Background
Heparin-induced thrombocytopenia (HIT) is an immune-mediated process that presents with a drop in platelet count related to proximate exposure to unfractionated or low molecular weight heparin. It results from the activation of platelets by heparin-dependent platelet factor 4 (PF4) antibodies, which become detectable 5 days or more after initiation of heparin therapy. 

There is associated increased generation of thrombin in vivo, and thus patients are hypercoagulable even though their platelet counts are low.[1,2] This means HIT often results in major venous and arterial thrombosis. 

The published HIT frequency rates range from 0.5% to 5.0% of all heparin-treated patients. HIT is much more commonly associated with un­fractionated heparin than with low molecular weight heparin (1% to 5% vs <1%). Other risk factors for HIT include recent surgery and female gender.[3,4]

If the diagnosis of HIT cannot be confirmed, then the treatment consists of immediate substitution of unfractionated or low molecular weight heparin with an alternative rapid-acting anticoagulant (e.g., argatroban), a direct thrombin inhibitor, or a factor Xa inhibitor (e.g., fondaparinux).[5]

Diagnosis of HIT is based on two criteria: appropriate clinical features (Table 1 [17]) and a positive laboratory test for platelet-activating HIT antibodies.[6] Clinical features include the following 4Ts: 

•    Thrombocytopenia or a greater than 50% fall in platelet count.
•    Appropriate Timing of the onset of thrombocytopenia after initiation of heparin.
•    Thrombosis, skin necrosis, or systemic reaction with exposure to heparin.
•    Exclusion of other causes of Thrombocytopenia.[7] 

Laboratory diagnosis requires dem­onstration of platelet-activating HIT antibodies in the sera of patients. Tests such as the enzyme-linked immuno­sorbent assay (ELISA) and a commercially available rapid particle gel immunoassay (PaGIA) are very sensitive for heparin-PF4 antibodies, but are not specific for HIT. 

These tests—particularly the ELISA—have a high negative predictive value and thus a negative test is very good at excluding HIT. In contrast, functional tests using washed platelets, such as the platelet 14C-serotonin release assay (SRA) and the heparin-induced platelet activation (HIPA) test, detect heparin-dependent platelet-activating anti­bodies. 

These assays have the highest diagnostic specificity for HIT, but are technically demanding and are offered only by reference centres. Consequently, both these tests have a long turnaround-time. To help determine when such testing is warranted, we designed a 4Ts pretest clinical assessment form and proceeded to study the impact of this form on ordering patterns and the frequency of positive results when testing for heparin-induced thrombocytopenia.  

Methods
A study to determine whether a pretest clinical assessment form could im­prove the utilization of laboratory resources was conducted at Vancouver General Hospital over 2.5 years. In the first 15-month retrospective phase of the study (December 2006 to February 2008), all requests for HIT testing were accepted without screening; patient characteristics and the results of HIT testing during this phase were reviewed by two of the coauthors (CJC and MH). 

The patients’ 4Ts scores were recorded retrospectively and cor­related with test results. In the second 15-month prospective phase (October 2008 to December 2009), the 4Ts form was introduced as a mandatory re­quirement for HIT testing. The changes in the testing pattern were then analyzed.

HIT testing in both study phases was performed using a rapid gel immunoassay (PaGIA, ID-HPF4 from DiaMed, Cressier sur Morat, Switzerland) followed by verification of all samples by the 14C-serotonin release assay, performed at the McMaster University Reference Coagulation La­boratory (Hamilton, ON, Canada). 

Data were analyzed using Minitab release 14.2 (Minitab Inc.), employing the chi-square test, Fisher exact test, and unpaired t-tests, as appropriate.

Results
A total of 145 tests were ordered during the study period, 65 before the mandatory 4Ts pretest form was introduced, and the remaining 80 after the mandatory 4Ts pretest form was introduced. 

Although 15% of patients (10/65) tested positive using the rapid PaGIA in the “before” phase compared with 21% of patients (17/80) in the “after” phase, the difference in the positive tests was not statistically significant. Significance was found, however, in other comparisons. For instance, only 4% of patients (2/53) with a low 4Ts score (< 3) tested positive for heparin-induced thrombocytopenia using rapid PaGIA (Table 2 [18]). 

By contrast, 17% of patients (12/69) with an intermediate 4Ts score (4–5) and 57% of patients (13/23) with a high 4Ts score (6–8) tested positive using the rapid PaGIA. No patients (0/53) with a low 4Ts score tested positive using SRA, while HIT was confirmed using SRA in 9% of patients (6/69) with an intermediate 4Ts score and in 39% of patients (9/23) with a high 4Ts score (P < .001).

The number of patients with a low 4Ts score who were tested for HIT dropped from 52% (34/65) to 24% (19/80)  (P < .001) after the 4Ts pretest form was introduced. Meanwhile, the number of patients with an intermediate 4Ts score who were tested rose from 40% (26/65) to 54% (43/80) (P = NS), and the number of patients with a high 4Ts score who were tested rose from 8% (5/65) to 23% (18/80) (P = .015) after the 4Ts pretest form was introduced.
 
Conclusions
Indiscriminate use of laboratory tests with high sensitivity but low specificity can result in confusion, misdiagnosis, inappropriate treatment, and additional expense. 

Introduction of a pretest based on a probability algorithm can prevent inappropriate testing of patients with low probability for heparin-induced thrombocytopenia, thus improving the specificity of test results, more appropriate treatment, and better utilization of laboratory resources. A good example of this is the use of a pretest clinical assessment with a D-dimer test for the exclusion of suspected deep vein thrombosis.[8] 

In a large European study of HIT, using a 4Ts score reduced false-positive results, thus enhancing the specificity of a very sensitive screening test.[9] In our study, the introduction of a 4Ts pretest clinical assessment form indirectly resulted in physician education: tests were ordered for a greater percentage of patients with a high rather than a low probability of HIT. 

Since we found no patients with an SRA-confirmed diagnosis of HIT in the low range of 4Ts scores (1–3), we support the recommendation made by Pouplard and colleagues[9] that HIT testing should be restricted to patients with intermediate and high 4Ts scores (4 and above). Our current version of the 4Ts pretest form incorporates this recommendation.

Acknowledgments
Dr Warkentin’s research is supported by the Heart and Stroke Founda­tion of Ontario (T.E.W.: #T6950).

Competing interests
Dr Warkentin has received honoraria/speaker fees from Pfizer Canada, Canyon Pharmaceuticals, GlaxoSmithKline, Organon Pharmaceuticals, Sanofi Aventis, and GTI Diagnostics; consultancy fees from medico­legal consultations,  Canyon Pharmaceuticals, GSK, Organon Pharmaceuticals, GTI Diagnostics, and ParinGenix; research support from the Heart and Stroke Foundation, GSK, GTI Diagnostics, and Diamed Lab Supplies.  

This article has been peer reviewed.

Dr Hudoba is a hematopathologist in the Department of Pathology and Laboratory Medicine at Vancouver General Hospital and clinical associate professor at the University of British Columbia. Dr Dalal is a hematopathologist in the Department of Pathology and Laboratory Medicine at VGH and clinical professor at UBC. Dr Warkentin is a professor in the Department of Pathology and Molecular Medicine at McMaster University, Hamilton, Ontario. Mr Pal is a data analyst in the Department of Pathology and Laboratory Medicine at VGH. Dr Chipperfield is a hematopathologist in the Department of Pathology and Laboratory Medicine at VGH and clinical associate professor at UBC. Dr Pi is a hematopathologist in the Department of Pathology and Laboratory Medicine at VGH and clinical associate professor at UBC. Dr Tsang is a hematologist in the Department of Medicine, VGH and clinical associate professor at UBC. Dr Coupland is a hematopathologist and department head of the Department of Pathology and Laboratory Medicine at VGH and clinical professor at UBC. Dr Carter is a hematopathologist, coagulation director in the Department of Pathology and Laboratory Medicine at St. Paul's Hospital and associate professor at UBC.