Targeting the pylorus in gastroparesis: From physiology to endoscopic pyloromyotomy

Abstract

Background

The pylorus plays a key role in the control of gastric content outflow. Impairment of pyloric physiology has been observed in gastroparesis, particularly when associated with diabetes mellitus or opioid intake or after antireflux surgery. New tools have been developed to identify pyloric dysfunction in routine care, including functional luminal impedance planimetry (FLIP). As such, a new therapeutic strategy targeting the pylorus, namely endoscopic pyloromyotomy (G‐POEM), has received increasing attention and emerged as a promising treatment for gastroparesis.

Purpose

The present review details the involvement of the pyloric pathophysiology in gastroparesis, as well as clinical results of G‐POEM according to the current literature.

Key points.

• Pyloric contractions and relaxations play a key role in the regulation of gastric emptying.

• Pyloric function has been shown to be altered in patients with gastroparesis, explored by manometric studies, or by new EndoFLIP tool.

• Per‐oral endoscopic pyloromyotomy (G‐POEM) has shown mid‐term clinical efficacy in 60 to 80% of patients with refractory gastroparesis.

• G‐POEM also leads to an acceleration of gastric emptying with an acceptable safety profile. Further studies to confirm the place of this treatment are needed.

1. INTRODUCTION

Gastroparesis is defined as delayed gastric emptying (GE) of solid food in the absence of mechanical obstruction. ¹ Cardinal symptoms include nausea, vomiting, gastric fullness, early satiety, bloating, and abdominal pain/discomfort. ² These symptoms can overlap with those of functional dyspepsia, but gastroparesis has to be confirmed by an objective measurement of GE, either by scintigraphy, or by ¹³C‐octanoic acid breath test. The prevalence of diagnosed gastroparesis has been evaluated in the United States at 0.24% of the population. ³ Recent epidemiological data from the United Kingdom confirm a similar prevalence of 0.13% in Europe based on the diagnostic code, although the disease appears to be underdiagnosed. ⁴ Gastroparesis leads to an economic burden, evaluated between US$4000 and US$9000 per patient per year in a recent study. ⁵ Moreover, delayed GE has been associated with increased mortality, independently of age, sex, or the etiology of the disease. ⁶

The first‐line treatment relies on dietary modification and medical treatment, with prokinetic drugs such as metoclopramide. ⁷ , ⁸ However, the efficacy of medical treatment is limited, and up to 40% of patients are considered to be refractory after appropriate treatment. ⁹ Severe refractory gastroparesis can lead to dehydration, weight loss, and impaired quality of life and could become a challenging condition. ¹⁰ Thus, invasive techniques have been developed to improve the symptoms of such patients.

Considering the role of pyloric dysfunction in gastroparesis, ¹¹ gastric peroral endoscopic pyloromyotomy (G‐POEM) has emerged as an alternative therapeutic technique in patients with refractory gastroparesis. The first human case was described in 2013 by Khashab et al. ¹² The technique is based on the principles of peroral endoscopic myotomy of the lower esophageal sphincter, used to treat achalasia. Since then, cohort studies have assessed the clinical efficacy and the safety of G‐POEM in refractory gastroparesis.

This review aims to detail the role of the pylorus in the pathophysiology of gastroparesis, and then to report the efficacy, modalities, and the place of G‐POEM, based on the results of published studies.

2. PYLORIC PHYSIOLOGY AND DYSFUNCTION IN GASTROPARESIS

The pylorus plays a key role in the regulation of GE. Postprandially, there is a first brief period of emptying of liquid and small size particles. Then, constriction of the pylorus restricts GE as the gastric peristaltic contraction reaches the distal antrum. Food is then retropulsed to the antrum, where antral contractions will triturate it into smaller particles (2–4 mm). ¹³ Then, pyloric relaxation in coordination with antro‐pyloro‐duodenal motor activity leads to GE of the food. ¹⁴

The pyloric sphincter is a narrow zone of thickened muscularis and increased luminal pressure. Studies in vitro reveal two functional areas in the pyloric musculature. ¹⁵ The first one is the circular muscle close to the myenteric plexus, which is dominated by the propagation of the slow wave from the antrum. ¹⁶ In this area, pyloric contraction is linked to the propagation of the gastric slow wave. The second area is the deeper circular muscle, which is regulated by motor neurons. Slow waves do not propagate into this area, and the two regions are under independent control. Despite clear anatomic connections between the muscular tissue of the stomach and the duodenum, electrical and mechanical behavior is independent on each side of the pylorus. ¹⁷ The pylorus has been characterized by a strong reduction in the number of interstitial cells of Cajal (ICCs), with a decreased kit positive cells in pyloric tissue, compared with antral or duodenal specimens. ¹⁸ This reduction in ICCs was related to a slow‐wave free gap. ¹⁸ Thus, in the absence of gastric emptying, there is no coordination of motor activities. But after the opening of the pylorus, antral activity propagates aborally and coordinates with duodenal activity. ¹⁸

Pyloric contractions and relaxations are regulated by both intrinsic and extrinsic innervation. ¹⁹ The intrinsic innervation is provided by myenteric plexus of the stomach which extends through the pylorus. Enteric motor neurons regulate pyloric function, through release of neurotransmitters. Transduction of motor signal is then performed through ICC, which conduct the signal to smooth muscle cells. Excitatory response is mediated by enkephalin and acetylcholine, whereas inhibitory response is mediated by nitric oxide (NO) and vasoactive intestinal peptide (VIP) ¹⁷ , ²⁰ (Figure 1B). The extrinsic nerves are mainly constituted by branches of the vagus nerve. The vagus exerts both inhibitory and excitatory effect on gastric motility in physiological condition, while sympathetic nerves have no major role. ²¹ Sensory nerve fibers are the most frequent and carry afferent signals to the nuclear tractus solitarius (NTS). Stretch receptors have a high density in the pylorus and respond to low levels of stretch. ²² Efferent motor fibers of the vagus start from dorsal motor nucleus and regulate pyloric response by interacting with the enteric nervous system. ²³ Denervation of the pylorus led to a decrease in the compliance and loss of relaxation of pylorus, leading to a decrease in transpyloric outflow. ²⁴ Stimulation of vagal fibers showed consistent results with a decrease in pyloric resistance. ²⁵ Thus, the inhibitory effect is thought to be the most active effect (Figure 1A). The regulation of contractions is then mediated through enterogastric vagovagal reflex, such as the pyloric contractions after acidification of duodenum. ²⁶

FIGURE 1.

Representation of extrinsic and intrinsic innervation of the pylorus. (A) Extrinsic innervation is mainly conducted by the vagus nerve, afferent sensitive nerves end in the nucleus tractus solitarius (NTS), and communicate with the dorsal motor nucleus (DMV). Efferent motor nerves of the vagus start from the DMV and play an excitatory and inhibitory role on gastric motility, but mostly an inhibitory role on pyloric relaxation. (B) Intrinsic innervation is transmitted by enteric motor neurons that release excitatory or inhibitory transmitters to the myenteric interstitial cell of Cajal (ICC‐MY). ICC will then transduct the signal to the smooth muscle cells, which are the effectors of pyloric contractions.

The nitrergic pathway has been shown to be the main inhibitory pathway. ²⁷ Both myenteric and submucosal layers of pyloric muscles are innervated by nitrergic neurons. ¹⁷ Lesions in the nitrergic pathway can be due to the loss of NO synthase (NOS), loss of NOS neurons, or loss of ICCs, which regulate the postjunctional nitrergic response. ²⁸ Alteration of this pathway can thus lead to an abnormal pyloric relaxation, disturbing the regulation of GE. Loss of ICCs has been reported in full‐thickness biopsies of patients with gastroparesis, even in the pylorus, and appears to have clinical relevance in gastroparesis. ²⁹ , ³⁰

Pyloric physiology has also been assessed by manometric studies, where it has been described as an area of 1.5 to 2 cm length, with a higher pressure of 5 to 9 mmHg in the fasting period. ³¹ , ³² Fisher and Cohen showed that pyloric pressure increased after duodenal acidification, leading to a diminution of the retrograde movement of duodenal content in stomach. ³¹ , ³³ These pyloric contractions also occurred after duodenal triglyceride infusion and seemed to interrupt GE. ³⁴ However, the timing and mechanistic steps of pyloric relaxation and contraction are still not completely understood. Camilleri et al. revealed that antral motility was related to emptying of solid food but that intestinal motility could also impact GE. ¹³ , ³⁵ Studies combining detailed imaging combined with manometry suggested that GE occurred during pyloric relaxation and mostly in periods without gastric contractions. ³⁶ In addition, antral propagated contractions coordinated with increased pyloric resistance. ³⁷ Thus, it has been suggested that antral contraction coordinated with pyloric contractions leaded to food trituration, and the contribution of a “pressure pump” rather than a “peristaltic pump” controlled by pylorus opening controlled GE. ³⁶ , ³⁷

The most common causes of gastroparesis are diabetes, post‐surgical, or drug‐induced, especially by opiates. ³⁸ However, idiopathic gastroparesis remains the most common diagnosis. Pyloric dysfunction is one of the key mechanisms of gastroparesis, but it may depend on the etiology of gastroparesis. It was first described in the context of diabetic gastroparesis by Mearin et al. in 1986. ¹¹ This study assessed the pyloric manometric profile in 24 diabetic patients and observed unusually tonic contractions described as “pylorospasm” in 14 of them. Evidence of extraintestinal autonomic neuropathy was also confirmed in 14 patients, and these results suggested that pylorospasm might be due to diabetic neuropathy. ¹¹ In post‐surgical gastroparesis, proximal gastric vagotomy has been shown to inhibit gastric tone and delay GE without altering antral contractions. ³⁹

The role of endogenous opiates in the regulation of pyloric function has also been shown. ⁴⁰ The presence of enkephalinergic nerves has been confirmed in the muscularis of the pylorus in dogs and humans. ⁴¹ , ⁴² The neural response to opiates has been characterized in dogs as inhibition of cholinergic excitatory and nitrergic inhibitory junction potential in pyloric muscle. ⁴³ Endogenous opiates have been shown to contribute to pyloric contraction induced by duodenal acidification. ⁴⁴ Moreover, this effect was reversed by naloxone, the opioid antagonist. ⁴³ , ⁴⁵ Exogenous opiates have been shown to induce stimulation of pyloric tone and phasic contractility. ⁴⁶ Thus, exogenous opioids may result in a delay in GE. These data are also consistent with the clinical data demonstrating the worst outcome and more symptoms related to GE in patients with gastroparesis with chronic use of opioids. ⁴⁷ These considerations are of particular interest in the United States, where the prevalence of opioid intake in gastroparesis could reach up to 40% in some series. ⁴⁸ But these epidemiological data seem to be different in Europe and Canada, where opioid use is at least four times lower. ⁴⁹ , ⁵⁰

Finally, recent physiologic tools were developed to assess the functionality of the pyloric sphincter. The functional luminal imaging probe known as EndoFlip® is a 240‐cm catheter with a bag mounted at its distal end. ⁵¹ , ⁵² This distal end has impedance electrodes enclosed in the bag that allow the different measures. Subjects are usually asked to fast at least 8 h before the procedure. ⁵³ The device is positioned in the pylorus either through the endoscope, under endoscopic visualization, ⁵² or either by fluoroscopy guidance. ⁵¹ Pyloric angulation has not been reported as a technical obstacle. The impact of filling in the stomach or using anesthetics has not been investigated. ⁵⁴ After positioning the device, the balloon is filled to set volumes of 10, 20, 30, 40, and 50 ml using a stepwise protocol. Measurement of the cross‐sectional area (CSA), the bag pressure, and the distensibility index are recorded. The pyloric distensibility index (P‐DI) is calculated in the zone with the narrowest CSA with the corresponding intra‐bag pressure. ⁵³ Normal and pathological values still need to be confirmed. In the study of Gourcerol et al., fasting P‐DI in 21 nonanesthetized volunteers was 25.2 ± 2.4 mm²/mmHg at 40 ml of bag inflation. ⁵¹ A second Indian study in 20 anesthetized healthy subjects reported a P‐DI of 8.4 ± 4.7 mm²/mmHg. ⁵⁵ The last study from the Mayo Clinic reported a P‐DI of 10.9 ± 4.8 mm²/mmHg in 24 unsedated healthy volunteers. ⁵⁶ Thus, the possibility of ethnical/racial or anesthetic‐induced differences has been suggested. ⁵³

In the French study, the lower 90th percentile was set at 10 mm²/mmHg at 40 ml of inflation, and this threshold was used in some studies to define decreased P‐DI. ⁵¹ The distensibility of the sphincter has been shown to be lower in patients with gastroparesis than in healthy volunteers. ⁵¹ Decreased P‐DI, according to this definition, has been found in 56.5% of patients with diabetic gastroparesis and 51.5% of patients with idiopathic gastroparesis, while it was present in only 10% of healthy volunteers. ⁵⁷ P‐DI was also decreased in 60% to 75% of patients after antireflux surgery, or after esophagectomy, but not after sleeve gastrectomy. ⁵⁸ Several studies have shown that distensibility was correlated with GE and gastroparesis symptoms. ⁵² , ⁵³ Finally, a decreased P‐DI has also been observed in patients with chronic nausea and vomiting without delayed GE, suggesting that pyloric dysfunction could explain “gastroparesis‐like” symptoms in this group of patients. ⁵⁹ However, patients with severely delayed GE still had a lower distensibility index. P‐DI was even better than GE test to correlate with symptoms in this study; however, these studies are retrospective and at risk of biases. ⁵³ Whether P‐DI could be more appropriate than GE or not remains unclear, no study has assessed EndoFlip tool in functional dyspepsia with normal gastric emptying. Therefore, EndoFlip still needs to be considered as an additional exploratory tool in the pathophysiology of gastroparesis. Based on these data, the pylorus has appeared as an important therapeutic target to accelerate GE.

3. ENDOSCOPIC THERAPIES TARGETING THE PYLORUS

3.1. Pyloric balloon dilation

The first endoscopic technique used to treat pyloric dysfunction was pyloric balloon dilation, first reported in children. ⁶⁰ Hydraulic balloon dilation up to 18 mm led to resolution of symptoms in 13 children out of 19 (68%). ⁶⁰ Retrospective studies with this technique have then been reported in adults. ⁶¹ , ⁶² , ⁶³ Studies focusing on refractory gastroparesis reported an acute efficacy in 50% of patients at 2 months, with a sustained efficacy in one third of patients at 2 years. ⁶³ , ⁶⁴ No complications were shown with the hydraulic balloon dilation, used up to 20 mm. Recently, the EsoFLIP, combining EndoFlip measurement, and hydraulic dilation at a diameter up to 30 mm have been assessed and led to a symptomatic improvement in 57% of patients at 4 months. ⁶⁵ The study by Jehangir et al. reported symptomatic improvement with pyloric dilation after a previous pylorotomy in 5 patients out of 13. ⁶³ Patients who had symptomatic improvement had a lower P‐DI before dilation (7.2 ± 1.0 vs. 13.9 ± 2.1 mm²/mmHg; p = 0.02).

3.2. Botulinum toxin

The most evaluated technique has been the injection of botulinum toxin into the pylorus. Several retrospective and open‐label studies reported an improvement in gastroparesis symptoms, including vomiting, nausea, postprandial fullness, and bloating, with a short‐term improvement in 50% to 77% of patients. ⁶⁶ , ⁶⁷ , ⁶⁸ , ⁶⁹ Most of these studies noted an acceleration of GE. ⁶⁶ , ⁶⁸ , ⁶⁹ However, two randomized placebo‐controlled trials did not confirm the superiority of botulinum toxin over a placebo regarding the primary endpoint, which was a decrease in the Gastric Cardinal Symptom Index (GCSI). ⁷⁰ , ⁷¹ The secondary endpoints revealed a significant decrease in postprandial fullness and nausea one month after toxin injection, as compared to the sham procedure, in the study by Arts et al. ⁷⁰ In these two randomized studies, both botulinum toxin and placebo resulted in an acceleration of GE. ⁷¹ As such, the European and American recommendations did not support the routine use of this treatment. ² , ⁷ Noteworthy, these trials included a limited number of patients, with 23 and 32 patients, respectively. In addition, none of these studies assessed pyloric physiology prior to injection, as distensibility measurements were not available at the time when they were conducted. More recently, one study identified that P‐DI before intrapyloric injection of botulinum toxin may predict the outcome of this treatment. ⁷² Indeed, patients with altered P‐DI had greater symptom improvement, a higher quality of life, and better GE after injection, than patients with normal P‐DI.

3.3. Peroral endoscopic pyloromyotomy (G‐POEM): technical considerations

The development of endoscopic techniques for submucosal dissection led to a new area of endoscopic procedures. The first peroral endoscopic myotomy was performed in the lower sphincter of the esophagus for the treatment of achalasia. ⁷³ The technique was then applied to the pyloric sphincter for the treatment of gastroparesis. ¹² Technically, the procedure is performed under general anesthesia and orotracheal intubation. The patient should be hospitalized, and the prophylactic administration of systemic antibiotics before the procedure is recommended. ⁷⁴ G‐POEM should always be performed with CO₂ insufflation. ⁷⁴ The different steps of the endoscopic procedure are shown in Figure 2. During the upper endoscopy, a mucosal incision is performed 3 to 5 cm prior to the pylorus. ¹² , ⁷⁵ The incision is usually 1.5 to 2 cm in length and may be longitudinal or transverse. Most operators perform the incision in the posterior wall or greater curvature; however, there is no clear recommendation on the best location for the mucosal incision. ⁷⁴ The endoscope is then maneuvered into the submucosal area, and a tunnel is created by injection of a colored preparation into the antral submucosa. Tunneling is performed until visualization of the white pyloric ring. The myotomy is then performed using an endoscopic knife. The length of the myotomy should be 2 to 3 cm, but no clinical study has assessed this. ⁷⁴ Local administration of antibiotics in the tunnel is not recommended. Finally, the tunnel is closed with clips, and the endoscopy checks for bleeding or other complications. There is no standardization of the equipment used, and the choice of a knife or specific endoscopic material is left to the operator, according to his personal experience.

FIGURE 2.

Endoscopic images of the gastric peroral endoscopic myotomy (G‐POEM) procedure. (A) Identification of the pylorus prior to the procedure; (B) mucosal incision in the greater curvature; (C) progression in the submucosal tunnel; (D) identification of the pyloric ring; (E) pyloromyotomy; (F) closure of the tunnel entry.

After the procedure, the patient remains fasted until Day 1. A control by a second upper endoscopy on Day 1 can be considered to exclude a leak and to confirm the effectiveness of the incision closure with the clips being still present. Some centers use a contrast study after the procedure. ⁷⁶ In the absence of complications, the patient might restart feeding progressively (liquid on the first day and then a soft diet) and could be discharged on day 1. Proton pump inhibitors should be given to the patient during the procedure and for 4 weeks after the procedure twice a day. The mean length of stay reported in the last studies after G‐POEM was 2 ± 2 days. ⁷⁵ , ⁷⁷ One study evaluated the possibility of a same‐day discharge 4 h after the procedure and observed a readmission rate of 7.4% within 30 days in a group of 54 patients, similar to that of a group with inpatient recovery. ⁷⁸

The technical success of the technique reached 100% in most of the studies, and there was no definitive contraindication to the procedure reported. ⁷⁹ These results could be reached after a learning curve for endoscopists to carry out the procedure under the best conditions. As such, in a study by Reja et al., operators showed a continuous improvement during the first 18 cases. ⁸⁰ The mean duration of the procedure was reduced during the learning process. Moreover, Rodriguez et al. suggested that the length of stay could also be reduced by the experience of the center. In their study, the last patients included had a shorter duration of hospitalization due to the greater experience of endoscopists and to better perioperative management. ⁸¹

4. EFFICACY AND SAFETY OF G‐POEM

4.1. Clinical efficacy of G‐POEM

Several open‐label studies reported clinical improvement after G‐POEM in refractory gastroparesis. ⁸¹ , ⁸² , ⁸³ , ⁸⁴ , ⁸⁵ , ⁸⁶ , ⁸⁷ The results of these trials are summarized in Table 1. The first retrospective trials reported a short‐term improvement ranging from 80% to 85% 3 months after the procedure. ⁸² , ⁸³ , ⁸⁶ Symptomatic improvement after G‐POEM has been reported in patients with diabetic, idiopathic, and post‐surgical gastroparesis. ⁸⁶ , ⁸⁸ The efficacy has also been confirmed in patients previously treated by toxin injection or by gastric electrical stimulation. ⁸⁹

TABLE 1.

Major studies assessing the efficacy of G‐POEM

Study and year	Number of patients	Etiology of gastroparesis	Definition of clinical success	Clinical responders	Gastric emptying Scintigraphy evolution
Khashab et al. 83 2016 multicenter retrospective	30	Diabetes 37%, Idiopathic 23% Post‐surgical 40%	Reduction in patient's self‐reported symptoms	86%, median follow‐up of 6 months (IQR 7–11)	Improvement 82% Normalization 47% (17 patients)
Gonzalez et al. 82 2017 single center retrospective	29	Diabetes 24%, Idiopathic 52% Post‐surgical 17%	Improvement in mean GCSI score	79% at 3 months 69% at 6 months	Normalization 70% (26 patients)
Dacha et al. 85 2017 single center, retrospective	16	Diabetes 56% Idiopathic 31% Post‐surgical 6%	Improvement in mean GCSI score and no hospitalization	81% at 3 months	Normalization 75% (12 patients)
Rodriguez et al. 125 2017 single center, retrospective	47	Diabetes 26% Idiopathic 57% Post‐surgical 17%	Symptom relief measured by GCSI score	Improvement in mean GCSI score at 3 months (31 patients)	Decrease in the mean 4 h gastric retention (16 patients)
Malik et al. 84 2018 single center, retrospective	13	Diabetes 8% Idiopathic 31% Post‐surgical 61%	Subjective self‐reported improvement	73% at 3 months	Improvement 66% Normalization 33% (6 patients)
Kahaleh et al. 86 2018 multicenter, retrospective	33	Diabetes 21% Idiopathic 36% Post‐surgical 36%	Improvement in mean GCSI score	85%, median follow‐up of 11.5 months (2–31 months)	Decrease in 2 h and 4 h gastric retention rate
Rodriguez et al. 81 2018 single center, retrospective	100	Diabetes 21% Idiopathic 56% Post‐surgical 19%	Improvement in mean GCSI score	Improvement in mean GCSI score at 3 months	Improvement 78% Normalization 57% (63 patients)
Mekaroonkamol et al. 87 2019, single center retrospective	30	Diabetes 40% Idiopathic 40% Post‐surgical 17%	Improvement in GCSI score of at least 1.0	71% at 6 months 57% at 12 months	Improvement 78% Normalization 48%
Jacques et al. 95 2018 single center Prospective	20	Diabetes 50% Idiopathic 20% Post‐surgical 5%	Improvement in GCSI score of at least 0.75	90% at 3 months	Normalization 30%
Vosoughi et al. 99 2020 multicenter retrospective	37	Diabetes 32% Idiopathic 27% Post‐surgical 27%	Improvement in GCSI score of at least 1.0	70% at 12 months	Improvement 46% Normalization 32% (28 patients)
Ragi et al. 75 2020 multicenter, retrospective	76	Diabetes 34% Idiopathic 36% Post‐surgical 20%	Improvement in GCSI score of at least 1.0	66% at 12 months	Decrease in 2 h and 4 h gastric retention rate (65 patients)
Tan et al. 91 2020 Single center Retrospective	79	Post‐surgical 100%	Improvement ≥25% in at least two of the GCSI subscales	77% at 6 months 59% at 12 months 34% at 24 months	Data not shown
Abdelfatah et al. 77 2021, Single center Retrospective	97	Diabetes 42% Idiopathic 47% Post‐surgical 11%	Improvement in GCSI score of at least 1.0	75% at 6 months 69% at 12 months	Improvement 90% Normalization 63% (74 patients)
Conchillo et al. 93 2021 Prospective	24	Diabetes 25% Idiopathic 46% Post‐surgical 29%	Improvement in GCSI score of at least 1.0	58% at 6 months 33% at 12 months	–
Gregor et al. 89 2021 Prospective	52	Diabetes 40% Idiopathic 40% Post‐surgical 19%	Improvement in GCSI score of at least 1.0	58% at 6 months 48% at 12 months	Decrease in 4 h gastric retention rate
Vosoughi et al. 92 2021 Multicenter, Prospective	75	Diabetes 24% Idiopathic 41% Post‐surgical 35%	Improvement in GCSI score of at least 1.0	56% at 12 months	Normalization 47% (53 patients)

The definition of success differed across studies, and this could have explained some discrepancies in the reported success rates. To define clinical success, studies used either self‐reported improvement or improvement in the GCSI score, which varies from 0 (no symptoms) to 5 (worst symptom severity) and is validated to assess symptom evolution in gastroparesis. ⁹⁰ In the latter studies, clinical success was defined as an improvement of at least 1 point in the average GCSI, with a decrease of at least 25% in two of its three subscales. Using self‐reported improvement or GCSI, studies reported a short‐term clinical efficacy between 70% and 80% of patients after 3 to 6 months. ⁷⁷ , ⁸² , ⁸³ , ⁸⁷ Long‐term efficacy beyond 12 months ranged from 53% to 70% of patients in retrospective studies, ⁷⁵ , ⁷⁷ , ⁹¹ and 33% to 56% in prospective trials, ⁸⁹ , ⁹² , ⁹³ based on GCSI variation. A recent meta‐analysis involving 10 studies and 482 patients calculated a pooled success rate of 61% at 1 year. ⁹⁴ Longer‐term data were reported in a study by Tan, with a sustained efficacy in only 33% of patients at 2 years, whereas Ragi et al. reported a sustained efficacy of 74% at 24 months. ⁷⁵ Abdelfatah et al. also reported a sustained efficacy at 3 years in only six out of seven patients. ⁷⁷ Secondary outcomes revealed that all GCSI subscales, namely satiety, bloating, and nausea, improved after G‐POEM by a mean reduction of 1.3 points. ⁷⁵ , ⁹⁵ Reduction of abdominal pain was also reported in the short‐term results in 50% of patients but was not further confirmed in long‐term series. ⁷⁷ , ⁸⁷ , ⁹² To date, only two randomized trials aimed to assess the clinical efficacy of G‐POEM. In the first study, Martinek et al. randomized 41 patients into two groups, the G‐POEM procedure or a sham procedure. Clinical success defined as a decrease of at least 50% in the GCSI score at 6 months reached 71% of patients in the G‐POEM group vs. 22% in the sham group (p = 0.005). ⁹⁶ After 6 months, patients in the sham group were allowed to perform a cross over, and 9 patients out of 12 (75%) reported clinical success of the procedure at 6 months. The efficacy was higher in patients with diabetic gastroparesis reaching an efficacy of 89% in this sub‐group. The second study by Gonzalez et al. was only reported in abstract form. Forty patients were randomized into two groups, the G‐POEM procedure or botulinum toxin injection. ⁹⁷ In this study, the success rate based on a 1‐point decrease in GCSI did not differ between the two treatments after 3 months (65% vs. 47%; p = 0.27) and 12 months (67% vs. 57%, p = 0.58). A third ongoing randomized trial against a sham procedure is currently recruiting subjects (Clinicaltrials.gov, number NCT04552184).

G‐POEM was shown to also be effective in healthcare use. A reduction of antiemetic medications was reported in 50% to 70% of patients who responded to the procedure. ⁷⁷ , ⁸⁷ G‐POEM also led to a reduction in the number of gastroparesis‐related hospitalizations. In two studies, the number of days of hospitalization decreased from 3.2 to 0.4 and the number of emergency visits decreased from 3.3 to 0.2 after G‐POEM, suggesting the procedure could be cost‐effective. ⁸⁷ , ⁸⁹ Nutritional status was improved, with body mass index increasing after G‐POEM (from 26.1 ± 6.1 to 26.8 ± 5.0). ⁸⁷ Four patients with nutrition dependency came off parenteral or tube feed nutrition in a study by Mekaroonkamol et al. ⁸⁷ , ⁹⁸ Finally, these improvements resulted in a significant improvement in the patients' quality of life, assessed by the Gastrointestinal Quality of Life Index (GIQLI), ⁹⁵ PAGI‐QoL scale (Patient Assessment of Upper Gastrointestinal Disorders‐Quality of Life), ⁹⁵ or SF‐36 questionnaires. ⁸⁷ , ⁹²

The impact of G‐POEM on GE is summarized in Table 1. An improvement in all scintigraphy parameters has been reported with a significant acceleration of the GE half time (T½). ⁷⁵ , ⁹⁵ An acceleration of the GE rate was noted in 46% to 90% of patients, ⁷⁷ , ⁹⁹ and a normalization of the 4‐h gastric retention rate was observed in 30% to 75% of patients. ⁸⁵ , ⁹⁵ Vosoughi et al. observed that the decrease in the 4‐h gastric retention and the decrease in the GCSI score were moderately but positively correlated (r = 0.29; p = 0.046), ⁹² suggesting that clinical efficacy of G‐POEM could be driven, at least in part, by an acceleration of GE. In fact, studies compared EndoFlip measures before and after G‐POEM and showed significant changes after the procedure as shown in Figure 3. ¹⁰⁰ , ¹⁰¹ G‐POEM led to an increase in the CSA and a decrease in the bag pressure, leading to an increase in the postprocedure P‐DI index measured immediately after the procedure that was sustained 3, ⁹³ , ⁹⁵ , ⁹⁹ 6, and 12 months after the procedure. ⁸⁹

FIGURE 3.

Representative recording of pyloric distensibility before and after G‐POEM procedure. Measurements were made in the same patient with the EndoFlip bag inflated at 40 ml.

A few studies addressed whether pyloromyotomy should be performed endoscopically or surgically. One study compared 30 patients treated with pyloroplasty to 30 patients treated with G‐POEM. ¹⁰² Both procedures resulted in a similar significant improvement in GCSI score and GE. Pyloroplasty was associated with a longer length of stay and a higher rate of complications (16.7% vs. 3.3%; p = 0.086), suggesting that the endoscopic procedure should be favored over the surgical approach. A second multicenter study compared G‐POEM with pyloromyotomy and included 102 patients (39 with G‐POEM and 63 with surgery). ¹⁰³ This study ended up with similar results, since the efficacy did not differ between the two procedures, although fewer adverse events were reported in the G‐POEM group (13% vs. 33%; p = 0.02). ¹⁰³ Finally, a meta‐analysis comparing 375 G‐POEM with 332 surgical pyloric procedures reported similar efficacies for the two techniques ¹⁰⁴ and no difference in the adverse event rate.

4.2. Predictive factors of clinical success

Two studies showed that neither age nor sex had any predictive impact on G‐POEM in a multivariate analysis. ⁸² , ¹⁰⁵ A higher body mass index was associated with a risk of G‐POEM failure in a study by Abdelfatah et al. (odds ratio (OR) 1.097 [1.02–1.17]). ⁷⁷ In a study by Ragi et al., the GCSI satiety subscale score was the best predictor of clinical success at 12 months (OR 3.41 [1.01–11.54]). ⁷⁵ On the contrary, patients with abdominal pain or receiving psychiatric medication had a higher risk of G‐POEM failure (OR 1.33 [0.11–1.0]). ⁷⁷ Pre‐therapeutic symptomatic severity was predictive of the clinical success of the procedure. ⁷⁵ , ⁹² As such, baseline GCSI higher than 2.6 was associated with clinical success at 12 months (OR 3.23 [1.06–9.9]). ⁹² By contrast, a longer duration of gastroparesis was associated with G‐POEM failure in two studies (OR 1.4 [1.07–1.8]). ⁷⁷ , ⁸⁸ Finally, the etiology of gastroparesis was not associated with the clinical outcome of G‐POEM. ⁸⁸ , ¹⁰⁵

Data on GE remain controversial. A study by Vosoughi et al. showed that severe gastroparesis based on a 4‐h gastric retention rate > 20% was associated with clinical success 12 months after G‐POEM (OR 3.65 [1.14–11.66]). ⁹² In a study by Ragi et al., a higher 4‐h retention rate was associated with failure of the procedure (OR 0.97 [0.95–1.00]). ⁷⁵ Interestingly, regional scintigraphy assessment identified that patients with more proximal retention patterns were the best responders to G‐POEM, as compared with patients with distal retention. ¹⁰⁶ Recent studies focused on the assessment of pyloric dysfunction, attempting to confirm its predictive role in the outcome of G‐POEM. Measurement of P‐DI can be performed before, during, and/or after the G‐POEM procedure. A study by Jacques et al. identified that a low pre‐therapeutic P‐DI (<9.2 mm²/mmHg) was predictive of clinical success at 3 months. ⁹⁵ This threshold predicted the clinical response with a sensitivity of 100% and a specificity of 72%. In another study, a similar P‐DI (<10 mm²/mmHg) predicted a symptomatic response to botulinum toxin. ⁷² Other studies failed to confirm the predictive outcome of pre‐operative EndoFlip metrics. ⁸⁹ , ⁹⁹ The study of Vosoughi et al showed that post‐procedural CSA could be the best predictor for success and acceleration of GE, with an odds ratio of 1.02 [1.01–1.04]; p = 0.008, but this is the only study that stained CSA as better than P‐DI. Post‐operative CSA >154 mm² with a distension volume of 40 ml was predictive of clinical success at 1 year with a sensitivity of 71% and a specificity of 91%. ⁹⁹ This result suggested a dose response to the treatment (the higher the post‐operative CSA, the better the clinical outcome). The study of Conchillo et al. reported that post‐operative P‐DI increase defined as at least 20% improvement of baseline was correlated with clinical success. ⁹³ These results suggest that this tool could be used before and/or after the procedure to predict and/or confirm the effectiveness of pyloromyotomy and may become a per‐procedural guiding tool, as suggested for lower esophageal sphincter in achalasia. ⁵⁷ To date, given the efficacy of G‐POEM, the lack of alternative treatment, and the current data on EndoFlip, this tool cannot be recommended as a screening procedure. Pre‐therapeutic antro‐duodenal manometry parameters were also evaluated, but no parameters have been associated with the clinical success or failure of G‐POEM. ⁹³

Finally, a previous response to an intrapyloric botulinum toxin injection has been proposed as a screening test before performing G‐POEM. In a meta‐analysis of 375 patients treated with G‐POEM, previous botulinum toxin administration had a positive influence on the clinical success of G‐POEM treatment, based on scintigraphic GE acceleration but not on symptomatic improvement. ¹⁰⁴ In addition, a study by Gregor et al. reported a better outcome for patients who previously responded to toxin injection in a total of 25 patients collected prospectively. ⁸⁹ In the same way, in a short series, three patients previously improved by toxin injection were improved by G‐POEM, whereas one patient who did not respond to toxin injection did not respond to G‐POEM. ¹⁰⁷

4.3. Safety and adverse events after G‐POEM

A multicenter review of adverse events related to G‐POEM was performed by Ichkhanian et al. ¹⁰⁸ The authors reported an adverse event rate of 14% between 2016 and 2018. Most of the adverse events (58%) were reported during the first 48 h after the procedure or during the procedure (29%), and late adverse events were rare. Intraprocedural complications included mucosal injury, which was treated endoscopically during the procedure, and capnoperitoneum, which could be decompressed percutaneously. ¹⁰⁸ Only clinical capnoperitoneum is considered, defined as significant abdominal distension leading to pain, or respiratory impact, whereas asymptomatic radiologic capnoperitoneum is not considered as a complication. The most frequent complication was pain (50% of complications), but it was almost always managed medically. The majority of late complications (>48 h) were bleeding and were managed endoscopically. ¹⁰⁸ No mortality was reported. Adverse events were classified as mild or moderate in 93% of cases, and severe adverse events accounted for only 6% of complications, according to the ASGE lexicon for endoscopic adverse events. ¹⁰⁹

The rate of adverse events dropped to between 6% and 7% in the last series, as shown in pooled analysis. ⁹⁴ , ¹¹⁰ The last prospective trials reported a similar rate of adverse events around 6%, which might be explained by the operators having more experience. ⁸⁹ , ⁹² In these series, all the complications were classified as mild to moderate. ⁸⁹ , ⁹² Capnoperitoneum was the most frequent, treated by needle decompression (3% of patients). ⁹² Bleeding accounted for 1% to 2% of patients and was managed endoscopically. ⁷⁵ , ⁸⁹ Finally, mucosotomy or mucosal injuries also accounted for 1% to 2% of patients and were endoscopically managed during the procedure. ⁸⁹ , ⁹² The experience of the endoscopist was shown to be the strongest predictive factor of complications, with an OR of 3.03 [1.03–8.94] for endoscopists who performed fewer than 20 previous myotomies. Also, in a study by Reja et al., endoscopists who achieved the learning curve of 18 patients had fewer complications. ⁸⁰

5. ALTERNATIVE TREATMENTS AND FUTURE DIRECTIONS

To date, the most validated alternative treatment to G‐POEM in severe refractory gastroparesis is gastric electrical stimulation (GES). ¹¹¹ In randomized controlled trials, GES has shown a significant reduction of nausea and vomiting in patients with refractory gastroparesis. ¹¹² There are no criteria to guide the choice of treatment, which therefore depends on the local possibilities and experience. One study from Shen et al. aimed to compare the issue of these two treatments. ¹¹³ This study used a propensity score‐matched analysis to compare two groups of 26 patients each, whatever the symptomatic profile. Their data suggest a higher clinical response rate at 24 months in the G‐POEM group (76% vs. 54%). However, it has been suggested that GES is performed preferentially in patients with symptoms of nausea and vomiting, whereas G‐POEM can be performed in patients with other symptoms related to delayed GE, such as early satiety or postprandial fullness. ¹¹⁴ Whether GES or G‐POEM is more effective at treating refractory gastroparesis associated with nausea and/or vomiting remains uninvestigated to date.

Combining GES procedure with pyloroplasty has first been assessed. This combined procedure was safe and led to an acceleration of GE as compared to GES alone. ¹¹⁵ A second study confirmed the symptomatic improvement based on total symptom score in 27 patients after a mean follow‐up of 17 months. ¹¹⁶ Therefore, performing G‐POEM instead of pyloroplasty after GES has been suggested as an augmentative therapy. Two series of 5 and 22 patients confirmed the feasibility and safety of G‐POEM after prior therapy with GES. ¹¹⁷ , ¹¹⁸ These two studies reported symptomatic improvement based on a reduction in the mean GCSI score and normalization of GE in 60% of the patients. In a prospective study by Gregor et al., all patients previously treated with GES were clinical responders to G‐POEM. ⁸⁹ Future controlled randomized studies are therefore required to address the superiority of performing both techniques over one or the other alone.

Technical issues have been assessed to optimize the endoscopic technique and improve its clinical outcome. In a study by Abdelfatah et al., performing a double pyloromyotomy resulted in a higher clinical response than the standard single pyloromyotomy technique. ¹¹⁹ Double pyloromyotomy could therefore be discussed in patients with insufficient improvement in P‐DI after a single myotomy. The 6‐month efficacy reached 86% in the 35 patients treated with double pyloromyotomy as compared to a 67% efficacy in the other 55 patients (p = 0.04). There were no differences in the procedure duration or adverse events. However, GE was not compared between the two groups, and whether this procedure leads to a long‐term improvement or not remains unknown. ¹²⁰ In addition, a decrease in the success rate was observed beyond 1 year. ⁸⁹ , ⁹² Performing another myotomy has appeared as an effective strategy in patients with achalasia, with relapse after a first esophageal myotomy. ¹²¹ This strategy of “redo G‐POEM” has only been reported to date in several case reports after relapse or lack of effectiveness after a first G‐POEM procedure. ¹²² , ¹²³ , ¹²⁴

6. CONCLUSION

G‐POEM is a safe technique with encouraging clinical efficacy as a treatment for patients with refractory gastroparesis. Targeting the pylorus seems to be effective at accelerating GE and thus leads to symptomatic improvement of nearly 60% to 80% of patients at mid‐term follow‐up. Validation of the technique in further randomized trials is highly expected to confirm the initial results. Future directions will encompass optimization of the technique using either predictive functional tools and/or combination with other techniques.

AUTHOR CONTRIBUTION

Heithem Soliman contributed to the design of the article, reviewed the litterature, drafted the article, and approved the final version. Guillaume Gourcerol contributed to the designe of the article, critically reviewed the article, and approved the final version.

CONFLICT OF INTEREST

The authors disclose no conflicts.

Soliman H, Gourcerol G. Targeting the pylorus in gastroparesis: From physiology to endoscopic pyloromyotomy. Neurogastroenterology & Motility. 2023;35:e14529. doi: 10.1111/nmo.14529