Ratio of antibody binding to PF4/polysaccharide complexes compared to PF4 alone by fluid-phase EIA. Results of fluid-phase EIA testing for sera from 15 patients who formed anti–PF4/heparin-IgG antibodies (detected using solid-phase EIA) while receiving enoxaparin (n = 6, •) or fondaparinux (n = 9, ○). The data are expressed as ratios of binding to PF4 in the presence of polysaccharide (UFH, 0.6 IU/mL; LMWH, 0.5 anti-Xa U/mL; danaparoid, 0.1 anti-Xa U/mL; and fondaparinux, 0.1, 0.4, 1.2, and 10.0 μg/mL) over the baseline (buffer). Horizontal bars indicate medians. * indicate the 4 samples that tested positive (in the presence of UFH) in the platelet activation assay. For comparison, results are also shown for 15 patients with clinical HIT (▪). Statistically significant increases in reactivity (null hypothesis, mean ratio of OD [presence of drug]/OD [presence of buffer] = 1) for the 15 sera obtained from patients in the orthopedic trials were observed for UFH (P = .003), LMWH (P < .001), danaparoid (P = .002), but not with fondaparinux at any concentration (P > .05). Whereas 14 of 15 sera from patients in the orthopedic trials exhibited more than 2-fold greater reactivity than baseline against PF4/LMWH, none reacted similarly against PF4/fondaparinux (P < .001 by the McNemar test, 2-tailed).