Author response: Endoscopic and surgical treatment options for gastroparesis: systematic review and network metaanalysis

Correspondence to: Daniel Eckhardt (email: daniel.eckhardt@klinikum-nuernberg.de)

Department of General, Visceral, and Thoracic Surgery

Klinikum Nürnberg

Paracelcus Medical University

Nuremberg

Germany

BJS, https://doi.org/10.1093/bjs/znaf183, published 29 September 2025

Thank you for giving us the opportunity to respond to the comments by Storr et al. We value their arguments and would like to clarify several points about the focus and methodology of our publication on endoscopic and surgical treatment options for gastroparesis.

We agree with Storr et al. that distinguishing between gastroparesis and functional dyspepsia remains challenging. To minimise misclassification, we applied strict inclusion criteria, requiring objective evidence of delayed gastric emptying1,2 and the use of validated, gastroparesis-specific symptom scores, such as the Total Symptom Score (TSS)3 and the Gastroparesis Cardinal Symptom Index (GCSI)4. These measures substantially reduced the likelihood of including studies that primarily enrolled patients with functional dyspepsia, despite diagnostic overlap being a well-known general limitation in the field.

Regarding the second point, Storr et al. noted that, although our initial aim included nutritional and pharmacological therapies, the final network focused on interventional treatments. While we agree that dietary and medical therapies are the primary treatments recommended in current guidelines2, most studies evaluating these approaches did not meet our predefined methodological criteria. This was particularly evident in the absence of standardised, gastroparesis-specific symptom scores and objectified delayed gastric emptying, unlike meta-analyses by other authors5. Consequently, they could not be incorporated into a coherent network. As stated in the title of our manuscript, the focus is on endoscopic and surgical treatment options, and our conclusions apply exclusively to these interventional approaches.

Finally, Storr et al. correctly highlight the heterogeneity of outcome measures and the prevalence of non-randomised designs. We assessed the risk of bias using the ROBINS-I tool and acknowledged the limitations relating to confounding and selection bias. Although TSS and GCSI differ, relative changes from baseline enabled reasonable comparisons to be made across studies. We agree entirely that future randomised, prospective trials using uniform diagnostic criteria and standardised symptom metrics are required to enable more definitive comparisons across all therapeutic domains.