Key points

Introduction

A growing body of literature supports the use of laparoscopic pyloroplasty as a minimally invasive treatment of refractory gastroparesis that has failed conservative measures and for benign gastric outlet obstruction. Endoscopic pyloric dilation, stent placement, and Botox have been described for similar indications, but often with transient or mixed results. With wider acceptance of endoscopic submucosal dissection and per-oral endoscopic myotomy (POEM) as a viable treatment for achalasia, it is not surprising that a similar tunneling approach would be applied to dysfunctions of the pylorus resulting in poor gastric emptying. Per-oral pyloromyotomy (POP) has recently been proposed as an endoscopic alternative to surgical pyloroplasty or pylormyotomy because it is less invasive by its nature and potentially more durable than current endoscopic treatments. Using technologies common to POEM, POP involves the creation of a submucosal tunnel followed by division of the pyloric sphincter muscles. Here the physiologic basis of pyloric disruption for various gastric disorders as well as the development of POP, its potential indications, the precise steps in this highly skill-dependent technique, and early data in humans is reviewed.

Introduction

A growing body of literature supports the use of laparoscopic pyloroplasty as a minimally invasive treatment of refractory gastroparesis that has failed conservative measures and for benign gastric outlet obstruction. Endoscopic pyloric dilation, stent placement, and Botox have been described for similar indications, but often with transient or mixed results. With wider acceptance of endoscopic submucosal dissection and per-oral endoscopic myotomy (POEM) as a viable treatment for achalasia, it is not surprising that a similar tunneling approach would be applied to dysfunctions of the pylorus resulting in poor gastric emptying. Per-oral pyloromyotomy (POP) has recently been proposed as an endoscopic alternative to surgical pyloroplasty or pylormyotomy because it is less invasive by its nature and potentially more durable than current endoscopic treatments. Using technologies common to POEM, POP involves the creation of a submucosal tunnel followed by division of the pyloric sphincter muscles. Here the physiologic basis of pyloric disruption for various gastric disorders as well as the development of POP, its potential indications, the precise steps in this highly skill-dependent technique, and early data in humans is reviewed.

The pylorus and pyloric disruption

The pylorus is frequently involved in benign digestive disorders, including gastric outlet obstruction and gastroparesis (delayed gastric emptying in the absence of mechanical obstruction). One of the most common examples of this is congenital hypertrophic pyloric stenosis. This idiopathic condition has a population incidence of 2 to 4 per 1000 live births and presents with gastric obstructive symptoms (vomiting and failure to feed) at a very early age. For more than 100 years, it has been effectively palliated by the division of the abnormally hypertrophic pyloric sphincter. Today, pediatric pyloromyotomy is 99% effective at relieving patient symptoms and is increasingly performed laparoscopically. In the adult population, gastric outlet obstruction is most commonly due to reactive pyloric stenosis—usually secondary to chronic peptic ulcer disease or caustic ingestion. As a purely mechanical problem, this diagnosis and its treatment are fairly straightforward and, if unresponsive to medical therapy including endoscopic dilation, responds well to surgical emptying treatments such as a pyloroplasty or distal gastrectomy. More importantly, pyloric dysfunction is implicated in the rather ill-defined category of gastroparesis. Gastroparesis, defined as a delay in gastric emptying not related to a mechanical outlet obstruction, can be idiopathic, diabetic, or postsurgical. Pyloric disruption has been used for this diagnosis with varying degrees of success.

Surgical pyloroplasty has been successfully used to treat benign gastric outlet obstruction for decades, and it at least theoretically makes sense. Pyloroplasty (per the Heineke-Mikulicz method) involves longitudinal division of the full thickness of the pyloric ring, thereby obliterating the cause of obstruction. This division is followed by transverse closure of the full-thickness defect, thereby assuring a geometrically enlarged lumen. An alternative procedure is a pyloromyotomy, which involves the longitudinal division of the serosa and muscular layers of the distal stomach, pylorus, and proximal duodenal bulb, leaving the mucosa intact. The choice of whether one uses a pyloromyotomy or a pyloroplasty depends somewhat on the procedure; for example, the current standard of care for congenital hypertrophic pyloric stenosis in children is pyloromyotomy, but the usual treatment of gastric outlet related to peptic ulcer disease is a pyloroplasty plus vagotomy (or antrectomy). Following esophagectomy with gastric interposition, benign gastric outlet obstruction (which affects approximately 15% of patients) has been described as benefitting from surgical pyloroplasty or pyloromyotomy, although it can also be treated with endoscopic balloon dilation. Pyloric disruption by endoscopic balloon dilation can also be used to treat obstruction from chronic ulcers or caustic stricture, although in these settings definitive treatment often requires surgery. Endoscopic self-expanding metallic stents are another intriguing possibility for treating both benign pyloric stenosis and gastroparesis. Studies to date are small and uncontrolled, but several have demonstrated symptomatic improvement and/or reduced gastric-emptying times with transpyloric stenting. Unfortunately, this technique is plagued by the issue of migration, which currently hinders any long-term use of stents.

Regarding gastroparesis, the role of pyloric disruption has only recently come to the fore with a growing body of literature citing benefit in this context. Gastroparesis is historically an ill-defined and highly complex disease. It is thought to stem from a combination of damage to the vagus nerve or myenteric plexus, reduced hormone secretion (motilin and grehlin), and diminished growth factors required by the Interstitial Cells of Cajal (the gastric pacemakers). Atony of the corpus, continuous spasm of the pyloric ring, and desynchronization between stomach, duodenum, and pylorus are often described. Because of the complex pathophysiology of gastroparesis, it is not surprising that symptoms correlate poorly with gastric emptying, resulting in multiple subtypes of the disease. Diagnosis requires the presence of chronic symptoms (either nausea, vomiting, early satiety, postprandial fullness, or upper abdominal pain) typically assessed via the validated Gastroparesis Cardinal Symptom Index, GCSI, delayed gastric emptying, as evidenced by scintigraphy, and a contrast study ruling-out mechanical obstruction. Symptoms, which directly influence the patient’s quality of life, are currently used to assess treatment success. Even in this quagmire of a poorly defined, complex disorder with modest overall incidence (1.8%–4% in the general population), treating gastroparesis is still worthwhile. Patients are significantly debilitated; the economic burden is profound, and the incidence of gastroparesis-related hospitalizations has increased 158% between 1995 and 2004.

Dietary modification and symptom control with antinausea and antiemetic medications remain first-line therapy. Failing these, prokinetics often are next. Metoclopramide, the only US Food and Drug Administration (FDA)-approved medication for gastroparesis treatment in the United States, currently carries a black-box warning of tardive dyskinesia, and prolonged use is associated with tachyphylaxis. Gastric electrical stimulation is currently FDA-approved on protocol (humanitarian device exemption) for diabetic and idiopathic gastroparesis. Benefits have been found in relieving nausea and vomiting, with conflicting results regarding the effect on gastric emptying. A major issue with this device is its high cost and orphan device status, which makes it difficult to access in North America.

Endoscopic intrapyloric injection of botulinum toxin A, which was promising in initial open-label studies, has since demonstrated no difference from placebo in 2 randomized controlled trials. As a result, The American Gastroenterological Society no longer recommends intrapyloric botulinum toxin injections in the treatment of gastroparesis.

As stated earlier, both antral and pyloric dysfunctions have been demonstrated in the pathogenesis of gastroparesis. Although improved gastric emptying does not consistently correlate with improved symptoms, it is known that improved gastric emptying frequently correlates with improved symptoms in children and adults. It is also known that pyloroplasty has a long history as an effective and permanent gastric drainage procedure in the context of mechanical obstruction and elective vagotomy. Moreover, as early as 2007, foregut surgeons began reporting improved gastric emptying and substantially improved nonreflux symptoms in gastrointestinal reflux disease (GERD) patients who had received pyloroplasty in conjunction with fundoplication. Currently, several groups perform simultaneous laparoscopic fundoplication and pyloroplasty as routine practice in patients meeting the diagnostic criteria for both GERD and gastroparesis. A similar synergy has been seen with the addition of laparoscopic pyloroplasty to the use of the gastric stimulator.

Current evidence for laparoscopic pyloroplasty and gastroparesis

Evidence-based medicine might dictate that laparoscopic pyloroplasty should be the primary treatment of gastroparesis. Hibbard and colleagues published a retrospective, single-center review of prospectively collected data. Twenty-eight patients with refractory gastroparesis (25% diabetic) underwent laparoscopic Heineke-Mikulicz pyloroplasty or, in 2 patients, a laparoscopic-assisted pyloroplasty using a trans-oral circular stapler. Patients with prior or concomitant gastric surgery were excluded. Refractory gastroparesis was defined by classic symptoms and an abnormal gastric-emptying study in the absence of mechanical obstruction. By 3-month follow-up, 92% of patients reported improved or resolved symptoms ( P ≤.013, albeit not assessed by GCSI). Half the patients had a postoperative gastric-emptying study, which was normalized in 71% ( P = .001). Fourteen percent of patients had early recurrent symptoms, resulting in subsequent interventions (3 gastric stimulation, 1 fundoplication). Nevertheless, these results were striking enough to inspire several similar studies in rapid succession.

In another retrospective review by Toro and colleagues, 50 gastroparetic patients (10% diabetic) also underwent laparoscopic Heineke-Mikulicz pyloroplasty. Major differences in this study were inclusion of patients with prior or concomitant foregut surgery (68% and 64%, respectively) and the use of the GCSI to evaluate symptoms postoperatively. Although long-term follow-up is also unavailable, 2-month follow-up showed similar rates of symptom improvement and improved gastric-emptying times (82% and 96%, respectively, both P <.001). More than half of the patients (54%) had normalized gastric-emptying times, and there was no incidence of dumping syndrome after pyloroplasty. Of note, the 5 diabetics included in the study had substantial improvement in gastric-emptying times, suggesting that pyloroplasty is indeed effective even in this patient population.

Mancini and colleagues retrospectively evaluated the effects of laparoscopic and open pyloroplasty in 46 patients with gastroparesis (30% diabetic). Patients with prior gastric surgery, including stimulator placement, were excluded. Preoperative and postoperative evaluations involved the GCSI and gastric-emptying scintigraphy with follow-up of at least 12 months. Improved gastric emptying was found in 90% of patients, with 60% demonstrating normalized emptying values ( P = .001). Symptoms were assessed preoperatively and postoperatively for 41 patients, all of whom demonstrated statistically significant improvement (mean difference = −2.3, P <.005). This study confirmed the efficacy of surgical pyloroplasty for treating gastroparetic symptoms and gut dysfunction, including in diabetics.

Shada and colleagues provided more definitive evidence by examining pyloroplasty as first-line surgical therapy for refractory gastroparesis in a retrospective review of 177 patients, 58% of whom had concurrent fundoplication for presumed overflow reflux. Overall, 90% of gastroparetic patients had concurrent surgery of some kind (Heller myotomy, paraesophageal hernia repair, percutaneous endoscopic gastrostomy [PEG], or J tube). Gastroparesis was defined by scintigraphy, endoscopic visualization of retained food after prolonged NPO (nothing by mouth) status, or clinical symptoms suspicious for vagal nerve injury after complex foregut surgery. In fact, 13 patients had normal gastric-emptying scintigraphy times, but were highly symptomatic with retained solid food found repeatedly on endoscopy. Postoperative scintigraphy was performed in almost half the patients (n = 70), 85% of whom had improved emptying times and 77% of whom had normalized ( P = .001). In place of the GCSI, a standardized Symptom Severity Scale was used for evaluating symptoms. Of the 9 major symptoms queried, all showed statistically significant improvement at 1 month ( P = .0001) except early satiety ( P = .14). Of interest, all symptoms showed small increases between the 3- and 6-month evaluation, underscoring the need for long-term follow-up in this complex patient population. It should be noted that 11% of patients went on to another surgical treatment (gastrectomy or gastric stimulator) due to severe, refractory symptoms.

In summary, for nearly 2 decades pyloroplasty has been shown to improve gastric emptying, yet it has only recently been evaluated as a primary treatment for gastroparesis. To date, findings are promising, suggesting that surgical pyloroplasty is as safe and effective (if not more so) as any other available treatment option for medically refractory gastroparesis. It is now clear that surgical pyloroplasty does not preclude subsequent or concomitant interventions if needed.

The evolution of per-oral pyloroplasty

With the increasing acceptance of POEM for the treatment of achalasia, it was hypothesized early on that an analogous procedure could be used to address diseases of the pylorus, perhaps replicating the results of surgical pyloroplasty in the treatment of benign gastric outlet obstruction and gastroparesis. POP was proposed as a less invasive potential alternative. Using different nomenclature and slightly different techniques, this idea has been explored for more than a decade.

In 2001, Hagiwara and colleagues reported results from 5 adults with postoperative pyloric stenosis who had persistent obstruction after treatment with oral prokinetics and repeated endoscopic balloon dilation. Endoscopically, 2 to 3 radial incisions were made through mucosa and muscle using a needle knife. These incisions were extended from antrum to duodenal bulb followed by low-pressure balloon dilation for 15 to 20 minutes. Patients returned for repeat endoscopic balloon dilation after 1 week. No complications were reported, and all patients remained obstruction-free (based on esophagogastroduodenoscopy, symptoms, and barium swallow) at 10- to 26-month follow-up.

In 2005, Ibarguen-Secchia published a single-center pilot study examining outcomes in children with congenital pyloric stenosis. In this population of patients, the practice of pyloromyotomy has a long, well-vetted and successful history. The investigator, an accomplished endoscopist, proposed that an endoscopic approach might allow these patients to avoid general anesthesia and return to oral intake more quickly. Ten infants (7 boys), between 3 and 7 weeks old, underwent endoscopic pylotomyotomy for congenital pyloric stenosis under conscious sedation. Using either a needle knife (7) or standard sphincterotome (3), 2 incisions were made at the pylorus, one anterior and one posterior, first through the mucosa and then through the circular muscle fibers. Muscle division was continued until longitudinal fibers were seen. Mucosotomies were not closed. All patients were taking per os fluids within 2 hours, and 80% were discharged home the same day. The remaining patients were kept overnight because of persistent vomiting, which subsequently resolved. Follow-up occurred between 6 months and 2 years, and all patients continued to do well. Zhang and colleagues reported similar outcomes with the same technique in 8 children. In this study, 1 postoperative patient developed persistent vomiting after 1 month of resolved symptoms. Repeat endoscopic pyloromyotomy was performed, and the patient was symptom free thereafter. Zhang’s remaining patients were similarly doing well at follow-up from 2 to 9 months after the operation.

Endoscopic submucosal tunneling and esophageal myotomy were first demonstrated in animals by Pasricha and colleagues 3 years before Inoue reported on his POEM technique in man. The experiment involved 4 pigs and a submucosal tunnel made after incising the mucosa 5-cm lower esophageal sphincter. The submucosal space was developed with electrocautery, and the gastroscope was introduced into this submucosal space. After creating a tunnel, the circular muscle fibers of the distal esophagus were then divided by needle knife and the mucosotomy closed with clips. There were no complications before sacrifice at 5 days. Manometry performed on postprocedural day 5 showed that average LES pressures had fallen from 16.4 mm Hg to 6.7 mm Hg.

The feasibility of applying this tunneling technique to the pylorus was first described in animals by Kawai and colleagues in 2012. Four pigs underwent submucosal tunneling followed by endoscopic pylomyotomy. The median resting pressure of the pyloric sphincter was reduced by 63% immediately following the procedure, and a 50% reduction was maintained at 2-week reassessment (just before sacrifice). Chaves and colleagues, in another animal study, did not measure sphincter resting pressure, but rather evaluated sphincter division post-mortem. They report similar results in their 6 pigs, which underwent a nearly identical procedure.