Laboratory Testing for Hit Antibodies: How Much Class Do We Need?.

Abstract

HIT is usually caused by platelet-activating antibodies of IgG class that bind to neoepitopes on platelet factor 4 (PF4) bound to heparin or certain other polyanions. Commercial enzyme-immunoassays (EIAs), however, measure PF4/polyanion-reactive antibodies of IgA and IgM class, in addition to IgG class antibodies. This raises the question: Does the detection of these IgA and IgM class antibodies improve HIT assay operating characteristics (perhaps because IgA and IgM are a marker for pathogenic HIT-IgG and/or for clinical HIT) or worsen operating characteristics (perhaps via detection of numerous non-HIT sera containing non-pathogenic IgA and/or IgM class antibodies)? We performed systematic serologic studies of 362 stored sera from a clinical trial of heparin therapy (Arch Intern Med 2003; 163: 2518) in which 12 patients developed HIT by clinical criteria (>50% platelet count fall) and 350 patients did not develop HIT. Assays used included: serotonin release assay (SRA); in-house PF4/heparin-EIA that individually detects IgG (EIA-G), IgA (EIA-A), and IgM (EIA-M) class antibodies; and commercial EIA from GTI (EIA-GTI). The 12 patients with clinical HIT tested strongly positive in the SRA, EIA-G, and EIA-GTI (Table 1). Positive tests among the 350 non-HIT patients (to calculate test specificity) were seen in 12 (SRA), 28 (EIA-G), and 69 (EIA-GTI) patients.Table 1. Serologic features of 12 patients with clinical HIT.Serotonin ReleaseIn-house EIA-IgGCommercial EIASensitivity (true-POS)12/12 (100%)12/12 (100%)12/12 (100%)Median (IQR) POS result98.5% (90.0, 99.5)1.669 (1.223, 2.056)1.802 (1.355, 2.344)Specificity (true-NEG)338/350 (96.6%)322/350 (92.0%)281/350 (80.3%)IQR=interquartile (25%, 75%) range; data are percent serotonin release using washed platelets and absorbance units (OD405) using EIA-G and EIA-GTI.In contrast, the EIA-A and EIA-M assays were positive in less than half of the HIT patients. Further, the magnitude of the IgA and IgM anti-PF4/heparin immune response did not differ between the 12 HIT patients, and the 69 non-HIT patients who had any PF4/polyanion immune response, as defined by a positive EIA-GTI without HIT (Table 2).Table 2. Comparison of the HIT and non-HIT Immune Response for IgA and IgM.IgA PositiveIgA: Median (IQR)IgM PositiveIgM: Median (IQR)HIT (n=12)5/12 (41.7%)0.400 (0.210, 1.421)3/12 (25.0%)0.340 (0.269, 0.522)Non-HIT immune response (n=69)*25/69 (36.2%)0.316 (0.208, 0.870)18/69 (26.1%)0.322 (0.193, 0.475)P value0.750.581.000.25* Non-HIT immune response defined as positive EIA-GTI but no clinical HIT.Similar observations were made when we compared the 24 SRA-positive patients (including the 12 HIT patients) against the 57 SRA-negative, EIA-GTI positive patients. CONCLUSION: Detection of PF4/polyanion-reactive IgA and IgM class antibodies worsens the operating characteristics of HIT assays through the detection of numerous non-pathogenic antibodies, without any offsetting advantages in the detection of pathogenic HIT antibodies. Optimal diagnostic laboratory testing for HIT antibodies should include a platelet activation assay and an EIA that detects only IgG class antibodies reactive against PF4/heparin (or PF4/polyanion).