Estimation of survival probability in cirrhotics. Evaluation of a computer model applicable for decision-making.

In 25 randomly selected, non-alcohol-induced cirrhotics, a simple mathematical model originally designed as the "Mayo-model" for the prediction of prognosis and survival in patients with primary biliary cirrhosis, was applied using a newly developed computer program. The calculated risk scores were 4.91 for surviving patients compared to 7.65 for those who died (P less than 0.001). The survival probability at the end of the follow-up period was 88% vs 47% (P less than 0.001) for these two groups. In 12 patients of the total of 25, multiple values were calculated at yearly intervals. Usually these R values increased except in one patient with chronic active hepatitis treated effectively with corticosteroids, and for 2 others during the correction of vitamin K deficiency. Our data suggest that this model, originally designed for and validated in patients with primary biliary cirrhosis, may also be applied to predict survival in other non-alcohol-induced cirrhotics and therefore may be applicable in prospective controlled trials in cirrhotics or in evaluating the need for and the timing of liver transplantation in the (near) future in these patients.