BACKGROUND & AIMS: Gastric per oral endoscopic pyloromyotomy (GPOEM) is a promising treatment for gastroparesis. There are few data on the long-term outcomes of this procedure. We investigated long-term outcomes of GPOEM treatment of patients with refractory gastroparesis. METHODS: We conducted a retrospective case-series study of all patients who underwent GPOEM for refractory gastroparesis at a single center (n = 97), from June 2015 through March 2019; 90 patients had more than 3 months follow-up data and were included in our final analysis. We collected data on gastroparesis cardinal symptom index (GCSI) scores (measurements of postprandial fullness or early satiety, nausea and vomiting, and bloating) and SF-36 questionnaire scores (measures quality of life). The primary outcome was clinical response to GPOEM, defined as a decrease of at least 1 point in the average total GCSI score with more than a 25% decrease in at least 2 subscales of cardinal symptoms. Recurrence was defined as a return to baseline GCSI or GCSI scores of 3 or more for at least 2 months after an initial complete response. The secondary outcome was the factors that predict GPOEM failure (no response or gastroparesis recurrence within 6 months). RESULTS: At initial follow-up (3 to 6 months after GPOEM), 73 patients (81.1%) had a clinical response and significant increases in SF-36 questionnaire scores (indicating increased quality of life) whereas 17 patients (18.9%) had no response. Six months after GPOEM, 7.1% had recurrence. At 12 months, 8.3% of patients remaining in the study had recurrence. At 24 months, 4.8% of patients remaining in the study had a recurrence. At 36 months, 14.3% of patients remaining in the study had recurrence. For patients who experienced an initial clinical response, the rate of loss of that response per year was 12.9%. In the univariate and multivariate regression analysis, a longer duration of gastroparesis reduced the odds of response to GPOEM (odds ratio [OR], 0.092; 95% CI, 1.04-1.3; P = .001). On multivariate logistic regression, patients with high BMIs had increased odds of GPOEM failure (OR, 1.097; 95% CI, 1.022-1.176; P = .010) and patients receiving psychiatric medications had a higher risk of GPOEM failure (OR, 1.33; 95% CI, 0.110-1.008; P = .052). CONCLUSIONS: In retrospective analysis of 90 patients who underwent GPOEM for refractory gastroparesis, 81.1% had a clinical response at initial follow-up of their procedure. 1 year after GPOEM, 69.1% of all patients had a clinical response and 85.2% of initial responders maintained a clinical response. Patients maintained a clinical response and improved quality of life for as long as 3 years after the procedure. High BMI and long duration gastroparesis were associated with failure of GPOEM.