The issue of small-for-size graft (SFSG) containing the middle hepatic vein in right liver living donor liver transplantation from 1996 to 2008 (n = 320) was studied. Characteristics of donors, grafts and recipients were comparable between Era I (first 50 cases) and Era II (next 270 cases) except that the median model for end-stage liver disease (MELD) score was higher in Era I (29 vs. 24; p = 0.024). The median graft to standard liver volume ratio (G/SLV) in Era I was 49.0% (range, 32.8-86.2%), versus 49.3% (range, 28.4-89.4%) in Era II (p = 0.498). Hospital mortality rate, the study endpoint, dropped from 16.0% (8/50) in Era I to 2.2% (6/270) in Era II (p = 0.000). Univariate analysis showed that MELD score (p = 0.002), pretransplant hepatorenal syndrome (p = 0.000) and Era I (p = 0.000) were significant in hospital mortality. Logistic regression analysis showed that only Era I (relative risk 9.758; 95% confidence interval, 2.885-33.002; p = 0.000) was significant. In Era I, G/SLV<40% had a relative risk of 7.8 (95% confidence interval, 1.225-49.677; p = 0.030). The hospital mortality rates for G/SLV<40% were 50% (3/6) and 1.9% (1/52) in Era I and II respectively. In conclusion, through accumulation of experience, SFSG became less important as a factor in hospital mortality.