Addition of Atrial Myopathy to HATCH Score for Predicting New‐Onset Atrial Fibrillation After Ablation of Atrial Flutter

Authors: Teerapat Nantsupawat, Yanhui Li, Stephanie Li, Neeraj Sathnur, Supavit Chesdachai, Selcuk Adabag, David G. Benditt, Venkatakrishna N. Tholakanahalli

Abstract

Background

The available risk prediction models are inadequate to identify a true low‐risk patient for developing new‐onset atrial fibrillation (AF) after typical atrial flutter (AFL) ablation. We aimed to determine whether adding markers of atrial myopathy to HATCH score (hypertension, age ≥75 years, transient ischemic attack/stroke, chronic obstructive pulmonary disease, and heart failure) can improve prediction of new‐onset AF after ablation of typical AFL.

Methods

The study included 208 consecutive patients who underwent successful ablation of typical AFL at Minneapolis VA Medical Center and University of Minnesota Medical Center. Patients with history of AF prior to ablation were excluded.

Results

Among the 208 patients, 76 (36.5%) developed new‐onset AF post AFL ablation. Mean follow‐up duration was 62 ± 31.8 months. HATCH score was not associated with new‐onset AF. When adding atrial myopathy (presence of at least one of the PTFV1 > 5000 µV*ms, interatrial block determined by biphasic inferior p wave with duration >120 ms, left atrium (LA) diameter ≥44 mm, or LA index ≥3 cm/m^2^) to HATCH score (HATCH‐A2), the combination was independently associated with new‐onset AF. The AF incidence between HATCH‐A2 score of 0–1 and ≥2 were 6.7% and 40.5%, respectively (Odds ratio 9.04, 95% confidence 2.05–39.81, p = 0.004). Particularly, when HATCH‐A2 score was 0, none of the patients developed AF.

Conclusions

Adding atrial myopathy to HATCH score improved predictability and could be used to delineate a true low‐risk patient of new‐onset AF after typical AFL ablation.

Introduction

Isthmus‐dependent atrial flutter (AFL) is commonly treated by cavotricuspid isthmus (CTI) ablation with a high long‐term success rate of 92%–95% [1], and consequently patients with isolated AFL without a prior history of atrial fibrillation (AF), are usually considered cured after successful AFL ablation. Further, although there are not as yet formal practice guideline recommendations, in such cases many physicians do not feel the need to continue oral anticoagulation (OAC). However, recent reports have shown a 13.9%–26.2% incidence of apparently new‐onset AF post CTI ablation, and up to 45% in a more intensively monitored subgroup of patients (i.e., either by >7 days/year Holter monitoring or by implanted cardiac devices) [2]. These latter findings have tended to shift practice toward continuation of OAC after ablation of isolated AFL. Inevitably, however, continued anticoagulation exposes not only those patients deemed at high‐risk of later developing AF, but also those thought to be at low‐risk, of both the expense of medications and potential bleeding complications.

Multiple studies have attempted to define markers that predict high AF‐risk following isolated AFL ablation; typically, these markers have included structural/valvular heart disease [2], increased left atrial (LA) diameter [2, 3, 4], obstructive sleep apnea [4], interatrial block (i.e., prolonged P‐wave) [5], ejection fraction <50% [1], chronic obstructive pulmonary disease [3, 6], and hypertension [2]. Other AF‐risk factors that might additionally be used increasing age, diabetes mellitus, heart failure, and obesity.

No single factor or risk score has been able to identify true low‐risk patients that may not need long‐term OAC post AFL ablation. In this regard, the HATCH score [7] consists of hypertension (1 point), age ≥ 75 (1 point), history of TIA/stroke (2 points), chronic obstructive pulmonary disease (1 point), and heart failure (2 points). It has been used in an attempt to predict AF post AFL ablation. However, the results have been imperfect.

Given that certain readily accessible markers of ‘atrial myopathy’ including LA enlargement, presence of interatrial block, and left/right atrial remodeling are well established risk factors for developing AF, we hypothesized that incorporating these parameters into HATCH score will improve the ability to differentiate individuals at high AF risk after AFL ablation from a patient subset at low‐risk.

Materials and Methods

Patient Selection

This study retrospectively reviewed findings in all patients who had isthmus‐dependent atrial flutter and underwent successful CTI ablation between January 1, 2006 and December 31, 2014 from University of Minnesota Medical Center in Minneapolis, and the Minneapolis VA Medical Center. Patients with prior history of AF or follow‐up duration less than 1 year were excluded from analysis to avoid selection bias due to loss to follow‐up.

Definitions

AF was defined to be present if documented by one or more of the (1) 12‐lead electrocardiogram (ECG); (2) Rhythm strip lasting > 30 s from telemetry, Holter recording, or an implantable cardiac monitor (ICM), with irregular RR intervals unrelated to evident ventricular or atrial ectopy and no discernable P‐waves; or (3) Electrogram (EGM) from a cardiac implantable electronic device (CIED) showing atrial high rate episode ≥175 bpm lasting ≥5 min [8]. The AF diagnoses were independently adjudicated by two electrophysiologists who were blinded of the patients’ profile.

Atrial myopathy definition in our study was ‘Any complex of structural, architectural, contractile or electrophysiological changes affecting the atria with the potential to produce clinically‐relevant manifestations’ [9]. Atrial myopathy was considered present if any one of the following criteria was LA anteroposterior (AP) diameter ≥44 mm, LA diameter index ≥3 cm/m^2^ of body surface area, PTFV1 > 5000 µV*ms, or interatrial block P‐wave [9, 10].

Interatrial block was defined by P‐wave duration >120 ms and biphasic morphology in inferior leads (Figure 1a) [5]. P‐wave terminal force (PTFV1) was defined as the duration in milliseconds (ms) of the terminal part (negative) of the P wave in lead V1 multiplied by its depth in microvolts (Figure 1b) [11].

Echocardiographic analysis used a LA AP diameter cutoff of ≥44 mm for LA dilatation (from averaging LA AP diameter cutoff from multiple studies) [2, 3, 4, 12].

Routine follow‐up was defined as clinic visit in 1–2 months post AFL ablation procedure and then every 3–6 months with 12‐lead ECG at each visit. Holter/event monitor follow‐up comprised routine follow‐up plus at least one 24–48 h Holter monitor or 14–30 days event monitor. CIED follow‐up incorporated routine follow‐up plus regular clinic visit/remote monitoring of pacemaker, defibrillator, or implantable loop recorder every 6 months.

Predictor Variables

We collected baseline characteristics and pre‐specified risk factors for developing new‐onset AF, including age, LA AP diameter, obstructive sleep apnea, left ventricular ejection fraction, at least moderate severity valvular disease, chronic obstructive pulmonary disease, hypertension, CHA2DS2VASc, interatrial block p wave, PTFV1, atrial myopathy, HATCH and proposed HATCH‐A2 score. All predicting variables were collected at the time of the typical flutter ablation. ECGs for the criteria of atrial myopathy were assessed before the CTI ablation as the ablation can affect ECG especially P wave in inferior leads.

For HATCH score, points were assigned to each factor according to the original recommendation [13]. In evaluating the proposed HATCH‐A2 score, we conducted a multivariable logistic regression analysis to compare its predictive power against the HATCH score and atrial myopathy for the incidence of new‐onset AF following AFL ablation. Interactions among variables in the final model were assessed as well. We assigned 2 points to atrial myopathy and 1 point for each risk factor in HATCH because atrial myopathy demonstrated stronger association to subsequent new‐onset AF than 2 other risk factors combined.

Catheter Ablation

Isthmus‐dependent AFL was confirmed prior to ablation by 3D activation mapping using CARTO system (Biosense Webster). Entrainment was performed by pacing at coronary sinus proximal and distal electrodes, cavotricuspid isthmus site, and lateral right atrium. CTI ablation was performed using either 8 mm non‐irrigated (50–60 W, 60°C) or 3.5 mm irrigated (30–40 W) ablation catheter. Successful AFL ablation was confirmed by evidence of bidirectional block and differential pacing activation mapping.

Outcome Variables

Primary outcome was incidence of AF in the population post CTI ablation. Secondary outcome was incidence of ischemic stroke/systemic thromboembolism.

Statistical Analysis

Given the previously reported incidence of new‐onset AF post AFL ablation was about 25%, if the risk prediction model is to be able to divide patients into low‐risk group (risk of developing new‐onset AF of ≤5%) versus high‐risk group (risk of developing new‐onset AF of ≥30%), then 170 patients will provide statistical power of 95%. Chi‐square and t‐test were used to compare baseline characteristics. We used logistic regression to identify association/odds ratio of risk factors and new‐onset AF post AFL ablation. We performed multivariate logistic regression analysis by using pre‐specified risk factors for AF from literature reviews (obstructive sleep apnea, valvular disease, obesity, and diabetes mellitus) except factors that are parts of HATCH or HATCH‐atrial myopathy score (HATCH‐A2). All analyses were performed using SPSS Windows version 21, SPSS Inc. Chicago, IL USA.

Results

Baseline Characteristics

There were 231 consecutive patients with isolated typical AFL who underwent successful CTI ablation between January 1, 2006 and December 31, 2014. We excluded 23 patients who had follow‐up duration less than 1 year. Therefore, 208 patients were included in the analysis. Among the 23 patients with less than 1‐year follow‐up, there was only one patient who developed AF. Mean age was 64.8 ± 12.1 years old. Majority of patients were male (91.3%). Mean LA AP diameter and ejection fraction were 45.2 ± 7.2 mm and 49.2 ± 11.8%, respectively. Mean CHA2DS2VASc was 2.8 ± 1.5 with 79.3% of patients had score ≥2. There were 147 (70.7%) patients taking beta blockers and 14 (6.7%) patients prescribed antiarrhythmic drugs. There were no significant differences in baseline characteristics between patients who developed new‐onset AF and those who did not (Table 1). Mean follow‐up duration was 62.0 ± 31.8 months; the follow‐up duration ranged from 12 to 136 months. Of the 208 patients, types of follow‐up were routine office visit in 124 (59.6%) patients, CIED clinic follow‐up in 44 (21.2%) patients, Holter monitor in 24 (11.5%) patients, and event monitor in 16 (7.7%) patients.

Incidence of AF

Among the 208 patients in the study population, there were 76 (36.5%) individuals who developed new‐onset AF during follow‐up. Mean duration from CTI ablation to AF onset was 31.5 ± 26.3 months; the AF onset varied from 1 day to 8.9 years post‐ablation. New‐onset AF was detected earlier in Holter/event monitor and CIED group (22.0 ± 22.7 and 28.2 ± 22.6 months, respectively), compared to routine follow‐up group (37.4 ± 29.1 months), presumably due to the intensity of ECG recording in the former groups. Rate of AF detection was 38/124 (30.6%) in routine follow‐up, 13/40 (32.5%) in Holter/event monitor, and 25/44 (56.8%) in CIED follow‐up.

Risk Factors Predicting AF

As demonstrated in Table 1, a number of possible risk factors proved not to be associated with increased risk of new‐onset AF. Similarly, except for LA AP diameter, individual markers of atrial myopathy including interatrial block p wave, PTFV1, and LA diameter index had a trend toward increased risk of new‐onset AF from the univariate logistic regression analysis but not statistically significant (Table 2). However, when any one of these individual markers was used to define one entity, ‘atrial myopathy’, sensitivity to AF occurrences was increased and the combined marker became an independent predicting factor of new‐onset AF (OR of 3.92; 95% CI 1.84–8.38, p value < 0.001) (Tables 2 and 3). On the other hand, there was no cumulative effects from four different markers of atrial myopathy in predicting AF (data not shown).

Neither CHA2DS2VASc score nor HATCH score was predictive of increased risk of new‐onset AF in both univariate and multivariate analysis. Two points were then assigned to the atrial myopathy factor in HATCH‐A2 score as described earlier in the method section. Adding atrial myopathy (2 points) to HATCH score (HATCH‐A2) with a cutoff value of 2, improved HATCH score predictability of new‐onset AF. When compared predictability of HATCH‐A2 score to atrial myopathy, HATCH‐A2 score with a cutoff value of 2 was overall better than atrial myopathy alone in identifying low‐risk patients (OR 9.04; 95% CI 2.05–39.81, p = 0.004 vs. OR 3.92; 95% CI 1.84–8.38, p < 0.001, and negative predictive value [NPV] of 93.3% vs. 82.8%, respectively) (Table 3).

When HATCH score was 0, there were five out of 23 patients (21.7%) who developed new‐onset AF as compared to 0 out of 11 patients (0%) when HATCH‐A2 score was 0 (Figure 2). AF incidence in patients with HATCH‐A2 0–1 was 6.7% (2/30), as compared to 40.5% (70/173) in HATCH‐A2 ≥2 group. Although HATCH‐A2 score have improved discriminating ability as compared to HATCH score in low‐risk group, HATCH‐A2 score did not affect the risk stratification in the higher risk population when HATCH or HATCH‐A2 score were ≥2 (Figure 2). As demonstrated in Figure 3b, patients who had atrial myopathy would automatically had HATCH‐A2 score of 2 or more, and therefore left the HATCH‐A2 0–1 group without atrial myopathy.

HATCH‐A2 cutoff value of 2 provided sensitivity and specificity in predicting new‐onset AF of 97.2% and 21.4%, respectively, with a negative predictive value of 93.3% and a positive predictive value of 34.5%.

Sensitivity Analysis

If we include 23 patients who had follow‐up duration of less than 1 year in the analysis. HATCH‐A2 ≥2 remained associated with a higher risk of new‐onset AF post AFL ablation (OR 8.66, 95% CI 2.01–37.32, p = 0.004).

Ischemic Stroke After CTI Ablation

Four patients (1.9%) or 3.72 per 1000 person‐years developed ischemic stroke with the onset of stroke from 0.2 to 40 months post ablation (Table 4). Two of them had new‐onset AF post AFL ablation but other two remained in sinus rhythm during follow‐up. All four patients had HATCH‐A2 score more than 2. Two patients had stopped OAC prior to ischemic stroke. Two patients developed ischemic stroke after ablation despite continuing OAC.

Discussion

Main Findings

This study provides four main findings regarding risk prediction of AF occurrence after successful AFL ablation. First, the incidence of new‐onset AF in patients post‐ablation of isolated AFL was 36.5% over 5 years which is similar to Chen et al. study with incidence of 39% [7]. Second, neither HATCH score nor CHA2DS2VASc score were able to predict new‐onset AF in patient after AFL ablation. Third, HATCH‐A2, a modified score system by incorporating the combined ‘atrial myopathy’ risk score to the HATCH score, greatly improved the sensitivity of predicting new‐onset AF post AFL ablation. Finally, in our study, mean duration post CTI ablation to AF onset was 31.5 months. This may represent actual delayed onset of AF, or delayed detection of AF, or both, but in either case emphasizes the importance of longer‐term follow‐up after apparently successful AFL ablation.

Relation of AF and AFL

When compared to AF incidence in a general population of 1.4%, the higher AF incidence in patients after ablation of typical AFL can be explained by the apparent close pathophysiologic interrelationship between AF and AFL. Typical AFL could be preceded by AF by either causing a functional line of block or initiating re‐entry around the preexistence anatomical line of block between the vena cava [14]. Atrial electrical and structural remodeling from atrial flutter, both of which could potentially be a substrate for AF, have also been shown in human studies [15].

HATCH and HATCH‐A2 Score

HATCH score was first developed as a tool to predict progression of paroxysmal AF to persistent AF [16]. Based on the multivariate analysis, clinical characteristics such as hypertension, age, previous transient ischemic attack or stroke, chronic obstructive pulmonary disease, and heart failure were shown to be independent factors in predicting AF progression and thus were included in the calculation of HATCH score. HATCH score was shown by Chen et al. to be a useful tool in predicting new‐onset AF following isolated AFL ablation [7]. However, in that study, patients with HATCH score of 0 and 1 still had AF incidence post ablation of 9.8% and 32.4%, respectively. In our study, 21.7% and 38.2% of patients with HATCH score of 0 and 1, respectively, had new‐onset AF post AFL ablation. There were significant portion of patients with HATCH score of 0 and 1 that had atrial myopathy (5.9% and 23.2%, respectively), which could explain why HATCH score failed to predict new‐onset AF post AFL ablation from our current study and the study from Garcia‐Seara et al. [13].

LA enlargement, presence of interatrial block, and left/right atrial remodeling are well established risk factors for developing recurrent AF post AF ablation, and accepted as markers of atrial myopathy. Our results showed that atrial myopathy was independently associated with new‐onset AF post CTI ablation. This is consistent with other atrial myopathy parameters studies [2, 3, 4, 5]. Although atrial myopathy alone was independently associated with increased AF incidence post AFL ablation, by adding atrial myopathy to HATCH score, we further improved the ability in identifying true low‐risk patients. Although with low specificity, HATCH‐A2 serves its purpose of providing a high sensitivity of 97.2% in predicting new‐onset AF as compared to HATCH score sensitivity of 59.2% when using a same cutoff value of ≥2.

Stroke

The incidence of ischemic stroke in our study was 1.9%, which is considerably lower than 5.85% reported by Exposito et al. on patients post AFL ablation [17]. Although mean CHADS2 score, post ablation OAC discontinuation rate, and AF incidence were similar, more patients in the study from Exposito et al. had prior stroke, which was the strongest risk factor for developing stroke in AF [18]. Despite of only four patients with ischemic stroke post AFL ablation, all of them had HATCH‐A2 score ≥2. The risk of developing new‐onset AF when HATCH‐A2<2 is 1.86 per 100 person‐years in our study, weighing against major bleeding from OAC of 3.57 per 100 person‐years [19]. Whether it is reasonable to consider discontinuing OAC post AFL ablation when HATCH‐A2<2, especially if patients have high bleeding risk, needs further validation in a large randomized control study.

Strengths and Limitations

The strengths of our study include relatively robust follow‐up methods post isolated AFL ablation incorporating routine clinic visits, Holter/event monitors, and when available CIED interrogation; these follow‐up techniques reflected a real‐world practice that has a mixture of clinical AF, subclinical AF, and also AF underdetection. A second strength was the relatively long mean follow‐up duration of 62 months, which increased the likelihood of detecting late‐onset AF post AFL ablation. Finally, the sample size and AF incidence were adequate to provide statistical power of >90%.

Several limitations of our analysis are important to consider. First, although the sample size had adequate statistical power, it was still relatively small skewed population with male predominance. Second, LA size was determined echographically by LA AP diameter in parasternal‐long axis, which is less accurate than LA volume. Third, post AFL ablation monitoring techniques were varied and lacked certainty regarding AF duration; the latter may have an impact on AF detection and stroke risk. Finally, about 60% of patients had routine clinic visit follow‐up and subclinical AF was possibly missed. Nonetheless, LOOP study demonstrated that detection of subclinical AF by implantable loop recorder and initiated OAC accordingly did not reduce the risk of stroke or systemic arterial embolism [20].

Conclusions

In our study, HATCH score did not predict new‐onset AF in patients after AFL ablation. Atrial myopathy is an independent risk factor for new‐onset AF post‐AFL ablation. When added atrial myopathy to HATCH score (HATCH‐A2), it further delineated true low‐risk patients in developing AF post‐AFL ablation.

Author Contributions

T.N. designed the study, collected data, performed statistical analysis, interpreted the data, and wrote the manuscript. Y.H., N.S., S.C., S.L. collected and re‐checked the data prior to the analysis. J.L., V.T. designed the study, reviewed, and revised the manuscript. D.B., S.A. performed data interpretation and critically revised the manuscript. All authors agreed to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Ethics Statement

The study protocol was approved by the Institutional Review Boards of all participating sites (STUDY00002032).

Consent

The requirement for individual consent was waived.

Conflicts of Interest

The authors declare that they have no pertinent commercial associations or sources of support that might pose a conflict of interest. Dr. Benditt was supported in part by a grant from the Dr. Earl E Bakken family in support of Heart‐Brain research.