If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Data and Research Services, Alberta Strategy for Patient Oriented Research Support Unit and Provincial Research Data Services, Alberta Health Services, Edmonton, Canada
Corresponding Author: Dr. Tammy J Bungard 8425 Aberhart Centre 11402 University Avenue University of Alberta Edmonton, AB, T6G 2J3 Phone: 780-407-8597 Fax: 780-407-8907
For patients with atrial fibrillation (AF) having ischemic stroke or transient ischemic attack (TIA) despite taking direct oral anticoagulants (DOAC), the optimal strategy for ongoing anticoagulation is unknown.
Methods
Using provincial administrative databases in Alberta, Canada, we compared anticoagulant use before/after the breakthrough stroke/TIA and assessed recurrence of stroke/TIA or bleeding, with consideration of medication adherence. Adherence was defined as the proportion of days covered (PDC) being >80%.
Results
Amongst 985 patients, median age was 80 (13) years with a mean CHADS2 score of 1.7+1 prior to index event. Patients were followed for 643 (836) days. Following the index stroke/TIA event, 623 patients (63%) filled the same DOAC regimen, 83 (8%) filled a different dose, 155 (16%) switched DOAC agents, 51 (5%) switched to warfarin and 73 (7%) filled no OAC. Patients who kept the same regimen more commonly had TIA index events (59%); patients that changed dose or drug more often had stroke index events (55-78%). During follow up, 135 (14%) had stroke/TIA recurrence and 46 (5%) had bleeding; rates of each did not differ between prescribing patterns. Post-index event, proportion of patients with PDC >80% improved from 55% to 80%.
Conclusions
While most maintained the same DOAC regimen after stroke/TIA, rates of recurrent stroke/TIA and bleeding were similar across prescribing patterns. Stroke/TIA severity may have influenced prescribing practices. DOAC adherence improved post-stroke/TIA and signals an opportunity for optimization in patients with AF.
Atrial fibrillation (AF) is the most commonly encountered cardiac arrhythmia and confers significant risk for ischemic stroke and transient ischemic attack (TIA).
When risk of stroke is sufficiently high based on risk stratification schemes such as CHADS2, CHADS2-VASc or CHADS-65, guidelines for AF recommend anticoagulation with direct oral anticoagulants (DOACs) over warfarin.
January CT, Wann LS, Calkins H et al. 2019 AHA/ACC/HRS Focused Update of the 2014 AHA/ACC/HRS Guideline for the Management of Patients With Atrial Fibrillation: A Report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines and the Heart Rhythm Society in Collaboration With the Society of Thoracic Surgeons. Circulation 2019;140(2):e125–e151.
However, despite anticoagulation of AF with DOACs, there remains approximately a 1 per 100 patient years residual risk of stroke and systemic embolism, as reported within the large DOAC clinical trials.
Atrial fibrillation guidelines recommend that patients who experience a stroke while on oral anticoagulant (OAC) therapy be managed with an emphasis on addressing medication adherence, ensuring guideline concordant dosing, and avoiding drug interactions.
Kirchhof P, Benussi S, Kotecha D et al. 2016 ESC Guidelines for the management of atrial fibrillation developed in collaboration with EACTS. Europace 2016;18(11):1609–1678.
there is however a lack of robust clinical evidence to describe and support such a strategy. While there are no randomized trials directly comparing the efficacy of DOACs, a strategy of changing the original DOAC to one with an alternative mechanism of action has been utilized by prescribers.
A recent retrospective cohort study suggested that changing OAC type is not associated with reduced rates of stroke recurrence compared to no change in OAC, though this study focused on patients anticoagulated with vitamin K antagonists.
Given the uncertainty of clinical outcomes with these varying treatment options, we sought to describe OAC prescribing in patients with AF following an ischemic stroke/TIA while already taking a DOAC and assess recurrent stroke/TIA and bleeding outcomes post-stroke with consideration of medication adherence.
METHODS
Study design
This was a retrospective review of patients with AF on a DOAC that were discharged from hospitals with a diagnosis of ischemic stroke/TIA from January 2011 to October 2019. The Alberta Strategy for Patient Oriented Research Unit linked administrative databases that capture all interactions within the single-payer, universal-access healthcare system in Alberta, Canada; including: Discharge Abstract Database for hospitalizations, National Ambulatory Care System for emergency department encounters, Physician Claims Database for outpatient physician visits, Pharmaceutical Information Network for outpatient medication dispensation data, Provincial Laboratory Database laboratory values, and Provincial registration and Vital Statistics Database for the vital statuses.
Patient population
Patients aged >18 years who were discharged alive from an acute care setting, emergency department or rehabilitation facility in Alberta with a diagnosis of ischemic stroke or TIA based the International Classification of Diseases, 10th Revision (ICD-10) codes H34.1, I63 and G45.9 (referred to as the index event) were included. Next, confirmation of AF or atrial flutter in the previous 10 years based on at least one diagnosis in a hospital setting or at least two diagnoses at least 30 days apart in an outpatient setting, using ICD-10 and International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM) codes was done. Accuracy of ICD-9 and ICD-10 codes for stroke and TIA has been previously validated against chart audit data in Alberta.
Patients had to have community pharmacy fills for dabigatran, apixaban or rivaroxaban within 120 days prior to the index event. Patients were excluded for moderate-severe mitral stenosis, mechanical heart valve, stroke/TIA within 90 days before index event, stroke/TIA while originally admitted for another diagnosis or procedure, lack of database entries suggesting lack of healthcare system contact within 120 days following discharge from index event, or primary residence outside of Alberta based on postal code. Edoxaban was not investigated in our study due to very limited use (n=2) during the study period.
Outcomes
The primary outcome was the proportion of patients within each anticoagulant prescribing strategy pre- versus post-index event, based on OAC pharmacy fills. Strategies were pre-defined as those that kept same DOAC regimen, kept same DOAC agent but changed dose, changed to an alternative DOAC, changed to warfarin or did not fill an OAC. Secondarily, we sought to determine the proportion having recurrent stroke/TIA, or an occurrence of a bleeding event requiring medical attention. A recurrent stroke/TIA was identified via ICD-10 codes for ischemic stroke or TIA within inpatient, emergency, and outpatient settings (Supplemental Table S1). Bleeding events were defined as either: fatal, intracerebral, subarachnoid, other critical area, gastrointestinal, genitourinary, or other bleeding based on ICD-10 codes within a hospital inpatient setting (Supplemental Table S1); or bleeding based on the ICD-10 codes in an emergency department requiring a transfusion of whole blood or packed red blood cells. Patient follow up ceased with a clinical event including recurrent stroke/TIA, bleeding, or death. In the absence of these outcomes, patients were followed to the end of the study period, which encompassed a minimum of 6 months for all patients.
Medication adherence was investigated using the proportion of days covered (PDC) method. PDC measures the ratio of total days covered by refills within a period to the total number of days from first fill to the end of the period.
The period was 1 year unless follow up ended due to a clinical event. Patients with a PDC ratio >80% were considered adherent, as a PDC <80% is a known risk factor for AF-related stroke.
Alberts MJ, Peacock WF, Fields LE et al. Association between once- and twice-daily direct oral anticoagulant adherence in nonvalvular atrial fibrillation patients and rates of ischemic stroke. Int J Cardiol;215:11–13.
Descriptive results are presented as proportions, mean ± standard deviation or median [interquartile range]. Creatinine clearance (CrCl) was based on the modified Cockcroft-Gault standardized to a weight of 72kg, using the last recorded creatinine value prior to discharge from the index event. A Cox regression model was used to generate hazard ratios (HRs) for recurrent stroke/TIA risk factors. Statistical tests were performed with level of significance of p<0.05 and 95% confidence intervals were calculated. Ethics approval was obtained from the University of Alberta Health Research Ethics Board (Pro00104912).
RESULTS
Baseline characteristics
Amongst 985 patients included (Figure 1, Table 1), the median age was 80 [13] years, 46% were female, and the majority (97%) had CrCl >30 ml/min. The mean CHADS2 score was 1.7±1 prior to the index event, with 11% having a prior history of stroke/TIA. A similar proportion of index events were identified as stroke (50%) versus TIA (50%), and the median length of admissions was 2 [7] days. Patients were followed up for a median of 643 [836] days, during which there were 316 deaths.
Figure 1Study flow diagram. 45,796 patients in Alberta, Canada were initially identified within administrative databases. 985 were analyzed within this study that had a history of AF and filled a DOAC prior to a stroke or TIA, and met the additional inclusion/exclusion criteria listed. AF: atrial fibrillation; TIA: transient ischemic attack.
Following the index event, 623 of 985 patients (63%) filled the same DOAC regimen, 83 (8%) filled a different dose of the same DOAC, 155 (16%) switched DOAC agents; 51 (5%) switched to warfarin and 73 (7%) filled no OAC (Figure 2). Of those who filled a different dose, 62 (75%) filled an increased dose and 21 (25%) filled a decreased dose.
Figure 2OAC prescribing post-index event. For patients with atrial fibrillation that have a stroke or TIA despite DOAC therapy, the prescribing practices of oral anticoagulation are depicted as proportions of the total cohort. 623 of 985 patients (63%) filled the same DOAC regimen, 83 (8%) filled a different dose of the same DOAC, 155 (16%) switched DOAC agents; 51 (5%) switched to warfarin and 73 (7%) filled no OAC. DOAC: direct oral anticoagulant; TIA: transient ischemic attack.
Prior to the index event, use of dabigatran (n=323, 33%), apixaban (n=317, 32%) and rivaroxaban (n=343, 35%) was similar (Figure 3). Patients who switched DOACs most commonly filled apixaban post-index event (72%). Amongst the patients who remained on a DOAC, the distribution of DOAC agents filled was significantly different pre- vs post-index event (p<0.05): dabigatran (33% vs 27%), apixaban (32% vs 43%) and rivaroxaban (35% vs 30%). Dabigatran was the single DOAC being filled from 2011 to 2012, with fills of apixaban and rivaroxaban emerging in 2013 and proportional use increasing thereafter (Supplemental Figure S1).
Figure 3Oral anticoagulant prescribing based on pre-index event DOAC therapy. Following a breakthrough stroke despite the use of dabigatran (n=323), apixaban (n=317) and rivaroxaban (n=343), the prescribing pattern for subsequent oral anticoaglant use was analyzed. Proportion of patients (%) per prescribng pattern for each DOAC used pre-index event is illustrated. DOAC: direct oral anticoagulant.
Differences in prescribing were seen based on type of index event. Of patients kept on the same DOAC regimen, a lower proportion had experienced stroke (41%) than TIA. Patients that changed to a different DOAC dose, different DOAC agent, warfarin or no OAC more commonly had stroke (55%, 72%, 78% and 60% respectively) than TIA. Compared to the overall cohort, patients that changed to warfarin had longer admissions (11 days), longer durations of follow up (915 days), a higher proportion experiencing death (10%) and more had either a CrCl <30 ml/min or no documented creatinine value (12%). Patients that did not fill an OAC post-index event had shorter durations of follow up (40 days) with a higher proportion experiencing death (63%).
Recurrent stroke or TIA, and bleeding outcomes
During follow up, 135 of 985 patients (14%) experienced a recurrence of stroke (49%) or TIA (51%), occurring a median of 244 days [439] from the index event (Table 2). The proportion of patients with recurrent stroke/TIA was similar across prescribing categories; however, the group that did not fill an OAC that had more recurrent stroke/TIAs (32%), with most of these events being strokes (65%) occurring at a shorter time to event (17 days [32]).
Table 2Characterization of recurrent strokes/TIAs, and bleeding events post-index event
Other: genitourinary, other critical areas including intraspinal, intraocular, retroperitoneal, intraarticular and pericardial hemorrhage, other areas including epistaxis and hemothorax.
∗ Percentages expressed as proportion of all patients or patients in each prescribing strategy with a recurrent stroke or TIA.
# Percentages expressed as proportion of all patients or patients in each prescribing strategy with a bleeding event.
$ Other: genitourinary, other critical areas including intraspinal, intraocular, retroperitoneal, intraarticular and pericardial hemorrhage, other areas including epistaxis and hemothorax.
46 of 985 (5%) experienced a bleeding event during follow up, of which the majority were gastrointestinal (63%), or intracranial/subarachnoid (23%) in nature (Table 2). Proportion of patients with bleeding events was similar across prescribing categories. Median time to event was 536 days [773], though patients that did not fill an OAC had a shorter time to event at 60 days [53].
Adherence
Prior to the index stroke/TIA event, 540 (55%) patients were considered adherent with a PDC >80% for DOACs, and this improved to 85% over the first year following the index event (Table 3). Adherence rates post-index event amongst those having recurrent stroke/TIA (87 of 104, 84%) and bleeding (30 of 38, 79%) were comparable to the entire cohort.
Table 3Proportion of adherent patients (PDC >80%) based on prescribing categories during follow up
DOAC fill properties including prescribing strategy, DOAC agent and adherence >80% were not found to be predictive of stroke/TIA recurrence based on our Cox regression model (Supplemental Table S2). Of the patient factors assessed, only age at time of index event was a significant predictor of recurrence (HR 1.032, 95% CI 1.005-1.06).
DISCUSSION
In this study, the most common strategy to manage anticoagulation after a stroke/TIA despite DOAC therapy was to keep the same DOAC regimen (63% of patients), followed by changing to another DOAC (16%), changing the dose (8%), changing to warfarin (5%) and no OAC (7%). Over a median follow up time of 643 days after their initial event, 7% of the cohort had a stroke, 8% had a TIA, and 5% had a bleeding event. The rates of stroke seen in this cohort are comparable to previous observational studies examining rates of AF-related strokes with prior anticoagulation,
Overall, proportions of patients with stroke/TIA recurrence and bleeding events were similar between the different prescribing strategies in this study. This result is similar to a previous analysis showing a change in OAC, primarily from vitamin K antagonists, was not associated with decreased risk of recurrent ischemic stroke compared to unchanged OAC.
While these results do not point to a preferred approach to managing anticoagulation, they may reflect the importance of an individualized decision-making process. For example, severity of index event appeared to have influenced prescribing patterns, since patients experiencing TIA and shorter hospital stays were more commonly maintained on the same DOAC regimen, while patients with stroke and comparatively longer length of stays more frequently had dose changes or were switched to an alternative anticoagulant therapy.
In clinical practice, patients presenting with stroke/TIA events despite anticoagulant therapy require assessment to determine if treatment failure occurred or if they were on inadequate therapy due to poor adherence, or other factors such as discordant dosing and drug-drug interactions. Adherence of our cohort improved post-index event with proportion with PDC >80%increasing from 55% to 85%. The proportion of patients with PDC <80% pre-index event is higher than that seen in previous studies,
Alberts MJ, Peacock WF, Fields LE et al. Association between once- and twice-daily direct oral anticoagulant adherence in nonvalvular atrial fibrillation patients and rates of ischemic stroke. Int J Cardiol;215:11–13.
suggesting that poor adherence was a factor for breakthrough strokes/TIAs requiring intervention in our cohort. With adherence improving across all prescribing categories post-index event, this aligns with a practice of encouraging adherence and implementing strategies such as compliance packaging regardless of therapy prescribed. However, those that have breakthrough stroke/TIA despite good adherence may carry intrinsically strong drivers of thrombogenesis that outmatch normally sufficient levels of anticoagulation, or exhibit pharmacokinetic variability that causes inadequate anticoagulation despite guideline concordant dosing, such as poor absorption or excessive metabolism. This risk-enriched population that has already failed stroke-prevention once may benefit greatly from DOAC steady state concentration measurements to rule out pharmacokinetic reasons for failure. Calibrated, DOAC-specific assays are becoming more widely available, such that their integration into initial assessment protocols may help individualize and improve upon overall stroke/TIA treatment.
Changing to an alternative anticoagulant in the absence of factors compromising anticoagulation is a recommendation included in guidelines based on expert opinion.
For patients in our population that changed DOAC agents, switching to apixaban was the predominant strategy. Although apixaban is associated with superiority over warfarin in reducing AF-related stroke or systemic embolism, bleeding and mortality
and is one of the most used DOACs in current practice, there are no randomized, head-to-head trials to date that compare the efficacy and safety of the available DOACs. Of interest, adherence post-index event improved despite most patients being switched to apixaban, which is criticized for being more difficult to adhere to due to twice daily dosing compared to rivaroxaban.
The year of index event was highly predictive of the strategy to change to warfarin. From 2011 to 2012 in our cohort, dabigatran and warfarin were the only available oral anticoagulant agents for stroke prevention in AF. From 2013 onwards, the annual proportion of patients switching to warfarin declined due to availability of apixaban and rivaroxaban, and across the last two years of follow up, apixaban and rivaroxaban were the most common DOACs and were used at similar rates (Supplemental Figure S1)
Patients that did not fill any OACs following their index event had substantially shorter follow up times with a median of 40 days [191], during which 63% died, 32% had a recurrent stroke/TIA and 7% had a bleeding event. This categorization likely represented patients with poor prognosis.
This study has a number of limitations associated with design as a retrospective, observational study of administrative data using diagnostic codes. All characteristics of breakthrough stroke/TIA during DOAC treatment could not be accounted for, including stroke severity data, alternate stroke etiologies and suboptimal risk factor management of alcohol use, tobacco use, sleep apnea, obesity, diabetes, hypertension, and dyslipidemia; such information could better rationalize decisions to continue or change DOAC regimens. Body weight, anti-Xa level and/or thrombin time on admission would also improve evaluation of DOAC use, while non-prescription medication use would allow thorough investigation of drug-drug interactions.
CONCLUSION
In our study, only about one-third of patients filled a different OAC regimen after a breakthrough stroke or TIA, including changing the original DOAC dose (8%), changing the DOAC agent (16%) or changing to warfarin (5%). Despite different approaches to managing OAC post-stroke, rates of stroke/TIA recurrence and bleeding events were similar. Patient factors such as preference and stroke vs TIA severity may have influenced prescribing decisions. DOAC adherence was improved post-index event and should be optimized in all patients with AF. Further research is required to confirm whether choice of oral anticoagulant reduces stroke risk in patients that experience DOAC treatment failure in the absence of factors causing inadequate anticoagulation.
January CT, Wann LS, Calkins H et al. 2019 AHA/ACC/HRS Focused Update of the 2014 AHA/ACC/HRS Guideline for the Management of Patients With Atrial Fibrillation: A Report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines and the Heart Rhythm Society in Collaboration With the Society of Thoracic Surgeons. Circulation 2019;140(2):e125–e151.
Kirchhof P, Benussi S, Kotecha D et al. 2016 ESC Guidelines for the management of atrial fibrillation developed in collaboration with EACTS. Europace 2016;18(11):1609–1678.
Alberts MJ, Peacock WF, Fields LE et al. Association between once- and twice-daily direct oral anticoagulant adherence in nonvalvular atrial fibrillation patients and rates of ischemic stroke. Int J Cardiol;215:11–13.
FUNDINGSOURCES: No funding was received for this study.
DISCLOSURES:
GJ received funding from Canadian Institutes of Health Research, Heart and Stroke Foundation of Canada, and National Institutes of Health, all unrelated to the current study.
TJB received unrestricted research funding from Pfizer, and advisory board honoraria from Pfizer, all not related to this manuscript.
00115-4&id=fx1.jpg){kind=link}
00115-4&id=gr1.jpg){kind=link}
00115-4&id=gr2.jpg){kind=link}
00115-4&id=gr3.jpg){kind=link}