Safety and feasibility of same-day discharge for per-oral endoscopic myotomy

Per-oral endoscopic myotomy (POEM) was pioneered by Inoue et al.¹ in 2010 for the treatment of achalasia. Since its inception, it has rapidly progressed to become the standard of care and an effective, minimally invasive alternative to laparoscopic Heller myotomy (LHM).² The procedure involves the creation of an oesophageal submucosal tunnel followed by dissection of the circular muscle bundle extending past the gastro-oesophageal junction (GOJ). This releases the lower oesophageal sphincter, specifically addressing the underlying pathophysiology of failure to relax the GOJ in achalasia.^1,3

Current literature shows high efficacy of POEM for achalasia, with a meta-analysis by Schlottmann et al.⁴ showing improved dysphagia in 93.2% of patients compared to 87.7% following LHM. However, in this analysis, the average length of stay (LOS) was 1 day longer in the POEM group (4.26 vs 3.23 days). Ofosu et al.⁵ studied pneumatic dilation (PD) and showed a clinical success rate of 76.9% at 12 months but a significantly shorter mean LOS of 1.9 days. Regardless of intervention, hospital LOS primarily relates to the management of pain and complications, establishment of oral intake and protocolised requirements for imaging. Inoue’s original post-POEM protocol involved computed tomography (CT) on the day of the procedure, followed by barium oesophagram on day 1, before dietary initiation.¹ In many centres, protocolised imaging remains the incumbent approach, leading to delayed discharge and increased costs.²

In direct contrast, recent studies show that routine oesophagram and hospital admission may be unnecessary. Cloutier et al. demonstrated successful same-day discharge (SDD) in 79.1% of patients following POEM in Canada,⁶and 62.4% by Benias et al.⁷in the United States of America (USA).This has significant implications for overburdened healthcare systems, and yet SDD has not evolved into mainstream practice.^2,8,9 Currently, there are no Australasian data regarding the safety and feasibility of SDD for POEM. Accordingly, we evaluated this across two centres in New Zealand.

Aim

The primary outcome of interest was success rates of planned SDD after POEM.

Complications and re-admission rates were reviewed to determine safety.

Methods

Participants

Consecutive patients referred for POEMs were all prospectively planned for SDD with follow-up in the community. Data were retrospectively reviewed, collected and analysed.

To be included, patients had to have a confirmed diagnosis (time barium swallow or preferably oesophageal manometry) and be fit to receive a general anaesthetic (initially physician assessed with subsequent anaesthetic assessment) to proceed to POEM. Some patients elected to have endoscopic dilatations. Prior treatment was allowed.

Design

All procedures were performed at two tertiary referral centres in New Zealand between 2020 and 2023. Referrals were received from various districts within New Zealand. One interventional endoscopist at each unit performed all procedures. All cases were performed under general anaesthetic with administration of peri-procedural antibiotics.

No routine oesophagrams, imaging or endoscopy were performed without endoscopist-determined clinical indication. Regular analgesia, including opiates, was available in post-anaesthetic recovery.

Post-procedurally, oral intake was established at 1 hour starting with clear oral fluids. If tolerated, this was followed by free oral fluids including nutritional supplements for 5 days. Solid food was re-introduced at 5 days post-procedure as tolerated.

Discharge criteria were 1) no concerning procedural complications as determined by the endoscopist, 2) minimal or no pain with or without oral analgesics, and 3) establishment of oral liquid intake. A prescription for a short duration of oral liquid opiates was supplied on discharge as required.

Procedure

Patient data including age, gender, ethnicity, manometry findings and Eckardt score were collected on Microsoft Excel. The Eckardt score is an internationally validated symptom assessment score¹⁰ and was collected to determine pre-procedure symptomatology. If information was missing, patients were individually contacted to calculate the Eckardt score.

Procedural findings of the length of myotomy, duration of procedure and use of antibiotics were reviewed. Quantitative statistical analysis was performed using Microsoft Excel.

Ethics

A retrospective analysis of the procedures was performed. Institutional ethics waiver was obtained and out of scope review obtained by Health and Disability Ethics Committees (review reference 2023 OOS 17904).

Results

Patient characteristics

Fifty consecutive POEM procedures were successfully completed in our study. All were performed for achalasia. Patients were referred from diverse regions, across the North and South Islands of New Zealand.

Patient characteristics are documented in Table 1. The mean age was 50.7 years, with 52% males. NZ European and Māori were the most common ethnicities. Type 2 achalasia was the most frequent manometric subtype, accounting for 66% of POEM indications. The mean Eckardt score prior to POEM was 6.86 (SD 2.75). Duration of symptoms varied from less than 1 year to greater than 10 years. Almost half (48%) of the cohort had previous interventions, with the majority being LHM (22%), followed by PD (16%), botulinum toxin (BT) injection (8%) and prior POEM (2%).

The median duration of follow-up was 19 months (interquartile range 9–29.75 months).

Procedural details and adverse events

Procedural details are listed in Table 2. The mean procedure duration was 36.7 minutes (SD 15.3) and the average myotomy length was 9.08cm (SD 2.88). Opiates were required in 16 (32%) patients in the post-procedural phase, and 10 (20%) patients received scripts on discharge.

View Table 1–2, Figure 1.

Adverse events were seen in a total of three patients, with two mucosal tears and one contained collection. Of the patients with mucosal tears, one was identified intra-procedurally and clipped without complications and discharged the same day. The other had pain post-procedurally with evidence of subcutaneous emphysema and therefore had a re-look gastroscopy with endoscopic clipping. There were no further complications from the tear. There were no deaths.

Outcomes

Discharge

A total of 41/50 (82%) patients were cleared for SDD (Figure 1). Of those, SDD was successfully achieved in 35 (70%) patients. Six patients (12%) were admitted due to social reasons alone (one due to lack of transport home, one due to breastfeeding issues and four due to miscommunication with the wards).

Admissions

Nine patients (18%) required hospital admission. The mean LOS for the whole population group, including those discharged the same day, was 0.36 (SD 0.63), or 1.2 days (SD 0.56) for those requiring admission.

Pain

Six (12%) were admitted for pain management; five of these were discharged within 24 hours. One patient in this group was discharged on day 3. This patient had a repeat gastroscopy to investigate the pain, which was unremarkable. All six patients admitted with pain were female and three were managed with titration of oral analgesia.

Dietary progression

Two patients (4%) were admitted due to failure of dietary progression and were both discharged the next day.

Re-look gastroscopy

One (2%) patient was admitted for a re-look gastroscopy due to a mucosal injury identified during the index procedure. This patient had an oesophageal ulcer that was clipped and was able to be discharged on day 3.

Re-admissions

There were two re-admissions, yielding a re-admission rate of 4%. One of these patients (presented 7 days post-procedure) required CT imaging due to pain, fevers and raised inflammatory markers. This showed a closed collection. This patient was discharged with oral antibiotics after 24 hours of observation. The other presented to the emergency department also 7 days post-procedure with chest pain. This patient had normal laboratory investigations and chest radiograph and discharged the same day without need for CT imaging or endoscopy.

No patients clinically required imaging or oesophagram in the immediate post-procedural phase, and two patients, as mentioned above, required gastroscopy.

Discussion

We have demonstrated that the majority of patients planned for SDD after POEM were able to successfully achieve this with our modified protocol of early feeding and regular, proactive analgesia. This was safe despite the lack of post-procedural imaging, with a low documented rate of immediate complications. The re-admission rate was also low, indicating that delayed complications were uncommon and therapeutic benefits were durable.

Our findings align with international data, albeit with greater SDD success in our cohort, possibly related to practice improvements over time.^6–8 Overall, this adds to accumulating evidence that SDD after POEM is possible and likely to become more commonplace with time.

This is relevant as it is still common practice to routinely admit patients post-POEM. For example, in 2021 Gupta and Sidhu et al.⁹ described three tertiary referral centres in Australia wherein all patients were admitted overnight and only discharged once pain-free. The mean admission length was 2.6 days without adverse events and 9.8 days with adverse events. Furthermore, a Canadian study describes routine endoscopy and oesophagram in all patients and a median stay of 1-day post-procedure.¹¹ Indeed, a meta-analysis of 2,276 patients in 2019 by Tan et al. showed a length of stay between 1 to 6.2 days,¹² and another meta-analysis conducted by Schlottman et al.⁴ in 2018 showed a mean stay of 4.26 days in 1,958 patients.

Notably, in 2021 Fujiyoshi et al.¹³ published a retrospective analysis of 497 patients who received routine second-look endoscopy 1 day following index procedure. They reported 12 (2.4%) patients in which the second-look endoscopy changed practice, either by intervention during endoscopy or more prolonged fasting and advocated for admission with second-look endoscopy. However, whether this impacted long-term outcome is not reported and is unclear. During our median follow-up of 19 months in our study, only one patient presented with a closed collection, and this was safely managed conservatively. International studies reflect similarly reassuring results. Cloutier et al. reported on 90-day follow-up outcomes with 10 (11% of patients) re-presentations to hospital. None, however, had complications in relation to collections or mediastinitis. Three did have upper gastrointestinal bleeding; however, only one had endoscopic findings of an ulcer, and the others did not have endoscopically relevant complications.⁶ Zhang and her colleagues furthermore reported no presentations to hospital during a 21-day median follow-up period on 17 patients total.⁸ Indeed, in the initial paper presented by Benias et al.⁷ in 2019, no patients who underwent SDD had serious adverse events during an average follow-up period of 10.1 months. Two patients were thought to have collections, although investigations revealed that only one truly did. Nonetheless, these patients were part of the group that did not achieve same-day discharge criteria, and importantly the complications did not affect long-term outcomes. This suggests that re-look endoscopy need not be mandated following per-oral endoscopic myotomy.

Our data challenge the need for such long admissions and support both the safety and feasibility of SDD. None of the 41 patients cleared for SDD underwent routine endoscopy or imaging and only one required re-admission due to a closed collection. This patient was able to be managed conservatively with outpatient antibiotics.

Furthermore, prior studies investigating SDD necessitated the absence of an opiate requirement for SDD, only allowing paracetamol,^6,7 or paracetamol and tramadol.⁸ In contrast, we allowed patients to have oral opiates peri-procedurally, and so long as pain was adequately controlled with oral medications, patients were cleared for discharge. Importantly, this did not result in any significant increase in adverse events or re-admissions, and we believe this allowed for increased rates of SDD.

Potential improvements could be made to encourage SDD within our own cohort. Firstly, four patients were admitted because of miscommunication regarding discharge with the wards, an issue that can be readily rectified. Additionally, three of the six patients admitted for pain only required oral analgesics. These patients were discharged within 24 hours of the procedure, suggesting that with more robust reviewing and discharge planning, SDD could have been facilitated.

Protocols necessitating admission, endoscopy and imaging lack strong evidence, and SDD is increasingly recognised. Our data represent 50 consecutive POEM procedures with a high rate of safe SDD, and although it is limited to only two centres in New Zealand, it provides a good sample size of demonstrated effectiveness.

Conclusion

POEM is an effective and safe procedure for achalasia, but protocols have required admission and routine imaging. This is the first study across Australasia reporting high success rates of planned SDD following POEM without routine post-procedural imaging. Adding to international literature around SDD, our data furthermore suggest that oral opiate requirements need not be criteria for admission.

Re-admission rate was low and adverse outcomes were considered mild and managed conservatively. Further data from other institutions will cement this practice change, resulting in cost-saving health systems without sacrificing safety.