Background. In intestinal transplantation, graft versus host disease (GVHD), while relatively rare, remains a major cause of morbidity and mortality posttransplant. Because of its rarity of occurrence, no multivariable analysis of risk factors for GVHD development has previously been reported. Methods. We used Cox stepwise regression to determine the significant multivariable predictors of the hazard rate of developing biopsy-proven GVHD during the first 60 months posttransplant among 445 consecutive intestinal transplant cases at our center since 1994. Results. GVHD was observed in 8.8% (39/445); median time-to-GVHD development (range) was 1.5 months (0.5-17.3 mo) posttransplant. Sites of GVHD included skin (N = 21), skin/gastrointestinal tract (N = 6), gastrointestinal tract/rectum (N = 4), skin/liver (N = 4), skin/lung (N = 2), skin/rectum (N = 1), and skin/bone marrow (N = 1). Three factors were selected into the Cox model offering significant protection from GVHD development (listed in order of selection): isolated intestine or liver-intestine (LI) (versus modified multivisceral [MV] or MV) allograft (P = 0.00003), alemtuzumab (versus no induction, anti-CD25, rabbit antithymocyte globulin, or rabbit antithymocyte globulin/rituximab) induction (P = 0.004), and liver inclusion (LI or MV) (P = 0.009). These results remained unchanged even after accounting for the propensity to receive alemtuzumab induction. Observed GVHD incidence was 2.4% (3/125), 0.0% (0/38), 17.9% (7/39), and 11.9% (29/243) for isolated intestine, LI, modified MV, and MV allografts, and 2.7% (3/113) versus 10.8% (36/332) for those receiving versus not receiving alemtuzumab induction, respectively. Conclusions. These results may help advance the current state of knowledge about risk factors for GVHD development following intestinal transplantation.
