Gastric peroral endoscopic pyloromyotomy versus gastric electrical stimulation in the treatment of refractory gastroparesis: methodologic issues

We read with interest the recent study by Shen et al. [1] and wish to congratulate the authors for their innovative work. The single-center retrospective study provided the first evidence of superiority of gastric peroral endoscopic pyloromyotomy (G-POEM) versus gastric electrical stimulation (GES) in the management of refractory gastroparesis. However, we have some concerns on the methodology that require clarification by the authors.

First, the implementation and reporting of the propensity score analysis is not standardized, which may introduce bias and misinterpretation. Propensity score methods have gained increasing popularity in the field of endoscopy, and a checklist and guideline for propensity score analysis have been proposed [2] [3]. Generally, logistic regression modeling is used to estimate the propensity score, yet this is not mentioned in this article. Even though a rule of thumb with 6 – 10 events per variable (EPV) is roughly accepted [2], the EPV in this study of 3.8 (7/27) is so small that it may lead to a biased propensity score and impaired power. The absolute caliper width of 0.1 appears to be loose and whether the sampling was with or without replacement is not described. Furthermore, it would have been much better if the characteristics before and after propensity score matching had been provided in their table 1.

Second, additional analyses in the study gild the lily. The main aim was to compare the clinical outcomes between G-POEM and GES. To further investigate the predictors of recurrence and the associations between etiology and recurrence in each group among 46 patients after matching is not advisable.

Third, the reasons for excluding patients with incomplete data need description. Six patients in each group represents a large proportion and whether the missingness is independent of the outcomes may affect results. Additionally, McNemar’s test is usually used on paired dichotomous data and may be not suitable for these data after matching. To avoid misunderstanding, analysis methods for the longitudinal data in figure 5 should be clarified and any interpretation of results needs to be done carefully.

Publication History

Article published online:
28 July 2020

• References

• 1 Shen S, Luo H, Vachaparambil C. et al. Gastric peroral endoscopic pyloromyotomy versus gastric electrical stimulation in the treatment of refractory gastroparesis: a propensity score-matched analysis of long term outcomes. Endoscopy 2020; 52: 349-358
• 2 Yang JY, Webster-Clark M, Lund JL. et al. Propensity score methods to control for confounding in observational cohort studies: a statistical primer and application to endoscopy research. Gastrointest Endosc 2019; 90: 360-369
• 3 Yao XI, Wang X, Speicher PJ. et al. Reporting and guidelines in propensity score analysis: a systematic review of cancer and cancer surgical studies. J Natl Cancer Inst 2017; DOI: 10.1093/jnci/djw323.