Introduction: Current evidence regarding the clinical outcomes of non-vitamin K oral anticoagulants (NOACs) versus warfarin in patients with atrial fibrillation (AF) and previous stroke is inconclusive, especially in patients with previous intracranial haemorrhage (ICrH). We aim to undertake a systematic review and meta-analysis assessing the effectiveness and safety of NOACs versus warfarin in AF patients with a history of stroke. Methods: We searched studies published up to December 10, 2022, on PubMed, Medline, Embase, and Cochrane Central Register of Controlled Trials. Studies on adults with AF and previous ischaemic stroke (IS) or IrCH receiving either NOACs or warfarin and capturing outcome events (thromboembolic events, ICrH, and all-cause mortality) were eligible for inclusion. Results: Six randomized controlled trials (RCTs) (including 19,489 patients with previous IS) and fifteen observational studies (including 132,575 patients with previous IS and 13,068 patients with previous ICrH) were included. RCT data showed that compared with warfarin, NOACs were associated with a significant reduction in thromboembolic events (odds ratio [OR]: 0.85, 95% confidence interval [CI]: 0.75–0.96), ICrH (OR: 0.57, 95% CI: 0.36–0.90), and all-cause mortality (OR: 0.88, 95% CI: 0.80–0.98). In analysing observational studies, similar results were retrieved. Moreover, patients with previous ICrH had a lower OR on thromboembolic events than those with IS (OR: 0.66, 95% CI: 0.46–0.95 vs. OR: 0.80, 95% CI: 0.70–0.93) in the comparison between NOACs and warfarin. Conclusions: Observational data showed that in AF patients with previous stroke, NOACs showed better clinical performance compared to warfarin and the benefits of NOACs were more pronounced in patients with previous IrCH versus those with IS. RCT data also showed NOACs are superior to warfarin. However, current RCTs only included AF patients who survived an IS, and further large RCTs focused on patients with previous ICrH are warranted.

The risk of stroke recurrence is particularly high in patients with atrial fibrillation (AF) and previous stroke [1]. Oral anticoagulation (OAC) therapy showed better clinical performance for secondary stroke prevention compared with no treatment in these patients [2‒4]. Warfarin used to be the most common OAC worldwide, but it has several limitations such as a narrow therapeutic window and the need for frequent blood tests to monitor coagulation levels regularly. Moreover, risk of warfarin-related intracranial haemorrhage (ICrH) also limits its usage, especially in patients with AF who survived after ICrH [4]. However, the situation changed with the availability of a newer class of OAC-non-vitamin K antagonists (NOACs). Previous meta-analysis showed that the rates of ICrH after NOACs were 0.55% versus 0.91% after warfarin [5]. Evidence suggest that compared to warfarin, NOACs showed comparable efficacy and superior safety of reducing ICrH risk by 50% in patients [6, 7]. As the guideline recommendation, NOACs are currently considered first-choice treatment for secondary prevention due to their comparable efficacy, better safety, and easier administration without the need for frequent blood tests [8, 9].

Despite strong recommendations on NOACs over warfarin in patients with non-valvular AF and previous ischaemic stroke (IS) for secondary prevention of all events, the effectiveness, and safety of NOACs compared with warfarin in patients with AF who survived an ICrH has rarely been evaluated. Previous systematic review and meta-analyses have reported the beneficial effects of OAC treatment on lowering the risk of IS without increasing ICrH recurrence in these populations [10, 11], but NOACs were not analysed as an anticoagulation treatment option in these studies. Therefore, data on whether NOACs are superior to warfarin in reducing the risk of recurrent ICrH in patients with a history of ICrH are lacking. In this study, a systematic review and meta-analysis was performed with the aim of comparing the effectiveness and safety of NOACs to warfarin in patients with non-valvular AF and previous IS or ICrH. We also aim to evaluate if patients with AF and a history of ICrH benefit more from NOACs when compared with patients with previous IS.

This review was conducted in accordance with the PRISMA guidelines. The review protocol was registered with the PROSPERO database of systematic reviews: CRD42022382732.

Search Strategy and Selection Criteria

The following four databases were searched for the systematic review from inception to December 10, 2022: PubMed, Medline, Embase, and Cochrane Central Register of Controlled Trials. There were no restrictions on language or duration of follow-up. Details of the search strategy are shown in the online supplementary materials (for all online suppl. material, see https://doi.org/10.1159/000534596). We included randomized controlled trials (RCTs) or cohort studies that recruited participants (aged ≥18 years) with AF and a history of IS/transient ischaemic attack or nontraumatic spontaneous ICrH of any size and any type. Patients with a diagnosis of post-operative AF, valvular AF, AF associated with mechanical valve malfunction, AF associated with mechanical complication of heart valve prosthesis, or rheumatic AF were excluded. This study chose NOAC as interventions in the experimental group. Warfarin treatment was chosen as the intervention for the control group, regardless of dose and frequency. Control groups treated with a placebo, with no intervention or antiplatelet drugs such as aspirin were excluded.

Outcomes

The outcomes of interest were thromboembolic events, ICrH, and all-cause mortality. Thromboembolic events were chosen to reflect the range of definitions used in the included studies (such as deep vein thrombosis, IS, myocardial infarction, or systemic embolism).

Data Extraction

One reviewer (M.S.) first scrutinized all titles and abstracts after removing duplicate papers and excluding clearly irrelevant articles. The remaining studies were read in full against the inclusion and exclusion criteria independently by the two investigators (M.S. and H.W.). Discussion between the two reviewers was used to resolve disagreements, and a third arbiter (Y.W.) was available if resolution could not be reached. Extracted data included study design, participant characteristics, sample size, type and doses of anticoagulant, initiation and duration of anticoagulant, outcome measurements, and length of follow-up.

Bias and Quality Assessment

The Newcastle-Ottawa Scale (NOS) tool was used to assess the quality of observational cohort studies [11] according to the selection of study groups, their comparability, and outcome assessment in the studies. The score could range between 0 and 9. Cochrane Collaboration’s tool was used for RCTs [12] for assessing risk of bias, assigning low, high, or unclear risk of bias based on the process of sequence generation, allocation concealment, blinding, data collection, and outcome reporting.

Data Synthesis and Statistical Analyses

We performed two meta-analyses; one included the observational studies comparing NOAC versus warfarin on the related outcomes, and the other included only the RCTs. This study used both random-effects model (when heterogeneity = 0) and fix-effects model (when heterogeneity >0) to conduct a typical pairwise meta-analysis for each pairwise comparison of treatments to estimate all primary and secondary outcomes as odds ratio (OR) with associated 95% confidence intervals (CIs). For Q statistics, the I2 statistic and p value were used to calculate the proportion of variability across studies that may be attributed to heterogeneity. An I2 value of less than 25% is viewed as low heterogeneity, 25–50% as moderate heterogeneity, and over 50% as high heterogeneity.

For studies with more than one follow-up point, results from the longest follow-up were included in the main analysis. With multiple publications from a single database, the study with the largest number of patients was selected to avoid duplication of data. Subgroup analyses were performed based on follow-up period (≤1 year and >1 year) and geographic locations (Asia, Europe, and North America).

The software Review Manager (RevMan) Version 5.4 was used to perform pairwise analysis, forest plots, and funnel plots. In this meta-analysis, a p value <0.05 was considered statistically significant for all comparisons.

A total of 8,448 articles were identified through the searches. After eliminating duplicate records, 5,709 articles were screened by titles and abstracts for eligibility and 268 articles remained. By full-text review, 21 articles were eventually included in the meta-analysis study. The flowchart is shown in Figure 1.

Fig. 1.

PRISMA flow diagram depicting the selection of included studies.

Fig. 1.

PRISMA flow diagram depicting the selection of included studies.

Close modal

Characteristics of the Included Studies

Randomized Controlled Trials

A total of 6 RCTs [13‒18] comprising 19,489 participants (mean age ranging from 70.1 to 79.4 years, 57.9–84.6% males, follow-up ranged from a median of 30 days to a median of 2.8 years) were included in the analysis. All the RCTs only included patients with AF and previous IS. Five studies [13, 15‒18] used CHADS2 to assess the risks of stroke recurrence. Two of them [13, 17] had a mean score >3 and the rest three [15, 16, 18] showed the proportion of patients with a score>3 ranged from 67% to 92%. The other study [14] showed that the mean HAS-BLED score <2. Baseline characteristics of each study are shown in online supplementary Table 1.

Observational Data

Fifteen observational studies [19‒33] (n = 145,643 patients) compared NOAC versus warfarin. The mean age of the patients ranged from 69.0 to 83.9 years, with male sex comprising 35.0–70.7%. Follow-up time were between a median of 16.1 days to a median of 5.4 years. Six studies [19, 25‒27, 31, 33] included participants with AF and a history of ICrH (one study only included AF patients with previous intracerebral haemorrhage (ICH) [31] and the other 4 studies included patients with both intracerebral and other types of ICrH), while nine studies [20‒24, 28‒30, 32] focused on AF patients with previous IS. One of the studies [33] included both patients with previous ICrH and those with previous IS. The risks of stroke recurrence differed significantly across the studies, as shown either by the mean (SD) [20‒22, 25‒29, 33] or the median (IQR) [19, 33] of CHA2DS2-VASc score. The medians of score ranged from 4 to 6 and the means were between 2.3 and 6.8. The risks of bleeding shown by HAS-BLED score varied from <2 to >4. NIHSS score were reported in 5 studies [20‒22, 28, 29] to assess stroke severity. There was a great difference in the medians of the score with a range from 3 to 11. Baseline characteristics of each study are shown in online suppl. Table 2.

Outcome of Interest

Randomized Controlled Trials

Thromboembolic Events. Six studies [13‒18] with 19,489 participants reported on the outcome of IS or systemic embolism. The results showed NOAC treatment was associated with a significantly reduced risk of thromboembolic events compared with warfarin (OR: 0.85, 95% CI [0.75–0.96]; p = 0.01; I2 = 0%) (shown in Fig. 2).

Fig. 2.

Meta-analysis of RCT data: forest plot depicting the risks of unfavourable outcomes (thromboembolic events, incident ICrH, and mortality) in patients with AF and previous IS receiving NOAC versus warfarin.

Fig. 2.

Meta-analysis of RCT data: forest plot depicting the risks of unfavourable outcomes (thromboembolic events, incident ICrH, and mortality) in patients with AF and previous IS receiving NOAC versus warfarin.

Close modal

Incident ICrH. Five studies [14‒18], including 18,677 participants, reported on the outcome of incident ICrH. The pooled analysis revealed a significant reduction in incident ICrH with NOAC compared with warfarin (OR: 0.57, 95% CI [0.36–0.90]; p = 0.02; I2 = 70%) (shown in Fig. 2).

All-Cause Mortality. Five studies [14‒18] included 18,677 participants who reported on all-cause mortality. The present results found that NOAC use was associated with a reduced risk of mortality compared with warfarin treatment (OR: 0.88, 95% CI [0.80–0.98]; p = 0.02; I2 = 0%) (shown in Fig. 2).

Subgroup Analysis. When stratifying the RCTs by follow-up, 2 studies [13, 14] (n = 995) at follow-up ≤1 year and 4 studies [15‒18] (n = 18,494) at >1 year compared NOAC versus warfarin on thromboembolic events, one study [14] (n = 183) at follow-up ≤1 year and 4 studies [15‒18] (n = 18,494) at >1 year on incident ICrH, and one study [14] (n = 183) at follow-up ≤1 year and 4 studies [15‒18] (n = 18,494) at >1 year on all-cause mortality. The pooled ORs demonstrate that NOAC treatment significantly reduced the risks of thromboembolic events (OR: 0.86, 95% CI [0.75–0.98]; p = 0.02; I2 = 0%) and incident ICrH (OR: 0.49, 95% CI [0.32–0.75]; p = 0.001; I2 = 60%) at >1 year follow-up, while at ≤1 year follow-up, the efficacy was comparable in these two treatments (thromboembolic events, OR: 0.72, 95% CI [0.45–1.15]; p = 0.17; I2 = 0%; incident ICrH, OR: 1.16, 95% CI [0.62–2.19]; p = 0.64). NOAC therapy was also associated with lower risks of all-cause mortality compared to warfarin at >1 year follow-up. There was no observation of mortality in both NOAC treatment and warfarin treatment in the analysis of ≤1-year follow-up. Results are shown in Figure 3. A subgroup analysis by geographic locations cannot be performed since four out of six RCTs [15‒18] were multi-centre studies which included patients from several countries, and the rest two studies only included Asian participants.

Fig. 3.

Meta-analysis of RCT data: forest plot depicting the risks of unfavourable outcomes (thromboembolic events, incident ICrH, and mortality) in patients with AF and previous IS receiving NOAC versus warfarin by follow-up periods.

Fig. 3.

Meta-analysis of RCT data: forest plot depicting the risks of unfavourable outcomes (thromboembolic events, incident ICrH, and mortality) in patients with AF and previous IS receiving NOAC versus warfarin by follow-up periods.

Close modal

Observational Data

Thromboembolic Events. Thirteen studies [19‒30, 32] with 142,304 participants (11,412 patients with a history of ICrH and 130,892 patients with a history of IS) reported on the outcome of any thromboembolic events. Four studies [19, 25‒27] included patients with previous ICrH, and nine studies [20‒24, 28‒30, 32] reported on patients with a history of IS. The results revealed a significant reduction in thromboembolic events with NOAC compared with warfarin (OR: 0.76, 95% CI [0.66–0.87]; p < 0.001; I2 = 78%) (shown in Fig. 4). Moreover, AF patients with previous IS could have a 20% reduction in thromboembolic events after receiving NOAC (OR: 0.80, 95% CI [0.70–0.93]; p = 0.003; I2 = 79%), while a 34% decrease in thromboembolic events was observed in AF patients with a history of ICrH (OR: 0.66, 95% CI [0.46–0.95]; p = 0.03; I2 = 75%) (shown in Fig. 4).

Fig. 4.

Meta-analysis of observational data: forest plot depicting the risks of unfavourable outcomes (thromboembolic events, incident ICrH, and mortality) in patients with AF and previous stroke receiving NOAC versus warfarin.

Fig. 4.

Meta-analysis of observational data: forest plot depicting the risks of unfavourable outcomes (thromboembolic events, incident ICrH, and mortality) in patients with AF and previous stroke receiving NOAC versus warfarin.

Close modal

Intracranial Haemorrhage. Ten studies [19, 20, 22, 23, 25‒28, 30, 31], including 119,817 participants, reported on the outcome of ICrH. Five studies [19, 25‒27, 31] included 11,756 patients with an index ICrH (intracerebral, subdural, subarachnoid, or epidural haemorrhage), and five studies [20, 22, 23, 28, 30] included 108,061 patients with a history of IS. Our results showed that NOAC treatment was associated with a significant decrease in the risk of IrCH recurrence compared to warfarin therapy in AF patients with previous stroke (OR: 0.55, 95% CI [0.44–0.68]; p < 0.001; I2 = 36%). NOAC resumption was associated with a 47% reduction of recurrent ICrH in AF patients with previous haemorrhagic stroke (OR: 0.53, 95% CI [0.40–0.70]; p < 0.001; I2 = 25%), which was similar to the effects (44% reduction in the risk of ICH) in patients with AF who survived IS (OR: 0.56, 95% CI [0.38–0.81]; p = 0.002; I2 = 34%). Results are presented in Figure 4.

All-Cause Mortality. Eleven studies [19, 20, 22, 23, 25, 27‒31, 33] including 133,949 participants (12,473 patients with a history of ICrH and 121,476 patients with a history of IS) reported on all-cause mortality. Six studies [20, 22, 23, 28–30] included patients with previous IS, four studies [19, 25, 27, 31] included patients with previous ICrH, and 1 study included both patients with a history of IS and those with ICrH [33]. The present results indicated that NOAC therapy could reduce mortality in AF patients with a history of stroke by 45% compared to warfarin therapy (OR: 0.55, 95% CI [0.44–0.70]; p < 0.001; I2 = 92%) (shown in Fig. 4). The reduced mortality from NOAC therapy is similar in AF patients with previous ICrH (OR: 0.58, 95% CI [0.40–0.83]; p = 0.003; I2 = 89%) and in patients with previous IS (OR: 0.53, 95% CI [0.39–0.74]; p < 0.001; I2 = 93%) (shown in Fig. 4).

Subgroup Analysis. Subgroup analyses were performed according to follow-up period (follow-up ≤1 year vs. follow-up >1 year) and geographic locations (Asia vs. Europe vs. North America). Seven studies [19, 20, 23, 26, 27, 30, 32] with 113,832 participants reported on any thromboembolic events at follow-up ≤1 year, while six studies [21, 22, 24‒26, 29] with 24,182 participants at follow-up >1 year. Five studies [19, 20, 23, 27, 30] reported on ICrH at follow-up ≤1 year, three studies [22, 25, 31] at follow-up >1 year, and one study [26] reported on the outcome at both periods. In total, 108,905 patients reported on ICrH at follow-up ≤1 year and 6,622 patients at >1 year. In terms of mortality, results at follow-up ≤1 year were reported in six studies [19, 20, 23, 27, 30, 33] (n = 111,375) and follow-up >1 year were reported in four studies [22, 25, 29, 31] (n = 129,037). The results showed that NOAC treatment was associated with a significant reduction in the risk of ICrH and all-cause mortality at both follow-up ≤1 year (ICrH: OR: 0.62, 95% CI [0.53–0.73]; p < 0.001; I2 = 4%; mortality: OR: 0.72, 95% CI [0.62–0.83]; p < 0.001; I2 = 45%) and follow-up >1 year (ICrH: OR: 0.61, 95% CI [0.42–0.90]; p = 0.01; I2 = 24%; mortality: OR: 0.42, 95% CI [0.26–0.68]; p < 0.001; I2 = 96%). NOAC had a significant decreased risks in thromboembolic events at follow-up ≤1 year (OR: 0.65, 95% CI [0.54–0.77]; p < 0.001; I2 = 72%), while this was not significant at follow-up >1 year (OR: 0.91, 95% CI [0.79–1.06]; p = 0.22; I2 = 44%). Results are shown in Figure 5.

Fig. 5.

Meta-analysis of observational data: forest plot depicting the risks of unfavourable outcomes (thromboembolic events, incident ICrH, and mortality) in patients with AF and previous stroke receiving NOAC versus warfarin by follow-up periods.

Fig. 5.

Meta-analysis of observational data: forest plot depicting the risks of unfavourable outcomes (thromboembolic events, incident ICrH, and mortality) in patients with AF and previous stroke receiving NOAC versus warfarin by follow-up periods.

Close modal

Seven studies [19, 21‒23, 25, 27, 30] included 113,481 participants from Asia, three studies [20, 24, 26] with 7,322 patients in Europe, and two studies [29, 32] with 16,589 participants from North America were included in the pooled analysis of assessing the thromboembolic events after NOAC treatment versus warfarin. Regarding ICrH, results on Asians were reported in six studies [19, 22, 23, 25, 27, 30] with 113,380 patients, and on Europeans were reported in three studies [20, 26, 31] with 1,525 participants. No study in North America was reported on ICrH. Six studies [19, 22, 23, 25, 27, 30] with 113,380 participants from Asia, three studies [20, 31, 33] included 3,995 patients, and one study [29] with 11,662 patients reported on all-cause mortality. The results demonstrated that NOAC treatment significantly reduce the risks of thromboembolic events (OR: 0.71, 95% CI [0.57–0.89]; p = 0.003; I2 = 69%), ICrH (OR: 0.56, 95% CI [0.45–0.70]; p < 0.001; I2 = 31%), and mortality (OR: 0.44, 95% CI [0.30–0.66]; p < 0.001; I2 = 94%) for Asian patients, while the observed reduced risks were not statistically significant for European patients (thromboembolic events: OR: 1.00, 95% CI [0.77–1.29]; p = 0.98; I2 = 25%; ICrH: OR: 0.79, 95% CI [0.49–1.28]; p = 0.34; I2 = 0%; mortality: OR: 0.88, 95% CI [0.73–1.06]; p = 0.17; I2 = 0%) for European patients. NOAC treatment was also associated with decreased risks of thromboembolic events (OR: 0.74, 95% CI [0.55–1.00]; p = 0.05; I2 = 92%) for patients in North America. Meta-analyses were not conducted in North America on ICrH and mortality because the data were not available or not sufficient for meta-analysis (no study on ICrH and only one study on mortality). Results are shown in Figure 6.

Fig. 6.

Meta-analysis of observational data: forest plot depicting the risks of unfavourable outcomes (thromboembolic events, incident ICrH, and mortality) in patients with AF and previous stroke receiving NOAC versus warfarin by geographic locations.

Fig. 6.

Meta-analysis of observational data: forest plot depicting the risks of unfavourable outcomes (thromboembolic events, incident ICrH, and mortality) in patients with AF and previous stroke receiving NOAC versus warfarin by geographic locations.

Close modal

Risk of Bias Assessment

Risk of bias was generally low in all studies. Among the 15 studies, low risk of bias was assigned to scale items ranging from 7 out of 9 to 9 out of 9 items. All 16 studies had low risk for ascertainment of exposure and assessment of outcome. One study scored 9, five studies scored 8 and the remaining nine studies scored 7. Low risk of bias of blinding outcome assessments was presented on all the RCTs except one. One study showed high risk of bias on blinding of participants, while the remaining 5 studies presented low risk of bias. In total, one study was of very high quality with low risk of bias in 6 items and one study was of low quality with low risk of bias in only 3 items. The other 4 studies showed low risk of bias in 5 items. The overall risk of bias assessment for all included online suppl. Table 3, Figures 1 and 2.

In the present meta-analysis, the pooled estimates of observational data revealed that NOAC use was associated with reduced risks of thromboembolic events, ICrH, and all-cause mortality in AF patients with previous stroke. The benefits of reduction in thromboembolic events were more pronounced in patients with previous ICrH in than those with previous IS after receiving NOACs. NOAC therapy significantly reduced the risk of thromboembolic events at ≤1 year follow-up but not at >1 year. The risks of ICrH and mortality did not differ by follow-up time. Analysing the RCT data, only patients with previous IS were included in the pooled estimates. The pooled ORs demonstrated that NOAC therapy was associated with significant reduction in thromboembolic events, incident ICrH, and all-cause mortality. The reduced risks of ICrH and mortality were observed at both ≤1 year and >1 year follow-up, which is similar to that in the pooled analyses of observational studies. In contrast to the results from observational data, the significant reduction in thromboembolic events was only observed at follow-up >1 year. Meta-analysis by geographic locations showed that the benefits of NOAC over warfarin were presented in Asian and American patients, while in European patients the efficacy was not significant.

Both pooled estimates of RCT data and observational studies found that NOACs are more effective than warfarin in preventing thromboembolic events and ICrH, which is consistent with current guidelines that NOAC is superior to warfarin in patients with AF and previous IS [8, 9]. The net benefits of NOACs in combination with the fact that international normalised ratio monitor, dose adjustment, and dietary restrictions are required for warfarin have made NOACs a better choice for stroke prevention in AF patients with previous IS. However, regarding the use of NOACs in patients with previous ICrH, there is no completed phase 3 RCT to prove its efficacy and safety. Currently, therefore, meta-analysis of observational studies provides the best evidence. Our review found that NOAC use was associated with a significantly decreased risk of thromboembolic events, recurrent ICH, and mortality. Most of previous meta-analyses investigated the safety and efficacy of restarting OAC therapy after ICrH with only warfarin used in most included studies [10, 11]. As NOACs become more widely available, research comparing NOAC and warfarin in patients with AF and previous ICrH is warranted to guide clinical practice. Recently, one review compared the effect of NOACs versus warfarin on recurrent ICH in AF patients with a history of ICrH and only included 3 observational studies with 8,711 participants [3]. In the present study, we updated the previous reviews with more studies and more patients included. Additionally, subgroup analyses by follow-up time and geographic locations were performed.

The benefits in reducing the risks of recurrent ICrH and all-cause mortality did not differ by follow-up time in both pooled analyses of RCT data and observational data. However, the reduction in thromboembolic events was only observed at ≤1 year follow-up in observational data while in RCT data, this reduction was only seen at >1 year follow-up. In RCT data, only two studies [13, 14] with a total of about 100 events were included in the pooled analysis at follow-up <1 year, which was not sufficiently powered to evaluate the efficacy. Moreover, in one [12] out of these two studies, most of the recurrent ischaemic lesions were asymptomatic ones on results of MRI within 4 weeks after first stroke, which may not be captured in population-based observational studies. In the pooled analysis of observational studies, the superior efficacy of NOACs was only observed at follow-up <1 year. Possible reasons include heterogeneity in stroke severity, variation in initial timing of OAC, and the possibility that high mortality rates in the warfarin group might have concealed the occurrence of the thromboembolic events. Although the benefits in reduction of thromboembolic events in subgroup analysis by follow-up period showed conflicting results, it also suggests a clear trend of an appealing effectiveness profile for NOACs in comparison with warfarin in the present study. In addition, we should note that these findings were from a combination of observational studies included AF patients with previous ICrH and previous IS, while the results from RCT data only included AF patients with previous IS. A recent meta-analysis (pooled analysis of a combination of observational and RCT data) comparing OAC (NOAC, warfarin, etc.) and no therapy in AF patients who survived an ICrH found the superior performance of OAC in reducing thromboembolic events was only observed at ≤1 year follow-up [3]. Moreover, most of the included studies in that review [3] were observational. In the present study, the number of observation studies on AF patients with previous ICrH is not large enough to stratify the studies by follow-up, and there is no phase 3 RCT on patients with ICrH. Therefore, further studies focus on AF patients survived an ICrH are warranted.

It is interesting that NOAC-associated reduced risks in thromboembolic events, ICrH, and all-cause mortality were shown in Asian patients, but these benefits were not statistically significant in European patients. According to previous literature, risk of warfarin-related ICrH is higher in the Asian population compared with non-Asians and NOACs seem to have a greater relative risk reduction of ICH in the Asian population than in non-Asians [34, 35]. This may partially account for the pronounced benefits of NOAC over warfarin in Asian patients. Another explanation may be the small number of ICrH cases in European patients, which results in limited statistical power to achieve significant.

To the best of our knowledge, this is the first meta-analysis comparing the effectiveness and safety of NOAC in AF patients with a prior history of ICH and a history of IS. It is interesting that the benefits of NOACs were more pronounced in patients with previous IrCH versus those with IS. ICrH survivors are at a high risk of not only sustaining haemorrhage but also experiencing further IS. Moreover, the risk of recurrent IS is even higher in patients with AF and a history of ICrH compared with those without ICrH [9]. NOACs showed better efficacy in anticoagulation and superior safety while reducing ICrH risk by 50% compared with warfarin [6, 7]. Therefore, patients with previous ICrH, who may be more prone to have recurrent any stroke than patients with IS, would have more benefits with the treatment of NOACs.

Current recommendations to inform optimal timing of anticoagulation after both IS and ICrH are based on expert consensus. The European Society of Cardiology (ESC) for the management of AF in patients who suffer a moderate-to-severe IS recommended that anticoagulation treatment should be interrupted for 3–12 days to allow a multidisciplinary assessment of acute stroke and bleeding risk [36]. For AF patients after ICrH, the optimal timing of anticoagulation should be delayed beyond the acute phase, probably for at least 4 weeks [36]. In the present study, we found only one study out of six ICrH studies reported the initiation time of NOAC (within 90 days after ICrH). To guide optimal timing of OAC initiation after an ICrH in patients with AF, well-designed RCTs are warranted.

Strength and Limitations

The key limitation of this review is that only observational cohort studies were included in the pooled estimates of comparing NOAC versus warfarin in AF patients with a history of ICrH. The clinical and methodological heterogeneity in non-randomized studies limits the results to general population. For example, differences in timing and dosage of re(initiation) of OAC therapy may result from factors associated with future bleeding risk such as age, stroke severity, or size of the haematoma. These specific factors are instances of confoundings by indication in observational studies, in which patients at higher perceived risks may be less likely to be restarted on NOAC or warfarin. Several ongoing phase 3 RCTs are comparing the efficacy and safety of NOAC versus warfarin for stroke prevention in patients with AF who survived an ICrH [37‒39]. It will be critical to better understand the benefits of NOACs in this patient population. Second, there was selection bias of individual studies in that some included only patients with intraparenchymal haemorrhage, while others reported on both intracerebral and other types of ICH combined. ICH is reported to be associated with a higher risk of thrombotic events than subarachnoid haemorrhage, while recurrence of subarachnoid haemorrhage is considered rare [40]. Third, the number of studies on AF patients with previous ICrH is not large enough to stratify the studies by follow-up. In addition, information on blood pressure control was not available, which is an important factor for ICH recurrence [41]. A strength of our study is the inclusion of more studies for meta-analysis. The greater number of included studies enabled us to undertake more subgroup analyses than previous studies. All the systematic reviews identified were hand-searched for relevant studies, which decreased the number of missed studies.

Meta-analysis of observational studies evaluating the effectiveness and safety of NOACs suggests that compared with warfarin; NOACs are associated with lower risks of thromboembolic events, recurrent ICH, and all-cause mortality in both AF patients with a history of IS and patients with previous ICH. Moreover, the benefits of reduction in thromboembolic events were more pronounced in patients with previous ICrH than in those with previous IS after receiving NOACs. The pooled analysis of RCT data also demonstrated the superior efficacy and safety of NOACs to warfarin in AF patients with previous IS. However, no phase 3 RCT was completed assessing the benefits of NOAC in AF patients with previous ICH. Because of the limitations of observational studies, further evidence from RCTs is warranted to better guide clinicians in making informed decisions.

The authors have no conflicts of interest to declare.

H.W. and Y.W. received funding from the European Union’s Horizon 2020 research and innovation programme under Grant agreement No. 754517. L.L. received financial support from the China Scholarship Council PhD Scholarship (CSC No. 202108310074). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Minglei Shi, conceptualization, data curation, formal analysis, and writing – original draft; Lu Liu, conceptualization, funding acquisition, and writing – review and editing; Hatem Wafa, funding acquisition and data curation; Vasa Curcin, conceptualization, supervision, and writing – review and editing; and Yanzhong Wang, conceptualization, funding acquisition, supervision, and writing – review and editing.

Additional Information

Minglei Shi and Lu Liu contributed equally to this work.Minglei Shi and Lu Liu are joint first authors.Vasa Curcin and Yanzhong Wang contributed equally to this work.Vasa Curcin and Yanzhong Wang are joint senior authors.

1.
Atrial Fibrillation Investigators
.
Risk factors for stroke and efficacy of antithrombotic therapy in atrial fibrillation: analysis of pooled data from five randomized controlled trials
.
Arch Intern Med
.
1994
;
154
:
1449
57
.
2.
January
CT
,
Wann
LS
,
Calkins
H
,
Chen
LY
,
Cigarroa
JE
,
Cleveland
JC
Jr
.
2019AHA/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
51
.
3.
Ivany
E
,
Ritchie
LA
,
Lip
GYH
,
Lotto
RR
,
Werring
DJ
,
Lane
DA
.
Effectiveness and safety of antithrombotic medication in patients with atrial fibrillation and intracranial hemorrhage: systematic review and meta-analysis
.
Stroke
.
2022
;
53
(
10
):
3035
46
.
4.
Diener
HC
,
Hankey
GJ
,
Easton
JD
,
Lip
GYH
,
Hart
RG
,
Caso
V
.
Non-vitamin K oral anticoagulants for secondary stroke prevention in patients with atrial fibrillation
.
Eur Heart J Suppl
.
2020
22
Suppl I
I13
I21
.
5.
Jiang
H
,
Jiang
Y
,
Ma
H
,
Zeng
H
,
Lv
J
.
Effects of rivaroxaban and warfarin on the risk of gastrointestinal bleeding and intracranial hemorrhage in patients with atrial fibrillation: systematic review and meta-analysis
.
Clin Cardiol
.
2021
;
44
(
9
):
1208
15
.
6.
Ruff
CT
,
Giugliano
RP
,
Braunwald
E
,
Hoffman
EB
,
Deenadayalu
N
,
Ezekowitz
MD
.
Comparison of the efficacy and safety of new oral anticoagulants with warfarin in patients with atrial fibrillation: a meta-analysis of randomised trials
.
Lancet
.
2014
;
383
(
9921
):
955
62
.
7.
Lip
G
,
Freedman
B
,
De Caterina
R
,
Potpara
TS
.
Stroke prevention in atrial fibrillation: past, present and future. Comparing the guidelines and practical decision-making
.
Thromb Haemost
.
2017
;
117
(
7
):
1230
9
.
8.
Klijn
CJ
,
Paciaroni
M
,
Berge
E
,
Korompoki
E
,
Kõrv
J
,
Lal
A
.
Antithrombotic treatment for secondary prevention of stroke and other thromboembolic events in patients with stroke or transient ischemic attack and non-valvular atrial fibrillation: a European Stroke Organisation guideline
.
Eur Stroke J
.
2019
;
4
(
3
):
198
223
.
9.
Hindricks
G
,
Potpara
T
,
Dagres
N
,
Arbelo
E
,
Bax
JJ
,
Blomström-Lundqvist
C
.
2020 ESC Guidelines for the diagnosis and management of atrial fibrillation developed in collaboration with the European Association for Cardio-Thoracic Surgery (EACTS): the Task Force for the diagnosis and management of atrial fibrillation of the European Society of Cardiology (ESC) Developed with the special contribution of the European Heart Rhythm Association (EHRA) of the ESC
.
Eur Heart J
.
2021
;
42
(
5
):
373
498
.,
10.
Murthy
SB
,
Gupta
A
,
Merkler
AE
,
Navi
BB
,
Mandava
P
,
Iadecola
C
.
Restarting anticoagulant therapy after intracranial hemorrhage: a systematic review and meta-analysis
.
Stroke
.
2017
;
48
(
6
):
1594
600
.
11.
Korompoki
E
,
Filippidis
FT
,
Nielsen
PB
,
Del Giudice
A
,
Lip
GYH
,
Kuramatsu
JB
.
Long- term antithrombotic treatment in intracranial hemorrhage survivors with atrial fibrillation
.
Neurology
.
2017
;
89
(
7
):
687
96
.
12.
Higgins
JPT
,
Green
S
, editors.
Cochrane handbook for systematic reviews of interventions V Version 5.1.0(updated March 2011)
The Cochrane Collaboration
2011
. www.handbook.cochrane.org.
13.
Tanahashi
N
,
Hori
M
,
Matsumoto
M
,
Momomura
S
,
Uchiyama
S
,
Goto
S
.
Rivaroxaban versus warfarin in Japanese patients with nonvalvular atrial fibrillation for the secondary prevention of stroke: a subgroup analysis of J-ROCKET AF
.
J Stroke Cerebrovasc Dis
.
2013
;
22
(
8
):
1317
25
.
14.
Hong
K-S
,
Kwon
SU
,
Lee
SH
,
Lee
JS
,
Kim
YJ
,
Song
TJ
.
Rivaroxaban vs warfarin sodium in the ultra-early period after atrial fibrillation–related mild ischemic stroke: a randomized clinical trial
.
JAMA Neurol
.
2017
;
74
(
10
):
1206
15
.
15.
Diener
H-C
,
Connolly
SJ
,
Ezekowitz
MD
,
Wallentin
L
,
Reilly
PA
,
Yang
S
.
Dabigatran compared with warfarin in patients with atrial fibrillation and previous transient ischaemic attack or stroke: a subgroup analysis of the RE-LY trial
.
Lancet Neurol
.
2010
;
9
(
12
):
1157
63
.
16.
Easton
JD
,
Lopes
RD
,
Bahit
MC
,
Wojdyla
DM
,
Granger
CB
,
Wallentin
L
.
Apixaban compared with warfarin in patients with atrial fibrillation and previous stroke or transient ischaemic attack: a subgroup analysis of the ARISTOTLE trial
.
Lancet Neurol
.
2012
;
11
(
6
):
503
11
.
17.
Hankey
GJ
,
Patel
MR
,
Stevens
SR
,
Becker
RC
,
Breithardt
G
,
Carolei
A
.
Rivaroxaban compared with warfarin in patients with atrial fibrillation and previous stroke or transient ischaemic attack: a subgroup analysis of ROCKET AF
.
Lancet Neurol
.
2012
;
11
(
4
):
315
22
.
18.
Rost
NS
,
Giugliano
RP
,
Ruff
CT
,
Murphy
SA
,
Crompton
AE
,
Norden
AD
.
Outcomes with edoxaban versus warfarin in patients with previous cerebrovascular events: findings from ENGAGE AF-TIMI 48 (effective anticoagulation with factor xa next generation in atrial fibrillation-thrombolysis in myocardial infarction 48)
.
Stroke
.
2016
;
47
(
8
):
2075
82
.
19.
Lee
S-R
,
Choi
EK
,
Kwon
S
,
Jung
JH
,
Han
KD
,
Cha
MJ
.
Oral anticoagulation in Asian patients with atrial fibrillation and a history of intracranial hemorrhage
.
Stroke
.
2020
;
51
(
2
):
416
23
.
20.
D’Anna
L
,
Filippidis
FT
,
Antony
S
,
Brown
Z
,
Wyatt
H
,
Malik
A
.
Early initiation of direct anticoagulation after stroke in patients with atrial fibrillation
.
Eur J Neurol
.
2020
;
27
(
11
):
2168
75
.
21.
Kanai
Y
,
Oguro
H
,
Tahara
N
,
Matsuda
H
,
Takayoshi
H
,
Mitaki
S
.
Analysis of recurrent stroke volume and prognosis between warfarin and four non–vitamin K antagonist oral anticoagulants' administration for secondary prevention of stroke
.
J Stroke Cerebrovasc Dis
.
2018
;
27
(
2
):
338
45
.
22.
Yoshimura
S
,
Koga
M
,
Sato
S
,
Todo
K
,
Yamagami
H
,
Kumamoto
M
.
Two-year outcomes of anticoagulation for acute ischemic stroke with nonvalvular atrial fibrillation: SAMURAI-NVAF study
.
Circ J
.
2018
;
82
(
7
):
1935
42
.
23.
Kumazawa
R
,
Jo
T
,
Matsui
H
,
Fushimi
K
,
Yasunaga
H
.
Direct oral anticoagulants versus warfarin for secondary prevention of cerebral infarction and bleeding in older adults with atrial fibrillation
.
J Am Geriatr Soc
.
2022
;
70
(
7
):
2029
39
.
24.
Larsen
TB
,
Rasmussen
LH
,
Gorst-Rasmussen
A
,
Skjøth
F
,
Lane
DA
,
Lip
GYH
.
Dabigatran and warfarin for secondary prevention of stroke in atrial fibrillation patients: a nationwide cohort study
.
Am J Med
.
2014
;
127
(
12
):
1172
8.e5
.
25.
Tsai
C-T
,
Liao
JN
,
Chiang
CE
,
Lin
YJ
,
Chang
SL
,
Lo
LW
.
Association of ischemic stroke, major bleeding, and other adverse events with warfarin use vs non: vitamin K antagonist oral anticoagulant use in patients with atrial fibrillation with a history of intracranial hemorrhage
.
JAMA Netw Open
.
2020
;
3
(
6
):
e206424
.
26.
Nielsen
PB
,
Skjøth
F
,
Søgaard
M
,
Kjældgaard
JN
,
Lip
GYH
,
Larsen
TB
.
Non–vitamin K antagonist oral anticoagulants versus warfarin in atrial fibrillation patients with intracerebral hemorrhage
.
Stroke
.
2019
;
50
(
4
):
939
46
.
27.
Lin
SY
,
Chang
YC
,
Lin
FJ
,
Tang
SC
,
Dong
YH
,
Wang
CC
.
Post-intracranial hemorrhage antithrombotic therapy in patients with atrial fibrillation
.
J Am Heart Assoc
.
2022
;
11
(
6
):
e022849
.
28.
Seiffge
DJ
,
Paciaroni
M
,
Wilson
D
,
Koga
M
,
Macha
K
,
Cappellari
M
.
Direct oral anticoagulants versus vitamin K antagonists after recent ischemic stroke in patients with atrial fibrillation
.
Ann Neurol
.
2019
;
85
(
6
):
823
34
.
29.
Xian
Y
,
Xu
H
,
O’Brien
EC
,
Shah
S
,
Thomas
L
,
Pencina
MJ
.
Clinical effectiveness of direct oral anticoagulants vs warfarin in older patients with atrial fibrillation and ischemic stroke: findings from the Patient-centered Research into Outcomes Stroke Patients prefer and Effectiveness Research (PROSPER) study
.
JAMA Neurol
.
2019
;
76
(
10
):
1192
202
.
30.
Yokoyama
M
,
Mizuma
A
,
Terao
T
,
Tanaka
F
,
Nishiyama
K
,
Hasegawa
Y
.
Effectiveness of nonvitamin K antagonist oral anticoagulants and warfarin for preventing further cerebral microbleeds in acute ischemic stroke patients with nonvalvular atrial fibrillation and at least one microbleed: CMB-NOW multisite pilot trial
.
J Stroke Cerebrovasc Dis
.
2019
;
28
(
7
):
1918
25
.
31.
Poli
D
,
Antonucci
E
,
Vignini
E
,
Martinese
L
,
Testa
S
,
Simioni
P
.
Anticoagulation resumption after intracranial hemorrhage in patients treated with VKA and DOACs
.
Eur J Intern Med
.
2020
;
80
:
73
7
.
32.
Yang
L
,
Brooks
MM
,
Glynn
NW
,
Zhang
Y
,
Saba
S
,
Hernandez
I
.
Real-world direct comparison of the effectiveness and safety of apixaban, dabigatran, rivaroxaban, and warfarin in Medicare beneficiaries with atrial fibrillation
.
Am J Cardiol
.
2020
;
126
:
29
36
.
33.
Komen
JJ
,
Forslund
T
,
Mantel-Teeuwisse
AK
,
Klungel
OH
,
von Euler
M
,
Braunschweig
F
.
Association of preceding antithrombotic therapy in atrial fibrillation patients with ischaemic stroke, intracranial haemorrhage, or gastrointestinal bleed and mortality
.
Eur Heart J Cardiovasc Pharmacother
.
2021
;
7
(
1
):
3
10
.
34.
Shen
AY
,
Yao
JF
,
Brar
SS
,
Jorgensen
MB
,
Chen
W
.
Racial/ethnic differences in the risk of intracranial hemorrhage among patients with atrial fibrillation
.
J Am Coll Cardiol
.
2007
;
50
(
4
):
309
15
.
35.
Wang
KL
,
Lip
GY
,
Lin
SJ
,
Chiang
CE
.
Non-vitamin K antagonist oral anticoagulants for stroke prevention in asian patients with nonvalvular atrial fibrillation: meta-analysis
.
Stroke
.
2015
;
46
(
9
):
2555
61
.
36.
Hindricks
G
,
Potpara
T
,
Dagres
N
,
Arbelo
E
,
Bax
JJ
,
Blomström-Lundqvist
C
.
2020 Esc guidelines for the diagnosis and management of atrial fibrillation developed in collaboration with the European Association for Cardio-Thoracic Surgery (EACTS): the task force for the diagnosis and management of atrial fibrillation of the European Society of Cardiology (ESC) developed with the special contribution of the European Heart Rhythm Association (EHRA) of the esc
.
Eur Heart J
.
2021
;
42
(
5
):
373
498
.
37.
ClinicalTrials.gov
. NCT03907046. Anticoagulation in ICH survivors for stroke prevention and recovery (ASPIRE).
2019
.
38.
ClinicalTrials.gov
. NCT03996772. Prevention of stroke in intracerebral haemorrhage survivors with atrial fibrillation (PRESTIGE-AF).
2019
.
39.
ClinicalTrials.gov
. NCT03950076. Edoxaban for intracranial hemorrhage survivors with atrial fibrillation (ENRICH-AF).
2019
.
40.
Simard
JM
,
Aldrich
EF
,
Schreibman
D
,
James
RF
,
Polifka
A
,
Beaty
N
.
Low-dose intravenous heparin infusion in patients with aneurysmal subarachnoid hemorrhage: a preliminary assessment
.
J Neurosurg
.
2013
;
119
(
6
):
1611
9
.
41.
Biffi
A
,
Anderson
CD
,
Battey
TW
,
Ayres
AM
,
Greenberg
SM
,
Viswanathan
A
.
Association between blood pressure control and risk of recurrent intracerebral hemorrhage
.
JAMA
.
2015
;
314
(
9
):
904
12
.
42.
Wells
GA
,
Shea
JB
,
O’Connell
D
,
Peterson
J
The Newcastle-Ottawa Scale (NOS) for assessing the quality of nonrandomized studies in meta-analyses
Oxford
2000
.