“Iceberg model” of HIT. Clinical HIT, comprising HIT with (HIT-T) or without thrombosis, is represented by the portion of the iceberg above the waterline; the portion below the waterline represents subclinical anti-PF4/heparin seroconversion. Three types of assays are highly sensitive for the diagnosis of HIT: the washed platelet activation assays (SRA and HIPA), the IgG-specific PF4-dependent EIAs (EIA-IgG), and the polyspecific EIAs that detects anti-PF4/heparin antibodies of the 3 major immunoglobulin classes (EIA-IgG/A/M). In contrast, diagnostic specificity varies greatly among these assays, being the highest for the platelet activation assays (SRA and HIPA) and lowest for the EIA-IgG/A/M. This is because the EIA-IgG/A/M is most likely to detect clinically irrelevant, non-platelet-activating anti-PF4/heparin antibodies. The approximate probability of SRA⁺ status in relation to a given EIA result, expressed in OD units, was obtained from the literature.⁷