Gastric peroral endoscopic myotomy is a minimally invasive endoscopic option for patients with refractory gastroparesis, but clinical response rates remain variable and factors to predict success have not been clearly elucidated. We assessed the outcomes of G-POEM for patients with refractory gastroparesis in our cohort, and the utility of intra-operative endoluminal functional luminal imaging probe (EndoFLIP). 

Retrospective, single tertiary centre, case-control study assessing the outcomes of G-POEM for refractory gastroparesis. Outcome measures include clinical response (decrease of 1 point in the total Gastroparesis cardinal Symptom Index (GCSI)) and complications. Impedance planimetry (EndoFLIP®, EF-325N, Medtronic Ltd) readings were obtained before and after myotomy, standardised to 50ml balloon inflation.

104 patients underwent G-POEM for refractory gastroparesis (n= 43 with type 1 diabetes). Technical success was 100%. 53/104 patients were able to be discharged on the same day (SDD), now adopted as standard of care protocol. 12/104 (11.5%) patients had hospital readmission within 30 days, of whom 5 required repeat endoscopy for further clipping. One patient required laparotomy for delayed leak at another institution. The median GCSI score reduced from 3.91 to 2.44 (p<0.001) at 12 months with a clinical success rate of 60%. Clinical success at 1 year was not different in patients with previous response to botulinum toxin treatment (p=0.773), any pre-procedural therapy (p=0.607), nasojejunal tube requirement prior to G-POEM (p=0.531) and a history of opioid use (25% vs. 54%; p=0.062).

There was no difference in outcome with use of lesser versus greater curvature approach (p=0.516) however procedural time (minutes) was significantly shorter with lesser curvature approach (35 (29-47.5) vs 60(38.8-60); p=0.047) and a trend towards less leaks (1 vs 5 were observed, p=0.053). There was also no difference in clinical outcome with supervised, trainee-completed procedures (n=46; p=0.301). Baseline GCSI, baseline gastric retention at 4 hours did not correlate with reduction in GCSI, however improvement in intra-operative EndoFLIP distensibility index from pre- to post myotomy had a weak correlation with change in GCSI from pre-POEM to >12 months post (r=-0.391; p=0.027).

Our results are consistent with literature confirming the need to establish pathways to improve clinical success rates post G-POEM for patients with refractory gastroparesis. Intra-operative EndoFLIP may be able to be utilised in addition to GCSI and gastric emptying however further prospective studies are required.

**AL and HA joint first author