Impact of chronic kidney disease on left atrial appendage occlusion: A meta-analysis of procedural outcomes and complications

Background: Patients with chronic kidney disease (CKD) experience atrial fibrillation more frequently. The balance of medical management for stroke prevention and bleeding events presents a challenging issue in CKD population. Left atrial appendage occlusion (LAAO) may be an effective solution for stroke prevention in patients who experience frequent bleeding with oral anticoagulants. However, the specific impact of CKD on the procedural success, complications, and outcomes of LAAO implantations remains underexplored. Methods: We conducted a search of various databases for articles published before October 31, 2023. This search yielded 7 studies, comparing outcomes between CKD and non-CKD cohorts undergoing LAAO implantation. Our analysis focused on CHA2DS2-VASc scores, average eGFR, use of oral anticoagulants, procedural success rates, procedural complications, and associated outcomes. Results: The meta-analysis included data from 2576 patients, with 1131 identified as having CKD. The CKD group also had higher CHA2DS2-VASc scores (4.7 ± 1.4 vs 4.0 ± 1.5; P < .001) and HAS-BLED scores (3.8 ± 1.1 vs 3.1 ± 1.0; P < .001) than the non-CKD group. CKD patients showed a nonreduction in procedural success rates and a nonsignificant increase in total complications. The risks of stroke and transient ischemic attack, major bleeding, and cardiovascular mortality were not significantly different between the 2 groups. However, a significantly lower rate of total mortality was observed in the non-CKD group (odds ratio: 0.43; 95% confidence interval, 0.32–0.60). Conclusion: While CKD is associated with a nonsignificant decrease in procedural success and a nonsignificant increase in complication risks, the outcomes of LAAO implantation are comparably favorable between CKD and non-CKD groups. Despite similar procedural outcomes, the CKD group exhibited a higher rate of all-cause mortality.


Introduction
[3] Chronic kidney disease (CKD), marked by kidney damage and a reduced filtration rate (estimated Glomerular filtration rate below 60 mL/min/1.73m 2 ), [4] heightens the risk of strokes in AF patients, especially as eGFR decreases. [5,6]Stroke prevention in non-dialysis CKD patients mirrors that in the general population.[13] For AF patients, most clots form in the left atrial appendage (LAA) due to its shape causing blood flow stasis. [14][17] LAAO patients may need less or no OACs, potentially requiring only single antiplatelet therapy.[20][21] In CKD patients, comorbidities significantly affect the success of coronary interventions and AF ablation, [22,23] but the impact on LAAO implantations remains underexplored.
Recent research has shown that LAAO is equally safe for people with CKD and those without, effectively reducing strokes, transient ischemic attacks (TIA), and bleeding incidents. [24,25]owever, some studies have observed a rise in adverse events among CKD patients. [26][29] Notably, one study pointed out a greater likelihood of thromboembolic and bleeding issues in those with CKD. [30]Considering the inconsistencies in previous studies, our meta-analysis aims to rigorously evaluate the effectiveness and safety of LAAO in both CKD and non-CKD populations.

Search strategies, trial selection, quality assessment, and data extraction
Two cardiologists, Wei-Chieh Lee and Hsiu-Yu Fang, conducted a comprehensive literature search in various databases, including PubMed, Embase, ProQuest, ScienceDirect, Cochrane Library, ClinicalKey, Web of Science, and ClinicalTrials.gov.They searched for articles published up to October 31, 2023, using key terms such as "atrial fibrillation or AF," "left atrial appendage occlusion or left atrial appendage occlude or LAAO," "chronic kidney disease or CKD," and "oral anticoagulants."The search was conducted without language restrictions to broaden the scope of relevant studies.In instances of disagreement, a third reviewer, Po-Jui Wu, was consulted for resolution.This analysis included randomized controlled studies and observational studies that compared the efficacy and safety outcomes of LAAO in groups with and without CKD.The inclusion criteria specified human studies with a parallel design, focusing on the comparison of efficacy and safety of LAAO between CKD and non-CKD groups.Exclusion criteria encompassed case reports or series, animal studies, review articles, conference abstracts, and unpublished data.No language limitations were set, aiming to increase the pool of eligible articles.Supplemental Figure 1, Supplemental Digital Content, http://links.lww.com/MD/N203, a PRISMA flow chart, depicts the literature search and screening protocol.

Outcomes
The outcomes of interest in this study included both shortterm and long-term results.Short-term outcomes encompassed procedural success, periprocedural stroke, cardiac tamponade, and overall complications.Long-term outcomes focused on stroke/TIA major bleeding events, CV mortality, and all-cause mortality.

Statistical analysis
We extracted the frequency of each assessed outcome from the included studies and presented the data as cumulative rates.A random effects model was utilized to aggregate individual odds ratios (ORs), with all analyses conducted using the Comprehensive Meta-Analysis software version 3 (Biostat, Englewood, NJ).To evaluate heterogeneity among the trials, the chi-square test was applied, considering P ≤ .10 as indicative of significant heterogeneity.We also used the I 2 statistics, categorizing heterogeneity as low to high ranging from 25% to 75%.To determine potential publication bias, we employed Egger test, visualizing the results through funnel plots, with statistical significance defined at P ≤ .10.For comparisons between CKD and non-CKD groups, statistical significance was established at P < .05.

Characteristics of included studies
The process of selecting studies for inclusion is illustrated in Supplemental Figure 1, Supplemental Digital Content, http:// links.lww.com/MD/N203.A total of 7 studies [24][25][26][27][28][29][30] met the inclusion criteria, encompassing 2576 participants.Details regarding the study period, participant characteristics, criteria for CKD, and the follow-up period are outlined in Table 1. ) compared to the non-CKD group (83.9 ± 18.9 mL/ min/1.73m 2 , P < .001).These demographic differences are crucial for understanding the diverse impacts of LAAO in patients with and without CKD.

Pooled OR of long-term outcomes
Figure 2 evaluates the long-term outcomes between non-CKD and CKD groups.The pooled OR for stroke and TIA is 1.48, with a 95% CI of 0.61 to 3.58, indicating no statistically significant difference between the groups, and demonstrating low heterogeneity (Cochran Q = 4.558; df = 4; I 2 = 12.24%; P = .336).

Discussion
This study compared outcomes between non-CKD and CKD populations in AF patients undergoing LAAO, aiming to explore the impact of CKD on the efficacy and safety of LAAO in these groups.The CKD population exhibited a higher prevalence of comorbidities, as well as increased CHA2DS2-VASc and HAS-BLED scores.The average age in the study population was 76 ± 7.2 years, with 30.2% having a history of stroke, 42.8% experiencing major bleeding previously, but only 35.0% of the patients had been on OACs.Between the non-CKD and CKD groups, CKD was associated with a nonsignificant reduction in procedural success and a nonsignificant increase in complication risks.Post-LAAO implantation, the incidence of stroke/TIA and major bleeding was similar between the 2 groups, despite a higher potential for stroke in the CKD group according to significantly higher CHA2DS2-VASc scores (non-CKD vs CKD; 4.0 ± 1.5 vs 4.7 ± 1.4; P < .001).However, a higher incidence of all-cause mortality was observed in the CKD population, which may be attributed to their complex baseline characteristics.
Based on 7 studies [24][25][26][27][28][29][30] that we reviewed, there were no significant differences in procedural success and complications between the non-CKD and CKD groups.Previous research [24][25][26]28] indicates that LAAO significantly reduces the rates of stroke/ TIA and major bleeding across stages of CKD, compared to the expected annual risk. Notly, the cessation of anticoagulant therapy following LAAO was particularly beneficial for patients at a very high risk of bleeding, where a significant reduction in major bleeding events was observed.In our study, the CKD population exhibited a higher predicted bleeding risk according to the HAS-BLED (non-CKD vs CKD; 3.1 ± 1.0 vs 3.8 1.1; P < .001),yet they showed similar rates of longterm major bleeding.Consequently, LAAO plays a crucial role in significantly reducing bleeding events in the CKD population, especially for those with a higher bleeding risk and who are intolerant to OACs.Previous meta-analyses have not standardized the criteria and of CKD, often including patients with end-stage renal disease (ESRD).[31] In contrast, our study defined CKD as an eGFR of <60 mL/min/1.73m² and excluded patients with ESRD.Patients undergoing dialysis face significant constraints regarding OAC use.Currently, there is no uniform guideline for anticoagulation following LAAO, and the practice of using post-procedural dual antiplatelet agents (excluding OACs) after LAAO is common.[32] Three of the studies we included provided detailed regimens of post-procedural OAC or antiplatelet agents.[24,27,28] Other important issue is device-related thrombosis after LAAO.Device-related thrombosis following LAAO is linked to ischemic events, and renal insufficiency increases the risk of such thrombosis by approximately fourfold.[33,34] Therefore, there is a significant need for larger studies to assess the long-term risk of ischemic events following LAAO in CKD patients, particularly those not protected by OACs or on lowdose OACs.

Limitations
This study is subject to several limitations.First, most of the included studies are observational cohort studies, which introduces a potential for selection bias.However, these studies collectively encompass 2576 participants, including the most recent research.Second, we were unable to access complete baseline characteristics for all participants in the studies we reviewed.Third, this meta-analysis did not analyze the post-procedural regimen of OAC or antiplatelet therapy, nor the incidence of device-related thrombosis after LAAO.Fourth, the studies included in our review had varying follow-up periods.Despite these limitations, this study provides valuable insights into the efficacy and safety of LAAO in the CKD population.

Conclusion
While CKD is associated with a slight, nonsignificant increase in complication risks and a marginal decrease in procedural success, the outcomes of LAAO are similarly favorable in both CKD and non-CKD groups.Despite higher rates of all-cause mortality and more comorbidities in the CKD group, LAAO shows comparable effectiveness in reducing stroke/TIA and major bleeding incidents when compared to the non-CKD population.

Figure 1 .
Figure 1.(A) Forest plots demonstrating the rate of procedural success between the non-CKD and CKD groups in 7 studies.(B) Forest plots demonstrating the risk of periprocedural stroke between the non-CKD and CKD groups in 4 studies.(C) Forest plots demonstrating the risk of cardiac tamponade between the non-CKD and CKD groups in 5 studies.(D) Forest plots demonstrating the risk of complications between the non-CKD and CKD groups in 5 studies.Abbreviation: CKD: chronic kidney disease.CKD = chronic kidney disease.

Figure 2 .
Figure 2. (A) Forest plots demonstrating the risk of stroke and TIA between the non-CKD and CKD groups in 5 studies.(B) Forest plots demonstrating the risk of major bleeding events between the non-CKD and CKD groups in 5 studies.(C) Forest plots demonstrating the risk of cardiovascular mortality between the non-CKD and CKD groups in 2 studies.(D) Forest plots demonstrating the risk of all-cause mortality between the non-CKD and CKD groups in 6 studies.CKD = chronic kidney disease, TIA = transient ischemic attack, CV = cardiovascular.

Table 1
Characteristics of the 7 included studies.Table 2 describes a comprehensive comparison of patient demographics between the non-CKD and CKD groups, encompassing 2576 participants divided into 1445 non-CKD and 1131 CKD patients.The data highlight key demographic differences, with the average age significantly higher in the CKD group (78 ± 6.3 years) compared to the non-CKD group (74 ± 7.4 years, P < .001).The percentage of males was also lower in the < .001).Coronary artery disease was also more prevalent in the CKD group (51.6% vs 33.3% in non-CKD, P < .001),while the incidence of prior stroke was lower (26.0% in CKD vs 33.4% in non-CKD, P < .001).The CHA2DS2-VASc score was higher in the CKD group (4.7 ± 1.4) compared to the non-CKD group (4.0 ± 1.5, P < .001),as was the HAS-BLED score (3.8 ± 1.1 in CKD vs 3.1 ± 1.0 in non-CKD, P < .001).The eGFR was significantly lower in the CKD group (42.2 ± 13.6 mL/ min/1.73m 2 CKD = chronic kidney disease, eGFR = estimated Glomerular filtration rate.

Table 2
Patients' demographics between non-CKD and CKD populations.
Data are expressed as mean ± standard deviation or as number (percentage).AF = atrial fibrillation, CKD = chronic kidney disease, eGFR = estimated Glomerular filtration rate.