D-dimer-An International Assessment of the Quality of Laboratory Testing: Implications for D-dimer Use in the Real World.

Abstract

D-dimer assessment has several established roles in venous thromboembolism (VTE) and disseminated intravascular coagulation diagnosis, and recently the risk stratification of coronavirus disease 2019 (COVID-19). D-dimer assays are neither standardized nor harmonized, use varying methodologies, and use different reporting units, all resulting in a lack of interchangeability and generalizability of assays. Using large multiyear datasets from an international laboratory quality assurance program, we assessed (1) common D-dimer assays in use worldwide, (2) differences in analytical performance between different methods, and (3) interlaboratory variability between positive samples. External proficiency testing results from laboratories participating in the External Quality Control for Assays and Tests (ECAT) Foundation were analyzed from 2017 to 2023. Annually, between 578 and 690 laboratories participated in the D-dimer sample surveys with response rates ranging from 88 to 97%. The three most common assays in use in 2023 were the Siemens Innovance D-dimer (42%), the IL HemosIL D-dimer HS 500 (20%), and the Diagnostica Stago (Stago) Liatest D-dimer Plus (10%)-all these are automated, quantitative, latex immunoassays expressed in fibrinogen equivalent units (FEU). The highest interlaboratory variability was observed around the typical VTE exclusion threshold of 0.5 mg/L FEU. Lower interlaboratory variability was observed at values above 0.8 mg/L FEU. Our study provides recent, international performance data on currently used D-dimer assays and describes the significant variability between assays and across D-dimer concentrations. We demonstrate that assays are not interchangeable and that using them interchangeably has the potential to result in clinically important errors. There is an urgent need to educate users about these issues and to work towards harmonizing D-dimer units and reporting.