This study was undertaken to evaluate an assay to assess the risk for drug-induced immediate hypersensitivity reactions. Groups of five to 10 guinea-pigs were given six ip injections of the test compound on days 1, 3, 5, 8, 10 and 12. Aluminium hydroxide was also given in the first injection. At day 33, the animals were given an iv injection of the test compound and the response was recorded by grading the severity of clinical symptoms. Cutaneous passive anaphylaxis was also evaluated in six naive guinea pigs using Evans blue and sera collected from treated animals on day 26. A panel of six positive model compounds (ovalbumin, aprotinin, chymopapain, tetracosactide, cyanocobalamin and procaine), and the negative compound Ribomunyl were tested. Positive systemic and/or cutaneous anaphylactic responses were observed with ovalbumin, aprotinin, chymopapain and tetracosactide whereas no responses were noted with cyanocobalamin, procaine and Ribomunyl. Our results suggest that this protocol can help differentiate positive model compounds (known to induce reactions in man) from negative model compounds, provided that their molecular weight is large enough, but that it is not applicable to substances of low molecular weight.