In a Latin square design, nine New Zealand white male rabbits (2.8-4.8 kg) each received intravenous indocyanine green (ICG) in doses of 2.5, 12.5, and 25 mg/kg. A period of 4 weeks separated consecutive experiments. A specific high-pressure liquid chromatographic (HPLC) assay and the traditional spectrophotometric (SPEC) method were used to monitor plasma ICG. The analyses demonstrated the ambiguous nature of the SPEC assay, and the HPLC procedure pointed to the presence of an unidentified ICG metabolite. Dose-dependent ICG disposition was evident from the ANOVA of the mean (+/- SD) clearances, namely, 17.09 (7.35), 4.50 (0.82), and 2.27 (0.57) mL X min-1 X kg-1 for the aforementioned doses, respectively. Analysis of variance of the clearances also demonstrated a significant ordering effect suggesting cumulative ICG toxicity. The mean ICG profiles for the three doses accorded with a novel three-compartment model containing two ICG distributional spaces in addition to a compartment (liver) responsible for ICG elimination from the circulation. A saturable uptake process into the eliminating compartment (Vmax = 0.93 mg X kg-1 X min-1; Km = 31.9 mg/L) accounted for the dose-dependent features. Additional studies in four unanesthetized rabbits with chronically catheterized bile ducts revealed no disparity between the SPEC and HPLC analyses of biliary ICG. The mean (+/- SD) ICG recovery in bile following an intravenous dose of about 12.5 mg/kg was 59.8 (16.3)%.