1. The clearance and biotransformation of caffeine (1,3,7-trimethylxanthine) were investigated in eight healthy control subjects and 16 patients with cirrhosis, by measuring serial serum caffeine concentrations and recoveries of methylxanthine metabolites in urine for 48 h after a 400 mg oral caffeine load.
2. In the control group, the mean (± sd) serum caffeine clearance was 1.3 ± 0.4 ml min−1 kg−1 and a mean of 56.4 ± 16.5% of the administered caffeine was recovered from the urine over 48 h as methyluric acids and methylxanthines. The majority of the metabolites were excreted in the first 24 h period and only 2.0 ± 1.4% of the administered caffeine was excreted unchanged.
3. Patients with compensated cirrhosis (n = 10) metabolized caffeine similarly to the control subjects. Thus the mean serum caffeine clearance was 1.4 ± 1.2 ml min−1 kg−1 and a mean of 57.2 ± 11.7% of the administered caffeine was recovered from the urine over 48 h. The majority of the metabolites were excreted in the first 24 h; the pattern of metabolic excretion was unaltered and only 2.2 ± 0.9% of the administered caffeine was excreted unchanged.
4. In the patients with decompensated cirrhosis (n = 6), significant changes were observed in caffeine metabolism. The mean serum caffeine clearance (0.4 ± 0.2 ml min−1 kg−1) was significantly impaired compared with controls (P < 0.01) and a significant delay was observed in metabolite excretion in the urine. Thus the mean recovery of metabolites in the urine during the first 24 h (25.0 ± 11.2%) was significantly reduced compared with controls (44.1 ± 12.4%, P = 0.03), whereas the mean urinary metabolite recovery in the second 24 h (20.9 ± 10.5%) was insignificantly increased compared with controls (12.3 ± 7.8%). Overall, the mean recovery of metabolites in the urine in 48 h (45.9 ± 15.4%) was similar to that in the control group. The overall recovery of unchanged caffeine was significantly greater than in controls (5.0 ± 2.8% vs 2.0 ± 1.4%, P = 0.04), but the pattern of metabolite excretion was otherwise unchanged.
5. In the patients with liver disease there were significant linear correlations between the degree of hepatocellular dysfunction and the serum caffeine elimination half-life (r = 0.774; P < 0.01) and the total recovery of methylxanthine metabolites in the urine, in the 0–24 h (r = 0.702; P = 0.002) and 0–48 h (r = 0.581; P = 0.018) periods.
6. Caffeine clearance is impaired in patients with decompensated cirrhosis either because of a reduction in hepatic caffeine uptake or else because of a reduction in ‘functioning hepatocyte mass’. However, the biotransformation of caffeine is unaltered in the presence of hepatic dysfunction.