Differences between the protocol and the present update:

• Data collection and analysis: we used Review Manager 5.4 and RevMan Web for data collection and analysis, updated from RevMan 5.1.

• Selection of studies: we added reprints of articles to the reference lists of included studies but did not consider them separately if they contained no new data. In the 2014 review, we discarded article reprints. We listed studies that were available only in abstract form, or were only identified in the ISRCTN register, as studies 'awaiting classification'.

• Outcomes: we redesignated the outcome of 'pH/impedance studies' to 'pH/impedance indices' to account for the range of pH measurements described in the available literature.

• Data extraction and management: three review authors (MT, IL, EA) independently extracted study data using a robust data extraction form and checked and entered the data into RevMan 5.4/RevMan Web, with MT, EA, and IL analysing the data and highlighting any discrepancies. In the 2014 review, two review authors extracted and entered study data into RevMan 5.1.

• Measures of treatment effect: we extracted continuous data (e.g. reflux index) for summary data: we used means and standard deviations to derive a standardised mean difference (SMD) with a 95% confidence interval using a fixed‐effect model. The latest NASPGHAN/ESPGHAN guidelines do not define normal values for pH‐metry and pH‐impedance (NASPGHAN‐ESPGHAN guidelines 2018). The values of reflux index mentioned in the 2014 review (> 10% in 24 hours in infants and > 4% in 24 hours in children > 12 months) have been modified here with a judgement regarding improvement/non‐improvement. Dichotomous data, such as improvement/non‐improvement in endoscopic appearance, produced outcome data we presented as risk ratios. In the 2014 review, we used reported data rather than extracting summary data.

• Unit of analysis issues: we considered issues related to multiple observations for the same outcome (e.g. repeated pH‐impedance measurements), and consulted the Cochrane Gut group if clarification was required. If we included multi‐arm studies, we would analyse multiple intervention groups to prevent arbitrary omission of relevant groups or double‐counting of participants. In the 2014 review, there was some overlap in reported data (e.g. according to age criteria); this was corrected in this review.

• Dealing with missing data: we contacted trial authors or sponsors of studies published from 2014 to 2022 to provide missing data, or to seek clarification when we were uncertain about the specifics of a trial pertinent to analysis. In the 2014 review, we contacted study authors of studies published within the previous 10 years (e.g. to 2004).

• Data synthesis: were unable to combine studies meaningfully due to the heterogeneity of studies in terms of outcomes, comparisons, and populations. For continuous measurements, we had planned to use weighted mean differences to pool results from studies using a common measurement scale. Where studies used different measurement scales, we planned to pool standardised mean differences. Instead, we have presented difference in means and 95% confidence intervals for individual studies and summary effects, using the following order: Population > Comparison > Outcome, following updated guidance. Given the individual study differences and heterogeneity in study design, we provided guidance based on individual treatments to give better focus for decision‐makers. This differs from the 2014 review.

• Sensitivity analysis: if meta‐analysis had been possible, we intended to undertake sensitivity analysis using RevMan Web, to ascertain whether any decisions regarding thresholds influenced result reporting (e.g. choosing age thresholds at 12 months influencing meta‐analytic robustness). We planned to integrate the findings into the results and conclusions. This was not considered in the 2014 review. However, a meta‐analysis was not possible and sensitivity analysis not required.

• Summary of findings and assessment of the certainty of the evidence: working independently, two authors used the five GRADE considerations (risk of bias, consistency of effect, imprecision, indirectness, and publication bias) to assess the certainty of the body of evidence for each outcome, and to draw conclusions about evidence certainty within the text of the review. We resolved disagreements through discussion, and involved all review authors involved if the initial two authors could not reach agreement. All authors then reviewed the GRADE considerations in assessing the certainty of evidence and integrated this into the summary of findings tables. The summary of findings tables distinguish results by age (infants, and children aged one to 16 years), then comparison, and the evidence is presented by outcome (symptoms, adverse events, pH‐impedance indices, and endoscopic findings), with clear rationales given where evidence was downgraded or upgraded according to GRADE criteria, including if the risk of bias was so great the evidence needed downgrading by two steps.

• Literature search in this update version: we did not search the Cochrane Review Group Specialised Register as it was not updated since the 2014 version and the included RCTs are included in Cochrane CENTRAL, which we did search. We did not search the Centralised Information Service for Complementary Medicine (CISCOM). This database did not yield additional eligible studies for our review in the 2014 version, and it was not available to us for this update. In the 2014 version, we handsearched published abstracts from conference proceedings. For this update, we did not handsearch proceedings from conferences that took place after 2014 because Embase now includes proceedings from these conferences (2000 onwards); these abstracts were searched electronically through our main electronic search. In the 2014 version, we searched the clinical trials register mRCT. In this updated version, we searched the World Health Organization's clinical trials register (ICTPR) and ClinicalTrials.gov (https://clinicaltrials.gov/). We also revisited the search strategies, and added some new terms to reflect the current practice of treatment in the updated search.

• We have used the terms GOR and GORD throughout the review, following NASPGHAN‐ESPGHAN guidelines 2018 and NICE 2019 definitions, and we acknowledge that different groups may have used different definitions for these terms in their studies. We have included some narrative where relevant and encourage readers to review the original articles if they wish to ascertain in more detail how other authors distinguish between GOR and GORD.