Different Coagulation Indicators in Predicting Clinical Outcomes for Patients With Direct Oral Anticoagulants: A Systematic Review and Meta-analysis

Purpose: There are many anticoagulant test indexes available for direct oral anticoagulants (DOACs), but how to select the appropriate index and the index cutoff values are still controversial. This is the first study, to our knowledge, to assess the association of different coagulation indicators with clinical outcomes among DOACs using a meta-analysis of observational studies.

Methods: A medical literature search was conducted using PubMed, Web of Science, EMBASE, ClinicalTrials.gov, and the Cochrane Library from inception to February 2020. Studies that reported relationships between coagulation indexes and clinical outcomes or the diagnostic value of coagulation assays were included in the analysis.

Findings: A total of 17 articles (7 meta-analyses and 10 systematic reviews) from 8904 citations were included in the analysis. In the analysis of bleeding events with coagulation indexes for DOACs, for peak prothrombin time level (cutoff value of 19-25 s), the pooled results found a sensitivity of 0.61 (95% CI, 0.44-0.75) and a specificity of 0.71 (95% CI, 0.49-0.86). For rivaroxaban, the trough anti-factor Xa concentration (AXA-C) (cutoff value of 400-500 ng/mL) had a sensitivity of 0.53 (95% CI, 0.16-0.87) and a specificity of 0.87 (95% CI, 0.71-0.94), with a diagnostic odds ratio of 7 (95% CI, 2-32). For apixaban, trough AXA-C had a sensitivity of 0.85 (95% CI, 0.60-0.96) and a specificity of 0.83 (95% CI, 0.52-0.95). The AUC of the AXA-C peak was higher than that of the trough AXA-C for apixaban, with a higher sensitivity and specificity. Compared with trough concentration of anti-factor IIa for dabigatran, the peak concentration had a higher specificity (98%) at the cutoff value of 484 ng/mL. In the analysis of thromboembolic events with coagulation indexes for DOACs, peak and trough prothrombin time values were not typically correlated with subsequent symptomatic venous thromboembolism, without a sensitivity or specificity higher than 90%. Trough AXA-C had a sensitivity of 100% and but a low specificity (<50%) for rivaroxaban-apixaban. Trough AXA-C had a sensitivity of 100% and a specificity of 32% with a cutoff value of 108 ng/mL for dabigatran.

Implication: Peak prothrombin time (19-25 s) and AXA-C had a better predictive value on bleeding outcomes for rivaroxaban and apixaban, whereas peak concentration of anti-factor IIa activity can be an indicator for dabigatran. Coagulation indexes might not be a good indicator of thromboembolic events of DOACs. Because the limited studies focused on association of coagulation indicators and clinical outcomes, more studies are needed to verify this in the future.