Clostridium difficile infection (CDI) recurs in ∼20% of patients. Prior studies indicated that antibody responses directed against the C. difficile toxins A and B were potentially associated with lower risk of recurrent CDI. Here we tested the hypothesis that circulating anti-toxin IgG antibody levels associate with reduced risk of recurrent CDI. A cohort study with prospective enrollment and retrospective data abstraction examined antibody levels in 275 adult patients at the University of Michigan with CDI. We developed an enzyme linked immunosorbent assay to detect IgG antibodies against toxin A and toxin B in sera obtained at the time of diagnosis. Logistic regression examined the relationship between antibody levels and recurrence, and sensitivity tests evaluated for follow-up and survivor biases, history of CDI, and PCR ribotype. Follow-up data were available for 174 subjects, of whom 36 (20.7%) had recurrence. Comparing antibody levels vs. recurrence and CDI history, anti-toxin A levels were similar, while anti-toxin B levels had a greater range of values. In unadjusted analysis, detection of anti-toxin A antibodies, but not anti-toxin B antibodies, associated with an increased risk of recurrence (OR 2.71 [1.06, 8.37], P = .053). Adjusting for confounders weakened this association. The results were the same in sensitivity analyses. We observed a borderline increased risk of recurrence in patients positive for anti-toxin A antibodies, and sensitivity analyses showed this was not simply a reflection of prior exposure status. Future studies are needed to assess how neutralizing antibody or levels after treatment associate with recurrence.