Of readily available methods to estimate the donor liver size, measurement of the body circumference at the xiphoid level (xiphoid measure) appeared to be the most accurate in the present prospective study of 60 donors and 57 recipients (r = 0.64, P = 0.0001). The estimated liver volume could be calculated using the equation: bloodless liver volume (1) = 1.44 x xiphoid measure (m). The difference between donor and recipient xiphoid measures was significantly higher in slowly recovering patients than in those recovering uneventfully (7 +/- 7 cm vs. - 5 +/- 8 cm, P < 0.001). The bloodless donor liver volume measured by water displacement averaged 1249 +/- 230 ml and had increased by 3 weeks posttransplant by 64 +/- 28% as determined using computed tomography. The volume of the liver graft seemed to adapt to the recipient as it correlated positively with body weight (r = 0.64, P < 0.01) and negatively with the age of the recipient (r = -0.42, P < 0.01). The liver graft volume seemed to increase less markedly in patients with a slow recovery than in those with an uncomplicated recovery (37% +/- 15% vs. 68% +/- 24%, P < 0.001). We conclude that a simple measurement of the body circumference at the xiphoid level can be used to estimate the donor liver volume. A gross mismatch of this parameter between the donor and the recipient seems to increase the risk of graft dysfunction. We also found that the change in the liver graft volume is influenced by the recipient's age and body weight.