The Extent of Mechanical Esophageal Deviation to Avoid Esophageal Heating During Catheter Ablation of Atrial Fibrillation

JACC: CLINICAL ELECTROPHYSIOLOGY

VOL.

ª 2017 PUBLISHED BY ELSEVIER ON BEHALF OF THE

-, NO. -, 2017

ISSN 2405-500X/$36.00

AMERICAN COLLEGE OF CARDIOLOGY FOUNDATION

http://dx.doi.org/10.1016/j.jacep.2017.03.017

The Extent of Mechanical Esophageal Deviation to Avoid Esophageal Heating During Catheter Ablation of Atrial Fibrillation Chandrasekar Palaniswamy, MD, Jacob S. Koruth, MD, Alexander J. Mittnacht, MD, Marc A. Miller, MD, Subbarao Choudry, MD, Rahul Bhardwaj, MD, Dinesh Sharma, MD, Jonathan M. Willner, MD, Sujata S. Balulad, MD, Elizabeth Verghese, MD, Georgios Syros, MD, Anurag Singh, MD, Srinivas R. Dukkipati, MD, FHRS, Vivek Y. Reddy, MD

ABSTRACT OBJECTIVES This study sought to determine the extent of lateral esophageal displacement required during mechanical esophageal deviation (MED) and to eliminate luminal esophageal temperature elevation (LETElev) during pulmonary vein (PV) isolation. BACKGROUND MED is a conceptually attractive strategy of minimizing esophageal injury while allowing uninterrupted energy delivery along the posterior left atrium during PV isolation. METHODS MED was performed using a malleable metal stylet within a plastic tube placed within the esophagus. Barium was instilled to characterize the trailing esophageal edge. For each MED attempt, the MEDEffective, defined as the distance from the trailing esophageal edge-to-ablation line, was correlated to occurrences of LETElev. RESULTS In 114 consecutive patients/221 PV pairs undergoing MED (age 62.1
11 years, 75% men, 62%/38% paroxysmal/persistent AF), esophageal stretching invariably occurred such that the esophageal edge trailed behind the plastic tube. MEDEffective distances of 0 mm to 10 mm, 10 mm to 15 mm, 15 mm to 20 mm or >20 mm were achieved in 60 (27.1%), 64 (29%), 48 (21.7%), and 49 (22.2%) attempts, respectively. Overall, LET elevation >38 C occurred in 81 of 221 (36.7%) PV pairs. The incidence of LETElev among the 4 groups was 73.3%, 35.9%, 25%, and 4.1%, respectively. MEDEffective distances were 9.1
6.5 mm and 18
7.6 mm in patients with and without LETElev, respectively (p < 0.0001). Three patients (2.6%) experienced clinically significant MED-related trauma, albeit only with a stiffer stylet. CONCLUSIONS Mechanical esophageal deviation >20 mm from the PV ablation line prevents significant esophageal heating during PV isolation, but this level of displacement was difficult to safely achieve with this off-the-shelf mechanical stylet approach. (J Am Coll Cardiol EP 2017;-:-–-) © 2017 Published by Elsevier on behalf of the American College of Cardiology Foundation.

C

atheter ablation has emerged as an effective

operators, the procedure is nonetheless associated

and widely adopted treatment strategy for

with a small but significant risk of injury to collateral

patients with symptomatic atrial fibrillation

structures. Foremost among these is the esophagus,

(AF). The procedure consists of pulmonary vein (PV)

which lies in close proximity to the posterior LA

isolation alone or in combination with other lesion

and, in particular, the PV antrum, making it suscepti-

sets, predominantly in the left atrium (LA). Although

ble to ablation-related injury. Thermal injury to

it is safe overall when performed by experienced

the esophagus resulting in atrioesophageal fistula

From the Helmsley Electrophysiology Center, Icahn School of Medicine at Mount Sinai, New York, New York. Dr. Reddy has reported owning stock options in Circa Scientific, Inc, a manufacturer of esophageal temperature probes. All other authors have reported that they have relationships relevant to the contents of this paper to disclose. Manuscript received September 6, 2016; revised manuscript received February 27, 2017, accepted March 5, 2017.

2

Palaniswamy et al.

JACC: CLINICAL ELECTROPHYSIOLOGY VOL.

-, NO. -, 2017 - 2017:-–-

Extent of Esophageal Deviation to Avoid Esophageal Heating

ABBREVIATIONS

remains one of the most feared complications

MED such that the trailing esophageal edge may yet

AND ACRONYMS

of AF ablation. Despite increased awareness

be in proximity to the thermal wave front emanating

in the electrophysiology and cardiology com-

from the point of RF energy delivery. We sought to

munities of the signs and symptoms associ-

determine the extent of effective lateral esophageal

ated

atrioesophageal

displacement during MED (MEDEffective, defined as

fistula, and despite prompt diagnosis and

the distance from the trailing esophageal edge-to-

treatment of the condition, the mortality of

ablation line), and correlate this with elevations of

AF = atrial fibrillation LA = left atrium LET = luminal esophageal temperature

MED = mechanical esophageal deviation

PV = pulmonary vein RF = radiofrequency

with

post-ablation

atrioesophageal

fistula

remains

55%

(1).

the esophageal temperature.

Beyond the fatal, but relatively rare, complication of atrioesophageal fistula, thermal damage to the vagus nerve plexus on the outside surface of the esophagus is thought to also lead to esophageal and gastric dysmotility issues such as gastroparesis (2). Recent data have emerged that these dysmotility issues are more common that previously appreciated, affecting 17% of patients after AF ablation (3). There is no universally accepted approach for minimizing thermal injury to the esophagus. Many physicians use a strategy of monitoring the luminal esophageal temperature (LET) to identify esophageal heating during energy delivery. Radiofrequency (RF) energy is typically interrupted for an LET of 38 to 39  ; however, this approach negatively affects procedural workflow because cessation of energy delivery (upon esophageal heating) must be followed by a waiting period before the esophageal temperature returns to baseline to allow subsequent lesions. Other strategies to minimize esophageal injury include:

METHODS PATIENT

POPULATION. Consecutive

patients un-

dergoing AF ablation procedures using an approved RF ablation catheter from November 2014 to October 2015 at our institution were studied. The study was approved by the Institutional Review Board of Mount Sinai Hospital. Patients were excluded if they had a prior history of severe esophagitis or ulcers, strictures, or esophageal surgery. Patients undergoing repeat ablation procedures where the PVs were persistently isolated from prior procedures were also excluded. All patients underwent ablation under general anesthesia with a strategy of uninterrupted oral anticoagulation with either warfarin or a nonwarfarin

oral

anticoagulant.

Double

transseptal

punctures were performed after intravenous unfractionated heparin was administered to maintain an activated clotting time of 350 to 400 seconds.

delivery of lower energy (usually #25 W), lesions of

MECHANICAL ESOPHAGEAL DEVIATION TECHNIQUE. MED

shorter duration, and planning more medial or lateral

was performed after approval by the Mount Sinai

ablation sets on the posterior wall to minimize

Hospital “Novel Procedures Oversight Committee”;

esophageal heating. However, these approaches may

in addition to the standard consent for AF ablation,

also negatively affect the long-term PV isolation rates

all patients were separately consented for MED. MED

and clinical success of the ablation procedure.

was performed after transseptal puncture and before

Furthermore, none of these techniques completely

creation of LA geometry with the electroanatomic

avoids esophageal heating leading to atrioesophageal

mapping system. This was performed by anesthesi-

fistula and dysmotility. Esophageal cooling using a

ologists or electrophysiologists with prior experience

cooled water-irrigated esophageal balloon has also

in MED during our earlier study (4). As previously

been described as a strategy to minimize thermal

described, a standard orogastric tube was inserted.

injury. Real-time visualization of the esophagus with

The tip was positioned at the distal end of the

intracardiac echocardiography, especially with the

esophagus, and 20 to 30 ml of oral barium sulfate

probe deployed in the LA, may also be used as a

contrast (Liquid E-Z-Paque, E-Z-EM Canada Inc.,

monitoring strategy to reduce esophageal injury.

Lake Success, New York) was injected to allow the

Alternatively, we had previously reported the

contrast to fill the mid and distal esophagus. Instil-

feasibility of mechanical esophageal deviation (MED)

lation of barium helped accurately characterize the

using an off-the-shelf malleable metal stylet deliv-

trailing edge of the esophagus. Next, a 32 Fr flexible

ered within a plastic tube to deviate the esophagus

polyvinyl chloride thoracic catheter (e.g., Atrium

during AF ablation as a means to completely avoid

Medical Corporation, Hudson, New Hampshire) was

esophageal injury (4). MED is an attractive strategy

inserted into the esophagus. The tip was positioned a

because esophageal injury could potentially be

few centimeters below the level of the lower PVs. A

completely avoided while nonetheless allowing un-

preshaped malleable metal stylet was inserted into

interrupted energy delivery along the posterior LA.

the lumen of the thoracic catheter to create a curve

Significant esophageal stretching may occur during

in the distal half. In this series, there were 2 versions

JACC: CLINICAL ELECTROPHYSIOLOGY VOL.

-, NO. -, 2017

Palaniswamy et al.

- 2017:-–-

Extent of Esophageal Deviation to Avoid Esophageal Heating

F I G U R E 1 Measurement of Effective Lateral Deviation of the Esophagus

(A) Fluoroscopy. (B) Electroanatomic map. Measurement of the distance between the trailing esophageal edge (blue dots) and the ablation lesion set with mechanical esophageal deviation. ABL ¼ ablation; AL ¼ pulmonary vein isolation ablation line; CS ¼ coronary sinus; TE ¼ trailing edge; TP ¼ esophageal temperature-monitoring probe; TS ¼ flexible thoracic tube in esophagus with stylet.

of metal stylets used: most commonly, a 14 Fr

Englewood, Colorado). The temperature probe was

aluminum intubation stylet (stylet 1), and infre-

manipulated to ensure that it was positioned lateral

quently, a somewhat stiffer stylet made of stainless

to the thoracic tube at the trailing edge (Figure 1A). RF

steel, the Gliderite rigid stylet (Verathon Inc.,

energy delivery on the posterior wall was terminated

Bothell, Washington) that is designed to work with

if LET exceeded 38 C; this was defined as an instance

Glidescope video laryngoscopes (stylet 2). With

of LET elevation (LET Elev ). The peak LET was recor-

either stylet, the proximal end was manipulated

ded for all temperature rises. Using fluoroscopy in the

(applying

to

anteroposterior view, reference points were placed

laterally displace the esophagus. Initially, stylet 1

on the electroanatomical map corresponding to the

was employed for deviation, but if the level of lateral

trailing esophageal edge (Figure 1B). The extent of

excursion was insufficient, the stiffer stylet 2 was

MED Effective was defined as the distance between the

used. However, because of trauma to the oropharynx

PV isolation ablation line and the esophageal refer-

(see Results), this stiffer stylet was only employed in

ence points corresponding to the trailing esophageal

4 patients and thereafter abandoned.

edge. Measurements were made using the tools

clockwise/counterclockwise

torque)

available on each of the mapping systems employed. METHODOLOGY

FOR

ASSESSING

ESOPHAGEAL

For patients in whom the esophageal position at

DEVIATION. After MED, the LA geometry was created

baseline was midline in reference to the PV antra,

using an electroanatomic mapping system and PV

MED was performed to the side contralateral to the

isolation was performed. MED can lead to distortion

PV isolation set (for example, rightward deviation

of the anatomy and affect the accuracy of the elec-

during ablation of the left PVs; Figure 2A). The

troanatomic map. MED was therefore maintained at

esophagus was then deviated to the contralateral

the same level throughout the ablation along a PV

side; if there was a significant displacement of the

pair to minimize these effects. After ipsilateral PV

atrial anatomy, the LA geometry was created again

isolation with MED, repeat MED was performed to

before ablation of the contralateral PVs (Figure 2B). In

enable

electro-

some instances where the baseline esophageal posi-

anatomic map was always checked for accuracy and

tion was not midline and substantially toward one of

reconstructed if necessary in case of map shifts. LET

the PV antra, further extreme deviation was per-

was monitored using the S-CATH esophageal tem-

formed to the same side to allow isolation of

perature monitoring probe (Circa Scientific Inc.,

both PV pairs in that single deviated position

contralateral

PV

isolation.

The

3

4

Palaniswamy et al.

JACC: CLINICAL ELECTROPHYSIOLOGY VOL.

-, NO. -, 2017 - 2017:-–-

Extent of Esophageal Deviation to Avoid Esophageal Heating

F I G U R E 2 Mechanical Esophageal Deviation to the Side Contralateral to Ablation

(A) Ablation catheter along the left pulmonary vein isolation set, esophagus deviated to right. (B) Ablation catheter along right pulmonary vein isolation set, esophagus deviated to left. SP ¼ spiral mapping catheter. Abbreviations as in Figure 1.

(Figure 3A and 3B). The ease of placement of the de-

as MED distances, the paired Student t test was used.

vice and deviation as reported by the operator were

We used a 2-sided p value <0.05 to assess for statis-

recorded on a Likert scale between 1 and 5, with

tical significance for all analyses. For the discrete

1 being easy and 5 being extremely difficult. Throat

outcomes that could not be assumed to be indepen-

discomfort and dysphagia post-MED as reported by

dent measurements (i.e., left/right readings per pa-

the patient were also recorded on a scale from 1 to 5.

tient), such as the incidence of total temperature rises

Any complications such as oral or pharyngeal

and MED >20 mm, generalized estimating equations

bleeding, reintubation, fever, upper gastrointestinal

(PROC GENMOD, SAS) was used. This linear regres-

bleed, or any other unexpected events were recorded.

sion technique corrects for the correlation among observations on the same individual by estimating

PULMONARY VEIN ISOLATION. Catheter mapping

the covariance structure of the data, thereby allowing

and manipulation was guided by either the Carto 3

for improved estimates of the standard errors of

(Biosense-Webster Inc., Diamond Bar, California) or

measurement.

NAVEX (St. Jude Medical Inc., Minneapolis, Minnesota) electroanatomical mapping system. Catheter ablation was performed using a force-sensing irri-

RESULTS

gated RF ablation catheter: either the Thermocool Smartouch (Biosense-Webster Inc.) or Tacticath (St.

PATIENT CHARACTERISTICS. Baseline characteris-

Jude Medical Inc.) ablation catheter. The power used

tics of our study population are summarized in

during ablation along the posterior left atrium was

Table 1. The population included 114 patients (221 PV

not different than that used at other locations: typi-

pairs: 112 right PVs and 109 left PVs) with paroxysmal

cally, 35 to 40 W with the Smartouch catheter and

AF (62%) or persistent AF (38%). The mean age was

25 to 30 W with the Tacticath catheter.

62.1
11 years, and 75% of patients were men. The median congenital heart failure, hypertension, age

STATISTICS. Statistical analysis was performed using

>75 years, diabetes mellitus, stroke, vascular disease,

SAS system software, version 9.3 (SAS Institute Inc.,

age 65 to 74 years, and sex score was 2. The ablation

Cary, North Carolina). Categorical variables are

procedure was performed with uninterrupted anti-

expressed as percentage and continuous variables as

coagulation in all patients: 20 (17.5%) with uninter-

mean
standard deviation. Categorical variables

rupted warfarin and the remaining 94 (82.5%) with

were analyzed using the chi-square or Fisher exact

uninterrupted

test; for continuous variables within a patient, such

Antiplatelet therapy was being administered in

non-warfarin

oral

anticoagulants.

JACC: CLINICAL ELECTROPHYSIOLOGY VOL.

-, NO. -, 2017

Palaniswamy et al.

- 2017:-–-

Extent of Esophageal Deviation to Avoid Esophageal Heating

F I G U R E 3 Extreme Mechanical Esophageal Deviation to 1 Side

(A) Ablation catheter along the left pulmonary vein isolation set, esophagus deviated to left. (B) Ablation catheter along right pulmonary vein isolation set, esophagus deviated to left. Abbreviations as in Figure 1.

19 patients (16.7%). The procedure was the first-ever

been appreciated without the use of either barium

AF ablation procedure in 103 patients (90.4%) and a

instillation to opacify the esophageal lumen or the

redo procedure in the remaining 11 patients (9.6%).

careful placement of the esophageal temperature

PROCEDURAL CHARACTERISTICS. In Table 2, the

procedural characteristics of the study cohort are described. Force-sensing irrigated ablation catheters

probe along the trailing edge. T A B L E 1 Demographics and Baseline Patient Characteristics

Characteristic

were used in all procedures; approximately twothirds of the procedures were performed using the Carto 3 system and the remaining one-third using the NavX system. PV isolation alone or in combination with cavotricuspid isthmus ablation was performed in

Value (n ¼ 114)

Age (yrs) (mean
SD)

62.1
11

Men (%)

85 (74.6)

Type of AF Paroxysmal

71 (62.3)

Persistent

43 (37.7)

94.7% of patients. Additional linear lesions were

CHA2DS2-VASc score (median)

placed in only 5.3% of patients. The mean fluoroscopy

Hypertension

time was 21.6
10.4 min and fluoroscopy dosage

Diabetes mellitus

15 (13.2)

(air kerma) was 83
48 mGy.

Prior TIA or CVA

15 (13.2)

2 60 (53.4)

Vascular disease

24 (21.1)

MECHANICAL ESOPHAGEAL DEVIATION. Of the 114

Heart failure or EF <35%

21 (18.4)

patients, 7 (6.1%) had only 1 PV pair ablated, hence

Anticoagulant used

MED was performed on 1 side only. Of the remaining

Warfarin

20 (17.5)

107 patients, 96 underwent esophageal deviation to

Rivaroxaban

67 (58.8)

Apixaban

25 (21.9)

both sides, whereas 11 (10.3%) underwent extreme deviation to 1 side only. Of these 11 patients, 9 had an

Dabigatran

2 (1.8)

Antiplatelet use

19 (16.7)

initial esophageal position that was leftward; the

Proton pump inhibitor use

19 (16.7)

esophagus was further deviated to the extreme left

Failed class I/III antiarrhythmic drugs

79 (69.3)

side, and ablation of both PV pairs was performed in

Repeat ablation procedure

11 (9.6)

this position. The 1 remaining patients underwent extreme deviation to the right side, and both PV isolation lesion sets were placed. Qualitatively, it was invariably true that MED was associated with a significant trailing esophageal edge that would not have

Data presented as mean
standard deviation (SD) or n (%). AF ¼ atrial fibrillation; CHA2DS2-VASc ¼ congenital heart failure, hypertension, age >75 years, diabetes mellitus, stroke, vascular disease, age 65 to 74 years, and sex; CVA ¼ cerebrovascular accident; EF ¼ ejection fraction; TIA ¼ transient ischemic attack.

5

6

Palaniswamy et al.

JACC: CLINICAL ELECTROPHYSIOLOGY VOL.

-, NO. -, 2017 - 2017:-–-

Extent of Esophageal Deviation to Avoid Esophageal Heating

T A B L E 2 Procedural Characteristics

T A B L E 4 Subjective Assessment and Complications of

Value (n ¼ 114)

Parameter

Mechanical Esophageal Deviation Characteristic

Type of electroanatomic mapping system Carto 3

78 (68.4)

NavX

36 (31.6)

Type of ablation catheter used Smartouch

78 (68.4)

Tacticath

36 (31.6)

Ease of placing the device*

1.4
0.6 (1)

Ease of deviation*

1.9
1.1 (2)

Throat pain or discomfort†

1.3
0.7 (1)

Dysphagia†

1.2
0.6 (1)

Complications

Ablation lesions performed

Value (n ¼ 114)

Subjective measures of the MED procedure

3 (2.6%)

PVI only

25 (21.9)

Oral/pharyngeal bleeding

1

PVI plus CTI ablation

83 (72.8)

Uvular hematoma

2

PVI, CTI, and additional ablation lesions

6 (5.3)

Total ablation time (min, mean
SD)

103.8
33.7

Fluoroscopy time (min, mean
SD)

21.6
10.4

Fluoroscopy dosage (mGy, mean
SD)

83.1
48

*Graded on scale of 1–5, with 1 being easy and 5 being extremely difficult (value expressed as mean
SD, median in parentheses). †Graded on scale of 1–5, with 1 being no discomfort and 5 extreme discomfort (value expressed mean
SD, median in parentheses). Abbreviation as in Table 3.

Data presented as mean
standard deviation or n (%). CTI ¼ cavotricuspid isthmus; PVI ¼ pulmonary vein isolation; total ablation time ¼ cumulative radiofrequency ablation time during each procedure.

mean MEDEffective distances were 9.1
6.5 mm and 18
7.6 mm in patients with and without LET Elev ,

EXTENT OF MED E f f e c t i v e . Based on the extent of

MED Effective as measured by the distance between the trailing esophageal edge and ablation lesion set, the instances of MED were stratified into 4 groups (Table 3). The extent of MEDEffective was 0 mm to 10 mm, 10 mm to 15 mm, 15 mm to 20 mm, or >20 mm in 60 (27.1%), 64 (29%), 48 (21.7%), and 49 (22.2%) instances of MED, respectively. Not surprisingly, the

respectively (p < 0.0001). Instances of LET Elev among the 4 groups occurred in 73.3%, 35.9%, 25%, and 4.1% instances, respectively. Thus, a MED Effective >20 mm largely eliminated LET Elev; however, this was achieved in only 22.2% of patients, and was greater during ablation of the right PVs with leftward deviation than during ablation of the left PVs with rightward deviation (32.1% vs. 11.9%, p ¼ 0.0006). Indeed, the average extent of MED Effective achieved was greater for leftward deviation than for rightward deviation (17.4
8.3 vs. 12.2
7.3, p < 0.0001).

T A B L E 3 Extent of Mechanical Esophageal Deviation and

Incidence of Luminal Esophageal Temperature Elevation (>38  C) Characteristic

PV pairs

Total

Right PV

Left PV

221

112

109

Extent of MED (per subgroup)

Similarly, although instances of LET Elev occurred during ablation of 81/221(36.7%) PV pairs, there was a marked variation in the incidence of LET Elev between the 2 directions of deviation: LET Elev occurred during 25% cases of leftward deviation for right PVs ablation,

0–10 mm

60 (27.1)

17 (15.2) 43 (39.4)

10.1–15 mm

64 (29.0)

33 (29.5) 31 (28.4)

15.1–20 mm

48 (21.7)

26 (23.3) 22 (20.2)

49 (22.2)

36 (32.1) 13 (11.9)

>20 mm 

Incidence of LET elevations >38 C Overall cohort

81/221 (36.7) 28 (25.0) 53 (48.6)

but approximately doubled to 48.6% of cases during rightward deviation for left PVs ablation (Table 3). In total, there were 316 ablation lesions that resulted in LET Elev during the 221 PV isolating lesion sets, translating to 1.4 ablations with LET Elev per ipsilateral lesion set. Again, the incidence of LETElev/patient was

Per subgroup 0–10 mm

44/60 (73.3) 14 (12.5) 30 (27.5)

higher for left PVs ablation with rightward deviation

10.1–15 mm

23/64 (35.9) 9 (8.0)

14 (12.9)

than right PVs ablation with leftward deviation (2.0 vs. 0.9 ablations/patient, respectively; p ¼ 0.01).

15.1–20 mm

12/48 (25)

5 (4.5)

7 (6.4)

>20 mm

2/49 (4.1)

0 (0)

2 (1.0)

316

100

216

226

60

166

10.1–15 mm

55

19

36

15.1–20 mm

33

21

12

SUBJECTIVE ASSESSMENT AND COMPLICATIONS. As

No. of ablation lesions with LETElev Overall cohort Per subgroup 0–10 mm

>20 mm

summarized in Table 4, from a subjective perspective, the MED device was easy to place in the esophagus

2

0

2

(mean score of 1.4 on Likert scale), but the actual deviation step was occasionally more difficult (mean score of 1.9 on Likert scale). On average, the patients experienced minimal throat pain or discomfort or

Data presented as n (%). LETElev ¼ instances of luminal esophageal elevation >38 C; MED ¼ mechanical esophageal deviation; PV ¼ pulmonary vein.

dysphagia (mean scores of 1.3 and 1.2, respectively). Three patients did develop clinically significant oropharyngeal discomfort as a result of MED-related

JACC: CLINICAL ELECTROPHYSIOLOGY VOL.

-, NO. -, 2017

Palaniswamy et al.

- 2017:-–-

Extent of Esophageal Deviation to Avoid Esophageal Heating

trauma. But in each of these cases, the stiffer stylet

inability to accurately visualize esophageal borders or

2 had been used and the trauma occurred not in the

because of design limitations of the LET probe itself

esophagus (all of these patients underwent endos-

(e.g., linear probes with single thermocouples have

copy), but rather the oropharynx. One patient experi-

narrow sensing capability).

enced bleeding from a pharyngeal laceration that

The LET probe in this study was chosen to address

required cauterization. Two other patients sustained

these limitations; specifically, by virtue of its width

uvular hematoma that resolved spontaneously. All

and multiple thermocouples, this probe is able to

patients were followed clinically and none experi-

interrogate a much larger extent of the esophagus. In

enced any long-term sequelae. No cases of fever, upper

addition, we used real-time imaging of the esophagus

gastrointestinal bleed, or any other unexpected events

with barium contrast to maximize the approximation

were noted. With use of the more malleable stylet 1,

of the LET probe with the borders of the esophagus.

there was no evidence of symptomatic MED-related

When combined with MED, this approach further im-

trauma. There were also no instances of atrioesopha-

proves the sensitivity of LET rises as a marker of

geal fistula or symptomatic gastric dysmotility.

esophageal injury. Although we did not perform routine endoscopy, the lack of LET rises in the setting

DISCUSSION

of optimal MED provides the physician confidence that the esophagus is not being heated during uninter-

In this study, we present the feasibility of mechanical

rupted posterior LA ablation. It is also of significance

esophageal deviation using an off-the-shelf malleable

that power titration along the posterior LA was not

metal stylet during AF ablation in a large consecutive

modified; that is, the same amount of energy used

patient cohort, with specific attention to the relation-

anteriorly was also used posteriorly (except in in-

ship between the extent of esophageal excursion and

stances of esophageal heating, in which case the power

LET rises (4). The major findings of this manuscript are

was down-titrated).

that 1) lateral displacement of the esophagus is feasible

In this report, we also defined the concept of the

in a diverse cohort of patients undergoing AF ablation,

trailing edge, its importance, and the extent of devi-

2) it is necessary to deviate the esophagus $20 mm

ation necessary to avoid LET rises. Our experience

from the site of ablation to optimally eliminate

clearly demonstrated that esophageal stretching oc-

esophageal heating, and 3) this optimal level of

curs during MED, therefore highlighting the impor-

esophageal deviation can be achieved in only a mi-

tance of identifying the trailing edge with either

nority of patients using this off-the-shelf stylet.

barium or the LET probe (provided the probe is truly

In a nonrandomized study, we had previously

placed at the trailing edge). If the trailing edge is not

demonstrated that posterior LA ablation guided

identified and the operator were to assume that the

by

esophagus was sufficiently deviated based on the

LET

monitoring

(interruption

of

RF

for

temperature $38.5  C) was associated with a lower

location the deviation tool along the “leading edge,”

incidence of esophageal injury (6% vs 36%, p < 0.006)

inadvertent delivery of RF energy to the esophagus

(5). This study was based on the concept that the LET

may occur.

probes, by virtue of their intraluminal location, can

Finally, our results indicate that achieving devia-

predict esophageal thermal injury by detecting LET

tion of 20 mm or more results in infrequent LET rises.

rises, providing operators with an opportunity to

This is consistent with a previous study of 106

interrupt or modify RF applications. Despite the

consecutive AF ablation patients in which the authors

positive outcome of this and other similar studies and

demonstrated that LET elevation was more frequent

the widespread adoption of LET monitoring by many

with a lesser degree of separation between the

operators, LET use does not universally prevent

esophagus and ablation sites along the posterior LA.

esophageal injury (6). This was recognized in several

They also described that at separations $20 mm, LET

reports of esophageal injury despite LET monitoring.

rises (>38  C) never occurred (8). Our results are

Several potential explanations can explain this

consistent with the results of this study. On the other

paradox: 1) suboptimal selection of LET cutoffs that

hand, in our study, this ideal degree esophageal de-

trigger RF interruption and 2) failure of the LET probe

viation was achievable in only 22% of patients. This

to sufficiently interrogate that lateralized portion of

appears to be driven in part by the limitations of the

esophagus that is in closest proximity to the thermal

technique used for deviation in this report in that the

source with the consequence of underdetection of

tool itself typically used, stylet 1, was unable to ach-

thermal injury (7). The latter can occur either because

ieve sufficient lateral displacement of the esophagus

the operator underestimates the proximity of the

in all patients. Interestingly, it was our qualitative

thermal source to the esophagus because of the

impression that a greater extent of lateral excursion

7

8

Palaniswamy et al.

JACC: CLINICAL ELECTROPHYSIOLOGY VOL.

-, NO. -, 2017 - 2017:-–-

Extent of Esophageal Deviation to Avoid Esophageal Heating

was achieved with the stiffer stylet 2; however, this

tools are needed; as such, the MED used in this study

stylet resulted in oropharyngeal trauma; indeed, 3 of

still needs further evaluation before being considered

the 4 patients in whom this was used experienced

for routine clinical use. In the future, such tools may

clinically significant trauma. Thus, there was a trade-

allow for greater and more consistent deviation, ease

off between efficacy and safety: the more malleable

of use, and a favorable safety profile. Although LET

stylet 1 did not result in clinically significant oropha-

elevations may indicate lack of sufficient lateral de-

ryngeal trauma, but was also a less effective deviation

viation, it is also possible that the MED itself displaces

tool.

the esophagus closer to the posterior LA, making LET

With regard to safety, we did not perform routine

rises more frequent when deviation was suboptimal.

endoscopy in all patients. Accordingly, it is not possible to reliably comment on the correlation of

STUDY LIMITATIONS. Although we did not have a

LET Elev , or lack thereof, and esophageal necrosis. In

comparative randomized cohort without MED, it was

our prior published experience with MED

(4),

clear that there was an immediate improvement in

complete endoscopy of the esophagus, stomach, and

workflow. That is, in previous cases when MED had

duodenum was performed routinely on all patients

not been used, it was invariably true that ablation le-

post-procedure.

that

sions were truncated before adequate lesion could be

although MED-related trauma did occur in 60% of

made because of esophageal temperature elevations;

subjects (as was expected), it did not result in any

this appeared to occur with much less frequency with

clinical sequelae. Based on these findings and our

MED. Of course, the most obvious safety advantage of

combined clinical experience with this technique, we

esophageal deviation is the potential for avoiding

no longer routinely perform post-MED endoscopy.

atrioesophageal fistula formation. Although it seems

The

study

demonstrated

Although endoscopy or imaging modalities such as

self-evident that fistulas would be avoided if the

computed tomography or magnetic resonance imag-

esophagus is diverted from the point of thermal abla-

ing may have provided additional information about

tion, the relative rarity of fistula formation will make it

subclinical mechanical esophageal injury, this was not

virtually impossible to definitively prove that MED can

performed in our study and is an important limitation

prevent this often fatal complication. In addition to

from a safety perspective. Finally, it is notable that all

thermal injury, other mechanisms such as ischemic

of the procedures in our experience were performed in

injury to the esophagus may also contribute to

the setting of uninterrupted oral anticoagulation with

atrioesophageal fistula formation. On the other hand,

either warfarin or, more frequently, a non-warfarin

it is interesting that in our series of >100 patients, we

oral anticoagulant. This highlights the safety of this

observed no instances of clinically evident gastric

mechanical deviation strategy. Indeed, this is consis-

dysmotility, although this has been reported to occur

tent with a recently published large consecutive series

only in about 1% to 2% of patients after AF ablation.

of 704 patients undergoing AF ablation in whom a

It is also notable that MED allowed us to use high

transesophageal echocardiogram probe was used to

power along the posterior wall (in the absence of

mechanically deviate the esophagus (9). Although the

esophageal heating); however, because of the thin

correlation to esophageal heating was investigated in

nature of the posterior left atrial wall, ablation with

only a minority of this cohort (n ¼ 25), this experience

high power should be performed with caution.

nonetheless speaks to the safety of mechanical

Although certainly not provable from our series, this

displacement of the esophagus.

raises the possibility that the ablation time for

The effectiveness of MED differed depending on

achieving PV isolation could be reduced. It is also

the lateral direction of excursion, favoring left-ward

possible that MED might decrease the frequency of PV

deviation (for right PV ablation): 1) an MED Effective

reconnection. This is particularly intriguing because

>20 mm was achieved almost 3 times more frequently

we have recently demonstrated an association be-

for leftward deviation (32.1% vs. 11.9%), 2) the average

tween esophageal heating and the subsequent devel-

extent of MEDEffective greater for leftward deviation

opment of PV reconnection (10). In a series of 142

(17.4
8.3 mm vs. 12.2
7.3 mm), and most important,

patients undergoing redo ablation procedures (in

3) LET Elev occurred almost twice as frequently during

whom we had performed esophageal temperature

rightward deviation (48.6% vs. 25%). The reasons for

monitoring during the initial procedure), PV recon-

this can be speculated to be related to thor-

nections occurred in 20% of the PV pairs; interest-

acoabdominal anatomical variations between patients

ingly, the point of reconnection for the majority (83%)

or could be technique specific. This suggests that,

of the lesion sets was along the posterior wall.

although the extent of deviation that can be achieved

Furthermore, there was a positive correlation between

may be limited by certain patient factors, dedicated

the occurrence of LETElev during the index ablation

JACC: CLINICAL ELECTROPHYSIOLOGY VOL.

-, NO. -, 2017

Palaniswamy et al.

- 2017:-–-

Extent of Esophageal Deviation to Avoid Esophageal Heating

procedure and the point of PV reconnection along the inferior aspects of the posterior ablation lines. Further studies will be required to determine whether esophageal deviation will decrease the presence of PV reconnections along the posterior wall.

PERSPECTIVES COMPETENCY IN MEDICAL KNOWLEDGE: Mechanical esophageal deviation is an attractive strategy to minimize esophageal heating during atrial fibrillation ablation. This has the potential to minimize thermal injury to the esophagus and

CONCLUSIONS

improve procedural work flow by reducing the frequency of

We demonstrate the ability of mechanical esophageal deviation during posterior LA ablation to achieve proportional reduction of LET rises with incremental degrees of esophageal deviation. Although esophageal deviation as performed in this study was able to achieve deviation, dedicated tools are required to consistently and safely achieve effective levels of esophageal deviation.

premature termination of ablation resulting from esophageal heating. In a study of 114 consecutive patients undergoing atrial fibrillation ablation, deviation performed using off-the-shelf mechanical stylet approach was demonstrated to be feasible and able to incrementally reduce esophageal heating depending on extent of lateral deviation. TRANSLATIONAL OUTLOOK: Mechanical esophageal deviation to a distance exceeding 20 mm from the ablation line was

ADDRESS FOR CORRESPONDENCE: Dr. Vivek Y.

Reddy, Helmsley Electrophysiology Center, Icahn School of Medicine at Mount Sinai, One Gustave L. Levy Place, Box 1030, New York, New York 10029.

optimal, whereas a distance <10 mm was largely ineffective. Achieving optimal deviation was difficult with this current technique. Better tools to achieve greater and more consistent esophageal deviation should be developed and evaluated.

E-mail: [email protected].

REFERENCES 1. Singh SM, d’Avila A, Singh SK, et al. Clinical outcomes after repair of left atrial esophageal fistulas occurring after atrial fibrillation ablation procedures. Heart Rhythm 2013;10:1591–7.

5. Singh SM, d’Avila A, Doshi SK, et al. Esophageal injury and temperature monitoring during atrial fibrillation ablation. Circ Arrhythm Electrophysiol 2008;1:162–8.

2. Shah D, Dumonceau JM, Burri H, et al. Acute pyloric spasm and gastric hypomotility: an extracardiac adverse effect of percutaneous radiofrequency ablation for atrial fibrillation. J Am Coll

6. Leite LR, Santos SN, Maia H, et al. Luminal

Cardiol 2005;46:327–30. 3. Knopp H, Halm U, Lamberts R, et al. Incidental and ablation-induced findings during upper gastrointestinal endoscopy in patients after ablation of atrial fibrillation: a retrospective study of 425 patients. Heart Rhythm 2014;11:574–8.

esophageal temperature monitoring with a deflectable esophageal temperature probe and intracardiac echocardiography may reduce esophageal injury during atrial fibrillation ablation procedures: results of a pilot study. Circ Arrhythm Electrophysiol 2011;4:149–56.

4. Koruth JS, Reddy VY, Miller MA, et al. Mechanical esophageal displacement during catheter

7. Knecht S, Sticherling C, Reichlin T, et al. Reliability of luminal oesophageal temperature monitoring during radiofrequency ablation of atrial fibrillation: insights from probe visualization and oesophageal reconstruction using

ablation for atrial fibrillation. J Cardiovasc Electrophysiol 2012;23:147–54.

magnetic resonance imaging. pii:euw129 [Epub ahead of print].

8. Maenosono R, Oketani N, Ishida S, et al. Effectiveness of esophagus detection by threedimensional electroanatomical mapping to avoid esophageal injury during ablation of atrial fibrillation. J Cardiol 2012;60:119–25. 9. Mateos JC, Mateos EI, Peña TG, et al. Simplified method for esophagus protection during radiofrequency catheter ablation of atrial fibrillation— prospective study of 704 cases. Rev Bras Cir Cardiovasc 2015;30:139–47. 10. Tran VN, Kusa S, Smietana J, et al. The relationship between esophageal heating during left atrial posterior wall ablation and the durability of pulmonary vein isolation. Europace 2017. http://dx.doi.org/ 10.1093/europace/euw232. [Epub ahead of print].

KEY WORDS esophageal deviation, esophageal protection, stylet

9