Relationship Between Endothelial Dysfunction and the Outcomes After Atrial Fibrillation Ablation

Authors: Keisuke Okawa, Masahiro Sogo, Takeshi Morimoto, Ryu Tsushima, Yuya Sudo, Eisuke Saito, Masatomo Ozaki, Masahiko Takahashi

Abstract

Background

Endothelial dysfunction (ED) is associated with cardiovascular events in patients with atrial fibrillation (AF). However, the utility of ED as a prognostic marker after AF ablation supplementary to the CHA2DS2‐VASc score is unclear. This study aimed to investigate the relationship between ED and 5‐year cardiovascular events in patients undergoing AF ablation.

Methods and Results

We conducted a prospective cohort study of patients who underwent a first‐time AF ablation and for whom the endothelial function was assessed by the peripheral vascular reactive hyperemia index (RHI) before ablation. We defined ED as an RHI of <2.1. Cardiovascular events included strokes, heart failure requiring hospitalization, arteriosclerotic disease requiring treatment, venous thromboses, and ventricular arrhythmias or sudden cardiac death. We compared the 5‐year incidence of cardiovascular events after AF ablation between those with and without ED. Among the 1040 patients who were enrolled, 829 (79.7%) had ED, and the RHI value was found to be associated with the CHA2DS2‐VASc score (P=0.004). The 5‐year incidence of cardiovascular events was higher among patients with ED than those without ED (98 [11.8%] versus 13 [6.2%]; log‐rank P=0.014). We found ED to be an independent predictor of cardiovascular events after AF ablation (hazard ratio [HR], 1.91 [95% CI, 1.04–3.50]; P=0.036) along with a CHA2DS2‐VASc score of ≥2 (≥3 for women) (HR, 3.68 [95% CI, 1.89–7.15]; P<0.001).

Conclusions

The prevalence of ED among patients with AF was high. Assessing the endothelial function could enable the risk stratification of cardiovascular events after AF ablation.

Clinical PerspectiveWhat Is New? The baseline vascular endothelial function assessed by peripheral reactive hyperemia arterial tonometry is associated with the 5‐year outcomes in patients with atrial fibrillation who underwent catheter ablation.The assessment of the endothelial function has the potential to enhance the standard risk stratification method, which is based on the CHA2DS2‐VASc score, especially for the stratification of the lowest‐ and highest‐risk patients. What Are the Clinical Implications? The assessment of the endothelial function before atrial fibrillation ablation is simple and highly reproducible and so would be beneficial to the daily clinical practice to provide useful information and guide an appropriate postablation management.A combination of the endothelial function and CHA2DS2‐VASc score might lead to developing a more reliable scoring system for patients with atrial fibrillation undergoing catheter ablation.

Atrial fibrillation (AF) is the most commonly occurring arrhythmia and requires medical treatment because of the increased rates of morbidity and mortality. ^1^ Catheter ablation has been reported to reduce the risk of cardiovascular events, such as strokes, heart failure, and death, in patients with AF. ^2^ , ^3^ However, the methods for the risk stratification of the outcomes after AF ablation are limited, despite the high frequency of underlying disease and comorbidities as well as AF‐related complications. The CHA2DS2‐VASc score is a standard tool for the stratification of the risk of ischemic strokes in patients with AF and has been reported to be a powerful predictor of cardiovascular outcomes after AF ablation, including recurrence and other cardiovascular events. ^4^ , ^5^

Endothelial dysfunction (ED), assessed by the peripheral vascular reactive hyperemia index (RHI), is an integrated index of arteriosclerosis and an established prognostic marker of cardiovascular events. ^6^ Recent articles have discussed the close relationship between the endothelial function and AF. ^7^ , ^8^ , ^9^ , ^10^ In particular, the endothelial function assessed by flow‐mediated dilation has been found to be associated with new‐onset AF as well as with cardiovascular events in patients with AF. ^11^ , ^12^

We hypothesized that the endothelial function assessed by the RHI before AF ablation could predict the occurrence of long‐term cardiovascular events after ablation. To test this, we investigated the 5‐year outcomes in patients who underwent AF ablation and compared the incidence of cardiovascular events between the patients with and without ED. We evaluated whether assessing the endothelial function could reinforce the established risk stratification method using the CHA2DS2‐VASc score.

Methods

Study Design and Patients

The data that support the findings of this study are available from the corresponding author upon reasonable request. We recruited all patients with nonvalvular AF who planned to undergo a first‐time AF ablation between May 2013 and June 2022 at Kagawa Prefectural Central Hospital, Kagawa, Japan, for this prospective cohort study. The inclusion criteria were as (1) a diagnosis of nonvalvular AF, (2) a first‐time AF ablation, and (3) available RHI data before the AF ablation. Valvular AF was defined as moderate or severe mitral stenosis or having undergone a mechanical valve replacement. Exclusion criteria were as (1) declined the RHI measurement, (2) end‐stage renal failure requiring hemodialysis, and (3) a follow‐up of <6 months after the ablation. We considered the RHI was not representative of the status of the endothelial function in patients with hemodialysis because of the presence of an arteriovenous shunt; thus, such patients were excluded from the cohort. The sample size was based on the registered patients during the study period, which was determined in advance considering the feasibility of the study.

This study conformed to the principles of the Declaration of Helsinki and was conducted after approval from the Clinical Ethics Committee of the Kagawa Prefectural Central Hospital. Written informed consent was substituted by the opt‐out method by announcing the handling of the personal data and right to withdraw consent on the website of the study institution.

Measurement of the Endothelial Function

As a standard protocol for the cardiovascular risk assessment before the AF ablation, we evaluated the endothelial function by assessing the RHI with reactive hyperemia peripheral arterial tonometry using an EndoPAT2000 (Itamar Medical, Caesarea, Israel), which measures the digital hyperemic response. We measured the RHI after stopping all medications, including the vasoactive medications, on the day of the measurement to avoid the effect of the medications.

The RHI is a well‐established tool for assessing the endothelial function and is approved by the US Food and Drug Administration. The RHI is an independent predictor of adverse cardiovascular events, beyond the traditional Framingham risk score. ^13^ The principle underlying RHI has been previously described,^14^ and its accuracy during AF has been validated in our previous study. ^15^ The threshold for the ED was taken as <2.1, based on the physiological diagnostic criteria. ^16^ The RHI values were concealed from the attending physicians who observed the patients.

AF Ablation

Anticoagulation therapy was initiated at least 1 month before the ablation. Transesophageal echocardiography was performed on the day of admission to confirm the absence of any left atrium thrombi in patients with persistent AF or a CHA2DS2‐VASc score of ≥2.

We performed AF ablation based on the standard method. ^17^ Briefly, electrical pulmonary vein isolation by radiofrequency ablation was performed using an open‐irrigation 3.5‐mm tip‐deflectable catheter (Thermocool; Biosense Webster, CA) with contact‐force sensing and a 3‐dimensional mapping system (CARTO; Biosense Webster). Heparin was administered to maintain the activated clotting time at ≥300 seconds during the procedure. A posterior‐wall isolation, superior vena cava isolation, and cavotricuspid isthmus ablation were performed as needed. General anesthesia was induced and maintained using a continuous propofol intravenous infusion with a subglottic device to keep the airway open. An artificial respirator was used to maintain effective ventilation.

Anticoagulation therapy was discontinued 6 months after the ablation in patients with a CHA2DS2‐VASc score of <2 or continued all their lives in patients with a CHA2DS2‐VASc score of ≥2 unless anticoagulation‐related adverse events occurred.

Data Collection and Outcomes

We collected data on the patient characteristics, including the systolic and diastolic blood pressures, heart rate, medical history, comorbidity, laboratory data, medication use, and electrocardiography and transthoracic echocardiography findings. The purpose of the transthoracic echocardiography was to measure the left atrial diameter and left ventricular ejection fraction. The latter was measured using the disk‐summation method and expressed as the average over 5 cardiac cycles if AF persisted. We investigated the outcomes from the medical records of the Kagawa Prefectural Central Hospital. We further investigated the outcomes in March 2021 by a mail‐in questionnaire or telephone call to the patients, their families, and their primary care physicians for patients who completed follow‐up clinical visits at the Kagawa Prefectural Central Hospital.

The primary outcome was the first cardiovascular event during the follow‐up period. We defined a “cardiovascular event” as a stroke, heart failure requiring hospitalization, arteriosclerotic disease requiring treatment, venous thrombosis, or ventricular arrhythmia or sudden cardiac death. We defined cardiovascular death as death attributable to these diseases. Strokes included cerebrovascular infarctions, transient ischemic attacks, and intracranial hemorrhages. Heart failure requiring hospitalization was defined as worsening heart failure requiring treatment in the hospital. Arteriosclerotic diseases included coronary artery disease (acute coronary syndrome and stable angina), aortic disease (aortic dissection and aneurysm), and peripheral artery disease. Venous thromboses included acute pulmonary thromboembolisms and deep vein thromboses. Ventricular arrhythmias included ventricular fibrillation, sustained ventricular tachycardia, and frequent premature ventricular contractions accounting for ≥20% of the total heartbeat. The outcomes were assessed by cardiologists who were uncertain of the RHI values.

Statistical Analysis

We performed descriptive statistics of the total cohort and compared the baseline characteristics and outcomes between the patients with (denoted as the “ED group”) and without (“non‐ED group”) ED. We also compared the outcomes between the patients with a CHA2DS2‐VASc score of <2 and those with a score of ≥2 and between the 4 groups created by dividing the patients according to a CHA2DS2‐VASc score <2/≥2 and with/without ED. However, women were divided according to a CHA2DS2‐VASc score <3/≥3 based on the latest guidelines. ^18^

Continuous variables are expressed as the mean and SD or median and interquartile range of their distribution. The t‐test or Wilcoxon rank‐sum test was used for between‐group comparisons, based on data distribution. Categorical variables are presented as numbers and percentages, and intergroup comparisons were conducted using the χ^2^ test. The outcomes are presented as the number of first events. The Kaplan‐Meier survival curve and log‐rank test were used to compare the incidences of the outcomes in the ED and non‐ED groups considering the censoring attributable to an end of the follow‐up without outcomes. The effect was expressed as the hazard ratio (HR) and its 95% CI, with the use of a Cox proportional hazard model. A multivariable Cox proportional hazard model was used to determine the factors associated with cardiovascular events. Clinically relevant candidate variables for cardiovascular events included a smoking habit, a drinking habit, ^19^ dyslipidemia, body mass index ≥25 kg/m^2^, ^20^ chronic kidney disease, ^21^ nonparoxysmal AF, ^22^ ED, CHA2DS2‐VASc score of ≥2 ^23^ (≥3 for women), ^18^ and AF recurrence. Subgroup analyses were also performed to estimate the HR of ED for age, sex, hypertension, diabetes, heart failure, chronic kidney disease, AF type, and history of systemic thromboembolisms. We estimated each HR after adjusting for the variables similar to the above variables, except for ED. We performed a sensitivity analysis after removing the events in the first year to ensure that the statistical association between the RHI and events was not confounded by common preceding factors. In addition, we plotted the adjusted HR by the follow‐up year.

The proportion of ED was graphed using each CHA2DS2‐VASc score as the bar variable. Scores of ≥5 were combined into 1 category because of the small number of patients with such scores, and the RHI values for each CHA2DS2‐VASc score were analyzed using an ANOVA.

In addition, we assessed the discriminatory power of the RHI for cardiovascular events when adding the RHI to the CHA2DS2‐VASc score by calculating the net reclassification improvement and integrated discrimination improvement. We also constructed a Brier score, receiver operating characteristic curve, and decision curve analyses. ^24^ We estimated the forecasts of the CHA2DS2‐VASc score for the 5‐year cardiovascular events in the Brier score analysis while referring to the stroke incidence of each score ^25^ , ^26^ (score 0=0%, 1=5%, 2=10%, 3=15%, 4=20%, and ≥5=25%). We regarded non‐ED as a score of –1 (score 0 remained 0) in the Brier score and receiver operating characteristic analyses.

We performed a cubic‐spline analysis to identify the cutoff value of the RHI for cardiovascular events after AF ablation and compared the outcomes between the 2 patient groups divided by the determined cutoff value. Finally, we estimated the 5‐year cardiovascular events and deaths in the excluded patients in whom the RHI could not be evaluated because of the possibility of selection bias.

Missing data were not imputed and were eliminated from the corresponding analyses. All reported P values were 2 sided. The statistical analyses were performed using JMP version 15 (SAS Institute Inc, Cary, NC), R version 4.2.2 (R Foundation for Statistical Computing, Vienna, Austria), and SAS 9.4 (SAS Institute Inc) software.

Results

Baseline Characteristics

We enrolled 1040 patients for this study (Figure 1). The mean age was 67±10 years, 33% were women, 58% had hypertension, 37% had heart failure, 27% had chronic kidney disease, 54% had nonparoxysmal AF, the mean left atrial diameter was 38±6 mm, and the mean left ventricular ejection fraction was 61±9%. There were 829 patients (79.7%) in the ED group.

The median follow‐up after AF ablation was 35 months (range, 7–60 months) in the ED group and 37 months (range, 9–60 months) in the non‐ED group (P=0.26). The prevalence of hypertension, heart failure, diabetes, dyslipidemia, and chronic kidney disease, body mass index, hs‐CRP (high‐sensitivity C‐reactive protein), and angiotensin‐converting enzyme inhibitor or angiotensin II receptor blocker or calcium channel blocker intake were significantly higher in the ED group than the non‐ED group (Table 1).

Outcomes

The 5‐year incidence of cardiovascular events was higher in the ED group than the non‐ED 98 (11.8%) versus 13 (6.2%) (Figure 2A and Table 2). The incidences of each specific cardiovascular event (strokes, heart failure requiring hospitalization, arteriosclerotic disease requiring treatment, venous thromboses, ventricular arrhythmias or sudden cardiac death, and cardiovascular death) were consistently higher in the ED group (Table 2).

Subgroup analyses of the cardiovascular events based on the endothelial function supported these findings, and no interactions were observed for each variable (Figure 3). The multivariable Cox proportional hazard model determined ED to be an independent predictor of cardiovascular events in addition to a CHA2DS2‐VASc score of ≥2 (≥3 for women) and chronic kidney disease (Table 3).

A sensitivity analysis by removing the cardiovascular events that occurred within the first year exhibited consistent results to the main results (log‐rank P=0.018; adjusted HR, 2.32 [95% CI, 0.99–5.43]; Figure S1). The adjusted HR of the cardiovascular events within each year was also consistent (1‐year HR, 1.70 [95% CI, 0.72–4.00]; 2‐year HR, 1.84 [95% CI, 0.88–3.86]; 3‐year HR, 2.17 [95% CI, 1.08–4.34]; 4‐year HR, 1.79 [95% CI, 0.98–3.30]; 5‐year HR, 1.91 [95% CI, 1.04–3.50]), and data are plotted in Figure S2.

The distribution of the CHA2DS2‐VASc scores and ED proportions are presented in Figure 4A. The ED proportion and CHA2DS2‐VASc scores were found to be associated (P=0.044). The RHI values were also associated with the CHA2DS2‐VASc score (P=0.004) (Figure 4B).

The incidence of cardiovascular events among the patients without ED was similar to that of the patients with ED, but the CHA2DS2‐VASc score was 1 value lower (Figure 2B) (CHA2DS2‐VASc score with ED versus without ED: score 0, 0% versus 0%; score 1, 5.5% versus 2.6%; score 2, 5.8% versus 5.6%; score 3, 14.6% versus 6.4%; score 4, 14.4% versus 12.1%; and score ≥5, 27.1% versus 13.3%, respectively).

The 5‐year incidence of cardiovascular events was higher among the patients with a CHA2DS2‐VASc score of ≥2 (≥3 for women) compared with those with a score of <2 (<3 for women) (Figure 5A). After dividing the cohort into 4 groups based on the endothelial function, the 5‐year incidence of cardiovascular events differed among all groups, with the highest and lowest incidence observed for the CHA2DS2‐VASc score ≥2 (≥3 for women) with ED and CHA2DS2‐VASc score <2 (<3 for women) without ED groups, respectively (Figure 5B). The prognostic value of ED defined by the RHI for cardiovascular events improved after adding it to the CHA2DS2‐VASc score (net reclassification improvement of 0.19 [95% CI, 0.06–0.32] [P=0.004]; integrated discrimination improvement of 0.004 [95% CI, 0.0002–0.008] [P=0.039]). The detailed results of the net reclassification improvement are shown in Table S1. The median Brier score adding the ED to the CHA2DS2‐VASc score was lower than that of only the CHA2DS2‐VASc 0.001 (interquartile range, 0.003–0.040) versus 0.023 (interquartile range, 0.003–0.040) (P=0.04). The concordance index adding the ED to the CHA2DS2‐VASc score was greater than that of only the CHA2DS2‐VASc 0.719 versus 0.713 (P=0.28) (Figure S3). The clinical decision curve improved after adding the RHI value to the CHA2DS~2−~VASc score (Figure S4).

The cutoff value of the RHI among this cohort was determined as 1.8 by a cubic‐spline analysis. After dividing the patients by the cutoff value, the 5‐year incidence of cardiovascular events was higher in the low RHI (≤1.8) group than high RHI (>1.8) group (log‐rank P=0.003; adjusted HR, 1.63 [95% CI, 1.04–2.55]; P=0.031) (Figure S5).

Among the 511 patients in whom the RHI could not be evaluated, after excluding 4 patients who were lost to follow‐up, 48 (9.5%) had cardiovascular events and 7 (1.4%) had cardiovascular deaths. The incidence was almost similar to that in the overall patients included in this study.

Discussion

We demonstrated that the baseline vascular endothelial function assessed by peripheral reactive hyperemia arterial tonometry was associated with the 5‐year outcomes in patients with AF who underwent catheter ablation. In addition, this assessment of the endothelial function had a potential of enhancing the standard risk stratification method, which is based on the CHA2DS2‐VASc score.

There have been limited studies on the prognostic markers for post‐AF ablation despite the high demand for this procedure in daily clinical practice. The CHA2DS2‐VASc has been shown to be a useful predictor of cardiovascular outcomes after AF ablation in 2 previous studies. One study involving 565 patients who underwent AF ablation and were followed up for 39 months found the CHADS2 and CHA2DS2‐VASc scores to be predictors of thromboembolic events, including ischemic strokes, transient ischemic attacks, peripheral embolisms, and pulmonary embolisms. ^4^ The CHA2DS2‐VASc score can be used to further stratify patients with CHADS2 scores of <2 into 2 groups with different event rates, using a cutoff value of 2. The other study involving 2179 patients who underwent a first‐time AF ablation found that both the CHADS2 and CHA2DS2‐VASc scores were excellent for stratifying patients for the 5‐year major adverse cardiovascular events with a composite end point of death, strokes, and heart failure hospitalizations. ^5^ Furthermore, the CHA2DS2‐VASc score was reportedly superior to the CHADS2 score for predicting AF recurrence and the occurrence of AF‐related morbidities. Our study confirmed the utility of the CHA2DS2‐VASc score for predicting cardiovascular events after AF ablation, and we provided new insight into the use of ED as a prognostic marker independent of the CHA2DS2‐VASc score.

Developing an index of the endothelial function could enable the identification of early‐stage arteriosclerosis. ^27^ The relationship between the endothelial function and AF has been discussed ^7^ , ^8^ , ^9^ , ^10^ ; ED could be a predictor of cardiovascular outcomes in patients with AF ^12^ because AF genesis is driven by comorbidities, which are also associated with a worsening endothelial function. Indeed, previous studies have identified an association between ED and the recurrence of AF after ablation. ^28^ , ^29^ To the best of our knowledge, this is the first study to identify that the baseline peripheral vascular ED assessed by the RHI was an independent predictor of long‐term cardiovascular events after AF ablation. In addition, we showed that this assessment of the endothelial function might enhance the risk stratification using the CHA2DS2‐VASc score, particularly for stratifying low‐ and high‐risk patients. This assessment is simple and highly reproducible and so would be beneficial to the daily clinical practice to provide useful information and guide an appropriate postablation management.

This study had several limitations that should be considered. First, the cohort included patients with nonvalvular AF who were indicated for AF ablation and who underwent RHI measurements; thus, selection bias was inevitable. Because the study was conducted in a high‐volume referral center where ≥400 AF ablation procedures are performed per year, patients with indications for AF were representative of the patients with nonvalvular AF who received AF ablation. Although almost 30% of eligible patients were excluded because of a lack of an RHI, the risk of events among the patients was similar to that in the included patients. Therefore, the bias could be small, if any. Second, the differences in the medication prescription rates might have affected the results of this study despite the medications having been stopped on the day of the RHI evaluation. Third, the value of the RHI for defining ED in patients with AF might be lower than that in the general population. Further study would be required to clarify a more precise cutoff value of the RHI in patients with AF. Fourth, the changes in the endothelial function and comorbidities, or changes in the prescribed medications during the follow‐up, may have affected our results. However, the outpatient care, including the changes in the medications, followed the system‐wide guidelines regardless of the endothelial function. Several studies have reported that the endothelial function can be improved by therapeutic intervention, especially in patients with persistent AF. ^15^ , ^28^ , ^30^ Therefore, further study is required to evaluate whether an improved endothelial function may translate into a reduced risk of cardiovascular events. Fifth, the degree of the incremental prognostic value of adding the RHI measurement to the CHA2DS2‐VASc score was small in our cohort. That might be because of the relatively small number of events despite the long‐term follow‐up. Finally, we enrolled a large number of patients with AF undergoing catheter ablation and investigated the detailed clinical courses and outcomes, and we are satisfied that our study provides reliable real‐world data. However, the study was nonrandomized, and the size of the effect will be biased with other unknown confounders. In addition, this was a single‐center observational study, and the generalizability of the findings is uncertain. Therefore, further larger‐scale multicenter studies are needed to validate the generalizability of our results.

Conclusions

The peripheral vascular endothelial function before AF ablation is a useful tool for the risk stratification of the incidence of 5‐year cardiovascular events. Further studies will provide insight into whether a change in the endothelial function after AF ablation is associated with the incidence of cardiovascular events.

Sources of Funding

None.

Disclosures

None.

Supporting information