Impact of D-Dimer Assays Performed at Local Labs Vs. Central Laboratory in the Evaluation of Apex Trial Outcomes

Introduction: Elevated D-dimer levels are expected to represent both patients at high risk for venous thromboembolism (VTE), and those benefiting from extended-duration antithrombotic therapy. Using D-dimer testing as a method for enrichment strategy, the APEX trial compared the clinical benefit of extended-duration betrixaban therapy with enoxaparin standard of care (NEJM 375:534, 2016). D-dimer test results from the local laboratories at the enrollment sites were used for study outcome analyses based on a ≥ 2xULN value to identify subjects with elevated D-dimer (Cohort 1).

Methods: Bloodsamples were collected from hospitalized medically ill patients (n=7,513) enrolled in the APEX study for D-dimer level analysis. A total of 27 differentD-dimer tests were used for D-dimer testing at the local laboratories. A duplicate sample was analyzed independently by a central laboratory using a single, quantitative method (STA-LITA-D-Di test). APEX study data are further analyzed using D-dimer test results obtained from the central laboratory and then compared with the D-dimer results obtained from the local laboratories.

Results: Analysis of APEX D-dimer results from local vs central labs was performed in randomized subjects in the modified intention to treat (mITT) population (which consists of all subjects who were administered at least one dose of study drug) (n=7441).In this population 1,365 D-dimer results (18.3%) were discordant. The percent discordance of central and local lab D-dimer results in the betrixaban (670/3721) and enoxaparin (695/3720) arms were 18.0% and 18.7%, respectively. The event rates using ≥ 2X ULN D-dimer cutoff value showed significant differences between the two treatments when central lab D-dimer value was used (5.3% vs 7.6%, p=0.002) and with the local D-dimer (5.7% vs 7.2%, p=0.038), but less so with the latter test. Event rates in subjects designated as positive by central lab but negative by local lab (false negatives) in enoxaparin (21/302, 7%) vs betrixaban (9/313, 2.9%) groups were significantly different while in subjects identified as negative by central lab and positive by local lab (false positives), there were no differences in the overall event rates for enoxaparin (18/393, 4.6%) and betrixaban (17/357, 4.8%) groups. Further analysis of the data based on the type of local D-dimer method utilized suggested that 9/27 tests resulted in false negatives more than false positives.

Conclusions: Variability in D-dimer testing as a prognostic indicator for VTE can significantly impact outcome evaluation in clinical trials. Standardization of D-dimer testing is necessary when utilizing the test for predictive enrichment of patient population.