The dynamic range of immunoassays for heparin-induced thrombocytopenia.

Abstract

Background: Following the current guidelines, immunoassays for the diagnosis of heparin-induced thrombocytopenia (HIT) are interpreted dichotomously, with test results categorized as either positive or negative. However, the extent to which test results hold diagnostic significance across the entire dynamic range remains unclear.

Objectives: We utilized data from the prospective TORADI-HIT study, comprising 1,393 consecutive patients with suspected HIT, to assess the diagnostic significance of two H/PF4 immunoassay test results across their respective dynamic ranges: HemoSil Acustar HIT IgG [CLIA] and Lifecodes PF4 IgG [ELISA].

Methods: HIT diagnosis was determined by a washed-platelet heparin-induced platelet activation assay (HIPA). For each measurement point in the dataset, we computed likelihood ratios (LR), sensitivities, and specificities. To provide post-test probabilities for individual test results, we calculated interval-specific likelihood ratios and integrated it into a web-based calculator.

Results: The prevalence of HIT was 8.5% (n = 119). A likelihood ratio of ≥ 10 was first achieved at 0.3% of the dynamic range (0.4 U/ml; CLIA) and 16% (0.64 OD; ELISA), respectively. A likelihood ratio of ≥ 100 was present at 9.4% (12 U/ml; CLIA) and 75.0% (3.0 OD; ELISA). The slope of the linear regression line (LR ∼ dynamic range) was 9.5 (CLIA) and 0.9 (ELISA).

Conclusion: Despite both immunoassays showing an association between results and diagnostic significance, the strength of that association varies by assay. CLIA has a larger increase per measurement unit. Post-test probabilities for individual patients can be estimated using a web-based calculator: https://pcd-research.shinyapps.io/BayesianCalculator/.