Peroral endoscopic myotomy (POEM) is a safe and effective treatment for achalasia; however, a notable proportion of patients experience therapeutic failure. Urakami et al. and Abe et al. developed pre-procedure scores to predict POEM failure at 12 and 6 months, respectively. The aim of this study was to evaluate the applicability of these scores in our patient population.

We conducted a retrospective cohort study with prospectively collected data, including patients who underwent POEM for achalasia between January 2017 and February 2023, with at least 12 months of follow-up. Therapeutic failure was defined as an Eckardt score >3 or the need for additional intervention. After calculating and stratifying the scores, their discriminative ability, sensitivity, and specificity were assessed.

Eighty-three patients (mean age 55 years, range 17–79; 56% female) underwent POEM, predominantly with type 2 achalasia (n=50). Thirty-seven percent had undergone prior treatment. Initial clinical response was achieved in 98% of patients (n=81). During long-term follow-up (median 49 months), 11 patients developed symptomatic recurrence, occurring at a median of 24 months after the procedure (range: 6–48 months). Aside from the pre-procedure Eckardt score (p = 0.04), no statistically significant association was observed between age, sex, manometric diagnosis, esophageal transit characteristics or prior treatment and therapeutic failure. The Urakami score demonstrated an AUC of 0.725, stratifying patients into low (n=41), intermediate (n=35), and high-risk (n=8) groups, with a sensitivity of 7.7% and specificity of 90%. The Abe score showed an AUC of 0.737, stratifying patients into low (n=70) and high-risk (n=13) groups, with a sensitivity of 30.8% and specificity of 87%.

Both scores demonstrated acceptable discriminative ability (AUC >0.7), indicating they can statistically distinguish patients at higher versus lower risk of POEM failure. The Urakami score had very low sensitivity but high specificity (90%), meaning it reliably identifies patients who are unlikely to fail POEM and may help avoid unnecessary interventions in low-risk individuals. In contrast, the Abe score showed higher sensitivity (30.8%) but lower specificity (87%), reflecting a trade-off between detecting at-risk patients and overestimating risk. These findings highlight that while pre-procedure scores can aid risk stratification, their clinical application should be complemented by careful patient evaluation and follow-up, and further refinement of predictive tools is needed.