Written by Dr. Roupen Odabashian MD, FRCPC
Hematologist-Oncologist | Founder, MeDucation AI | Updated November 2026
Conflict of interest, up front: I am a practicing hematologist-oncologist and I founded MeDucation AI, which sells a physician-written heme/onc question bank and other board-prep tools. Nothing below depends on that product. The clinical content here is drawn from ASH's 2018 VTE guideline on HIT and its companion pocket guide, FDA drug labels, and peer-reviewed diagnostic accuracy studies, all cited by number.
Step | What to do | Key number |
|---|---|---|
1. Suspect | Calculate the 4T score in any patient on or recently exposed to heparin with a falling platelet count | Score 0-8 |
2. Risk-stratify | Low (0-3), intermediate (4-5), high (6-8) probability | Low-probability 4T has a negative predictive value of 0.998[1] |
3. Act before results | Stop heparin, start a non-heparin anticoagulant if intermediate/high probability | ASH recommends this even before antibody testing returns[2] |
4. Confirm | PF4/heparin immunoassay, then a functional assay (SRA or HIPA) if immunoassay positive | Immunoassay sensitivity >95%; functional assay specificity 90-98%[2] |
5. Treat | Argatroban, bivalirudin, danaparoid, fondaparinux, or a DOAC | Treat 3-6 months if thrombosis present[3] |
What is HIT in medical terms, and why does the timing matter?
HIT is a prothrombotic, antibody-driven drug reaction, not a simple drop in platelets from heparin toxicity[4]. IgG antibodies form against complexes of platelet factor 4 and heparin, and those immune complexes activate platelets through FcγRIIA receptors, triggering platelet aggregation and a hypercoagulable state[5]. The clinical signature is a triad: thrombocytopenia, a tendency toward severe thrombosis, and a characteristic timing of the platelet fall after heparin exposure[5]. One third to one half of HIT cases are complicated by thrombosis, which can be limb- or life-threatening[4], and in untreated patients roughly half go on to develop severe thromboembolism, with high associated mortality[5]. That combination, a common drug exposure plus a rare but dangerous antibody reaction, is exactly why fellows are tested on this so often: the diagnosis has to be made on clinical grounds, fast, before confirmatory serology is back.
How does the 4T score work, and what do the points mean?
The 4T score grades four features to produce a pretest probability of HIT[1][2]. It was built to replace gestalt judgment with a structured estimate, and ASH's guideline panel explicitly recommends using it rather than clinical impression alone[2].
Feature | 2 points | 1 point | 0 points |
|---|---|---|---|
Thrombosis or other sequelae | New confirmed thrombosis, skin necrosis at injection site, anaphylactoid reaction to IV bolus, adrenal hemorrhage | Progressive/recurrent thrombosis, non-necrotizing skin lesions, suspected thrombosis | None |
Other causes of thrombocytopenia | None apparent | Possible | Definite |
Total scores sort into low, intermediate, or high probability categories[2]. A systematic review and meta-analysis by Cuker and colleagues validated the score's main strength on the low end: a low-probability 4T score carries a negative predictive value of 0.998 (95% CI 0.970-1.000)[1], meaning it is genuinely good at ruling HIT out. It is much weaker at ruling HIT in. In that same analysis, the positive predictive value of an intermediate score was only 0.14, and even a high-probability score was 0.64[1]. Other series report the 4T model's positive predictive value in the 9-17% range[1]. A large retrospective diagnostic study found the 4T score alone produced an area under the ROC curve of 81.3%, versus 97.7% for a chemiluminescent immunoassay[5]. ASH's own guideline panel is blunt about the consequence: missing or inaccurate input data can produce a faulty score and the wrong management decision, so when information is uncertain, the panel advises erring toward a higher score[2]. A separate ICU-based retrospective study found that 8.6% of patients with a low-probability score were still HIT-positive on lab testing, and 57% of patients scored high-probability (6-8) turned out to be HIT-negative, which is the clearest argument on record for never treating or testing on the 4T score alone[1].
What is the HIT diagnosis workflow: PF4 ELISA vs serotonin release assay?
The confirmed-diagnosis pathway is a two-step algorithm: an immunoassay first, then a functional assay if that immunoassay is positive[2]. For patients with a low-probability 4T score, ASH recommends against ordering HIT lab testing at all, unless there is real uncertainty about the score itself, for example from missing data[2]. For intermediate- or high-probability patients, the panel recommends an immunoassay (PF4/heparin ELISA or a rapid immunoassay), and if that is positive and a functional assay is available, suggests confirming with the functional assay[2].
The two assay families test different things. Immunoassays, including polyspecific ELISA, IgG-specific ELISA, and rapid immunoassays such as chemiluminescence, lateral flow, latex agglutination, or particle gel immunoassay, detect the presence of antibodies against PF4/heparin complexes regardless of whether those antibodies actually activate platelets; sensitivity for IgG-specific and polyspecific ELISA exceeds 95%, while rapid immunoassays run 80-90%[2]. Functional platelet activation assays, the serotonin release assay (SRA) and the heparin-induced platelet activation assay (HIPA), instead detect antibodies that functionally induce heparin-dependent platelet activation, with reported sensitivity 90-98% and specificity 90-95%[2]. The washed-platelet HIPA, like the SRA, is recognized as the reference standard by major scientific societies[5]. Functional assays are not widely available and typically require referral to a reference lab[2].
