Laboratory diagnosis of heparin-induced thrombocytopenia.

Abstract

Heparin-induced thrombocytopenia (HIT) is a clinical-pathological disorder; thus, laboratory testing for the pathogenic platelet-activating antiplatelet factor 4 (PF4)/heparin antibodies is central for diagnosis. The "iceberg" model summarizes the inter-relationship between platelet activation assays and PF4-dependent immunoassays, with platelet-activating antibodies comprising a subset of anti-PF4/heparin antibodies. The platelet serotonin-release assay (SRA), performed by reference laboratories, has high sensitivity and specificity for HIT (~95% each), and is especially suited for detecting highly pathogenic HIT sera containing both heparin-dependent and heparin-independent platelet-activating antibodies; this latter subgroup of antibodies explains "autoimmune HIT" disorders (delayed-onset, persisting, spontaneous, heparin "flush," fondaparinux-associated). Recently, SRA-negative HIT has become recognized, in which serum from some HIT patients contains subthreshold levels of platelet-activating antibodies (by SRA) that become detectable using a PF4-enhanced platelet activation assay. Unusual immunologic features of HIT include early antibody detectability (at onset of platelet count fall) and antibody transience (seroreversion). Widely available PF4-dependent enzyme immunoassays (EIAs) have high sensitivity but poor specificity for HIT, although specificity is enhanced with IgG-specific EIAs and strong positive results; unfortunately, EIA results are usually not available in real time. Automated rapid immunoassays, such as the chemiluminescence immunoassay (CLIA) and latex immunoturbidimetric assay (LIA), facilitate real-time laboratory diagnosis. Recently available likelihood ratio (LR) data for positive (LR+) and negative (LR-) test results allow clinicians to adjust their pretest probabilities for HIT, using Bayesian analysis, into real-time posttest probabilities that are dramatically increased (test positive) or decreased (test negative). Moreover, (semi-)quantitative CLIA- and LIA-positive results (weak, moderate, strong positive) can further refine the posttest probability of HIT.