10 1056@NEJMoa1905380

The n e w e ng l a n d j o u r na l of m e dic i n e
Original Article
Endoscopic or Surgical Myotomy in Patients

with Idiopathic Achalasia
Yuki B. Werner, M.D., Bengt Hakanson, M.D., Jan Martinek, M.D.,
Alessandro Repici, M.D., Burkhard H.A. von Rahden, M.D.,
Albert J. Bredenoord, M.D., Raf Bisschops, M.D., Helmut Messmann, M.D.,
Marius C. Vollberg, M.Sc., Tania Noder, R.N., Jan F. Kersten, M.Sc.,
Oliver Mann, M.D., Jakob Izbicki, M.D., Alexander Pazdro, M.D.,
Uberto Fumagalli, M.D., Riccardo Rosati, M.D., Christoph‑Thomas Germer, M.D.,
Marlies P. Schijven, M.D., Alice Emmermann, M.D., Daniel von Renteln, M.D.,
Paul Fockens, M.D., Guy Boeckxstaens, M.D., and Thomas Rösch, M.D.
A BS T R AC T
BACKGROUND
Pneumatic dilation and laparoscopic Heller’s myotomy (LHM) are established treatments The authors’ affiliations are listed in the
for idiopathic achalasia. Peroral endoscopic myotomy (POEM) is a less invasive therapy Appendix. Address reprint requests to Dr.
Rösch at the Department of Interdisci-
with promising early study results. plinary Endoscopy, University Hospital
METHODS Hamburg-Eppendorf, Martinistr. 52, 20246
Hamburg, Germany, or at t.roesch@uke.de.
In a multicenter, randomized trial, we compared POEM with LHM plus Dor’s fundoplica-
tion in patients with symptomatic achalasia. The primary end point was clinical success, N Engl J Med 2019;381:2219-29.
DOI: 10.1056/NEJMoa1905380
defined as an Eckardt symptom score of 3 or less (range, 0 to 12, with higher scores indi- Copyright © 2019 Massachusetts Medical Society.
cating more severe symptoms of achalasia) without the use of additional treatments, at the
2-year follow-up; a noninferiority margin of −12.5 percentage points was used in the pri-
mary analysis. Secondary end points included adverse events, esophageal function, Gas-
trointestinal Quality of Life Index score (range, 0 to 144, with higher scores indicating
better function), and gastroesophageal reflux.
RESULTS
A total of 221 patients were randomly assigned to undergo either POEM (112 patients)
or LHM plus Dor’s fundoplication (109 patients). Clinical success at the 2-year follow-up
was observed in 83.0% of patients in the POEM group and 81.7% of patients in the LHM
group (difference, 1.4 percentage points; 95% confidence interval [CI], −8.7 to 11.4;
P = 0.007 for noninferiority). Serious adverse events occurred in 2.7% of patients in the
POEM group and 7.3% of patients in the LHM group. Improvement in esophageal func-
tion from baseline to 24 months, as assessed by measurement of the integrated relax-
ation pressure of the lower esophageal sphincter, did not differ significantly between
the treatment groups (difference, −0.75 mm Hg; 95% CI, −2.26 to 0.76), nor did improve-
ment in the score on the Gastrointestinal Quality of Life Index (difference, 0.14 points;
95% CI, −4.01 to 4.28). At 3 months, 57% of patients in the POEM group and 20% of
patients in the LHM group had reflux esophagitis, as assessed by endoscopy; at 24
months, the corresponding percentages were 44% and 29%.
CONCLUSIONS
In this randomized trial, POEM was noninferior to LHM plus Dor’s fundoplication in
controlling symptoms of achalasia at 2 years. Gastroesophageal reflux was more common
among patients who underwent POEM than among those who underwent LHM. (Funded
by the European Clinical Research Infrastructure Network and others; ClinicalTrials.gov
number, NCT01601678.)
n engl j med 381;23 nejm.org December 5, 2019 2219

The New England Journal of Medicine
Downloaded from nejm.org on December 5, 2019. For personal use only. No other uses without permission.
Copyright © 2019 Massachusetts Medical Society. All rights reserved.
A
chalasia is an esophageal motor Patients
disorder consisting of defective relaxation Patients with symptomatic achalasia and a med-
of the lower esophageal sphincter and ical indication for surgical myotomy or pneu-
disturbed esophageal peristalsis. The clinical matic dilation were eligible for inclusion in the
symptoms associated with the condition include trial if they were 18 years of age or older, had
dysphagia, regurgitation, chest pain, and weight an Eckardt symptom score11 higher than 3 (a de-
loss.1 Current treatment options include endo- scription of the scale is provided below), and
scopic botulinum toxin injection, endoscopic had findings on preinterventional manometry
pneumatic dilation, and surgical laparoscopic that were consistent with the diagnosis of acha-
Heller’s myotomy (LHM).2 These therapies allevi- lasia (classified as subtypes I to III).12 Patients
ate symptoms, mainly by reducing lower esopha- who had undergone previous surgery of the
geal sphincter pressure and lowering the resis- stomach or esophagus, including surgical ther-
tance to flow at the esophagogastric junction. On apy of achalasia, or who had received a diagno-
the basis of findings from randomized trials,3,4 sis of secondary achalasia or other organic
recent consensus statements and guidelines sup- causes of dysphagia were excluded. Eligible pa-
port both pneumatic dilation and LHM as effec- tients who had previously undergone endoscopic
tive primary therapies for achalasia without a treatment were not excluded. All patients pro-
clear preference for either method.2,5 vided written informed consent. Further details
Peroral endoscopic myotomy (POEM) is a of the inclusion and exclusion criteria are pro-
purely endoscopic (scarless) method of myotomy6 vided in Tables S1 through S3 in the Supple-
that has been used in clinical practice since mentary Appendix.
2010.7 Since then, a multitude of studies, most of
which were retrospective,5,8-10 have shown the high Interventions
clinical efficacy and safety of POEM. The aim of The patients at each center were randomly as-
this randomized, noninferiority trial was to com- signed in a 1:1 ratio to undergo either POEM or
pare POEM with LHM plus Dor’s fundoplication LHM plus Dor’s fundoplication. Randomly per-
in patients with idiopathic achalasia. muted blocks of varying sizes, with separate
blocks for each center, were used to balance
group assignments according to center. A trial
Me thods
nurse who was unaffiliated with the research
Trial Design group and otherwise not involved with the trial
This was a prospective, multicenter, randomized, performed the randomization.
open-label, noninferiority trial performed at eight POEM was performed as previously described
centers in six European countries. The trial was by Inoue et al.7 and involved the creation of an
approved by the institutional review board at esophageal submucosal tunnel, which was ex-
each participating center. On-site data monitor- tended 2 to 3 cm into the gastric cardia (Fig. 1).
ing was provided by Clinical Trial Center North The length of myotomy was adjusted according
at the University Hospital Hamburg-Eppendorf to the achalasia subtype (i.e., 6 to 7 cm above
and the European Clinical Research Infrastruc- the lower esophageal sphincter in type I and II
ture Network (ECRIN). Monitoring visits were and calibrated according to manometric extent
performed (details are provided in Table S9 in in type III). The endoscopists had sufficient
the Supplementary Appendix, available with the experience in therapeutic endoscopy, including
full text of this article at NEJM.org). A safety esophageal interventions such as endoscopic mu-
board reviewed all adverse events. Various public cosal resection and submucosal dissection (each
foundations and Olympus Europa supported the had performed more than 50 procedures), and
trial. None of the sponsors had any role in the received formal POEM training that included the
design of the trial or in the analysis or interpre- use of animal or endoscopic training models.
tation of the data. The authors vouch for the The first 4 to 5 procedures were supervised by a
completeness and accuracy of the data and for tutor, and each endoscopist had to independently
the fidelity of the trial to the protocol, available perform 8 to 10 run-in POEM procedures before
at NEJM.org. the start of the trial. At each center, POEM was
2220 n engl j med 381;23 nejm.org December 5, 2019

Endoscopic or Surgical Myotomy for Idiopathic Achalasia
A POEM B LHM
LIVER
DIAPHRAGM
DIAPHRAGM
Lower esophageal
sphincter STOMACH STOMACH
Figure 1. Alternative Methods of Treatment for Achalasia.

Panel A shows the peroral endoscopic myotomy (POEM) approach, in which transection of the muscular fibers of the
lower esophageal sphincter is performed entirely with a standard endoscope. Panel B shows the surgical procedure
by laparoscopic Heller’s myotomy (LHM); a fundoplication is added to this procedure.
performed by one endoscopist (six centers) or two tion by means of endoscopy, manometry, and
endoscopists (two centers). esophageal pH monitoring (at least 1 week after
LHM was performed according to current the discontinuation of a proton-pump inhibitor)
standards.3 Surgical myotomy was performed by was planned at 3 and 24 months (Table S4).
dividing the muscle fibers of the lower esopha-
geal sphincter and extending the division to at Trial End Points
least 6 cm into the esophageal side and at least The primary end point was clinical success, de-
2 to 3 cm into the gastric cardia (Fig. 1). Ante- fined as an Eckardt symptom score of 3 or less
rior fundoplication with the use of the technique without the use of additional treatments, at the
described by Dor was routinely performed. At 2-year follow-up. The Eckardt symptom score11 is
each center, LHM was performed by one surgeon a validated questionnaire and is calculated as the
(five centers) or two surgeons (three centers); sum of symptom scores of four components of
each surgeon had performed at least 20 LHM achalasia including dysphagia, regurgitation,
procedures. Further details of POEM and LHM chest pain, and weight loss. Each component
plus Dor’s fundoplication are provided in the can be graded from 0 to 3 points. The maximum
protocol. Eckardt symptom score is 12 points, with higher
scores indicating more severe symptoms. A good
Trial Follow-up outcome is classified as an Eckardt symptom score
Clinical data were collected at follow-up visits at of 3 or less. All patients reported an Eckardt
3, 6, 12, and 24 months. Patient-reported out- symptom score higher than 3 at baseline. The
comes were assessed by means of telephone primary hypothesis was that POEM would be
calls, mail, or follow-up appointments by dedi- noninferior to LHM plus Dor’s fundoplication
cated trial personnel who were aware of the with respect to the primary end point.
treatment-group assignments. Objective evalua- Prespecified secondary end points included

clinical measures regarding symptoms, Gastro- POEM and LHM plus Dor’s fundoplication, we
intestinal Quality of Life Index score (range, 0 to estimated that each treatment group would re-
144, with higher scores indicating better func- quire 99 patients who could be evaluated in order
tion),13 and gastroesophageal reflux.14 Objective for the trial to have 80% power to detect non-
measures included the grading of endoscopic inferiority with the use of noninferiority margin
reflux lesions according to the Los Angeles Clas- of −12.5 percentage points with respect to the
sification15; assessment of esophageal function, primary end point. Interventional gastroenterolo-
as measured by calculation of the integrated re- gists and surgeons involved in this trial consid-
laxation pressure of the lower esophageal sphinc- ered a margin of −12.5 percentage points clini-
ter by means of manometry; and 24-hour pH cally acceptable. Feasibility was another factor,
monitoring. Gastroesophageal reflux was con- given the low incidence of achalasia (0.3 to 1.6
sidered abnormal if the acid exposure time (total cases per 100,000 persons per year among adults1)
percentage of time with pH <4) was greater than and the willingness of patients to participate in
4.5%.3 Further details of these clinical measures trials comparing endoscopy with surgery. We
are provided in Tables S5 through S7. Additional initially aimed to enroll 220 patients to compen-
secondary end points included adverse events sate for a 10% dropout rate. In August 2015,
(i.e., complications such as bleeding, perforation after 214 patients had undergone randomization,
or mucosal damage, ulcerations, prolonged pain we increased the sample size to 240 because of
that occurred during or immediately after the a dropout rate that was higher than anticipated.
procedure) and serious adverse events (i.e., com- We performed analyses primarily in the mod-
plications of the procedure that led to relevant ified intention-to-treat population, which includ-
additional interventional measures or reinterven- ed all patients who underwent randomization
tions during or after the procedure, led to or pro- and the assigned treatment. The per-protocol
longed inpatient hospitalization, or led to ad- population included all patients in the modified
mission to the intensive care unit or death) intention-to-treat population who completed
(Table S8),16-18 procedure time, length of myot- follow-up without any major protocol deviation.
omy, hospitalization time after intervention, We used multiple imputation to account for
laboratory data, and treatment failure. Additional missing Eckardt symptom scores for patients
procedural data, use of proton-pump inhibitors, who were lost to follow-up or withdrew from the
and endoscopic findings at baseline were evalu- trial during follow-up; predictive mean match-
ated post hoc. ing was used to account for the skewed distribu-
The statistical analysis plan (available with tion of Eckardt symptom scores (100 imputa-
the protocol) specified that clinically relevant tions with 1000 iterations each; five cases per
exploratory subgroup analyses would be per- match set). The following variables were used to
formed. Exploratory subgroups were defined ac- impute missing Eckardt symptom scores: avail-
cording to age (<40 or ≥40 years), sex, achalasia able Eckardt symptom scores (including base-
subtype (I, II, or III), and previous treatment for line), achalasia subtype, previous treatment, body-
achalasia (yes or no). Secondary end points and mass index, age, and sex.
post hoc outcomes are listed in Table S16. The primary analysis compared the percent-
ages of patients in the POEM group who had
Statistical Analysis clinical success at the 2-year follow-up with that
Sample size was calculated on the basis of a in the LHM group. POEM was considered non-
noninferiority design in which POEM was com- inferior to LHM if the lower limit of the two-
pared with LHM plus Dor’s fundoplication. Avail- sided 95% confidence interval of the absolute
able literature on treatments for achalasia showed between-group difference (POEM minus LHM)
a success rate of 86% (95% confidence interval in the percentage of patients who had clinical
[CI], 85 to 88) after LHM.19 The clinical success success was above −12.5 percentage points. Sev-
rate of LHM plus Dor’s fundoplication was ob- eral sensitivity analyses involving all patients in
served to be 90% (95% CI, 84 to 96) at the 2-year the modified intention-to-treat population were
follow-up in the European Achalasia Trial in performed, including a complete-case analysis,
2011.3 Assuming success rates of 90% for both a last-observation-carried-forward analysis, and

a maximum-likelihood analysis with an expecta- during the trial period but who did not consent
tion-maximization algorithm. Identical analyses to participate in the trial is provided in Table S9.
were applied to the per-protocol data set (see the Demographic data, symptom scores, and Gas-
Supplementary Appendix). trointestinal Quality of Life Index scores at base-
In exploratory subgroup analyses, we used line were similar in the two treatment groups
logistic-regression models to evaluate interactions (Table 1). Overall, 108 of 112 patients (96.4%) in
between subgroup variables and treatment group. the POEM group and 104 of 109 patients (95.4%)
Time courses of dichotomous and continuous in the LHM group had complete follow-up results
variables were investigated with the use of mixed- with respect to the primary end point.
effects regression models and generalized esti-
mating equations, which were chosen on the Clinical Success
basis of goodness of fit. Considering the distri- In the modified intention-to-treat population, 93
bution of the outcome variable, baseline levels, of 112 patients (83.0%) in the POEM group and
and repeated measurements per participant, we 89 of 109 patients (81.7%) in the LHM group had
included age, sex, achalasia subtype, and previ- clinical success at the 2-year follow-up (the pri-
ous treatment as covariates in these models. Post mary end point). Missing data on clinical suc-
hoc analyses included calculation of the relative cess at 2 years were imputed for 4 patients in the
differences in the rate of clinical success, sensi- POEM group and 5 patients in the LHM group.
tivity analysis involving all patients who under- The percentage of patients who had clinical suc-
went randomization (tipping-point analysis in cess at 3 months after the assigned intervention
the intention-to-treat population), and an inter- was 94.6% (95% CI, 88.2 to 97.8) in the POEM
action test comparing Eckardt symptom scores group and in 89.0% (95% CI, 81.2 to 93.9) in the
over time. LHM group. The percentages of patients who
Confidence intervals are not adjusted for mul- had clinical success over time are shown in Fig-
tiple comparisons and cannot be used to draw ure 3A. Similar results were obtained in the per-
inferences about effects. Analyses were conducted protocol population, with 82.4% of patients in
with the use of statistical software package R, the POEM group and 80.6% of patients in the
version 3.5.1 (R Foundation for Statistical Com- LHM group having clinical success at 2 years
puting), and SPSS software, version 22 (IBM). (Table S10).
In the modified intention-to-treat population,
the absolute between-group difference in the per-
R e sult s
centage of patients who had clinical success at
Patients the 2-year follow-up was 1.4 percentage points
Between December 7, 2012, and October 9, 2015, (95% CI, −8.7 to 11.4) in favor of POEM
a total of 241 patients were enrolled at eight (P = 0.007 for noninferiority). The lower bound-
European centers and underwent randomization ary of the 95% confidence was above the pre-
after providing written informed consent (Fig. 2). specified noninferiority margin of −12.5 per-
Of the 241 patients enrolled, 20 were excluded centage points (Fig. 3B). In the per-protocol
from the trial, and 221 underwent the assigned population, the corresponding between-group
treatment (112 in the POEM group and 109 in difference was 1.8 percentage points (95% CI,
the LHM group [the modified intention-to-treat −8.7 to 12.3) in favor of POEM. These results
population]). In the modified intention-to-treat were maintained in prespecified sensitivity analy-
population, 4 patients were excluded from the ses, and a post hoc tipping-point analysis in the
POEM group and 6 patients (including 1 patient intention-to-treat population (i.e., all patients who
who had manometric findings at 3 and 24 months underwent randomization) showed that noninfe-
that were not consistent with achalasia) from the riority would be supported in 88% of imputation
LHM group; thus, the per-protocol population scenarios (Fig. S2).
comprised 108 patients in the POEM group and In an analysis of clinical success across time
103 patients in the LHM group. An overview of points, which was performed with the use of a
the distribution of patients across trial centers mixed-effects logistic-regression model, the odds
and of the patients who underwent treatment ratio for clinical success in the POEM group, as

241 Patients underwent randomization

(1:1 ratio)
120 Were assigned to the POEM group 121 Were assigned to the LHM group
8 Were excluded
12 Were excluded
2 Withdrew consent
6 Withdrew consent
2 Did not undergo treatment
5 Did not undergo treatment
4 Had exclusion criteria
1 Underwent acupuncture
discovered after
randomization
112 Underwent POEM 109 Underwent LHM

(modified intention-to-treat population) (modified intention-to-treat population)
6 Were excluded
4 Were lost to follow-up
4 Were excluded
1 Withdrew consent
2 Were lost to follow-up
1 Had exclusion criteria
2 Withdrew consent
discovered after
randomization
108 Had no major protocol deviation 103 Had no major protocol deviation
and completed follow-up and completed follow-up
(per-protocol population) (per-protocol population)
112 Were included in the modified 109 Were included in the modified
intention-to-treat analyses intention-to-treat analyses
108 Were included in the per-protocol 103 Were included in the per-protocol
analyses analyses
Figure 2. Enrollment, Randomization, and Follow-up.

Patients from eight European centers were randomly assigned to undergo either POEM (the POEM group) or LHM
plus Dor’s fundoplication (the LHM group). Analyses involved 221 patients in the modified intention-to-treat popu-
lation, which included all patients who underwent randomization and the assigned treatment, and 211 in the per-
protocol population, which included all patients in the modified intention-to-treat population who completed follow-
up without any major protocol deviation. A comparison of patients who withdrew from the trial with those who were
included in the primary analysis is provided in Table S15.
compared with the LHM group, was 1.28 (95% ceived any treatment for achalasia, 65 of 73 (89%)
CI, 0.85 to 1.92). Exploratory subgroup analyses in the POEM group and 58 of 69 (84%) in the
of the primary end point did not suggest evi- LHM group had clinical success at 2 years, and
dence of any interaction on the basis of age, sex, the corresponding proportions of patients who
achalasia subtype, and previous treatment (Fig. had clinical success at 2 years among those who
S3). The results in the modified intention-to-treat had previously received any treatment (botulinum
population, as summarized descriptively, showed toxin injection, pneumatic dilation, or both) (Ta-
that 10 of 12 patients (83%) with achalasia sub- ble 1) were 28 of 39 (72%) and 31 of 40 (78%).
type III in the POEM group and 7 of 9 such pa- To address the concern that lower-than-
tients (78%) in the LHM group had clinical suc- expected rates of clinical success may bias non-
cess at 2 years after the assigned intervention. inferiority analyses when only absolute differ-
Among the patients who had not previously re- ences were considered, a post hoc analysis of the

relative rate of clinical success revealed that the Table 1. Demographic and Clinical Characteristics of the Patients at Baseline.*
lower boundary of the 95% confidence interval
of the odds of clinical success at the 2-year fol- POEM Group LHM Group
Characteristic (N = 112) (N = 109)
low-up was 0.90 (odds ratio, 1.02; 95% CI, 0.90
to 1.15). This lower boundary was above the Age — yr 48.6±14.9 48.6±14.6
fixed ratio of 0.86 that corresponded with the Male sex — no. (%) 68 (60.7) 60 (55.0)
prespecified noninferiority margin of the abso- Body-mass index† 24.8±4.6 24.5±4.5
lute between-group difference in the percentage
Esophageal function according to integrated 26.8±11.4 26.0±10.9
of patients who had clinical success when a suc- relaxation pressure — mm Hg‡
cess rate of 90% was assumed and was thus
Achalasia subtype — no. (%)
consistent with noninferiority.
I 15 (13.4) 21 (19.3)
Secondary End Points II 85 (75.9) 78 (71.6)
The between-group difference (POEM minus III 12 (10.7) 9 (8.3)
LHM) in the mean Eckardt symptom scores Subtype unclassified 0 1 (0.9)
across time points was 0.16 points (95% CI, Previous therapy — no. (%)
−0.04 to 0.36) in favor of POEM (Fig. S4). A total
None 73 (65.2) 69 (63.3)
of 11 patients had persistent symptoms after
Endoscopic pneumatic dilation 27 (24.1) 31 (28.4)
undergoing the assigned intervention. A reinter-
vention was performed in 2 of 3 patients in the Endoscopic botulinum toxin injection 7 (6.2) 8 (7.3)
POEM group in whom the initial intervention Pneumatic dilation and botulinum toxin 5 (4.5) 1 (0.9)
had failed and in all 8 patients in the LHM injection
group in whom the initial intervention had Eckardt symptom score§ 6.8±2.0 6.7±2.0
failed. Recurrence of symptoms of achalasia was Gastrointestinal Quality of Life Index score¶ 89.2±23.1 90.4±18.1
recorded in an additional 16 patients in the
POEM group and 12 patients in the LHM group; * Plus–minus values are means ±SD. Baseline characteristics were similar in
the two treatment groups. Percentages may not total 100 because of round-
in 1 patient in the POEM group, these symptoms ing. LHM denotes laparoscopic Heller’s myotomy, and POEM peroral endo-
occurred because of reflux, and a Dor’s fundopli- scopic myotomy.
cation was planned at the 2-year visit (Table S11). † The body-mass index is the weight in kilograms divided by the square of the
height in meters.
Improvement in esophageal function, as as- ‡ Esophageal function was assessed objectively by measurement of the integrated
sessed objectively by measurement of the inte- relaxation pressure of the lower esophageal sphincter at the esophagogastric
grated relaxation pressure of the lower esopha- junction with the use of high-resolution manometry.
§ The Eckardt symptom score is a validated questionnaire and is calculated as
geal sphincter with the use of high-resolution the sum of symptom scores of four components of achalasia including dys-
manometry, did not differ significantly between phagia, regurgitation, chest pain, and weight loss. Each component can be
the treatment groups (difference, −0.75 mm Hg; graded from 0 to 3 points. The maximum Eckardt symptom score is 12 points,
with higher scores indicating more severe symptoms. A good outcome is
95% CI, −2.26 to 0.76). Similarly, improvement classified as an Eckardt symptom score of 3 or less. All patients reported
in Gastrointestinal Quality of Life Index scores an Eckardt symptom score higher than 3 at baseline.
between baseline and 24 months did not differ ¶ The Gastrointestinal Quality of Life Index is a validated questionnaire and
is calculated as the sum of subscores of 36 items regarding gastrointestinal
significantly between the groups (difference, function, emotion, physical function, social function, and medical treatment.
0.14 points; 95% CI, −4.01 to 4.28). Additional Each subscore ranges from 0 to 4 points. The maximum Gastrointestinal
details are provided in Figures S5 and S6. Quality of Life Index score is 144 points, with higher scores indicating better
function. Data were missing for two patients in the LHM group.
Procedure time was shorter in POEM group
than in the LHM group by 13.81 minutes (95%
CI, 6.26 to 21.36), but the length of hospital stay
did not differ significantly between the groups group (absolute between-group difference, 4.6
(difference, 0.26 days; 95% CI, −0.12 to 0.63). percentage points; 95% CI, −1.1 to 10.4). All seri-
Procedural data are provided in Table S12. ous and nonserious adverse events are listed in
No deaths occurred during the trial. A seri- Table S13.
ous adverse event occurred in 3 patients (2.7%) The development of gastroesophageal reflux
in the POEM group, and nine serious adverse disease was assessed clinically and by endoscopy
events occurred in 8 patients (7.3%) in the LHM and pH monitoring (Table 2 and Tables S14 and

A Clinical Success
POEM LHM
100
94.6
Percentage of Patients with

90 89.0 89.3 89.0
Clinical Success
84.8 83.5 83.0 81.7
80
70
60
0
0 3 6 12 24
Follow-up (mo)
No. at Risk
POEM (no. of imputed observations) 112 (4) 112 (6) 112 (5) 112 (4)
LHM (no. of imputed observations) 109 (3) 109 (7) 109 (7) 109 (5)
B Absolute between-Group Difference in the Percentage of Patients Who Had Clinical Success at 2 Years
Prespecified
noninferiority LHM more successful POEM more successful
margin
1.4
−12.5 −10.0 −5.0 0.0 5.0 10.0 12.5

Absolute Difference (percentage points)
Figure 3. Rates of Clinical Success over Time and Results of the Primary Analysis.
Panel A shows the estimated percentage of patients in each treatment group (modified intention-to-treat popula-
tion) who had clinical success, defined as an Eckardt symptom score of 3 or less (range, 0 to 12, with higher scores
indicating more severe symptoms of achalasia) without the use of additional treatments, over the follow-up period.
The primary end point was the percentage of patients who had clinical success at 2 years after the assigned inter-
vention, which was 83.0% in the POEM group and 81.7% in the LHM group. The number of imputed observations
at each time point is provided in parentheses. I bars indicate 95% confidence intervals. Panel B shows that the re-
sults of the primary analysis in the modified intention-to-treat population supported the hypothesis that POEM was
noninferior to LHM. The absolute between-group difference in the percentage of patients who had clinical success
at the 2-year follow-up was 1.4 percentage points (95% CI, −8.7 to 11.4) in favor of POEM (P = 0.007 for noninferiority).
S15 and Fig. S7). Among 196 and 165 patients in months in 4 of 87 patients (5%) in the POEM
the modified intention-to-treat population who group and 5 of 78 patients (6%) in the LHM
underwent endoscopy at 3 months and 2 years, group. Esophageal pH monitoring showed simi-
respectively, the incidence of reflux esophagitis lar proportions of patients with abnormal reflux
(all grades) was higher in the POEM group than at both time points (at 3 months, 41 of 93 [44%]
in the LHM group, both at 3 months (57% vs. in the LHM group and 27 of 82 [33%] in the
20%; odds ratio, 5.74; 95% CI, 2.99 to 11.00) and LHM group, and at 24 months, 21 of 70 [30%]
at 24 months (44% vs. 29%; odds ratio, 2.00; in the POEM group and 17 of 56 [30%] in the
95% CI, 1.03 to 3.85). As summarized descrip- LHM group). A post hoc analysis of the use of
tively, high-grade esophagitis (Los Angeles Clas- proton-pump inhibitors showed that a higher
sification grade C or D according to the Lyon percentage of patients in the POEM group than
Consensus20) was observed at 3 months in 6 of in the LHM group were receiving low-dose pro-
100 patients (6%) in the POEM group and 3 of ton-pump inhibitors across time points after
96 patients (3%) in the LHM group and at 24 baseline (at baseline, 28 of 112 [25.0%] vs. 33 of

Table 2. Clinical and Objective Evaluation of Gastroesophageal Reflux Disease (a Secondary End Point) over Time.*
Measure 3 Months 2 Years
POEM Group LHM Group POEM Group LHM Group

(N = 112) (N = 109) (N = 112) (N = 109)
Clinical scores
Mean DeMeester clinical score (95% CI) 0.9 (0.6–1.1) 0.5 (0.3–0.7) 1.2 (0.9–1.5) 1.0 (0.6–1.0)
Daily reflux symptoms — no./total no. (%) 5/108 (4.6) 2/105 (1.9) 7/107 (6.5) 2/103 (1.9)
Occasional reflux symptoms — no./total no. (%) 42/108 (38.9) 29/105 (27.6) 49/107 (45.8) 45/103 (43.7)
Daily proton-pump inhibitor use — no./total no. (%) 25/108 (23.1) 16/105 (15.2) 41/106 (38.7) 20/103 (19.4)
Occasional proton-pump inhibitor use — no./total no. (%) 8/108 (7.4) 13/105 (12.4) 15/106 (14.2) 8/103 (7.8)
LA Classification grade of reflux esophagitis — no./total no. (%)†
Overall, grades A to D 57/100 (57) 19/96 (20) 38/87 (44) 23/78 (29)
Grade A 32/100 (32) 13/96 (14) 18/87 (21) 13/78 (17)
Grade B 19/100 (19) 3/96 (3) 16/87 (18) 5/78 (6)
Grade C 5/100 (5) 2/96 (2) 4/87 (5) 2/78 (3)
Grade D 1/100 (1) 1/96 (1) 0/87 3/78 (4)
Esophageal acid exposure‡
Mean acid exposure time (95% CI) — % 7.1 (5.4–8.9) 6.7 (4.1–9.3) 5.7 (2.8–8.5) 5.4 (2.2–8.5)
Acid exposure time >4.5% — no./total no. (%) 41/93 (44) 27/82 (33) 21/70 (30) 17/56 (30)
* The 95% confidence intervals have not been adjusted for multiplicity and should not be used to draw inferences about effects.
† Los Angeles (LA) Classification grade was assessed on endoscopy: grade A indicates one or more mucosal breaks of 5 mm in length or less;
grade B, one or more mucosal breaks of longer than 5 mm; grade C, mucosal breaks that extend between two or more mucosal folds (but in-
volve <75% of the circumference of the esophagus); and grade D, mucosal breaks which involve at least 75% of the esophageal circumference.
‡ Esophageal acid exposure was assessed with the use of 24-hour pH monitoring; acid exposure time is the total percentage of time with a
pH lower than 4.
109 [30.3%]; at 3 months, 33 of 108 [30.6%] vs. trials of LHM or POEM.3,21 A possible explanation
29 of 105 [27.6%]; and at 24 months, 56 of 106 may be the inclusion of patients who had previ-
[52.8%] vs. 28 of 103 [27.2%]). ously received treatment (endoscopic pneumatic
dilation or botulinum toxin injection) in our
trial, whereas previous treatment for achalasia
Discussion
was an exclusion criterion in the other trials. In
Our randomized trial prospectively compared the our trial, one third of patients in each treatment
two different approaches to esophageal myotomy group at baseline had already been unsuccess-
— POEM and LHM — in patients with idio- fully treated. Although the descriptive differences
pathic achalasia. Our results provide evidence for in the proportions of patients who had clinical
the noninferiority of POEM to LHM plus Dor’s success at 2 years after LHM between those who
fundoplication in controlling symptoms of acha- had previously received treatment and those who
lasia at 2 years but showed that gastroesopha- had not are consistent with the findings in a
geal reflux was more common among patients previous trial,22 the influence of having received
who underwent POEM than among those who previous treatment on the clinical success of
underwent LHM. This finding of noninferiority POEM is less clear.23,24
was corroborated by manometric data on esoph- Serious adverse events were reported in 3% of
ageal function over time. patients in the POEM group and 7% of patients
The rates of clinical success at 2 years for in the LHM group; the trial was not powered to
both approaches to esophageal myotomy were assess differences in infrequent adverse events.
lower in our trial than in previous randomized The percentage of patients who had minor ad-

verse events was similar in the treatment groups, pain medications; in general, the findings at the
although definitions of minor adverse events 2-year follow-up in this trial suggest that there
vary in the literature.16 The need for additional was no between-group difference in improve-
surgery to treat adverse events after POEM has ments in patient-reported quality of life. Patients
been reported,25 including by our research and trial personnel were aware of the treatment-
group,16,26 but in the current trial, there were no group assignments because blinding was not
adverse events of POEM that required surgical possible. This was a potential source of bias
repair. Treatment conversion from POEM to given that the primary end point was based on
LHM was not needed in any patient in our trial, patients’ reports of symptoms; however, objec-
but we currently would not consider practicing tive assessment by manometry corroborated the
POEM without a thoracic surgeon as an on-site primary finding.
back-up. In conclusion, in this multicenter, random-
Gastroesophageal reflux has been a major ized trial, the less invasive POEM approach was
concern with POEM, especially because, unlike noninferior to LHM in controlling symptoms of
LHM, no antireflux procedure is added to this achalasia but resulted in more cases of gastro-
procedure. Meta-analyses on this topic have esophageal reflux. Our results could assist in
shown that the incidence of gastroesophageal shared decision making in selecting the appro-
reflux is much higher after POEM than after priate individual therapy for patients with pri-
LHM, ranging from a doubling27 to an increase mary achalasia.
by a factor of nine,9 but the meta-analyses were The content of this article is solely the responsibility of the
based mostly on retrospective case series. Our authors and does not necessarily represent the official views of
randomized trial showed that more cases of re- Harvard Catalyst, Harvard University and its affiliated academic
health care centers, or the National Institutes of Health.
flux esophagitis occurred among the patients in A data sharing statement provided by the authors is available
the POEM group than in the LHM group, with with the full text of this article at NEJM.org.
a larger between-group difference at 3 months Supported by the European Clinical Research Infrastructure
Network (ECRIN); Hamburgische Stiftung für Wissenschaften,
than at 2 years. In addition, a higher percentage Entwicklung und Kultur Helmut und Hannelore Greve (Ham-
of patients in the POEM group than in the LHM burg, Germany); by Dr. med. Carl-August Skröder Stiftung
group used proton-pump inhibitors during the (Hamburg, Germany); Dr. Gerhard Büchtemann Stiftung (Ham-
burg, Germany); Agnes-Graefe Stiftung (Hamburg, Germany);
follow-up period, even at 2 years. The results of Georg und Jürgen Rickertsen Stiftung (Hamburg, Germany);
pH monitoring did not differ significantly be- Reinhard Frank Stiftung (Hamburg, Germany); Johann Max
tween the treatment groups at the 2-year follow- Böttcher Stiftung (Hamburg, Germany); Richard und Annema-
rie Wolf Stiftung (Knittlingen, Germany); Olympus Europa
up, although not all patients agreed to undergo (Hamburg, Germany), which provided funding for the training
this test. The issue of gastroesophageal reflux of personnel at the European centers through POEM Grants
after achalasia treatments should be examined 206/7 and 2017/18; the German Society for Gastroenterology
and Metabolism and Olympus Europe Foundation (Endoscopy
further, especially regarding possible long-term Research Award 2012); a prize awarded at the United European
consequences, such as development of Barrett’s Gastroenterology Week 2010 conference; and the Olympus Euro-
esophagus and its complications.28,29 NOTES Research Fund Program. The trial was also supported by
Harvard Catalyst, the Harvard Clinical and Translational Science
Our trial has several limitations. Less than Center (National Center for Advancing Translational Sciences, Na-
50% of the eligible patients participated in the tional Institutes of Health Award UL 1TR002541), and by finan-
trial, mainly because of a lack of patient consent cial contributions from Harvard University and its affiliated aca-
demic health care centers.
to undergo randomization. The surgeons were Disclosure forms provided by the authors are available with
more experienced in performing LHM plus Dor’s the full text of this article at NEJM.org.
fundoplication than the endoscopists were in per- We thank Ms. Rie Koiwa (Frankfurt, Germany) for her assis-
tance with an earlier version of Figure 1 and Dr. Haru Inoue
forming POEM. We did not analyze treatment (Tokyo) for his initial assistance in training personnel at several
effects on postoperative pain or on the use of sites on how to use the peroral endoscopic myotomy technique.
Appendix
The authors’ affiliations are as follows: the Departments of Interdisciplinary Endoscopy (Y.B.W., T.N., T.R.), Epidemiology and Statis-
tics (J.F.K.), and General, Visceral, and Thoracic Surgery (O.M., J.I.), University Hospital Hamburg-Eppendorf, and the Department of
Surgery, Israelitic Hospital (A.E.), Hamburg, the Department of Surgery, University Hospital Würzburg, Würzburg (B.H.A.R., C.-T.G.),
and the Department of Gastroenterology, University Hospital Augsburg, Augsburg (H.M.) — all in Germany; the Department of Clini-
cal Sciences, Danderyd Hospital, Karolinska Institutet, and the Department of Surgery, Ersta Hospital, Stockholm (B.H.); the Depart-

ment of Hepatogastroenterology, Institute for Clinical and Experimental Medicine (J.M.), and the 3rd Department of Surgery, 1st Fac-
ulty of Medicine, Charles University in Prague and University Hospital Motol (A.P.) — all in Prague, Czech Republic; the Departments
of Gastroenterology (A.R.) and Digestive Surgery (U.F.), Istituto Clinico Humanitas Rozzano, and the Department of Digestive Surgery,
IRCCS Ospedale San Raffaele (R.R.) — both in Milan; Amsterdam University Medical Centers location AMC, Amsterdam Gastroenterol-
ogy and Metabolism (A.J.B., P.F.) and Surgery (M.P.S.), University of Amsterdam, Amsterdam; the Department of Gastroenterology and
Hepatology, University Hospitals Leuven, and Translational Research Center for Gastrointestinal Disorders, Katholieke Universiteit
Leuven, Leuven, Belgium (R.B., G.B.); the Department of Psychology, Harvard University, Cambridge, MA (M.C.V.); and the Division of
Gastroenterology, Montreal University Hospital and Research Center, Montreal (D.R.).
References
1. Boeckxstaens GE, Zaninotto G, Rich- Predictors of outcome in patients with naive patients with achalasia: a random-
ter JE. Achalasia. Lancet 2014;383:83-93. achalasia treated by pneumatic dilation. ized clinical trial. JAMA 2019;322:134-44.
2. Zaninotto G, Bennett C, Boeckxstaens Gastroenterology 1992;103:1732-8. 22. Smith CD, Stival A, Howell DL, Swaf-
G, et al. The 2018 ISDE achalasia guide- 12. Pandolfino JE, Kwiatek MA, Nealis T, ford V. Endoscopic therapy for achalasia
lines. Dis Esophagus 2018;31(9). Bulsiewicz W, Post J, Kahrilas PJ. Achala- before Heller myotomy results in worse
3. Boeckxstaens GE, Annese V, Bruley sia: a new clinically relevant classification outcomes than Heller myotomy alone.
des Varannes S, et al. Pneumatic dilation by high-resolution manometry. Gastroen- Ann Surg 2006;243:579-84.
versus laparoscopic Heller’s myotomy for terology 2008;135:1526-33. 23. Nabi Z, Ramchandani M, Chavan R,
idiopathic achalasia. N Engl J Med 2011; 13. Eypasch E, Williams JI, Wood-Dauphi et al. Peroral endoscopic myotomy in
364:1807-16. nee S, et al. Gastrointestinal Quality of Life treatment-naïve achalasia patients versus
4. Moonen A, Annese V, Belmans A, Index: development, validation and appli- prior treatment failure cases. Endoscopy
et al. Long-term results of the European cation of a new instrument. Br J Surg 1995; 2018;50:358-70.
achalasia trial: a multicentre randomised 82:216-22. 24. Liu ZQ, Li QL, Chen WF, et al. The
controlled trial comparing pneumatic dila- 14. Demeester TR, Johnson LF, Joseph GJ, effect of prior treatment on clinical out-
tion versus laparoscopic Heller myotomy. Toscano MS, Hall AW, Skinner DB. Pat- comes in patients with achalasia under-
Gut 2016;65:732-9. terns of gastroesophageal reflux in health going peroral endoscopic myotomy. En-
5. Kahrilas PJ, Katzka D, Richter JE. The and disease. Ann Surg 1976;184:459-70. doscopy 2019;51:307-16.
use of per-oral endoscopic myotomy in 15. Armstrong D, Bennett JR, Blum AL, 25. Haito-Chavez Y, Inoue H, Beard KW,
achalasia: expert review and best practice et al. The endoscopic assessment of esoph- et al. Comprehensive analysis of adverse
advice from the AGA Institute. Gastroen- agitis: a progress report on observer agree- events associated with peroral endoscopic
terology 2017;153:1205-11. ment. Gastroenterology 1996;111:85-92. myotomy in 1826 patients: an interna-
6. Pasricha PJ, Hawari R, Ahmed I, et al. 16. Werner YB, von Renteln D, Noder T, tional multicenter study. Am J Gastroen-
Submucosal endoscopic esophageal myot- et al. Early adverse events of per-oral en- terol 2017;112:1267-76.
omy: a novel experimental approach for doscopic myotomy. Gastrointest Endosc 26. Kumagai K, Tsai JA, Thorell A, Lun-
the treatment of achalasia. Endoscopy 2017;85(4):708.e2-718.e2. dell L, Håkanson B. Per-oral endoscopic
2007;39:761-4. 17. Crespin OM, Liu LWC, Parmar A, et al. myotomy for achalasia: are results com-
7. Inoue H, Minami H, Kobayashi Y, et al. Safety and efficacy of POEM for treatment parable to laparoscopic Heller myotomy?
Peroral endoscopic myotomy (POEM) for of achalasia: a systematic review of the Scand J Gastroenterol 2015;50:505-12.
esophageal achalasia. Endoscopy 2010;42: literature. Surg Endosc 2017;31:2187-201. 27. Repici A, Fuccio L, Maselli R, et al.
265-71. 18. Barbieri LA, Hassan C, Rosati R, Ro- GERD after per-oral endoscopic myotomy
8. Andolfi C, Fisichella PM. Meta-analy- mario UF, Correale L, Repici A. System- as compared with Heller’s myotomy with
sis of clinical outcome after treatment for atic review and meta-analysis: efficacy fundoplication: a systematic review with
achalasia based on manometric subtypes. and safety of POEM for achalasia. United meta-analysis. Gastrointest Endosc 2018;
Br J Surg 2019;106:332-41. European Gastroenterol J 2015;3:325-34. 87(4):934.e18-943.e18.
9. Schlottmann F, Luckett DJ, Fine J, 19. Campos GM, Vittinghoff E, Rabl C, 28. Zendehdel K, Nyrén O, Edberg A, Ye W.
Shaheen NJ, Patti MG. Laparoscopic Hell- et al. Endoscopic and surgical treatments Risk of esophageal adenocarcinoma in
er myotomy versus peroral endoscopic for achalasia: a systematic review and achalasia patients, a retrospective cohort
myotomy (POEM) for achalasia: a system- meta-analysis. Ann Surg 2009;249:45-57. study in Sweden. Am J Gastroenterol 2011;
atic review and meta-analysis. Ann Surg 20. Gyawali CP, Kahrilas PJ, Savarino E, 106:57-61.
2018;267:451-60. et al. Modern diagnosis of GERD: the Lyon 29. Leeuwenburgh I, Scholten P, Alderlies
10. Akintoye E, Kumar N, Obaitan I, Consensus. Gut 2018;67:1351-62. ten J, et al. Long-term esophageal cancer
Alayo QA, Thompson CC. Peroral endo- 21. Ponds FA, Fockens P, Lei A, et al. Effect risk in patients with primary achalasia:
scopic myotomy: a meta-analysis. Endos- of peroral endoscopic myotomy vs pneu- a prospective study. Am J Gastroenterol
copy 2016;48:1059-68. matic dilation on symptom severity and 2010;105:2144-9.
11. Eckardt VF, Aignherr C, Bernhard G. treatment outcomes among treatment- Copyright © 2019 Massachusetts Medical Society.


10 1056@NEJMoa1905380

Uploaded by

Copyright:

Available Formats

10 1056@NEJMoa1905380

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

10 1056@NEJMoa1905380

Uploaded by

Copyright:

Available Formats

The n e w e ng l a n d j o u r na l of m e dic i n e

Endoscopic or Surgical Myotomy in Patients

n engl j med 381;23 nejm.org December 5, 2019 2219

2220 n engl j med 381;23 nejm.org December 5, 2019

The New England Journal of Medicine

Figure 1. Alternative Methods of Treatment for Achalasia.

n engl j med 381;23 nejm.org December 5, 2019 2221

2222 n engl j med 381;23 nejm.org December 5, 2019

The New England Journal of Medicine

n engl j med 381;23 nejm.org December 5, 2019 2223

241 Patients underwent randomization

112 Underwent POEM 109 Underwent LHM

Figure 2. Enrollment, Randomization, and Follow-up.

2224 n engl j med 381;23 nejm.org December 5, 2019

The New England Journal of Medicine

n engl j med 381;23 nejm.org December 5, 2019 2225

Percentage of Patients with

−12.5 −10.0 −5.0 0.0 5.0 10.0 12.5

2226 n engl j med 381;23 nejm.org December 5, 2019

The New England Journal of Medicine

Measure 3 Months 2 Years

POEM Group LHM Group POEM Group LHM Group

n engl j med 381;23 nejm.org December 5, 2019 2227

2228 n engl j med 381;23 nejm.org December 5, 2019

The New England Journal of Medicine

n engl j med 381;23 nejm.org December 5, 2019 2229

You might also like