Antimony potassium tartrate (APT) is a complex salt that until recently was used worldwide as an antischistosomal drug. Treatment was efficacious only if APT was administered intravenously to humans at a near lethal total dose of 36 mg/kg. Because unconfirmed epidemiologic studies suggested there might be an association between APT treatment and bladder cancer, we initiated prechronic toxicity studies with the drug to select a route of administration and doses in the event that chronic studies of APT were needed. The toxicity and concentration of tissue antimony levels were compared in 14-d studies with F344 rats and B6C3F1 mice administered APT in the drinking water or by ip injection to determine the most appropriate route for longer term studies. Drinking water doses estimated by water consumption were 0, 16, 28, 59, 94 and 168 mg/kg in rats and 0, 59, 98, 174, 273, and 407 mg/kg in mice. APT was poorly absorbed and relatively nontoxic orally, whereas ip administration of the drug caused mortality, body weight decrements, and lesions in the liver and kidney at doses about one order of magnitude below those in drinking water. Because of these data and the dose-related accumulation of antimony in the target organs, an ip dose regimen was selected for subsequent studies. Both sexes of F344 rats and B6C3F1 mice were given 0, 1.5, 3, 6, 12, and 24 mg/kg doses of APT every other day for 90 d by ip injection. There were no clinical signs of toxicity nor gross or microscopic lesions in mice that could be attributed to toxicity of APT, although elevated concentrations of antimony were detected in the liver and spleen of mice. Rats were more sensitive than mice to the toxic effects of APT, exhibiting dose-related mortality, body weight decrements, and hepatotoxicity. The concentrations of antimony measured in liver, blood, kidney, spleen, and heart of rats were proportional to dose, but there were no biochemical changes indicative of toxicity except in the liver. Hepatocellular degeneration and necrosis occurred in association with dose-related elevations in activities of the liver-specific serum enzymes sorbitol dehydrogenase and alanine aminotransferase. By alternating the site of abdominal injection and the days of treatment, mesenteric inflammation at the site of administration was minimized in the rats and mice, indicating that the ip route would be suitable for chronic studies.(ABSTRACT TRUNCATED AT 400 WORDS)