Continue or Discontinue Anticoagulation Early After Stroke

Study Questions:

In patients already on anticoagulation who have an acute ischemic stroke, should anticoagulation be continued or discontinued?


This is a post-hoc analysis of PASS (Preventive Antibiotics in Stroke Study), a randomized controlled trial that enrolled patients in the Netherlands from 2010-2014. The “continue” group had systemic anticoagulation continued during their stroke admission. The “discontinue” group had systemic anticoagulation discontinued upon admission.


Of 2,124 acute ischemic stroke patients in PASS, 192 (9%) were on anticoagulation at the time of admission. Of these, 186 (97%) were on a vitamin K antagonist (like warfarin). Anticoagulation was continued in 157 of 192 (82%) patients. Anticoagulation was discontinued temporarily in 24/192 (12.5%) and permanently in 11/192 (5.7%) patients. Among patients with a severe stroke (NIHSS >15), anticoagulation was stopped in 14/27 (52%) compared to 21/165 (13%) in patients without a severe stroke (National Institutes of Health Stroke Scale [NIHSS] ≤15) (p < 0.001).

Discontinue patients were more likely to have a thrombotic event within 90 days (11 vs. 3%; p = 0.038), but this difference became nonsignificant after adjustment for sex and NIHSS. There were no major bleeding events in either group. Mortality was higher in the discontinue group than in the continue group (31% vs. 15%; p = 0.019), but this difference became nonsignificant after adjustment for sex and NIHSS. A favorable modified Rankin scale score (mRS 0-2) at 90 days was more common in the continue group than in the discontinue group (55% vs. 20%, p < 0.001); however, this difference also became nonsignificant after adjustment for sex and NIHSS.


After adjustment for stroke severity, there was no difference in frequency of thrombotic events, mortality, or 90-day favorable outcome between the continue and discontinue anticoagulation groups. No major bleeding events were observed in either group.


Providers appeared more likely to discontinue anticoagulation in patients with severe neurologic deficits (NIHSS >15), likely because of larger infarct size and a perceived higher risk of hemorrhagic transformation. Given that anticoagulation was discontinued in patients with larger strokes, the low rate of major bleeding events in the continue (less sick) group should be interpreted with caution. The more frequent thrombotic events, less frequent favorable 90-day outcome, and higher mortality observed in the discontinue group was probably a consequence of these patients having more severe strokes, given that the difference did not persist after adjustment for NIHSS. Note that 97% of the patients in this post-hoc analysis were on vitamin K antagonists, limiting generalizability of these results to patients on direct oral anticoagulants.

Clinical Topics: Anticoagulation Management, Prevention

Keywords: Anticoagulants, Brain Ischemia, Embolism, Fibrinolytic Agents, Hemorrhage, Secondary Prevention, Stroke, Thrombolytic Therapy, Thrombosis, Vascular Diseases, Vitamin K, Warfarin

< Back to Listings