Circulating levels of hyaluronan (HA) may be a clinically useful non-invasive test in liver disease to evaluate, for example, the degree of fibrosis or cirrhosis. Healthy subjects exhibit a two-fold or greater postprandial elevation of plasma HA. We undertook this work to determine the effect of eating on plasma HA in patients with cirrhosis, to define mechanisms underlying the changes observed and to examine whether using fasting samples improves the discriminating value of plasma HA as a clinical test. Plasma HA was measured using a protein-binding assay in serial samples obtained over 2.5 h, during which a 2100 kJ meal and/or intravenous infusion of octreotide (25 microg stat, 25 microg/h) were given, or in paired random and fasting samples from patients with cirrhosis. In six patients with Pugh's A cirrhosis, plasma HA increased more than two-fold after food ingestion, peaking after 75-90 min. In four patients (one Pugh's A), no response was observed. Octreotide transiently increased plasma HA by a mean of 20%, but, in combination with eating, reduced the peak plasma HA by 80% in responding patients. In 30 patients with cirrhosis, random plasma HA was abnormal in only 70% compared with 93% of autologous fasting samples, using the respective normal ranges (P=0.01). Ingestion of food markedly increased plasma HA in patients with compensated but not decompensated cirrhosis, most likely due to a postprandial increase in splanchnic blood flow. Fasting patients prior to taking blood improves the clinical discriminating value of plasma HA.