Endoscopy 2017; 49(07): 625-628
DOI: 10.1055/s-0043-112490
© Georg Thieme Verlag KG Stuttgart · New York

Will Reflux Kill POEM?

Referring to Kumbhari V et al. p. 634–642
Thomas Rösch
1  Department of Interdisciplinary Endoscopy, University Hospital Hamburg-Eppendorf, Germany
Alessandro Repici
2  Digestive Endoscopy Unit, Division of Gastroenterology, Humanitas Research Hospital, Milan, Italy
Guy Boeckxstaens
3  Department of Gastroenterology, University Hospital Leuven, Leuven, Belgium
› Author Affiliations
Further Information

Publication History

Publication Date:
28 June 2017 (online)

Peroral endoscopic myotomy, which was once given the nice and poetic acronym POEM [1] [2], has aroused enormous interest in the gastrointestinal endoscopy community as an endoscopic therapeutic option for idiopathic achalasia [3] [4] [5] and subsequently for spastic motility disorders. As perhaps the only remnant of the natural orifice transluminal endoscopic surgery (NOTES) hype, it is a procedure whereby the lower esophageal sphincter is dissected via a submucosal tunnel, working almost entirely outside the lumen. When performed carefully, it has its own aesthetics during endoscopic live demonstrations, but perhaps we should start questioning whether it really benefits our patients (or maybe even harms them).

“We should keep a close and critical eye on the issue of reflux in POEM”

As so often in interventional medical research, an interesting topic leads to a plethora of retrospective success stories, called studies, or – even worse – registries, eventually mashed up in so-called meta-analyses, the scientific epidemic of our times. Not surprisingly; there are very few really prospective studies and no single randomized trial, yet already 10 meta-analyses or systematic reviews have been reported [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16]. Also not surprisingly, all conclude that POEM is excellent.

To start with efficacy data: Although clinical assessment using the Eckardt score has become the international standard, symptoms remain an insensitive tool as patients either minimize symptoms or adapt their diet. Hence, certainly at the early stage of introducing a new technique, additional functional testing is fundamental. High resolution manometry (HRM) and timed barium esophagogram to assess lower esophageal sphincter function and emptying should therefore be performed preferentially at different time points during follow-up [16]. Initial clinical success (defined as Eckardt score ≤ 3) is achieved in the vast majority of patients with any therapeutic approach; however objective functional data may be better early predictors of long-term success. Mid-term outcomes at 2 years have been shaky for POEM, with substantial loss of efficacy found in some studies [17], better outcome in others [18] [19] [20], and superb results in international multicenter surveys, the latter raising some questions about case selection. Indeed, reported case series may represent a fraction of those treated, given the high volume names in the author list [21]. This all shows that beyond marketing research, we clearly need randomized trials, and at least three such trials comparing POEM with Heller myotomy or balloon dilatation are under way. The benchmark in this area is the randomized trial comparing Heller myotomy with balloon dilation [22] [23]. With regards to POEM and adverse events, major complication rates seem to be low [7] [8] [24], and with minor complications, we clearly have a definition problem [24].

What about reflux? Is it a (mid-/long-term) complication or an unavoidable sequela? How frequent and serious is gastroesophageal reflux? Should we neglect it and simply treat with low dose proton pump inhibitors (PPIs)? Or do we actually create a “human model” of Barrett’s esophagus? Honestly, we do not know. The methodological quality of most studies is limited, mainly because the objective assessment used (symptom questionnaires, PPI consumption, reflux esophagitis, 24-hour pH-metry) varies significantly between studies. Not unexpectedly, the outcome of different meta-analyses differs enormously with 24-hour pH-metry as the single stable parameter, probably because of low numbers of studies included ( [ Table1 ]). Barrett’s esophagus has been reported in a few cases, mostly evident only on biopsy (i. e., very short Barrett’s), but systematic studies are lacking.

Table 1

Reflux rates according to articles summarizing peroral endoscopic myotomy (POEM) papers. (Those articles reviewing comparisons of POEM with laparoscopic Heller myotomy are not included.)

Studies, n

Cases, n





Akintoye 2016 [6]



8.5 %



13 %



47 %

Crespin 2016 [8]





47 %

Patel [11] 2016




33 %


19 %


43 %

Barbieri [7] 2015



13 %

* Of the 16 studies included, only 8 reported reflux according to Table 3 of the paper; other reflux parameters are not included.

The paper on reflux and POEM in this issue of Endoscopy [25] specifically deals with this question. It is a retrospective compilation of patient data from 7 centers. Of note, more than 50 % of cases were excluded as no 24-hour pH-metry data were available. In 22.3 % no Chicago classification was registered. The authors report abnormal acid exposure, defined as a DeMeester score > 14.72, in 58 % of patients, a figure amongst the highest ever reported. We can argue that this may be a selection of cases, a negative selection perhaps (measuring pH only in those who were symptomatic), but the paper also points towards the issue of pathological gastroesophageal reflux after POEM.

How does the competitor fare: what about reflux after surgical Heller myotomy? The advantage of this approach is the introduction of an antireflux procedure, albeit not the classical Nissen fundoplication for reflux disease. Although the addition of an antireflux procedure is accepted to greatly reduce esophageal acid exposure, only one small randomized trial (n = 43) has been published on this topic, showing reflux rates of 9.1 % versus 47.6 % in the groups with and without Dor fundoplication, respectively [26]. Before that trial, a somewhat curious meta-analysis, including 15 studies that involved 532 patients with an antireflux procedure and only 69 without, had not shown significantly reduced overall reflux rates (determined only in subgroups), namely 7.9 % versus 10 % on pH-metry [27]. In 2009, in a more comprehensive meta-analysis on various achalasia treatments, the abovementioned randomized trial was not included. This meta-analysis showed reflux reductions from 31.5 % (without fundoplication) to 8.8 % (with fundoplication) in 2507 cases for the laparoscopic procedure, with similar differences from older papers using the open approach [28]. The methods of reflux assessment might have been different in this analysis, but details are not specified.

The large randomized trial of laparoscopic Heller myotomy (LHM) versus balloon dilation did not confirm these low reflux rates of < 10 %, but showed abnormal acid exposure in 23 % for LHM and in 15 % for balloon dilation at 2 years [22], and in 34 % versus 12 % at 5 years [29]. One of the smaller randomized studies comparing open myotomy versus balloon dilation reported abnormal 24-hour pH-metry results in 28 % for surgery versus 8 % for balloon dilation, in 81 patients at 5 years’ follow-up [30]. In contrast, another small randomized trial (n = 50) showed much higher reflux symptom rates for balloon dilation (28 %) compared to LHM (16 %) [31].

So – what to learn from this confusing compilation of evidence of variable quality? Reflux rates (pH-metry) following LHM are in the range of 20 %, perhaps up to 30 % mid-term. In long-term studies, patient numbers are limited, but data on gastroesophageal reflux are available. Remarkably, however, clinical efficacy has not been assessed using the Eckardt score. Each study has developed its own score, mostly centered around different degrees of dysphagia [32] [33] [34] [35]. Reflux rates in the five most significant mid- and long-term studies are detailed in  [Table 2] [32] [33] [34] [35] [36]. On mid-term follow-up of 2 – 5 years (three studies), reflux rates went up to 50 % and 45 % – 65 % required some form of antacid medication. One of the two very long-term studies [35] on open Heller myotomy with Dor hemifundoplication included 67 patients, divided into three small subgroups according to the length of follow-up (80 – 119,120 – 239, and ≥ 240 months). Heartburn increased over time, in rate and severity from 23 % of patients with “occasional” occurrence and 8 % with “frequent” occurrence to 18 % “occasional” and 29 % “frequent.” In the three different groups with increasing follow-up times, 24-hour pH-metry showed abnormal acid reflux in 15 %, 28 %, and 53 %, respectively. A total of 9 patients (13 %) developed Barrett’s esophagus, with occurrence again increasing over time, with 3 patients showing long-segment Barrett’s on endoscopy. Of note, 3 squamous cell carcinomas were detected. In the other surgical paper, the series with the longest follow-up to date [36], results on reflux are reported for Heller myotomy (two thirds of cases initially) and balloon dilation in combination. Only 62 of 150 included patients completed the reflux questionnaire, probably explaining the somewhat deviant and apparently high frequency of acid reflux ([Table 2]). Notably, not only squamous cell, but also, to a lesser extent, adenocarcinoma has been reported in long-term follow-up and database studies [3] [37] [38].

Table 2

Reflux rates in mid- and long-term follow-up series of laparoscopic [32] [33] [34] and open [35] Heller myotomy.


Patients, n


Antacid medication

Endoscopic signs

pH – metry


Vela 2006 [34] [1]

38 months


56 %[1]

56 %

Carter 2011 [33]

62 months


12 %

45 %

Popoff 2012 [32]

59 months


65 %


Csendes 2006 [35]

80 to 240 months


6 % – 29 %


6 – 29 %

31 %

Sawas 2017 [36] [2]

Mean 17.5 years
(range 10 – 40)


61 %

72.5 %

1 88 % were laparoscopic Heller myotomies. Reflux symptoms were 65 % without fundoplication vs. 39 % with fundoplication

2 Series includes 150 patients, with myotomy (n = 112, probably mostly open) and balloon dilatation (n = 38). Only combined results are reported for reflux, while it is stated that Heller myotomy patients had a lower risk (0.37); however in the table provided, heartburn is more frequent with Heller myotomy (37 % vs. 19 %).

So, cutting the lower esophageal sphincter inevitably causes reflux; this can be reduced by adding a hemifundoplication that is done as part of LHM. The surgical long-term data summarized above, however, seem to indicate that this may only be a temporary solution for a variable number of years in the majority of patients. As shown, in some surgical studies reflux rates eventually increase to around 60 %. With POEM, no antireflux procedure is included. Thus, even in the short term follow-up studies, higher initial reflux rates are reported than for LHM, as can be deduced by comparing the different case series and studies available in the literature. These reflux rates may eventually remain steady over time, not exceeding in the long run the incidence of postsurgical myotomy. But this is speculation at present.

Could we thus conclude that reflux finally occurs with any form of myotomy, but earlier (possibly much earlier) with POEM? Yes, probably, but hopefully some answers will be given by the 2-year results of the ongoing randomized trial(s). In an interim analysis of the randomized trial of POEM versus balloon dilation, presented at DDW 2017, endoscopic reflux rates were substantially higher in the POEM group (48.3 % vs. 13.1 %) while efficacy was better at 1 year (92.2 % vs. 70 %; n = 133) [39]. The relevance of the triad of acid reflux, development of Barrett’s esophagus, and perhaps cancer risk for patients treated for achalasia, and the subsequent management of these problems have yet to be determined. We are not sure whether adding laparoscopic antireflux measures (or perhaps endoscopic ones, if proven efficacious) at a later stage in selected POEM patients with more severe reflux will be the solution. In any case, we should keep a close and critical eye on this issue in further studies. Reflux threatens the POEM success story: let’s see whether we can balance out the negative reflux effects, and if so, how we can best achieve this goal.