Determination of protective efficacy after natural rotavirus infection is important as a basis for evaluating rotavirus vaccines. Therefore, placebo recipients in a large 2-year rotavirus vaccine trial conducted across the United States were followed to determine the protection afforded by natural rotavirus infection. Serotype 1 rotaviruses predominated (93% year 1, 66% year 2), but isolates of all four major human rotavirus serotypes circulated during both years. Of the 45 placebo recipients with documented rotavirus illnesses in year 1, 1 developed rotavirus disease in year 2 compared with 29 of the other 235 placebo recipients (P = .03). Serologic data were available for 171 placebo recipients, and 37 of 140 without rotavirus illnesses in year 1 had evidence of asymptomatic rotavirus infection. None of these 37 experienced rotavirus disease in year 2 compared with 22 of the remaining 103 (P < .001). Overall efficacy after natural rotavirus infection was 93% (95% confidence interval, 50%-99%).