- Email: [email protected]
Outcomes of per-oral endoscopic myotomy for treatment of esophageal achalasia with a median follow-up of 49 months
ORIGINAL ARTICLE
Outcomes of per-oral endoscopic myotomy for treatment of esophageal achalasia with a median follow-up of 49 months Quan-Lin Li, MD,1,* Qiu-Ning Wu, MD,1,* Xiao-Cen Zhang, MD,1,* Mei-Dong Xu, MD, PhD,1 Wei Zhang, PhD,2 Shi-Yao Chen, MD, PhD,1 Yun-Shi Zhong, MD, PhD,1 Yi-Qun Zhang, MD, PhD,1 Wei-Feng Chen, MD, PhD,1 Wen-Zheng Qin, MD, PhD,1 Jian-Wei Hu, MD,1 Ming-Yan Cai, MD, PhD,1 Li-Qing Yao, MD,1 Ping-Hong Zhou, MD, PhD1 Shanghai, China
Background and Aims: Per-oral endoscopic myotomy (POEM) has received wide acceptance as a highly effective and safe treatment for esophageal achalasia. Short-term and small-scale studies are ample, but long-term large-scale studies are few. The aim of the study was to systematically analyze our long-term results of POEM, with particular emphasis on POEM failures and associated risk factors. Methods: In this single-center study, consecutive patients treated with POEM between August 2010 and December 2012 were included. The Kaplan-Meier survival function was used to estimate clinical success rate at each year. The Cox proportional hazards model was used to analyze risk factors related to recurrence. Results: A total of 564 patients were included. Major perioperative adverse events occurred in 36 patients (6.4%). After a median follow-up of 49 months (range, 3-68), the Eckardt score and lower esophageal sphincter (LES) pressure were significantly decreased (median Eckardt score, 8 to 2 [P < .05]; median LES pressure, 29.7 mm Hg to 11.9 mm Hg [P < .05]). Fifteen failures occurred within 3 months, 23 between 3 months and 3 years, and 10 after 3 years. The estimated clinical success rates at 1, 2, 3, 4, and 5 years were 94.2%, 92.2%, 91.1%, 88.6%, and 87.1%, respectively. Multivariate Cox regression revealed long disease duration (10 years) and history of prior interventions to be risk factors for recurrence. Clinical reflux occurred in 37.3% of patients (155/ 416). Conclusions: POEM is a highly safe and effective treatment for esophageal achalasia with favorable long-term outcomes. (Gastrointest Endosc 2017;-:1-8.)
Per-oral endoscopic myotomy (POEM) in its current form1 is the fruition of a long succession of trials2-4 exploring the submucosal space as a means to access body cavities. It is also probably the most successful member in the family of
natural orifice transluminal endoscopic surgery today, drawing phenomenal attention within its short history.5-12 Conceptually based on the surgical prototype Heller myotomy, POEM offers a less-invasive method of breaking
Abbreviations: AE, adverse event; CI, confidence interval; IQR, interquartile range; LES, lower esophageal sphincter; POEM, per-oral endoscopic myotomy.
0016-5107/$36.00 https://doi.org/10.1016/j.gie.2017.10.031
DISCLOSURE: All authors disclosed no financial relationships relevant to this publication. Research support for this study was provided by grants from the National Natural Science Foundation of China nos. 81302098 (Li QL), 81370588 (Xu MD), 81470811 (Zhou PH), 81401930 (Chen WF), 81570595 (Xu MD), and 81670483 (Zhou PH); Major Project of Shanghai Municipal Science and Technology Committee nos. 16411950400 (Zhou PH), 14441901500 (Xu MD), and 15JC1490300 (Xu MD); Chen Guang Program of Shanghai Municipal Education Commission no. 15CG04 (Li QL); and Outstanding Young Doctor Training Project of Shanghai Municipal Commission of Health and Family Planning, no. 2017YQ026 (Li QL).
Current affiliations: Endoscopy Center and Endoscopy Research Institute, Zhongshan Hospital (1), Department of Biostatistics, Shanghai Medical College (2), Fudan University, Shanghai, China.
*Drs Li, Wu, and Zhang contributed equally to this article.
Received May 25, 2017. Accepted October 13, 2017.
Present address: Qiu-Ning Wu: Gastrointestinal Endoscopy Center, The Sixth Affiliated Hospital of Sun Yat-Sen University, Guangzhou, China; Xiao-Cen Zhang: Mount Sinai St. Luke’s-West Hospital, New York, USA. Reprint requests: Ping-Hong Zhou, MD, PhD, Endoscopy Center and Endoscopy Research Institute, Zhongshan Hospital, Fudan University, 180 FengLin Road, Shanghai, 200032, P. R. China. If you would like to chat with an author of this article, you may contact Dr Zhou at [email protected].
Copyright ª 2017 by the American Society for Gastrointestinal Endoscopy
www.giejournal.org
Volume
-,
No.
-
: 2017 GASTROINTESTINAL ENDOSCOPY 1
long-term results of POEM treatment for esophageal achalasia
Li et al
the pressurized zone of the esophageal gastric junction. Because both POEM and Heller myotomy render basically the same result anatomically (myotomy across the esophageal gastric junction, leaving esophageal mucosa intact), it is reasonable to infer they would lead to a similar clinical response. Indeed, the effect of POEM has been proven to be comparable with Heller myotomy by small-scale short-term studies.13-16 Nevertheless, concerns exist regarding POEM’s durability, whereas Heller myotomy has stood the test of time.17 Inoue et al,18 who developed and named POEM, reported their favorable 3-year results in 2015. As 1 of the earliest centers to perform POEM and the highest in volume, we herein present our 5-year results regarding the long-term outcomes of POEM.
METHODS Patients Patients were diagnosed with established methods: clinical symptoms, barium swallow, EGD, manometry, and/or chest CT scan. Exclusion criteria included coagulopathy and systemic disorders that precluded safe general anesthesia. The procedures and study were conducted in accordance with the Declaration of Helsinki and had approval from the Ethical Board of Zhongshan Hospital. Written informed consent was obtained from all patients before the procedure.
Outcome measurements The primary outcome of the study was the clinical success rate of POEM (Eckardt score 3) during follow-up. The secondary outcomes included procedure-related adverse events (AEs), lower esophageal sphincter (LES) pressure on manometry pre- and post-POEM, reflux symptoms (grade 0, absent; grade 1, <2 days a week; grade 2, 2-4 days a week; grade 3, >4 days a week),19 reflux esophagitis on EGD, and procedure parameters such as operation time, length of hospital stay, and myotomy length. Baseline and postmyotomy LES pressures were recorded using a high-resolution manometry system (Sierra Scientific Instruments Inc, Los Angeles, Calif) as previously described.20
POEM procedures The POEM procedure we used was largely the same as the original protocol raised by Inoue et al1 and has been reported by our previous publication.21 Four major steps were involved: submucosal injection, submucosal tunneling, myotomy, and closure of the mucosotomy. Several important technical modifications22 were put forward by our group and have become our standard practice over time: myotomy at the posterior (5-6 o’clock) rather than anterior esophageal wall, (2) full thickness rather than selective circular muscle myotomy, and (3) push and pull myotomy: in the esophagus anchor the knife tip at the muscle fibers from the extraluminal side and PUSH the scope, and at the cardia 2 GASTROINTESTINAL ENDOSCOPY Volume
-,
No.
-
: 2017
lay the knife on top of the muscle fibers intraluminally and PULL the scope.23 We were also among the first to apply a water jet–assisted knife in POEM.24 One technical drawback was that the CO2 insufflator did not become readily available in our center until late 2011, and this had an effect on our gas-related AEs.
Adverse events Major perioperative AEs were defined as conditions that resulted in vital-sign instability, intensive care unit stay, hospital readmission, conversion to open surgery, invasive postoperative procedures, blood transfusion, or hospitalization >5 days because of functional impairment of the patient.25 Minor perioperative AEs were defined as AEs that called for clinical interventions but did not qualify for major AE, including bouts of intraoperative subcutaneous emphysema/pneumoperitoneum that required needle decompression only and prolonged intraoperative bleeding (>200 mL) that did not require transfusion. Accidental mucosal injury was recorded given its call for special management (eg, clipping, prolonged fasting, and/or nasogastric tube placement) but not graded major or minor AEs per se. Long-term AEs were associated with acid reflux, including gastroesophageal reflux symptoms, strictures, and Barrett’s esophagus. Mucosal edema and mucosal injury, 2 risk factors for perioperative AEs, were graded and categorized.25 Mucosal edema was characterized by mucosa surface texture abnormality (cobblestone sign, milky color of the injection wheal, jellylike adhesions, etc) and/or increased cutting-edge thickness (inward folding, high tension during clipping, etc.), and graded in severity accordingly.25 Mucosal injuries were categorized as 1 of 2 types based on difficulty of repair. Type I injuries were easily repairable injuries (small, whitish, linear/dotted, or mucosal color change), whereas type II injuries were difficult-to-repair injuries (big, scorched, round/unevenly bordered, and often related to submucosal adhesion).25
Follow-up Patients were scheduled to follow-up at the center 1 month, 3 months, 6 months, and 1 year postoperation and yearly afterward, during which symptom assessment, physical examination, and objective tests including EGD and barium esophagram were performed. High-resolution manometrywas recommended to be done at least once postoperatively. Clinical response was evaluated using the Eckardt score.26 Post-POEM Eckardt score 3 was considered the benchmark for treatment success. Those with Eckardt scores >3 were grouped according to time of failure: within 3 months, nonresponders; between 3 months and 3 years, early recurrence; and after 3 years, late recurrence. Barium swallow was performed to document treatment results objectively. EGD was recommended because it provided both treatment outcome assessment and esophageal cancer surveillance.26 Manometry, although strongly advised, tended to be poorly welcomed by the patients www.giejournal.org
Li et al
because it is uncomfortable and available in only a few number of centers. For patients from faraway provinces and unwilling to come back for follow-up, detailed telephone interviews, including symptoms and treatments or tests received at outside hospitals, were conducted. The latest database checkup and phone interviews were carried out in October 2016.
Statistical analysis Categorical variables were reported as counts and percentages. Continuous data were reported as median with interquartile range (IQR) and categorized according to commonly used clinical criteria when necessary. Data were compared using (paired) the Student t test or Wilcoxon signed-rank test as appropriate. Long-term clinical success rate and risk of recurrence were estimated and graphed using the Kaplan-Meier survivor function and Nelson-Aalen cumulative hazard function, respectively. The Cox proportional hazards model was used to evaluate independent risk factors for clinical recurrence. Sensitivity analysis based on multiple imputation was conducted to check for influence of lost-to-follow-up.27 Statistical significance level was set at .05, 2-sided. All analyses were made in SAS (version 9.3; SAS Institute, Cary, NC).
long-term results of POEM treatment for esophageal achalasia
TABLE 1. Patient characteristics and procedural parameters (n [ 564) Characteristic
Value
Median age, y [IQR] (range) Gender, M/F
274/290
Median disease duration, y [IQR] (range) Sigmoid esophagus
Patient characteristics and procedure parameters A total of 567 patients received POEM at Zhongshan Hospital, Fudan University, Shanghai, China, between August 2010 and December 2012. Three POEM procedures were aborted, 2 because of severe submucosal fibrosis and 1 because of atrial fibrillation related to the electrosurgical knife. Therefore, 564 patients were analyzed in this study. The median age was 38 years, and the median disease duration was 4 years. Forty-eight patients (8.5%) had sigmoid esophagus, and 193 patients (34.2%) had received prior interventional treatment, including 20 with prior Heller myotomy and 3 with prior POEM. The median prePOEM Eckardt score was 8, and the median pre-POEM LES pressure was 29.7 mm Hg (Table 1). Regarding the procedure, the median operation time was 45 minutes. Full-thickness myotomy was performed in 352 patients (63.9%). Mucosal edema was present in 54 patients (9.6%), including 42 (7.4%) grade I and 12 (2.1%) grade II edema (Table 1).
Perioperative AEs Mucosal injuries occurred in 93 patients (16.5%), including 78 (13.8%) type I injuries and 15 (2.7%) type II injuries. Forty-eight patients (8.5%) required nasogastric tube placement at the end of the procedure. Conditions that called for nasogastric tube placement included insecure mucosal closure, severe bleeding, and a long procedwww.giejournal.org
4 [2, 9] (.1-60) 48 (8.5)
Prior treatment
193 (34.2)
Balloon/bougie dilation
134 (23.8)
Botulinum toxin injection
30 (5.3)
Esophageal stent
34 (6.0)
Heller myotomy
20 (3.6)
POEM
3 (.5)
Median pre-POEM Eckardt score [IQR] (range) Median pre-POEM LES pressure, mm Hg [IQR] (range) Median operation time, min [IQR] (range)
8 [7, 9] (4-12) 29.7 [21.4, 39.8] (5.1-70.9) 45 [32, 66] (15-202)
CO2 insufflation
340 (60.3)
Mucosa edema
54 (9.6)
Grade I
42 (7.4)
Grade II
RESULTS
38 [28, 49] (6-77)
12 (2.1)
Submucosal fibrosis
44 (7.8)
Median tunneling length, cm [IQR] (range)
13 [13, 13] (7-23)
Above EGJ
10 [10, 10] (4-20)
Below EGJ
3 [3, 3] (0-6)
Full-thickness myotomy Median myotomy length, cm [IQR] (range) Above EGJ
352 (63.9) 10 [10, 10] (5-17) 8 [8, 8] (3-13)
Below EGJ
2 [2, 2] (0-6)
Tunneling knife, TK/Hybrid/HK/others
370/106/75/6
Myotomy knife, TK/Hybrid/HK/others
437/91/23/6
Median length of stay, days [IQR] (range)
3 [2, 4] (1-29)
Values are n (%), unless otherwise noted. IQR, Interquartile range; POEM, per-oral endoscopic myotomy; LES, lower esophageal sphincter; EGJ, esophageal gastric junction; TK, triangle tip knife (KD-640L; Olympus, Japan); Hybrid, HybridKnife (Erbe, Tübingen, Germany); HK, hook knife (KD-620LR; Olympus, Tokyo, Japan).
ure time. Other minor AEs included 3 patients with estimated blood loss >200 mL (stabilized after endoscopic hemostasis and fluid resuscitation), 1 patient with subcutaneous emphysema, and 1 patient with pneumoperitoneum (both relieved by needle decompression). Major perioperative AEs occurred in 36 patients (6.4%). Three patients had delayed mucosal barrier failure (2 tunnel entrance dehiscence and 1 delayed mucosal flap perforation). Three had delayed bleeding. Six had hydrothorax. Twenty-one had pneumothorax, all of whom had received Volume
-,
No.
-
: 2017 GASTROINTESTINAL ENDOSCOPY 3
long-term results of POEM treatment for esophageal achalasia
Li et al
TABLE 2. Perioperative AEs (n [ 564) Adverse event
Value
Mucosal injury
93 (16.5)
Type I
78 (13.8)
Type II
15 (2.7)
Nasogastric tube inserted at end of POEM
48 (8.5)
Major perioperative AEs*
36 (6.4)
Delayed mucosal barrier failure
3 (.5)
Delayed bleeding requiring interventions and/or transfusion
3 (.5)
Hydrothorax requiring drainage
6 (1.1)
Pneumothorax requiring drainage
21 (3.7)
Othery
3 (.5)
Minor perioperative AEs
5 (.9)
Intraoperative bleeding >200 mL, no transfusion or special management
3 (.5)
Intraoperative subcutaneous emphysema requiring needle decompression
1 (.2)
Intraoperative pneumoperitoneum requiring needle decompression
1 (.2)
Values are n (%). AE, Adverse event; POEM, per-oral endoscopic myotomy. *When a patient had inter-related major and minor AEs, only the major AE is listed. These included 3 intraoperative and delayed bleeding, 1 intraoperative subcutaneous emphysema and pneumothorax requiring continuous drainage, and 2 intraoperative pneumoperitoneum and pneumothorax requiring continuous drainage. yOne 40-year-old man had status epilepticus on postoperative day 2 and required intensive care unit stay. One 45-year-old man had delayed extubation and intensive care unit stay because of chronic lung infection possibly complicated by pharyngeal edema. One 61-year-old woman had postoperative lung infection and transient hypoxemia.
air rather than CO2 insufflation. Three patients had other miscellaneous major AEs (Table 2).25 After initiation of CO2 insufflation, the AE rate dropped to 2.4%.
Long-term symptomatic relief Of the 564 patients, 420 (74.5%) had follow-up results for a median of 49 months (range, 3-67); 357 patients (63.3%) were followed-up for 3 years and 237 (42.0%) for 4 years. The median Eckardt score for all patients at their latest follow-up was 2 (IQR [1, 3]; range, 0-10), which was significantly lower than their corresponding pre-POEM score (median, 8; IQR [7, 9]; range, 4-12; P < .05). When the Kaplan-Meier survival function was used, the estimated clinical success rates at 1, 2, 3, 4, and 5 years were 94.2% (95% confidence interval [CI], 91.6%-96.0%), 92.2% (95% CI, 89.2%-94.4%), 91.1% (95% CI, 87.9%-93.5%), 88.6% (95% CI, 85.0%-91.4%), and 87.1% (95% CI, 82.8%-90.4%), respectively (Table 3, Fig. 1).
Manometry outcomes A total of 265 patients (47.0%) underwent per-protocol high-resolution manometry postoperatively. Other patients did not receive follow-up manometry because of the difficulty in traveling to major medical centers capable of motility 4 GASTROINTESTINAL ENDOSCOPY Volume
-,
No.
-
: 2017
tests, discomfort related to the test, and other personal reasons. The median LES pressure at follow-up was 11.9 mm Hg (IQR [9.2, 15.6]; range, 4.0-24.6), which was significantly lower than their corresponding pre-POEM value (median, 29.7 mm Hg; IQR [21.4, 39.8]; range 5.1-70.9; P < .05).
Gastroesophageal reflux and other long-term AEs Gastroesophageal reflux was evaluated in 416 patients. Of the 375 patients who were evaluated for heartburn and regurgitation, 137 patients (36.5%) reported presence of reflux symptoms, with 126 (92.0%) grade 1 and 11 (8.0%) grades 2 to 3 by frequency. Of the 341 patients who received followup EGD, 58 patients (17.0%) showed signs of reflux esophagitis. Forty-three patients were evaluated for severity of esophagitis, of whom 26 (60.5%) had mild esophagitis (Los Angeles classification grade A), 14 (32.6%) had moderate esophagitis (grades B-C), and only 3 (7.0%) had severe esophagitis (grade D). When combining both symptomatic reflux and reflux esophagitis, clinical reflux occurred in 155 patients (37.3%). There was no confirmed reflux-related post-POEM stricture that required interventional treatments (Supplementary Table 1, available at www.giejournal.org). Five known deaths occurred during follow-up. One 48-year-old woman (the fourth patient in the whole series) had persistent symptoms and died of cachexia 15 months after POEM. The other 4 patients died of acute cholangitis, cholangiocarcinoma, lymphoma, and lung cancer, respectively. The 2 patients who died of lymphoma and lung cancer had negative noncontrast-enhanced chest CT during the perioperative period.
POEM failures Of the 420 patients with follow-up results, 48 (11.4%) failures occurred. Fifteen were nonresponders (failure within 3 months), 23 were early recurrence (between 3 months and 3 years), and 10 were late recurrence (after 3 years). The Nelson-Aalen estimated cumulative hazards for recurrence at 1, 2, 3, 4, and 5 years were 5.9% (95% CI, 4.0%-8.7%), 8.0% (95% CI, 5.7%-11.3%), 9.3% (95% CI, 6.7%-12.8%), 12.0% (95% CI, 8.9%-16.2%), and 13.7% (95% CI, 10.0%18.7%), respectively (Fig. 2, Table 3). The median age of these patients was 47 years, and the median disease duration was 6.8 years. Eight (16.7%) had sigmoid esophagus, and 24 (50.0%) had a history of prior interventional treatment. Their median post-POEM Eckardt score was 5 (IQR [4, 6]; range, 4-10), which was significantly lower than the corresponding pre-POEM value (median, 8; IQR [7, 9]; range, 4-11; P < .05). There was no significant difference in the pre-POEM Eckardt score between patients with and without recurrence (P Z .88). Clinical reflux was evaluated in 46 patients, of whom 23 (50.0%) were positive, including 47.7% (21/44) with symptomatic reflux and 20.5% (8/39) with reflux esophagitis on EGD (Table 4). When the multivariate Cox proportional hazards model was used, disease duration 10 years (hazard ratio, 2.45; 95% CI, www.giejournal.org
Li et al
long-term results of POEM treatment for esophageal achalasia
TABLE 3. Estimated clinical success (Eckardt score ≤3) and cumulative recurrence
No. at risk*
No. of recurrences within period
End of follow-upy within period
Estimated clinical success rate* (95% CI)
Estimated cumulative recurrence* (95% CI)
376
354
27
183
.942 (.916-.960)
.059 (.040-.087)
358
330
7
17
.922 (.892-.944)
.080 (.057-.113)
2-3
357
323
4
3
.911 (.879-.935)
.093 (.067-.128)
3-4
237
215
8
100
.886 (.850-.914)
.120 (.089-.162)
4-5
48
47
2
166
.871 (.828-.904)
.137 (.100-.187)
No. of patients under follow-up*
0-1 1-2
Time, year
*Patients under follow-up, at risk, estimated clinical success, and cumulative recurrence at the end of the period. Success rate estimated by Kaplan-Meier survival function and cumulative recurrence by Nelson-Aalen function. yThe patient recruitment period was August 2010 to December 2012, and the last round of data collection lasted throughout 2016. As a result, many patients receiving their POEM during 2012 have not yet been contacted at 4/5 years and are not loss-to-follow-up.
DISCUSSION Introduced as a minimally invasive alternative to the long-accepted criterion standard, Heller myotomy, POEM needs to undergo comprehensive assessments regarding its safety and effectiveness before it is truly accepted. In www.giejournal.org
1 .75
95% CI Survivor function
.5
A total of 144 patients were lost to follow-up, most of whom (137, 95.1%) were because of changed or absent contact information; 7 (4.9%) declined phone follow-up. Baseline characteristics, occurrence of major perioperative AEs, and length of hospital stay were comparable between patients with and without follow-up except that those without follow-up were more likely to be male and have a shorter disease duration (median disease duration, 3 years vs 4 years in those with follow-up; P Z .03) (Supplementary Table 2, available online at www.giejournal.org), indicating a possible higher willingness to follow-up in female patients and patients with longer disease duration. Because missing observations might lead to biased parameter estimates, sensitivity analysis by multiple imputation was applied to the Cox regression model to test for stability of results. Under the assumption that independent factors were missing at random and had multivariate normal distribution, multiple imputations for postoperative Eckardt score and time-to-event were performed 100 times, and the results from 100 Cox regression models were combined. Result showed that history of prior interventions increased the risk of recurrence (hazard ratio, 1.20; 95% CI, 1.091.32; P < .01). Meanwhile, although long disease duration (10 years) seemed to have a more prominent hazard ratio, the results after multiple imputations were not statistically significant (hazard ratio, 2.10; 95% CI, .99-4.44; P Z .05). However, both hazard ratio estimates were close to those before multiple imputations, indicating a nonsignificant impact of missing values on risk factor analysis.
.25
Lost to follow-up and sensitivity analysis
Kaplan-Meier survival estimate for clinical success rate
0
1.35-4.46) and prior interventions (hazard ratio, 1.12; 95% CI, 1.02-1.23) were independently related to a higher chance of recurrence (Table 5).
20
0
40
60
80
Follow-up time (months)
Figure 1. Kaplan-Meier survival estimate for clinical success rate. CI, Confidence interval.
spite of the favorable results reported by several highquality studies,18,28-33 large cohorts that can give a more precise outlook on long-term outcomes are few. Extended follow-up shows decreased effectiveness over time for all types of achalasia treatments, including Heller myotomy, which is considered the most enduring.34-37 Reasons for recurrence include avoidable causes such as inadequate myotomy, tight fundoplication and strictures caused by reflux esophagitis, and unavoidable causes such as fibrosis, progression of disease, and development of megaesophagus.36-39 As forerunners in the field of POEM, we are by no means assuming that POEM is a permanent cure for all. Rather, our assumption is that POEM should have long-term outcomes comparable with Heller myotomy. This assumption has indeed been supported by nonrandomized studies.13,15,40 Two randomized controlled trials comparing POEM with Heller myotomy are also underway and are due to be complete in 2017 (NCT02138643) and 2019 (NCT01601678). Initial clinical success rates higher than 90% have been reported by all studies with a volume of over 100 (single or multicenter)18,28-33,41 (Supplementary Table 3, available online at www.giejournal.org), which is again shown by Volume
-,
No.
-
: 2017 GASTROINTESTINAL ENDOSCOPY 5
long-term results of POEM treatment for esophageal achalasia
Li et al
TABLE 4. Summary of patients with recurrence (n [ 48)
.2
Nelson-Aalen cumulative hazard estimate for recurrence
Characteristic
.15
Median age, y [IQR] (range) Gender, M/F
.1
Median disease duration, y [IQR] (range)
95% CI
0
.05
Cumulative hazard
0
20
40 60 Follow-up time (months)
80
our study (clinical success rate of 96.4% at 3 months). Decreasing effectiveness over time has been indicated by several mid-term studies. Inoue et al18 reported a success rate of 91.3% at 2 months, 91.0% at 1 to 2 years, and 88.5% at 3 years in their first 500 patients. A multicenter international trial conducted by Werner et al31 reported a clinical success rate of 93.7% at 3 to 6 months, 88.5% at 12 to 18 months, and 77.5% at 24 to 41 months. Under the Kaplan-Meier survival function, our result showed a success rate of 94.2% at 1 year, 91.1% at 3 years, and 87.1% at 5 years. These results are comparable with that of Heller myotomy performed in state-of-the-art centers.36,37,42-44 According to our experience, incomplete myotomy is the most likely cause for treatment failure within 12 months, a viewpoint shared by Heller myotomy experts.38,44 In the adult Chinese population we generally try to guarantee a myotomy of at least 8 cm above the esophageal gastric junction and 2 cm below. In addition, we prefer full-thickness over partialthickness myotomy because it is in accordance with the Heller myotomy principle and is faster while no more likely to cause procedure-related AEs.23 Although a study has shown full-thickness myotomy to be predictive of clinically relevant gastroesophageal reflux,41 this should not necessarily be viewed negatively because endoscopic reflux signs have been shown to be an independent predictor of treatment success.31 Under the Cox proportional hazards model, long disease duration (10 years) and prior interventional treatment history are identified as risk factors for POEM failures. Considering that all existing achalasia treatments aim to lower the LES pressure, it is not surprising to see that patients who have failed 1 treatment are also more likely to fail a following treatment, regardless of which is used. For instance, it is well documented that Heller myotomy tends to have a poorer response in patients with advanced disease and treatment history.44,45 Meanwhile, robust evidence has shown the benefits of POEM in these patients.46-49 In addition, our study showed that even in patients who are considered failures (Eckardt score >3), there was still a significant decrease in the score compared 6 GASTROINTESTINAL ENDOSCOPY Volume
-,
No.
-
: 2017
47 [34, 53] (23-71) 18/30 6.8 [3, 15] (.3-40)
Sigmoid esophagus
8 (16.7)
Prior treatment
24 (50.0)
Median operation time, min [IQR] (range) Full-thickness myotomy
Figure 2. Nelson-Aalen cumulative hazard estimate for recurrence. CI, Confidence interval.
Value
Median myotomy length, cm [IQR] (range)
45 [34, 60] (19-120) 33 (68.8) 10 [10, 11] (6-16)
Above EGJ
8 [8, 9] (3-13)
Below EGJ
2 [2, 3] (2-5)
Clinical reflux Reflux symptoms, yes/total Grade 1 Grades 2-3 Reflux esophagitis on EGD, yes/total
23/46 (50) 21/44 (47.7) 19 (90.5) 2 (9.5) 8/39 (20.5)
Median pre-POEM Eckardt score [IQR] (range)
8 [7, 9] (4-11)
Median post-POEM Eckardt score [IQR] (range)
5 [4, 6] (4-10)
Values are n (%), unless otherwise noted. IQR, Interquartile range; EGJ, esophageal gastric junction; POEM, per-oral endoscopic myotomy.
with baseline. Considering the above, it is reasonable to offer POEM as first-line therapy to these patients.12 Post-POEM reflux and its management is another interesting topic. Long-term acid exposure is 1 of the major arguments against POEM because currently it cannot be paired with antireflux procedures and Heller myotomy can. Previous studies have reported post-POEM symptomatic reflux rate varying from 8.6% to 37% and endoscopic reflux esophagitis rate from to 14.8% to 66% (Supplementary Table 3). Symptoms and reflux esophagitis do not always overlap, and it is interesting to note that many centers reported a higher rate of reflux esophagitis than reflux symptoms. This is not the case for normal GERD, in which most patients with heartburn and regurgitation will not have erosions under endoscopy.50 Evaluating reflux in postPOEM patients is difficult for various reasons, including the overlap of symptoms between reflux and undertreated achalasia (chest pain, regurgitation, dysphagia, and even heartburn); stasis, which may falsely prolong reflux time under pH testing; and ischemic/postsurgical changes in the esophageal gastric junction that can mimic erosive esophagitis. A better approach might be to look at the clinical consequences of reflux, in which case POEM seemed to have fared fairly well. In our series there was no confirmed POEM failure caused by reflux-related strictures. All patients www.giejournal.org
Li et al
long-term results of POEM treatment for esophageal achalasia
TABLE 5. Risk factor analysis for recurrence using Cox proportional hazards model Univariate Variable Age 60 y Male gender
Multivariate
Hazard ratio (95% CI)
P value
Hazard ratio (95% CI)
P value
1.20 (.51-2.83)
.67
NA
NS
.67 (.38-1.21)
.19
NA
NS
Disease duration 10 y
2.46 (1.39-4.37)
<.01
2.45 (1.35-4.46)
<.01
Sigmoid esophagus
2.36 (1.10-5.05)
.03
NA
NS
Any type of prior interventional treatments*
1.16 (1.05-1.27)
<.01
1.12 (1.02-1.23)
.02
Balloon/bougie dilation
1.14 (1.03-1.27)
.02
d
d
Botulinum toxin injection
1.22 (.61-2.43)
.57
d
d
Esophageal stent
1.14 (.61-2.10)
.68
d
d
Heller myotomy
2.94 (1.06-8.20)
.04
d
d
POEM
12.7 (3.06-52.5)
<.01
d
d
Pre-POEM Eckardt score 7
1.34 (.66-2.69)
.42
NA
NS
Full-thickness myotomy
1.40 (.76-2.60)
.28
NA
NS
Myotomy length above esophageal gastric junction 8 cm
.93 (.45-1.92)
.84
NA
NS
Clinical reflux
1.66 (.93-2.95)
.09
NA
NS
Symptomatic reflux
1.64 (.91-2.97)
.10
d
d
Reflux esophagitis on EGD
1.27 (.58-2.76)
.55
d
d
NA, Not assessed; NS, not significant; d, not included for analysis; POEM, per-oral endoscopic myotomy. *Defined as total number of prior interventions, including dilation, botulinum toxin injection, stent, Heller myotomy, and POEM.
with reflux symptoms and signs, including those with Los Angeles grade D esophagitis, were successfully managed with proton pump inhibitors, and no patient required rescue fundoplication. Our study is the largest single-center, long-term outcome study of POEM worldwide to date. Nevertheless, we also face challenges such as high loss-to-follow-up rate, poor patient compliance at diagnostic tests, and difficulties in accessing records from outside hospitals. These result in a lack of in-depth analysis of causes of POEM failures, especially regarding the role reflux played. The late initiation of CO2 insufflation also made the AE rate seem unrealistically high. Despite the limitations we believe this study brings important information regarding POEM given the size of the study, in-depth analysis, detailed recording, and long follow-up. In addition, because a longer disease history was found to be a risk factor for recurrence and was related to a higher rate of follow-up, our estimation on long-term success rate is likely a modest one. In conclusion, our large cohort showed excellent outcome of POEM after a medium follow-up of 49 months (estimated clinical success rate at 5 years, 87.1%). Even for those considered to have failed POEM (Eckardt score 4), there was often still a significant improvement of symptoms compared with their baselines. Long disease duration and interventional treatment history are related to recurrence of achalasia symptoms but do not deny POEM as first-line therapy in patients with advanced disease. Postoperative acid reflux is currently an underexplored area and deserves www.giejournal.org
further study to testify for the safety of POEM in the long-term.
REFERENCES 1. Inoue H, Minami H, Kobayashi Y, et al. Peroral endoscopic myotomy (POEM) for esophageal achalasia. Endoscopy 2010;42:265-71. 2. Sumiyama K, Gostout CJ, Rajan E, et al. Transgastric cholecystectomy: transgastric accessibility to the gallbladder improved with the SEMF method and a novel multibending therapeutic endoscope. Gastrointest Endosc 2007;65:1028-34. 3. Pasricha P, Hawari R, Ahmed I, et al. Submucosal endoscopic esophageal myotomy: a novel experimental approach for the treatment of achalasia. Endoscopy 2007;39:761-4. 4. Woodward TA, Jamil LH, Wallace MB. Natural orifice trans-luminal endoscopic surgery in the esophagus. Gastrointest Endosc Clin North Am 2010;20:123-38. 5. Barbieri LA, Hassan C, Rosati R, et al. Systematic review and metaanalysis: efficacy and safety of POEM for achalasia. United Eur Gastroenterol J 2015;3:325-34. 6. Wei M, Yang T, Yang X, et al. Peroral esophageal myotomy versus laparoscopic Heller’s myotomy for achalasia: a meta-analysis. J Laparoendosc Adv Surg Techn 2015;25:123-9. 7. Talukdar R, Inoue H, Reddy DN. Efficacy of peroral endoscopic myotomy (POEM) in the treatment of achalasia: a systematic review and meta-analysis. Surg Endosc 2015;29:3030-46. 8. Bechara R, Ikeda H, Inoue H. Peroral endoscopic myotomy: an evolving treatment for achalasia. Nat Rev Gastroenterol Hepatol 2015;12:410-26. 9. Crespin OM, Liu LWC, Parmar A, et al. Safety and efficacy of POEM for treatment of achalasia: a systematic review of the literature. Surg Endosc 2017;31:2187-201. 10. Akintoye E, Kumar N, Obaitan I, et al. Peroral endoscopic myotomy: a meta-analysis. Endoscopy 2016;48:1059-68.
Volume
-,
No.
-
: 2017 GASTROINTESTINAL ENDOSCOPY 7
long-term results of POEM treatment for esophageal achalasia
Li et al
11. Patel K, Abbassi-Ghadi N, Markar S, et al. Peroral endoscopic myotomy for the treatment of esophageal achalasia: systematic review and pooled analysis. Dis Esoph 2016;29:807-19. 12. Stavropoulos SN, Friedel D, Modayil R, et al. Diagnosis and management of esophageal achalasia. BMJ 2016;354:i2785. 13. Teitelbaum EN, Rajeswaran S, Zhang R, et al. Peroral esophageal myotomy (POEM) and laparoscopic Heller myotomy produce a similar short-term anatomic and functional effect. Surgery 2013;154:885-92. 14. Ujiki MB, Yetasook AK, Zapf M, et al. Peroral endoscopic myotomy: a short-term comparison with the standard laparoscopic approach. Surgery 2013;154:893-900. 15. Bhayani NH, Kurian AA, Dunst CM, et al. A comparative study on comprehensive, objective outcomes of laparoscopic Heller myotomy with per-oral endoscopic myotomy (POEM) for achalasia. Ann Surg 2014;259:1098-103. 16. Schneider AM, Louie BE, Warren HF, et al. A matched comparison of per oral endoscopic myotomy to laparoscopic Heller myotomy in the treatment of achalasia. J Gastrointest Surg 2016;20: 1789-96. 17. Campos GM, Vittinghoff E, Rabl C, et al. Endoscopic and surgical treatments for achalasia. Ann Surg 2009;249:45-57. 18. Inoue H, Sato H, Ikeda H, et al. Per-oral endoscopic myotomy: a series of 500 patients. J Am Coll Surg 2015;221:256-64. 19. Vigneri S, Termini R, Leandro G, et al. A comparison of five maintenance therapies for reflux esophagitis. N Engl J Med 1995;333: 1106-10. 20. Pandolfino JE, Kwiatek MA, Nealis T, et al. Achalasia: a new clinically relevant classification by high-resolution manometry. Gastroenterology 2008;135:1526-33. 21. Li Q, Yao L, Xu X, et al. Repeat peroral endoscopic myotomy: a salvage option for persistent/recurrent symptoms. Endoscopy 2016;48:134-40. 22. Li Q, Zhou P. Perspective on peroral endoscopic myotomy for achalasia: Zhongshan experience. Gut Liver 2015;9:152-8. 23. Li Q, Chen W, Zhou P, et al. Peroral endoscopic myotomy for the treatment of achalasia: a clinical comparative study of endoscopic fullthickness and circular muscle myotomy. J Am Coll Surg 2013;217: 442-51. 24. Cai M, Zhou P, Yao L, et al. Peroral endoscopic myotomy for idiopathic achalasia: randomized comparison of water-jet assisted versus conventional dissection technique. Surg Endosc 2014;28:1158-65. 25. Zhang X, Li Q, Xu M, et al. Major perioperative adverse events of peroral endoscopic myotomy: a systematic 5-year analysis. Endoscopy 2016;48:967-78. 26. Eckardt AJ, Eckardt VF. Treatment and surveillance strategies in achalasia: an update. Nat Rev Gastroenterol Hepatol 2011;8:311-9. 27. Rubin DB. Multiple imputation after 18þ years. J Am Stat Assoc 1996;91:473-89. 28. Sharata AM, Dunst CM, Pescarus R, et al. Peroral endoscopic myotomy (POEM) for esophageal primary motility disorders: analysis of 100 consecutive patients. J Gastrointest Surg 2015;19:161-70. 29. Ramchandani M, Reddy DN, Darisetty S, et al. Peroral endoscopic myotomy for achalasia cardia: treatment analysis and follow up of over 200 consecutive patients at a single center. Dig Endosc 2016;28:19-26. 30. Familiari P, Gigante G, Marchese M, et al. Peroral endoscopic myotomy for esophageal achalasia. Ann Surg 2016;263:82-7.
8 GASTROINTESTINAL ENDOSCOPY Volume
-,
No.
-
: 2017
31. Werner YB, Costamagna G, Swanström LL, et al. Clinical response to peroral endoscopic myotomy in patients with idiopathic achalasia at a minimum follow-up of 2 years. Gut 2016;65:899-906. 32. Shiwaku H, Inoue H, Yamashita K, et al. Peroral endoscopic myotomy for esophageal achalasia: outcomes of the first over 100 patients with short-term follow-up. Surg Endosc 2016;30:4817-26. 33. Ngamruengphong S, Inoue H, Chiu PW, et al. Long-term outcomes of per-oral endoscopic myotomy in patients with achalasia with a minimum follow-up of 2 years: an international multicenter study. Gastrointest Endosc 2016;85:927-33. 34. Zaninotto G, Annese V, Costantini M, et al. Randomized controlled trial of botulinum toxin versus laparoscopic Heller myotomy for esophageal achalasia. Ann Surg 2004;239:364-70. 35. Vela MF, Richter JE, Khandwala F, et al. The long-term efficacy of pneumatic dilatation and Heller myotomy for the treatment of achalasia. Clin Gastroenterol Hepatol 2006;4:580-7. 36. Ortiz A, de Haro LFM, Parrilla P, et al. Very long-term objective evaluation of Heller myotomy plus posterior partial fundoplication in patients with achalasia of the cardia. Ann Surg 2008;247:258-64. 37. Kilic A, Schuchert MJ, Pennathur A, et al. Long-term outcomes of laparoscopic Heller myotomy for achalasia. Surgery 2009;146:826-33. 38. Patti MG, Molena D, Fisichella PM, et al. Laparoscopic Heller myotomy and Dor fundoplication for achalasia: analysis of successes and failures. Arch Surg 2001;136:870-7. 39. Gockel I. Persistent and recurrent achalasia after Heller myotomy. Arch Surg 2007;142:1093. 40. Hungness ES, Teitelbaum EN, Santos BF, et al. Comparison of perioperative outcomes between peroral esophageal myotomy (POEM) and laparoscopic Heller myotomy. J Gastrointest Surg 2013;17:228-35. 41. Ren Y, Tang X, Chen Y, et al. Pre-treatment Eckardt score is a simple factor for predicting one-year peroral endoscopic myotomy failure in patients with achalasia. Surg Endosc 2017;31:3234-41. 42. Frantzides CT, Moore RE, Carlson MA, et al. Minimally invasive surgery for achalasia: a 10-year experience. J Gastrointest Surg 2004;8:18-23. 43. Bonatti H, Hinder RA, Klocker J, et al. Long-term results of laparoscopic Heller myotomy with partial fundoplication for the treatment of achalasia. Am J Surg 2005;190:883-7. 44. Zaninotto G, Costantini M, Rizzetto C, et al. Four hundred laparoscopic myotomies for esophageal achalasia. Ann Surg 2008;248:986-93. 45. Smith CD, Stival A, Howell DL, et al. Endoscopic therapy for achalasia before Heller myotomy results in worse outcomes than Heller myotomy alone. Ann Surg 2006;243:579-86. 46. Sharata A, Kurian AA, Dunst CM, et al. Peroral endoscopic myotomy (POEM) is safe and effective in the setting of prior endoscopic intervention. J Gastrointest Surg 2013;17:1188-92. 47. Hu J, Li Q, Zhou P, et al. Peroral endoscopic myotomy for advanced achalasia with sigmoid-shaped esophagus: long-term outcomes from a prospective, single-center study. Surg Endosc 2015;29:2841-50. 48. Orenstein SB, Raigani S, Wu YV, et al. Peroral endoscopic myotomy (POEM) leads to similar results in patients with and without prior endoscopic or surgical therapy. Surg Endosc 2015;29:1064-70. 49. Lv L, Liu J, Tan Y, et al. Peroral endoscopic full-thickness myotomy for the treatment of sigmoid-type achalasia. Eur J Gastroenterol Hepatol 2016;28:30-6. 50. Katz PO, Gerson LB, Vela MF. Guidelines for the diagnosis and management of gastroesophageal reflux disease. Am J Gastroenterol 2013;108: 308-28.
www.giejournal.org
Li et al
long-term results of POEM treatment for esophageal achalasia
SUPPLEMENTARY TABLE 1. Follow-up results on gastroesophageal reflux Result
Value
Clinical reflux, yes/total
155/416 (37.3)
Reflux symptoms, yes/total
137/375 (36.5)
Grade 1 Grades 2-3 Reflux esophagitis on EGD, yes/total*
126 (92.0) 11 (8.0) 58/341 (17.0)
Grade A
26 (60.5)
Grade B-C
14 (32.6)
Grade D
3 (7.0)
Values in parentheses are percents. Clinical reflux: either reflux symptoms of reflux esophagitis on EGD. Grading of reflux symptom frequency: Grade 1, <2 days a week; grade 2, 2-4 days a week; grade 3, >4 days a week. Grading of reflux esophagitis: Los Angeles Classification. *Fifteen patients were recorded to have esophagitis but not graded for severity.
www.giejournal.org
Volume
-,
No.
-
: 2017 GASTROINTESTINAL ENDOSCOPY 8.e1
long-term results of POEM treatment for esophageal achalasia
Li et al
SUPPLEMENTARY TABLE 2. Comparison of baseline characteristics and postoperative parameters between patients with and without follow-up Patients with follow-up (n [ 420)
Patients lost to follow-up (n [ 144)
P value
39 [29, 51] (6-77)
37 [26, 48] (12-76)
.12
193/227
81/63
.03
4 [2, 9] (.1-60)
3 [1, 7] (.1-40)
.03
36 (8.6)
12 (8.3)
1.00
151 (36.0)
42 (29.2)
.16
8 [6, 9] (4-12)
8 [7, 9] (5-12)
.53
260 (63.0)
92 (66.7)
.47
Median myotomy length above EGJ, cm [IQR] (range)
8 [8, 8] (3- 13)
8 [8, 8] (3-13)
.89
Median myotomy length below EGJ, cm [IQR] (range)
2 [2, 3] (0-6)
2 [2, 2] (0-5)
.08
Median age, y [IQR] (range) Gender, M/F Median disease duration, y [IQR] (range) Sigmoid esophagus Prior treatment Median pre-POEM Eckardt score [IQR] (range) Full-thickness myotomy
Major perioperative adverse events Median length of stay, days [IQR] (range)
28 (6.7)
8 (5.6)
.84
3 [2, 4] (1-29)
2 [2, 4] (1-29)
.40
Values are n (%), unless otherwise noted. IQR, Interquartile range; POEM, per-oral endoscopic myotomy; EGJ, esophageal gastric junction.
8.e2 GASTROINTESTINAL ENDOSCOPY Volume
-,
No.
-
: 2017
www.giejournal.org
Li et al
long-term results of POEM treatment for esophageal achalasia
SUPPLEMENTARY TABLE 3. Literature review outcomes of POEM in studies with a volume of over 100 (single-center, multicenter trials and multicenter pooled analyses) Inoue et al18
Sharata et al28
Ramchandani et al29
Familiari et al30
Werner et al31
Shiwaku et al32
Ngamruengphong et al33
Ren et al41
Country origin
Japan
USA
India
Italy
Multicenter, USA and Europe
Japan
Multicenter, USA, Europe, Asia
China
No. of patients
500
100
212
94
80
103
205
115
2 mo, 1-2 y, 3y
16 mo (mean)
6 mo, 1 y
11 mo (mean)
29 mo (mean)
3 mo
31 mo (median)
At 1-2 y: 77.3% (286/370) At 3 y: 58.1% (61/105)
100%
At 6 mo: 98.0% (149/152)
97.9% (92/94)
1 patient lost at 6 mo 1 patient lost at 18 mo
100%
100%
At 2 mo: 91.3% (386/423) At 1-2 y: 91% (260/286) At 3 y: 88.5% (54/61) (Eckardt score <2 or reduction >4)
92% (92/100)
94.5% (87/92)
At 3-6 mo: 93.7% (74/79) At 12-18 mo: 88.5% (69/78) At 24-41 mo: 77.5% (62/80)
99% (99/ 100)
Follow-up time Follow-up rate
Clinical success rate (Eckardt score 3)
AEs
At 6 mo: 94% (140/149) At 1 y: 92% (94/102)
3.2% (total)*
6% (total)y
No major AEs
No major AE
Reflux symptoms
At 2 mo: 16.8% (71/423) At 3 y: 21.3% (13/61)
Heartburn 8.6% (7/81), regurgitation 9.9% (8/81)
21.6% (22/102)
24.3% (17/73)
37% (27/73)
Reflux esophagitis
At 2 mo: 64.7% (268/414) At 3 y: 56.3% (9/16)
27.4% (20/73)
16.6% (14/84)
27.4% (20/73)
37.5% (37/72)
Positive 24-hour pH testing
d
38.2% (26/68) (DeMeester score >14.7)
d
d
93.9% (108/ 115)
93% (107/ 115)
8.2% (total)z
-
9% (9/100)
26.8% (55/197)
d
66% (66/ 100)
18% (26/144)
14.8% (17/115)
25.6% (22/86) (monitoring time with pH < 4)
37.5% (12/32) (DeMeester score >14.72)
-
1 readmission 10 AEs (mucosa (Clavien-Dindo ulcer) I –IIIa) No other major
53.4% (39/73) (total reflux time >5%)
At 6 mo: 98% (185/189) At 12 mo: 98% (142/144) At 24 mo: 91% (187/205)
1y
POEM, Per-oral endoscopic myotomy; AE, adverse event. *One pneumothorax with mediastinal emphysema, 1 severe bleeding, 8 mucosal injuries, 3 postoperative hematomas from submucosal bleeding, 1 inflammation in the lesser omentum, and 2 pleural effusions. yThree leaks diagnosed on routine esophagram, 1 postoperative intratunnel bleeding, 1 Ogilvie’s syndrome, and 1 prolonged intubation for CO2 retention. zEight inadvertent mucosotomies, 2 symptomatic pleural effusions, 2 aspiration pneumonia, 2 deep vein thrombosis and pulmonary embolism, 1 symptomatic pneumothorax, 1 delayed bleeding, 1 esophageal leak, and 1 mediastinitis.
www.giejournal.org
Volume
-,
No.
-
: 2017 GASTROINTESTINAL ENDOSCOPY 8.e3
Outcomes of per-oral endoscopic myotomy for treatment of esophageal achalasia with a median follow-up of 49 months Quan-Lin Li, MD,1,* Qiu-Ning Wu, MD,1,* Xiao-Cen Zhang, MD,1,* Mei-Dong Xu, MD, PhD,1 Wei Zhang, PhD,2 Shi-Yao Chen, MD, PhD,1 Yun-Shi Zhong, MD, PhD,1 Yi-Qun Zhang, MD, PhD,1 Wei-Feng Chen, MD, PhD,1 Wen-Zheng Qin, MD, PhD,1 Jian-Wei Hu, MD,1 Ming-Yan Cai, MD, PhD,1 Li-Qing Yao, MD,1 Ping-Hong Zhou, MD, PhD1 Shanghai, China
Background and Aims: Per-oral endoscopic myotomy (POEM) has received wide acceptance as a highly effective and safe treatment for esophageal achalasia. Short-term and small-scale studies are ample, but long-term large-scale studies are few. The aim of the study was to systematically analyze our long-term results of POEM, with particular emphasis on POEM failures and associated risk factors. Methods: In this single-center study, consecutive patients treated with POEM between August 2010 and December 2012 were included. The Kaplan-Meier survival function was used to estimate clinical success rate at each year. The Cox proportional hazards model was used to analyze risk factors related to recurrence. Results: A total of 564 patients were included. Major perioperative adverse events occurred in 36 patients (6.4%). After a median follow-up of 49 months (range, 3-68), the Eckardt score and lower esophageal sphincter (LES) pressure were significantly decreased (median Eckardt score, 8 to 2 [P < .05]; median LES pressure, 29.7 mm Hg to 11.9 mm Hg [P < .05]). Fifteen failures occurred within 3 months, 23 between 3 months and 3 years, and 10 after 3 years. The estimated clinical success rates at 1, 2, 3, 4, and 5 years were 94.2%, 92.2%, 91.1%, 88.6%, and 87.1%, respectively. Multivariate Cox regression revealed long disease duration (10 years) and history of prior interventions to be risk factors for recurrence. Clinical reflux occurred in 37.3% of patients (155/ 416). Conclusions: POEM is a highly safe and effective treatment for esophageal achalasia with favorable long-term outcomes. (Gastrointest Endosc 2017;-:1-8.)
Per-oral endoscopic myotomy (POEM) in its current form1 is the fruition of a long succession of trials2-4 exploring the submucosal space as a means to access body cavities. It is also probably the most successful member in the family of
natural orifice transluminal endoscopic surgery today, drawing phenomenal attention within its short history.5-12 Conceptually based on the surgical prototype Heller myotomy, POEM offers a less-invasive method of breaking
Abbreviations: AE, adverse event; CI, confidence interval; IQR, interquartile range; LES, lower esophageal sphincter; POEM, per-oral endoscopic myotomy.
0016-5107/$36.00 https://doi.org/10.1016/j.gie.2017.10.031
DISCLOSURE: All authors disclosed no financial relationships relevant to this publication. Research support for this study was provided by grants from the National Natural Science Foundation of China nos. 81302098 (Li QL), 81370588 (Xu MD), 81470811 (Zhou PH), 81401930 (Chen WF), 81570595 (Xu MD), and 81670483 (Zhou PH); Major Project of Shanghai Municipal Science and Technology Committee nos. 16411950400 (Zhou PH), 14441901500 (Xu MD), and 15JC1490300 (Xu MD); Chen Guang Program of Shanghai Municipal Education Commission no. 15CG04 (Li QL); and Outstanding Young Doctor Training Project of Shanghai Municipal Commission of Health and Family Planning, no. 2017YQ026 (Li QL).
Current affiliations: Endoscopy Center and Endoscopy Research Institute, Zhongshan Hospital (1), Department of Biostatistics, Shanghai Medical College (2), Fudan University, Shanghai, China.
*Drs Li, Wu, and Zhang contributed equally to this article.
Received May 25, 2017. Accepted October 13, 2017.
Present address: Qiu-Ning Wu: Gastrointestinal Endoscopy Center, The Sixth Affiliated Hospital of Sun Yat-Sen University, Guangzhou, China; Xiao-Cen Zhang: Mount Sinai St. Luke’s-West Hospital, New York, USA. Reprint requests: Ping-Hong Zhou, MD, PhD, Endoscopy Center and Endoscopy Research Institute, Zhongshan Hospital, Fudan University, 180 FengLin Road, Shanghai, 200032, P. R. China. If you would like to chat with an author of this article, you may contact Dr Zhou at [email protected].
Copyright ª 2017 by the American Society for Gastrointestinal Endoscopy
www.giejournal.org
Volume
-,
No.
-
: 2017 GASTROINTESTINAL ENDOSCOPY 1
long-term results of POEM treatment for esophageal achalasia
Li et al
the pressurized zone of the esophageal gastric junction. Because both POEM and Heller myotomy render basically the same result anatomically (myotomy across the esophageal gastric junction, leaving esophageal mucosa intact), it is reasonable to infer they would lead to a similar clinical response. Indeed, the effect of POEM has been proven to be comparable with Heller myotomy by small-scale short-term studies.13-16 Nevertheless, concerns exist regarding POEM’s durability, whereas Heller myotomy has stood the test of time.17 Inoue et al,18 who developed and named POEM, reported their favorable 3-year results in 2015. As 1 of the earliest centers to perform POEM and the highest in volume, we herein present our 5-year results regarding the long-term outcomes of POEM.
METHODS Patients Patients were diagnosed with established methods: clinical symptoms, barium swallow, EGD, manometry, and/or chest CT scan. Exclusion criteria included coagulopathy and systemic disorders that precluded safe general anesthesia. The procedures and study were conducted in accordance with the Declaration of Helsinki and had approval from the Ethical Board of Zhongshan Hospital. Written informed consent was obtained from all patients before the procedure.
Outcome measurements The primary outcome of the study was the clinical success rate of POEM (Eckardt score 3) during follow-up. The secondary outcomes included procedure-related adverse events (AEs), lower esophageal sphincter (LES) pressure on manometry pre- and post-POEM, reflux symptoms (grade 0, absent; grade 1, <2 days a week; grade 2, 2-4 days a week; grade 3, >4 days a week),19 reflux esophagitis on EGD, and procedure parameters such as operation time, length of hospital stay, and myotomy length. Baseline and postmyotomy LES pressures were recorded using a high-resolution manometry system (Sierra Scientific Instruments Inc, Los Angeles, Calif) as previously described.20
POEM procedures The POEM procedure we used was largely the same as the original protocol raised by Inoue et al1 and has been reported by our previous publication.21 Four major steps were involved: submucosal injection, submucosal tunneling, myotomy, and closure of the mucosotomy. Several important technical modifications22 were put forward by our group and have become our standard practice over time: myotomy at the posterior (5-6 o’clock) rather than anterior esophageal wall, (2) full thickness rather than selective circular muscle myotomy, and (3) push and pull myotomy: in the esophagus anchor the knife tip at the muscle fibers from the extraluminal side and PUSH the scope, and at the cardia 2 GASTROINTESTINAL ENDOSCOPY Volume
-,
No.
-
: 2017
lay the knife on top of the muscle fibers intraluminally and PULL the scope.23 We were also among the first to apply a water jet–assisted knife in POEM.24 One technical drawback was that the CO2 insufflator did not become readily available in our center until late 2011, and this had an effect on our gas-related AEs.
Adverse events Major perioperative AEs were defined as conditions that resulted in vital-sign instability, intensive care unit stay, hospital readmission, conversion to open surgery, invasive postoperative procedures, blood transfusion, or hospitalization >5 days because of functional impairment of the patient.25 Minor perioperative AEs were defined as AEs that called for clinical interventions but did not qualify for major AE, including bouts of intraoperative subcutaneous emphysema/pneumoperitoneum that required needle decompression only and prolonged intraoperative bleeding (>200 mL) that did not require transfusion. Accidental mucosal injury was recorded given its call for special management (eg, clipping, prolonged fasting, and/or nasogastric tube placement) but not graded major or minor AEs per se. Long-term AEs were associated with acid reflux, including gastroesophageal reflux symptoms, strictures, and Barrett’s esophagus. Mucosal edema and mucosal injury, 2 risk factors for perioperative AEs, were graded and categorized.25 Mucosal edema was characterized by mucosa surface texture abnormality (cobblestone sign, milky color of the injection wheal, jellylike adhesions, etc) and/or increased cutting-edge thickness (inward folding, high tension during clipping, etc.), and graded in severity accordingly.25 Mucosal injuries were categorized as 1 of 2 types based on difficulty of repair. Type I injuries were easily repairable injuries (small, whitish, linear/dotted, or mucosal color change), whereas type II injuries were difficult-to-repair injuries (big, scorched, round/unevenly bordered, and often related to submucosal adhesion).25
Follow-up Patients were scheduled to follow-up at the center 1 month, 3 months, 6 months, and 1 year postoperation and yearly afterward, during which symptom assessment, physical examination, and objective tests including EGD and barium esophagram were performed. High-resolution manometrywas recommended to be done at least once postoperatively. Clinical response was evaluated using the Eckardt score.26 Post-POEM Eckardt score 3 was considered the benchmark for treatment success. Those with Eckardt scores >3 were grouped according to time of failure: within 3 months, nonresponders; between 3 months and 3 years, early recurrence; and after 3 years, late recurrence. Barium swallow was performed to document treatment results objectively. EGD was recommended because it provided both treatment outcome assessment and esophageal cancer surveillance.26 Manometry, although strongly advised, tended to be poorly welcomed by the patients www.giejournal.org
Li et al
because it is uncomfortable and available in only a few number of centers. For patients from faraway provinces and unwilling to come back for follow-up, detailed telephone interviews, including symptoms and treatments or tests received at outside hospitals, were conducted. The latest database checkup and phone interviews were carried out in October 2016.
Statistical analysis Categorical variables were reported as counts and percentages. Continuous data were reported as median with interquartile range (IQR) and categorized according to commonly used clinical criteria when necessary. Data were compared using (paired) the Student t test or Wilcoxon signed-rank test as appropriate. Long-term clinical success rate and risk of recurrence were estimated and graphed using the Kaplan-Meier survivor function and Nelson-Aalen cumulative hazard function, respectively. The Cox proportional hazards model was used to evaluate independent risk factors for clinical recurrence. Sensitivity analysis based on multiple imputation was conducted to check for influence of lost-to-follow-up.27 Statistical significance level was set at .05, 2-sided. All analyses were made in SAS (version 9.3; SAS Institute, Cary, NC).
long-term results of POEM treatment for esophageal achalasia
TABLE 1. Patient characteristics and procedural parameters (n [ 564) Characteristic
Value
Median age, y [IQR] (range) Gender, M/F
274/290
Median disease duration, y [IQR] (range) Sigmoid esophagus
Patient characteristics and procedure parameters A total of 567 patients received POEM at Zhongshan Hospital, Fudan University, Shanghai, China, between August 2010 and December 2012. Three POEM procedures were aborted, 2 because of severe submucosal fibrosis and 1 because of atrial fibrillation related to the electrosurgical knife. Therefore, 564 patients were analyzed in this study. The median age was 38 years, and the median disease duration was 4 years. Forty-eight patients (8.5%) had sigmoid esophagus, and 193 patients (34.2%) had received prior interventional treatment, including 20 with prior Heller myotomy and 3 with prior POEM. The median prePOEM Eckardt score was 8, and the median pre-POEM LES pressure was 29.7 mm Hg (Table 1). Regarding the procedure, the median operation time was 45 minutes. Full-thickness myotomy was performed in 352 patients (63.9%). Mucosal edema was present in 54 patients (9.6%), including 42 (7.4%) grade I and 12 (2.1%) grade II edema (Table 1).
Perioperative AEs Mucosal injuries occurred in 93 patients (16.5%), including 78 (13.8%) type I injuries and 15 (2.7%) type II injuries. Forty-eight patients (8.5%) required nasogastric tube placement at the end of the procedure. Conditions that called for nasogastric tube placement included insecure mucosal closure, severe bleeding, and a long procedwww.giejournal.org
4 [2, 9] (.1-60) 48 (8.5)
Prior treatment
193 (34.2)
Balloon/bougie dilation
134 (23.8)
Botulinum toxin injection
30 (5.3)
Esophageal stent
34 (6.0)
Heller myotomy
20 (3.6)
POEM
3 (.5)
Median pre-POEM Eckardt score [IQR] (range) Median pre-POEM LES pressure, mm Hg [IQR] (range) Median operation time, min [IQR] (range)
8 [7, 9] (4-12) 29.7 [21.4, 39.8] (5.1-70.9) 45 [32, 66] (15-202)
CO2 insufflation
340 (60.3)
Mucosa edema
54 (9.6)
Grade I
42 (7.4)
Grade II
RESULTS
38 [28, 49] (6-77)
12 (2.1)
Submucosal fibrosis
44 (7.8)
Median tunneling length, cm [IQR] (range)
13 [13, 13] (7-23)
Above EGJ
10 [10, 10] (4-20)
Below EGJ
3 [3, 3] (0-6)
Full-thickness myotomy Median myotomy length, cm [IQR] (range) Above EGJ
352 (63.9) 10 [10, 10] (5-17) 8 [8, 8] (3-13)
Below EGJ
2 [2, 2] (0-6)
Tunneling knife, TK/Hybrid/HK/others
370/106/75/6
Myotomy knife, TK/Hybrid/HK/others
437/91/23/6
Median length of stay, days [IQR] (range)
3 [2, 4] (1-29)
Values are n (%), unless otherwise noted. IQR, Interquartile range; POEM, per-oral endoscopic myotomy; LES, lower esophageal sphincter; EGJ, esophageal gastric junction; TK, triangle tip knife (KD-640L; Olympus, Japan); Hybrid, HybridKnife (Erbe, Tübingen, Germany); HK, hook knife (KD-620LR; Olympus, Tokyo, Japan).
ure time. Other minor AEs included 3 patients with estimated blood loss >200 mL (stabilized after endoscopic hemostasis and fluid resuscitation), 1 patient with subcutaneous emphysema, and 1 patient with pneumoperitoneum (both relieved by needle decompression). Major perioperative AEs occurred in 36 patients (6.4%). Three patients had delayed mucosal barrier failure (2 tunnel entrance dehiscence and 1 delayed mucosal flap perforation). Three had delayed bleeding. Six had hydrothorax. Twenty-one had pneumothorax, all of whom had received Volume
-,
No.
-
: 2017 GASTROINTESTINAL ENDOSCOPY 3
long-term results of POEM treatment for esophageal achalasia
Li et al
TABLE 2. Perioperative AEs (n [ 564) Adverse event
Value
Mucosal injury
93 (16.5)
Type I
78 (13.8)
Type II
15 (2.7)
Nasogastric tube inserted at end of POEM
48 (8.5)
Major perioperative AEs*
36 (6.4)
Delayed mucosal barrier failure
3 (.5)
Delayed bleeding requiring interventions and/or transfusion
3 (.5)
Hydrothorax requiring drainage
6 (1.1)
Pneumothorax requiring drainage
21 (3.7)
Othery
3 (.5)
Minor perioperative AEs
5 (.9)
Intraoperative bleeding >200 mL, no transfusion or special management
3 (.5)
Intraoperative subcutaneous emphysema requiring needle decompression
1 (.2)
Intraoperative pneumoperitoneum requiring needle decompression
1 (.2)
Values are n (%). AE, Adverse event; POEM, per-oral endoscopic myotomy. *When a patient had inter-related major and minor AEs, only the major AE is listed. These included 3 intraoperative and delayed bleeding, 1 intraoperative subcutaneous emphysema and pneumothorax requiring continuous drainage, and 2 intraoperative pneumoperitoneum and pneumothorax requiring continuous drainage. yOne 40-year-old man had status epilepticus on postoperative day 2 and required intensive care unit stay. One 45-year-old man had delayed extubation and intensive care unit stay because of chronic lung infection possibly complicated by pharyngeal edema. One 61-year-old woman had postoperative lung infection and transient hypoxemia.
air rather than CO2 insufflation. Three patients had other miscellaneous major AEs (Table 2).25 After initiation of CO2 insufflation, the AE rate dropped to 2.4%.
Long-term symptomatic relief Of the 564 patients, 420 (74.5%) had follow-up results for a median of 49 months (range, 3-67); 357 patients (63.3%) were followed-up for 3 years and 237 (42.0%) for 4 years. The median Eckardt score for all patients at their latest follow-up was 2 (IQR [1, 3]; range, 0-10), which was significantly lower than their corresponding pre-POEM score (median, 8; IQR [7, 9]; range, 4-12; P < .05). When the Kaplan-Meier survival function was used, the estimated clinical success rates at 1, 2, 3, 4, and 5 years were 94.2% (95% confidence interval [CI], 91.6%-96.0%), 92.2% (95% CI, 89.2%-94.4%), 91.1% (95% CI, 87.9%-93.5%), 88.6% (95% CI, 85.0%-91.4%), and 87.1% (95% CI, 82.8%-90.4%), respectively (Table 3, Fig. 1).
Manometry outcomes A total of 265 patients (47.0%) underwent per-protocol high-resolution manometry postoperatively. Other patients did not receive follow-up manometry because of the difficulty in traveling to major medical centers capable of motility 4 GASTROINTESTINAL ENDOSCOPY Volume
-,
No.
-
: 2017
tests, discomfort related to the test, and other personal reasons. The median LES pressure at follow-up was 11.9 mm Hg (IQR [9.2, 15.6]; range, 4.0-24.6), which was significantly lower than their corresponding pre-POEM value (median, 29.7 mm Hg; IQR [21.4, 39.8]; range 5.1-70.9; P < .05).
Gastroesophageal reflux and other long-term AEs Gastroesophageal reflux was evaluated in 416 patients. Of the 375 patients who were evaluated for heartburn and regurgitation, 137 patients (36.5%) reported presence of reflux symptoms, with 126 (92.0%) grade 1 and 11 (8.0%) grades 2 to 3 by frequency. Of the 341 patients who received followup EGD, 58 patients (17.0%) showed signs of reflux esophagitis. Forty-three patients were evaluated for severity of esophagitis, of whom 26 (60.5%) had mild esophagitis (Los Angeles classification grade A), 14 (32.6%) had moderate esophagitis (grades B-C), and only 3 (7.0%) had severe esophagitis (grade D). When combining both symptomatic reflux and reflux esophagitis, clinical reflux occurred in 155 patients (37.3%). There was no confirmed reflux-related post-POEM stricture that required interventional treatments (Supplementary Table 1, available at www.giejournal.org). Five known deaths occurred during follow-up. One 48-year-old woman (the fourth patient in the whole series) had persistent symptoms and died of cachexia 15 months after POEM. The other 4 patients died of acute cholangitis, cholangiocarcinoma, lymphoma, and lung cancer, respectively. The 2 patients who died of lymphoma and lung cancer had negative noncontrast-enhanced chest CT during the perioperative period.
POEM failures Of the 420 patients with follow-up results, 48 (11.4%) failures occurred. Fifteen were nonresponders (failure within 3 months), 23 were early recurrence (between 3 months and 3 years), and 10 were late recurrence (after 3 years). The Nelson-Aalen estimated cumulative hazards for recurrence at 1, 2, 3, 4, and 5 years were 5.9% (95% CI, 4.0%-8.7%), 8.0% (95% CI, 5.7%-11.3%), 9.3% (95% CI, 6.7%-12.8%), 12.0% (95% CI, 8.9%-16.2%), and 13.7% (95% CI, 10.0%18.7%), respectively (Fig. 2, Table 3). The median age of these patients was 47 years, and the median disease duration was 6.8 years. Eight (16.7%) had sigmoid esophagus, and 24 (50.0%) had a history of prior interventional treatment. Their median post-POEM Eckardt score was 5 (IQR [4, 6]; range, 4-10), which was significantly lower than the corresponding pre-POEM value (median, 8; IQR [7, 9]; range, 4-11; P < .05). There was no significant difference in the pre-POEM Eckardt score between patients with and without recurrence (P Z .88). Clinical reflux was evaluated in 46 patients, of whom 23 (50.0%) were positive, including 47.7% (21/44) with symptomatic reflux and 20.5% (8/39) with reflux esophagitis on EGD (Table 4). When the multivariate Cox proportional hazards model was used, disease duration 10 years (hazard ratio, 2.45; 95% CI, www.giejournal.org
Li et al
long-term results of POEM treatment for esophageal achalasia
TABLE 3. Estimated clinical success (Eckardt score ≤3) and cumulative recurrence
No. at risk*
No. of recurrences within period
End of follow-upy within period
Estimated clinical success rate* (95% CI)
Estimated cumulative recurrence* (95% CI)
376
354
27
183
.942 (.916-.960)
.059 (.040-.087)
358
330
7
17
.922 (.892-.944)
.080 (.057-.113)
2-3
357
323
4
3
.911 (.879-.935)
.093 (.067-.128)
3-4
237
215
8
100
.886 (.850-.914)
.120 (.089-.162)
4-5
48
47
2
166
.871 (.828-.904)
.137 (.100-.187)
No. of patients under follow-up*
0-1 1-2
Time, year
*Patients under follow-up, at risk, estimated clinical success, and cumulative recurrence at the end of the period. Success rate estimated by Kaplan-Meier survival function and cumulative recurrence by Nelson-Aalen function. yThe patient recruitment period was August 2010 to December 2012, and the last round of data collection lasted throughout 2016. As a result, many patients receiving their POEM during 2012 have not yet been contacted at 4/5 years and are not loss-to-follow-up.
DISCUSSION Introduced as a minimally invasive alternative to the long-accepted criterion standard, Heller myotomy, POEM needs to undergo comprehensive assessments regarding its safety and effectiveness before it is truly accepted. In www.giejournal.org
1 .75
95% CI Survivor function
.5
A total of 144 patients were lost to follow-up, most of whom (137, 95.1%) were because of changed or absent contact information; 7 (4.9%) declined phone follow-up. Baseline characteristics, occurrence of major perioperative AEs, and length of hospital stay were comparable between patients with and without follow-up except that those without follow-up were more likely to be male and have a shorter disease duration (median disease duration, 3 years vs 4 years in those with follow-up; P Z .03) (Supplementary Table 2, available online at www.giejournal.org), indicating a possible higher willingness to follow-up in female patients and patients with longer disease duration. Because missing observations might lead to biased parameter estimates, sensitivity analysis by multiple imputation was applied to the Cox regression model to test for stability of results. Under the assumption that independent factors were missing at random and had multivariate normal distribution, multiple imputations for postoperative Eckardt score and time-to-event were performed 100 times, and the results from 100 Cox regression models were combined. Result showed that history of prior interventions increased the risk of recurrence (hazard ratio, 1.20; 95% CI, 1.091.32; P < .01). Meanwhile, although long disease duration (10 years) seemed to have a more prominent hazard ratio, the results after multiple imputations were not statistically significant (hazard ratio, 2.10; 95% CI, .99-4.44; P Z .05). However, both hazard ratio estimates were close to those before multiple imputations, indicating a nonsignificant impact of missing values on risk factor analysis.
.25
Lost to follow-up and sensitivity analysis
Kaplan-Meier survival estimate for clinical success rate
0
1.35-4.46) and prior interventions (hazard ratio, 1.12; 95% CI, 1.02-1.23) were independently related to a higher chance of recurrence (Table 5).
20
0
40
60
80
Follow-up time (months)
Figure 1. Kaplan-Meier survival estimate for clinical success rate. CI, Confidence interval.
spite of the favorable results reported by several highquality studies,18,28-33 large cohorts that can give a more precise outlook on long-term outcomes are few. Extended follow-up shows decreased effectiveness over time for all types of achalasia treatments, including Heller myotomy, which is considered the most enduring.34-37 Reasons for recurrence include avoidable causes such as inadequate myotomy, tight fundoplication and strictures caused by reflux esophagitis, and unavoidable causes such as fibrosis, progression of disease, and development of megaesophagus.36-39 As forerunners in the field of POEM, we are by no means assuming that POEM is a permanent cure for all. Rather, our assumption is that POEM should have long-term outcomes comparable with Heller myotomy. This assumption has indeed been supported by nonrandomized studies.13,15,40 Two randomized controlled trials comparing POEM with Heller myotomy are also underway and are due to be complete in 2017 (NCT02138643) and 2019 (NCT01601678). Initial clinical success rates higher than 90% have been reported by all studies with a volume of over 100 (single or multicenter)18,28-33,41 (Supplementary Table 3, available online at www.giejournal.org), which is again shown by Volume
-,
No.
-
: 2017 GASTROINTESTINAL ENDOSCOPY 5
long-term results of POEM treatment for esophageal achalasia
Li et al
TABLE 4. Summary of patients with recurrence (n [ 48)
.2
Nelson-Aalen cumulative hazard estimate for recurrence
Characteristic
.15
Median age, y [IQR] (range) Gender, M/F
.1
Median disease duration, y [IQR] (range)
95% CI
0
.05
Cumulative hazard
0
20
40 60 Follow-up time (months)
80
our study (clinical success rate of 96.4% at 3 months). Decreasing effectiveness over time has been indicated by several mid-term studies. Inoue et al18 reported a success rate of 91.3% at 2 months, 91.0% at 1 to 2 years, and 88.5% at 3 years in their first 500 patients. A multicenter international trial conducted by Werner et al31 reported a clinical success rate of 93.7% at 3 to 6 months, 88.5% at 12 to 18 months, and 77.5% at 24 to 41 months. Under the Kaplan-Meier survival function, our result showed a success rate of 94.2% at 1 year, 91.1% at 3 years, and 87.1% at 5 years. These results are comparable with that of Heller myotomy performed in state-of-the-art centers.36,37,42-44 According to our experience, incomplete myotomy is the most likely cause for treatment failure within 12 months, a viewpoint shared by Heller myotomy experts.38,44 In the adult Chinese population we generally try to guarantee a myotomy of at least 8 cm above the esophageal gastric junction and 2 cm below. In addition, we prefer full-thickness over partialthickness myotomy because it is in accordance with the Heller myotomy principle and is faster while no more likely to cause procedure-related AEs.23 Although a study has shown full-thickness myotomy to be predictive of clinically relevant gastroesophageal reflux,41 this should not necessarily be viewed negatively because endoscopic reflux signs have been shown to be an independent predictor of treatment success.31 Under the Cox proportional hazards model, long disease duration (10 years) and prior interventional treatment history are identified as risk factors for POEM failures. Considering that all existing achalasia treatments aim to lower the LES pressure, it is not surprising to see that patients who have failed 1 treatment are also more likely to fail a following treatment, regardless of which is used. For instance, it is well documented that Heller myotomy tends to have a poorer response in patients with advanced disease and treatment history.44,45 Meanwhile, robust evidence has shown the benefits of POEM in these patients.46-49 In addition, our study showed that even in patients who are considered failures (Eckardt score >3), there was still a significant decrease in the score compared 6 GASTROINTESTINAL ENDOSCOPY Volume
-,
No.
-
: 2017
47 [34, 53] (23-71) 18/30 6.8 [3, 15] (.3-40)
Sigmoid esophagus
8 (16.7)
Prior treatment
24 (50.0)
Median operation time, min [IQR] (range) Full-thickness myotomy
Figure 2. Nelson-Aalen cumulative hazard estimate for recurrence. CI, Confidence interval.
Value
Median myotomy length, cm [IQR] (range)
45 [34, 60] (19-120) 33 (68.8) 10 [10, 11] (6-16)
Above EGJ
8 [8, 9] (3-13)
Below EGJ
2 [2, 3] (2-5)
Clinical reflux Reflux symptoms, yes/total Grade 1 Grades 2-3 Reflux esophagitis on EGD, yes/total
23/46 (50) 21/44 (47.7) 19 (90.5) 2 (9.5) 8/39 (20.5)
Median pre-POEM Eckardt score [IQR] (range)
8 [7, 9] (4-11)
Median post-POEM Eckardt score [IQR] (range)
5 [4, 6] (4-10)
Values are n (%), unless otherwise noted. IQR, Interquartile range; EGJ, esophageal gastric junction; POEM, per-oral endoscopic myotomy.
with baseline. Considering the above, it is reasonable to offer POEM as first-line therapy to these patients.12 Post-POEM reflux and its management is another interesting topic. Long-term acid exposure is 1 of the major arguments against POEM because currently it cannot be paired with antireflux procedures and Heller myotomy can. Previous studies have reported post-POEM symptomatic reflux rate varying from 8.6% to 37% and endoscopic reflux esophagitis rate from to 14.8% to 66% (Supplementary Table 3). Symptoms and reflux esophagitis do not always overlap, and it is interesting to note that many centers reported a higher rate of reflux esophagitis than reflux symptoms. This is not the case for normal GERD, in which most patients with heartburn and regurgitation will not have erosions under endoscopy.50 Evaluating reflux in postPOEM patients is difficult for various reasons, including the overlap of symptoms between reflux and undertreated achalasia (chest pain, regurgitation, dysphagia, and even heartburn); stasis, which may falsely prolong reflux time under pH testing; and ischemic/postsurgical changes in the esophageal gastric junction that can mimic erosive esophagitis. A better approach might be to look at the clinical consequences of reflux, in which case POEM seemed to have fared fairly well. In our series there was no confirmed POEM failure caused by reflux-related strictures. All patients www.giejournal.org
Li et al
long-term results of POEM treatment for esophageal achalasia
TABLE 5. Risk factor analysis for recurrence using Cox proportional hazards model Univariate Variable Age 60 y Male gender
Multivariate
Hazard ratio (95% CI)
P value
Hazard ratio (95% CI)
P value
1.20 (.51-2.83)
.67
NA
NS
.67 (.38-1.21)
.19
NA
NS
Disease duration 10 y
2.46 (1.39-4.37)
<.01
2.45 (1.35-4.46)
<.01
Sigmoid esophagus
2.36 (1.10-5.05)
.03
NA
NS
Any type of prior interventional treatments*
1.16 (1.05-1.27)
<.01
1.12 (1.02-1.23)
.02
Balloon/bougie dilation
1.14 (1.03-1.27)
.02
d
d
Botulinum toxin injection
1.22 (.61-2.43)
.57
d
d
Esophageal stent
1.14 (.61-2.10)
.68
d
d
Heller myotomy
2.94 (1.06-8.20)
.04
d
d
POEM
12.7 (3.06-52.5)
<.01
d
d
Pre-POEM Eckardt score 7
1.34 (.66-2.69)
.42
NA
NS
Full-thickness myotomy
1.40 (.76-2.60)
.28
NA
NS
Myotomy length above esophageal gastric junction 8 cm
.93 (.45-1.92)
.84
NA
NS
Clinical reflux
1.66 (.93-2.95)
.09
NA
NS
Symptomatic reflux
1.64 (.91-2.97)
.10
d
d
Reflux esophagitis on EGD
1.27 (.58-2.76)
.55
d
d
NA, Not assessed; NS, not significant; d, not included for analysis; POEM, per-oral endoscopic myotomy. *Defined as total number of prior interventions, including dilation, botulinum toxin injection, stent, Heller myotomy, and POEM.
with reflux symptoms and signs, including those with Los Angeles grade D esophagitis, were successfully managed with proton pump inhibitors, and no patient required rescue fundoplication. Our study is the largest single-center, long-term outcome study of POEM worldwide to date. Nevertheless, we also face challenges such as high loss-to-follow-up rate, poor patient compliance at diagnostic tests, and difficulties in accessing records from outside hospitals. These result in a lack of in-depth analysis of causes of POEM failures, especially regarding the role reflux played. The late initiation of CO2 insufflation also made the AE rate seem unrealistically high. Despite the limitations we believe this study brings important information regarding POEM given the size of the study, in-depth analysis, detailed recording, and long follow-up. In addition, because a longer disease history was found to be a risk factor for recurrence and was related to a higher rate of follow-up, our estimation on long-term success rate is likely a modest one. In conclusion, our large cohort showed excellent outcome of POEM after a medium follow-up of 49 months (estimated clinical success rate at 5 years, 87.1%). Even for those considered to have failed POEM (Eckardt score 4), there was often still a significant improvement of symptoms compared with their baselines. Long disease duration and interventional treatment history are related to recurrence of achalasia symptoms but do not deny POEM as first-line therapy in patients with advanced disease. Postoperative acid reflux is currently an underexplored area and deserves www.giejournal.org
further study to testify for the safety of POEM in the long-term.
REFERENCES 1. Inoue H, Minami H, Kobayashi Y, et al. Peroral endoscopic myotomy (POEM) for esophageal achalasia. Endoscopy 2010;42:265-71. 2. Sumiyama K, Gostout CJ, Rajan E, et al. Transgastric cholecystectomy: transgastric accessibility to the gallbladder improved with the SEMF method and a novel multibending therapeutic endoscope. Gastrointest Endosc 2007;65:1028-34. 3. Pasricha P, Hawari R, Ahmed I, et al. Submucosal endoscopic esophageal myotomy: a novel experimental approach for the treatment of achalasia. Endoscopy 2007;39:761-4. 4. Woodward TA, Jamil LH, Wallace MB. Natural orifice trans-luminal endoscopic surgery in the esophagus. Gastrointest Endosc Clin North Am 2010;20:123-38. 5. Barbieri LA, Hassan C, Rosati R, et al. Systematic review and metaanalysis: efficacy and safety of POEM for achalasia. United Eur Gastroenterol J 2015;3:325-34. 6. Wei M, Yang T, Yang X, et al. Peroral esophageal myotomy versus laparoscopic Heller’s myotomy for achalasia: a meta-analysis. J Laparoendosc Adv Surg Techn 2015;25:123-9. 7. Talukdar R, Inoue H, Reddy DN. Efficacy of peroral endoscopic myotomy (POEM) in the treatment of achalasia: a systematic review and meta-analysis. Surg Endosc 2015;29:3030-46. 8. Bechara R, Ikeda H, Inoue H. Peroral endoscopic myotomy: an evolving treatment for achalasia. Nat Rev Gastroenterol Hepatol 2015;12:410-26. 9. Crespin OM, Liu LWC, Parmar A, et al. Safety and efficacy of POEM for treatment of achalasia: a systematic review of the literature. Surg Endosc 2017;31:2187-201. 10. Akintoye E, Kumar N, Obaitan I, et al. Peroral endoscopic myotomy: a meta-analysis. Endoscopy 2016;48:1059-68.
Volume
-,
No.
-
: 2017 GASTROINTESTINAL ENDOSCOPY 7
long-term results of POEM treatment for esophageal achalasia
Li et al
11. Patel K, Abbassi-Ghadi N, Markar S, et al. Peroral endoscopic myotomy for the treatment of esophageal achalasia: systematic review and pooled analysis. Dis Esoph 2016;29:807-19. 12. Stavropoulos SN, Friedel D, Modayil R, et al. Diagnosis and management of esophageal achalasia. BMJ 2016;354:i2785. 13. Teitelbaum EN, Rajeswaran S, Zhang R, et al. Peroral esophageal myotomy (POEM) and laparoscopic Heller myotomy produce a similar short-term anatomic and functional effect. Surgery 2013;154:885-92. 14. Ujiki MB, Yetasook AK, Zapf M, et al. Peroral endoscopic myotomy: a short-term comparison with the standard laparoscopic approach. Surgery 2013;154:893-900. 15. Bhayani NH, Kurian AA, Dunst CM, et al. A comparative study on comprehensive, objective outcomes of laparoscopic Heller myotomy with per-oral endoscopic myotomy (POEM) for achalasia. Ann Surg 2014;259:1098-103. 16. Schneider AM, Louie BE, Warren HF, et al. A matched comparison of per oral endoscopic myotomy to laparoscopic Heller myotomy in the treatment of achalasia. J Gastrointest Surg 2016;20: 1789-96. 17. Campos GM, Vittinghoff E, Rabl C, et al. Endoscopic and surgical treatments for achalasia. Ann Surg 2009;249:45-57. 18. Inoue H, Sato H, Ikeda H, et al. Per-oral endoscopic myotomy: a series of 500 patients. J Am Coll Surg 2015;221:256-64. 19. Vigneri S, Termini R, Leandro G, et al. A comparison of five maintenance therapies for reflux esophagitis. N Engl J Med 1995;333: 1106-10. 20. Pandolfino JE, Kwiatek MA, Nealis T, et al. Achalasia: a new clinically relevant classification by high-resolution manometry. Gastroenterology 2008;135:1526-33. 21. Li Q, Yao L, Xu X, et al. Repeat peroral endoscopic myotomy: a salvage option for persistent/recurrent symptoms. Endoscopy 2016;48:134-40. 22. Li Q, Zhou P. Perspective on peroral endoscopic myotomy for achalasia: Zhongshan experience. Gut Liver 2015;9:152-8. 23. Li Q, Chen W, Zhou P, et al. Peroral endoscopic myotomy for the treatment of achalasia: a clinical comparative study of endoscopic fullthickness and circular muscle myotomy. J Am Coll Surg 2013;217: 442-51. 24. Cai M, Zhou P, Yao L, et al. Peroral endoscopic myotomy for idiopathic achalasia: randomized comparison of water-jet assisted versus conventional dissection technique. Surg Endosc 2014;28:1158-65. 25. Zhang X, Li Q, Xu M, et al. Major perioperative adverse events of peroral endoscopic myotomy: a systematic 5-year analysis. Endoscopy 2016;48:967-78. 26. Eckardt AJ, Eckardt VF. Treatment and surveillance strategies in achalasia: an update. Nat Rev Gastroenterol Hepatol 2011;8:311-9. 27. Rubin DB. Multiple imputation after 18þ years. J Am Stat Assoc 1996;91:473-89. 28. Sharata AM, Dunst CM, Pescarus R, et al. Peroral endoscopic myotomy (POEM) for esophageal primary motility disorders: analysis of 100 consecutive patients. J Gastrointest Surg 2015;19:161-70. 29. Ramchandani M, Reddy DN, Darisetty S, et al. Peroral endoscopic myotomy for achalasia cardia: treatment analysis and follow up of over 200 consecutive patients at a single center. Dig Endosc 2016;28:19-26. 30. Familiari P, Gigante G, Marchese M, et al. Peroral endoscopic myotomy for esophageal achalasia. Ann Surg 2016;263:82-7.
8 GASTROINTESTINAL ENDOSCOPY Volume
-,
No.
-
: 2017
31. Werner YB, Costamagna G, Swanström LL, et al. Clinical response to peroral endoscopic myotomy in patients with idiopathic achalasia at a minimum follow-up of 2 years. Gut 2016;65:899-906. 32. Shiwaku H, Inoue H, Yamashita K, et al. Peroral endoscopic myotomy for esophageal achalasia: outcomes of the first over 100 patients with short-term follow-up. Surg Endosc 2016;30:4817-26. 33. Ngamruengphong S, Inoue H, Chiu PW, et al. Long-term outcomes of per-oral endoscopic myotomy in patients with achalasia with a minimum follow-up of 2 years: an international multicenter study. Gastrointest Endosc 2016;85:927-33. 34. Zaninotto G, Annese V, Costantini M, et al. Randomized controlled trial of botulinum toxin versus laparoscopic Heller myotomy for esophageal achalasia. Ann Surg 2004;239:364-70. 35. Vela MF, Richter JE, Khandwala F, et al. The long-term efficacy of pneumatic dilatation and Heller myotomy for the treatment of achalasia. Clin Gastroenterol Hepatol 2006;4:580-7. 36. Ortiz A, de Haro LFM, Parrilla P, et al. Very long-term objective evaluation of Heller myotomy plus posterior partial fundoplication in patients with achalasia of the cardia. Ann Surg 2008;247:258-64. 37. Kilic A, Schuchert MJ, Pennathur A, et al. Long-term outcomes of laparoscopic Heller myotomy for achalasia. Surgery 2009;146:826-33. 38. Patti MG, Molena D, Fisichella PM, et al. Laparoscopic Heller myotomy and Dor fundoplication for achalasia: analysis of successes and failures. Arch Surg 2001;136:870-7. 39. Gockel I. Persistent and recurrent achalasia after Heller myotomy. Arch Surg 2007;142:1093. 40. Hungness ES, Teitelbaum EN, Santos BF, et al. Comparison of perioperative outcomes between peroral esophageal myotomy (POEM) and laparoscopic Heller myotomy. J Gastrointest Surg 2013;17:228-35. 41. Ren Y, Tang X, Chen Y, et al. Pre-treatment Eckardt score is a simple factor for predicting one-year peroral endoscopic myotomy failure in patients with achalasia. Surg Endosc 2017;31:3234-41. 42. Frantzides CT, Moore RE, Carlson MA, et al. Minimally invasive surgery for achalasia: a 10-year experience. J Gastrointest Surg 2004;8:18-23. 43. Bonatti H, Hinder RA, Klocker J, et al. Long-term results of laparoscopic Heller myotomy with partial fundoplication for the treatment of achalasia. Am J Surg 2005;190:883-7. 44. Zaninotto G, Costantini M, Rizzetto C, et al. Four hundred laparoscopic myotomies for esophageal achalasia. Ann Surg 2008;248:986-93. 45. Smith CD, Stival A, Howell DL, et al. Endoscopic therapy for achalasia before Heller myotomy results in worse outcomes than Heller myotomy alone. Ann Surg 2006;243:579-86. 46. Sharata A, Kurian AA, Dunst CM, et al. Peroral endoscopic myotomy (POEM) is safe and effective in the setting of prior endoscopic intervention. J Gastrointest Surg 2013;17:1188-92. 47. Hu J, Li Q, Zhou P, et al. Peroral endoscopic myotomy for advanced achalasia with sigmoid-shaped esophagus: long-term outcomes from a prospective, single-center study. Surg Endosc 2015;29:2841-50. 48. Orenstein SB, Raigani S, Wu YV, et al. Peroral endoscopic myotomy (POEM) leads to similar results in patients with and without prior endoscopic or surgical therapy. Surg Endosc 2015;29:1064-70. 49. Lv L, Liu J, Tan Y, et al. Peroral endoscopic full-thickness myotomy for the treatment of sigmoid-type achalasia. Eur J Gastroenterol Hepatol 2016;28:30-6. 50. Katz PO, Gerson LB, Vela MF. Guidelines for the diagnosis and management of gastroesophageal reflux disease. Am J Gastroenterol 2013;108: 308-28.
www.giejournal.org
Li et al
long-term results of POEM treatment for esophageal achalasia
SUPPLEMENTARY TABLE 1. Follow-up results on gastroesophageal reflux Result
Value
Clinical reflux, yes/total
155/416 (37.3)
Reflux symptoms, yes/total
137/375 (36.5)
Grade 1 Grades 2-3 Reflux esophagitis on EGD, yes/total*
126 (92.0) 11 (8.0) 58/341 (17.0)
Grade A
26 (60.5)
Grade B-C
14 (32.6)
Grade D
3 (7.0)
Values in parentheses are percents. Clinical reflux: either reflux symptoms of reflux esophagitis on EGD. Grading of reflux symptom frequency: Grade 1, <2 days a week; grade 2, 2-4 days a week; grade 3, >4 days a week. Grading of reflux esophagitis: Los Angeles Classification. *Fifteen patients were recorded to have esophagitis but not graded for severity.
www.giejournal.org
Volume
-,
No.
-
: 2017 GASTROINTESTINAL ENDOSCOPY 8.e1
long-term results of POEM treatment for esophageal achalasia
Li et al
SUPPLEMENTARY TABLE 2. Comparison of baseline characteristics and postoperative parameters between patients with and without follow-up Patients with follow-up (n [ 420)
Patients lost to follow-up (n [ 144)
P value
39 [29, 51] (6-77)
37 [26, 48] (12-76)
.12
193/227
81/63
.03
4 [2, 9] (.1-60)
3 [1, 7] (.1-40)
.03
36 (8.6)
12 (8.3)
1.00
151 (36.0)
42 (29.2)
.16
8 [6, 9] (4-12)
8 [7, 9] (5-12)
.53
260 (63.0)
92 (66.7)
.47
Median myotomy length above EGJ, cm [IQR] (range)
8 [8, 8] (3- 13)
8 [8, 8] (3-13)
.89
Median myotomy length below EGJ, cm [IQR] (range)
2 [2, 3] (0-6)
2 [2, 2] (0-5)
.08
Median age, y [IQR] (range) Gender, M/F Median disease duration, y [IQR] (range) Sigmoid esophagus Prior treatment Median pre-POEM Eckardt score [IQR] (range) Full-thickness myotomy
Major perioperative adverse events Median length of stay, days [IQR] (range)
28 (6.7)
8 (5.6)
.84
3 [2, 4] (1-29)
2 [2, 4] (1-29)
.40
Values are n (%), unless otherwise noted. IQR, Interquartile range; POEM, per-oral endoscopic myotomy; EGJ, esophageal gastric junction.
8.e2 GASTROINTESTINAL ENDOSCOPY Volume
-,
No.
-
: 2017
www.giejournal.org
Li et al
long-term results of POEM treatment for esophageal achalasia
SUPPLEMENTARY TABLE 3. Literature review outcomes of POEM in studies with a volume of over 100 (single-center, multicenter trials and multicenter pooled analyses) Inoue et al18
Sharata et al28
Ramchandani et al29
Familiari et al30
Werner et al31
Shiwaku et al32
Ngamruengphong et al33
Ren et al41
Country origin
Japan
USA
India
Italy
Multicenter, USA and Europe
Japan
Multicenter, USA, Europe, Asia
China
No. of patients
500
100
212
94
80
103
205
115
2 mo, 1-2 y, 3y
16 mo (mean)
6 mo, 1 y
11 mo (mean)
29 mo (mean)
3 mo
31 mo (median)
At 1-2 y: 77.3% (286/370) At 3 y: 58.1% (61/105)
100%
At 6 mo: 98.0% (149/152)
97.9% (92/94)
1 patient lost at 6 mo 1 patient lost at 18 mo
100%
100%
At 2 mo: 91.3% (386/423) At 1-2 y: 91% (260/286) At 3 y: 88.5% (54/61) (Eckardt score <2 or reduction >4)
92% (92/100)
94.5% (87/92)
At 3-6 mo: 93.7% (74/79) At 12-18 mo: 88.5% (69/78) At 24-41 mo: 77.5% (62/80)
99% (99/ 100)
Follow-up time Follow-up rate
Clinical success rate (Eckardt score 3)
AEs
At 6 mo: 94% (140/149) At 1 y: 92% (94/102)
3.2% (total)*
6% (total)y
No major AEs
No major AE
Reflux symptoms
At 2 mo: 16.8% (71/423) At 3 y: 21.3% (13/61)
Heartburn 8.6% (7/81), regurgitation 9.9% (8/81)
21.6% (22/102)
24.3% (17/73)
37% (27/73)
Reflux esophagitis
At 2 mo: 64.7% (268/414) At 3 y: 56.3% (9/16)
27.4% (20/73)
16.6% (14/84)
27.4% (20/73)
37.5% (37/72)
Positive 24-hour pH testing
d
38.2% (26/68) (DeMeester score >14.7)
d
d
93.9% (108/ 115)
93% (107/ 115)
8.2% (total)z
-
9% (9/100)
26.8% (55/197)
d
66% (66/ 100)
18% (26/144)
14.8% (17/115)
25.6% (22/86) (monitoring time with pH < 4)
37.5% (12/32) (DeMeester score >14.72)
-
1 readmission 10 AEs (mucosa (Clavien-Dindo ulcer) I –IIIa) No other major
53.4% (39/73) (total reflux time >5%)
At 6 mo: 98% (185/189) At 12 mo: 98% (142/144) At 24 mo: 91% (187/205)
1y
POEM, Per-oral endoscopic myotomy; AE, adverse event. *One pneumothorax with mediastinal emphysema, 1 severe bleeding, 8 mucosal injuries, 3 postoperative hematomas from submucosal bleeding, 1 inflammation in the lesser omentum, and 2 pleural effusions. yThree leaks diagnosed on routine esophagram, 1 postoperative intratunnel bleeding, 1 Ogilvie’s syndrome, and 1 prolonged intubation for CO2 retention. zEight inadvertent mucosotomies, 2 symptomatic pleural effusions, 2 aspiration pneumonia, 2 deep vein thrombosis and pulmonary embolism, 1 symptomatic pneumothorax, 1 delayed bleeding, 1 esophageal leak, and 1 mediastinitis.
www.giejournal.org
Volume
-,
No.
-
: 2017 GASTROINTESTINAL ENDOSCOPY 8.e3