[Prognostic factors in pediatric liver transplantation. Multivariate analysis].

Aim: To analyze independent risk factors associated with poor graft and patient survival in a series of 292 pediatric liver transplants (PLT) performed in 234 children during a 15 years period. MATERIAL AND METHODS. 1. Univariate graft and patient survival analysis in 45 variables related to pretransplant patient status, surgical technique and donor conditions. 2. Variables found with univariate analysis to be associated with outcome were entered into a stepwise backward proportional hazard model (Cox), to determine independent prediction of outcome.

Results: 11 variables influence the graft survival: recipient age, z-score recipient height, UNOS status, recipient and donor weight, transplant for immune hepatitis, platelet transfusion during the transplant, blood index > 4 during the surgery, type of arterial reconstruction, retransplantation and era of the transplant (first er: 1986-1990; 2nd. era: 1991-1995; 3rd. era: 1996-2000). Four of those variables are independent in the multivariate analysis: UNOS 1 status (Odds Ratio, OR = 2.82, 95% confidence interval = 1.36-5.85), recipient < 3 years (OR = 3.76, 95% CI = 2.13-6.63), transplants for autoimmune hepatitis and era (OR of first and second versus third era respectively 3.93 and 2.81). The independent variables influencing the patient survival were: children receiving more than one graft children less than 3 years old and transplant era.

Conclusions: Liver transplant in small children is associated with an increased risk of graft loss and patient dead. The experience of the hospital in pediatric liver transplantation improves the results, particularly in small children.