A prospective, randomized, double-blind clinical trial of bovine and porcine heparin was conducted. One hundred forty-one patients were randomized, of whom 89 received heparin treatment for six or more days (mean, ten days). Two developed severe thrombocytopenia (platelet count less than 20,000 per μl); both were randomized to the bovine heparin group, but one inadvertently received two doses of porcine heparin. Laboratory investigation suggested that the thrombocytopenia in these two patients was immunologically mediated, and platelet reactivity to both bovine and porcine heparin was demonstrated. Twenty patients had a decline in platelet count of greater than 50,000 from baseline, although the total count remained above 150,000 per μl. In seven of these subjects, the platelet count returned to its original level while heparin therapy was continued. Of the 13 patients with a persistent decrease in platelet count, ten had received bovine heparin and their counts decreased by an average of 88,000 per μl; the reduction in the three porcine-treated patients was 68,000; this difference was not statistically significant. In the remaining patients, the post-treatment platelet counts in both groups were significantly higher than pre-treatment values (p < .005), perhaps indicating a cessation of the platelet consumption that accompanied the original thrombotic event. We conclude that a modest and persistent reduction in platelet count commonly occurs during heparin therapy (13/89 receiving treatment for six or more days), but that severe thrombocytopenia is relatively rare (2/89).