There Is No Such Thing as a “Positive” Antibody Test

Abstract

“We had a new consult late last night, Dr Rice. Heparin-induced thrombocytopenia.”“Really, I wasn't called. Did you have to go in?”“No, no. When I heard the OD was 0.6, I asked them not to call you. They had already started argatroban and said we could see the patient in the morning.” The 65-year-old man with hypertensive renal disease had been on hemodialysis for 3 years. Three days earlier, he underwent procedures for coronary artery bypass grafts and placement of a bioprosthetic aortic valve. Admission platelet count was 210 × 103; now it was 105 × 103. Extubated that morning, he was considered to be doing well. I reviewed the chart while my fellow and the nurse were peeling the “No Heparin” sign off the head of the bed. The onset of heparin-induced thrombocytopenia (HIT) occurs 5 to 12 days after beginning heparin exposure. Patients may develop HIT in the first 2 weeks after dialysis initiation, but it is very unlikely to occur later with the chronic heparin exposure. (In relatively rare cases, the massive heparin and platelet factor 4 release during cardiac surgery “resets the clock.”) Heart surgery may be the most common precipitating factor for HIT, but there are challenges to establishing the diagnosis in this scenario.1Rice L Heparin-induced thrombocytopenia in the cardiac patient: 10 points to help the physician.Methodist Debakey Cardiovasc J. 2011; 7: 6-9Crossref PubMed Scopus (4) Google Scholar One challenge is that on-pump heart surgery is expected to lower platelets 25% to 50% with a nadir 2 to 3 days post operation.2Nader ND Khadra WZ Reich NT Bacon DR Salerno TA Panos AL Blood product use in cardiac revascularization: comparison of on- and off-pump techniques.Ann Thorac Surg. 1999; 68: 1640-1643Abstract Full Text Full Text PDF PubMed Scopus (112) Google Scholar So-called “dilutional thrombocytopenia” likely results from platelets battered in the extracorporeal circuit being removed from the circulation before megakaryocytes react. With long pump runs (such as with valve replacement), platelet counts may fall lower. The patient clearly had a very low risk of HIT (and did not need argatroban). Key clinical features of HIT were recognized 40 years ago: Platelets fall 1 week after beginning heparin, accompanied by an extreme risk of thrombotic complications, arterial or venous. Like case series in the 1950s and 1960s reporting repetitive thromboses in patients receiving increasingly more heparin, I collected 50 patients over 10 years in my medical center with catastrophic thrombotic diatheses caused by HIT.3Rice L Huffman DM Waddell CC Luper WE Udden MM Levine ML Therapy of thromboembolic disease: the heparin thrombocytopenia/thrombosis syndrome.in: Thrombosis, Anticoagulants and Antiplatelet Agents in Clinical Practice. Park Row Publishers, New York, NY1988: 31-36Google Scholar My worst nightmare did not anticipate that I would soon see a similar number of these patients every year.4Rice L Cases of heparin-induced thrombocytopenia elucidate the syndrome.Chest. 2005; 127: 21S-26SAbstract Full Text Full Text PDF PubMed Scopus (10) Google Scholar A small number of we HIT-ophiles took to the road to lecture at Grand Rounds and any available venues, spreading the word that this problem was not rare and was being routinely missed with disastrous consequences to patients. (Were we overly successful?). The antibody-mediated pathogenesis of HIT was clear from the beginning. When platelet factor 4 modified by heparin was discovered to be the relevant antigen, “antigen immunoassays” quickly supplanted previously available functional assays based on heparin-induced platelet aggregation or serotonin release (SRA). Immunoassays are rapid, reproducible, and inexpensive. Their negative predictive value is 98% to 100% using the traditional optical density (OD) cutoff of 0.4.5Warkentin TE Sheppard JI Moore JC Sigouin CS Kelton JG Quantitative interpretation of optical density measurements using PF4-dependent enzyme-immunoassays.J Thromb Haemost. 2008; 6: 1304-1312Crossref PubMed Scopus (278) Google Scholar Unfortunately, in the good majority, “positive” tests do not have platelet-activating antibodies confirmed. The positive predictive value of enzyme-linked immunosorbent assays (ELISAs) depends on the prevalence of true HIT in the population to which they are applied, but generally only 10% to 30% of “positive” patients have HIT. Studies have shown that the degree of positivity (ie, the height of the OD) strongly correlates with true HIT and thrombotic risks. This became immediately clinically clear to us, so for well over a decade we have insisted that our laboratory report out OD values. HIT is a clinicopathologic diagnosis: Patients with suspicious clinical manifestations should have laboratory tests to eventually confirm the presence of pathogenic antibodies. When HIT is reasonably suspected, all heparin should be discontinued, alternative anticoagulants started, and antibody test results awaited. With moderate to high clinical suspicion, holding heparin is inadequate because of the high rate of early thrombotic complications.6Rice L Heparin-induced thrombocytopenia: myths and misconceptions (that will cause trouble for you and your patient).Arch Intern Med. 2004; 164: 1961-1964Crossref PubMed Scopus (74) Google Scholar To aid physicians on when to suspect HIT, scoring systems have been validated; the 4T system is the most widely used.7Cuker A Gimotty PA Crowther MA Warkentin TE Predictive value of the 4Ts scoring system for heparin-induced thrombocytopenia: a systematic review and meta-analysis.Blood. 2012; 120: 4160-4167Crossref PubMed Scopus (307) Google Scholar Points (0-2) are awarded based on whether the platelet decline is typical, whether the timing is right, whether concomitant thrombotic events have occurred, and whether there are other likely explanations for the thrombocytopenia. Patients are divided into high, intermediate, or low probability. Similar to antibody testing, low probability clinical scoring has a very high negative predictive value for HIT (> 99%)7Cuker A Gimotty PA Crowther MA Warkentin TE Predictive value of the 4Ts scoring system for heparin-induced thrombocytopenia: a systematic review and meta-analysis.Blood. 2012; 120: 4160-4167Crossref PubMed Scopus (307) Google Scholar; such patients require no antibody testing and may continue heparin therapy as needed. When told 10 years ago that some medical centers were experiencing overdiagnosis, I laughed, but I do not laugh anymore. The epidemic of overdiagnosis8Cuker A Cines DB How I treat heparin-induced thrombocytopenia.Blood. 2012; 119: 2209-2218Crossref PubMed Scopus (113) Google Scholar is such that when I am called about HIT suspicion, only one in three patients may actually have it. It is easy to appreciate the factors underlying this epidemic. Physicians have become attuned to the problem, whether from publications, lectures, experiencing a catastrophe or two in their own patients, or even from involvement in lawsuits (there have been plenty); they are imbued with the need not to miss this diagnosis. The availability of serologic tests for HIT (just check a box) and the high “false-positive” rates of the tests have fed overdiagnosis. Face it, poorly informed and/or lazy physicians would like to believe a test would give them definitive results, rather than that they have to learn about disease characteristics, about when suspicion is appropriate, and about properly interpreting test results. Overdiagnosis is further fueled by the lack of appreciation of the serious negative ramifications this diagnosis conveys. The economic costs associated with HIT overdiagnosis are substantial.9Smythe MA Koerber JM Fitzgerald M Mattson JC The financial impact of heparin-induced thrombocytopenia.Chest. 2008; 134: 568-573Abstract Full Text Full Text PDF PubMed Scopus (27) Google Scholar IV direct thrombin inhibitors are not only costly, but the risks of bleeding have been as high as 20%. I have seen life-threatening bleeding when doctors unfamiliar with these agents and their nuances have felt forced to use them in patients who really did not require them. Patients have had urgent surgeries delayed, have been culled from transplant-eligible lists, and have been denied acceptance to the outpatient dialysis units or to the rehabilitation or long-term care units they need. Patients who never really had HIT have been “branded,” denying them optimal anticoagulant therapy for future problems. In this issue of CHEST (see page 55), Chan et al10Chan CM Woods CJ Warkentin TE Sheppard J-AI Shorr AF The role for optical density in heparin-induced thrombocytopenia: a cohort study.Chest. 2015; 148: 55-61Abstract Full Text Full Text PDF PubMed Scopus (40) Google Scholar explore the use of an OD cutoff of 1.0 instead of the traditional 0.4 to determine a “positive” HIT ELISA. In about 500 patients who had heparin-antibody tests ordered, none with an ELISA OD between 0.4 and 1.0 had positive platelet-activating antibodies on SRA testing, so the positive predictive value of the test was improved with no sacrifice to the 100% negative predictive value. The findings are not unexpected given that similar results have been obtained in hundreds of patients studied retrospectively.5Warkentin TE Sheppard JI Moore JC Sigouin CS Kelton JG Quantitative interpretation of optical density measurements using PF4-dependent enzyme-immunoassays.J Thromb Haemost. 2008; 6: 1304-1312Crossref PubMed Scopus (278) Google Scholar Yet the prospective data of Chan et al10Chan CM Woods CJ Warkentin TE Sheppard J-AI Shorr AF The role for optical density in heparin-induced thrombocytopenia: a cohort study.Chest. 2015; 148: 55-61Abstract Full Text Full Text PDF PubMed Scopus (40) Google Scholar are important to assure physicians that “weak-positive” ELISA results do not suggest HIT and actually greatly mitigate the possibility of this diagnosis. These data should certainly help calm the overdiagnosis epidemic. One should not, however, be misled that there is zero trade-off of sensitivity. Although these authors encountered no one with true HIT with an OD between 0.4 and 1.0, I and others certainly have. In the past 2 years, I have seen three such patients; one with florid fatal HIT had an OD of 0.8, a very high 4T score, and a positive SRA. Such patients are not the rule, and raising the OD for positive reporting would certainly improve the overdiagnosis rate, but Chan et al10Chan CM Woods CJ Warkentin TE Sheppard J-AI Shorr AF The role for optical density in heparin-induced thrombocytopenia: a cohort study.Chest. 2015; 148: 55-61Abstract Full Text Full Text PDF PubMed Scopus (40) Google Scholar should be careful of claiming no loss of sensitivity in a study limited by the fact that only 10 patients (out of 500 tested) were felt to have true HIT. Recent studies have called attention to the fact that HIT antibody tests are being ordered “willy-nilly” in patients with a very low probability of the disorder. Having surveyed my own laboratory and those in other major medical centers and referral centers, I find that, just as Chan et al10Chan CM Woods CJ Warkentin TE Sheppard J-AI Shorr AF The role for optical density in heparin-induced thrombocytopenia: a cohort study.Chest. 2015; 148: 55-61Abstract Full Text Full Text PDF PubMed Scopus (40) Google Scholar report, only 10% of samples tested for HIT have a positive ELISA even using 0.4 as the OD cutoff; only 2% have a positive SRA for platelet-activating antibodies. HIT antibody tests are not designed for “screening” in any population. Whenever studied for screening (eg, prior to cardiovascular surgery), they have not been beneficial. Care teams or pharmacists should not walk bed to bed in an ICU calculating 4T scores, finding “I get a 4 on this one, send the test.” Rather, testing is appropriate only when HIT is reasonably suspected. As in other recent publications, we surveyed 131 consecutive heparin antibody tests ordered in the ICUs and found that 85% had a low pretest probability (4T score ≤ 3); chart review revealed no heparin exposures at all in 27%.11Wanat M Fitousis K Hall J Rice L PF4/heparin antibody testing and treatment of heparin-induced thrombocytopenia in the intensive care unit.Clin Appl Thromb Hemost. 2013; 19: 297-302Crossref PubMed Scopus (8) Google Scholar Few of these tests should ever have been ordered. So, do I support the call from Chan et al10Chan CM Woods CJ Warkentin TE Sheppard J-AI Shorr AF The role for optical density in heparin-induced thrombocytopenia: a cohort study.Chest. 2015; 148: 55-61Abstract Full Text Full Text PDF PubMed Scopus (40) Google Scholar to raise the positive OD cutoff to 1.0? No, I do not. We do not report an ANA test “positive” for lupus or a rheumatoid factor as “positive for rheumatoid arthritis,” nor should heparin antibodies be denoted as “positive” or “negative.” Rather, the numerical titer should be reported and interpreted by the physician in the clinical context of his patient. So then, how do you diagnose HIT in 2015? •Consider HIT whenever a patient exposed to heparin has an unexpected fall in platelet count, new thrombotic events, or both.•Timing is critical. HIT occurs 5 to 12 days after beginning heparin exposure (but delayed onset with recent exposure or rapid onset with reexposure should also be considered).•The test is your friend, but only when ordered appropriately with reasonable suspicion for HIT (it is not for “screening”), and it must be interpreted appropriately. Low titer (less than OD 1.0) almost always rules out HIT; moderate titer must be interpreted in the clinical context. In 2015, there is much to celebrate about the occurrence of HIT. Even as awareness and concern peak, the incidence of true HIT is declining, owing largely to the decreasing use of unfractionated heparin (more low-molecular-weight heparin and nonheparin anticoagulants are being used) and the downturn in cardiovascular surgeries. Safer and more refined anticoagulant strategies for HIT are coming to the fore, and even nonanticoagulant interventions are under study.12Joglekar MV Quintana Diez PM Marcus S et al.Disruption of PF4/H multimolecular complex formation with a minimally anticoagulant heparin (ODSH).Thromb Haemost. 2012; 107: 717-725Crossref PubMed Google Scholar Diagnostic assays are emerging with higher sensitivity/specificity profiles and real-time availability,13Cuker A Rux AH Hinds JL et al.Novel diagnostic assays for heparin-induced thrombocytopenia.Blood. 2013; 121: 3727-3732Crossref PubMed Scopus (36) Google Scholar, 14Padmanabhan A Jones CG Bougie DW et al.Heparin-independent, PF4-dependent binding of HIT antibodies to platelets: implications for HIT pathogenesis.Blood. 2015; 125: 155-161Crossref PubMed Scopus (63) Google Scholar but it remains difficult to conceive of a test that would not require careful consideration of when to order it and how to interpret it. If we can address the overdiagnosis problem, some of we HIT-ophiles may begin to contemplate retirement.