Peroral Endoscopic Myotomy for the Treatment of Achalasia: a Clinical Comparative Study of Endoscopic Full-Thickness and Circular Muscle Myotomy

Peroral Endoscopic Myotomy for the Treatment of Achalasia: A Clinical Comparative Study of Endoscopic Full-Thickness and Circular Muscle Myotomy

Quan-Lin Li, MD, Wei-Feng Chen, MD, Ping-Hong Zhou, MD, PhD, Li-Qing Yao, MD, Mei-Dong Xu, MD, PhD, Jian-Wei Hu, MD, Ming-Yan Cai, MD, Yi-Qun Zhang, MD, PhD, Wen-Zheng Qin, MD, Zhong Ren, MD, PhD

BACKGROUND: A circular muscle myotomy preserving the longitudinal outer esophageal muscular layer is often recommended during peroral endoscopic myotomy (POEM) for achalasia. However, because the longitudinal muscle fibers of the esophagus are extremely thin and fragile, and completeness of myotomy is the basis for the excellent results of conventional surgical myotomy, this modification needs to be further debated. Here, we retrospectively analyzed our prospectively maintained POEM database to compare the outcomes of endoscopic full-thickness and circular muscle myotomy. STUDY DESIGN: According to the myotomy depth, 103 patients with full-thickness myotomy were assigned to group A, while 131 patients with circular muscle myotomy were assigned to group B. Symptom relief, procedure-related parameters and adverse events, manometry outcomes, and reflux complications were compared between groups. RESULTS: The mean operation times were significantly shorter in group A compared with group B (p ¼ 0.02). There was no increase in any procedure-related adverse event after full-thickness myotomy (all p < 0.05). During follow-up, treatment success (Eckardt score 3) persisted for 96.0% (95 of 99) of patients in group A and for 95.0% (115 of 121) of patients in group B (p ¼ 0.75). There were no statistically significant differences in pre- and post-treatment D-value of symptom scores and lower esophageal sphincter pressures between groups (both p > 0.05). The overall clinical reflux complication rates were also similar (21.2% vs 16.5%, p ¼ 0.38). CONCLUSIONS: Short-term symptom relief and manometry outcomes of each method were comparable. Full- thickness myotomy significantly reduced the procedure time but did not increase the procedure-related adverse events or clinical reflux complications. (J Am Coll Surg 2013;217: 442e451. 2013 by the American College of Surgeons)

Achalasia is a primary esophageal motor disorder charac- esophagus.1 Diagnosis and surgical treatment of achalasia terized by the absence of peristalsis and a defective relax- have been established, even if still debated, for a long ation of the lower esophageal sphincter (LES), resulting in time. The extent of open or laparoscopic Heller myotomy impaired bolus transport and stasis of food in the was deﬁned on the basis of clinical and manometric

CME questions for this article available at Presented at Digestive Disease Week 2013, Orlando, FL, May 2013. http://jacscme.facs.org Received February 6, 2013; Revised April 1, 2013; Accepted April 23, 2013. Disclosure Information: Authors have nothing to disclose. Timothy J Eberlein, From the Endoscopy Center and Endoscopy Research Institute, Zhongshan Editor-in-Chief, has nothing to disclose. Hospital, Fudan University, Shanghai, China. Financial support: This study was supported by the grants from the Correspondence address: Ping-Hong Zhou, MD, PhD, Endoscopy Medical Leading Project of Shanghai Municipal Science and Technology Center and Endoscopy Research Institute, Zhongshan Hospital, Fudan Committee (10411969600), and the Major Project of Shanghai Munic- University, 180 FengLin Rd, Shanghai 200032, P R China. email: ipal Science and Technology Committee (10411955900, 11411950502 [email protected] and 11DZ2280400). Drs Quan-Lin Li and Wei-Feng Chen contributed equally to this work.

ª 2013 by the American College of Surgeons ISSN 1072-7515/13/$36.00 Published by Elsevier Inc. 442 http://dx.doi.org/10.1016/j.jamcollsurg.2013.04.033 Vol. 217, No. 3, September 2013 Li et al Endoscopic Full-Thickness Myotomy 443

symptom remission. For these considerations, POEM Abbreviations and Acronyms with a full-thickness myotomy has been performed in EGD ¼ esophagogastroduodenoscopy about half of the achalasia patients in our center. In EGJ ¼ esophagogastric junction this study, we retrospectively analyzed our prospec- ¼ GER gastroesophageal reﬂux tively maintained POEM database to compare the out- HRM ¼ high-resolution manometry LES ¼ lower esophageal sphincter comes of endoscopic full-thickness and circular muscle POEM ¼ peroral endoscopic myotomy myotomy, hypothesizing that full-thickness myotomy PPI ¼ proton pump inhibitor would be superior to circular muscle myotomy alone.

METHODS results.2 The standard surgical procedure involves complete division of both the longitudinal and circular Study design muscles of the esophagus from the distal esophagus This retrospective analysis was approved by the local onto the proximal stomach while preserving the mucosa. research ethics committee. The analysis included consecu- Despite improvement in perioperative outcomes after tive patients who had primary achalasia and were treated laparoscopic myotomy, it is still an invasive procedure, with POEM successfully by a single operator (PH Zhou) causes significant pain and discomfort, and is relatively at the authors’ institution between August 2010 and expensive. March 2012. Patients were eligible for enrollment in the Peroral endoscopic myotomy (POEM) has recently study if they had an Eckardt symptom score 4. The diag- been described as a scarless and less invasive surgical nosis of achalasia was made on the basis of the absence of myotomy option for treatment of achalasia.3-8 Initial pub- peristalsis and on impaired relaxation of the LES on established experience in humans is more than encouraging lished methods; barium swallow and esophagogastroduo- despite a relatively short follow-up. A circular muscle denoscopy (EGD) were performed in all enrolled myotomy preserving the longitudinal outer esophageal patients before POEM, and high resolution manometry muscular layer is often recommended during POEM, (HRM) was performed in 127 of 234 of them. Exclusion which is different from the usual full-thickness myotomy criteria were severe cardiopulmonary disease or other performed surgically. Although selective circular muscle serious disease leading to unacceptable surgical risk, pseu- > myotomy is designed in order to avoid entering the doachalasia, megaesophagus (diameter 7 cm), and hiatal > pleural space and decrease morbidity, it is often hard to hernia ( 2 cm). All included patients provided written achieve because the longitudinal muscle fibers of the informed consent to undergo POEM after detailed spoken esophagus are extremely thin, which frequently leads to and written explanations concerning the POEM proce- an unintentional splitting of the muscle fibers during dure and other possible treatment options. POEM. Either minor electrocautery damage, mechanical Procedural details were recorded prospectively in a data- trauma from maneuvering the endoscope in the tunnel, base. After thoroughly investigating the database and or air insufflation alone can result in spreading of the their medical records, patients were grouped in 2 sets longitudinal muscle fibers and adventitia and transmural according to myotomy depth: Group A underwent endo- openings into the mediastinum.8 In fact, a clear separa- scopic full-thickness myotomy, including the internal tion of circular (to be completely dissected) and longitu- circular and longitudinal muscular layer; group B under- dinal (to be left) muscular layers cannot be found any went endoscopic circular muscle myotomy alone, more at the esophagogastric junction (EGJ) and stomach. involving only the internal circular muscular layer. Moreover, although dissection of the circular muscle Patients who had only limited damage of the longitudinal fibers alone led to a significant reduction of LES pressure muscle fibers were also assigned to group B. In this study, as demonstrated by postoperative manometry in some an inadvertent full-thickness myotomy was performed in studies,3,9 in the experience of others, completeness of 37 of 154 enrolled patients before October 2011; an myotomy is a prerequisite for sufficient and long-term intentional full-thickness myotomy was performed in reduction of LES pressure and is the basis for the excellent 66 of 80 enrolled patients subsequently after considering result of conventional surgical myotomy.8,10 Given the the hypothesis that full-thickness myotomy would be fact that an incomplete myotomy with possible fibrotic superior for the excellent result. healing may be considered a major reason for postoperative recurrence,11,12 a full-thickness myotomy may reduce Outcomes measurements the incidence of fusion by further spreading of the cut The main outcomes measures that were compared muscular wall and theoretically inducing long-term between the 2 groups were: 444 Li et al Endoscopic Full-Thickness Myotomy J Am Coll Surg

1. Therapeutic success at the follow-up assessment, Patients were admitted 1 day before the POEM proce- defined as a reduction in the Eckardt score to 3;8 dure to undergo EGD, cleaning out food remnants from 2. Procedure-related parameters such as procedure time, the esophagus. Prophylactic intravenous antibiotics and hospital stay, and myotomy length; proton pump inhibitors (PPI) were introduced 30 3. Procedure-related adverse events such as mucosal minutes before the procedure. The POEM procedure injury/perforation, subcutaneous emphysema, pneu- was performed under general anesthesia, as described mothorax, pneumoperitoneum, pneumonia, focal atel- previously.3 In brief, submucosal injection and an initial ectasis, pleural effusion, fever (temperature 38C), mucosal incision were first done in the 5 to 6 o’clock severe postoperative pain (narcotic treatment such as position on the posterior esophagus approximately fortanodyn and sauteralgyl), submucosal infection, 10 cm proximal to the EGJ. Then, a submucosal tunnel delayed bleeding, etc. Pneumomediastinum occurred was created, passing over the EGJ, and about 3 cm into in nearly all patients with full-thickness myotomy, so the proximal stomach. The myotomy was begun at 2 cm it wasn’t considered a complication in this study; distal to the mucosal entry point, approximately 6 to 8 4. Lower esophageal sphincter pressure on manometry cm above the EGJ, and was extended for a distance of pre- and post POEM; and 2 to 3 cm to the stomach in all patients. Selective dissec- 5. Clinical reflux complications at the follow-up assess- tion of the circular muscle bundles was attempted in ment, defined as reflux symptoms or esophagitis on group B and the longitudinal muscle bundles were care- EGD. fully protected and left intact at the limit of the dissected area. In group A, a full-thickness myotomy, including the internal circular and longitudinal muscular layer, High resolution manometry was done. Substantial reduction of LES tonus was Baseline and postmyotomy LES pressures were recorded confirmed by opening the EGJ with gentle insufflation using an HRM system (Sierra Scientific Instruments through the endoscope and easily passing an endoscope 13 Inc), as previously described. Briefly, the HRM through the lumen of the esophagus. After careful hemo- assembly was placed transnasally and the manometric stasis, the mucosal incision site was closed with 4 to 6 catheter positioned to record from the hypopharynx to hemostatic clips. A mucosal injury or perforation, which the stomach with approximately 5 intragastric sensors. was suspected on endoscopic control, was also clipped Studies were performed with the patients in a supine posi- and a nasogastric tube was placed when necessary. An tion after at least a 6-hour fast. The manometric protocol example of the endoscopic full-thickness myotomy included a 5-minute period to assess basal EGJ pressure; procedure is shown in Fig. 1 and Video 1. 10 water swallows of 5 mL; and 1 water swallow each of 1 mL (dry), 10 mL, and 20 mL. All manometric analysis was done using Mano-View software applied to the data Postprocedure management tracings viewed in the color pressure topography mode To evaluate procedure-related adverse events after and referenced to intragastric pressure. POEM, such as subcutaneous emphysema, pneumothorax, and pleural effusion, a chest CT scan was carried Peroral endoscopic myotomy procedures out on the first day after operation in most patients. If Peroral endoscopic myotomy was attempted with a single- patients developed a pneumothorax after the procedure, channel gastroscope (GIF-H260, Olympus Medical instead of inserting chest tube drainage, we inserted Systems Co) and a hybrid knife (ERBE, Erbe Elektromedi- a central venous catheter at the third or fourth inter- zin GmbH), triangle-tip knife (KD-640L; Olympus), or costal space. A 20-gauge needle was used to relieve the a hook knife (KD-620LR, Olympus). A transparent cap severe pneumoperitoneum during and after the proce- (D-201-11802, Olympus) was attached to the tip of the dure. Postoperative observations included analysis of gastroscope. Other equipment included injection needle chest pain, dyspnea, abdominal pain or distention, and (NM-4L-1, Olympus), hot biopsy forceps (FD-410LR, cyanosis or signs of peritonitis. Postoperative medica- Olympus), clips (HX-610-90, HX-600-135, Olympus; tions included a double-dose PPI, antibiotics, and hemo- Resolution), a high-frequency generator (VIO 200D, coagulase injection. Patients were kept npo for 24 hours ERBE), and an argon plasma coagulation unit (APC300, after POEM, and on a liquid diet for an additional 24 ERBE). Room air was used for insufflations during the hours. Patients were discharged with double-dose PPI procedure before December 2011 and carbon dioxide gas and a soft diet for 2 weeks. Patients were asked to dis- was used for insufflations with a CO2 insufflator (UCR; continue PPI medication and start a regular diet 4 to Olympus) subsequently. 8 weeks after POEM. Vol. 217, No. 3, September 2013 Li et al Endoscopic Full-Thickness Myotomy 445 FPO = print & web 4C Figure 1. Peroral endoscopic myotomy (POEM) with a full-thickness myotomy. (A, B, C, D) Sequence of a full-thickness myotomy being performed at the right lateroposterior wall. Note after clinging to the longitudinal muscle fibers and lifting them up toward the esophageal lumen, the full-thickness muscle bundles being sectioned. (D) The arrow indicates the external coat of the distal esophagus.

Follow-up RESULTS Patients were scheduled for a follow-up visit at 1 month During the study period, POEM was attempted in a total after POEM for HRM. They also underwent follow-up of 238 consecutive patients with achalasia. In 4 patients EGD at 1, 3, 6, and 12 months after POEM and annually (1.7%), submucosal tunneling failed for the following thereafter to view the healing of the wound and to check reasons: serious inflammation and adhesion (n ¼ 2); sus- any objective sign of reflux esophagitis. Patients were con- pected cardia cancer (n ¼ 1); and absence of the submu- tacted via telephone every 3 months postoperatively to cosal layer (n ¼ 1). So, 234 patients of those underwent assess for complications and to obtain a current Eckardt POEM successfully and were finally included in this anal- score. Symptomatic gastroesophageal reflux (GER) was ysis. According to myotomy depth, 103 patients with assessed every 3 months postoperatively using the GerdQ full-thickness myotomy were included in group A, and questionnaire.14 A score 7 (scale 0 to 18) was considered 131 patients with circular muscle myotomy were positive for GER, in line with established usage of the ques- included in group B. tionnaire. A final checkup was performed in May 2012. Patient characteristics Statistical analysis Table 1 summarizes the clinical characteristics of those Statistical analysis was performed with SPSS 16.0 soft- patients. Patients in group A were younger than those ware. Measurement values were expressed as means and in group B (37.6 13.2 vs 41.5 16.3 years, respec- standard deviation. Statistical significance was evaluated tively, p ¼ 0.04). The mean duration of symptoms before using Student’s t-test for independent or paired samples, undergoing POEM was similar between groups (7.2 chi-square test, or Fisher’s exact test as appropriate. All 7.7 vs 7.9 9.4 years, respectively, p ¼ 0.53). Before reported p values were 2-tailed, and p values <0.05 POEM, 39 patients in group A and 45 patients in group were considered to indicate statistical significance. B had received other endoscopic therapy (dilation only, 446 Li et al Endoscopic Full-Thickness Myotomy J Am Coll Surg

Table 1. Preoperative Patient Characteristics Characteristic Group A (full-thickness myotomy) Group B (circular muscle myotomy) p Value n 103 131 Sex, M/F, n 45/58 67/64 0.26 Age, mean SD, y 37.6 13.2 41.5 16.3 0.04 Duration of symptoms, mean SD, y 7.2 7.7 7.9 9.4 0.53 Previous treatment, n (%) Other endoscopic therapy 39 (37.9) 45 (34.4) 0.58 Heller myotomy 6 (5.8) 8 (6.1) 0.93 POEM 0 1 (0.8) 1.00 Sigmoid type 1 esophagus, n (%) 11 (10.7) 7 (5.3) 0.13 POEM, peroral endoscopic myotomy.

30 vs 23; Botox only, 3 vs 2; temporary stenting, 3 vs 9; 2 emphysema and bilateral pneumothorax, which were or more methods, 3 vs 11, respectively; p ¼ 0.58). Four- treated with subcutaneous puncture and closed thoracic teen patients after failed Heller myotomy (6 in group A vs drainage intraoperatively. After these treatments, opera- 8 in group B, p ¼ 0.93), 18 patients with sigmoid type 1 tions were continued and successfully completed. No esophagus (11 in group A vs 7 in group B, p ¼ 0.13), and patient had massive hemorrhage during the operation. 1 patient after failed initial POEM (in group B) were also A total of 84 patients in group A (81.6%) and 112 included in the study. No patient was with sigmoid type patients in group B (85.5%) underwent CT scans after 2 esophagus in the study. the procedure. As judged on CT scan, there were no significant differences between groups A and B in postop- Procedure-related parameters erative pneumothorax (20.2% [17 of 84] vs 24.1% [27 of As indicated in Table 2, mean operation times were 112]; p ¼ 0.52), pneumoperitoneum (42.9% [36 of 84] significantly shorter in group A compared with group B vs 47.3% [53 of 112]; p ¼ 0.53), pleural effusion (66.7% (41.7 18.9 vs 48.9 28.6 minutes, p ¼ 0.02). The [56 of 84] vs 59.8% [67 of 112]; p ¼ 0.33), pneumonia total lengths of the endoscopic myotomy were similar (48.8% [41 of 84] vs 54.5% [61 of 112]; p ¼ 0.43), and between groups (8.1 1.3 vs 7.9 1.3 cm in the esoph- focal atelectasis (14.3% [12 of 84] vs 17.9% [20 of 112]; agus and 2.2 0.6 vs 2.3 0.5 cm in the stomach; p ¼ 0.50), respectively. However, similar to intraopera- p ¼ 0.41 and p ¼ 0.07, respectively). Twenty-seven tive subcutaneous emphysema, the incidence of postoper- patients in group A (26.2%) and 23 patients in group ative subcutaneous emphysema on CT scan was also less B (17.6%) underwent POEM with the use of the hybrid in group A (26.2% [22 of 84] vs 44.6% [50 of 112]; knife (p ¼ 0.11). p ¼ 0.01). Among those who had pneumothorax, 7 patients in group A and 7 patients in group B needed Procedure-related adverse events closed thoracic drainage. Only 1% of patients (2 of As shown in Table 2, during the operation, 13 patients in 196) with pleural effusion needed intervention, and spon- group A (12.6%) and 27 patients in group B (20.6%) had taneous absorption of the effusion was observed in other mucosal injury or small mucosal perforations (p ¼ 0.11). patients. Other findings on CT scans were minor prob- All perforations were clipped after myotomy and nasogas- lems without apparent clinical effect and did not need tric tubes were placed in 10 patients. The submucosal further intervention. penetration did not induce any clinical complications. In our study, room air was used for insufflation in Intraoperative subcutaneous emphysema was less in fewer patients in group A (63 of 103 [61.2%] vs 124 group A compared with group B (7.8% [8 of 103] vs of 131 [94.7%]; p ¼ 0.00), which seemed to be respon- 22.1% [29 of 131]; p ¼ 0.00). (Note that the study sible for less intraoperative and postoperative subcuta- authors rounded down p values of < 0.005 to 0.00 in neous emphysema. When we subdivided the patients order to keep the same number of decimal places for all via the insufflated gas during the procedure (room air p values.) Of those, 1 patient in group A and 1 patient vs CO2), we found that all intraoperative subcutaneous in group B suffered from pneumoperitoneum simulta- emphysema occurred in patients with room air insuffla- neously, and Veress needle decompression of the pneu- tion. In patients with room air insufflation, there moperitoneum was required intraoperatively; 1 patient were no significant differences between groups A and B in group B developed both severe subcutaneous in postoperative subcutaneous emphysema (41.3% Vol. 217, No. 3, September 2013 Li et al Endoscopic Full-Thickness Myotomy 447

Table 2. Comparison of Procedure-Related Parameters and Adverse Events Between Group A and Group B Variable Group A (full-thickness myotomy) Group B (circular muscle myotomy) p Value Mean operation time, mean SD, min 41.7 18.9 48.9 28.6 0.02 Myotomy length, mean SD, cm Esophagus 8.1 1.3 7.9 1.3 0.41 Stomach 2.2 0.6 2.3 0.5 0.07 Use of hybrid knife, n (%) 27 (26.2) 23 (17.6) 0.11 Room air insufﬂations, n (%) 63 (61.2) 124 (94.7) 0.00 Intraoperative complications, n (%) Mucosal injury 13 (12.6) 27 (20.6) 0.11 Subcutaneous emphysema 8 (7.8) 29 (21.1) 0.00 Room air insufﬂation 8 29 0.08

CO2 insufﬂation 0 0 1.00 Pneumothorax 0 1 (0.8) 1.00 Pneumoperitoneum 1 (1.0) 1 (0.8) 1.00 Postoperative complications, n (%)

Judged on CT scan (air vs CO2 insufﬂation), n 46/38 107/5 Subcutaneous emphysema 22 (26.2) 50 (44.6) 0.01 Room air insufﬂation 19 50 0.54

CO2 insufﬂation 3 0 1.00 Pneumothorax 17 (20.2) 27 (24.1) 0.52 Room air insufﬂation 16 27 0.23

CO2 insufﬂation 1 0 1.00 Pneumoperitoneum 36 (42.9) 53 (47.3) 0.53 Room air insufﬂation 29 52 0.10

CO2 insufﬂation 7 1 1.00 Pleural effusion 56 (66.7) 67 (59.8) 0.33 Pneumonia 41 (48.8) 61 (54.5) 0.43 Focal atelectasis 12 (14.3) 20 (17.9) 0.50 Other Fever 2 (1.9) 4 (3.1) 0.59 Severe postoperative pain 13 (12.6) 11 (8.4) 0.29 Submucosal infection 1 (1.0) 0 0.44 Delayed bleeding 0 1 (0.8) 1.00 Length of antibiotics treatment, mean SD, d 2.4 1.0 3.2 2.0 0.00 Length of postoperative hospital stay, mean SD, d 2.7 1.1 3.6 2.7 0.00

[19 of 46] vs 46.7% [50 of 107]; p ¼ 0.54), pneumo- and 10.1 3.6 109/L post-treatment in group A thorax (34.8% [16 of 46] vs 25.2% [27 of 107]; p ¼ (p ¼ 0.00), and was 6.0 2.2 109/L and 10.8 0.23), and pneumoperitoneum (63.0% [29 of 46] vs 3.5 109/L, respectively, in group B (p ¼ 0.00). A 48.6% [52 of 107]; p ¼ 0.10). These results indicated significant decrease in hemoglobin level also occurred that full-thickness myotomy did not increase the risk of after POEM. Mean hemoglobin level decreased from ¼ any of these events and CO2 insufflation was vitally 13.7 1.4 g/dL to 12.4 1.4 g/dL in group A (p important to reduce gas-related compilations. 0.00) and from 13.0 2.0 g/dL to 12.2 1.7 g/dL in Two patients in group A (1.9%) and 4 patients in group B (p ¼ 0.00). Thirteen patients in group A group B (3.1%) developed signs of fever and temperature (12.6%) and 11 patients in group B (8.4%) had severe 38C(p¼ 0.59). A total of 92 patients had paired postoperative pain on the day of operation and received blood test pre- and post-treatment (34 in group A vs narcotic treatment (p ¼ 0.29). 58 in group B, p ¼ 0.08). There were significant increases There were no differences in other complications in leukocyte count after POEM in both groups. Mean between groups. One patient in group A had submucosal leukocyte count was 6.0 2.8 109/L pretreatment tunnel infection because of an opening at the submucosal 448 Li et al Endoscopic Full-Thickness Myotomy J Am Coll Surg entry and accumulation of a large amount of food; 1 Eckardt scores from 8.0 2.0 (range 5 to 11) pretreat- delayed bleeding in the submucosal tunnel and 1 epileptic ment to 5.5 2.0 (range 4 to 10) post-treatment (p ¼ seizure occurred in group B. 0.00). Of them, 5 underwent additional therapies with Both the average lengths of antibiotics treatment and endoscopic balloon dilation, and the other patients didn’t postoperative hospital stay were significantly shorter for request any further treatments until the final checkup. group A compared with group B (antibiotics treatment, One patient with a treatment failure (with the same Eck- 2.4 1.0 days vs 3.2 2.0 days, p ¼ 0.00; hospital ardt score of 10, before and after POEM) died from stay, 2.7 1.1 days vs 3.6 2.7 days, p ¼ 0.00; cachexia in group A, while 1 patient with a treatment respectively). success died from lymphoma in group B. Symptom relief The symptom score follow-up rate was extremely high, Manometry outcomes 96.1% (99 of 103) for group A and 92.4% (121 of As shown in Table 3, more patients in group A had per- 131) for group B (p ¼ 0.23). As shown in Table 3, the protocol HRM at 1 month postoperatively compared average observation period was significantly shorter for with group B (68 of 103 [66.0%] vs 51 of 131 group A (6.1 4.3 months vs 10.5 3.8 months, [38.9%]; p ¼ 0.00). Other patients didn’t undergo p ¼ 0.00). Treatment success (Eckardt score 3) per- follow-up manometry, due to the actual condition of sisted for 96.0% (95 of 99) of patients in group A and our hospital (too many patients, shortage of equipment for 95.0% (115 of 121) of patients in group B (p ¼ and professionals), loss to follow-up, discomfort related 0.75). Mean symptom score decreased from 7.6 2.0 to the manometry procedure, and other personal reasons. to 1.2 1.5 in group A (p ¼ 0.00), and from 8.0 Mean LES pressure decreased from a mean of 30.5 1.9 to 1.1 1.3 in group B (p ¼ 0.00). There was no 14.5 mmHg to 12.4 3.6 mmHg after POEM in group statistical significant difference in pre/post-treatment A(p¼ 0.00); in group B, it decreased from 29.6 11.5 D-value of symptom scores between the 2 groups (6.5 mmHg to 12.2 4.3 mmHg (p ¼ 0.00). There was no 2.3 vs 7.0 2.2, p ¼ 0.11). The average gain in weight statistical significant difference of pre- or post-treatment post-treatment was also similar between groups (5.7 D-values of LES pressure between the 2 groups (18.1 5.0 kg vs 6.7 5.2 kg, p ¼ 0.18). 13.9 mmHg vs 17.4 10.2 mmHg, p ¼ 0.75). In 6 During follow-up, 4 patients in group A and 6 patients patients with treatment failure and per-protocol HRM in group B had treatment failure. These patients had mentioned earlier, LES pressures were 25.0 8.9

Table 3. Comparison of Symptom Relief, Manometry Outcomes, and Reflux Complications Between Group A and Group B Group A Group B Variable (full-thickness myotomy) (circular muscle myotomy) p Value Mean follow-up period, mean SD, mo 6.1 4.3 10.5 3.8 0.00 Symptom score follow-up rate, n (%) 99 (96.1) 121 (92.4) 0.23 Eckardt score, mean SD Pretreatment 7.6 2.0 8.0 1.9 Post-treatment 1.2 1.5 1.1 1.3 Pre/post-treatment D-value 6.5 2.3 7.0 2.2 0.11 Treatment success (Eckardt score 3), n (%) 95 (96.0) 115 (95.0) 0.75 Average gain in weight post-treatment, mean SD, kg 5.7 5.0 6.7 5.2 0.18 Manometry follow-up rate, n (%) 68 (66.0) 51 (38.9) 0.00 Lower esophageal sphincter pressure, mean SD, mmHg Pretreatment 30.5 14.5 29.6 11.5 Post-treatment 12.4 3.6 12.2 4.3 Pre/post-treatment D-value 18.1 13.9 17.4 10.2 0.75 Clinical reflux complications, n (%) Symptomatic reflux only (GerdQ score 7) 8 13 Esophagitis on EGD only 2 0 Both 11 7 Overall 21 (21.2%) 20 (16.5%) 0.38 EGD, esophagogastroduodenoscopy. Vol. 217, No. 3, September 2013 Li et al Endoscopic Full-Thickness Myotomy 449 mmHg pretreatment and 12.0 4.9 mmHg post- Because recurrent dysphagia, which is most commonly treatment (p ¼ 0.06). caused by late scarring or fibrosis at the site of the myotomy, typically occurs more than 6 months after the initial Reflux complications procedure,12 the relatively short follow-up period (6.1 At the same follow-up interval, 19 patients in group A 4.3 months in the full-thickness myotomy group and and 20 patients in group B had a GerdQ score 7, 10.5 3.8 months in the circular muscle myotomy indicating symptomatic GER. Thirteen patients in group), to some extent, seems to be responsible for these group A and 7 patients in group B were found to short-term results. Further observations and follow-up to have mild or moderate esophagitis (Los Angeles classi- evaluate long-term outcomes have been initiated. In addi- fication B or C) on follow-up EGD. Among them, 11 tion, other objective evaluations such as timed barium 8,16 patients in group A and 7 patients in group B had swallow and endoscopic functional luminal imaging 9,17 both esophagitis on EGD and a GerdQ 7. So the probe, should be used to determine esophageal overall clinical reflux complication rates were similar emptying and EGJ distensibility for predicting clinical between groups (21.2% [21 of 99] vs 16.5% [20 of success. 121]; p ¼ 0.38; Table 3). Generations of surgeons were taught to stay away from the esophagus due to its “unforgiving” nature and the fact that this organ has no serosal lining. So, selective DISCUSSION circular muscle myotomy is designed to avoid entering Peroral endoscopic myotomy incorporates concepts of the pleural space and decrease morbidity during POEM. natural orifice transluminal endoscopic surgery and However, in this study, we found similar intra- and perio- expands on techniques used in endoscopic submucosal perative complication rates between the 2 procedures, dissection in order to achieve a division of the esophageal providing initial evidence that full-thickness myotomy circular muscle fibers across the EGJ and onto the has a safety profile comparable to that of circular muscle stomach. The technical modification proposed by this myotomy alone. During circular muscle myotomy, a slight previously published endoscopic method consists of push of the tip of the endoscope could result in splitting the partial-thickness myotomy involving the internal circular outer longitudinal muscle because of its thin and fragile muscular fibers to decrease morbidity, which is different nature. This might explain the occurrence of pneumome- from the usual full-thickness myotomy performed surgi- diastinum pneumothorax, pneumoperitoneum, and/or cally. However, because completeness of myotomy is subcutaneous emphysema in patients with circular muscle the basis for the excellent result of conventional surgical myotomy alone. So, although the aim is to leave the layer of myotomy, this modification needs to be further debated. longitudinal muscle fibers intact, some degree of pneumo- Two recent studies reported that endoscopic full- mediastinum, and potentially, pneumoperitoneum, pneu- thickness myotomy did not seem to be superior to mothorax, and/or subcutaneous emphysema, will likely partial-thickness myotomy, as demonstrated by postoper- result regardless of technique. During full-thickness myot- ative manometry in pigs,9,15 and because this was a physi- omy, visible complete transmural openings into the medi- ologic, nonachalasic animal model, comparisons to the astinum and into the peritoneal cavity did occur and clinical setting may not be valid. In achalasia patients, active gas insufflation during endoscopy was therefore ex- von Renteln D and colleagues8 demonstrated that pected to leak into the mediastinum and peritoneal cavity. a complete myotomy (full-thickness myotomy) at the However, pneumomediastinum, pneumoperitoneum, cardia may be superior in providing rapid esophagogastric and/or subcutaneous emphysema alone are often of limited emptying, as demonstrated by postoperative barium clinical significance. In fact, pneumomediastinum would swallow, despite the fact that there were no significant usually not be considered a complication because during differences in the symptom score and a reduction in laparoscopic dissection at the hiatus, as it has a reported LES pressures between groups; however, their study pop- incidence as high as 85%, but resolves within few days.18 ulation was not large enough (only 16 patients). In this So care must be taken not to injure the underlying pleura study we compared the outcomes of endoscopic full- and cause pneumothorax. We advocate using CO2 instead thickness and circular muscle myotomy in a total of of room air for insufflation, reducing air insufflation as 234 patients, which represented the largest POEM series much as possible, and clinging to the longitudinal muscle to date. Our results demonstrated that short-term fibers and lifting them up toward the esophageal lumen symptom relief and manometry outcomes of each method for dissection during full-thickness myotomy. were comparable, which mirrored the finding of previous Postoperative infectious complication is a critical series.8,9,15 concern for patients undergoing full-thickness myotomy. 450 Li et al Endoscopic Full-Thickness Myotomy J Am Coll Surg

However, full-thickness myotomy was not associated with However, as a consequence of tissue adhesions and any infectious complications in our study. Two patients limited space, mucosal injury or even perforation some- after full-thickness myotomy and 4 patients after circular times occurs, especially at the cardia. This area can be muscle myotomy developed signs of fever. Only 1 patient clipped with metallic clips after myotomy. When neces- after full-thickness myotomy had submucosal tunnel sary, endoscope-assisted nasogastric tube placement can infection, but it was due to an opening at the submucosal also be used for decompression. After mucosal integrity entry and accumulation of a large amount of food. This is repaired, the mucosal injury often does not induce supports the hypothesis that endoscopic 2-level perfora- any clinical complications, even in patients who have tion of the gastrointestinal wall using the submucosal a full-thickness myotomy. technique, even into the mediastinum, is possible as The use of blunt dissection for submucosal tunneling, long as the mucosal opening is sufficiently closed and either balloon22 or other methods, may be considered an ongoing spillage is avoided.19,20 Once the anatomic loca- alternative to decrease the risk of incidental mucosal tion of the mucosal perforation is separated from the injury, especially for operators without sufficient endo- muscular perforation, healing of both sites is possible scopic submucosal dissection experience. However, the without violating the principles of a full-thickness perfo- disadvantage of balloon dilation is the inability to posi- ration of the esophagus. tion the balloon accurately within the wall layers which, In our study, full-thickness myotomy created myoto- without visual control, may lead to unforeseen damage mies of similar length with significantly shorter operative to esophageal structures. Recently, Khashab and associ- times when compared with circular muscle myotomy. ates23 described their initial experience with the use of However, the advantage of full-thickness myotomy in a novel gel to facilitate esophageal submucosal tunneling terms of operative time may be not great enough and it during POEM. However, “auto-tunneling” by the dis- is unknown whether these differences will translate into secting gel stopped at the level of the LES, and further any concrete clinical benefits for patients. Although tunneling into the gastric cardia using the endoscopic most circular muscular myotomies occurred early in the submucosal dissection technique was needed. Moreover, series, a learning curve effect may not be considered an because most mucosal injury occurs in difficult patients important factor for the shorter procedure duration of with significant submucosal adhesions, sigmoid type full-thickness myotomy because all POEM procedures esophagus, or other previous treatments, the use of blunt were performed by a single senior surgeon (PH Zhou) dissection for submucosal tunneling is often difficult to in this study, and the learning curve may plateau at about achieve. 20 cases for experienced endoscopists.21 In addition, we Gastroesophageal reflux is the other critical concern for have demonstrated, in a prospective randomized trial patients undergoing full-thickness myotomy. As opposed (data not shown, discussed in another upcoming paper), to Heller myotomy, POEM does not include an antire- that compared with the conventional electric knife, the flux procedure, which could potentially result in increased use of the hybrid knife can lead to a significant decrease rates of GER postoperatively. Then, preservation of the in POEM procedure time due to less replacement of longitudinal muscle fibers is considered as an option to accessories. However, the use of the hybrid knife was help prevent GER after POEM. However, because the comparable between groups in this study and therefore integrity of natural antireflux mechanisms around the did not contribute to the significantly shorter duration EGJ is important to prevent reflux,24 POEM with for the full-thickness myotomy group. In our experience, a full-thickness myotomy also achieves myotomy from full-thickness myotomy can reduce operative times signif- the luminal side with minimal disturbance of these mech- icantly because much time can elapse when we carefully anisms. Similar to circular muscle myotomy alone, elim- protect the fragile longitudinal muscles. ination of dissection at the level of the diaphragmatic One refinement of POEM is the application of the hiatus thereby preserves the phrenoesophageal ligaments, endoscopic submucosal dissection technique to create which may be beneficial to avoid postmyotomy reflux in the submucosal tunnel. This precaution is vital to avoid patients undergoing full-thickness myotomy. In this injuring the overlying mucosa while tunneling because study, the overall clinical reflux complication rates were mucosal injury may put the patient at the risk for a poten- similar between groups (21.2% vs 16.5%; p ¼ 0.38). tial fistula with mediastinal sepsis, especially in patients However, a study limitation is that no postoperative who have a full-thickness myotomy. It is important to 24-hour pH studies were performed. Although the create the tunnel as close as possible to the muscular layers POEM procedure potentially indicated a lower reflux not only to avoid injury to the mucosal flap but also incidence, because more than 50% of laparoscopic Heller due to the lower vascularity adjacent to the muscle. myotomy patients without a fundoplication will have Vol. 217, No. 3, September 2013 Li et al Endoscopic Full-Thickness Myotomy 451

pathologic reflux, and as many as 10% of patients with 8. von Renteln D, Inoue H, Minami H, et al. Peroral endoscopic a Heller and partial fundoplication will have problems,25 myotomy for the treatment of achalasia: a prospective single e continued clinical follow-up is essential to assess the long- center study. Am J Gastroenterol 2012;107:411 417. 9. Perretta S, Dallemagne B, Donatelli G, et al. Transoral endo- term risk of GER after POEM. scopic esophageal myotomy based on esophageal function Weaknesses of our study include its retrospective design testing in a survival porcine model. Gastrointest Endosc and current lack of some objective evaluations and long- 2011;73:111e116. term outcomes. Not all patients underwent manometry, 10. Gutschow CA, Ho¨lscher AH. Myotomy for esophageal acha- timed-barium swallow, and 24-hour pH testing due to lasia - laparoscopic versus peroral endoscopic approach. Endos- copy 2010;42:318e319. the medical condition of our hospital (too many patients, 11. Wang L, Li YM. Recurrent achalasia treated with Heller myot- shortage of equipment and professionals), loss to follow- omy: a review of the literature. World J Gastroenterol 2008; up, discomfort related to the procedure, and other 14:7122e7126. personal reasons. We will continue to accrue data and 12. Petersen RP, Pellegrini CA. Revisional surgery after Heller are in the process of obtaining comprehensive objective myotomy for esophageal achalasia. Surg Laparosc Endosc Per- cutan Tech 2010;20:321e325. physiology studies, as well as long-term follow-up. 13. Pandolfino JE, Kwiatek MA, Nealis T, et al. Achalasia: a new clinically relevant classification by high-resolution manometry. CONCLUSIONS Gastroenterology 2008;135:1526e1533. 14. Jones R, Junghard O, Dent J, et al. Development of the In summary, according to our short-term results of this GerdQ, a tool for the diagnosis and management of gastro- large study, symptom relief and manometry outcomes oesophageal reflux disease in primary care. Aliment Pharmacol were comparable between patients undergoing full- Ther 2009;30:1030e1038. thickness and circular muscle myotomy. Full-thickness 15. Bonin EA, Moran E, Bingener J, et al. A comparative study of endoscopic full-thickness and partial-thickness myotomy using myotomy significantly reduced the procedure time but submucosal endoscopy with mucosal safety flap (SEMF) tech- did not increase the procedure-related adverse events or nique. Surg Endosc 2012;26:1751e1758. clinical reflux complications. 16. de Oliveira JM, Birgisson S, Doinoff C, et al. Timed barium swallow: a simple technique for evaluating esophageal Author Contributions emptying in patients with achalasia. AJR Am J Roentgenol 1997;169:473e479. Study conception and design: Li, Zhou 17. Rohof WO, Hirsch DP, Kessing BF, et al. Efficacy of treat- Acquisition of data: Li, Chen, Zhou, Yao, Xu, Hu, ment for patients with achalasia depends on the distensibility Cai, Zhang, Qin, Ren of the esophagogastric junction. Gastroenterology 2012;143: Analysis and interpretation of data: Li, Zhou 328e335. Drafting of manuscript: Li, Chen, Zhou 18. Harris JA, Gallo CD, Brummett DM, et al. Extra-abdominal pneumodissection after laparoscopic antireflux surgery. Am Critical revision: Zhou, Yao Surg 2001;67:885e889. 19. Willingham FF, Gee DW, Lauwers GY, et al. Natural orifice REFERENCES transesophageal mediastinoscopy and thoracoscopy. Surg e 1. Eckardt AJ, Eckardt VF. Treatment and surveillance strategies Endosc 2008;22:1042 1047. in achalasia: an update. Nat Rev Gastroenterol Hepatol 2011; 20. Rattner DW, Hawes R, Schwaitzberg S, et al. The Second 8:311e319. SAGES/ASGE White Paper on natural orifice transluminal 2. Campos GM, Vittinghoff E, Rabl C, et al. Endoscopic and endoscopic surgery: 5 years of progress. Surg Endosc 2011; e surgical treatments for achalasia: a systematic review and 25:2441 2448. meta-analysis. Ann Surg 2009;249:45e57. 21. Kurian AA, Dunst CM, Sharata A, et al. Peroral endoscopic 3. Inoue H, Minami H, Kobayashi Y, et al. Peroral endoscopic esophageal myotomy: defining the learning curve. Gastrointest e myotomy (POEM) for esophageal achalasia. Endoscopy Endosc 2013;77:719 725. 2010;42:265e271. 22. Pasricha PJ, Hawari R, Ahmed I, et al. Submucosal endoscopic 4. Swanstrom LL, Kurian A, Dunst CM, et al. Long-term esophageal myotomy: a novel experimental approach for the e outcomes of an endoscopic myotomy for achalasia: the treatment of achalasia. Endoscopy 2007;39:761 764. POEM procedure. Ann Surg 2012;256:659e667. 23. Khashab MA, Sharaiha RZ, Saxena P, et al. Novel technique 5. Chiu PW, Wu JC, Teoh AY, et al. Peroral endoscopic myot- of auto-tunneling during peroral endoscopic myotomy (with e omy for treatment of achalasia: from bench to bedside (with video). Gastrointest Endosc 2013;77:119 122. video). Gastrointest Endosc 2013;77:29e38. 24. Patti MG, Molena D, Fisichella PM, et al. Laparoscopic Heller 6. Hungness ES, Teitelbaum EN, Santos BF, et al. Comparison of myotomy and Dor fundoplication for achalasia: analysis of e perioperative outcomes between Peroral Esophageal Myotomy successes and failures. Arch Surg 2001;136:870 875; discus- e (POEM) and laparoscopic Heller myotomy. J Gastrointest sion 875 877. Surg 2013;17:228e235. 25. Richards WO, Torquati A, Holzman MD, et al. Heller myot- 7. Zhou PH, Li QL, Yao LQ, et al. Peroral endoscopic remyot- omy versus Heller myotomy with Dor fundoplication for acha- omy for failed Heller myotomy: a prospective single-center lasia: a prospective randomized double-blind clinical trial. Ann e e study. Endoscopy 2013;45:161e166. Surg 2004;240:405 412; discussion 412 415.