A Novel PF4-Dependent Platelet Activation Assay Identifies Patients Likely to Have Heparin-Induced Thrombocytopenia/Thrombosis (HIT)

Antibodies (abs) specific for heparin/platelet factor 4 (H/PF4) complexes are a hallmark of heparin-induced thrombocytopenia and thrombosis (HIT) and are thought to play a role in predisposing patients to thrombocytopenia and thrombosis. A solid phase assay that uses H/PF4 complexes as targets for ab detection (PF4 ELISA) is almost invariably positive in patients with HIT. This test can be performed quickly and a negative result makes HIT very unlikely; however, many patients testing positive in this assay do not develop HIT. A second test, the serotonin release assay (SRA), considered the "gold standard" for HIT diagnosis, selectively detects HIT abs that activate platelets and are presumptively pathogenic. A positive SRA test correlates well with clinical HIT but the SRA is technically demanding and is routinely performed only by a few reference laboratories, precluding its use for early patient diagnosis and management. This can have adverse consequences since both over- and under-diagnosis of HIT can have serious medical and economic impact.

We recently showed that platelet-activating HIT abs preferentially recognize PF4 bound to platelet glycosaminoglycans in the absence of heparin (Blood 2015; 125(1): 155-61). Based on this finding we developed a technically simple assay (PF4-dependent p-selectin expression assay [PEA]) for detection of platelet-activating HIT abs. In the PEA, target platelets are first "primed" for activation by being incubated with PF4 and are then exposed to patient serum. P-selectin expression (percent of maximum) is measured as an indicator of platelet activation (Thromb Haemost. 2015; 114(5) epub). Specificity of a positive test result is confirmed by inhibition with high dose heparin (HDH).

To compare diagnostic performance of the PEA and the SRA, 91 sera from patients referred for HIT testing who had been assigned clinical scores for HIT ("4T scores") ranging from 0 to 8 (low to high likelihood of HIT, respectively) were tested in both assays. Samples from patients with Intermediate or High 4T scores (4-8) and PF4 ELISA OD values >1.0 were considered "HIT-positive " given that neither criterion by itself has a high positive predictive value for HIT. Intermediate to High 4T-scored patients with PF4 ELISA ODs <1.0 were considered "HIT-negative ", as were samples associated with low 4T scores, which are known to have a high negative predictive value for HIT.

Receiver operating curve (ROC) analysis showed that the PEA had greater accuracy for identification of samples from HIT-positive patients than the SRA (AUC 0.93 vs. 0.82, respectively; p=0.01). When both sensitivity and specificity of an assay are important, one way to define a threshold for positivity is to determine the point on the ROC curve where the sum of sensitivity and specificity is maximized. This point in the PEA ROC curve corresponded to >24% p-selectin expression. For this analysis, a sample was considered PEA-positive if p-selectin expression was >24%, and inhibited >50% with HDH. Results are shown in Fig 1. The apparent superiority of the PEA over the SRA for identifying HIT-positive patients suggested that the PEA might be intrinsically more sensitive than the SRA for detecting platelet activating abs. This was confirmed by testing serial dilutions of 3 arbitrarily selected High-4T HIT abs against identical target platelets. A representative result is shown in Fig 2.

The many attractive features of the PEA, especially its technical simplicity and its ability to identify patients judged to have a high likelihood of HIT suggest that its use can facilitate early diagnosis and improve management of this diagnostically challenging disorder.

In Figures 1 and 2, the solid and dotted lines represent positive cut-offs of the SRA and PEA, respectively.

Figure 1.

View largeDownload slide

Thirty patients judged to be HIT-positive were tested in the PEA and SRA. The PEA was positive in 27 and the SRA in 16. Fig 1B. The PEA was positive in 8 and the SRA in 3 among 61 HIT-negative patients. The two assays had similar specificity (87% [PEA] vs. 95% [SRA]; p=ns), however, the PEA had significantly higher diagnostic sensitivity than the SRA (90% [PEA] vs. 53% [SRA]; p<0.01). Solid lines connect SRA and PEA results of each sample.

Figure 1.

View largeDownload slide

Thirty patients judged to be HIT-positive were tested in the PEA and SRA. The PEA was positive in 27 and the SRA in 16. Fig 1B. The PEA was positive in 8 and the SRA in 3 among 61 HIT-negative patients. The two assays had similar specificity (87% [PEA] vs. 95% [SRA]; p=ns), however, the PEA had significantly higher diagnostic sensitivity than the SRA (90% [PEA] vs. 53% [SRA]; p<0.01). Solid lines connect SRA and PEA results of each sample.

Close modal

Figure 2.

View largeDownload slide

A High 4T-scored platelet-activating HIT serum was adjusted to the fold-dilutions shown on the abscissa. PEA and SRA results are expressed on the ordinate. HDH results demonstrate specificity of platelet activation by HIT abs.

Figure 2.

View largeDownload slide

A High 4T-scored platelet-activating HIT serum was adjusted to the fold-dilutions shown on the abscissa. PEA and SRA results are expressed on the ordinate. HDH results demonstrate specificity of platelet activation by HIT abs.

Close modal

Figure 3.

View largeDownload slide

Figure 3.

View largeDownload slide

Close modal