Performance evaluation of a reflexive algorithm for suspected heparin-induced thrombocytopenia

Area of Science:

• Clinical diagnostics
• Immunology
• Hematology

Background:

• Heparin-induced thrombocytopenia (HIT) diagnosis relies on screening tests like ELISA and LIA, followed by platelet activation assays.
• High false-positive rates with LIA alone necessitate improved diagnostic algorithms.
• A reflex algorithm using LIA followed by IgG-specific ELISA confirmation was adopted to enhance HIT diagnostic accuracy.

Purpose of the Study:

• To evaluate the diagnostic performance of a reflex LIA-first ELISA algorithm for suspected HIT.
• To compare this algorithm's operational impact against ELISA or LIA alone.
• To assess the algorithm's effectiveness in improving diagnostic accuracy and reducing unnecessary confirmatory testing.

Main Methods:

• Retrospective comparison of ELISA alone, LIA alone, and LIA with reflex ELISA confirmation across three eras.
• Performance assessment against SRA-confirmed HIT.
• Calculation of sensitivity, specificity, PPV, NPV, and accuracy; assessment of assay reactivity and SRA positivity correlation.

Main Results:

• The reflex ELISA algorithm significantly improved positive predictive value (PPV) from 25.3% (LIA alone) to 61.2%.
• Accuracy increased from 33.3% (LIA alone) to 63.5% with the reflex algorithm.
• Reflex testing reduced potential SRA send-outs by 61.4%, from 15.8 to 6.1 per 100 patients.

Conclusions:

• Reflex ELISA confirmation enhances HIT diagnostic efficiency by increasing PPV and accuracy.
• The algorithm effectively reduces the number of patients requiring confirmatory SRA testing.
• Assay reactivity can stratify HIT likelihood, guiding confirmatory testing and potentially speeding up diagnosis and management.