Background

Direct oral anticoagulants (DOACs) have become the standard treatment for stroke prevention in nonvalvular atrial fibrillation (AF), and use of these medications is growing as the US population ages. Despite decreased risk of stroke while taking DOACs, acute ischemic stroke (AIS) remains common patients with AF, but current guidelines do not support the administration of intravenous thrombolysis (IVT) in patients who have taken a DOAC within 48 hours because of concern for increased risk of symptomatic intracerebral hemorrhage (sICH). This is based on data from patients taking vitamin K antagonists (VKAs), but DOACs have lower rates of ICH overall and these patients could benefit from IV tPA.

Methods

Meinel and colleagues performed a retrospective cohort study across 64 primary and comprehensive stroke centers across Europe, Asia, and Oceania comparing rates of symptomatic ICH (sICH) in patients on DOACs who received IVT versus a control group that was not on DOACs and received IVT. Additional selection strategies for IVT in the setting of DOAC use were studied including patients with DOAC levels, DOAC reversal prior to IVT, and IVT administration without DOAC levels or reversal. The primary outcome was sICH, defined as an NIHSS decline of ³4 points, within 36 hours after IVT. Secondary outcome was any ICH within 36 hours after IVT and functional status measured by modified Rankin Scale (mRS). The authors hypothesized the rate of sICH would not be higher in patients with recent DOAC ingestion.

Statistical analysis was a multilevel mixed effects logistic regression analysis which allowed the authors to adjust for known sICH risk factors such as age, hypertension, baseline NIHSS, premorbid poor functional status, admission blood pressure, and admission glucose. Selection strategy was also a variable in the model with the different strategies as discussed above. With DOAC ingestion as the independent variable in the model, four sensitivity analyses were performed based on geographic location, availability of institutional protocols, DOAC level >100ng/mL or proven ingestion <12h before IVT, and patients with known time of ingestion.

Results

The study enrolled 33,207 patients (43.5% female) with a median age of 73 (IQR 62-80) and median NIHSS of 9 (IQR 5-16). 832 patients had taken a DOAC and 355 (42.7%) of those patients received IVT without reversal or measured DOAC levels. 225 (27.0%) patients had DOAC levels measured, and 252 (30.3%) patients received DOAC reversal prior to IVT. The DOAC group differed from the control group in that patients were older with higher levels of pre-stroke disability, and they suffered more severe strokes including large vessel occlusions (LVO). Anti-platelet use was more common in the control group.

For the primary outcome, the unadjusted rate of sICH was lower in the DOAC group versus the control group at 2.5% (95% CI, 3.9-4.3) and 4.1% (95% CI, 3.9-4.4), respectively. When adjusted for stroke severity and sICH predictors, the data again showed that the DOAC group had lower odds of sICH (adjusted OR, 0.57; 95% CI, 0.36-0.92; P=0.02). This finding held across the different selection strategies and sensitivity analyses previously described.

For the secondary outcomes, there was no difference in the unadjusted rate of any ICH in the DOAC group versus the control group at 18% (95% CI, 15.4-20.9) and 17.4% (95% CI, 16.9-18.0), respectively. After adjustment there remained no difference between the groups. The DOAC group had a lower proportion of functional independence at 45% (95% CI, 41-49) versus 57% (95% CI, 56-57) for the control group (OR 0.62, 95% CI 0.53-0.73; p<0.001) in the unadjusted analysis, although there was no significant difference in the adjusted analysis (adjust OR 1.13, 95% CI 0.94-1.36; P=0.20). The DOAC group had worse baseline pre-stroke function and suffered larger strokes, but after adjustment the DOAC group had higher odds of functional independence. Furthermore, DOAC ingestion was not significantly associated with disability categories in the ordinal regression shift analysis.

Commentary

In this international, multicenter, retrospective cohort study, DOAC use within 48 hours was not associated with increased intracerebral hemorrhage after thrombolysis for AIS. This finding may seem counterintuitive, but there was prior data to suggest that IVT after DOAC use does not increase sICH. However, these studies did not verify ingestion or document laboratory evidence of DOAC activity. The strengths of this study are that it included patients with confirmed DOAC ingestion, specifically addressing the weaknesses of previous literature, and that it has generalizability given its geographical breadth and inclusion of different levels of stroke center.

The main limitation of this study is its retrospective and observational design which may have allowed selection bias. One potential explanation for the lower rate of sICH in the DOAC group could be selection bias due to clinicians cautiously selecting patients for IVT that they felt would be at a lower risk for sICH (i.e., earlier in the time window or younger age). However, a comparison of baseline characteristics between the DOAC group and control group does not support this. In fact, the DOAC group was older and suffered more severe strokes, factors that would favor increased sICH rather than decreased sICH. Furthermore, the odds ratio adjusting for stroke severity and sICH predictors also showed a lower rate of sICH in the DOAC group. The authors proposed multiple explanations for this seemingly counterintuitive finding including that DOAC use may result in smaller infarcts with a lower risk for hemorrhagic transformation, that thrombin inhibition may minimize blood-brain barrier disruption, and that DOAC use may assist with vessel recanalization.  

There was also variation in the dose of alteplase used since Asian centers administer a lower dose. Given the accelerating transition to tenecteplase, more data is needed with this thrombolytic agent. In this study, only 51 patients received tenecteplase, all of whom were in New Zealand, and 46 of them received idarucizumab for dabigatran reversal. This is a small sample size from one country using a less common DOAC.

In summary, these finding challenge current guidelines recommending that IVT should not be given to patients taking DOACs (last dose within 48 hours). IVT is one of only two proven interventions for AIS, so it is important to maximize the number of patients who may benefit from IVT. This study’s retrospective and observational design are significant limitations, and the findings should be confirmed with prospective studies that include tenecteplase.