The incidence of antibodies to heparin-PF4 complexes (H-PF4) has been evaluated in patients who were under heparin therapy for more than 7 days: 109 patients treated with unfractionated heparin (UH) and 100 patients with low-molecular-weight heparin (LMWH). The presence of antibodies was identified in 17% of the former group and 8% of the latter. In both the UH and the LMWH groups, IgM antibodies were found in all but four patients who showed IgA antibodies. IgG isotypes were only detected in five patients and were consistently associated to either IgM or IgA antibodies. The follow-up of H-PF4 antibodies in 76 patients treated with UH from 1 to > or = 12 days showed a relationship between the incidence of antibodies and the duration of therapy. Despite the presence of anti-H-PF4 antibodies there was no thrombocytopenia (<150 10(9)/L) in the patients. A significant drop of platelets requiring the discontinuation of heparin was observed, however, in three patients, but their platelet count consistently remained >150 10(9)/L. Our study demonstrates that the induction of antibodies to H-PF4 is a frequent phenomenon in patients treated with UH or with LMWH. The absence of thrombocytopenia and of clinical complications in these patients demonstrates that other conditions must be associated with H-PF4 antibodies for inducing type II HIT: optimal concentrations of heparin and PF4 in the blood circulation to allow the formation of macromolecular H-PF4 complexes, presence of activated platelets that present an increased binding of H-PF4 complexes, increased expression of FcgammaRIIA receptors, or presence of their H 131 phenotype. We conclude that the measurement of antibodies to H-PF4 complexes allows the detection of heparin-treated patients at risk of developing type II HIT.