Patients of either sex, older than 18 years of age, with a suspected diagnosis of esophageal achalasia based on upper endoscopy, barium esophagogram, and clinical evaluation findings and confirmed through high-resolution manometry were included in the study. Patients that presented with any pathology in which the procedure would be contraindicated, such as portal hypertension, premalignant conditions in the esophagus, some type of epiphrenic diverticulum, pregnant patients, or patients that rejected the procedure were excluded.

Prior to the procedure, all patients were evaluated through upper endoscopy, timed barium esophagogram, high-resolution manometry, and thoracoabdominal tomography. In addition, each patient filled out the Eckardt clinical questionnaire and a quality of life questionnaire (AE-18). In all cases, Chagas disease was ruled out through a blood test.

• •

  Patients were admitted 24h before the procedure.

• •

  Antimicrobial prophylaxis was used 24h before the procedure with 1g cefotaxime IV or 1g ciprofloxacin IV in patients allergic to penicillin or cephalosporin, both as single doses.

• •

  Fasting 24h prior to the procedure. Patients with megaesophagus or severe dilation received a liquid diet for 72h and fasted 24h before the procedure.

• •

  If a subsequent adverse event presented, the decision was to continue antibiotic treatment for at least 5 days, either IV or oral.

The POEM procedure was performed according to that described by Inoue et al.,10 with the following accessories: a conventional endoscope model EG590WR (diameter 9.8mm, channel 2.8mm); a hood model DH-28GR (Fujinon, Tokyo, Japan); an ERBE VIO 200D electrosurgical unit with an I-type Hybrid Knife (Tübingen, Germany) with the following parameters: injection (ERBEJET, effect 50), incision (ENDOCUT Q, effect 3, cutting duration 3, and cutting interval 3), tunnelization (SWIFT COAG, effect 3 to 70 w), myotomy (ENDOCUT Q, effect 2, cutting duration 3, cutting interval 3), hemostasis (SOFT COAG, effect 3 to 40 w); Resolution Clip hemoclips (Boston Scientific, USA) or cyanoacrylate (B-Braun, USA) were used for incision closure. An insufflator (ENDO STRATUS, Medivators; Minneapolis, Minnesota, USA) and carbon dioxide were used in all the procedures.

The internationally described technique we used for the POEM consisted of 5 steps (fig. 1):

• 1.

  Examination and injection: Through upper diagnostic endoscopy, we located the esophagogastric junction (EGJ) and documented the distance from the incisors to the EGJ. The site for performing the anterior POEM was located (at the 2 o’clock position). A posterior POEM (6 to 8 o’clock position) was carried out on the previously-treated patients. A mixture of normal saline solution and methylene blue at 0.5% was injected with the hybrid knife at 13-15cm proximal to the EGJ in all the cases, except those with type III achalasia, in whom the injection was applied at 20cm from the incisors to perform a longer myotomy.

• 2.

  Incision: A longitudinal incision, 10 to 15-mm long, was made parallel to the major axis of the esophagus and served as the entrance site to the submucosal tunnel.

• 3.

  Tunnel: An injection was applied to elevate the submucosal space, which was then dissected from the entrance up to 2-3cm below the EGJ.

• 4.

  Myotomy: It was begun 2cm under the incision, inside the submucosal tunnel as a partial myotomy (only cutting the internal circular layer of the esophagus) or a total myotomy (cutting both muscle layers [the last 20 cases]) and was continued up to the EGJ and extended 2cm below it. There were 3 methods for corroborating adequate myotomy: by measuring the location of the EGJ and 2cm below it, compared with the same measurement through the submucosal tunnel; through operator perception, feeling a decrease in resistance upon advancing the equipment to the level of the EGJ; and by performing a short retrovision in the stomach and observing the Z line or the opening of that region during moderate CO2 insufflation. Pediatric equipment is not available to us for performing the double-endoscope technique.

• 5.

  Closure: Six to 8 hemoclips were used to close the incision. In the cases in which they could not be placed, or in which there was an adverse event at the entrance site, such as dehiscence due to the detachment of one of the clips, 1ml of undiluted cyanoacrylate was applied. The application was exactly at the entrance site, initially preloading it in a 3-ml syringe and then pushing it inside the tunnel with 3ml of injectable water. Obturation was observed immediately and adequate closure was corroborated through the instillation of water and the absence of “bubbles” coming from the tunnel at that site.

Figure 1.

Procedural technique: A) Injection. B) Incision. C) Submucosal tunnel. D and E) Complete myotomy. F) Incision closure with cyanoacrylate.

(0.48MB).

After the procedure, a thoracoabdominal x-ray was taken to document pneumoperitoneum or pneumomediastinum. If either was found, and did not have any clinical repercussions, no maneuver was carried out. If there was a manifest clinical repercussion, a decompression puncture was made with a 14-F catheter. Patients fasted for 24h and then a water-soluble contrast esophagogram was performed to rule out leakage into the submucosal tunnel. If there was no contraindication, liquid diet was begun, and patients were released 48-72h after the procedure. During hospital stay, patients received analgesic management with nonsteroidal anti-inflammatory drugs and antiemetics. Antibiotic therapy was initially administered IV and then completed PO for 5-7 days. The services of general surgery, cardiothoracic surgery, and intensive care were available 24/7 for attending to any complications.

Follow-up was planned for one month after surgery, to be continued at postoperative months 3, 6, 12, 18, and 24, and then annually for 10 years. The following studies were carried out at each evaluation: upper endoscopy, high-resolution manometry, pH-study, a clinical evaluation of symptoms using the Eckardt scale, the AE-18 quality of life questionnaire, the gastroesophageal reflux disease questionnaire, and timed barium esophagogram, which was performed starting in May 2016 (n = 21).

Clinical success: when the patient achieved an Eckardt score ≤ 3.

Manometric success: when an IRP < 15mmHg was attained post-procedure.

Minor complications: non-life-threatening complications that presented during or after the procedure. Resolved medically, endoscopically, or spontaneously, they included: mucosotomy (communication between the lumen of the gastrointestinal tract and the submucosal tunnel), minor bleeding during the procedure that did not require packed red blood cell administration, subcutaneous emphysema, pneumoperitoneum or pneumomediastinum with or without puncture during or after the procedure, dysphagia, and mild abdominal pain.

Major complications: life-threatening complications, those that required surgery, or whose resolution was difficult through any modality. They included: hemodynamic instability during the procedure, the need for blood transfusion due to bleeding during or after the procedure (even when there was adequate hemostasis), admission to the intensive care unit for surveillance, prolonged hospitalization (> 5 days), pneumonia, conversion to open surgery, an invasive postoperative procedure (chest tube placement, thoracotomy, etc.), and when the patient required hospital readmission due to signs of submucosal tunnel leakage, bleeding, or signs suggestive of mediastinitis.

For sample size, we used the mean difference method, considering the pre-procedure and post-procedure delta in the IRP value. A mean pre-procedure IRP value was set at 26 ± 5.5mmHg as the reference and a post-procedure change of at least 16 ± 3.5mmHg was regarded as significant. Considering a type I error of 5% and a type II error of 20%, a total of 40 patients were needed.

The quantitative variables with normal distribution were expressed as means, using standard deviation as the measure of dispersion. The median and interquartile range was used for the quantitative variables with skewed distribution. Normality was evaluated through the mental methods test and the Kolmogorov-Smirnov test. Qualitative and ordinal variables were expressed as proportions. Different tests were employed for the bivariate analysis: the Student's t test for the quantitative variables with normal distribution; the Wilcoxon test for the related variables (before-after); and the χ2 test for the qualitative variables or the Fisher's exact test (in the case of having an expected value < 5). The ANOVA, Kruskal-Wallis, and Pearson's χ2 or Fisher's exact test were respectively used for the same group comparisons at different post-procedure times, depending on the variable. The statistical analysis was carried out with the SPSS version 22.0 (Chicago, Illinois, USA) program. A p < 0.05 was considered statistically significant, and the corresponding confidence intervals were 95%.

A total of 50 procedures were performed within the time frame of November 2014 and February 2017 at the Department of Gastrointestinal Endoscopy of the Hospital de Especialidades Centro Médico Nacional Siglo XXI. Thirty-one of the patients were women (62%) and 19 were men (38%). Mean patient age was 48.8 ± 14.1 years and the time from first symptom onset to final disease diagnosis was a median of 24 months (9-140). Our cohort included: 41 treatment-naïve patients (82%), 7 previously-treated patients (14%) (all with LHM), and 2 patients (4%) that partially responded to initial POEM and were candidates for a second POEM procedure (redo-POEM).

The median pre-procedure Eckardt symptom score values were 9 (1-12) and the post-procedure values were 2 (0-4). The pre-POEM IRP values had a median of 24.3mmHg (15.5-53.6) and at 3 months after the procedure, the median was 11.2mmHg (0-24). Eight patients had type I achalasia (16%), 28 had type II (56%) (the most frequent in our cohort), and 6 patients had type III (12%). In relation to motor disorders, 7 patients had changes due to previous myotomy (14%) and one patient had signs of distal esophageal spasm (DES) (2%). Median procedure duration was 80min (31-147) and mean myotomy length was 12.6 ± 2.75cm. Median hospital stay was 3 days (2-7) (Table 1).

After comparing the different types of patients included in the study, there were no statistically significant differences in relation to age, sex, the pre-POEM Eckardt scores, and the post-POEM Eckardt scores at 3 months, procedure duration, and post-POEM IRP values at 3 months. However, there were statistically significant differences among groups with respect to achalasia subtype or motor disorder; most of the treatment-naïve patients were type II (27 = 65.9%), followed by type I (8 = 19.5%), and finally type III (5 = 12.2%), as well as the case of DES (1 = 2.4%) included in the late group. The previously-treated cases were statistically different from those in the treatment-naïve group. The former had a total of only 7 cases. The previously-treated group had no type I cases, there was one (14.3%) type II case, one (14.3%) type III, and 5 cases could not be correctly classified through the Chicago classification and were catalogued as changes due to previous myotomy. The redo-POEM group consisted of only 2 cases (100%) and they had undergone a previous failed POEM. Other statistically significant differences between those two groups were the pre-POEM IRP levels. The treatment-naïve patients had a higher IRP value with a median of 28.7mmHg (15.7-53.6), compared with 18.7mmHg of the previously-treated patients and 19.9mmHg of the redo-POEM group. Thus, only the comparison between the naïve-treatment patients versus the previously-treated patients was statistically significant (p=0.007), but the treatment-naïve group versus the redo-POEM group was not, nor was the previously-treated group versus the redo-POEM group (p = NS). There was a statistically significant difference in relation to the direction of the myotomy. Obviously, it was anterior in all the treatment-naïve patients (41 = 100%) and posterior in the previously-treated patients (7 = 100%) and redo-POEM patients (2 = 100%). Finally, myotomy length was longer in the treatment-naïve patients and the previously-treated patients, with 12.8 ± 2.6cm and 13.0 ± 1.9cm, respectively, compared with 7.0 ± 1.4cm in the redo-POEM patients, demonstrating statistical differences between the redo-POEM group and the other 2 groups (treatment-naïve group vs redo-POEM group [p = 0.008] and previously-treated group vs redo-POEM group [p = 0.016]) (Table 2).

Several adverse events presented after the procedure: subcutaneous emphysema was the most frequent, presenting in 15/50 (27%) of the patients immediately after the procedure, followed by intraoperative bleeding in 13/50 (24%), with endoscopic control in all the cases. Pneumoperitoneum presented in 12/50 (22%) and pneumomediastinum in 5/50 (10%). Large-caliber needle puncture for decompression was necessary in only 5 of those patients, and it was performed post-procedure in all of them. Mucosotomy presented in 3 cases, requiring endoscopic treatment with cyanoacrylate. In one of the patients, it was evidenced through an upper gastrointestinal x-ray with water-soluble contrast medium and in another, it presented 48h after the procedure as a Mallory-Weiss tear. One patient presented with pneumonia, most likely due to aspiration during intubation, and it was medically resolved. One patient was admitted to the intensive care unit for 72h due to suspected pulmonary thromboembolism, which was then ruled out. Finally, one patient presented with cardiac arrhythmia during the study and was hospitalized and evaluated by the cardiology service. POEM completion was not contraindicated. There were no deaths associated with the procedure (fig. 2).

Figure 2.

Adverse events that presented in our patient cohort.

(0.07MB).

The median Eckardt score prior to the procedure was 9 and it decreased to a median of 2 after the procedure, resulting in clinical success at 3 months in all patients except 2. The median remained unchanged at post-POEM follow-up months 6, 12, 18, and 24. At 24 months only 22/50 (44%) patients continued in follow-up and were exclusively treatment-naïve patients, given that at the time of this writing, all of the previously-treated patients and redo-POEM patients had fewer than 18 months of follow-up. However, no differences in clinical response were found between the 3 treatment groups, and the Eckardt scores of the patient cohort remained unchanged at the short term and at the medium term. Statistically significant differences were observed only between the pre-procedure scores and the scores obtained after the procedure, with the exception of the post-procedure Eckardt scores at 24 months (fig. 3).

Figure 3.

Box graph showing the behavior of the Eckardt score throughout the follow-up period.

*p = 0.000 Kruskal-Wallis test; ** NS.

(0.07MB).

The IRP decreased from a pre-procedure median of 24.3mmHg (15.5-53.6) to 11.2mmHg (0-24) at 3 months after the procedure. The decrease in the IRP remained unchanged up to month 24 of the follow-up, where a slight increase in the median to 13.5mmHg was observed. Prior to that value, no statistically significant differences between post-procedure IRP values were found. At present, all the evaluated patients have been treatment-naïve patients, given that no previously-treated patients or redo-POEM patients has yet reached follow-up month 24 (fig. 4).

Figure 4.

Box graph showing the behavior of the IRP throughout the follow-up period.

*p = 0.000 Kruskal-Wallis test; ** NS.

(0.05MB).

All the patients were treated with a standard dose of a proton pump inhibitor (PPI) for 30 days, after which the patients had no medical treatment and were evaluated at 3, 6, 12, 18, and 24 months after the procedure. At post-procedure month 3, 35/50 (70%) patients presented with a positive pH-study (> 14.73), 8/50 (16%) had clinical reflux, and 15/50 (30%) had esophagitis (grades A and B in 80% of the cases). We then decided to begin a double dose of PPI for 4 weeks, and at the evaluation at 6 months, of the 43/50 (86%) patients that reached that point of the follow-up, 65% did not present with esophagitis, but 35% did; 12.5% of those patients had grade A esophagitis, 20% had grade B, and 2.5% had grade C, according to the Los Angeles classification. The pH-study was positive in 47.5% of the patients and 15% had clinical reflux. The decision was made to give those patients a standard PPI dose for an indefinite period of time, and at the evaluation at month 24, only 5% of the patients still had a positive endoscopy (grade A or B Los Angeles classification erosive esophagitis), 10% had clinically evident gastroesophageal reflux (heartburn), and 10% had a positive pH-study. Thus, there was no correlation between the instrumented pH measurement and the clinical method for evaluating that parameter, but it could be medically controlled in all cases through PPI dose readjustments.

The authors declare that no experiments were performed on humans or animals for this study.

The authors declare that no patient data appear in this article.

The authors declare that no patient data appear in this article.