Incidence and risk factors for diarrhea following kidney transplantation and association with graft loss and mortality.

BACKGROUND Gastrointestinal complications after kidney transplantation are associated with inferior graft outcomes. We examined the incidence, risk factors, and outcomes of posttransplantation diarrhea. STUDY DESIGN Historic cohort study. SETTING & PARTICIPANTS We examined first kidney transplant recipients in the United States from 1995 to 2002, with follow-up through December 2002. Recipients of multiple organs were excluded. We limited our study population to Medicare beneficiaries. PREDICTORS Recipient, donor, and transplant characteristics were ascertained by means of US Renal Data System database inquiry. OUTCOMES Incidence of diarrhea, graft loss, and death after transplantation. First episodes of diarrhea after transplantation were ascertained by using International Classification of Disease, Ninth Revision, Clinical Modification codes using Medicare billing data. Cause of diarrhea was classified as infectious or not and according to specific cause. Graft loss and death were ascertained from the date of the first diarrhea episode. RESULTS We enrolled 41,442 patients. Mean follow-up was 758 +/- 399 days. We observed 7,103 diarrhea cases and 8,104 graft losses (4,201 deaths). The 3-year cumulative incidence of diarrhea was 22%, with 18% diagnosed as noninfectious diarrhea with an unspecified cause. Using multivariate Cox proportional hazards analysis, factors associated with increased risk of unspecified noninfectious diarrhea were female sex (hazard ratio [HR], 1.40; 95% confidence interval, 1.33 to 1.48), type 1 diabetes (HR, 1.20; 95% confidence interval, 1.06 to 1.37), and regimens containing tacrolimus and mycophenolate mofetil (HR, 1.37; 95% confidence interval, 1.28 to 1.46). Unspecified noninfectious diarrhea was associated with increased risk of graft failure (HR, 2.13; 95% confidence interval, 1.98 to 2.28) and patient death (HR, 2.04; 95% confidence interval, 1.85 to 2.24). LIMITATIONS Use of claims data to ascertain patient characteristics and events; inability to make causal inference based on retrospective designs. CONCLUSIONS Regimens containing tacrolimus and mycophenolate mofetil were associated with increased risk of noninfectious diarrhea. Episodes of noninfectious diarrhea doubled the hazard of graft loss and patient death.