Patients with primary biliary cirrhosis (PBC), despite excellent outcomes after liver transplantation (LT), may develop recurrent primary biliary cirrhosis (rPBC). The impact of immunosuppression and HLA mismatches on rPBC is unclear. We evaluated 103 consecutive PBC patients who underwent transplantation (follow-up > or = 10 months) with serial protocol biopsies. Cox regression was used to evaluate factors associated with rPBC: the Model for End-Stage Liver Disease score pre-LT, year of transplantation, age and gender of the recipient and donor, cold and warm ischemic times, HLA mismatches, rejection, infections, and immunosuppression (initial/maintenance). The mean follow-up was 108 months (10-239 months), rPBC occurred in 36, and the mean was 44 months (10-200 months). Immunosuppression was cyclosporine-based in 38 (10 initially on monotherapy) and tacrolimus-based in 62 (19 initially on monotherapy). Steroids were discontinued in all but 7. Azathioprine was part of the initial immunosuppression in 70, 26 discontinued it, and 33 were never exposed to it. rPBC was associated independently with nonuse/discontinuation of azathioprine (P = 0.015, hazard ratio = 0.046, 95% confidence interval = 0.008-0.261). The mean time to rPBC was 74 months with azathioprine, 43 months when AZA was discontinued, and 31 months if no azathioprine was used. Cyclosporine or tacrolimus alone had no impact on rPBC, but cyclosporine with azathioprine was protective for rPBC in comparison with tacrolimus/azathioprine (0/18 versus 7/25, respectively; P < 0.001). rPBC was not affected by HLA mismatches. Azathioprine use in PBC patients who underwent transplantation was associated with less disease recurrence and a longer time to rPBC. Tacrolimus or cyclosporine individually had no effect, but cyclosporine and azathioprine in combination resulted in the least rPBC.