HPi: A Novel Parameter to Predict Graft-related Outcome in Adult Living Donor Liver Transplant.

Abstract

Portal hyperperfusion is frequently associated with early allograft dysfunction (EAD). It is imperative to identify patients who would require portal inflow modulation. We aimed to identify factors associated with hyperperfusion-related graft injury and develop a predictive index for the same. Prospectively maintained database was queried to identify 135 adult living donor liver transplant recipients between September 2016 and July 2020. According to the calculated sample size, 96 patients were randomly selected for "test cohort". The remaining 39 patients made the "validation cohort." EAD was defined according to the A2ALL study. "Hyperperfusion index (HPi)," defined as posttransplant portal pressure gradient (ΔPpost)/graft-to-recipient splenic volume ratio (GRSVR), was devised on the basis of laws of flow dynamics and regression analysis. Overall, 40 patients (29.6%) had EAD, six 90-d mortalities (4.4%) were attributable to EAD. In the test cohort, EAD patients (n = 29, 30.2%) had lower GRSVR (1.00 versus 2.22, P < 0.001), higher ΔPpost (14.8 versus 11.9, P = 0.004), and HPi (20.89 versus 8.67, P < 0.001). Multivariate analysis revealed GRSVR, ΔPpost, and HPi as significant factors to predict EAD. Receiver operating characteristic determined cutoff of HPi ≥9.97 could predict EAD with sensitivity of 90% and specificity of 73% (F-score = 0.712). HPi ≥16.25 predicted 90-d mortality with sensitivity of 100% and specificity of 78.9%. Patients with higher HPi had delayed graft-related recovery. Non-EAD patients had a higher 1-y (96% versus 79%) and 2-y (88% versus 79%) survival. The cutoff of HPi was validated well in the validation cohort (F-score = 0.645) (Hosmer-Lemeshow test, P = 0.89). While predicted GRSVR may help identify at-risk patients preoperatively, intraoperatively calculated HPi is more accurate in identifying patients who would require portal inflow modulation. Achieving an HPi below target cutoff significantly decreases the risk of EAD even in low-GRSVR patients.