Implementation of a Fully Automated Latex Immunoturbidimetric Assay for the Diagnosis of Heparin-Induced Thrombocytopenia in a High-Volume University Hospital Laboratory

Abstract

Background Heparin-induced thrombocytopenia (HIT) is an immune-mediated, adverse reaction to heparin in which heparin binds platelet factor 4 (PF4), triggering the development of heparin-PF4 antibodies (HITAb). HITAbs bind and activate platelets, causing thrombosis, platelet consumption, and thrombocytopenia. Heparin is replaced with relatively costly nonheparin anticoagulants until HIT can be ruled out. HIT diagnosis consists of a HITAb immunoassay with reflex to a serotonin release assay (SRA) for confirmation of positive results. Recently, a fully automated latex immunoturbidimetric assay (LIA) for detection of HITAbs received FDA clearance. We sought to verify the performance characteristics of the LIA with the aim of implementing the test in a high-volume university hospital laboratory. Methods The in-house HITAb LIA was performed on the Instrumentation Laboratory TOP700 analyzer using HemosIL HIT-Ab(PF4-H) reagent. The comparator method, a HITAb ELISA, was performed at a reference laboratory with positive results reflexed to SRA. All samples (36 total) sent to the reference laboratory for HIT testing from December 2017 to November 2018 were aliquoted and run in parallel by LIA. Intra-assay precision was assessed by running manufacturer-provided low and high control samples 10 times in succession, while interassay precision was assessed by running low and high samples every day for 10 days. Turnaround time to HITAb result was retrieved from the electronic medical records for HIT testing performed 60 days before and after in-house test implementation. Results The agreement between the LIA and ELISA was 92% (33/36). One discordant sample tested negative by ELISA and was not assessed by SRA. Another tested positive by ELISA and negative by LIA and was confirmed negative by SRA. The final discordant sample tested negative by LIA but positive by ELISA and was confirmed positive by SRA. Thirty-three percent (12/36) of samples tested positive for HITAb by ELISA and were reflexed to SRA. Both the ELISA and LIA showed 83% agreement (10/12) with the SRA. The coefficient of variance (CV) for the intra-assay precision studies was 18% and 5% for the low and high controls, respectively. The CV for the interassay precision studies was 28% and 5% for the low and high controls, respectively. Postimplementation quality control data revealed 61% and 20% imprecision on the low and high level controls, respectively, which declined significantly when reagents were removed from the instrument and refrigerated within 2 hours. The turnaround time for HITAb results was reduced by 74% (10.5 vs 41 hours) after in-house test implementation, significantly reducing the need for administration of nonheparin anticoagulants. Conclusion The LIA and ELISA methods compared favorably, allowing for clinical implementation of the LIA. The shortened turnaround time of the LIA significantly reduced the time to rule out HIT, enhancing patient care and reducing drug costs. The assay imprecision warrants further investigation regarding reagent stability.