False-negative immunoassays in Heparin Induced Thrombocytopenia

Heparin-induced thrombocytopenia (HIT) is a rare but serious complication of heparin therapy characterized by antibodies against complexes of platelet factor 4 (PF4) and heparin.[1, 2] Clinically, HIT manifests as thrombocytopenia with or without concurrent thrombosis.[2] Diagnosis typically relies on identifying anti-PF4/heparin antibodies by enzyme immunoassays (EIAs) followed by confirmation with platelet activation studies, such as the serotonin release assay (SRA), P-selectin expression assay (PEA)[3] or heparin-induced platelet activation (HIPA) assay.[1, 4] EIAs have low specificity, due to the high frequency of asymptomatic anti-PF4/heparin antibodies after heparin exposure, but they are extraordinarily sensitive and reliably exclude HIT when negative. [5] As the SRA has both high sensitivity (100%) and specificity (97%),[6] EIA false negative (EIA-FN) cases of HIT, defined as a negative immunoassay and a positive SRA, are exceedingly rare.[7, 8]

Clinicians at our institution have recently reported anecdotal cases of EIA-FN HIT. To investigate the occurrence of EIA-FN HIT and document its prevalence, we reviewed the electronic health record (EHR) at our academic medical center over a ten-year period to identify SRA-positive cases that were missed by anti-PF4/heparin IgG assays.

Laboratory records for Duke University Hospital patients from 01/01/2015–01/01/2025 were reviewed for anti-PF4/heparin antibody testing under an IRB-approved protocol (Duke IRB#:Pro00012901). The Duke Coagulation Laboratory provides on-site EIAs for poly-specific anti-PF4/heparin antibodies (detecting IgG, A, M isotypes) with reflex testing for IgG if the poly-specific assay is positive. The commercial assays used were the HemosIL^® HIT-Ab_[PF4-H] (positive >1 U/ml; Werfen, Bedford, MA) and the Hyphen BioMed Zymutest HIT IgG (positive >0.5; Aniara Diagnostica, West Chester, OH), using manufacturer’s cut-offs. Functional testing by SRA was requested by the treating providers and not reflexively ordered for patients with a positive immunoassay and was performed by LabCorp (Burlington, NC).

A review of Duke’s EHR over the ten-year period identified 6,769 patients who underwent poly-specific anti-PF4/heparin antibody screening (HIT screening). In this cohort, there were 5,873 negative HIT screening samples while 1,200 samples (20%) showed positivity. Reflex testing revealed positive anti-PF4/heparin IgG antibodies in 584 patients, while 616 were IgG negative. 342 patients with negative IgG testing subsequently had SRA testing, which was positive in six patients (~ 1.75% of IgG-negative patients tested by SRA, 0.5% of poly-specific-positive patients; and 0.09% of all screened patients). These patient records were manually reviewed and were deemed eligible for inclusion in the final case series (Figure 1).

Figure 1 –

Search Strategy and patient identification. Abbreviations: HIT = heparin induced thrombocytopenia.

Medical records of the six IgG-negative/SRA-positive patients were reviewed for medication history (including heparin), laboratory results, and thrombosis to determine the presence of clinical HIT. Table 1 summarizes the demographic, clinical, and laboratory information. Average age was 58 years old (range: 53–86 years), and patients were predominantly male (4M/2F). They were treated with heparin for underlying cardiac disease or thromboembolic disease, with four of the six patients having prior heparin exposure (1.5–8 years before index hospitalization). All six cases were suspected of HIT based on intermediate to high clinical 4T’s scores (4–8) and thrombocytopenia (mean % platelet count decline: 84%; range: 66–96%), with three of the six patients experiencing thrombosis. Poly-specific anti-PF4/heparin antibodies were positive in all patients (average: 1.70 U/mL, range: 1.28–2.04), with negative IgG testing. SRA testing revealed high positivity with low-dose (LD) heparin (mean SRA ± standard deviation: 76 ± 14%) and decreased serotonin release at high doses (5/6 patients with <1% and 1 patient with 18%).

Table 1 –

Patient / Clinical Characteristics. D0 indicates day of initial heparin exposure, with D# indicating number of days following initial exposure.

Patient	Age & Sex	Heparin indication	Prior Heparin Exposure (Y/N)	Platelet Count (Pre-Heparin)	Platelet Nadir (% decrease)	4T Score	HIT Screening Day	HIT Screen Polyspecific+ IgG Result	SRA Test Day (Result Day)	SRA Result	Re-Exposed to Heparin (Y/N)	Complications / Outcomes	Alternative Anti-Coagulant Used
1	82 M	VTE PPx: SQH D0 CPB: IV UFH D3	N	157	21 (87%)	6	D8	Poly: 1.655 IgG: 0.049	D11 (D15)	LD: 88% HD <1%	Y: UFH given for 6d following negative IgG prior to SRA result	DVT diagnosed D9; IVIG given for presumed immunologic thrombocytopenia. Discharged home.	Bivalirudin D8–D9, D15-D19 Apixaban D19 – D/C
2	38 M	VTE PPx: LMWH D0 IABP: IV UFH D2, followed by UFH gtt D2–D9. ECMO: UFH gtt D11-D16	Y: UFH 2y prior for cardiac surgery.	188	8 (96%)	4	D9 D16	Poly: 1.422 IgG: 0.133 Poly: <0.600	D10 (D16)	LD: 51% HD: 18%	Y: UFH used during heart & renal transplant surgeries	Cannulated for ECMO D8 due to decompensated heart failure; fibrin strands developed in ECMO circuit while on UFH. Received Heart Transplant D10 and Renal Transplant D12, both with UFH. Persistent thrombocytopenia and organ dysfunction thereafter. Transitioned to comfort care & died in the ICU.	Bivalirudin D9–11, D16 – Death
3	51 M	PE: UFH gtt D0–D10	Y: SQ UFH 6y prior for VTE PPx.	183	51 (72%)	4	D10	Poly: 2.040 IgG: 0.163	D14 (D19)	LD: 77% HD: <1%	N	None. Discharged home.	Argatroban D11 – D25 Warfarin D17 – D/C
4	43 F	VTE PPx: LMWH D0 Intracardiac thrombi: SQH D1 ECMO: IV UFH D6 followed by UFH gtt D6 -D9	Y: SQH 1.5y prior for VTE PPx.	210	27 (87%)	8	D9	Poly: 1.289 IgG: 0.138	D9 (D15)	LD: 71% HD: <1%	N	PEA cardiac arrest D6 resulting in ECMO cannulation; PE diagnosed D8 requiring pulmonary thrombectomy. Eventual Recovery with discharge to ARF.	Bivalirudin D9 – D19 Apixaban D19 – D/C
5	86 F	Atrial fibrillation: UFH gtt D0 - D1, D8 – D15 VTE PPx: LMWH D5–D8	Y: UFH 8y prior for acute PE.	287	99 (66%)	5	D14	Poly: 1.716 IgG: 0.363	D15 (D25)	LD: 76% HD: 7%	N	Multiple RRTs due to hypotension and hypercapnic respiratory failure. No thrombotic complications. Eventual recovery with discharge to SNF.	Bivalirudin D14 – D21 Apixaban D2–D5, D21 – D/C
6	49 M	VTE PPx: SQH D0 - D4 PE / ECMO: UFH D4 – D6	N	356	17 (95%)	6	D6	Poly: 2.049 IgG: 0.300	D17 (D21)	LD: 91% HD: <1%	N	PE diagnosed D4. Cannulated for ECMO due to RV Failure D4. SRA sent D6 not performed due to lab error, re-sent D17. Required multiple platelet transfusions. Eventual recovery with discharge home.	Bivalirudin D6 – D30 Warfarin D27 – D/C

Abbreviations: ARF = acute rehab facility; CPB = cardiopulmonary bypass; D/C = discharge; DVT = deep vein thrombosis; ECMO = extracorporeal membranous oxygenation; gtt = continuous IV infusion; HD = High Dose Heparin; IABP = Intra-aortic Balloon Pump; ICU = intensive care unit; LD = Low Dose Heparin; LMWH = low molecular-weight heparin; M/F = Male/Female sex; Poly = poly-specific HIT screening test (IgG/M/A); PE = pulmonary embolism; PEA = pulseless electrical activity; PPx = prophylaxis; RRT = rapid response team; SNF = skilled nursing facility; SQH = subcutaneous heparin; SRA = Serotonin Release Assay, UFH = unfractionated heparin; VTE = venous thromboembolism

These six EIA-FN HIT cases represent 1.75% of IgG- and SRA-tested samples (n=342), which may be an overestimate of the true prevalence, as not all IgG- samples (n=616) or screened samples were tested by SRA. Based on recent studies, the prevalence of EIA-FN HIT is exceedingly low; estimates vary between 0.02–0.7%. [7–9] In a survey of 1017 samples tested by EIA and SRA, Shelat and colleagues observed 7 EIA-negative/SRA-positive cases (0.7%), though clinical information was not available to determine if these cases had HIT or not.[9] In a review of 8456 cases, Warkentin et al. identified only 16 EIA-FN patients (0.2%) with only 1 of the cases confirmed to have HIT on chart review, while others were found to be have lab/clerical error (7/16) or low likelihood of HIT (8/16). [7] In a recently published series, Kanack et al. described a prevalence of 0.6%, identifying 3 patients with HIT among 500 EIA- patients.[8] In this report, the authors extensively characterized these 3 patients as well as an additional 3 EIA- HIT patients by additional serologic and functional assays. Reactivity was variable when tested with other immunoassay platforms, with 2/5 positive by the Lifecodes PF4 IgGAM (Immucor) and 5/6 were positive on the platelet lysate/heparin IgG-specific ELISA-Zymutest HIA IgG immunoassay (Hyphen BioMed). However, these samples were more reactive in functional assays in the presence of PF4, with all patients showing reactivity in the PEA assay and 3/3 patients showing PF4-dependent release of thrombospondin 1 from cryo-preserved platelets.[8] Recent studies by Zhou et al., showing distinct population of HIT antibodies that are reactive with PF4 bound to platelets rather than PF4 bound to microtiter plates, and not cross-reactive with IL-8 or NAP-2, further reinforce the observations that EIA-negative/ SRA-positive HIT cases may represent this class of PF4/platelet reactive antibodies.[10]

Similar to studies by Kanack et al., our case series demonstrates the occurrence of EIA-negative patients with true HIT. Although anti-PF4/heparin IgG was low in these patients, their presentations were nonetheless consistent with HIT based on clinical and laboratory features, including intermediate to high 4T scores (4–8), significant platelet decline (>80% platelet count decline), high serotonin release (mean SRA: 76%) with LD heparin and thrombotic complications (4/6). Based on the false-negative IgG, two patients were re-exposed to UFH, with both patients experiencing worsening clinical outcomes as a result. In three cases, the diagnosis of HIT was pursued based on clinical suspicion, while in the three others HIT was reconsidered based on discordant EIA/SRA results.

We acknowledge several limitations of this study, including the retrospective single institution design, evaluation of only IgG-negative/SRA-positive cases rather than poly-specific negative/SRA positive cases, and lack of correlative testing to explain the negative serologies in these patients. Given the heterogeneity of immunoassays for HIT testing and the marked variability in diagnostic approaches to HIT testing at most medical centers,[11] multi-institutional studies may be challenging to perform. By design, our study was restricted to only IgG-negative/SRA-positive cases, as nearly all cases with negative poly-specific results do not undergo further testing, and therefore such a study could not be practically performed. Of the 5,873 patients with negative poly-specific testing, only 335 had subsequent testing (either IgG or SRA) which was negative in all cases. Lastly, due to the lack of stored patient samples, we were unable to further investigate the mechanisms by which IgG-patients could elicit platelet activation, possibly in a PF4-dependent manner as shown in recent studies. [8, 10]

Despite these study limitations, our case series adds to the growing awareness of rare cases of EIA-negative/SRA-positive cases of HIT.[7–9] While the rarity of these cases reinforce published data demonstrating the high negative predictive value of immunoassays for HIT, the findings of adverse outcomes due to heparin re-exposure highlight the need for continued clinical vigilance. For such cases, where clinical suspicion remains high in the setting a negative IgG immunoassay, we recommend additional testing with PF4-dependent platelet-activating assays.