Timeline of heparin‐induced thrombocytopenia seroconversion in serial plasma samples tested using an automated latex immunoturbidimetric assay

Abstract Introduction HIT is caused by platelet‐activating IgG that recognize multimolecular PF4/heparin complexes. HIT antibodies are generally detectable by PF4‐dependent enzyme immunoassay (EIA) and by platelet serotonin‐release assay (SRA) at the beginning of the HIT‐related platelet count fall. We determined whether an automated immunoassay for HIT, the latex immunoturbidimetric assay (LIA), also detects antibodies early during the course of HIT. The LIA was also used to evaluate a patient with putative SRA‐negative HIT. Methods We evaluated the timing and magnitude of LIA reactivity in serial plasma samples obtained from 19 SRA‐positive patients (17 with abnormal platelet count changes indicating HIT; two with subclinical seroconversion) and one putative SRA‐negative HIT patient, all obtained from patients who participated in a clinical trial of heparin thromboprophylaxis. We determined LIA status at the onset of the HIT‐related platelet count fall. Results The LIA was positive in all 19 SRA‐positive patients (median value, 7.3 U/mL [range, 1.2‐35.5]; cutoff, 1.0 U/mL); for all 13 evaluable patients for whom an informative plasma sample was available at (or shortly before) the onset of the HIT‐related platelet count fall, LIA reactivity was positive. Heterogeneity in seroconversion using the LIA was observed; some patients exhibited gradual increases in reactivity, whereas other patients showed rapid increase in reactivity over a few days. The single clinical trial patient who met clinical‐pathological criteria for “SRA‐negative HIT” tested LIA‐positive. Conclusion The LIA detects HIT antibodies at the beginning of the HIT‐associated platelet count fall. The LIA was also positive in a patient with SRA‐negative HIT.


| INTRODUC TI ON
Immune heparin-induced thrombocytopenia (HIT) is caused by platelet-activating antibodies that recognize multimolecular complexes of a cationic protein, platelet factor 4 (PF4), bound either to heparin or to platelet-associated polyanions. [1][2][3] An important feature of HIT is the high sensitivity (≥97%) of PF4-dependent immunoassays for detecting pathogenic antibodies, 4,5 in contrast to the much lower sensitivity of platelet antibody testing for other immune-mediated thrombocytopenic disorders, such as idiopathic thrombocytopenic purpura [6][7][8] or "classic" (non-HIT) drug-induced immune thrombocytopenia. 9, 10 We have previously reported that PF4-dependent enzyme immunoassays (EIAs) and the platelet serotonin-release assay (SRA) are generally positive at the onset of the HIT-related platelet count fall, including during the early phase of the platelet count decline when HIT would not even be suspected. 11,12 We have proposed 13 that the high sensitivity of serum/plasma-based assays for HIT antibodies may have implications for the pathogenesis of HIT; for example, significant levels of free (unbound) HIT antibodies might be required to produce the dynamic conditions essential to form multimolecular PF4-polyanion complexes on platelet surfaces needed to engage and crosslink platelet FcγIIa receptors, resulting in Fc receptor-mediated platelet activation characteristic of HIT. [14][15][16] We capitalized on the availability of archived plasma samples from a clinical trial of heparin thromboprophylaxis to examine whether early detectability of HIT antibodies is also seen with an automated, rapid immunoassay, known as the latex immunoturbidimetric assay or LIA. To our knowledge, studies using serial plasma samples during acute seroconversion have not been reported using this assay. The LIA is performed on a coagulation instrument (ACL TOP ® Family, Instrumentation Laboratory), and thus, plasma, rather than serum, is used for testing. 17 As most testing for HIT is performed using serum, the availability of these well-characterized plasma samples provided us a unique opportunity to evaluate LIA reactivity associated with HIT seroconversion.
Given the potential for increased use of this assay to provide realtime, on-demand testing for HIT antibodies 18,19 -including incorporation into real-time Bayesian diagnostic analysis 20 -we sought to determine the changes in LIA reactivity during acute seroconversion and, particularly, whether assay positivity occurs during the earliest phase of HIT, at a time when the platelet count has begun to decline but usually too early for a diagnosis of HIT to be contemplated.
During the study, we became aware of one study patient who had a high clinical probability of HIT, and who tested EIA-positive. However, this patient had tested negative in the platelet serotonin-release assay (SRA) and, thus, was previously considered by our group not to have had HIT. In recent years, the concept of SRA-negative HIT has been proposed, [21][22][23] in which subthreshold levels of heparin-dependent platelet-activating antibodies (by SRA) can be detected using modifications to platelet activation assays such as addition of PF4 24,25 aimed to increasing test sensitivity. Therefore, we investigated whether this patient met clinical and laboratory criteria for SRA-negative HIT. As serial plasma samples were available for this patient, we also studied the LIA seroconversion profile of this patient with putative SRA-negative HIT.

| Archived blood samples from clinical trial
Archived plasma samples were available from a clinical trial of heparin thromboprophylaxis. 26 We have previously tested these plasma samples by SRA, 27 in-house IgG-specific EIA, 28 commercial IgG-specific EIA (LIFECODES PF4 IgG assay; Immucor GTI Diagnostics), 29 and commercial polyspecific EIA that detects antibodies of IgG, IgA, and IgM classes (LIFECODES PF4 Enhanced assay; Immucor GTI Diagnostics), 30 to elucidate certain clinical and laboratory features of HIT 11,12,31,32 ; however, we have not previously utilized these samples for LIA testing, nor for investigating patients with putative SRA-negative HIT. In this study, we focused on examining LIA seroconversion among SRA-positive patients, as well as determining whether any clinical trial patients met criteria for SRA-negative HIT. As described subsequently, we therefore identified all available archived patient plasma samples from the aforementioned thromboprophylaxis clinical trial in which patients were either known to have previously tested SRA-positive or had tested SRA-negative/EIA-positive but who a clinical picture suggesting a plausible diagnosis of HIT (discussed later in section, Investigations for SRA-negative HIT).

Permission was received from the Hamilton Integrated Research
Ethics Board to perform these studies (#1288-T).

| Definitions of HIT
Patients from the clinical trial met one of the following previously published definitions for HIT 12,31,32 : (a) classic (standard) definition of thrombocytopenia, that is, a platelet count falls to less than 150 × 10 9 /L that began five or more days after starting heparin 31 ; (b) proportional definition of thrombocytopenia, that is, platelet count fall of at least 50% 32 or 30.0% to 49.9% 12 from the postoperative peak that began five or more days after starting heparin; or (c) blunted platelet count recovery, that is, a platelet count that did not show the expected platelet count rise during the second postoperative week, and which also fell below the bounds of an appropriate control antibody-negative patient population with similar baseline (preoperative) platelet counts. 12         c No evidence of HIT by platelet count fall or thrombosis, but the patient did present approximately 2 mo later with symptomatic pulmonary embolism.
d First day that positive SRA was observed.
e Venogram report showed "valve clot in superficial femoral vein".

| Patient samples studied
Although many of the archived plasma samples had been expended during these and other studies, we were able to evaluate remaining samples available from each of the following patients groups for this study: (a) classic (standard) definition of thrombocytopenia, n = 5; (b) proportional thrombocytopenia, n = 8 (with six patients meeting >50% fall and two patients with 30.0%-49.9% platelet count fall, and with none of these eight patients attaining a platelet count nadir less than 150 × 10 9 /L); (c) blunted platelet count recovery, n = 4; and (d) subclinical seroconversion, n = 2. In addition, we had available serial plasma samples from a patient with (e) putative SRA-negative HIT (see next section).

| Investigations for SRA-negative HIT
We used the following criteria for SRA-negative HIT 24

| Timing of LIA vs EIA-GAM seroconversion
For patients with available/informative blood samples, we compared the timing of seroconversion between the LIA and the commercial

| Comparison with LIA reactivities from other studies
We compared the LIA reactivities of 20 patients from our study with Hospital is a regional trauma center that also has large medical, cardiology, cardiac surgery, and vascular surgery patient populations).

| D ISCUSS I ON
We describe a study of LIA seroconversions in patients with serial plasma samples obtained in a historic clinical trial of post-hip replacement heparin thromboprophylaxis. 26 We found that the The sensitivity of the LIA for HIT might be similar to that for the SRA, given that some patients with HIT may test SRA-negative, that is, so-called "SRA-negative HIT." Indeed, we identified such a patient in our study (Figure 3). This patient, who had a high 4Ts score and who tested EIA-positive, demonstrated LIA seroconver-  units. 35 It is important for clinicians to be aware of this important difference between the LIA and EIAs; otherwise, there is a risk of a clinician discounting a "weak" LIA test result as being unlikely to be HIT, whereas we found that approximately one-third of patients with confirmed HIT had "weak" LIA results (between 1.0 and 4.9 U/mL).
In conclusion, our study evaluating LIA seroconversion profiles using archived plasma samples from patients with HIT identified in a historic post-hip replacement surgery trial has shown that positive LIA test results are present at the onset of HIT-related platelet count fall. Further, our study identified a single patient who met criteria for SRA-negative HIT; this patient also demonstrated LIA seroconversion, indicating that LIA testing may be useful for evaluating patients suspected of having SRA-negative HIT.