In a 14-day multiple-dose study the pharmacokinetics of paroxetine was investigated in 12 patients with alcoholic cirrhosis and in 6 subjects without liver disease. The dose of 20-30 mg paroxetine daily was adjusted to the reduction in liver function, as assessed by the galactose elimination capacity. Accordingly, all but two of the cirrhotic patients received 20 mg, while all six control subjects received 30 mg. Dose-corrected, trough drug concentration at steady state (CSSmin) and dose-corrected AUC24h were significantly higher in the patients with liver diseases than in the control subjects [3.4 vs 1.5 ng.ml-1 per mg paroxetine and 89 vs 43 h (ng).ml-1 per mg paroxetine]. The elimination t1/2 was prolonged [83 vs 36 h], but the difference was not statistically significant, and the cirrhotic patients were still able to clear almost all the paroxetine by metabolism. All but two patients with cirrhosis experienced nausea during the first two or three days after the first dose, while none of the controls had this symptom. The study showed slower elimination of paroxetine and consequently higher plasma levels in patients with cirrhosis, suggesting that in the latter the dose of paroxetine should be in the lower end of the therapeutic range.