Heparin-induced thrombocytopenia (HIT), a relatively uncommon but difficult-to-diagnose condition, carries patient safety risks both from delays in diagnosis and from treatment. Commonly used enzyme-linked immunosorbent assays (ELISA) for HIT have low specificity, but higher specificity functional assays are not available in many labs, thereby increasing processing turnaround times and exposing some patients unnecessarily to costly and risky direct thrombin inhibitor (DTI) therapy. Researchers recently developed and implemented a comprehensive HIT protocol aimed at improving HIT-related outcomes, and their findings are the subject of this issue of Strategies.

HIT, an immune-mediated reaction to heparin, occurs in an estimated 0.2–5% of patients who receive heparin, but identifying them can be tricky, according to Maureen Smythe, PharmD, a clinical professor of pharmacy practice at Wayne State University and a pharmacist and coordinator of student and resident education at Beaumont Hospital in Royal Oak, Mich. "The diagnosis of HIT is confounding in that thrombocytopenia and thrombosis are the hallmark clinical presentation findings, but these can be caused by other etiologies. So you have to evaluate each patient for a specific temporal relationship between the development of these findings and exposure to a heparin product," she explained. "Ideally you'd like the clinical presentation to match the established temporal relationship with heparin exposure and you'd like to confirm the presence of HIT antibodies using serology testing."

Clinicians typically suspect HIT when a patient experiences an otherwise unexplained ≥50% drop in platelet count and/or a new thrombosis that begins 5–14 days after starting heparin, or when the individual's platelet count falls abruptly after he or she restarts heparin therapy following exposure to heparin within the preceding 3 months. Two types of tests support the diagnosis of HIT, antigen assays or functional assays. Antigen assays like the polyspecific platelet factor 4 (PF4)/heparin ELISA, which detects IgG, IgA, and IgM antibodies, have good sensitivity but limited specificity. Functional assays, such as the heparin-induced platelet activation test or the serotonin release assay, have higher specificity, but are not available in most hospitals.

PF4/heparin ELISAs often are the initial test in the diagnostic work-up of HIT, and those with optical density values <0.4 are considered negative, while those with optical density values 0.4–1 are considered weakly positive. These weakly positive results typically are followed by functional assay testing, which offers higher specificity, but is not available in most hospitals. Sent out as a reference test, functional assays can have days-long turnaround times. Many clinicians use a bedside clinical probability test such as the 4Ts score to focus diagnostic testing on patients with a higher pre-test probability of HIT. In the absence of such tools, physicians face a diagnostic and treatment challenge, as studies have shown that the majority of patients with weakly positive ELISA results do not have HIT.

"The problem is once you have a positive antigen test result, which happens in a lot of patients who don't have the clinical syndrome of HIT, it is difficult to ignore that result, and it's very hard not to put such patients on DTI, at which point you're exposing them to a lot more risk," observed Michael Fischer, MD, MS, an associate professor of medicine at Harvard Medical School in Boston.

Coagulation specialists have argued that this window between receipt of weakly positive ELISA results and higher specificity functional assay results leads to over-treatment of HIT, a little-appreciated issue. Smythe and her colleagues at Beaumont Hospital realized, however, that they could effect some improvements in this arena. "We showed in a previous study that many patients we treat with a DTI for HIT either had negative or weakly positive ELISA results. That made us realize we had some education to do," she recalled. "If the patient has a negative ELISA result for HIT, the chances of he or she having HIT are one percent or less, yet we were treating them. So we felt we had some room to improve our practice of whom we're actually treating for HIT."

A multidisciplinary team at Beaumont collaborated extensively to develop a protocol combining both diagnostic testing and clinical assessment to guide HIT recognition, laboratory testing, and treatment. The primary goal of this effort, which the authors reported previously, was to keep patients from being unnecessarily exposed to DTI therapy. After implementing the protocol, the investigators tracked its impact on HIT-related treatment, outcomes, and costs (Thromb Haemost 2012 108:992-8). Overall, they found the protocol both helped discontinue DTI therapy more rapidly when it wasn't warranted, while also starting it faster when indicated.

DTI therapy was discontinued within 24 hours of a negative PF4/heparin ELISA in 61% of patients post-implementation versus 30% pre-implementation. In addition, after Beaumont implemented the protocol, DTI therapy was more likely to start within 12 hours of HIT antibody testing in patients with a high- to moderate suspicion of HIT—77.4% post-implementation versus 25.8% beforehand. Thrombotic events were more common in the pre- versus post-implementation period, 34.4% versus 13%, while major bleeding incidents dropped by 6.6% after implementation. The authors also calculated that Beaumont would save at least $450,000 annually from the decline in inappropriate DTI use.

"This project was a significant undertaking, years of work by so many people, and we were thrilled with the results," said Smythe. "We were excited to find that we more than doubled our ability to stop DTI therapy within 24 hours of getting that negative ELISA result, because from a safety perspective it means we're not exposing patients to a medication they don't need. This also has cost implications, because DTI is an expensive treatment."

Based upon these results, both Smythe and Fischer encouraged laboratorians to consider whether an HIT protocol would benefit their institutions. "Lab directors should look at the authors' pre-protocol analysis in terms of understanding patterns of testing and patterns of use. That alone would be helpful in understanding how well hospitals are doing in diagnosing and treating this condition. What volume of testing are they doing? What volume of DTI are they using?" said Fischer. "The protocol itself is complicated, so hospitals would want to look carefully at whether it is something they could implement practically given their own clinical processes and work flows."

Smythe echoed Fischer's recommendations. "Reach out to your pharmacy partners if you don't have an HIT protocol in your hospital and explore the ability to collaborate and develop one, because patients are the winners here."