Outcomes of 100 Patients Greater than 4 Years after POEM for Achalasia – Original Study

Abstract

Objective:

We aim to describe the long-term follow up data from our institution’s POEM experience.

Summary Background Data:

Per-oral endoscopic myotomy (POEM) is a well-established endoscopic therapy for achalasia with excellent short-term efficacy, but long-term outcomes data are limited.

Methods:

Patients greater than 4 years removed from POEM for treatment of achalasia were studied. Clinical success was defined as an Eckardt Symptom (ES) score ≤ 3 and freedom from reintervention for achalasia. Patients underwent esophagogastroduodenoscopy (EGD), high-resolution manometry, impedance planimetry, and timed barium esophagram (TBE) pre-operatively and at least 4 years post-operatively. Objective GERD was defined LA Grade B or worse esophagitis on EGD.

Results

One hundred and nineteen consecutive patients were included. Five patients died or had catastrophic events unrelated to achalasia or POEM. One hundred of the remaining patients (88%, 100/114) had long-term data available. Clinical follow up for all patients was greater than 4 years post-operatively and the mean was 55 months. Mean current ES was significantly improved from preop (n=100, 1 ± 1 vs 7 ± 2, p <0.001). Overall clinical success was 88% and 92%. 5 patients had a current ES > 3 and 4 patients required procedural reintervention on the lower esophageal sphincter. Reinterventions were successful in 75% of patients (3/4), with current ES ≤ 3. The rate of objective GERD was 33% (15/45). Esophageal physiology was improved with a decrease in median IRP (11 ± 4 vs 33 ± 15 mmHg, p <0.001), a decrease in median TBE column height (3 ± 3 vs 13 ± 8 cm, p <0.001), and an increase in median distensibility index (5.1 ± 2 vs 1.1 ± 1 mm²/mmHg, p<0.001).

Conclusions

POEM provides durable symptom relief and improvement in physiologic EGJ relaxation parameters over 4.5 years postoperatively. Reinterventions are rare and effective.

MINI-ABSTRACT

In this study, we examined the long term clinical and physiologic outcomes of patients with achalasia at least 4 years removed from POEM.

INTRODUCTION

Achalasia is a rare neurodegenerative disease of the esophagus with an incidence of 1.5 / 100,000 people worldwide ^1–3. Patients present with progressive dysphagia, regurgitation, chest pain, and/or weight loss. Diagnostic studies including timed barium esophagography (TBE), esophagogastroduodenoscopy (EGD), impedance planimetry, and high-resolution manometry (HRM) reveal a non-relaxing lower esophageal sphincter and absent or greatly altered esophageal peristalsis. There is no cure, but effective symptom palliation is possible via disruption of the lower esophageal sphincter with pneumatic balloon dilation (PD) or surgical myotomy.

In 2010, Inoue and colleagues described the first series of patients treated with endoscopic surgical division of the lower esophageal sphincter.⁴ Their technique, coined per-oral endoscopic myotomy (POEM), quickly gained traction in several large centers of excellence. Multiple studies with moderate-term follow-up (2–4 years) have demonstrated significant symptom relief in 80–95% of patients post-POEM^5–7. Unfortunately, there are limited reports with long-term (>4 years) clinical and physiologic follow-up after POEM^8–10, thereby propagating this debate. Despite this limitation, multiple international guidelines support and encourage the use of POEM as an effective alternative to PD and LHM^11–13. Herein, we present greater than four year outcomes of our institutional POEM experience for treatment of achalasia and associated esophageal motility disorders¹⁴.

METHODS

We performed a retrospective review of our prospectively maintained achalasia database. Patients who were greater than 4 years removed from POEM were eligible. Patients from our 15-case learning curve were not included. All study activities were carried out with the approval of the Northwestern University Institutional Review Board.

Pre-Operative Evaluation

Symptoms were evaluated with the Eckhardt score (ES) and GerdQ scoring system^{15, 16}. Standard pre-operative physiologic testing (Figure 1) included EGD, timed barium esophagram (TBE), high-resolution manometry, and impedance planimetry via the functional luminal imaging probe (FLIP). TBE barium column heights were recorded at 5 minutes post-swallow. Manometric patterns were characterized with the Chicago Classification of esophageal motility disorders, version 3.0¹⁷. FLIP measurements were collected with 8-cm or 16-cm catheters using the EndoFLIP 1.0 or 2.0 system (Medtronic, Minneapolis, MN). Distensibility measures of intrabag pressure (P), cross-sectional area (CSA), and the esophagogastric junction distensibility index (DI = CSA/P) were collected at 40 mL and 60 mL for the 8-cm and 16-cm catheters, respectively.

Figure 1:

Standard pre-operative evaluation for patients with suspected achalasia. The results from a patient with type I achalasia are shown here, including timed barium esophragram (top left), esophagogastroduodenoscopy (top right), high-resolution manometry (bottom left), and impedance planimetry data from the functional luminal imaging probe (FLIP, bottom right)

Operative Technique

Our operative technique has been described in detail^{18, 19}. Briefly, after mucosotomy (in an anterior, 12 to 1 o’clock position) and submucosal tunnel creation, a selective myotomy of the circular fibers is completed along the esophagus, EGJ, and stomach. The extent of our standard myotomy is 6 cm proximal to the EGJ and 2–3 cm distal to the EGJ. Thus, the total length of the myotomy is 8–9 cm. The proximal extent of the myotomy is extended for patients with type III achalasia or spastic disorders of the esophagus, guided by the pre-operative HRM in order to ablate the segment of esophageal body spasm). The mucosotomy is closed with clips. Patients are discharged when tolerating a liquid diet. Esophagrams were not obtained routinely in the immediate postoperative period. Patients were discharged on routine daily PPI therapy which they were recommended to continue until objective evaluation for gastroesophageal reflux (GER) at 6–12 months post-operatively.

Post-Operative Evaluation

At 6–12 months post-postoperatively, patients were evaluated with repeat symptoms scores (ES and GerdQ) and physiologic testing including EGD, TBE, HRM, and FLIP. Patients were instructed to discontinue PPI therapy 2 weeks prior to endoscopic evaluation. Symptom scores were collected on the day of endoscopy. ES > 3 and GerdQ > 8 were considered abnormal¹⁶. Objective evidence of GER was defined as Los Angeles Grade B, C. or D esophagitis on EGD. FLIP assessments were completed during upper endoscopy according to our previously described protocols^{20, 21}. Following evaluation, proton pump inhibitor therapy was discontinued in the absence of symptomatic or objective GER. Patients were re-evaluated at least 4 years post-operatively, during which a repeated physiologic workup was recommended. pH studies were obtained for cases in which the diagnosis of GER was unclear by clinical data and EGD. A positive study was defined as > 4.0% total acid exposure time. Patients who were unavailable for in-person evaluation were contacted via telephone to obtain symptom scores. Clinical success was defined as an ES ≤ 3 and freedom from reintervention for achalasia throughout the course of the study.

Statistical Analysis

Data were analyzed with SPSS Statistics v25/26 (IBM, Armonk, NY). Continuous variables collected at multiple time points were analyzed with a paired t test. Comparisons between two groups and a continuous variable were completed with an independent t test. Dichotomous variables were compared with a Fischer Exact Test. Correlative analysis was completed with a Spearman’s rank-order correlation. A two-tailed p value < 0.05 was considered significant. All data are presented as mean ± standard deviation unless otherwise noted.

RESULTS

Patient Demographics

One hundred and nineteen consecutive patients underwent POEM from January 2012 to April 2015. Five patients were deceased or severely disabled from ailments unrelated to POEM and were excluded. The demographics of the remaining 114 patients are seen in Table 1. The majority of patients (67/114) were male. The mean body mass index was 26.7 kg/m² and the median American Society of Anesthesiologists (ASA) physical classification system score was 2. Twenty-three percent of patients (26/114, 23%) had pneumatic dilation or botulinum toxin injection prior to POEM. Patients with type II achalasia were most common, representing 51% of the cases. 20 patients (20/114, 18%) underwent POEM with an extended proximal myotomy (ePOEM) for type III or achalasia variants. 71 patients had intraoperative FLIP measurements available. POEM led to significant changes in cross-sectional area (157.7 ± 56 vs 40.0 ± 12 mm²), intrabag pressure (36.7 ± 13 vs 23.1 ± 7 mmHg), and DI (8.2 ± 4 vs 1.3 ± 1 mm²/mmHg, all p<0.001).

Table 1:

Pre-Operative Patient Demographics

Patients
Total	114
Male	67 (59%)
Female	47 (41%)
Pre-Operative Characteristics (Mean)
Age (Years)	53
Body Mass Index (kg/m2)	27
Duration of Symptoms (Years)	4
Median ASA Score	2
Prior Treatment+	26 (23%)
Achalasia Subtype
Type I	27 (24%)
Type II	58 (51%)
Type III	16 (14%)
EGJOO	8 (7%)
Jackhammer	4 (3%)
Distal Esophageal Spasm	1 (1%)
Operation
POEM	94 (82%)
Extended POEM	20 (18%)

ASA – American Society of Anesthesiologists physical classification system

EGJOO – Esophagogastric Junction Outflow Obstruction

⁺Prior pneumatic dilation or botulinum toxin injection

Long-Term Clinical Outcomes

One hundred patients (100/114, 88%) had current clinical outcomes data from a mean follow-up of 55 months post-POEM (Table 2). Twelve patients reported ES > 3 at some point during the follow up period. Therefore, clinical success was observed in 88% of patients (88/100). Mean ES (1 ± 1 vs 7 ± 2, p <0.001) was significantly improved from pre-operative values. Seven patients had improvement in their ES following medical or surgical therapy. Thus, at the time of their long-term follow-up evaluation, 95% of patients (95/100) has an ES ≤ 3. Clinical success as a function of achalasia subtype is depicted in Figure 2. Patients with types I and II achalasia had a significantly higher long-term clinical success rate than those with type III achalasia, EGJ outflow obstruction, and other spastic disorders of the esophagus (92% vs 74%, p < 0.05). Mean GerdQ (n=100, 6 ± 2 vs 11 ± 3, p <0.001) improved significantly. 53% of patients (53/100) were taking a PPI at last follow-up and 23% (23/100) had an abnormally elevated GerdQ.

Table 2:

Long-Term Post-Operative Symptomatic Outcomes

Patients	100/114 (88%)
Clinical Success (Eckhart ≤ 3, no re-intervention)	88/100 (88%)
Type I	23/24 (96%)
Type II	48/53 (91%)
Type III	10/13 (77%)
EGJOO/Other+	7/10 (70%)
Re-Intervention	4/100 (4%)
Pneumatic Dilation	3
Laparoscopic Heller Myotomy	1
Re-Intervention Success	3/4 (75%)
Daily PPI Use	53% (53/100)
Current Eckhart Score ≤ 3	95/100 (95%)

EGJOO – Esophagogastric Junction Outflow Obstruction

⁺Jackhammer, Distal Esophageal Spasm

Figure 2:

Cumulative clinical success (Eckart Score ≤ 3) after POEM

Long-Term Physiologic Outcomes

Table 3 includes follow-up physiologic data stratified by achalasia subtype. The mean time to physiologic follow-up was 55 months. EGD was completed in 45% of patients (45/100). Twenty-three patients who underwent EGD (23/45, 51%)) were taking a PPI prior to cessation for endoscopic surveillance. Objective GER was noted in 33% of patients (15/45). Ten patients had LA B esophagitis (10/45, 22%) and 5 patients had LA C esophagitis (5/45, 11%). There were no cases of Barrett’s esophagus. Nine patients with objective GER (9/15, 60%) were not taking a PPI. The majority of patients who were not on PPI therapy had no evidence of GER (13/22, 60%). Objective GER was noted in 46% (5/11) of patients with type I achalasia and 15% (4/26) of patients with type II achalasia. There were no patients with type III achalasia, EGJOO, or other spastic disorders who demonstrated esophagitis (0/8). Five patients (5/100, 5%) completed pH studies, of which 2 were positive (2/5, 40%).

Table 3:

Long-Term Post-Operative Physiologic Outcomes at Mean 55 Months Follow-Up

	IRP (mmHg)			DI (mm2/mmHg)			TBE Column Height at 5 minutes (cm)			Esophagitis on EGD
	n	Pre	Post	n	Pre	Post	n	Pre	Post	n	Yes	No	%
All Subtypes	41	33±15	11±4*	26	1±1	5±2*	18	13±8	3±3**	45	15	30	33%
Type I	9	33±17	10±3**	6	1±1	4±2*	3	3±3	2±2	11	5	6	46%
Type II	24	34±14	12±5**	17	1±1	5±2**	14	15±7	3±3**	27	8	19	30%
Type III	2	51±5	10 ±2*	-	-	-	-	-	-	3	2	1	67%
EGJOO/Other	6	23±15	10±2	3	1±1	5±1**	1	14	4	4	0	4	0%

IRP – Integrated Relaxation Pressure DI – Distensibility Index TBE – Timed Barium Esophagram Pre – Pre-operative EGJOO – Esophagogastric Junction Outflow Obstruction EGD – Esophagogastroduodenoscopy Post – Post-Operative

^*p <0.05

^**p<0.01

43 patients (43/100, 43%) had manometric data available. The mean basal expiratory EGJ pressure and 4-second IRP were significantly reduced from their preoperative values (n=41, 8.8 ± 3 vs 28.2 ± 15 mmHg, 11.1 ± 4 vs 33.3 ± 15, both p < 0.001). FLIP data were obtained in 45% of patients (45/100). Two patients had unusable data due to catheter malfunction. Mean follow-up CSA (129.2 ± 48 vs 35.7 ± 21 mm²), intrabag pressure (26.2 ± 8 vs 37.0 ± 12 mmHg), and DI (5.1 ± 2 vs 1.1 ± 1 mm²/mmHg) remained significantly improved from their pre-operative values (n=26, all p<0.001).

Timed barium esophagogram studies were obtained in 30% of patients (30/100). 70% of patients (21/30) had a TBE column height < 5 cm. Mean TBE column height was significantly reduced from pre-operative values (n=18, 3.2 ± 3 vs 13.2 ± 8 cm, p<0.05). Patients with an abnormal follow-up TBE column height (≥ 5 cm) had a significantly lower pre-operative BMI than those with a normal column height (n=30, 22.4 ± 3 vs 28.6 ± 7 kg/m², p <0.01).

Clinical Failure Case Details

Detailed characteristics of the 12 patients who experienced symptomatic failures (ES > 3) are listed in Table 4. Patients with type III achalasia, EGJOO, or other spastic disorders of the esophagus constitute 50% (6/12) of the failure patient pool. The mean time to failure was 12 months. Median Eckhardt score at the time of failure was 5, and half of patients (6/12) experienced a failure with an ES = 4. Five patients experienced a persistent failure in both the early and late-term follow up periods; two had subsequent improvement in symptoms with medical therapy and have current ES <3.

Table 4:

Clinical Characteristics and Management of Post-POEM Patients with Eckardt Score >3

Patient	Age (Years)	Achalasia Subtype	Operation	Time to Failure (months)	ES at Failure	Management	Success
1	36	Type 2	POEM	6, 47	4, 7	Medical Therapy	Yes
2	72	Type 2	POEM	7, 45	5, 5	Medical Therapy	No
3	68	Type 3	ePOEM	16	4	1x TTS Dilation	Yes
4	62	Jackhammer	POEM	14, 53	5, 5	1x TTS Dilation	Yes
5	21	Type 2	POEM	4, 49	4, 5	2x PD	No
6	41	EGJOO	POEM	7	6	LHM	Yes
7	66	Type 3	ePOEM	9	5	2x PD	Yes
8	77	Type 3	POEM	5	4	1x PD	Yes
9	26	Type 2	POEM	39	4	Medical Therapy	Yes
10	43	Type 2	POEM	32, 56	6, 6	Medical Therapy	No
11	32	EGJOO	POEM	47	8	Forthcoming PD	-
12	59	Type 1	POEM	55	4	Medial Therapy	-

LHM – Laparoscopic Heller Myotomy PD – Pneumatic Dilation TTS – Through the Scope ES - Eckardt Score ePOEM – Extended POEM

Four patients (4/100, 4%) underwent a re-intervention on the lower esophageal sphincter following POEM. Mean time to re-intervention was 6 months post-POEM. Three patients underwent pneumatic dilation and one patient had a laparoscopic Heller myotomy. The Eckardt score normalized in 75% of patients (3/4) after their re-intervention. Two additional patients with ES > 3 required through the scope dilation of a peptic stricture, and both experienced symptom relief following the procedure and PPI therapy. Medical therapy alone was pursued in four patients with symptom recurrence and was successful in 50% (2/4) of cases.

DISCUSSION

This study examined the clinical and physiologic outcomes of patients who were at least 4 years removed from POEM for treatment of achalasia. There are several key findings. Firstly, POEM resulted in durable symptom relief with an 88% clinical success rate at a mean follow-up of 55 months. Several patients recovered from clinical failure, either spontaneously or after reintervention, and 95% of patients had an ES ≤ 3 at the time of their long-term evaluation. Secondly, physiologic assessment revealed excellent sustained palliation of the LES, as evidenced by a persistent decrease in median IRP, FLIP distensibility index, and TBE column height. The objective post-operative GER rate was 33%, and there were no cases of Barrett’s esophagus. Finally, re-interventions targeting the lower esophageal sphincter were rare, and 75% (3/4) were effective in relieving symptoms.

The clinical success rate observed in this study reinforces the body of literature on POEM’s safety and efficacy. Contemporary institutional POEM experiences are listed in Table 5^{5–7, 9, 10, 14}.

Table 5:

POEM Series with Moderate- and Long-Term Follow-Up

Report (Year)	Patients	Follow-Up (Months)	Clinical Success+	EGJ Relaxation Pressure (mmHg) (Pre vs Post)	Distensibility Index (mm2/mmHg) (Pre vs Post)	Objective GERD++
Long-Term (>4 Years)
Campagna (2020)◊	114	55	88% (88/100)	33 vs 11	5 vs 1	33% (15/45)
Teitelbaum (2018)	36	65	83% (19/23)	23 vs 9	-	13% (2/16)
Li (2018)	564	49	87% (366/420)	28 vs 12	-	17% (58/341)
Moderate-Term (2–4 Years)
He (2019)	115	36	89% (69/78)	27 vs 13	-	-
Inoue (2015)	500	36	89% (54/61)	25 vs 12	-	24% (45/191)
Werner (2015)	80	29	79% (62/79)	32 vs 10	-	-

EGJ – Esophagogastric Junction GERD – Gastroesophageal Reflux Disease

^◊Current Study

⁺Eckardt Score ≤ 3 for entirety of study period

⁺⁺Endoscopic esophagitis or positive pH study

Li and Teitelbaum report similar rates of overall clinical success (87% and 83%), bolstering our finding that POEM provides durable palliation of symptoms^{9, 10} Our series offers additional clinical insights that are novel to the long-term POEM literature. As depicted in Figure 2, patients with type I/II achalasia experienced clinical failures at a low, steady rate across the study period. Conversely, patients with type III achalasia, EGJ outflow obstruction, or other spastic disorders experienced a higher failure rate and 83% of the failures (5/6) occurred within 18 months of POEM. Fortunately, 4/5 patients in this group of early type III/EGJOO failures experienced expeditious symptomatic recovery. Three patients underwent achalasia re-interventions (2 PD, 1 LHM), and all three had normal ES on subsequent follow-up. Two patients underwent through-the-scope dilation for peptic strictures. Both patients had GerdQ scores > 8 and esophagitis on endoscopy. One patient recovered immediately, while another required a second dilation late in the study prior to recovery. The literature offers conflicting data on the efficacy of POEM for patients with type III achalasia^{22, 23} Regardless of the method for LES disruption, these patients have less favorable outcomes than those with type I/II achalasia. However, POEM may offer better clinical success than PD or LHM for this subtype. This study suggests that patients with spastic achalasia may have increased rates of clinical failure in the early post-operative period, but many can experience long-lasting relief following early re-intervention.

Our rate of objectively measured post-POEM GER was 33%, with only 5 cases of LA C esophagitis. Two patients with LA C esophagitis were re-evaluated endoscopically after initiation of PPI therapy, and both demonstrated resolution of their esophagitis. The incidence of GER after POEM has varied widely in previously reported series from 13 to 55%²⁴. Regardless of the reported rate, most patients with post-POEM reflux are readily treated with PPI therapy²⁵. Myriad factors might influence the post-operative rate of GER, including operative technique. A recent meta-analysis by Mota et al examined 18 studies and found that an anterior circular-selective myotomy may lead to a lower incidence of post-POEM GER²⁶. Given the wide practice variation seen among early adopters of POEM²⁷, a more definitive answer may be elucidated as more centers of excellence report long-term data. Another factor that likely contributes to the reported variance in post-POEM GER is the poor correlation between symptoms and objective measures of reflux. Patients with post-POEM GER are often asymptomatic, with some series reporting “silent reflux” in up to 60% of patients with endoscopic esophagitis or an abnormal pH study²⁸. Conversely, Familiari et al have reported that up to 1/3 of patients with abnormal pH studies neither have esophagitis nor reflux symptoms²⁹. These conflicting findings are further evidence that a widely adopted standardized definition for post-POEM GER is much needed.

Other notable physiologic outcomes were seen in the manometry and TBE findings. Patients experienced continued palliation of the LES as evidenced by a persistently decreased mean TBE column height, IRP, and basal expiratory EGJ pressure. Column height on timed barium esophagography correlates with DI after therapeutic LES disruption³⁰. Moreover, esophageal stasis ≥ 5 cm is a sensitive predictor of symptomatic failure and need for re-intervention. In this study, patients with TBE columns ≥ 5 cm on long-term evaluation had a significantly lower pre-operative mean BMI than those with normal column heights. A lower pre-operative BMI could represent more significant anorexia and weight loss prior to POEM. This is turn may reflect more severe disease pre-operatively. However, it is important to note that esophageal stasis is only one portion of a complex physiologic mechanism that drives symptom recurrence.

There is ongoing controversy on the efficacy of POEM versus LHM for achalasia. A recent multicenter randomized trial by Werner et al made decisive inroads on this issue.³¹ They noted similar symptom control among POEM and LHM patients at 2 years. POEM had a higher rate of esophagitis at both 3 months and two years post-procedure; however, there were no differences in esophageal acid exposure measured by 24-hour pH monitoring. Studies by Ortiz and Csendes indicated that patients who were > 5 years removed from LHM may experience progressively worsening dysphagia or GER.^{32, 33} Conversely, in the current study cohort, a higher percentage of patients had a normal ES at 55 months post-POEM than at 28 months (95% vs 92%). This is attributable to successful recovery of moderate-term clinical failures. There are also multiple recent meta-analyses comparing LHM to POEM^34–36. These reports support POEM as a safe and effective method of symptomatic relief for achalasia patients. Awaiz et al noted a higher early clinical failure rate and comparable prevalence of GER following LHM³⁵. Docimo et al, among others, have noted a decreased hospital length of stay and reduced post-operative narcotic usage with POEM.³⁶ Unfortunately, all of the aforementioned comparative studies lack significant long-term follow-up data for POEM patients.

This study has several important limitations. Firstly, it was a retrospective analysis that lacks comparative treatment groups in pneumatic dilation and/or LHM. Secondly, we were limited in our ability to recruit patients for full physiologic testing on late-term assessment, which is a commonly reported challenge in this patient population^{8, 10, 33}. As a result, there may be some degree of bias related to which patients underwent follow-up testing, that may not be representative of the study population as a whole. Finally, reflective of the typical subtype distribution, our study population had relatively few patients with type III achalasia, EGJOO, and other spastic disorders, limiting interpretability of outcomes data from these patient subsets.

In conclusion, we observed excellent treatment efficacy and durability at long-term follow-up after POEM for achalasia. Overall clinical success was 88% across the entire study, and 95% of patients reported an ES ≤ 3 on the most recent follow-up. Patients with type I/II achalasia had a significantly higher long-term clinical success rate than those with type III achalasia and other spastic disorders of the esophagus (92% vs 74%). Physiologic studies indicated excellent long-term compliance of the lower esophageal sphincter, as evidenced by both manometry and FLIP measurements. One third of patients had evidence of GER on EGD, but there were few instances of severe esophagitis and no Barrett’s esophagus. These data support the ongoing use of POEM as standard therapy for patients with achalasia.