Impact of laboratory testing for heparin-induced antibodies: using Bayes' rule to prevent overdiagnosis of heparin-induced thrombocytopenia/Bedeutung von Laboruntersuchungen von Heparin-induzierten Antikörpern: Einsatz des Bayes Wahrscheinlichkeitstheorems zur Prävention der Überdiagnose einer Heparin-induzierten Thrombozytopenie
Journal Articles
Overview
Research
Identity
Additional Document Info
View All
Overview
abstract
AbstractHeparin-induced thrombocytopenia (HIT) is a clinical-pathological syndrome, i.e., criteria for diagnosis include a compatible clinical picture and laboratory detectability of heparin-dependent, platelet-activating antibodies of IgG class (“HIT antibodies”), and the lack of a more compelling alternative diagnosis. Heparin administration frequently leads to formation of antibodies of one or several immunoglobulin classes (IgG, IgA, IgM) that recognize a “self” protein, platelet factor 4 (PF4), when PF4 forms multimolecular complexes with heparin. A practical problem is that only a small minority of patients who form heparin-dependent antibodies also develop clinically evident HIT; serum from such patients typically contains IgG antibodies that are strongly platelet-activating. In addition, poorly characterized patient-dependent factors also influence risk of HIT, and thus even a strong positive in vitro test for HIT antibodies does not necessarily mean that HIT will occur. Given the possibility of non-HIT thrombocytopenia among heparin-treated patients, a positive test for heparin-dependent antibodies in such a patient might well lead to a false diagnosis of HIT. One scenario with considerable potential for “overdiagnosis” of HIT is the post-cardiac surgery patient in whom early postoperative thrombocytopenia and/or thrombosis of non-HIT etiology triggers testing for heparin-dependent antibodies a few days later. In this situation, “incidental” seroconversion, rather than confirmation of HIT, is a frequent outcome. This review summarizes the utility of Bayes' rule in making or refuting a diagnosis of HIT. Here, we suggest a pre-test odds of HIT (based on the clinical context) should be revised using the HIT antibody test result – including the strength of any positive result – through an appropriate likelihood ratio. This post-test odds of HIT yields a more reliable assessment of HIT status, potentially minimizing HIT overdiagnosis.
