fed microbials. Two hundred forty steers received one of four treatments throughout the feeding period: (i) control: no added microbials; (ii) HNP51: high dose of Lactohacillius acidophilus strain NP 51 (10(9) CFU per steer daily) and Propionibacterium freudenreichii (10(9) CFU per steer daily); (iii) HNP51+45: high dose of NP 51 (10(9) CFU per steer daily), P. freudenreichii (10(9) CFU per steer daily), and L. acidophilus NP 45 (10(6) CFU per steer daily); or (iv) LNP51+45: low dose of NP 51 (10(6) CFU per steer daily), P. freudenreichii (10(9) CFU per steer daily), and NP 45 (10(6) CFU per steer daily). Samples were collected from each animal and analyzed for the presence of E. coli O157 using immunomagnetic separation methods on day 0 (feces), 7 days before harvest (feces), and at harvest (feces and hide). At the end of the feeding period, cattle receiving HNP51 were 57% less likely to shed detectable E. coli O157 in their feces than were the controls (P < 0.01). For animals receiving HNP51+45 and LNP51+45, fecal prevalence did not differ from that of the controls. The prevalence of positive hide samples was least among cattle receiving HNP51+45 (3.3%); these animals were 79% less likely (P < 0.06) to have a positive hide sample than were the controls (prevalence = 13.8%). There was poor agreement of the culture results between fecal and hide samples collected from the same animal (kappa = 0.08; confidence interval = -0.05 to 0.2). Cattle supplemented with a high dose of NP 51 had reduced E. coli O157 prevalence in both fecal and hide samples, indicating that this treatment may be an efficacious preharvest intervention strategy.