The purpose of this study is to determine the incidence of heparin-associated thrombocytopenia (HAT) for various subgroups of heparin exposed patients and the impact of study quality on the reported incidence. Articles were identified using a Medline search, a manual search of the Index Medicus, and a review of article references. Key data included heparin type, administration route, indication, treatment duration, outcome criteria, incidence, and platelet count reliability. The pooled incidence estimate in studies requiring a repeatedly abnormal platelet count was compared with estimates from studies not requiring a repeated platelet count. The results showed that there were no adequately designed studies to estimate the incidence of HAT-related thrombosis or hemorrhage. The pooled incidence of HAT in studies requiring a reproducibly lowered platelet count (< 100,000/microL) was 3/281 (1.1%) with intravenous porcine heparin and 4/140 (2.9%) with intravenous bovine heparin. This difference was not statistically significant. The incidence of HAT with intravenous bovine heparin was significantly lower in studies that required a repeated platelet count. The incidence of HAT with heparin administered subcutaneously was small (0%) and in those studies requiring a repeatedly abnormal platelet count, there was no difference between porcine and bovine heparin. The authors concluded that the incidence of HAT is < 3% with intravenous heparin and extremely low for subcutaneous heparin. Study quality may influence the reported incidence of HAT.