Diet protein increases whereas carbohydrates decrease urea synthesis. Traditionally, these effects have been explained by changes in substrate supply. Diet protein intake increases whereas carbohydrate decreases blood amino acid concentration. However, glucose also decreases urea synthesis by a hepatic mechanism independent of the decrease in blood amino acid concentration. Whether this is due to an effect of glucose in itself, or whether the fall in glucagon or the rise in insulin is responsible, was not known. This survey deals with the effect of an increase in diet protein intake and of the separate effects of glucose, glucagon and insulin on functional hepatic nitrogen clearance in normal man and in patients with cirrhosis of the liver. The functional hepatic nitrogen clearance is calculated as the slope of the linear regression analysis of alanine-stimulated urea synthesis rate and blood alpha-amino nitrogen concentration, and expresses urea synthesis independent of changes in blood amino acid concentration. In patients with cirrhosis, hepatic nitrogen clearance is reduced in parallel with liver cell mass, despite high glucagon concentration that would normally up-regulate the process. In both healthy subjects and in patients with cirrhosis, an increase in diet protein intake (plus approximately 50 g/day) for 14 days increases hepatic nitrogen clearance by 40%. Thus, in addition to the substrate effect, protein intake increases urea synthesis by an effect in the liver, probably by enzyme formation. What induces this is not clear but high postprandial levels of glucagon may be involved. Although the effect is qualitatively intact in the patients, the response relative to the increase in protein intake is reduced by two-thirds. The effect may be important to control blood amino acid concentration during a high protein diet and may partly explain why patients with cirrhosis usually tolerates protein hyperalimentation without developing hepatic encephalopathy. It is shown that the reduction of hepatic nitrogen clearance by glucose depends on hyperglycaemia, and is accomplished by the additive effects of a direct hormone-independent action of glucose, and indirectly via suppression of glucagon. Insulin is not a direct controller of hepatic nitrogen clearance, but is still considered an important regulator of urea synthesis by its reducing effects on blood amino acid concentration. High experimental glucagon levels overrule the normal suppressive effect of glucose. In contrast, it is shown that the sugar-alcohol xylitol normalises the glucagon induced increase in hepatic nitrogen clearance. During normal glucagon levels xylitol exerts only a very little decrease in hepatic nitrogen clearance. In patients with cirrhosis, glucose does not down-regulate hepatic nitrogen clearance. However, when the spontaneous high glucagon levels are normalised by somatostatin, glucose decreases hepatic nitrogen clearance. This shows that the direct hormone-independent effect of glucose is intact. These findings indicate that the high glucagon levels during spontaneous hormone responses overrule the suppressive effect of glucose. Incomplete glucose suppression of glucagon secretion during alanine infusion contributes to the high glucagon levels. The removal of the high glucagon levels decreases hepatic nitrogen clearance in itself. Thus, the hyperglucagonaemia may be a compensatory mechanism by which the cirrhotic liver to some extent reestablishes its capacity to produce urea. The consequence is the defective down-regulation of hepatic nitrogen clearance by glucose. The reduction in urea synthesis by glucose, i.e. its nitrogen sparing effect, is accomplished by two different mechanisms: A hepatic component (reduction of the hepatic nitrogen clearance) and a peripheral component (reduced substrate availability mediated by the insulin response). This is an extension of former thoughts according to which glucose reduces urea synthesis due solely to