Author + information
- Received August 1, 2016
- Revision received January 17, 2017
- Accepted January 20, 2017
- Published online June 28, 2017.
- Andreas Rillig, MDa,
- Boris Schmidt, MDa,
- Luigi Di Biase, MD, PhDb,
- Tina Lin, MBBS, BMedScia,
- Leonie Scholz, MDa,
- Christian H. Heeger, MDa,
- Andreas Metzner, MDa,
- Daniel Steven, MDc,
- Peter Wohlmuth, PhDd,
- Stephan Willems, MDc,
- Chintan Trivedi, MD, MPHb,
- Joseph G. Galllinghouse, MDb,
- Andrea Natale, MDb,
- Feifan Ouyang, MDa,
- Karl-Heinz Kuck, MDa and
- Roland Richard Tilz, MDe,∗ ()
- aDepartment of Cardiology, Asklepios Klinik St. Georg, Hamburg, Germany
- bTexas Cardiac Arrhythmia Institute at St. David’s Medical Center, Austin, Texas
- cUniversitäres Herzzentrum Hamburg, Hamburg, Germany
- dAsklepios Proresearch, Asklepios Klinik St. Georg, Hamburg, Germany
- eMedical Clinic II, University Heart Center Lübeck, Lübeck, Germany
- ↵∗Address for correspondence:
Dr. Roland Richard Tilz, II. Med. Abteilung, Asklepios Klinik St. Georg, Lohmühlenstrasse 5, 20099 Hamburg, Germany.
Objectives Circumferential pulmonary vein isolation (CPVI) using irrigated radiofrequency is the most frequently used ablation technique for the treatment of atrial fibrillation worldwide.
Background To date, no large randomized multicenter trials have evaluated the efficacy and safety of CPVI using robotic navigation (RN) systems compared with the current gold standard of manual ablation (MN).
Methods In this prospective, international multicenter noninferiority trial, 258 patients with paroxysmal or persistent atrial fibrillation were randomized for CPVI using either RN (RN group, n = 131) or manual ablation (MN group, n = 127). In all patients, CPVI was performed using irrigated radiofrequency ablation in combination with a 3-dimensional mapping system. The primary endpoint was the absence of atrial arrhythmia recurrence on or off antiarrhythmic drugs during a 12-month follow-up period. Secondary endpoints were the evaluation of periprocedural complications and procedural data such as procedure time, fluoroscopy time, and incidence of esophageal injury.
Results Baseline characteristics were comparable between the RN group and MN group. Procedure time was significantly shorter in the MN group (129.3 ± 43.1 min vs. 140.9 ± 36.5 min; p = 0.026). 247 patients completed the 12-month follow-up (RN group, n = 123; MN group, n = 124). Recurrence rate was comparable between the RN and MN groups (n = 29 of 123 [23.6%] vs. 25 of 124 [20.2%]). The incidence of procedure-related major complications did not differ significantly between ablation arms (RN group, n = 8 [6.1%] vs. MN group, n = 6 [4.7%]; p = 0.62). One patient from the RN group developed a fatal atrioesophageal fistula.
Conclusions This study demonstrated that robotic ablation is noninferior to the current gold standard of manual ablation for CPVI with respect to success and complication rates. Procedure times were significantly longer in the RN group. (Alster Man and Machine: Comparison of Manual and Mechanical Remote Robotic Catheter Ablation for Drug-Refractory Atrial Fibrillation; NCT00982475)
Pulmonary vein isolation (PVI) is now established as the standard ablation approach for patients with paroxysmal and persistent atrial fibrillation (AF) (1,2). In patients with paroxysmal AF, success rates of up to 80% after multiple procedures using catheter ablation with radiofrequency current (RFC) have been reported during long-term follow-up (FU) in experienced centers (1). The major determinant for AF recurrence is still pulmonary vein reconduction (3).
The remote robotic navigation system (RNS) (Sensei, Hansen Medical, Mountain View, California) was developed to perform catheter ablation within the left atrium (LA) with enhanced catheter stability and more precise catheter navigation. Optimized catheter-to-tissue contact might lead to more effective lesions. On the other hand, it can lead to a potentially higher risk of LA wall damage, LA perforation and cardiac tamponade, or to an increased incidence of esophageal injury (4,5).
To date, only 1 randomized single-center study exists comparing the success rate of robotically navigated PVI and manually performed ablation (6). In this previous study, the patient number was limited and the study was terminated early after an interim analysis. Furthermore, the operators had limited experience with the RNS, thus the learning curve may have influenced the results significantly (6).
The Man and Machine trial is the first randomized, international, prospective multicenter controlled noninferiority trial comparing robotic navigation to the current gold standard of manually performed RF ablation for circumferential PVI with evaluation of complication rates and overall success rates on or off antiarrhythmic drugs (AADs) after a 12-month FU period as the primary endpoint.
The methods of the Man and Machine trial have been described in detail in the study protocol published previously (7).
In brief, this study was a prospective, randomized, controlled, international multicenter noninferiority trial comparing the efficacy and safety of robotically assisted PVI (RN group) with manually performed AF ablation (MN group). Patients with paroxysmal or persistent AF were included. Atrial arrhythmia lasting at least 30 s on or off AADs taking into consideration a blanking period of 3 months (i.e., within the FU period from month 3 to month 12) was chosen for the primary endpoint.
Secondary endpoints were procedural duration, fluoroscopy time and dose, procedure-related complications, and thermal esophageal injury as assessed by endoscopy.
All patients gave written informed consent. Funding was granted equally by St. Jude Medical and Hansen Medical. The study is approved by the local ethics committee. All authors had full access to the data and have read and agreed to the manuscript as written.
Eligibility and exclusion criteria
Patients aged 30 to 75 years, with symptomatic, drug-refractory paroxysmal or persistent AF defined as ≤2 months were included.
Patients with prior LA ablation or LA surgery, as well as patients with confirmed LA thrombus, left ventricular ejection fraction <50%, LA diameter >50 mm, or patients participating in another clinical trial were excluded. Furthermore, patients with contraindications to systemic anticoagulation, acute or chronic renal failure (stage ≥II), any reversible cause for AF, pregnancy, systemic infection, or other medical conditions with a life expectancy of <12 months were not accepted for participation in the trial.
Patients were recruited from 3 international high-volume ablation centers (Asklepios Klinik St. Georg, Hamburg, Germany; Universitäres Herzzentrum Hamburg, Hamburg, Germany; Texas Cardiac Arrhythmia Institute at St. David’s Medical Center, Austin, Texas). Patients were randomly assigned to 1 of the 2 treatment groups (Figure 1). All centers involved in the treatment of the study patients had performed at least 50 cases with the remote robotic navigation system before starting the study.
The sensei RN-system
The RNS has been described previously (8,9). In brief, it is an electromechanical system that facilitates catheter navigation via 2 steerable sheaths (Artisan/Lynx, Hansen Medical) incorporating an ablation catheter and is manipulated via a pull-wire mechanism. The robotic arm obeys the commands of the physicians’ workstation. The system continuously monitors contact force exerted by the catheter tip using the IntelliSense (Hansen Medical) software (7).
Electrophysiological studies were performed under deep sedation using fentanyl, midazolam, and a continuous infusion of propofol.
In the MN group, double transseptal punctures were performed in all patients, whereas in the RN group, a single or double transseptal puncture was performed at the discretion of the operator. The Lynx catheter or the Artisan sheath carrying the ablation catheter was used for ablation in the RN group. In both groups, an irrigated tip ablation catheter (Cool Path Duo, St. Jude Medical [St. Paul, Minnesota], or Celsius Thermocool, Biosense Webster) was used for ablation.
Following transseptal catheterization, intravenous heparin was administered and activated clotting time was measured every 30 minutes targeting an activated clotting time level of above 250 to 300 s.
A 3-dimensional (3D) map of the LA was performed using the Ensite NavX or Ensite Velocity system (St. Jude Medical) and pulmonary vein ostia were marked on the 3D map with 3D tags using the mapping catheter, as identified by selective angiographies (see study protocol for details) (7).
In all patients, a circumferential ablation line was deployed around the ipsilateral pulmonary vein ostia, aiming at complete bidirectional conduction block confirmed with the circular mapping catheter placed within the pulmonary veins. Irrigated RFC was delivered at a target temperature of 43°C, a power limit of 30W, and an infusion rate of 17 ml/min along the posterior LA wall, and with 40W and an infusion rate of 25 ml/min in the remaining LA in the manual group. When using the RNS, RFC application was performed with a temperature limit of 43°C, power limit of 40 W at the anterior wall and 20 W at the posterior wall, and with a contact force of 20 g to 40g as assessed by the IntelliSense software.
The procedural endpoint was persistent PV isolation confirmed by a circular mapping catheter within the PVs. after a 20-min waiting period.
Esophageal temperature monitoring
The luminal esophageal temperature was monitored continuously throughout the procedure in all patients (SensiTherm, St. Jude Medical). A significant rise in the esophageal luminal temperature was defined as an endoluminal temperature rise >41°C. Whenever the esophageal temperature rose above 41°C, RFC ablation was stopped immediately. When necessary, RFC was reapplied at the same site after the temperature returned to baseline.
All patients were bridged with low-molecular-weight heparin at a therapeutic dose until an international normalized ratio of 2 to 3 was reached. Oral anticoagulation was resumed the day after the procedure. Before discharge, all patients underwent transthoracic echocardiography and chest x-ray to rule out pericardial effusion and pneumothorax. Additionally, 24-h Holter electrocardiogram and 12-lead electrocardiograms were obtained.
Oral anticoagulation was maintained for a minimum of 3 months following ablation and thereafter according to the individual’s congestive heart failure, hypertension, age, diabetes mellitus, stroke, vascular disease, and sex (CHA2DS2-VASc) score regardless of the underlying rhythm (7).
Previously ineffective AADs were recommended for a maximum of 3 months.
All patients were treated with proton pump inhibitors for 6 weeks after discharge.
Endoscopy was performed within 1 to 3 days after the ablation. Repeat endoscopy was to be performed within 1 week if ulcerations were detected at the initial endoscopy. Findings were classified as no lesion, erythema (minimal lesion with intact mucosa), ulceration, and perforation. If esophageal ulceration was seen, patients were treated with intravenous proton pump inhibitors until repeat endoscopy revealed healing of the ulceration.
After discharge, clinical FU visits consisted of medical history-taking and rhythm screening using 72-h Holter monitoring at 3, 6, and 12 months. At all of these FU intervals patients were contacted via telephone to interrogate for clinical symptoms. In-office review of patients was not routinely performed; however, it was recommended and part of normal clinical practice that all patients were seen by a medical practitioner at the FU intervals. A 72-h Holter monitor was sent to the patients by our study nurse.
Whenever symptoms suggestive of arrhythmia recurrence were reported, patients received an external event recorder for closer FU. According to the guidelines, AF recurrences within the first 3-month blanking period were documented and treated when necessary, but were not considered as a primary endpoint. Re-ablation was not generally performed, except if drug-refractory, highly symptomatic arrhythmia recurrence occurred.
Definition of clinical endpoints and complications
The definition of endpoints and complications are described in detail in the previously published study protocol (7).
The primary objective of this trial was defined as atrial tachyarrhythmia lasting at least 30 s on or off AAD therapy outside of the 3-month blanking period. Right-sided typical atrial flutter was not defined as arrhythmia recurrence.
Fluoroscopy time, radiation dose, procedure time, as well as procedure-related complications, the incidence of esophageal injury and complication rate during the FU period were assessed as secondary endpoints.
Serious adverse events
Serious adverse events were defined as events potentially resulting in life-threatening consequences or severe/permanent impairment of patient health (7).
Esophageal lesions were classified as nonserious adverse events. All serious adverse events were adjudicated by an independent data safety committee, whose members were unaware of the treatment assignments.
The determination of the sample size of this trial is explained in detail in the study protocol (7).
This investigation is an international, multicenter, prospective, noninferiority trial examining the freedom from AF recurrence after catheter ablation. A total of 258 patients were randomized to a robotic-assisted or manual ablation group in a 1:1 fashion.
The primary endpoint was the recurrence rate at the first year after procedures on or off AADs. The aim of this trial was to examine noninferiority of the robotic ablation group compared with the manual ablated patients, whereas an equivalence limit of 15% was assumed. The primary hypothesis was proven with a noninferiority test for proportion differences between the ablation groups. Only patients that completed 12-month FU were included (n = 247 evaluable patients).
Baseline data of the patients (gender, age, body mass index), anamnestic data (hypertension, diabetes mellitus, history of AF, coronary artery disease), risk scores (CHA2DS2-VASC), procedural data (fluoroscopy and procedure time, number of RF applications), and FU data (occurrence of AF, adverse events) were recorded. The data were analyzed for all recruited and for all evaluable patients in the following way.
Continuous data were described as means and standard deviations, if the variables were normally distributed, or as medians, minima, first and third quartiles, and maxima if they were not. Differences of continuous variables between the 2 ablation groups (and patients with and without AF recurrences) were analyzed with t tests, if the data were approximately normally distributed, and with Wilcoxon Mann-Whitney tests otherwise.
Categorical data were described with absolute and relative frequencies. Differences between categorical variables were evaluated with the chi-square test or with Fisher exact test in case of small expected cell frequencies. Freedom from AF recurrence was estimated with the Kaplan-Meier method and shown graphically.
All p values are 2-sided (except for the analysis of the primary objective) and p < 0.05 was considered significant. All calculations were performed with the statistical analysis software SAS (SAS Institute Inc., version 9.3, Cary, North Carolina).
Baseline characteristics of the overall group (n = 258) are displayed in Table 1. There were no differences between the baseline characteristics of the RN group and the MN group. Importantly, between-center variance was small enough to be ignored and did not significantly affect the results. Two patients died during the FU period from nonprocedure-related reasons (cancer).
Procedural details and acute success rate
Procedural details are displayed in Table 1. Mean procedural time was significantly longer in the RN group compared with the MN group (140.9 ± 36.5 min vs. 129.3 ± 43.1 min; p = 0.026), whereas fluoroscopy times did not differ significantly (23.8 [12.8 to 42.7] min vs. 18.5 [12.0 to 43.2] min; p = 0.42).
Furthermore, the mean number of overall RF applications was higher in the RN group compared with the MN group (29.9 ± 9.0 vs. 24.9 ± 10.4; p < 0.001) and the overall RF application duration was longer (2,501.1 ± 760.7 s vs. 2,059.1 ± 779.3 s; p < 0.001).
The acute PVI success rate was 100% in both ablation arms (p = 1.00). In the RN group, 2 of 131 (1.5%) patients required a switch to manual ablation for technical reasons (RNS malfunction).
All patients with 12 months’ FU and patients with arrhythmia recurrence and therefore shorter FU periods were included in the analysis. Eight patients were excluded from the analysis because they were lost to FU to prevent bias of the ablation results. All 8 of these patients did not have recurrent arrhythmia before becoming lost to FU.
At hospital discharge, 156 patients were treated with AADs, including 110 patients treated with class I AADs and 46 patients treated with class III AADs.
Recurrence rates were similar in the RN group and the MN group in the 247 patients after 12 months of FU (RN group: n = 29 of 123 [23.6%] vs. MN group: 25 of 124 [20.2%] reaching a level of significance for noninferiority [p = 0.014]) (Figure 2). Recurrences were AF in 49 patients (RN group, n = 26, MN group n = 23; p = 0.61) and LA tachycardia (LAT) in 5 patients (RN group n = 3; MN group, n = 2; p = 0.64). Of the 193 patients remaining in sinus rhythm after 12 months of FU, 18 of 193 (9%) were on AADs, including 11 patients on class I and 7 patients on class III AADs (RN group: 11 of 94 [12%]; class I, n = 7; class III, n = 4; MN group, 7 of 99 [7%]; class I, n = 4; class III, n = 3; p = 0.27).
Temperature monitoring and incidence of esophageal lesions
Episodes of esophageal temperature rise >41°C occurred in 158 of 258 (61.2%) patients (RN group, n = 78; MN group, n = 80) with a mean peak esophageal temperature of 43.2°C (RN group, 41.9°C; MN group, 44.4°C).
Endoscopy after the ablation procedure was performed in 200 of 258 (77.5%) patients (RN group, n = 103; MN group, n = 97). The remaining patients refused to undergo endoscopy after the procedure. Esophageal injury was observed in 36 of 200 (18%) patients (erythema/minimal lesions, n = 23; ulceration, n = 13). The incidence of esophageal injury was comparable between the RN group (n = 19) and MN group (n = 17) (erythema/minimal lesion: RN group, n = 11, vs. MN group n = 12; ulceration: RN group, n = 8, vs. MN group, n = 5; p = 0.72). All lesions recovered without sequelae except in 1 patient (see subsequent description).
A female patient from the minimal lesion group developed a hemorrhagic lesion (size, 10 ± 3 mm) without ulceration, identified during the index endoscopy after the ablation procedure. Endoscopic clipping of the lesion was performed using an over the scope clip (Ovesco Endoscopy, Tübingen, Germany). Supportive/protective antibiotic therapy was initiated and good results were confirmed on repeat endoscopy at days 1 and 5 after the clipping procedure. A light diet was started during the subsequent days and continued during the hospital stay. At day 6 after the clipping procedure, bilateral pleural effusions, periesophagel air, and signs of mediastinal infection were diagnosed via computed tomography (CT) scan. Repeat CT scan the following day showed less periesophageal air, signs of ongoing mediastinitis, and a slight pericardial effusion without hemodynamic compromise. No pain during swallowing or any neurological symptoms were reported by the patient, but clinical signs of pneumonia were present. Antibiotic therapy was adjusted during the clinical course. The patient remained asymptomatic and pneumonia gradually resolved. At day 17 after esophageal clipping, the patient was discovered unconscious on the ward with bleeding from the nose and mouth. Another CT scan confirmed the esophageal clip in the initial location and revealed pericardial effusion not requiring drainage. In addition, splenic infarction was detected. In an additional cerebral CT on the same day, no signs of cerebral infarction or bleeding were evident. On day 18 after esophageal clipping, the thoracal and cerebral CT scans were repeated. Early signs of a large media infarction were detected, the splenic infarction was increased, and an atrioesophageal fistula was detected. Immediate surgery revealed a ruptured esophagus and a perforated LA. After surgical suture, the patient was admitted to the intensive care unit and died the next day.
Procedure-related complications were observed in 16/258 (6.2%) patients, including 14 of 258 patients (5.4%) with major complications and 2 with minor complications (both RN group). The incidence for both overall procedure-related complications (RN group: n = 10 [7.6%], vs. MN group: n = 6 [4.7%]; p = 0.33) as well as the incidence of major complications (RN group: n = 8 [6.1%] vs. MN group: n = 6 [4.7%]; p = 0.62) did not show a significant statistical difference. Major complications are summarized in Table 2.
Distribution of complications in the RN group
Major complications (n = 8)
One patient died from atrioesophageal fistula (described previously). Two patients developed pericardial effusion requiring pericardial puncture. One patient developed pneumonia requiring antibiotic treatment with prolonged hospital stay. One patient developed femoral aneurysma spurium requiring prolonged hospital stay, and 2 patients developed femoral aneurysma spurium (1 with hematoma) requiring thrombin injection. One patient developed femoral aneurysm with hematoma 7 months after the ablation procedure requiring surgery.
Minor complications (n = 2)
In 1 patient, impaired gastric motility was seen during endoscopy requiring prolonged hospital stay. Control gastroscopy 3 days later revealed normal gastric function. Another patient developed slight groin hematoma not resulting in prolonged hospitalization.
Distribution of complications in the MN group
Major complications (n = 6)
One patient developed hematoma of the left pectoral muscle after subclavian puncture requiring prolonged hospital stay. In 1 patient, groin bleeding occurred 3 days after discharge. In 2 patients, groin hematoma resulted in prolonged hospital stay. Two further patients were readmitted to hospital for groin hematoma.
To the best of our knowledge, this is the first randomized, prospective, international multicenter controlled noninferiority trial comparing RN-controlled PVI to the current gold standard of manually performed circumferential RF ablation.
The main findings of this study are:
1. The success rate of RNS-guided AF ablation was noninferior to manually guided catheter ablation.
2. The acute procedure related complication rate in the RN group was not statistically different to the incidence in the MN group.
3. The incidence of esophageal lesions was comparable between RN and manually guided CPVI.
4. Procedure times were significantly longer in the RN group compared with the MN group.
Acute efficacy and 12-month outcome
PVI in patients with paroxysmal AF (PAF) is an established ablation strategy with success rates up to 80% even after long-term FU, but to a large extent this is dependent on the competency of the operator (1). The RNS was introduced to facilitate catheter navigation and to improve tip-to-tissue contact, enabling the creation of sufficient ablation lesions (8,10). Initial reports have shown that RNS-guided AF ablation is associated with comparable acute, midterm, and long-term success rates (6,9–11). This randomized multicenter study confirms that the success rate using RNS is noninferior to a manual ablation approach in experienced centers after a 12-month FU period.
The acute and long-term efficacy may be significantly influenced by a learning curve when using the RNS. This has been demonstrated by several investigators (10–12). All sites participating in this trial have performed at least 50 RN-guided PVI procedures before taking part in this study and therefore have passed the expected learning curve (10,11). This is 1 of the major differences of our study as compared with a previously published single-center randomized study by Ullah et al,, in which a high crossover rate to manual ablation was observed in the RNS group, which was most likely related to limited operator experience (6).
Procedure times and flouroscopy times
Procedure times of robotic ablation and manual ablation have been compared in previous studies. Although in some of these studies manual ablation was more time-consuming (12,13) and was associated with shorter fluoroscopy times (13), in general, RNS-guided ablation is commonly associated with slightly longer procedure and fluoroscopy times for catheter ablation of AF (6,13). This was confirmed by the findings of our study, in which procedure times were significantly longer in the RN group; however, fluoroscopy times were not statistically different.
Contact force parameters for the RNS
Electrode-to-tissue contact, RF power, and duration are the key determinants of RF lesion size (14–16). In general, complication rates when using the RNS were initially thought to be potentially higher because of improved tip-to-tissue contact and lack of tactical feedback to the operator. Electrophysiologists using the RNS were particularly afraid of the risk of LA perforation leading to cardiac tamponade.
In a previous study by Di Biase et al. (17), a contact force of >40g was associated with a significantly increased incidence of steam pop or crater formation compared to a contact force of 20 to 30g. Although recent studies revealed a comparable risk profile of the RNS compared with manual ablation (6,10), several modifications of contact force and power settings had to be considered (4,5,17,18). Therefore, for safety reasons in the Man and Machine trial, the catheter contact force at the posterior LA wall when using the RNS was controlled by the Intellisense software throughout all ablation procedures and maintained between a predefined threshold of 20 to 40g (7).
No contact force measurement was used in the manual ablation group; however, the impact of contact force on clinical outcome is still under investigation. In recent studies, contact force–guided manual AF ablation procedures were not superior to manual ablation procedures without contact force (19). Also, Weiss et al. (25) showed that force-sensing catheters and magnetic irrigated tip catheters yielded similar outcomes.
Acute procedure-related major complications as defined in the study protocol (7) were observed in 14 of 258 (5.4%) patients in the Man and Machine trial. No significant difference was seen between the RN and the MN groups with regard to the incidence of major complications (RN group: n = 8 [6.1%] vs. MN group: n = 6 [4.7%]; p = 0.62). These findings are in line with and confirm previous reports (10,12).
Esophageal temperature monitoring and esophageal injury
Esophageal injury after catheter ablation has been described previously for both manual and RNS-assisted catheter ablation (4,5).
In general, power reduction at the LA posterior wall in combination with esophageal temperature monitoring is recommended to minimize the risk of esophageal injury when using several of the novel ablation tools (2,4,5,20,21). However, routine use of esophageal temperature monitoring remains controversial and the use of a temperature probe may even enhance the risk for esophageal lesions (22). The optimal temperature cutoff for esophageal temperature monitoring is still under investigation.
In manually performed PVI, power reduction alone, with a maximum power of 30W at the posterior wall, appears to be associated with an acceptable safety profile (23). Recent publications have shown that a maximum power of 30 W at the posterior LA wall using RNS, without an esophageal temperature cutoff, is associated with an unacceptably high occurrence of esophageal ulcerations, leading to esophageal perforation in 1 patient (5). Even with a maximum power of 25 W at the posterior wall during PVI with RNS combined with esophageal temperature monitoring, the incidence of esophageal lesions remains significant (4). In contrast, other studies have demonstrated that an esophageal temperature cutoff of 41°C is associated with an acceptable incidence of esophageal lesions when RNS was used with a power limit of 20 W at the posterior LA wall (5). The overall incidence of esophageal injury in this study (18%) was comparable to the described incidence of esophageal lesions during RF ablation procedures in the current data (4,5). This was the rationale for the temperature cutoff of 41°C for the Man and Machine trial (7).
This hypothesis was confirmed by the results of our trial, in which the incidence of esophageal injury was comparable between the robotic and manual ablation groups. Unfortunately, 1 patient of the RN group developed an atrioesophageal fistula at day 18 after treatment of an esophageal lesion with an over the scope clip. This patient died despite surgery. It remains speculative, whether esophageal clipping favored the development of an atrioesophageal fistula or if this was the fatal course after esophageal injury.
This was a prospective, randomized, international multicenter trial. Success rates were evaluated with 72-h Holter monitoring. Although 3-day Holter monitoring actually is considered to be the standard of care (2), the success rate may be overestimated in this trial because no 7-day Holter monitoring or continuous monitoring with an implantable loop recorder was performed. The success rate was evaluated after a 12-month FU period; therefore, the long-term success rate has to be assessed in studies with a longer FU duration.
AF patterns and minimal PAF event rates were not documented for the patients in our study. This may have provided more detailed information about the patient population treated in our study and might be of interest in future studies.
No adenosine was administered to assess dormant PV reconduction. This may have led to a higher PV reconduction rates in this study population. This study, besides patients with paroxysmal AF, only included patients with persistent AF not lasting longer than 2 months and therefore excluded patients at higher risk for AF recurrence.
As this study was started in 2009, patients on phenprocoumon or warfarin were bridged with low-molecular-weight heparin before and after the ablation procedure. Although bleeding events were rare in this trial, less bleeding complications might have occurred if a protocol of uninterrupted oral anticoagulation would have been used.
The Man and Machine trial demonstrates that the success rate for robotically guided AF ablation is comparable to the current gold standard of manual ablation. Procedure times are significantly longer in robotic ablation procedures. The incidence of major complications did not statistically differ between robotic and manual ablation.
COMPETENCY IN MEDICAL KNOWLEDGE: This study demonstrated that robotic ablation is noninferior to the current gold standard of manual ablation for CPVI with respect to success and complication rates. Whereas the incidence of major complications did not statistically differ between robotic and manual ablation, procedure times were significantly longer in the robotic group.
TRANSLATIONAL OUTLOOK: Advances in catheter ablation are needed to further improve success rate in atrial fibrillation treatment. According to our findings, robotic navigation might be an opportunity to optimize atrial fibrillation ablation in the future.
The authors thank the data safety committee members Dietrich Andresen, Leif-Hendrick Boldt, and Charalampos Kriatselis for their important work.
This study is supported by a grant from St. Jude Medical and Hansen Medical. Dr. Tilz has received travel grants from St. Jude Medical, Topera, Biosense Webster, Daiichi Sankyo, and Sentrheart; and speakers bureau honoraria from Biosense Webster, Biotronik, Pfizer, Topera, Bristol-Myers Squibb, Bayer, and Sanofi Aventis. Dr. Rillig has received travel grants from Hansen Medical and St. Jude Medical; speaker honoraria from St. Jude Medical and Boehringer Ingelheim; and fellowship grants from Boston Scientific. Dr. Di Biase is a consultant for Stereotaxis, St Jude Medical, and Biosense Webster; and has received honoraria/travel from Medtronic, EpiEP, Biotronik, Medtronic, Janssen, and Pfizer. Dr. Natale has received speaker honoraria from Boston Scientific, Biosense Webster, St. Jude Medical, Biotronik, and Life Watch; and is a consultant to Biosense Webster Lifewatch and St. Jude Medical. All other authors have reported that they have relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- antiarrhythmic drug
- atrial fibrillation
- computed tomography
- left atrium
- manual ablation
- radiofrequency current
- robotic navigation
- robotic navigation system
- pulmonary vein isolation
- Received August 1, 2016.
- Revision received January 17, 2017.
- Accepted January 20, 2017.
- 2017 American College of Cardiology Foundation
- Ouyang F.,
- Tilz R.,
- Chun J.,
- et al.
- Calkins H.,
- Kuck K.H.,
- Cappato R.,
- et al.
- Rillig A.,
- Meyerfeldt U.,
- Birkemeyer R.,
- et al.
- Ullah W.,
- McLean A.,
- Hunter R.J.,
- et al.
- Saliba W.,
- Reddy V.Y.,
- Wazni O.,
- et al.
- Schmidt B.,
- Tilz R.R.,
- Neven K.,
- Julian Chun K.R.,
- Fürnkranz A.,
- Ouyang F.
- Rillig A.,
- Lin T.,
- Schmidt B.,
- et al.
- Nakagawa H.,
- Wittkampf F.H.,
- Yamanashi W.S.,
- et al.
- Rillig A.,
- Schmidt B.,
- Feige B.,
- et al.
- Ullah W.,
- McLean A.,
- Tayebjee M.H.,
- et al.
- Rillig A.,
- Lin T.,
- Burchard A.,
- et al.
- Singh S.M.,
- D'Avila A.,
- Doshi S.K.,
- et al.
- Kuwahara T.,
- Takahashi A.,
- Takahashi Y.,
- et al.
- Ouyang F.,
- Bänsch D.,
- Ernst S.,
- et al.