Platelet count monitoring and laboratory testing for heparin-induced thrombocytopenia.

Abstract

To the Editor.—Congratulations on the interesting data of the College of American Pathologists Consensus Conference XXXVI: Diagnostic Issues in Thrombophilia, which appeared in the November issue of the Archives. The recommendations in general are well balanced. However, I was surprised by Dr Warkentin's statements on laboratory testing for the antibodies that cause heparin-induced thrombocytopenia (HIT).1Dr Warkentin correctly states that clinical and pathologic criteria should be present to assure a diagnosis of HIT. However, the data presented on the sensitivity and specificity of the different presently available antibody tests are misleading. While the sensitivity of most tests may be relatively high, it is difficult to understand why specificity is claimed to be around 90%. Since there is no standard test available to obtain comparative data on specificity and since the fall in platelet count certainly is not specific for HIT, the figures on specificity cannot be accepted. It remains unclear in the article how Dr Warkentin defines specificity.As far as the diagnostic specificity is concerned, Dr Warkentin correctly states that about 15% of orthopedic surgery patients develop HIT antibodies, while only 5% develop clinical HIT, and although 50% of cardiac surgery patients form HIT antibodies, clinical HIT is observed only in about 2% to 3%. We have obtained similar data in patients treated for venous thrombosis.2 Thus, the diagnostic specificity of the present tests varies between 2% and 30%.Clinicians are mainly interested in a test that is specific for the suspected disease, in this case, HIT. It seems clear also from Dr Warkentin's data that the presently available tests have no or very little diagnostic value and that they cannot be used to establish the diagnosis of HIT. Dr Warkentin correctly states that physicians should not wait for the results of antibody tests before starting alternative anticoagulation in patients with clinically suspected HIT. A simple statement that the presently available antibody tests are not helpful in the diagnosis of HIT would be more accurate.Possibly, antibody tests will be improved in the future to become more specific for the clinical diagnosis of HIT. At present, clinicians have to rely on the drop of platelet count and on clinical symptoms to suspect HIT, even though in some cases the drop of platelet count is not specific for HIT.