The limitations of donor age, graft size, and the Model for End-Stage Liver Disease (MELD) score have not been apparent in living donor liver transplantation (LDLT). Our team developed a formula for predicting graft survival after LDLT; the formula includes the graft weight, donor age, MELD score, and portosystemic shunt status. The aims of this study were to re-evaluate the reliability of our formula and to assess whether our modified treatment strategy has improved 6-month graft survival. Two hundred seventeen patients were allocated into 2 groups: patients with predictive scores 1.30 (n 5 162) and patients with predictive scores < 1.30 (n 5 55). The latter group was also divided into subgroups of patients with scores of 1.15 to 1.30 (n 5 37) and patients with scores < 1.15 (n 5 18). Survival rates for patients with scores < 1.30 were significantly worse than rates for patients with scores 1.30 (P 5 0.006). Survival rates for patients with scores < 1.15 were significantly worse than rates for patients with scores of 1.15 to <1.30 (P < 0.001). A multivariate analysis showed that a predictive score < 1.15 (odds ratio 5 7.87, P 5 0.006) and a body mass index 30 kg/m 2 (odds ratio 5 13.3, P < 0.001) were independent risk factors for 6-month graft mortality. In conclusion, predictive scores reliably predict 6-month graft survival and could allow a widening of the safe ranges for donor ages and graft sizes. Liver Transpl 20:323-332, 2014. V C 2013 AASLD.