Effectiveness of Metformin in Preventing Type 2 Diabetes in Children and Adolescents with Overweight or Obesity: A Protocol for a Systematic Review and Meta-Analysis
and Marina Ybarra 5,*,†Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsThe systematic review protocol outlined in Materials and Methods demonstrates a structured approach aligned with established guidelines (PRISMA-P 2015) and registered in PROSPERO, which enhances transparency and methodological rigour. However, there are several aspects to consider regarding its robustness:
Potential Limitations and Areas Impacting Robustness:
Search Strategy Details: While the broad database coverage is a strength, the protocol states that the complete search strategy will be available in Supplementary File S1 and PROSPERO. The robustness depends heavily on the specificity and sensitivity of these strategies, which are not detailed here. Without access to the actual search strings, it's difficult to assess whether the search terms are comprehensive enough to minimize missed studies.
Inclusion of Study Designs: Including quasi-experimental and prospective cohort studies alongside RCTs enhances comprehensiveness but may introduce heterogeneity and bias, thereby affecting the internal validity of the review. Clear plans for addressing potential heterogeneity or bias (e.g., risk-of-bias assessments, subgroup analyses) are not specified.
Exclusion of Retrospective and Cross-Sectional Studies: While justified, this exclusion limits the evidence base to prospective designs, which are generally more robust but may reduce the total number of eligible studies, especially in pediatric populations where RCTs are less common.
Outcome Definitions: The primary outcome is defined according to ISPAD and ADA guidelines, which promote standardisation; however, variability in how studies report incidence and diagnostic criteria may introduce heterogeneity. The protocol does not mention plans to address potential inconsistencies.
Study Quality and Bias Assessment: The protocol does not specify how the quality or risk of bias of included studies will be assessed, which is critical for evaluating the robustness and validity of synthesised evidence.
Handling of Heterogeneity: There is no mention of plans to assess heterogeneity (e.g., via statistical tests or the I² statistic) or to perform meta-analyses, both of which are important for evaluating the robustness of pooled estimates.
Potential for Bias in Selection and Data Extraction: Although dual review is planned, details on reviewer calibration or training and on procedures for resolving persistent disagreements are not provided.
Author Response
London, December 30th, 2025
Response to Reviewers
We sincerely thank the reviewers for their careful evaluation of our protocol and for their constructive and insightful comments. Below we provide a point-by-point response and describe the corresponding revisions made to the manuscript. In addition, the revised manuscript has been carefully proofread by two native English speakers to improve clarity and language quality.
Reviewer 1
Comment 1: “Potential Limitations and Areas Impacting Robustness: Search Strategy Details: While the broad database coverage is a strength, the protocol states that the complete search strategy will be available in Supplementary File S1 and PROSPERO. The robustness depends heavily on the specificity and sensitivity of these strategies, which are not detailed here. Without access to the actual search strings, it's difficult to assess whether the search terms are comprehensive enough to minimize missed studies.”
We thank the reviewer for this comment. The search strategy is already described in the main manuscript (Section 2.2), where we outline the databases searched, the core search concepts, and the use of both MeSH terms and free-text keywords. In addition, the full database-specific search strings, including Boolean operators and controlled vocabulary, were provided with the original submission as Supplementary File S1 and are publicly available in the PROSPERO registration and can promptly be accessed. Therefore, no further modification of the search strategy or manuscript text was required.
Comment 2: “Inclusion of Study Designs: Including quasi-experimental and prospective cohort studies alongside RCTs enhances comprehensiveness but may introduce heterogeneity and bias, thereby affecting the internal validity of the review. Clear plans for addressing potential heterogeneity or bias (e.g., risk-of-bias assessments, subgroup analyses) are not specified.”
We acknowledge this concern. The inclusion of quasi-experimental and prospective cohort studies was intentional, reflecting the limited availability of randomized controlled trials in pediatric populations. To address potential heterogeneity and bias, we have clarified our analytical approach in the revised manuscript. Specifically, risk of bias will be assessed using the Cochrane Risk of Bias 2 (RoB 2) tool for randomized trials and the Newcastle–Ottawa Scale (NOS) for non-randomized studies. In addition, subgroup and sensitivity analyses will explicitly explore the influence of study design, and non-randomized studies will be excluded from quantitative synthesis when appropriate, contributing instead to narrative synthesis. Therefore, we revised sections 2.7 and 2.8 accordingly and now it reads clearly. In addition, we added references to RoB 2, NOS and GRADE framework.
Comment 3: “Exclusion of Retrospective and Cross-Sectional Studies: While justified, this exclusion limits the evidence base to prospective designs, which are generally more robust but may reduce the total number of eligible studies, especially in pediatric populations where RCTs are less common.”
We agree that excluding retrospective and cross-sectional studies may limit the number of eligible studies. However, these designs do not allow assessment of temporal relationships between exposure and outcome, which is essential for evaluating preventive effects on type 2 diabetes. This exclusion was therefore maintained to preserve methodological rigor. We have clarified this rationale in the Study Design section (2.4) of the revised manuscript.
Comment 4: “Outcome Definitions: The primary outcome is defined according to ISPAD and ADA guidelines, which promote standardisation; however, variability in how studies report incidence and diagnostic criteria may introduce heterogeneity. The protocol does not mention plans to address potential inconsistencies.”
We agree with the reviewer that variability in outcome definitions may introduce heterogeneity. To address this, we have clarified our analytical approach in the revised manuscript in section 2.8. Specifically, outcome definitions will be harmonized where possible using standardized diagnostic thresholds. When multiple diagnostic criteria are applied (e.g., ISPAD, ADA, or author-defined definitions), sensitivity analyses will be conducted according to the diagnostic criteria used. Where harmonization is not feasible, results will be summarized narratively.
Comment 5: “Study Quality and Bias Assessment: The protocol does not specify how the quality or risk of bias of included studies will be assessed, which is critical for evaluating the robustness and validity of synthesised evidence.”
We thank the reviewer for highlighting the importance of risk-of-bias assessment. We respectfully note that this information was already included in Section 2.7 of the original manuscript. Nevertheless, to enhance clarity and visibility, we have strengthened this section in the revised version. Risk of bias will be assessed using the Cochrane Risk of Bias 2 (RoB 2) tool for randomized controlled trials and the Newcastle–Ottawa Scale (NOS) for non-randomized studies, with independent assessment by two reviewers and resolution of disagreements by consensus or adjudication by a third reviewer.
Comment 6: “Handling of Heterogeneity: There is no mention of plans to assess heterogeneity (e.g., via statistical tests or the I² statistic) or to perform meta-analyses, both of which are important for evaluating the robustness of pooled estimates.”
We thank the reviewer for this comment. We have clarified our approach to assessing heterogeneity and performing meta-analyses in the revised manuscript. Specifically, statistical heterogeneity will be evaluated using the I² statistic and Cochran’s Q test, and quantitative synthesis will be performed using a random-effects meta-analysis when sufficient homogeneity exists. When heterogeneity is substantial or pooling is not appropriate, findings will be summarized narratively. These methodological details are now explicitly described in Section 2.8 (Data Synthesis and Statistical Analysis).
Comment 7: “Potential for Bias in Selection and Data Extraction: Although dual review is planned, details on reviewer calibration or training and on procedures for resolving persistent disagreements are not provided.”
We thank the reviewer for this important comment. To further strengthen transparency and methodological rigor, we have clarified the procedures for reviewer calibration and disagreement resolution in the revised manuscript. Specifically, reviewers will complete a pilot calibration exercise prior to full screening and data extraction, and any persistent disagreements will be resolved through discussion or adjudication by a third senior reviewer.
Reviewer 2
“This protocol addresses a timely and clinically relevant question regarding the role of metformin in the prevention of type 2 diabetes among children and adolescents with overweight or obesity. The rationale is clearly presented, the objectives are well defined, and the protocol is appropriately registered in PROSPERO and aligned with PRISMA-P guidelines.
The eligibility criteria, outcomes of interest, and planned analytical approaches are generally well described and methodologically sound. The inclusion of both randomized and prospective observational studies is justified, and the proposed subgroup and sensitivity analyses are appropriate given the anticipated heterogeneity of the available literature. Some minor clarifications could further strengthen the protocol.”
We sincerely thank the reviewer for the positive and encouraging evaluation of our protocol. We are grateful for the constructive suggestions, which have helped us improve clarity, transparency, and editorial consistency. Below we address each comment point by point and describe the corresponding revisions made to the manuscript.
Comment 1: Providing a brief summary of the planned search strategy within the main manuscript (in addition to the PROSPERO registration and supplementary materials) would enhance transparency and reproducibility.
We thank the reviewer for this suggestion and agree that transparency of the search strategy is important. As requested, we have ensured that the main manuscript includes a concise description of the databases searched and the core search concepts (e.g., metformin, type 2 diabetes/prediabetes, insulin resistance, overweight/obesity, and pediatric populations), while retaining the full database-specific search strategies in Supplementary File S1 and the PROSPERO registration.
Comment 2: In addition, further clarification regarding how studies including concomitant pharmacological interventions will be handled analytically would be helpful.
We agree and appreciate this comment. We have clarified that studies in which metformin is administered in combination with other pharmacological interventions will be included, but their results will be handled analytically through subgroup or sensitivity analyses. When the independent effect of metformin cannot be reasonably isolated, such studies will be excluded from quantitative synthesis and summarized narratively.
Comment 3: Minor language refinements could improve clarity and readability in some sections. In particular, the authors are encouraged to ensure consistent use of acronyms throughout the manuscript. Several terms (e.g., type 2 diabetes, T2D) are appropriately defined at first mention but are later alternated with the full term, which may affect readability. Standardizing terminology after initial definition would improve editorial consistency.
We thank the reviewer for this editorial suggestion. To improve clarity and readability, we have revised the manuscript to consistently use the full term “type 2 diabetes” throughout the text, avoiding the use of abbreviations. This decision was made to enhance readability and ensure terminological consistency across sections.
Comment 4: The evidence table is comprehensive and well structured; however, its clarity and readability could be improved. The authors may consider simplifying or better defining certain columns (e.g., risk of bias rating and comments) and optimizing the layout to enhance usability, particularly given the protocol nature of the manuscript.
We thank the reviewer for this helpful suggestion. To improve clarity and usability, we revised the evidence table by removing the “Comments” column, as interpretative remarks will be addressed in the narrative Results and Discussion sections. In addition, the table layout was optimized by grouping variables under thematic headings, thereby enhancing readability while preserving the comprehensive scope appropriate for a protocol.
Comment 5: The sections describing expected results and discussion are informative; however, further emphasizing their exploratory and hypothesis-generating nature may improve conceptual clarity.
We thank the reviewer for this valuable suggestion. We agree that emphasizing the exploratory nature of these sections is important for conceptual clarity. Accordingly, we have revised both the Expected Results and Discussion sections to explicitly state that they are exploratory and hypothesis-generating, avoiding overinterpretation and reflecting the protocol-based nature of the manuscript.
Overall, this is a well-designed and carefully prepared protocol that is likely to generate valuable evidence to inform future pediatric diabetes prevention research.
We are grateful to both reviewers for their thoughtful comments, which have substantially improved the clarity, transparency, and methodological rigor of this protocol. We added a recent review addressing metabolic interventions in pediatric obesity to contextualize the frequent consideration of metformin in clinical practice, while clearly distinguishing narrative evidence from the primary prevention outcomes assessed in this protocol. We believe the revised manuscript now more clearly reflects its exploratory intent and adherence to best practices for systematic review protocols.
​​
Author Response File: 
Author Response.docx
Reviewer 2 Report
Comments and Suggestions for AuthorsThis protocol addresses a timely and clinically relevant question regarding the role of metformin in the prevention of type 2 diabetes among children and adolescents with overweight or obesity. The rationale is clearly presented, the objectives are well defined, and the protocol is appropriately registered in PROSPERO and aligned with PRISMA-P guidelines.
The eligibility criteria, outcomes of interest, and planned analytical approaches are generally well described and methodologically sound. The inclusion of both randomized and prospective observational studies is justified, and the proposed subgroup and sensitivity analyses are appropriate given the anticipated heterogeneity of the available literature.
Some minor clarifications could further strengthen the protocol. Providing a brief summary of the planned search strategy within the main manuscript (in addition to the PROSPERO registration and supplementary materials) would enhance transparency and reproducibility. In addition, further clarification regarding how studies including concomitant pharmacological interventions will be handled analytically would be helpful.
Minor language refinements could improve clarity and readability in some sections. In particular, the authors are encouraged to ensure consistent use of acronyms throughout the manuscript. Several terms (e.g., type 2 diabetes, T2D) are appropriately defined at first mention but are later alternated with the full term, which may affect readability. Standardizing terminology after initial definition would improve editorial consistency.
The evidence table is comprehensive and well structured; however, its clarity and readability could be improved. The authors may consider simplifying or better defining certain columns (e.g., risk of bias rating and comments) and optimizing the layout to enhance usability, particularly given the protocol nature of the manuscript.
The sections describing expected results and discussion are informative; however, further emphasizing their exploratory and hypothesis-generating nature may improve conceptual clarity. Overall, this is a well-designed and carefully prepared protocol that is likely to generate valuable evidence to inform future pediatric diabetes prevention research.
Comments on the Quality of English LanguageThe English language is generally clear and appropriate for a scientific protocol. Minor language refinements may improve clarity and readability in some sections.
Author Response
London, December 30th, 2025
Response to Reviewers
We sincerely thank the reviewers for their careful evaluation of our protocol and for their constructive and insightful comments. Below we provide a point-by-point response and describe the corresponding revisions made to the manuscript. In addition, the revised manuscript has been carefully proofread by two native English speakers to improve clarity and language quality.
Reviewer 1
Comment 1: “Potential Limitations and Areas Impacting Robustness: Search Strategy Details: While the broad database coverage is a strength, the protocol states that the complete search strategy will be available in Supplementary File S1 and PROSPERO. The robustness depends heavily on the specificity and sensitivity of these strategies, which are not detailed here. Without access to the actual search strings, it's difficult to assess whether the search terms are comprehensive enough to minimize missed studies.”
We thank the reviewer for this comment. The search strategy is already described in the main manuscript (Section 2.2), where we outline the databases searched, the core search concepts, and the use of both MeSH terms and free-text keywords. In addition, the full database-specific search strings, including Boolean operators and controlled vocabulary, were provided with the original submission as Supplementary File S1 and are publicly available in the PROSPERO registration and can promptly be accessed. Therefore, no further modification of the search strategy or manuscript text was required.
Comment 2: “Inclusion of Study Designs: Including quasi-experimental and prospective cohort studies alongside RCTs enhances comprehensiveness but may introduce heterogeneity and bias, thereby affecting the internal validity of the review. Clear plans for addressing potential heterogeneity or bias (e.g., risk-of-bias assessments, subgroup analyses) are not specified.”
We acknowledge this concern. The inclusion of quasi-experimental and prospective cohort studies was intentional, reflecting the limited availability of randomized controlled trials in pediatric populations. To address potential heterogeneity and bias, we have clarified our analytical approach in the revised manuscript. Specifically, risk of bias will be assessed using the Cochrane Risk of Bias 2 (RoB 2) tool for randomized trials and the Newcastle–Ottawa Scale (NOS) for non-randomized studies. In addition, subgroup and sensitivity analyses will explicitly explore the influence of study design, and non-randomized studies will be excluded from quantitative synthesis when appropriate, contributing instead to narrative synthesis. Therefore, we revised sections 2.7 and 2.8 accordingly and now it reads clearly. In addition, we added references to RoB 2, NOS and GRADE framework.
Comment 3: “Exclusion of Retrospective and Cross-Sectional Studies: While justified, this exclusion limits the evidence base to prospective designs, which are generally more robust but may reduce the total number of eligible studies, especially in pediatric populations where RCTs are less common.”
We agree that excluding retrospective and cross-sectional studies may limit the number of eligible studies. However, these designs do not allow assessment of temporal relationships between exposure and outcome, which is essential for evaluating preventive effects on type 2 diabetes. This exclusion was therefore maintained to preserve methodological rigor. We have clarified this rationale in the Study Design section (2.4) of the revised manuscript.
Comment 4: “Outcome Definitions: The primary outcome is defined according to ISPAD and ADA guidelines, which promote standardisation; however, variability in how studies report incidence and diagnostic criteria may introduce heterogeneity. The protocol does not mention plans to address potential inconsistencies.”
We agree with the reviewer that variability in outcome definitions may introduce heterogeneity. To address this, we have clarified our analytical approach in the revised manuscript in section 2.8. Specifically, outcome definitions will be harmonized where possible using standardized diagnostic thresholds. When multiple diagnostic criteria are applied (e.g., ISPAD, ADA, or author-defined definitions), sensitivity analyses will be conducted according to the diagnostic criteria used. Where harmonization is not feasible, results will be summarized narratively.
Comment 5: “Study Quality and Bias Assessment: The protocol does not specify how the quality or risk of bias of included studies will be assessed, which is critical for evaluating the robustness and validity of synthesised evidence.”
We thank the reviewer for highlighting the importance of risk-of-bias assessment. We respectfully note that this information was already included in Section 2.7 of the original manuscript. Nevertheless, to enhance clarity and visibility, we have strengthened this section in the revised version. Risk of bias will be assessed using the Cochrane Risk of Bias 2 (RoB 2) tool for randomized controlled trials and the Newcastle–Ottawa Scale (NOS) for non-randomized studies, with independent assessment by two reviewers and resolution of disagreements by consensus or adjudication by a third reviewer.
Comment 6: “Handling of Heterogeneity: There is no mention of plans to assess heterogeneity (e.g., via statistical tests or the I² statistic) or to perform meta-analyses, both of which are important for evaluating the robustness of pooled estimates.”
We thank the reviewer for this comment. We have clarified our approach to assessing heterogeneity and performing meta-analyses in the revised manuscript. Specifically, statistical heterogeneity will be evaluated using the I² statistic and Cochran’s Q test, and quantitative synthesis will be performed using a random-effects meta-analysis when sufficient homogeneity exists. When heterogeneity is substantial or pooling is not appropriate, findings will be summarized narratively. These methodological details are now explicitly described in Section 2.8 (Data Synthesis and Statistical Analysis).
Comment 7: “Potential for Bias in Selection and Data Extraction: Although dual review is planned, details on reviewer calibration or training and on procedures for resolving persistent disagreements are not provided.”
We thank the reviewer for this important comment. To further strengthen transparency and methodological rigor, we have clarified the procedures for reviewer calibration and disagreement resolution in the revised manuscript. Specifically, reviewers will complete a pilot calibration exercise prior to full screening and data extraction, and any persistent disagreements will be resolved through discussion or adjudication by a third senior reviewer.
Reviewer 2
“This protocol addresses a timely and clinically relevant question regarding the role of metformin in the prevention of type 2 diabetes among children and adolescents with overweight or obesity. The rationale is clearly presented, the objectives are well defined, and the protocol is appropriately registered in PROSPERO and aligned with PRISMA-P guidelines.
The eligibility criteria, outcomes of interest, and planned analytical approaches are generally well described and methodologically sound. The inclusion of both randomized and prospective observational studies is justified, and the proposed subgroup and sensitivity analyses are appropriate given the anticipated heterogeneity of the available literature. Some minor clarifications could further strengthen the protocol.”
We sincerely thank the reviewer for the positive and encouraging evaluation of our protocol. We are grateful for the constructive suggestions, which have helped us improve clarity, transparency, and editorial consistency. Below we address each comment point by point and describe the corresponding revisions made to the manuscript.
Comment 1: Providing a brief summary of the planned search strategy within the main manuscript (in addition to the PROSPERO registration and supplementary materials) would enhance transparency and reproducibility.
We thank the reviewer for this suggestion and agree that transparency of the search strategy is important. As requested, we have ensured that the main manuscript includes a concise description of the databases searched and the core search concepts (e.g., metformin, type 2 diabetes/prediabetes, insulin resistance, overweight/obesity, and pediatric populations), while retaining the full database-specific search strategies in Supplementary File S1 and the PROSPERO registration.
Comment 2: In addition, further clarification regarding how studies including concomitant pharmacological interventions will be handled analytically would be helpful.
We agree and appreciate this comment. We have clarified that studies in which metformin is administered in combination with other pharmacological interventions will be included, but their results will be handled analytically through subgroup or sensitivity analyses. When the independent effect of metformin cannot be reasonably isolated, such studies will be excluded from quantitative synthesis and summarized narratively.
Comment 3: Minor language refinements could improve clarity and readability in some sections. In particular, the authors are encouraged to ensure consistent use of acronyms throughout the manuscript. Several terms (e.g., type 2 diabetes, T2D) are appropriately defined at first mention but are later alternated with the full term, which may affect readability. Standardizing terminology after initial definition would improve editorial consistency.
We thank the reviewer for this editorial suggestion. To improve clarity and readability, we have revised the manuscript to consistently use the full term “type 2 diabetes” throughout the text, avoiding the use of abbreviations. This decision was made to enhance readability and ensure terminological consistency across sections.
Comment 4: The evidence table is comprehensive and well structured; however, its clarity and readability could be improved. The authors may consider simplifying or better defining certain columns (e.g., risk of bias rating and comments) and optimizing the layout to enhance usability, particularly given the protocol nature of the manuscript.
We thank the reviewer for this helpful suggestion. To improve clarity and usability, we revised the evidence table by removing the “Comments” column, as interpretative remarks will be addressed in the narrative Results and Discussion sections. In addition, the table layout was optimized by grouping variables under thematic headings, thereby enhancing readability while preserving the comprehensive scope appropriate for a protocol.
Comment 5: The sections describing expected results and discussion are informative; however, further emphasizing their exploratory and hypothesis-generating nature may improve conceptual clarity.
We thank the reviewer for this valuable suggestion. We agree that emphasizing the exploratory nature of these sections is important for conceptual clarity. Accordingly, we have revised both the Expected Results and Discussion sections to explicitly state that they are exploratory and hypothesis-generating, avoiding overinterpretation and reflecting the protocol-based nature of the manuscript.
Overall, this is a well-designed and carefully prepared protocol that is likely to generate valuable evidence to inform future pediatric diabetes prevention research.
We are grateful to both reviewers for their thoughtful comments, which have substantially improved the clarity, transparency, and methodological rigor of this protocol. We added a recent review addressing metabolic interventions in pediatric obesity to contextualize the frequent consideration of metformin in clinical practice, while clearly distinguishing narrative evidence from the primary prevention outcomes assessed in this protocol. We believe the revised manuscript now more clearly reflects its exploratory intent and adherence to best practices for systematic review protocols.
​​
Author Response File: Author Response.docx
