Hemophilia Liver Transplant Observational Study (HOTS)

Introduction: Hepatitis C is the leading cause of liver disease in individuals with hemophilia (H). HIV co-infection in 80% accelerates HCV disease progression. Although antiretroviral therapy improves HIV survival and slows HCV progression, some ultimately require transplantation. In contrast to non-hemophilic (NH) risk groups, H have 1) longer duration HCV infection with HCV acquisition via clotting factor in the first year of life; and 2) greater bleeding risk due to factor VIII (IX) deficiency in the setting of thrombocytopenia in end-stage liver disease. These risks are important as 1) duration of HCV infection is a significant predictor of liver disease progression; and 2) increased bleeding risk, including gastrointestinal bleeding, increases post-transplant mortality. Several studies have shown that, while post-transplant survival is similar, pre-transplant survival is shorter in HIV/HCV+ H than NH candidates. These studies were small, limiting analysis for predictors of pre-transplant survival.

Methods: We conducted a retrospective observational study utilizing United Network for Organ Sharing (UNOS) national transplant registry data, comparing HCV+ H and HCV+ NH candidates. A 7-item case report form was completed on each subject by eight respective transplant centers including 1) hemophilia status (hemophilia A, ICD 286.0; hemophilia B, ICD 286.1); 2) hepatocellular carcinoma status; 3) HIV status; 4) CD4+ count; 5) HIV RNA PCR; 6) HCV RNA PCR; and 7) platelet count, the latter four variables at listing and at transplantation. We performed univariate proportional hazards models for pre-transplant mortality within 90 days of listing, including group and baseline factors. Variables with p<0.1 from univariate models were included in an initial multivariate model. Kaplan-Meier time to-event curves were constructed for transplant candidates, including time to death, time to transplant, and time to MELD 25; and for transplant recipients, including time to death and time to graft loss. Statistical analysis was by SAS version 9.3, Cary NC.

Results: We identified 2,502 HCV+ liver transplant candidates, including 144 (5.8%) with HIV infection, 36 (1.4%) with hemophilia, and 1,213 (48.5%) undergoing liver transplantation from eight U.S. university-based transplant centers for the period January 1, 2004 to December 31, 2010. Other than male predominance, 100% vs. 78.0%, and younger age, 41 vs. 48 years, both p<0.0001, baseline data were similar between H and NH. A total of 1,289 candidates were not transplanted; 771 remained alive or were removed from listing pre-transplant; and 518 expired pre-transplant. The most common causes of pre-transplant death included multi-organ failure in 113 (21.8%) and sepsis or infection in 86 (16.6%). Among transplant candidates, the overall pre-transplant patient survival at 90-days and 6-months from listing did not differ between H and NH. In the first 90 days post-listing, time-to-event curves for death and elevated MELD 25 were significantly different between H and NH groups, log-rank, p=0.0001 and 0.02, respectively. In univariate proportional hazards models, pre-transplant mortality within 90-days of listing was associated with higher baseline MELD, HR=1.15, p<0.001, and lower platelet baseline platelet count, HR=1.11 per 25k/µL, p=0.04, and HIV+ status, p=0.003. The hazard of pre-transplant mortality was marginally worse in HIV+ H than HIV+NH candidates, HR=2.8, p=0.08. In multivariate analysis, higher MELD score was significantly associated with pre-transplant mortality (p<0.0001), and the hazard of pre-transplant mortality remained marginally worse in HIV+ H than HIV+ NH, HR=2.0, p=0.24.

Conclusion: This observational study confirms MELD is a significant predictor not only of post-transplant survival, but also pre-transplant survival. Despite longer duration HCV infection and greater bleeding risk, there was no significant difference in pre-transplant survival between HIV+ H and NH, likely due to small numbers of hemophilia candidataes missed by retrospective ICD classification. We suggest implementing ICD classification for H in the UNOS system, to enable robust data collection to determine the impact of HCV antiviral therapies on pre- and post-transplant outcomes, and the potential utility of a combination MELD classification that incorporates age at HCV acquisition and platelet count.