Management of oral anticoagulation after cardioembolic stroke and stroke survival data from a population based stroke registry (LuSSt)

Atrial fibrillation (AF) is the main cause of cardioembolic stroke (CES), the dominant ischemic stroke subtype in the elderly [1],[2]. Additionally, non-diagnosed AF is likely to be responsible for many cryptogenic strokes [3]. Prevalence of AF increases with age and lifetime risk of developing AF is estimated to be 25% after reaching the age of 40 [4]. Due to demographic changes in industrial countries, prevalence of AF is predicted to increase in the future [5]. Concomitant to the incidence of AF, incidence of CES is likely to increase. CES is associated with higher stroke severity and mortality [6],[7]. Stroke recurrence is associated with much worse outcome [6],[7]. Early stroke recurrence is frequent in ischemic stroke due to large artery atherosclerosis [8], in contrast, it is less common in CES [8],[6]. However, recurrence risk of CES is highest among stroke subtypes in the long term [7]. Oral anticoagulation (OAC) with vitamin K antagonists (VKA) and new oral anticoagulants (NOAC) is effective in primary and secondary stroke prevention and in reducing mortality [9],[10]. Both, VKAs and NOACs increase risk of intracerebral hemorrhage (ICH) [11]. As early hemorrhagic transformation occurs in CES in up to 17% within 5 days [12], and occurence of ICH is reported to be up to 12% [13], timing of OAC initiation after acute CES is still a controversial issue and its impact on mortality is not clear.

We hypothesized that initiation of OAC in hospital after first-ever CES is associated with a significant reduction on mortality compared to OAC non-treatment and OAC initiation post discharge. Implementation of OAC recommendation was additionally analysed.

The “Ludwigshafen Stroke Study” (LuSSt) is a population-based, prospective registry of stroke and transient ischemic attack (TIA) in the City of Ludwigshafen (Germany), starting January 1st, 2006. A detailed description of LuSSt has been published recently [14]. All patients with first-ever ischemic strokes (FEIS) due to AF until December 31st, 2010 were included in the present analysis.

During study period, 502 patients had a FEIS or TIA due to AF. ICH was ruled out in all patients by neuroimaging. As survival status was unknown in 23 patients (4.8%), 479 patients with FEIS (n = 394) and TIA (n = 85) were finally included in the present analysis. Mean age was 79.3 years +/- 10.1 (women: 80.7 years +/- 10.5; men: 77.4 years +/- 9.1). Median of hospital stay was 10.8 days (range 1-35 days). 332 patients (69.3%) were initially admitted to a stroke unit, whereas 29 patients (6.1%) received intensive care treatment. Overall, 18 patients (3.8%) were treated ambulatory. Thirteen patients died within first week. One year mortality rate was 28.4% and mortality rate after 500 days was 30.7%.

Neuroimaging was performed in all patients with 106 patients (22.1%) receiving cranial MRI. AF was newly diagnosed in 62 (12.9%) patients. Acute ischemic tissue damage could be observed in 20 patients (23.5%) with TIA based on neuroimaging. Due to study period, no patient received NOAC treatment. Baseline characteristics of all patients are displayed in Tables 1 and 2. Overall, OAC with VKA was started in hospital in 184 patients and it was continued post hospital in 181 patients (98.4%). Starting OAC post discharge was recommended in total in 130 patients. In 20 out of these patients OAC status post discharge was unknown and OAC-recommendation was implemented in 71 patients (64.5%). Figure 1 shows Kaplan-Meier estimate of survival after ischemic stroke according to OAC management. Antiplatelet therapy was prescribed in 173 (36.1%) patients. There was no patient who received a combination of OAC and antiplatelet agent. Main reasons for non-OAC recommendation were bad general condition (28%, n = 41), risk of falling (27%, n = 39) and dementia (13%, n = 19). As compared to those patients with recommendation to start OAC treatment post discharge, patients receiving OAC treatment in hospital had signficant lower initial stroke severity as indicated by NIHSS (p > 0.0001), lower HASBLED-score (p = 0.001), less signs of hemorrhagic transformation (p = 0.002) and territorial infarction (p > 0.0001), high rate of premedication with VKA (p < 0.0001), low rate of premedication with antiplatelets (p < 0.0001) and significant lower mRS-score (p < 0.0001) on discharge. Recurrent stroke within 500 days occurred in 43 patients (9%) with 13 patients (30,2%) being on OAC (started in hospital: n = 8 (61.5%) vs. started post discharge: n = 5 (38.5%), of which 6 patients (46.1%) had an INR-value below 2. Three out of 5 patients (60%) with OAC having an INR > 2, had a recurrent ischemic stroke within 30 days after OAC. One year recurrence rate was 7.1% (n = 34), whereas rate of early stroke recurrence within 14 days was 0.8% (n = 4) and within 7 days 0.4% (n = 2). OAC associated ICH was observed in three patients (0.6%), two (66.7%) occuring within 30 days after initial event.

Results of Cox regression analysis are displayed in Table 3. Higher age, higher mRS-score at discharge, diagnosis of CAD, and OAC treatment post stroke were independently associated with stroke mortality. Figure 2 additionally presents Kaplan-Meier estimate of survival with regard to stroke recurrence. Poststroke survival was highest in those patients receiving OAC irrespective of stroke recurrence.

We analysed impact of risk factors on mortality (500 days) in patients after first ever AF-associated CES or TIA in a population based stroke registry (LuSSt) particulary with regards to the effect of different OAC timings after FEIS. One-year mortality rate was 28.4%, which is in line with former study reports, where patients having a stroke associated with AF had a high risk of death in the year after FEIS [21].

Our data show that OAC non-treatment after AF-associated ischemic stroke is the main predictor for stroke mortality. Mortality was 5.4 times higher in patients not receiving OAC as compared to those that already received OAC in hospital. This fact again underlines the need for OAC treatment after CES [22]. Alarmingly, OAC-recommendation was only implemented in 71 out of 110 patients calling for improvement in outpatient care (e.g. by the implementation of special interventional programs) [23]. Interestingly, OAC initiation post discharge showed a trend (p = 0.1) towards lower mortality as compared to those that received OAC in hospital (HR 0.3; 95% CI 0.07-1.4) in multivariate analysis. Increased risk of bleeding under early treatment with OAC in more severe strokes might be one explanation [24]. Increased risk of ischemic stroke during first 30 days of treatment might also partly account for this [25]. In addition to OAC treatment, we found higher age and a higher mRS-score at discharge as independent predictors of mortality after CES. Recent data from the Canadian Stroke Network demonstrated association between older age and worse outcome after CES. Interestingly, there was no such association in those patients receiving therapeutic OAC treatment prior to stroke emphasizing the importance of OAC treatment in older patients [26]. It is well known that myocardial infarction is the leading cause of death in patients with cerebrovascular diseases [27]. We found CAD to be independently associated with mortality after CES. Antiplatelet agents such as Aspirin or Clopidogrel have shown to reduce ischemic events in patients with CAD [28]. Unfortunately, concomitant use of antiplatelet agents and OAC subsequently increases risk of bleeding [29]. Randomized trials have shown VKA to be as effective as a combination of VKA and Aspirin and more effective as Aspirin alone in CAD patients considering myocardial infarction [30]. Although recent observational studies showed decreased rate of MI in AF patients receiving Dabigatran as compared to Warfarin [31], certain subgroup analyses and metanalyses of randomized controlled trials demonstrated an increase of myocardial infarction [32],[33]. However, data of end-point driven randomized controlled trials on NOACs are missing. Right now, OAC with VKA might remain important for CES patients having CAD. Recurrence rates were similar to those reported in previous trials [34],[35]. Although being limited by small sample size, stroke recurrence did not influence survival in those patients receiving OAC, whereas there was a clear interaction in those patients that did not receive OAC. Clinical benefit of OAC in AF is again underlined.

Our study has strengths and limitations. Being based on registry data, study design was primarily descriptive and besides other confounding factors, quality of treatment of risk factors was not implemented, especially considering time of therapeutic range of VKA treatment. Furthermore we only analysed all-cause mortality. However, LuSSt is a population-based stroke registry without any age restrictions. Multiple overlapping methods were used to ensure nearly complete case ascertainment and high case completeness was achieved with regard to stroke mortality.

In our study OAC non-treatment was the main predictor for mortality after FEIS. A later initiation of OAC treatment post hospital was not inferior to early initiation during hospital stay, however, one third of those patients who were recommended to be treated with OACs never received treatment. Special secondary preventive programs may increase the proportion of anticoagulated patients after stroke due to AF.

All authors have made substantive intellectual contributions to this study. FP and MK have made substantial contributions to data acquisation, data analysis and data interpretation. AS and HB substantially contributed to data analysis and interpretation. AG was responsible for study conception and design as well as data interpretation. All authors critically drafted and revised the manuscript. They all agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. All authors read and approved the final manuscript.