Portal de Periódicos da CAPES

Immunoassays are not created equal

2009; Elsevier BV; Volume: 7; Issue: 8 Linguagem: Inglês

10.1111/j.1538-7836.2009.03489.x

1538-7933

Theodore E. Warkentin, Lori‐Ann Linkins,

Intramuscular injections and effects

There is an emerging consensus regarding the definition of immune heparin-induced thrombocytopenia (HIT), a ‘clinical–pathological’ syndrome [1Warkentin T.E. Chong B.H. Greinacher A. Heparin-induced thrombocytopenia: towards consensus.Thromb Haemost. 1998; 79: 1-7Crossref PubMed Scopus (558) Google Scholar]. Patients require a clinical picture broadly compatible with this diagnosis, usually an otherwise unexplained platelet count fall (often complicated by thrombosis) that bears a temporal relationship to an immunizing exposure to heparin. Detectability of heparin-dependent platelet-activating antibodies is also a criterion. However, as these antibodies are almost always directed against multimolecular complexes of platelet factor 4 (PF4) bound to heparin [2Greinacher A. Pötzsch B. Amiral J. Dummel V. Eichner A. Mueller-Eckhardt C. Heparin-associated thrombocytopenia: isolation of the antibody and characterization of a multimolecular PF4-heparin complex as the major antigen.Thromb Haemost. 1994; 71: 247-51PubMed Google Scholar], PF4-dependent immunoassays are often used to infer the presence of such platelet-activating antibodies. Indeed, as high-quality platelet activation assays are difficult to perform, and commercial immunoassays are widely available and (relatively) easy to perform, immunoassays are increasingly employed. But how good are the different immunoassays at detecting these pathological ‘HIT antibodies’? This is the topic addressed by a new study by Bakchoul and colleagues [3Bakchoul T. Giptner A. Bein G. Santoso S. Sachs U.J.H. Prospective evaluation of immunoassays for the diagnosis of heparin-induced thrombocytopenia.J Thromb Haemost. 2009; 7: 1260-5Abstract Full Text Full Text PDF PubMed Scopus (0) Google Scholar]. Ideally, an immunoassay should have high sensitivity, as one would not want to miss a diagnosis of HIT. On the other hand, an immunoassay should not detect too many clinically-irrelevant antibodies, as ‘overdiagnosis’ of HIT [4Lo G.K. Sigouin C.S. Warkentin T.E. What is the potential for overdiagnosis of heparin-induced thrombocytopenia?.Am J Hematol. 2007; 82: 1037-43Crossref PubMed Scopus (188) Google Scholar] could have important adverse clinical and even medicolegal consequences. Indeed, heparin exposure causes many patients to form anti-PF4/heparin antibodies that are unable to activate platelets, and thus the potential for overdiagnosis of HIT exceeds that seen in other situations of suspected drug-induced immune thrombocytopenia. Two different types of PF4-dependent immunoassays are available: enzyme-linked immunosorbent assays (ELISAs) and particle-based immunoassays [for review: 5Warkentin T.E. Sheppard J.I. Testing for heparin-induced thrombocytopenia antibodies.Transfus Med Rev. 2006; 20: 259-72Crossref PubMed Scopus (107) Google Scholar]. ELISAs – also known as enzyme-immunoassays (EIAs) – deploy PF4/polyanion antigen onto the plastic surfaces of microtiter plate wells, into which patient serum or plasma is added. After washing, an enzyme-labeled anti-antibody is added, with an enzyme–substrate reaction producing a color change that is used to detect the anti-PF4/heparin antibodies in a semi-quantitative fashion. ELISAs are ‘equilibrium’ assays, as the sequential antibody-antigen reactions are allowed sufficient time to reach an equilibrium state in which antibody binding is maximal prior to their detection. Three ELISAs are commercially available. One assay (from Diagnostica Stago, Asnières sur Seine, France) utilizes recombinant PF4 bound to heparin [6Amiral J. Bridey F. Dreyfus M. Vissoc A.M. Fressinaud E. Wolf M. Meyer D. Platelet factor 4 complexes to heparin is the target for antibodies generated in heparin-induced thrombocytopenia.Thromb Haemost. 1992; 68: 95-6Crossref PubMed Google Scholar]. Another (from GTI Inc., Waukesha, WI, USA) makes use of purified PF4 (obtained from outdated platelets) bound to the non-heparin polyanion, polyvinyl sulfonate [7Visentin G.P. Moghaddam M. Beery S.E. McFarland J.G. Aster R.H. Heparin is not required for detection of antibodies associated with heparin-induced thrombocytopenia/thrombosis.J Lab Clin Med. 2001; 138: 22-31Abstract Full Text Full Text PDF PubMed Scopus (112) Google Scholar]. A more recent assay (Zymutest HIA, from HYPHEN BioMed, Neuville-Sur-Oise, France) uses platelet lysate (source of PF4 and, possibly, other heparin-binding antigenic proteins) [8Amiral J. Vissac A.M. Role of heparin-dependent antigens in immune heparin-induced thrombocytopenia.in: Warkentin TE Greinacher A Heparin-Induced Thrombocytopenia. 4th edn. Informa Healthcare USA, Inc., 2007: 131-47Crossref Google Scholar] to bind to surface-immobilized heparin. In all of these ELISAs, the enzyme-labeled secondary antibody can be directed against all or any particular class of human immunoglobulins, thus allowing for targeted detection of anti-PF4/heparin antibodies of IgG class. This is an important advantage, as only the IgG class of antibodies – which are able to activate platelets through their Fcγ receptors [9Kelton J.G. Sheridan D. Santos A. Smith J. Steeves K. Smith C. Brown C. Murphy W.G. Heparin-induced thrombocytopenia: laboratory studies.Blood. 1988; 72: 925-30Crossref PubMed Google Scholar] – appears to cause HIT. Thus, detection of anti-PF4/heparin IgG likely improves assay specificity without loss of sensitivity [10Lindhoff-Last E. Gerdsen F. Ackermann H. Bauersachs R. Determination of heparin-platelet factor 4-IgG antibodies improves diagnosis of heparin-induced thrombocytopenia.Br J Haematol. 2001; 113: 886-90Crossref PubMed Scopus (86) Google Scholar], and IgG-specific anti-PF4/heparin antibody assays are now commercially available. However, as some investigators have suggested that IgM or IgA antibodies could occasionally cause HIT [11Meyer O. Aslan T. Koster A. Kiesewetter H. Salama A. Report of a patient with heparin-induced thrombocytopenia type II associated with IgA antibodies only.Clin Appl Thromb Hemost. 2006; 12: 373-5Crossref PubMed Scopus (0) Google Scholar], ‘poly-ELISAs’ that detect antibodies of all three major immunoglobulin classes (IgG, IgA, IgM) continue to be marketed. In contrast, particle-based immunoassays detect anti-PF4/heparin antibodies through their agglutination of particles bearing PF4 or PF4/heparin complexes [for review: 5Warkentin T.E. Sheppard J.I. Testing for heparin-induced thrombocytopenia antibodies.Transfus Med Rev. 2006; 20: 259-72Crossref PubMed Scopus (107) Google Scholar]. Here there is no equilibrium state achieved; rather, the antibodies cross-link the PF4-coated particles in a dynamic fashion. Unlike ELISAs, particle-based assays are relatively non-quantitative, as they typically yield a ‘yes–no’ result (although some semblance of quantitation can be achieved by performing the assay at different dilutions of patient serum or plasma). They are also called ‘rapid’ assays, as a result is available within approximately 15 min after preparing patient serum or plasma. One particle-based immunoassay [from DiaMed (Division of Bio-Rad Laboratories), Cressier sur Morat, Switzerland], known as the particle gel immunoassay (PaGIA) [12Meyer O. Salama A. Pittet N. Schwind P. Rapid detection of heparin-induced platelet antibodies with particle gel immunoassay (ID-HPF4).Lancet. 1999; 354: 1525-6Abstract Full Text Full Text PDF PubMed Scopus (100) Google Scholar], uses PF4/heparin complexes affixed to red, high-density polystyrene beads; if anti-PF4/heparin antibodies are present within patient serum or plasma, these (together with a secondary antihuman immunoglobulin akin to a Coombs’ reagent) will produce agglutination of the beads, which prevents passing of the beads through a tube containing sephacryl gel during the subsequent centrifugation step. Previous evaluations of this assay suggest reasonable – though not 100%– sensitivity – and so together with use of a clinical scoring system may be helpful in ruling out the diagnosis of HIT in low or intermediate probability situations [13Pouplard C. Gueret P. Fouassier M. Ternisien C. Trossaert M. Régina S. Gruel Y. Prospective evaluation of the ‘4Ts’ score and particle gel immunoassay specific to hearin/PF4 for the diagnosis of heparin-induced thrombocytopenia.J Thromb Haemost. 2007; 5: 1373-9Abstract Full Text Full Text PDF PubMed Scopus (0) Google Scholar, 14Bryant A. Low J. Austin S. Joseph J.E. Timely diagnosis and management of heparin-induced thrombocytopenia in a frequent request, low incidence single centre using clinical 4T’s score and particle gel immunoassay.Br J Haematol. 2008; 143: 721-6Crossref PubMed Scopus (0) Google Scholar]. Another rapid assay, The ‘HealthTEST Heparin/Platelet factor 4 Antibody Assay’ (Akers Biosciences, Inc., Thorofare, NJ, USA) makes use of ‘Particle ImmunoFiltration Assay’ (PIFA) technology. Here, a membrane filter is used to separate non-agglutinated blue-colored particles from agglutinated ones, and therefore a blue color in the result well indicates a negative test result (no color indicates a positive result). However, this assay fared poorly in comparative studies [15Warkentin T.E. Sheppard J.I. Raschke R. Greinacher A. Performance characteristics of a rapid assay for anti-PF4/heparin antibodies, the particle immunofiltration assay.J Thromb Haemost. 2007; 5: 2308-10Abstract Full Text Full Text PDF PubMed Scopus (29) Google Scholar], although a second generation assay is now marketed. To date, relatively few studies [13Pouplard C. Gueret P. Fouassier M. Ternisien C. Trossaert M. Régina S. Gruel Y. Prospective evaluation of the ‘4Ts’ score and particle gel immunoassay specific to hearin/PF4 for the diagnosis of heparin-induced thrombocytopenia.J Thromb Haemost. 2007; 5: 1373-9Abstract Full Text Full Text PDF PubMed Scopus (0) Google Scholar, 16Eichler P. Raschke R. Lubenow N. Meyer O. Schwind P. Greinacher A. The new ID-heparin/PF4 antibody test for rapid detection of heparin-induced antibodies in comparison with functional and antigenic assays.Br J Haematol. 2002; 116: 887-91Crossref PubMed Scopus (95) Google Scholar] have compared ELISAs directly with particle-based immunoassays. In this issue of the Journal, Bakchoul and colleagues [3Bakchoul T. Giptner A. Bein G. Santoso S. Sachs U.J.H. Prospective evaluation of immunoassays for the diagnosis of heparin-induced thrombocytopenia.J Thromb Haemost. 2009; 7: 1260-5Abstract Full Text Full Text PDF PubMed Scopus (0) Google Scholar] from Justus Liebig University in Giessen, Germany, have performed a comprehensive such study, comparing two ELISAs from GTI (IgG-ELISA, Poly-ELISA) with the PaGIA. Besides providing some insights into the relative performance of these distinct types of immunoassays, their study should also be considered a near-ideal model for how such studies should be performed. First, they used objective clinical–pathological criteria for diagnosis of HIT, namely the use of a scoring system, the 4Ts [17Lo G.K. Juhl D. Warkentin T.E. Sigouin C.S. Eichler P. Greinacher A. Evaluation of pretest clinical score (4 T’s) for the diagnosis of heparin-induced thrombocytopenia in two clinical settings.J Thromb Haemost. 2006; 4: 759-65Crossref PubMed Scopus (762) Google Scholar], to gauge the clinical likelihood of HIT, plus a platelet activation assay (the ‘heparin-induced platelet activation’ test (HIPA) [18Greinacher A. Michels I. Kiefel V. Mueller-Eckhard C. A rapid and sensitive test for diagnosing heparin-induced thrombocytopenia.Thromb Haemost. 1991; 66: 734-6Crossref PubMed Google Scholar], used to detect the presence of heparin-dependent platelet-activating antibodies. Second, they systematically studied sera obtained from a 500-patient cohort with widely varying clinical probabilities of HIT, evaluated prospectively. This allowed comparative assessment of detectability of anti-PF4/heparin antibodies of varying clinical significance, ranging from clearly pathological ones (testing positive in the HIPA) to non-platelet-activating antibodies of IgG and non-IgG isotypes (depending on their reaction profile in the IgG-ELISA and Poly-ELISA). Figure 1 shows that 35 (7%) out of the 500 study patients tested positive in the HIPA test, i.e. the patient’s blood contained heparin-dependent platelet-activating antibodies. This low frequency of HIT antibodies mirrors that found in other studies [5Warkentin T.E. Sheppard J.I. Testing for heparin-induced thrombocytopenia antibodies.Transfus Med Rev. 2006; 20: 259-72Crossref PubMed Scopus (107) Google Scholar, 19Greinacher A. Juhl D. Strobel U. Wessel A. Lubenow N. Selleng K. Eichler P. Warkentin T.E. Heparin-induced thrombocytopenia: a prospective study on the incidence, platelet-activating capacity and clinical significance of anti-PF4/heparin antibodies of the IgG, IgM, and IgA classes.J Thromb Haemost. 2007; 5: 1666-73Abstract Full Text Full Text PDF PubMed Scopus (198) Google Scholar], indicating that only a small proportion of patients who undergo serological investigations for HIT actually have this diagnosis. These patients with platelet-activating antibodies represent a subset among those who have anti-PF4/heparin antibodies of IgG class, per the IgG-ELISA. This latter group itself represents a subset among those testing positive in the Poly-ELISA. These data are consistent with the ‘iceberg model’ of HIT, which views clinically-manifest HIT as the ‘tip of the iceberg’ (protrusion above the waterline) among a much larger group with non-platelet-activating anti-PF4/heparin antibodies. Figure 1 also shows the frequency of a positive PaGIA for four different groups of patients: HIPA-positive (94%); IgG-ELISA-positive but HIPA-negative (51%); Poly-ELISA-positive but HIPA- and IgG-ELISA-negative (37%), and ELISA-negative (5%). The data clearly show that the PaGIA has much lower sensitivity for detecting antibodies that are otherwise detectable by either ELISA. This is not necessarily a disadvantage, as the key question is: what is the sensitivity for detecting clinically-relevant antibodies, for example those that test positive in the HIPA? Here, the sensitivity was only 94%, as two of 35 patients with strong evidence for HIT tested negative in the PaGIA. Thus, the study by Bakchoul and colleagues clearly shows that the PaGIA will occasionally yield a false-negative test result in a patient who has strong evidence for clinical HIT. This is of concern, as high sensitivity is a feature of the competing ELISAs. Moreover, approximately 5% of patients who have no detectable antibody by ELISA will test falsely-positive in the PaGIA, which could lead to unnecessary expense and, perhaps, adverse outcomes in patients who might be treated as presumptive HIT when clearly they do not have this diagnosis. These findings suggest that the PaGIA lacks the sensitivity to be a stand-alone test for ruling out HIT. However, it may be that the combination of a PaGIA result with a 4T’s score would safely exclude HIT (none of the HIPA-positive patients in the current study had a low 4T’s score). This hypothesis is being evaluated in an ongoing prospective clinical trial. The study by Bakchoul and coworkers also shows that the IgG-ELISA is superior to the Poly-ELISA, as its diagnostic specificity is greater (with identical 100% sensitivity) at the manufacturer’s recommended cutoff of 0.40 optical density units (89% vs. 81%). However, by using ‘optimal’ cutoffs for both ELISAs (0.65 for IgG-ELISA and 1.185 for Poly-ELISA), the specificities are similar (92% and 91%, respectively). Another useful feature of the ELISAs (vis-à-vis the PaGIA) is the quantitative information they provide: the greater the optical density value in the ELISA, the greater the probability that the patient has platelet-activating antibodies [20Warkentin T.E. Sheppard J.I. Moore J.C. Sigouin C.S. Kelton J.G. Quantitative interpretation of optical density measurements using PF4-dependent enzyme-immunoassays.J Thromb Haemost. 2008; 6: 1304-12Abstract Full Text Full Text PDF PubMed Scopus (276) Google Scholar]. This increases the diagnostic value of this assay, if employed using a Bayesian approach, in which the pretest probability is estimated from the clinical picture, and the magnitude of a positive ELISA is used to estimate the post-test probability. And given its very high sensitivity, a negative ELISA – unlike a negative PaGIA – essentially rules out the diagnosis of HIT (100% negative predictive value). One final issue: earlier in this commentary, we stated that the study by Bakchoul and colleagues represented a ‘near-ideal’ model for evaluating different assays. Is there any hint of a problem? Figure 1 shows that there were 51 patients who tested negative in the HIPA, but positive in the EIA-IgG: nearly half of these patients (24/51) had a high-probability 4Ts score. This raises an important question: could the ‘gold standard’ HIPA assay itself have failed to detect platelet-activating antibodies in some of these patients? In Germany, testing for platelet-activating antibodies by HIPA is usually performed in transfusion centers using blood obtained from routine blood donors. As platelets obtained from some donors are not readily activated by HIT sera [21Warkentin T.E. Hayward C.P.M. Smith C.A. Kelly P.M. Kelton J.G. Determinants of donor platelet variability when testing for heparin-induced thrombocytopenia.J Lab Clin Med. 1992; 120: 371-9PubMed Google Scholar], the HIPA test is usually performed using platelet from four different donors, with each reaction performed separately. Only two (or more) individual reactions need to yield a positive result for the patient’s blood to be considered to have tested positive for HIT antibodies. However, what if through chance three or even all four of the blood donors’ platelets do not react well to HIT antibodies? Or what if there are other technical limitations that could have limited optimal detection of platelet-activating antibodies? Another platelet activation test – the platelet serotonin-release assay (SRA) [22Sheridan D. Carter C. Kelton J.G. A diagnostic test for heparin-induced thrombocytopenia.Blood. 1986; 67: 27-30Crossref PubMed Google Scholar] – is used in certain North American and French reference centers, and usually employs platelets obtained from donors known to react well to HIT antibodies [21Warkentin T.E. Hayward C.P.M. Smith C.A. Kelly P.M. Kelton J.G. Determinants of donor platelet variability when testing for heparin-induced thrombocytopenia.J Lab Clin Med. 1992; 120: 371-9PubMed Google Scholar]. The SRA also utilizes a hard quantitative end point (percent serotonin release). Given these methodologic differences, it would have been interesting to see how these 51 HIPA-negative/IgG-ELISA-positive sera would have tested in the SRA, performed in another reference laboratory. To date, relatively few ‘serum exchange’ projects have been performed to compare performance of platelet activation assays in different laboratories [23Eichler P. Budde U. Haas S. Kroll H. Loreth R.M. Meyer O. Pachmann U. Pötzsch B. Schabel A. Albrecht D. Greinacher A. First workshop for detection of heparin-induced antibodies: validation of the heparin-induced platelet-activation test (HIPA) in comparison with a PF4/heparin ELISA.Thromb Haemost. 1999; 81: 625-9Crossref PubMed Scopus (0) Google Scholar]. This might be a worthy aim of future studies. Over the past two decades, HIT has evolved from a purely ‘clinical diagnosis’ in which the different laboratory assays were viewed with some skepticism, to the current situation, where the operating characteristics of a large variety of platelet activation assays and different types of immunoassays are becoming increasingly well-defined. Bakchoul and colleagues have given clarity to some of these important diagnostic considerations. The IgG-ELISA is emerging as the current ‘gold standard’ among the immunoassays, although it remains clear that a drawback of this assay remains its frequent detection of non-platelet-activating antibodies and, hence, the potential for considerable overdiagnosis of HIT. T.E. Warkentin author has consulted for GTI Inc. and for Akers BioSciences, and is an investigator in a study for which GTI has provided research funding, and another (together with L.-A. Linkins) for which DiaMed has provided assay kits.