Author + information
- Received July 19, 2017
- Revision received November 27, 2017
- Accepted December 4, 2017
- Published online February 7, 2018.
- Amir Jadidi, MDa,b,∗ (, )
- Björn Müller-Edenborn, MDa,
- Juan Chen, MDb,c,
- Cornelius Keyl, MDd,
- Reinhold Weber, MDa,b,
- Jürgen Allgeier, MDa,b,
- Zoraida Moreno-Weidmann, MDb,
- Dietmar Trenk, PhDe,
- Franz-Josef Neumann, MDa,
- Heiko Lehrmann, MDa,b and
- Thomas Arentz, MDa,b
- aDepartment of Cardiology, University Heart Center Freiburg–Bad Krozingen, Bad Krozingen, Germany
- bDepartment of Electrophysiology, University Heart Center Freiburg–Bad Krozingen, Bad Krozingen, Germany
- cCardiovascular Department, The First People’s Hospital of Jingmen, Hubei, China
- dDepartment of Anesthesiology, University Heart Center Freiburg–Bad Krozingen, Bad Krozingen, Germany
- eDepartment of Pharmacology, University Heart Center Freiburg–Bad Krozingen, Bad Krozingen, Germany
- ↵∗Address for correspondence:
Dr. Amir Jadidi, Department of Rhythmology, University Heart Center Freiburg–Bad Krozingen, Südring 15, 79189 Bad Krozingen, Germany.
Objectives Left atrial (LA) low-voltage substrate (LVS) potentially slows intra-atrial conduction, which might identify patients at risk for arrhythmia recurrence following pulmonary vein isolation (PVI).
Background Up to 50% of patients with persistent atrial fibrillation (AF) have arrhythmia recurrence following PVI, mostly due to arrhythmogenic LA LVS.
Methods Seventy-two patients with persistent AF underwent electrocardioversion to sinus rhythm and high-density voltage mapping of the left atrium. Invasively measured LA activation time and P-wave duration (PWD; total PWD and LA PWD [measured from −dV/dt in leads V1 and V2 until the end of the P-wave]) on amplified (40 to 50 mm/mV, 100 to 200 mm/s) digitized 12-lead electrocardiography (ECG) were compared with the extent of LA LVS (<0.5 and <1. 0mV). Freedom from arrhythmia following PVI was evaluated in 143 patients with persistent AF stratified according to amplified PWD before ablation.
Results LA LVS resulted in regional conduction delay, which increased LA activation time (r = 0.79). LA PWD strongly correlated with LA activation time (r = 0.96) and LA LVS (r = 0.80). As the first (right atrial) portion of the P wave (from P-wave beginning until −dV/dt in leads V1 and V2) was not affected by LA LVS, total PWD correlated with LA LVS (r = 0.84). PWD ≥150 ms identified advanced LA LVS with 94.3% sensitivity and 91.7% specificity. One-year arrhythmia freedom following PVI-only was significantly higher in patients with PWD <150 ms (n = 73) compared with those with prolonged PWD ≥150 ms (n = 70) (72.0% vs. 40.8%, p = 0.003).
Conclusions Advanced arrhythmogenic LVS is associated with significant intra-atrial conduction delay, which is accurately measurable by prolongation of PWD on amplified 12-lead ECG. PWD ≥150 ms during sinus rhythm measured prior to ablation identifies patients with persistent AF who are at increased risk for arrhythmia recurrence following PVI.
Current ablation strategies for the treatment of persistent atrial fibrillation (AF) have shown limited success rates, with recurrence after 1 year occurring in about 50% of patients following pulmonary vein isolation (PVI) (1,2). Arrhythmia recurrence essentially occurs in patients with fibrofatty left atrial (LA) remodeling, which represents an additional arrhythmogenic substrate besides the pulmonary veins (3–5). The affected parts of the left atrium can be identified using voltage mapping, in which a lower-than-normal bipolar voltage is found (3,5,6); local delayed enhancement on gadolinium-contrast magnetic resonance imaging (MRI) (7,8); and histopathology (9).
Improved success rates have been reported when the arrhythmogenic low-voltage substrate (LVS) was targeted by ablation in addition to PVI (3,5,6,10). Patients with persistent AF and normal LA voltage, in contrast, demonstrated low arrhythmia recurrence rates following PVI (5,6). Similarly, identification of the arrhythmogenic substrate using gadolinium-contrast MRI demonstrated a clear correlation between the extent of LA remodeling and freedom from arrhythmia following PVI (8).
The real-world application of these approaches in everyday use is, however, limited by the need for an intra-atrial mapping system or MRI with a dedicated LA sequence and gadolinium injection.
Recent studies in patients with paroxysmal AF revealed that prolonged P-wave duration (PWD) in sinus rhythm (SR), recorded mostly after PVI, was associated with higher arrhythmia recurrence rates (11–15). However, it remains unclear to what extent the observed P-wave prolongations were induced or affected by the ablation procedure itself and whether these findings can be extrapolated to patients with presumed electrostructural remodeling of the left atrium, as in persistent AF. Moreover, the cutoff values for reported PWD suggested to predict future arrhythmia recurrences were highly variable, ranging from 120 to >165 ms (11–15).
We hypothesized that progressive LA remodeling in persistent AF with development of LA LVS leads to intra-atrial conduction slowing that should be measurable using intracardiac activation mapping and be reflected by prolonged PWD on amplified digitized 12-lead electrocardiography (ECG). If so, marked prolongation of the amplified SR P-wave might predict arrhythmia recurrences in patients with persistent AF scheduled for PVI.
The study protocol was approved by the local ethics committee, and patients gave written informed consent. Consecutive patients with symptomatic persistent AF (lasting >7 days and <12 months) scheduled for first PVI between July 2015 and December 2016 were included. Patients underwent high-density LA electroanatomic mapping (>1,200 points) in SR. All patients underwent electric cardioversion 6 to 10 weeks prior to PVI, and antiarrhythmic therapy was established in most patients to favor atrial reverse remodeling in SR (16). Patients with AF recurrence underwent cardioversion directly prior to PVI.
Outcomes following PVI were evaluated in a validation cohort of 143 consecutive patients who underwent first PVI without additional LA ablations for symptomatic persistent AF between January 2013 and December 2015. These patients underwent the same approach with cardioversion prior to PVI as described previously and were stratified according to the amplified PWD in SR that was recorded prior to PVI. Antiarrhythmic drugs, if applicable, were continued for 3 months following ablation and then stopped in all patients. All patients underwent routine ambulatory cardiologic control, including 12-lead ECG 1 month following PVI. Twenty-four-hour ECG was scheduled at 6 and 12 months following PVI. Symptomatic patients (experiencing palpitations, dyspnea, or fatigue) underwent additional ambulatory cardiologic examinations, including 12-lead ECG and 24-h Holter recordings. If 12-lead ECG and 24-h ECG revealed no arrhythmia recurrences, an event recorder (a device that is activated by the patient during symptoms and records the surface ECG) was given to the patient to record symptomatic episodes. AF or atrial tachycardia lasting >30 s or recorded on event-triggered ECG past the initial 3-month blanking period and during the first 12 months after PVI were considered recurrences.
Amplified P-wave analysis from 12-lead ECG
Twelve-lead ECGs in standard position were recorded digitally in all patients prior to ablation using a Schiller AT-104 PC (Schiller, Baar, Switzerland) and Bard Labsystem Pro (Boston Scientific, Marlborough, Massachusetts) during the procedure (0.05 to 100 Hz, no additional noise filtering, sample rate 1,000 Hz). For best identification of the earliest and the latest component of the P wave in 1 of the 12 electrocardiographic leads, SR electrocardiographic recordings were amplified to 0.2 to 0.25 mV/cm (corresponding to 40 to 50 mm/mV) with a sweep speed of 100 to 200 mm/s (Online Figure 1). Measurements of PWD were performed manually by 2 independent cardiologists, each blinded to the patients’ characteristics and outcomes, using a digital caliper. PWD was measured from the earliest initial deflection from the isoelectric line in any lead to the time of the latest activation in any lead, with particular attention paid to leads exhibiting terminal forces: D1, V5 and V6, and aVR and in some cases V1 (sagittal force) as representative for latest activated LA sites.
Electrophysiological study and high-density voltage mapping in SR
Patients underwent high-density electro-anatomic mapping (>1,200 mapped LA sites) using a 20-pole circumferential catheter (1-mm electrode size, 2-5-2 mm spacing, and variable diameter of 15 to 25 mm; Optima [St. Jude Medical, St. Paul, Minnesota] or Lasso [Biosense-Webster, Diamond Bar, California]) in combination with the navigation system Ensite Velocity (St. Jude Medical) or Carto 3 (Biosense-Webster). Additionally, the following catheters were used: 1) a decapolar catheter in the coronary sinus (CS); 2) a tip-irrigated quadrupolar catheter with a distal 3.5-mm tip and 2-5-2 mm spacing (ThermoCool SmartTouch Navistar [Biosense-Webster] or TactiCath [St. Jude Medical]) for ablation.
Activation and voltage-map settings in SR
Map acquisition was performed by sequential multielectrode mapping during SR. Peak-to-peak intracardiac bipolar electrograms were recorded at 15 to 250 Hz and amplified to 0.1 to 0.2 mV/cm to display low-voltage electric activity. To achieve the highest accuracy of low-voltage identification, the electroanatomic maps were acquired after respiratory gating with low interpolation settings (Carto 3 system, 17 to 19; Ensite Velocity system, 12) and exclusion of mapping points >5 mm distant from the LA geometry. To discriminate LVS from poor wall contact, low-voltage electrograms were reconfirmed by contact-sensing enabled catheters (contact force >3 g).
LA activation time (LAT) in SR was measured on the electroanatomic map as the earliest activation of the left atrium (usually at the anteroseptum/roof or at the inferoseptal left atrium in case of Bachmann’s bundle block) to the latest activation of the lateral left atrium or LA appendage.
Quantification of low-voltage areas
LVS was quantified using the area measurement software available with the navigation mapping system. Two bipolar voltage thresholds in SR were chosen to assess the extent of low-voltage area: 1) areas with bipolar voltage <0.5 mV; and 2) areas with bipolar voltage <1.0 mV. The pulmonary vein ostia and mitral annulus were excluded. LVS confined to the peri–mitral annulus area <2 cm2 at a voltage threshold of <0.5 mV was considered physiological because of reduction in electrogram amplitude or voltage at the mitral annulus.
The procedure was performed under heparin anticoagulation with a targeted activated clotting time of 300 to 350 s prior to transseptal access. The procedural endpoint of ablation was achievement of PVI by circumferential ablation around the pulmonary vein ostia (17). Irrigated-tip catheter ablation was performed at 20 to 25 W for up to 20 s (contact force 4 to 12 g) at the posterior left atrium and within the CS and 28 to 35 W for up to 45 s (contact force >8 g) at other areas.
Statistical analysis was performed using SPSS version 22.0 for Macintosh (IBM Corporation, Armonk, New York) or GraphPad Prism version 6 for Macintosh (GraphPad Software, La Jolla, California). Data were checked for normal distribution using the Shapiro-Wilk test. Normally distributed data are expressed as mean ± SD, and skewed data are expressed as median (interquartile range). Scaled variables were compared between groups using Student t test or nonparametric testing depending on normality. Categorical variables were compared using the Fisher exact test. Linear regression was used to predict LA low-voltage area from LAT, LA PWD (LAP) or total PWD. The Bland-Altman method with a 95% limit of agreement was used to assess the pairwise agreement of LAT and LAP. Intraclass correlation coefficient estimates and their 95% confidence intervals on the basis of PWD measurements by 2 independent cardiologists were calculated on the basis of a 2-way mixed-effects model for consistency. Receiver-operating characteristic curve analysis was used to determine the optimal PWD regarding sensitivity and specificity in predicting the absence (<2 cm2 at 0.5 mV) or presence of LVS.
Arrhythmia-free survival (freedom from AF and atrial tachycardia) in the validation cohort is demonstrated using Kaplan-Meier curves for patients with short (<150 ms) and long (≥150 ms) PWD. Arrhythmia-free survival was compared between both groups using the log-rank test. Recurrence of arrhythmia during the blanking period of 90 days following PVI was not considered recurrence in the statistical analysis.
Covariates of potential clinical interest (PWD, age, sex, atrial hypertension, LA diameter, structural cardiomyopathy, and antiarrhythmic therapy) were documented for all patients. All baseline characteristics and potential confounders were tested initially in univariate analysis. Covariates with p values <0.10 in the univariate model, as well as covariates with p values >0.10 but with a potential clinical impact on prognosis, were included in the multivariate model. Cox proportional hazards models were performed using a backward-stepwise approach. A 2-sided p value ≤0.05 was considered to indicate statistical significance in all tests.
Patient characteristics and extent of LVS
Seventy-two consecutive patients scheduled for first PVI for persistent AF underwent electroanatomic mapping. All patients were successfully electrically cardioverted to SR 6 to 10 weeks prior to PVI. Nine patients (13%) presented with AF recurrence and were cardioverted to SR at procedure beginning. The clinical characteristics of all patients are given in Table 1. Most patients were men, with preserved left ventricular systolic function. Structural or ischemic cardiomyopathy was present in 7 (9.7%) and 13 (18%) patients, respectively. Sixty-four percent of patients were on antiarrhythmic drugs, mainly amiodarone. Although all patients presented with the clinical phenotype of persistent AF (AF lasting for >7 days and <12 months), we found high variability in the extent of LVS independent of the pre-defined voltage threshold (10.2 [15.2] cm2 at 0.5 mV and 20.6 [22.8] cm2 at 1.0 mV).
Identification of LA activation on surface ECG
In 50 of 72 mapped patients, the earliest activated areas of the left atrium during SR were the anteroseptal and septal roof regions. In these patients, early activation coincided with the second negative portion of the SR P wave, which begins at the maximum negative dV/dt in the precordial lead V1 or V2, indicating activation of the left atrium through the Bachmann bundle (Figure 1).
Interatrial block with biphasic P waves in the inferior leads due to conduction slowing over Bachmann’s bundle was present in the remaining patients. In these patients, the left atrium was activated via the CS, with the earliest activation of the LA region at the inferoposterior wall and hence a lateral and superior axis (from septal to lateral and from the CS/inferior LA toward the high posterior left atrium/LA roof), leading to a negative deflection of the initially positive P-wave portion in the inferior leads (Figure 2) (18). Independent of the extent of LVS, we observed the latest activation of the left atrium at the LA appendage and lateral left atrium. The latest component of the P wave is therefore best identified in leads detecting lateral (D1, V5 and V6, and aVR) and sagittal forces (V1 and V2).
Invasive mapping demonstrates correlation of LVS with LAT
At both voltage thresholds that were set to detect LA LVS (<0.5 and <1.0 mV), increasing extent of LA LVS was associated with a decreased intra-LA conduction velocity within the low-voltage area, leading to a prolongation of LAT (Figures 3A and 3B). The best correlation for LAT and LVS was found with a definition for LVS as areas displaying bipolar voltages <1.0 mV in SR (r = 0.79 for <1.0-mV cutoff, r = 0.73 for <0.5-mV cutoff; p < 0.001).
Amplified PWD on 12-lead ECG correlates with the extent of LA LVS
As described earlier, LAP was measured as maximum negative dV/dt in leads V1 and V2 (when lead V1 or V2 was biphasic positive-negative) or maximum negative dV/dt in the inferior leads (when interatrial block was present and in the absence of biphasic P waves in leads V1 and V2) until the end of the P-wave in any of the amplified 12 electrocardiographic leads. LAP strongly correlated with invasively measured LAT (r = 0.96, p < 0.0001) (Online Figure 2A). A Bland-Altman plot for LAT versus LAP calculated for difference versus average is shown in Online Figure 2B. Consequently, LAP showed a high correlation with the extent of LA LVS (r = 0.80, p < 0.0001) (Online Figure 2C). As the right atrial P-wave (measured from the initial deflection of the P-wave to dV/dt in leads V1 and V2 or the inferior leads) showed only weak correlation with the extent of LA LVS (r = 0.46), simple measurement of the amplified total PWD was sufficient to predict the extent of LA LVS (r = 0.84 for <1.0 mV as LVS threshold and r = 0.78 for <0.5 mV as LVS threshold, p < 0.0001) (Figures 4A and 4B). With proper amplification and sweep speed (40 to 50 mm/mV and 100 to 200 mm/s, respectively), the interobserver reproducibility of PWD measured by 2 independent and blinded investigators was excellent (intraclass correlation coefficient = 0.945; 95% confidence interval: 0.896 to 0.971).
Amplified PWD in SR: diagnostic value and accuracy for identification of arrhythmogenic LA LVS
To determine the optimal threshold value of the amplified PWD in SR that identifies significant LVS, receiver-operating characteristic curve analysis was used on the basis of the absence or presence of LVS. Receiver-operating characteristic curve analysis revealed that duration of amplified P-wave in SR ≥150 ms identified the presence of LA LVS with 94.3% sensitivity and 91.7% specificity (Figure 5).
Notably, patients with amplified PWD <150 ms had significantly less LVS (using both 0.5 and 1.0 mV as cutoff values for LVS detection), compared with patients with prolonged P waves (p < 0.0001 for <150 ms vs. ≥150 ms at 0.5 and 1.0 mV) (Figure 6).
PWD in patients with and without antiarrhythmic drugs did not differ (Online Figure 3A). Also, correlation of PWD and extent of LVS was not affected by antiarrhythmic drugs (Online Figure 3B). Additional antiarrhythmic medications during the blanking period of 90 days following PVI had no influence on arrhythmia recurrence (Online Figure 3C).
Amplified sinus-PWD measured prior to PVI predicts arrhythmia freedom in persistent AF
The prognostic value of the baseline amplified PWD on arrhythmia freedom within 1 year after PVI was evaluated in a validation cohort of 143 consecutive patients undergoing first PVI (without additional LA ablation) for persistent AF. The mapping and the validation cohorts were comparable regarding baseline characteristics (age, sex, echocardiographic parameters, and key characteristics of AF) (Table 2).
Patients were stratified using amplified PWD prior to PVI, with a cutoff value of 150 ms. Table 3 gives the clinical characteristics of stratified patients. The anteroposterior LA diameter was smaller in patients with total PWD <150 ms (42.2 vs. 45.6 mm, p < 0.001), while structural cardiomyopathies were encountered more commonly in patients with prolonged PWD exceeding ≥150 ms (4% vs. 17%, p < 0.01).
Arrhythmia freedom rate (freedom from AF and atrial tachycardia) within 12 months after PVI was significantly higher in patients with short amplified PWD (<150 ms) compared with those with prolonged amplified PWD (≥150 ms) recorded prior to PVI (72.0% for <150 ms vs. 40.8% for ≥150 ms, p = 0.0003) (Figure 7).
Of these 143 patients, 110 were in SR at the beginning of PVI following cardioversion 6 to 10 weeks earlier. In this subpopulation, freedom from arrhythmia was also higher when the amplified PWD was <150 ms (70.5% for <150 ms vs. 44.9% for ≥150 ms, p = 0.012). Thirty-one of the 143 patients had recurrence of AF following prior cardioversion and were electrically cardioverted to SR at the beginning of PVI. Again, considering the most recent electrocardiogram in SR prior to PVI, 1-year freedom from arrhythmia was higher in patients with amplified PWD <150 ms (72.7% for <150 ms vs. 30% for ≥150 ms, p = 0.029). Two patients presented in atrial tachycardia and had no recurrences at 1-year follow-up.
Results of multivariate Cox regression analysis of amplified PWD including factors that are known to affect freedom from arrhythmia following PVI (age, sex, LA diameter >45mm, structural cardiomyopathy, antiarrhythmic drugs, and arterial hypertension) are given in Table 4. Of these factors, only amplified PWD was predictive of freedom from arrhythmia in patients with persistent AF undergoing PVI-only ablation (for categorical PWD, hazard ratio: 1.914; p = 0.024 with PWD ≥150 ms; for continuous PWD, hazard ratio: 1.009 per millisecond increase; p = 0.019). PWD in SR (on amplified 12-lead ECG) of <150 ms was associated with a reduced risk for arrhythmia recurrence within 1 year by 57% (hazard ratio: 0.43; 95% CI: 0.246 to 0.751; p = 0.003) in patients with persistent AF.
Four major findings can be drawn from the present study: 1) despite a common clinical phenotype that includes several prior cardioversions and a duration of AF episodes of more than 1 month, the extent of LA LVS varies importantly in patients with persistent AF; 2) extensive LA LVS slows conduction within the left atrium, which is revealed both by an increase of LAT at invasive mapping and by the PWD in SR on amplified 12-lead surface ECG; 3) invasively measured LAT is reflected on amplified surface ECG by the second portion of the P-wave (LA P-wave beginning at maximum −dV/dt in leads V1 and V2 or inferior leads until the end of P-wave) and correlates with the extent of LA LVS; and 4) prolongation of amplified P-wave exceeding 150 ms identifies patients with advanced LA LVS who are at high risk for arrhythmia recurrence following PVI.
Correlation of P-wave with extent of LA LVS
The present study demonstrates that significant LA LVS, as a likely correlate of atrial fibrofatty structural remodeling, is present in only some patients with persistent AF. Independent of the cutoff value defining low voltage in SR (bipolar voltage of <0.5 mV vs. <1.0 mV), the extent of LA LVS was highly variable, ranging from <5 cm2 to more than 50 cm2 of the left atrium.
Significant LA LVS is associated with conduction delay within LVS regions. This results in prolongation of LAT on invasive LA mapping, which is reflected by a linear prolongation of the PWD on noninvasive amplified 12-lead surface ECG.
PWD in SR of ≥150 ms thereby identifies patients with significant LA LVS with high sensitivity and specificity. Patients with short P waves (<150 ms) demonstrated only a mean 1.1 and 4.3 cm2 of LA LVS (at 0.5 and 1.0 mV, respectively), mainly corresponding to the physiological decrease in atrial voltage at the mitral annulus.
This is in line with previous work by Bayes de Luna et al. (18) describing an association of prolonged PWD with interatrial block and Bachmann bundle block with typical biphasic P-wave configuration in the inferior leads. Further studies described normal PWD in healthy adults without AF of <140 ms (19), as well as shorter P waves in paroxysmal AF compared with persistent AF (20). Lin et al. (21), in contrast, found that the extent of LA LVS increases during the course of AF, with little to no LA LVS in early paroxysmal AF and extensive LA LVS in long-standing or permanent forms. Platonov et al. (22) postulated that P-wave prolongation in AF may be due to conduction slowing and/or atrial enlargement.
Our present work adds to these observations and demonstrates that regional LA LVS as an electrophysiological correlate of local fibrofatty remodeling leads to conduction slowing, prolonged atrial activation time, and P-wave prolongation. This finding links these previous independent observations, as differences in PWD between healthy patients and those with paroxysmal or persistent AF are likely dependent on the extent of LA LVS in these patient groups.
PWD as a predictor of arrhythmia recurrence after PVI
Bayes de Luna et al. (18) suggested that fibrosis of the anteroseptal LA region accounts for the morphological changes of the P-wave, which is compatible with conduction slowing of the anterior interatrial connecting fibers (Bachmann’s bundle). In line with those findings, conduction slowing as evidenced by increased PWD was later demonstrated to identify patients at risk for paroxysmal AF (23). More recent studies have focused on the predictive value of the PWD in patients with AF for arrhythmia recurrences after PVI. These studies uniformly demonstrated that prolonged PWD was associated with a higher risk for arrhythmia recurrence (11–15) in paroxysmal AF. The reported cutoff values for PWD predicting AF recurrence post-PVI, however, were highly variable and mostly arbitrarily chosen (ranging from 120 to >165 ms) (11–15). A very recent meta-analysis of these studies concluded that a PWD >149.5 ms in SR was predictive for increased arrhythmia recurrences in patients with paroxysmal AF following PVI (24). Importantly, PWD was generally measured only after PVI or cardioversion alone, as the respective studies focused on the potential LA reverse remodeling reflected by a decrease in PWD during the follow-up period (12,16,25,26).
The unique nature of our present study lies in the effort taken at our institution to cardiovert patients with persistent AF to promote favorable reverse electric remodeling, which allows measurement of the P-wave in SR prior to PVI (27). We used the individual extent of LA LVS to establish a cutoff value of 150 ms to predict arrhythmia freedom rate following PVI-only in persistent AF, which confirms prior findings and links the arrhythmia risk to the presence of significant LA LVS in these patients.
PWD to tailor the individual ablation strategy in patients undergoing AF ablation
Although persistent AF and atrial remodeling seem intertwined, the clinical criterion of persistency often proves inadequate to classify patients regarding ablation outcome. Recent studies revealed that arrhythmia freedom crucially depends on the presence or absence of LA structural abnormalities, as identified either by delayed enhancement cardiac MRI (7,8) or LA voltage mapping (3,5,6).
In the presence of significant LA LVS, PVI plus additional ablation of target areas within the LA LVS was shown to improve arrhythmia freedom rates 1 year after the AF ablation procedure (3,5,6). The obvious limitation of this approach is that the expected success and complexity of ablation can be judged only with the patient already undergoing the intervention. In addition, it requires the use and availability of a dedicated mapping system.
Quantification of LA fibrosis using delayed enhancement cardiac MRI, in contrast, was demonstrated to be feasible using specific MRI sequences. In the DECAAF trial, the extent of delayed enhancement was a predictor of outcomes after PVI-only (8). Although allowing patient stratification prior to ablation, it is time-consuming, is expensive, requires intravenous injection of magnetic resonance contrast agents (gadolinium), and requires dedicated MRI sequence/fibrosis detection software that is likely unavailable at many hospitals.
The approach we present in the present work does not require extensive additional equipment and should be freely available to most cardiologists and electrophysiologists involved in the management of patients undergoing AF ablation. We suggest that patients with short PWD (<150 ms) should be considered to have pulmonary vein–dependent persistent AF amenable to a PVI-only approach with high anticipated success rates using either radiofrequency or cryoballoon ablation. In contrast, patients with prolonged PWD (≥150 ms) are likely to have persistent AF with additional, non–pulmonary vein–dependent foci, which likely requires more intensive atrial mapping and targeted ablation to achieve satisfying arrhythmia freedom rates (Figure 8).
Several single-center studies in more than 1,200 patients with persistent AF demonstrated higher arrhythmia freedom rates when PVI plus additional ablation of LVS was performed (3,5,6,10). Further studies involving simultaneous biatrial mapping of ongoing AF wavelets using basket catheters or body surface mapping using Cardio-Insight electrocardiographic imaging also reported higher arrhythmia freedom rates in patients with persistent AF compared with previous ablation or PVI-only strategies (10,28,29).
Large-scale prospective randomized multicenter studies (SOLVE-AF and DECAAF II) investigating the best therapeutic approach in patients with persistent AF who have atrial LVS or delayed gadolinium enhancement are currently recruiting patients.
Prolonged PWD as a potential cardiovascular risk marker
Previous observational studies have reported an increased incidence of cerebral ischemic events in patients with or without histories of AF who presented with interatrial conduction block in SR (30). A very recent study demonstrated increased cerebral ischemic event rates on MRI in patients with AF who had significant LA LVS (31). Our present study clearly relates a prolonged PWD (≥150 ms) to the presence of LA LVS, connecting the observation in the aforementioned studies. Future large-scale studies may answer if patients with prolonged PWD ≥150 ms should receive oral anticoagulant therapy independent of AF history, to prevent cerebral and cardiac ischemic events.
Although advanced LA LVS was found in a large proportion of our study patients, the present study population represents a favorable pre-selection of patients with persistent AF who mostly maintained SR under antiarrhythmic drug therapy for several weeks. Patients not cardiovertable to SR or with early AF recurrence during voltage mapping might have a higher degree of LA LVS. In our experience, however, this constitutes a minority of patients with persistent AF as long as a rhythm-control approach using antiarrhythmic drugs and electric cardioversion is applied early in the clinical history of AF. This includes patients with long-standing AF who may require a loading dose of amiodarone before electric cardioversion to enable measurement of PWD in SR.
The present study demonstrates that the PWD on amplified 12-lead surface ECG in SR correlates with the extent of LA LVS. PWD ≥150 ms identifies patients at increased risk for arrhythmia recurrence following a PVI-only ablation approach for persistent AF.
COMPETENCY IN MEDICAL KNOWLEDGE: The present work establishes amplified P-wave duration ≥150ms as a novel noninvasive tool that identifies AF patients with arrhythmogenic left atrial low voltage substrate. These patients are at increased risk for arrhythmia recurrences following PVI. Using PWD as a pre-procedural screening tool, cardiologists and patients can tailor the appropriate individual ablation strategy, enhancing procedural success rates.
TRANSATIONAL OUTLOOK: The significance of prolonged P-wave duration - as a marker of atrial fibrosis - for other cardiovascular diseases such as embolic stroke needs to be evaluated in further studies.
The authors have reported that they have no relationships relevant to the contents of this paper to disclose. Drs. Jadidi and Müller-Edenborn contributed equally to this work and are joint first authors. Drs. Lehrmann and Arentz contributed equally to this work and are joint senior authors.
All authors attest they are in compliance with human studies committees and animal welfare regulations of the authors’ institutions and Food and Drug Administration guidelines, including patient consent where appropriate. For more information, visit the JACC: Clinical Electrophysiology author instructions page.
- Abbreviations and Acronyms
- antiarrhythmic drugs
- atrial fibrillation
- coronary sinus
- left atrial
- left atrial activation time
- left atrial P-wave duration
- low-voltage substrate
- magnetic resonance imaging
- pulmonary vein isolation
- P-wave duration
- sinus rhythm
- Received July 19, 2017.
- Revision received November 27, 2017.
- Accepted December 4, 2017.
- 2018 The Authors
- Scherr D.,
- Khairy P.,
- Miyazaki S.,
- et al.
- Yang G.,
- Yang B.,
- Wei Y.,
- et al.
- Haissaguerre M.,
- Shah A.J.,
- Cochet H.,
- et al.
- Jadidi A.S.,
- Lehrmann H.,
- Keyl C.,
- et al.
- Oakes R.S.,
- Badger T.J.,
- Kholmovski E.G.,
- et al.
- Platonov P.G.,
- Mitrofanova L.B.,
- Orshanskaya V.,
- Ho S.Y.
- Yagishita A.,
- De Oliveira S.,
- Cakulev I.,
- et al.
- Aytemir K.,
- Amasyali B.,
- Kose S.,
- et al.
- Wu J.T.,
- Long D.Y.,
- Dong J.Z.,
- et al.
- Mugnai G.,
- Chierchia G.-B.,
- De Asmundis C.,
- et al.
- Khan A.,
- Mittal S.,
- Kamath G.S.,
- Garikipati N.V.,
- Marrero D.,
- Steinberg J.S.
- Arentz T.,
- Weber R.,
- Bürkle G.,
- et al.
- Surawicz B.
- Platonov P.G.
- Fukunami M.,
- Yamada T.,
- Ohmori M.,
- et al.
- Wang Y.S.,
- Chen G.Y.,
- Li X.H.,
- Zhou X.,
- Li Y.G.
- Rivard L.,
- Hocini M.,
- Rostock T.,
- et al.
- Haissaguerre M.,
- Hocini M.,
- Denis A.,
- et al.
- Knecht S.,
- Sohal M.,
- Deisenhofer I.,
- et al.
- Bayés de Luna A.,
- Baranchuk A.,
- Martínez-Sellés M.,
- Platonov P.G.
- ↵Müller P, Makimoto H, Dietrich JW, et al. Association of left atrial low-voltage area and thromboembolic risk in patients with atrial fibrillation. Europace. In press.