Pleiotropic effects of direct oral anticoagulants on the coagulation-inflammation-endothelium axis: A phenotype-stratified narrative review.

AIMS: Direct oral anticoagulants (DOACs) - dabigatran (a thrombin inhibitor) and rivaroxaban, apixaban, and edoxaban (factor Xa [FXa] inhibitors) - modify oxidative stress, endothelial inflammation, and barrier integrity beyond anticoagulation. We ask whether this reflects distinct pharmacology or silencing of proteases at protease-activated receptors (PARs), and propose a three-condition framework for when it becomes clinically measurable.

METHODS

Narrative review (PubMed, Scopus, Web of Science; primary search 2019-2026, with foundational earlier references retained) with claims graded on a six-tier hierarchy; A‡ marks trials formally negative or inconclusive owing to insufficient statistical power.

RESULTS

DOAC effects largely converge on one node - reduced PAR-1 signalling under thrombin inhibition and dual PAR-1/PAR-2 blockade under FXa inhibition - rather than independent targets; the FXa-PAR-2 axis is cross-validated across macrophages, liver sinusoidal endothelial cells, and neutrophils; and dabigatran plausibly differs qualitatively rather than only in degree - a working hypothesis whose mechanistic linchpin (residual exosite-I signalling) currently rests on a single unreplicated in-vitro study (level D). Clinically the signal sorts by phenotype and disease phase, not class: low-dose rivaroxaban benefits stable atherosclerosis (COMPASS) but not heart failure in sinus rhythm (COMMANDER-HF), and COVID-19 benefit appears only in convalescence (MICHELLE), not acutely or in outpatients (ACTIV-4B, A‡). A nationwide cohort linked DOACs to lower acute kidney injury and chronic kidney disease progression than vitamin K antagonists, pending separation from their nephrotoxicity.

CONCLUSIONS

Pleiotropy emerges where three conditions coincide - the protease is available at PARs, its generation is chronic, and PAR signalling is rate-limiting. This supports within-indication molecule selection (selection, not extension) and biomarker co-primary trials, not extension of indications.