We thank the authors1 for their thoughtful comments, on our systematic review and meta-analysis comparing percutaneous endoscopic gastrostomy with laparoscopic techniques in children.2 We acknowledge that most included studies were non-randomized and retrospective, a limitation clearly stated in the manuscript and common in pediatric surgical research due to ethical and practical constraints. Certainty of evidence for each outcome was assessed using the GRADE (Grading of Recommendations Assessment, Development, and Evaluation) approach with GRADEpro GDT,3 recommended by the Cochrane Collaboration. In line with GRADE guidance, certainty assessment involves a balanced judgment of factors that may both decrease and increase confidence in effect estimates, rather than a purely additive process. Although non-randomized evidence initially starts at low certainty, upgrading was applied selectively when criteria such as large effect sizes and consistent findings across studies outweighed concerns related to heterogeneity or imprecision.4, 5 Regarding risk of bias assessment, the moderate overall risk assigned to all studies mainly reflects unavoidable confounding by indication, balanced by strengths such as objective outcomes and, in several cohorts, complete follow-up, in accordance with the ROBINS-I (Risk Of Bias In Non-randomized Studies—of Interventions) guidance.6 Although heterogeneity was substantial for some outcomes, this was expected given clinical diversity and institutional practices. Random-effects models and sensitivity analyses confirmed the robustness of findings. Finally, our conclusions were intended as a synthesis of current evidence rather than prescriptive recommendations, suggesting fewer major complications and reinterventions with laparoscopic approaches. We agree that high-quality prospective studies are needed and have highlighted this as a priority for future research. We believe this exchange contributes to improving evidence-based decision-making in pediatric gastrostomy placement. The authors declare no conflicts of interest.
Cassaro et al. (Mon,) studied this question.