Introduction
Peroral endoscopic myotomy (POEM) is an accepted treatment modality for achalasia
cardia [1 ]
[2 ]
[3 ]
[4 ]. Results of POEM have demonstrated excellent medium- and long-term success rates
[4 ]
[5 ]
[6 ]. The procedure is minimally invasive, has an excellent safety profile, and can be
performed even in cases of prior treatment failure [7 ]. Results of POEM are superior to single-session endoscopic balloon dilation (EBD)
and are comparable to laparoscopic Heller myotomy (LHM) [8 ]
[9 ]. However, in recent years, there has been increasing concern about post-POEM gastroesophageal
reflux (GER) [10 ]
[11 ]
[12 ]
[13 ]. All treatment modalities for achalasia cardia induce loss of control of the lower
esophageal sphincter, and are therefore prone to post-procedure GER. Conventional
teaching recommends that LHM is always combined with partial or full 360° fundoplication
(LHM + F), more often anterior (Dor fundoplication), to protect against postoperative
GER; overall reported post-LHM + F GER rates have been acceptable [14 ]
[15 ]
[16 ].
To date, fundoplication has not been standard after POEM. Early results of POEM reported
low GER rates (6 %–21 %) [5 ]
[6 ]
[17 ], possibly due to selection bias and limited follow-up data without objective evidence
such as pH studies [4 ]. Recent studies have reported high incidence of post-POEM GER (15 %–88 %) [10 ]
[11 ]
[12 ]
[18 ]. A recent meta-analysis reviewing POEM, LHM, and EBD demonstrated that POEM had
the highest efficacy for relief of dysphagia but also the highest incidence of post-POEM
GER [19 ]. A prospective randomized trial comparing POEM with LHM also demonstrated high incidence
of post-POEM GER [10 ]. Although most post-POEM GER occurrences are mild (grade A or B), severe GER (grade
C or D) or development of Barrett’s esophagus has also been reported [10 ]
[12 ]
[20 ]
[21 ]. Currently, most POEM centers recommend long-term endoscopic surveillance to identify
such changes [22 ]
[23 ].
In an attempt to reduce post-POEM GER, Inoue et al. described a novel endoscopic fundoplication
following anterior POEM, either as a single-stage concomitant procedure (POEM + F)
or later as a separate subsequent procedure (POEF) [24 ]
[25 ]. The procedure aimed to replicate surgical (laparoscopic) Dor partial fundoplication.
In their first series of 21 patients, the authors reported reduction in reflux symptoms
and optimum maintenance of the fundoplication wrap at 1-month follow-up. It is possible
that POEM + F or POEF may help to reduce or control post-POEM GER. The aim of the
current study was to evaluate short-term outcomes of a single-center case series of
patients with achalasia cardia undergoing single-session POEM + F.
Methods
The study was a retrospective analysis of a prospectively maintained database of all
patients with achalasia cardia who underwent POEM + F. Achalasia cardia was diagnosed
by esophagogastroduodenoscopy (EGD) and high-resolution manometry. We previously reported
short-term outcomes of POEM + F in four patients [26 ]; the current study includes these four patients. POEM + F was offered and performed
for naïve achalasia cardia and for all post-EBD recurrences. The following patients
were not offered POEM + F but were offered standard POEM without fundoplication: those
with prior history of LHM; those with a high anesthesia risk (American Society of
Anesthesiologists [ASA] Class III or higher); those with sigmoid achalasia and a grossly
tortuous esophagus (on earlier EGD assessment) in whom POEM was anticipated to be
technically difficult; and those who refused consent for fundoplication.
The detailed technique of POEM + F has been described previously [24 ]
[26 ]. A short description is presented here to ensure understanding in an evolving procedure
([Fig. 1 ], [Video 1 ]). Standard anterior POEM is first performed with a 2–3 cm full-thickness gastric
side myotomy. An ultraslim transnasal endoscope is introduced and passed into the
gastric fundus alongside the gastroscope in the tunnel ([Fig. 1a ]). Using double endoscope transillumination as a guide along with other anatomical
landmarks (diaphragmatic crus), perigastric fat and serosa overlying the gastric myotomy
is dissected and opened to gain entry into the peritoneal cavity ([Fig. 1b ]). The gastroscope is advanced through the peritoneal opening and angled upwards
and leftwards to reach the gastric fundus. The fundus is grasped using a grasper and
retracted into the tunnel to simulate the wrap. Wrap formation is continuously monitored
using the transnasal endoscope retroflexed in the stomach. Once a satisfactory wrap
is confirmed, the corresponding spot on the fundic serosa is marked using diathermy
([Fig. 1c ]). The gastroscope is withdrawn. A 3-cm nylon endoloop is clasped within a reconstrainable
clip and the assembly, along with the gastroscope, is reintroduced through the tunnel
into the peritoneal cavity. The endoloop is fixed to the gastric fundus using four
endoclips ([Fig. 1 d ]). The proximal end of the loop is fixed to the distal end of the myotomy using a
further four clips ([Fig. 1e ]). The endoloop is gradually tightened while monitoring the wrap formation by the
ultraslim scope ([Fig. 1f ]). After complete loop closure, the tails are trimmed using a Loop-cutter (Olympus
Corporation, Tokyo, Japan). After confirming hemostasis, the mucosal entry is closed
using endoclips.
Fig. 1 Salient steps of peroral endoscopic myotomy with fundoplication (POEM + F). a Double endoscope transillumination illustrating that the scope in the tunnel (identified
by the light) is beyond the gastroesophageal junction. Note the post-POEM lax lower
esophageal sphincter on retroflexion. b Peritoneal dissection deep into the gastric myotomy. Note the adjacent perigastric
fat. c Marking the optimum spot on the gastric fundus using a triangular tip knife and soft
coagulation current. d Endoloop being fixed to the fundus using endoclips. e Endoloop being fixed to the distal end of the myotomy using endoclips. f Immediate post-procedure retroflexed view of the completed wrap.
Video 1 Step-by-step peroral endoscopic myotomy with fundoplication (POEM + F) procedure.
After anterior full-thickness myotomy, the anterior peritoneum is opened. The wrap
is simulated under guidance from a transnasal endoscope positioned in parallel in
the stomach, and subsequently completed by fixing the fundus using endoloop and clips
to the distal end of the myotomy.
All patients were admitted to hospital for overnight observation. All patients were
prescribed proton pump inhibitors (PPI) for 2 weeks post-procedure with subsequent
instructions to consume PPI on demand if they experienced GER symptoms. First follow-up
was between 4 and 6 weeks. At follow-up, patients were questioned about dysphagia
and GER symptoms. Eckardt scores and Gastroesophageal Reflux Questionnaire (GerdQ)
scores were calculated. PPI usage was documented including the drug, dosage, and frequency
of administration. All patients were recommended EGD in order to document intactness
of the wrap and presence of endoscopic evidence of esophagitis, and 24-hour ambulatory
pH studies for objective documentation of GER. Both EGD and pH studies were performed
in patients following a minimum of 1 week off PPI medication.
Subsequent follow-up was by periodic telephone standardized questionnaire at 3, 6,
9, and 12 months post-procedure to assess Eckardt and GerdQ scores. Patients experiencing
symptomatic GER and with evidence of erosive esophagitis or abnormal esophageal acid
exposure time (EAET) were recommended long-term PPI therapy.
Outcome assessments and definitions
The primary outcome of the study was the incidence of significant post-POEM + F GER.
Secondary outcomes were clinical resolution of dysphagia, development of intraprocedural
or delayed adverse events, and integrity of the wrap at follow-up.
Technical success was defined as successful completion of POEM + F. Clinical success
was defined as resolution of dysphagia as determined by post-procedure Eckardt score
≤ 3 and absence of post-POEM + F GER. Post-POEM + F GER was defined according to the
Lyon Consensus as EAET > 6 % on pH studies or endoscopic evidence of Grade C or D
esophagitis. Presence of Grade A or B esophagitis and EAET 4 %–6 % was considered
as borderline evidence [27 ]. Presence of significant GER symptoms (GerdQ score ≥ 8) was recorded [28 ]. Adverse events were defined according to the American Society for Gastrointestinal
Endoscopy lexicon for endoscopic adverse events [29 ].
Statistical analysis
Continuous variables were recorded as mean and standard deviation (SD) or median and
interquartile range (IQR), and categorical variables were recorded as counts and percentages.
A P value of < 0.05 was considered statistically significant. Wilcoxon sign rank test
was used to calculate the P value. All statistical analyses were computed using SPSS software version 22 (IBM
Corp, Armonk, New York, USA).
Results
During an 11-month period (March 2019 to January 2020), 25 of 36 patients with achalasia
cardia underwent POEM + F. The patient recruitment, procedure, and follow-up algorithm
is detailed in Fig. 1 s in the online-only Supplementary material. A total of 11 patients were only offered
standard POEM because they were unsuitable for POEM + F (5 patients had post-LHM recurrent
achalasia, 4 patients were high risk for anesthesia [ASA Class III or higher], 1 patient
refused consent, and another patient had severely tortuous sigmoid achalasia for which
POEM was anticipated to be technically difficult and time consuming, and therefore
POEM + F was considered unsuitable because it was likely to prolong the procedure
time significantly).
Baseline demographics and clinical characteristics of patients are summarized in Table 1 s . Mean age was 40.1 (SD 13.7) years and there were 12 females. Mean duration of symptoms
before POEM + F was 12.7 (SD 10.2) months. A total of 23 patients (92.0 %) had type
II achalasia cardia, whereas type I and type III achalasia cardia were only seen in
one patient each. Two patients had recurrent achalasia cardia after prior failed EBD.
All other 23 patients (92.0 %) were naïve achalasia cardia. Mean pre-procedure Eckardt
score was 8.21 (SD 1.08).
Table 1
Perioperative characteristics and outcomes of patients who underwent peroral endoscopic
myotomy with fundoplication.
Successful POEM, n/N (%)
25/25 (100)
Technical success of POEM + F, n/N (%)
23/25 (92.0)
Total procedure time, minutes
115.6 (27.2)
105 (90–270)
Additional time for fundoplication, minutes
46.7 (12.4)
45 (35– 90)
Baseline Eckardt score (pre-POEM + F)
8.21 (1.08)
8
Eckardt score, mean (SD) (post-POEM + F)
0.1 (0.3) (P = 0.001)
Adverse events, n/N (%)
0
3/25 (12.0) (minor)
Hospital stay, mean (SD), days
2.1 (0.5)
POEM, peroral endoscopic myotomy; POEM + F, fundoplication after POEM; SD, standard
deviation; IQR, interquartile range.
Perioperative characteristics and outcomes are detailed in [Table 1 ]. Although POEM was successful in all patients, POEM + F was successful in 23/25
patients (technical success 92.0 %). In two patients, fundoplication could not be
performed after POEM due to device failure in one patient (breaking of the endoloop
after fixation while tightening) and unfavorable anatomy in the other (sliding hiatal
hernia with thick perigastric pad of fat). Mean and median total procedure times for
POEM + F were 115.6 (SD 27.2) minutes and 105 (IQR 90–270) minutes, respectively,
of which mean 46.7 (SD 12.4) minutes and median 45 (IQR 35–90) minutes were additionally
required for fundoplication. The moving average curve demonstrated significant improvement
in total procedure time after the first five procedures (88 [SD 23.4] minutes for
first five cases vs. 51.2 [SD 9.1] minutes for subsequent cases; P < 0.05) (Fig. 2 s ). No intra- or immediate post-procedure adverse events related to POEM + F were encountered.
Mean post-procedure hospital stay was 2.1 (SD 0.5) days.
Clinical follow-up was available for all 25 patients. Median duration of follow-up
was 12 months (IQR 9–13). Significant improvement in dysphagia was seen in all patients
(mean post-procedure Eckardt score 0.1 [SD 0.3] vs. mean pre-procedure score 8.21
[SD 1.08]; P = 0.001) ([Table 1 ]).
GER was evaluated in 23 patients who underwent successful POEM + F ([Table 2 ]). Two patients who underwent only POEM were excluded from the GER evaluation. GerdQ
scores for all 23 patients were obtained at a median of 12 months (IQR 9–13) after
POEM + F. Scores were < 8 in 22/23 patients (95.7 %) signifying low probability of
GER [28 ]
[30 ]. Regular PPI use was not reported by any patients. Post POEM + F 24-hour pH studies
were obtained in 18/23 patients (78.3 %) at a median of 2 months (IQR 1.5–6.75). All
pH studies were performed with patients after 1 week off PPI. DeMeester scores and
EAET were positive for GER in 2 patients (11.1 %) (EAET > 6 %) whereas these values
were normal in the remaining 16 patients (88.9 %) (EAET < 6 %). The symptom index
demonstrated 100 % and 80 % correlation, respectively, in the two patients with positive
scores. For the remaining 16 patients, mean EAET was 1.12 (SD 0.7). Meaningful symptom
index calculation could not be performed in these patients due to the low values.
Table 2
Evaluation of gastroesophageal reflux during follow-up in patients undergoing peroral
endoscopic myotomy with fundoplication.
GER Parameter
n/N (%)
Total N
23
GerdQ score ≥ 8
1/23 (4.3)
EGD findings available
22/23 (95.7)
Esophagitis LA Grade A
4/22 (18.2)
Wrap integrity
19/23 (82.6)
1/23 (4.3)
2/23 (8.7)
1/23 (4.6)
24-hour ambulatory pH studies
18/23 (78.3)
2/18 (11.1)
2/18 (11.1)
GER, gastroesophageal reflux; EGD, esophagogastroduodenoscopy; EAET, esophageal acid
exposure time.
Follow-up EGD was performed in 22 patients (95.7 %) at a mean of 2.78 (SD 3.54) months
post-procedure (median interval 1.5 months, IQR 1–2). All EGDs were performed after
patients had been off PPI for at least 1 week. One patient refused EGD because of
lack of symptoms. EGD revealed an intact wrap in 19/22 patients (86.4 %) ([Fig. 2a ]). The wrap was not distinctly visible (indistinct) in two patients ([Fig. 2b ]), whereas it appeared loose with an open lower esophageal sphincter in one patient
([Fig. 2c ]). Overall, the wrap was found to be intact in 19/23 patients who underwent successful
POEM-F (82.6 %). Erosive esophagitis (Grade A) was seen in 4/22 patients (18.2 %),
two of whom had an indistinct wrap. Notably, all four patients had normal EAET and
Demeester scores, and none reported GER symptoms.
Fig. 2 Retroflexed view of the gastric fundus during follow-up esophagogastroduodenoscopy.
a An intact wrap (image comparable to the image of peroral endoscopic myotomy with
fundoplication immediately post-procedure (Fig. 1f above) – elevated transverse ridge
of the wrap is visible; minimal or no esophageal mucosa is seen alongside the endoscope
exiting the gastroesophageal junction. b An indistinct wrap – transverse ridge corresponding to the wrap is not visible or
only very faintly visible. c A loose wrap – transverse ridge is visible but esophageal mucosa can be easily seen
alongside endoscope.
Three delayed adverse events were reported (12.0 %): EGD revealed that an endoclip
(one patient) and the cut ends of the endoloop (two patients) had eroded through the
mucosa overlying the submucosal tunnel in the distal esophagus or at the gastroesophageal
junction. In all three patients, surrounding esophageal mucosa had completely healed
around the clip or loop without any residual visible mucosal breach. No patient reported
any directly attributable symptoms to these findings.
Follow-up of the two patients with failed POEM + F revealed an elevated GerdQ score
and Grade B esophagitis in one patient, whereas the other patient was asymptomatic
and did not have erosive esophagitis on EGD. Both patients did not undergo pH studies.
Long-term PPI was prescribed to the patient reporting GER.
Post-POEM + F GER as per the Lyon Consensus was identified in two patients (11.1 %;
95 % confidence interval 10.6–11.5). One of these patients had a loose wrap on EGD
and was also significantly symptomatic (GerdQ score 11). Neither patient demonstrated
erosive esophagitis on EGD. Borderline evidence of GER as evidenced by Grade A esophagitis
was additionally observed in four patients but pH studies and symptom scores were
all normal (Table 2 s ).
Discussion
Although POEM is a well-established treatment modality for achalasia cardia [1 ]
[2 ]
[3 ], post-POEM GER has been a significant challenge in recent years [10 ]
[11 ]
[12 ]
[18 ]. POEM has therefore been criticized by surgical experts and several other sources
[10 ]
[11 ]
[12 ]
[13 ]. Post-POEM GER is often asymptomatic, may remain undiagnosed, and surveillance endoscopies
and long-term PPI have been recommended for all post-POEM patients [23 ]. Long-term effects of such GER are largely unknown, but occasional reports of post-POEM
Barrett’s esophagus and cancer have raised significant alarm [31 ].
In comparison, LHM – the surgical counterpart of POEM – has always incorporated a
partial or complete 360° fundoplication as an integral part of the surgical procedure.
The addition of fundoplication to LHM has demonstrated significant reduction in the
incidence of post-operative GER [16 ]
[32 ]
[33 ]. GER in post-LHM + F patients is significantly less frequent compared with that
after POEM [19 ]
[21 ].
Based on the principle of LHM + F, Inoue et al. reported POEM + F and POEF as endoscopic
antireflux procedures following POEM [24 ]. The fundoplication wrap in POEM + F closely mimics the Dor partial fundoplication
– a standard recommended procedure following LHM. The initial series of Inoue et al.
included 21 patients and reported safety, feasibility, and short-term outcomes following
POEM + F; results demonstrated an intact wrap in 95 % of patients at 1-month follow-up. The
current study reports short-term outcomes following POEM + F. POEM was effective in
all 25 patients as evidenced by significant improvement in dysphagia and pre- and
post-POEM Eckardt scores, whereas POEM + F was successful in 23 patients (92.0 %).
The primary aim of this study was to calculate the incidence of significant post-POEM + F
GER. Our results demonstrate that significant GER (EAET > 6 %) was documented in only
2/18 patients (11.1 %) with excellent symptom index correlation. This is much lower
than that documented in other POEM studies [10 ]
[11 ]
[12 ]
[18 ]. These results are promising and suggest that POEM + F could protect against post-POEM
GER, although additional studies are required to confirm these results.
Wrap integrity at follow-up was another parameter assessed during the current study.
The wrap appeared to be intact in 19 patients (82.6 %) at follow-up EGD at a median
1.5 months. POEM + F is based on surgical principles that depend on tissue healing,
fibrosis, and adhesion formation to maintain wrap integrity. As tissue healing is
usually completed by 6 weeks, it is likely that a wrap intact at 6 weeks will maintain
integrity in the long term. The wrap was loose in one patient (abnormal EAET > 6 %
and elevated GerdQ score confirming GER) and indistinct in two patients (both reported
Grade A esophagitis). It is therefore important to note that all three patients with
wrap failures had either conclusive or borderline evidence of GER on follow-up evaluation.
On the other hand, the fact that all 19 patients with an intact wrap reported normal
GerdQ scores at a median of 12 months’ follow-up suggests that the wrap could be an
effective barrier against GER.
Grade A erosive esophagitis was found in four patients, two of whom had wrap failure
whereas it was intact in the other two. However, as GerdQ scores and EAET were normal
in all four patients, it is impossible to determine whether the esophagitis was due
to GER or food stasis.
Diagnosis of post-POEM GER is challenging, and the guidelines presented by the Lyon
Consensus are possibly most relevant in this regard [27 ]. The results of the present study demonstrate conclusive evidence of GER in 11.1 %
patients and if both conclusive and borderline evidence were to be considered – 26.1 %.
Of these, only one patient was symptomatic (1/23, 4.3 %), four had endoscopic evidence
(4/22, 18.2 %), whereas EAET was abnormal in 2/18 (11.1 %). In comparison, symptomatic
post-POEM GER has been reported in 19 %, by pH studies in 39 %, and on endoscopy in
29.4 % in a recent large meta-analysis [11 ]. It appears, therefore, that POEM + F could be protective against post-POEM GER.
Fundoplication has been a standard procedure for prevention of postoperative GER after
LHM. In a large meta-analysis of 64 articles reporting outcomes of 4871 LHM + F procedures,
Campos et al. reported a mean postoperative GER incidence of 8.8 % (range 0–44) [16 ]. The current study demonstrated 11.1 % post-POEM + F GER incidence, which is comparable
to that reported for LHM + F. Further validation by larger randomized studies is necessary
to confirm these findings.
Is POEM + F reproducible? We previously reported an initial series of four patients
undergoing successful POEM + F [26 ]. Further to this, the current series reported 92.0 % (23/25) technical success of
POEM + F. No intraprocedural adverse events were encountered, although technical failures
occurred in two patients. In the first patient, the endoloop snapped while tightening
the wrap, possibly owing to friction from an endoclip. Although it was technically
feasible to redo the fundoplication using a new endoloop and clips, we preferred not
to pursue this in the interest of patient safety owing to the increased anesthesia
time that would be required. The other patient had a sliding hiatal hernia with an
angulated anatomy at the gastroesophageal junction, and a thick pad of fat at the
point of entry into the peritoneal cavity. Dissection and entry into the peritoneal
cavity and subsequent visualization was considered difficult and could likely compromise
procedural safety. Fundoplication was therefore aborted to avoid inadvertent intraperitoneal
injury or adverse events. Both these patients were among our initial five patients
when experience with this technique was limited. No technical challenges were encountered
in the remaining 23 patients. Additional operative time was necessary to perform POEM + F;
however, the learning curve was identified to be short (five cases). POEM + F can
therefore be considered a safe procedure with acceptably low failure rates, and these
may further reduce with increased experience.
Delayed adverse events were reported in our study, involving a clip or endoloop tail
eroding through the mucosa in three patients (12.0 %); no other adverse events were
encountered. No endoscopically visible mucosal breach was identified in any of these
patients and the mucosa had completely healed around these inclusions. Patients had
no related symptoms and therefore these findings have doubtful clinical significance.
Currently no consensus exists regarding the optimal approach for the prevention of
post-POEM GER. Most patients are maintained on PPI with good symptom control [34 ]. Endoscopic antireflux procedures have been reported to be used in conjunction with
POEM in an attempt to reduce post-POEM GER. Tyberg et al. reported a series of five
patients treated with second-session transoral incisionless fundoplication performed
after POEM, which demonstrated impressive short-term results [35 ]. The same group then also reported a single-session approach in a solitary case
report [36 ]. Toshimori et al. reported on POEF (POEM + F performed over two independent sessions)
as a subsequent second procedure to control post-POEM GER [25 ]. The authors also demonstrated use of endoscopic suturing instead of endoloop and
clips to create the wrap. It will be interesting to evaluate which of these techniques
can provide a more robust and secure fundoplication in the long term.
Although the two-session approaches are plausible alternatives, they have the inherent
drawback of requiring two separate procedures, thereby increasing procedure costs,
invasiveness, and potential morbidity. They may therefore be less preferred to POEM + F
or LHM + F in routine practice. In contrast, POEM + F offers the advantage of a single-session
fundoplication akin to what is practiced and recommended for surgical myotomy. The
procedure is also less device dependent, and therefore costs are likely to be lower.
Although it may be argued that not all POEM patients develop GER and therefore an
antireflux procedure may not be required in all, the reported high incidence of post-POEM
GER (especially asymptomatic GER), along with the fact that fundoplication has been
a standard procedure following LHM for several decades, surely justify the approach
of single-session POEM + F. Laparoscopic fundoplication has also been described after
POEM to treat post-POEM GER [34 ]. Although effective, this approach defeats the primary purpose of performing a less-invasive
procedure such as POEM if a more invasive approach is later required to control its
side-effects.
The current study has several limitations. First, it was a retrospective case series
from a single center, and therefore these results need confirmation in larger multicenter
trials. Second, nearly one-third of eligible patients (11/36, 30.6 %) were not offered
POEM + F, and therefore possible bias exists regarding patient selection. Third, although
clinical follow-up was obtained in all patients, follow-up EGD and pH studies could
not be obtained for all. It is possible that asymptomatic GER could have been missed
in these patients. Furthermore, we understand that nonavailability of impedance pH
may limit interpretation of some of the results. Fourth, the study was a single-arm
study without a POEM-only comparator arm. Future comparative studies will be required
to address this issue. Finally, POEM + F is an evolving procedure. It is therefore
likely that with increasing experience, the technique may be simplified, and that
the results and outcomes may demonstrate more solidarity in future.
In conclusion, this single-center study demonstrated excellent short-term outcomes
of POEM + F for the control of post-POEM GER. The procedure was reproducible and safe,
and the integrity of the wrap was well maintained at follow-up. Incidence of post-POEM + F
GER at a median 12-month follow-up was low and acceptable. POEM + F offers the advantage
of a single-session endoscopic solution for achalasia cardia treatment with possible
control of GER. These findings require further validation in larger randomized studies
with longer follow-up.
