Twenty-two cirrhotic patients who underwent hepatic resection from July 1989 to March 1991 at Vancouver General Hospital were analyzed prospectively in order to determine whether there was any preoperative evaluation of liver function that would help identify those patients who would not survive hepatic resection. Patients were analyzed on the basis of age, type of resection, and a variety of so-called standard liver function tests. In addition, all patients were evaluated preoperatively with an indocyanine green (ICG) clearance test, a relatively new test that we attempted to evaluate in terms of its role in hepatic resection in cirrhotic patients. Parametric statistical evaluation used included Student's t-test and multivariate regression, as well as discriminate analysis. The nonparametric evaluation used was the Wilcoxon rank sum test. Overall, the 30-day mortality rate was 18%, with those patients who did not survive resection having a significantly lower ICG clearance than those who underwent successful resection (p less than 0.0001). No other liver function test was useful in determining the outcome of resection. Similarly, neither age nor type of resection appeared to influence outcome. Use of discriminate analysis (p = 0.0029) allowed the identification of a cutoff point for ICG clearance below which hepatic resection should not be attempted.