The relationship between the impairment in hepatic and renal function in cirrhosis has not been well established. This study investigated urinary sodium excretion in comparison with quantitative parameters of liver function in 75 patients with various degrees of cirrhosis kept on a constant salt diet of 120 mmol/d for 5 days before the start of the study. The aminopyrine breath test (ABT), indocyanine green (ICG) elimination, galactose elimination capacity (GEC), and hepatic sorbitol elimination (HSE) served as quantitative parameters of liver function. Results for the quantitative tests were compared with those for the Child-Pugh score. Urinary sodium excretion showed a significant nonlinear relationship to ABT (r = .70; P < .0001). Less-significant correlations were observed for ICG (r = .60), the Child-Pugh score (r = -.57), GEC (r = .44), and HSE (r = .34). Because a number of significant correlations were observed between the different liver function tests, multivariate analysis was used to further elucidate the relationship between hepatic function and sodium excretion. Only one independent predictor of urinary sodium excretion could be identified, and that was the ABT (P < .02). More than half of the nonascitic patients showed a urinary sodium excretion of less than 80% of dietary sodium intake, indicating impaired renal sodium handling in preascitic cirrhosis. Based on the 95% confidence interval (CI) for ABT of nonascitic patients with normal (mean ABT 0.56% dose x kg/mmol CO2; 95% CI: 0.44 to 0.69) and reduced urinary sodium excretion (mean ABT 0.26% dose x kg/mmol CO2; 95% CI: 0.18 to 0.35), a threshold level of ABT of about 0.4 (% dose x kg/mmol CO2) for conservation of normal urinary sodium excretion in cirrhosis can be defined. This ABT value reflects an approximate 50% reduction in function compared with the mean of cirrhotic patients with normal liver and kidney function (0.81% dose x kg/mmol CO2). The presence of ascites was also associated with a reduction in ABT to below 0.4 (% dose x kg/mmol CO2), while, for all other parameters, either the cut-off point was close to the lower limit of normal or no cut-off level could be detected. In conclusion, the results of the present study provide further evidence that the impairment in urinary sodium excretion in cirrhosis is related to hepatic function. The data suggest a nonlinear relationship. Because ABT has been shown to reflect functional hepatocellular mass, the occurrence of sodium retention and ascites appears to be related to a threshold of an approximate 50% reduction in functional liver cell mass.