Catheter ablation and cognitive function in atrial fibrillation: A systematic review and meta-analysis

Background Atrial fibrillation (AF) is a risk factor for cognitive dysfunction. Although catheter ablation (CA) is one of the main treatments for AF, whether it can improve cognitive function in patients with AF remains unclear. We conducted a systematic review and meta-analysis to evaluate the cognitive outcome post-CA procedure. Methods Two investigators independently searched the PubMed, EMBASE, Web of Science, CNKI, WanFang, and VIP databases from inception to September 2021 for all the potentially eligible studies. The outcomes of interest included dementia or cognitive disorder through scoring or recognized classification criteria. Heterogeneity was determined by using Cochrane's Q test and calculating the I2. A random-effects model was used to incorporate the potential effects of heterogeneity. The Newcastle-Ottawa Scale (NOS) was used to assess the methodological quality of each included study, and the Grading of Recommendations Assessment, Development and Evaluation (GRADE) method was adopted to evaluate the quality of evidence. Result Thirteen studies including 40,868 patients were included, among which 12,086 patients received AF ablation. Meta-analysis indicated that patients with AF ablation had a lower risk of dementia incidence in comparison to patients with AF without ablation [hazard ratio (HR): 0.60, 95% CI: 0.43 to 0.84, p = 0.003 I2 = 40%]. Significant differences were observed in the incidence of new-onset dementia [risk ratio (RR): 0.43, 95% CI: 0.28 to 0.65, p < 0.0001 I2 = 84%]; the changes in the Montreal Cognitive Assessment (MoCA) score [weighted mean difference (WMD): 1.00, 95% CI: 0.36 to 1.64, p < 0.005 I2 = 0%] and Mini-Mental State Examination (MMSE) score (WMD: 0.98, 95% CI: 0.69 to 1.26, p < 0.00001 I2 = 0%]. However, in subgroup analysis, we did not observe significant changes in MoCA score at < 3 months (WMD: 1.20, 95% CI: −0.19 to 2.58, p = 0.09 I2 = 50%) and changes in cognitive function scores between the radiofrequency group and cryoballoon group [standard mean difference (SMD): 0.39, 95% CI: −0.47 to 1.24, p = 0.38 I2 = 87%]. The NOS indicated that included studies were moderate to high quality, while the quality of evidence assessed by GRADE was low in 2 and very low in 2. Conclusion We analyzed the related cognitive outcomes after AF ablation. In the overall population, AF ablation had a positive trend for improving cognitive function at >3 months post-procedure. However, AF ablation might not be related to the improvement of cognitive function at < 3 months. Systematic review registration https://www.crd.york.ac.uk/PROSPERO/, identifier: CRD42021285198.


Introduction
Atrial fibrillation (AF) is the most common of all sustained arrhythmia with a worldwide prevalence of around 46.3 million individuals in 2016, the majority of whom are older adults (1,2). Dementia is another major cause of morbidity in older adults, and more than 50 million people are living with dementia worldwide (3,4). It is estimated that by 2050, as the population of the United States ages, the prevalence of AF and dementia will increase by between 2.5-and 3.0-fold (5).
There is increasing evidence pointing to dementia and cognitive disorder as additional adverse outcomes associated with AF. A recent meta-analysis showed that patients with AF had a 36% increased risk of developing dementia (6). An increased risk of stroke resulting from AF could partly mediate this association (7)(8)(9). Other adverse cerebrovascular effects associated with AF included cerebral hypoperfusion (10,11), microbleeds (12,13), and systemic inflammation (14), which might play a role but are not well characterized. The impact of cognitive dysfunction on healthcare and society will only increase along with the significant disease burden of AF.
Catheter ablation (CA) represents the first-line therapy for treating symptomatic and drug-refractory AF (15). In addition, a recent meta-analysis showed that CA as a first-line strategy in patients with paroxysmal AF had potential utility compared with anti-arrhythmic drugs (16). CA is superior to drug therapy in suppressing AF and improving symptoms, exercise capacity, and quality of life in patients with AF (17). The development and refinement of AF ablation have emerged as an effective therapy for AF and again raises the question of whether CA could attenuate a cognitive impairment. AF ablation on one side might reduce the risk of stroke, cerebral thromboembolism, and hypoperfusion with long-term sinus rhythm maintenance (18), and CA could also reduce the antiarrhythmic drug burdens used for rhythm control (19). However, on the other side, silent cerebral lesions (SCLs) during the AF ablation procedure might adversely increase the risk of post-procedural dementia (20). Consequently, there is a contradiction in the association of AF ablation with cognitive function. Therefore, we aimed to perform a systematic review and meta-analysis to evaluate the cognitive outcome post-CA procedure.

Research design and registration
Our systematic review and meta-analysis were reported according to the criteria outlined in the Meta-Analysis of Observational Studies in Epidemiology (MOOSE) and the PRISMA 2020 (21). This systematic evaluation program was registered in the PROSPERO International Prospective Registration for Systematic Evaluation (PROSPERO number: CRD42021285198).

Data sources and search strategy
Two investigators (Peng-fei Chen and Deng Pan) independently and systematically searched the PubMed, EMBASE, Web of Science, CNKI, WanFang, and VIP databases from inception to 28 September 2021. The search MESH term and keywords used included "atrial fibrillation, " "catheter ablation, " "radiofrequency ablation, " "cryoablation, " "dementia, " "dementia, vascular, " "Alzheimer's disease, " "cognitive dysfunction, " "cognition disorder, " and "mental status test." Detailed search strategies are shown in the Supplementary material. No restrictions on language, publication date, or publication status were set in our study. In addition, we examined the relevant reviews and reference lists of the included articles for further eligible studies. All the disagreements were resolved by consulting a third investigator (Ming Guo).

Study selection
Two investigators independently screened titles, abstracts, and full-text material to select studies that met the following

Data extraction
Prespecified data variables were extracted independently by two investigators. General characteristics included the author, year, country, study design, sample size of participants, followup duration in months, history of stroke, and maximum adjusted covariates. Baseline characteristics included demographic data (age and gender), combined diseases (hypertension, diabetes, and stroke/transient ischemic attack), combined drugs (anticoagulant and antiplatelet), and CHA2DS2-VASc score. Baseline characteristics of pooled study populations were reported as median values and their interquartile ranges (IQRs).

Quality evaluation
The methodological quality of the included studies was assessed according to the Newcastle-Ottawa Scale (NOS) (22) with scores ranging from 0 to 9. We evaluated quality concerning patient selection, comparability of studies, and assessment of outcomes or exposures. Studies with a total NOS score of ≥8 stars were defined as high quality, NOS score of 6-8 stars as moderate quality, and NOS score of < 6 stars as low quality.
The Grading of Recommendations Assessment, Development and Evaluation (GRADE) method (23) was adopted to evaluate the quality of evidence. The GRADE working group rated the certainty of outcome evidence as high, moderate, low, or very low certainty of evidence based on the study design, risk of bias, inconsistency, indirectness, imprecision, and other considerations.

Cognitive outcomes
The incidence of new-onset dementia including dementia Alzheimer's type, vascular dementia, senile dementia, frontotemporal dementia, dementia with Lewy bodies, and individual cognitive impairment reported in this study.

Statistical analysis
Hazard ratios (HRs) with 95% confidence intervals (CIs) for the incidence of dementia were extracted from published data. If adjustments were made for HRs, the most adequately adjusted HRs were extracted. For dichotomous variables, risk ratios (RRs) with 95% CIs were calculated. Continuous variables were calculated and expressed as weighted mean differences (WMDs) or standard mean differences (SMDs). Heterogeneity was assessed by using the Cochrane Q statistics, (p < 0.1 was considered with statistical heterogeneity), and I 2 Statistics (25, 50, and 75% were considered to represent low, medium, and high heterogeneity, respectively). We adopted a random-effect model for the meta-analysis because it incorporates the potential effects of heterogeneity and therefore allows for the retrieval of more generalizable results. Sensitivity analyses by removing one individual study at a time to confirm the robustness of the results. All statistical analyses were carried out using the Review Manager 5.4 software.

Study search
The process of the database search and study identification is presented in Figure 1. A total of 655 records were retrieved from 6 databases, 531 were duplicates, and 98 studies were excluded based on title and abstract primarily because they were irrelevant to the study purpose. The remaining 26 articles were evaluated for eligibility by full-text screening. Of these, 13 studies were further excluded because 3 studies were reviews, 7 studies did not report the cognitive outcomes, and the other 3 were studies of incomplete data. Finally, 13 studies (24-36) were included in our systematic review and meta-analysis. Table 1 displays the general characteristics, comprising 40,868 individuals (12,086 patients with AF ablation) included in the meta-analysis. The median follow-up time ranged from 3 months to 9 years. In this systematic review and meta-analysis, 4 studies (24-27) were retrospective cohort studies and 9 (28-36) were controlled clinical trials. Six studies (26, 30-32, 34, 35) were conducted in China, 2 (24,29) in South Korean, 2 (25, 27) in the United States, 1 (28) in Germany, 1 (36) in Japan, and 1 (33) in Australia. Four studies (24)(25)(26)(27) reported the incidence of newonset dementia, 4 (28-31) reported the changes in MoCA score, .

Study quality
The NOS showed that the quality scores of all the included studies ranged from 6 to 9 (mean score: 7.3), indicating moderate to high quality. Table 3 shows the study quality.
Among these outcome indicators, the quality of evidence was low in 2 and very low in 2. Certainty assessment ratings and the summary of findings are presented in Table 4.

Results of the meta-analysis New-onset dementia
Three studies (24-26) evaluated the HRs of developing dementia, including 18,849 patients. We adopted a randomeffects model to perform the meta-analysis, and the overall adjusted pooled HR of developing dementia was 0.60 (95% CI: 0.43 to 0.84, p = 0.003 I 2 = 40%; Figure 2), which showed that patients with AF ablation compared with patients with AF without ablation had a 40% lower risk of developing dementia during follow-up. The sensitivity analysis results were consistent (HR: 0.47 to 0.69, p all < 0.05). We also conducted a metaanalysis of 4 studies (24-27) by dichotomous variables (randomeffect RR: 0.43, 95% CI: 0.28 to 0.65, p < 0.0001 I 2 = 84%; Figure 3). The sensitivity analysis results were consistent (RR: 0.39 to 0.53, p all < 0.05).

MoCA score
The changes from the baseline of the MoCA score were reported in 4 studies (5 analyses) (27-30). A significant improvement of the MoCA score was identified between the AF with ablation group and AF without ablation group, which favored the AF with ablation group (random-effect WMD: 1.00, 95% CI: 0.36-1.64, p = 0.002 I 2 = 0%; Figure 4). The sensitivity analysis results were consistent (WMD: 0.19-1.13). However, no statistical difference was found after removing the study of

MMSE score
The changes from the baseline of MMSE score were reported in 3 studies (5 analyses) (27, 35,36). The MMSE score after .

TICS-m score
One study (31) that included 139 patients compared changes in TICS-m score in the AF ablation group and drug treatment group. The result indicated that TICS-m scores in the AF ablation group (39.56 ± 3.198) were significantly improved compared with the drug treatment group (34.44 ± 3.271) (p < 0.00001) at 12-month follow-up.

Incidence of POCD
Two studies (33,34) Table 2). b Appropriate population generalizability and outcomes applicability. c The score was downgraded because the proportion of patients in jin's study is too high (weigh>50%). d The heterogeneity was considered to represent low. e The score was downgraded because substantial heterogeneity between studies was detected (I 2 >75%). f Narrow 95% confidence interval. g The score was downgraded because fewer studies were included and there may have been greater publication bias. h The score was downgraded because all included studies in this meta-analysis were observational studies, we cannot rule out that some residual factors may reduce the demonstrated effect. CI confdence interval, HR hazard ratio, RR risk ratio, WMD weighted mean difference, SMD standardized mean difference, AF atrial fibrillation.

FIGURE
Forest plot present the meta-analysis for the association between the risk of dementia incidence and AF ablation.

FIGURE
Forest plot present the meta-analysis for the incidence rate of dementia comparing AF with ablation vs. AF without ablation groups.

FIGURE
Forest plot present the meta-analysis for the changes of MoCA score comparing AF with ablation vs. AF without ablation groups.
13.7% (26/190) of patients with AF had POCD 2 days after post-procedure, and the global cognitive scores decreased 2 days after postoperation tests and improved significantly at 6 months postoperation. These two studies suggested that the higher incidence of POCD 2-day post-procedure may in part reflect the reversible effect of anesthesia on cognitive function. At long-time follow-up, AF ablation might be associated with cognitive function improvement.

FIGURE
Forest plot present the meta-analysis for the changes of MMSE score comparing before AF ablation vs. after AF ablation groups.

FIGURE
Forest plot present the subgroup analysis for the changes of MoCA score between the follow-up time with < months and > months.

Radiofrequency and CY ablation
Two studies reported changes in cognitive function scores between the RF group and the CY group. We found no statistically significant differences (p>0.05) in cognitive function scores between the RF group and CY group (random-effect SMD: 0.39, 95% CI: −0.47 to 1.24, p = 0.38, I 2 = 87%; Figure 7).

Discussion
This meta-analysis and systematic review of 13 studies examined the relationship between AF ablation and cognitive function. We found that patients with AF ablation were associated with a lower risk of developing dementia in comparison with patients without AF ablation. Moreover, AF ablation had a positive trend for improving cognitive evaluation scales as a whole. MoCA, MMSE, and TICS-m are common indicators for evaluating cognitive function. A significant improvement in the MoCA, MMSE, and TICS-m scores was identified. In the subgroup analysis, no significant statistical differences were found in changes in MoCA score at < 3 months and changes in cognitive function scores between the RF and CY group. Of note, 2 studies reported the prevalence rate of POCD. They compared post-procedural cognitive function at 2 days and more than 3 months with cognitive function immediately before AF ablation. Cognitive impairment seemed to be only temporary .

FIGURE
Forest plot present the subgroup analysis for the changes of cognitive function score between the RF group and the CY group.
2 days after the post-procedure, and late cognitive function improvement may be related to time-dependent improvement.
Previous studies had reported that AF ablation was associated with cognitive decline and acute brain lesions. A small prospective study enrolling 23 patients showed worse neuropsychological outcomes in verbal memory in the AF ablation group (37). Previously, new SCLs detected after CA were a common occurrence in magnetic resonance imaging (MRI) of the brain and were reported in 4.3-38.9% of patients (32,38,39). However, during MRI follow-up of more than 90% of patients 1 year after AF ablation, SCLs had been proven to be resolved (40,41). The neurological impact of these SCLs is unclear and is likely to be determined by their size, number, and anatomic region. The sequelae of SCLs may include subtle neurocognitive impairment, which is in turn associated with an increased lifetime risk of cognitive impairment. Several aspects of the AF ablation may lead to acute cerebral injury and post-procedural cognitive dysfunction, for example, periprocedural thrombus, cerebral hypoperfusion, and anesthesia (42). Besides, catheters are a source of embolization, such as air embolization and carbonization, which may increase the risk of thromboembolic events, and lead to cognitive impairment (43).
In contrast to early observational studies, other emerging research had shown that AF ablation might improve cognitive function by reducing the AF burden and restoring sinus rhythm (44-46). Jin's study (29) suggested that AF ablation could reduce the possibility of left atrial thrombosis caused by atrial asynchronism and hemodynamic changes by relieving clinical symptoms and maintaining sinus rhythm in patients, thus improving long-term cognitive function. A recent randomized controlled trial (47) by Haeusler et al. has reported that chronic white matter damage and acute ischemic lesions detected by MRI were found frequently after first-time CA for paroxysmal AF using uninterrupted oral anticoagulation. The median of MoCA was similar in patients with or without acute brain damage at 3 months after CA, and acute ischemic lesions detected by high-resolution diffusion-weighted imaging were not associated with cognitive function at 3 months after ablation. We inferred that the restoration and maintenance of sinus rhythm were an important mechanism, rather than the AF ablation itself, which has been associated with at least transient worsening of cognitive function. The recovery of sinus rhythm after AF ablation improved atrioventricular synchronization and systolic and diastolic function, which may enhance cerebral perfusion, promoting the recovery and improvement of cognitive function.
Two previous meta-analyses also examined the relationship between AF ablation and dementia. A meta-analysis of 4 studies by Saglietto et al. (48) showed that AF ablation is associated with a nearly 50% reduction in dementia occurrence. Another meta-analysis by Bodagh et al. (49) found that AF ablation was associated with a lower risk of subsequent dementia diagnosis. However, the above two meta-analyses did not include sufficient studies and did not analyze other cognitive outcome indicators except dementia. Our study included all the current studies on the relationship between AF ablation and cognitive outcomes. We analyzed more comprehensive cognitive outcome indicators and conducted a subgroup analysis on follow-up time and ablation type, which provided more evidence information.
The advantages of our meta-analysis may include the following. First, the results of this study were relatively stable and reliable because the meta-analysis covered studies from different countries and had a large sample size. Second, only cohort studies and controlled clinical trials were included, and the results showed a sequential association between AF ablation and improvement in cognitive function. Third, the most adequately adjusted HRs were extracted, which reduced clinical heterogeneity to a certain extent. Fourth, the sensitivity analyses that omitted a study at a time had no significant impact on the results, suggesting that the outcomes were credible. Fifth, subgroup analyses of followup time and ablation type were conducted to assess the potential study characteristics of the relationship between AF ablation and cognitive function. Finally, the NOS was used to assess the methodological quality of the studies, and the GRADE method was adopted to evaluate the quality of evidence.
However, this meta-analysis also had some limitations. First, as a meta-analysis of observational studies, we were unable to . /fneur. . determine whether the association between AF ablation and dementia was causal. Second, we cannot exclude that some residual factors may confound the association between AF ablation and cognitive function improvement, although we included studies with multivariate-adjusted HRs only. Third, in the real world, various drugs are commonly used to treat AF.
Most of the included studies did not mention specific treatment regimens, which to some extent leads to an unavoidable clinical heterogeneity.

Conclusion
We analyzed the related cognitive outcome post-CA procedure. In the overall population, AF ablation had a positive trend for declining the risk of developing dementia and improving cognitive function at >3 months post-procedure. However, AF ablation might not be related to the improvement of cognitive function at < 3 months and changes in cognitive function scores between the RF group and the CY group.

Data availability statement
The original contributions presented in the study are included in the article/Supplementary material, further inquiries can be directed to the corresponding author.