Comparison of Adhesive Strategies with Different Etching Approaches on the Clinical Performance of Restorations in Non-Carious Cervical Lesions: A Systematic Review and Network Meta-Analysis

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This is a clinically relevant and potentially valuable systematic review and network meta-analysis on restorative strategies for NCCLs. The large evidence base and the strategy-based node definition are strengths. However, I recommend major revision before the manuscript can be considered further.

1. Please clarify study eligibility and risk-of-bias assessment. The Methods indicate that randomized and non-randomized controlled clinical trials were eligible, but RoB 2 was used for the included studies. If non-randomized trials were included, they should be assessed with an appropriate tool, or the review should be restricted to randomized studies.

2. Please make the outcome operationalization fully reproducible. The manuscript states that marginal adaptation and retention loss were assessed with USPHS/FDI/comparable criteria and analyzed as event-based risk ratios, but the threshold used to convert ordinal clinical scores into “failure” is not sufficiently explicit in the main Methods. A clear mapping table or supplementary appendix is needed.

3. The primary analytical strategy needs stronger justification. The main NMA uses the longest available follow-up from each study despite marked variation in follow-up duration. In addition, the retention network contains many zero-event arms handled with a 0.5 continuity correction. Sensitivity analyses by follow-up interval and sparse-data handling would strengthen the robustness of the findings.

4. Please resolve the interpretation inconsistencies. The Results state that only marginal adaptation and retention loss were suitable for quantitative NMA, yet Figure 5 presents an overall P-score ranking across clinical outcomes. In the same section, clinical importance is defined as any RR different from 1.0, which is not an adequate clinical threshold. This framework should be clarified/revised, and the conclusions should be toned down so that they remain closer to the statistically supported contrasts and the remaining imprecision/heterogeneity.

Comments on the Quality of English Language

A light English/style revision is also advisable.

Author Response

We would like to sincerely thank the Editor and the reviewers for their careful evaluation of our manuscript and for their insightful and constructive comments. We greatly appreciate the time and expertise invested in reviewing our work. The suggestions provided were highly pertinent and have significantly contributed to improving the clarity and presentation of the results and the overall quality of our manuscript. All comments have been carefully addressed and the manuscript has been revised accordingly. For ease of identification, all modifications introduced in the revised manuscript have been highlighted in yellow.

The authors

 

Reviewer 1

Please clarify study eligibility and risk-of-bias assessment. The Methods indicate that randomized and non-randomized controlled clinical trials were eligible, but RoB 2 was used for the included studies. If non-randomized trials were included, they should be assessed with an appropriate tool, or the review should be restricted to randomized studies.

R/ Many thanks to the reviewer for this pertinent comment and indeed, there was an error of inconsistency that was amended by homogenizing the text clarifying that only RCTs were established as eligible in the study.

 

Please make the outcome operationalization fully reproducible. The manuscript states that marginal adaptation and retention loss were assessed with USPHS/FDI/comparable criteria and analyzed as event-based risk ratios, but the threshold used to convert ordinal clinical scores into “failure” is not sufficiently explicit in the main Methods. A clear mapping table or supplementary appendix is needed.

R/ The equivalences on the definition of failure and acceptable outcomes according to FDI and USPHS were clarified through an explanatory text in subsection 2.5 of the methodology and a table was added that clarifies it in detail.

 

The primary analytical strategy needs stronger justification. The main NMA uses the longest available follow-up from each study despite marked variation in follow-up duration. In addition, the retention network contains many zero-event arms handled with a 0.5 continuity correction. Sensitivity analyses by follow-up interval and sparse-data handling would strengthen the robustness of the findings.

R/ We appreciate this suggestion and have added an explicit description of sensitivity analyses in the Statistical Analysis section. These analyses explored alternative follow-up intervals and examined the potential influence of sparse data handling procedures. The additional analyses confirmed that the direction of the treatment effects remained consistent with the primary analysis based on the longest available follow-up.

A paragraph explaining the above was added at the end of subsection 2.5 of the methodology.

 

Please resolve the interpretation inconsistencies. The Results state that only marginal adaptation and retention loss were suitable for quantitative NMA, yet Figure 5 presents an overall P-score ranking across clinical outcomes. In the same section, clinical importance is defined as any RR different from 1.0, which is not an adequate clinical threshold. This framework should be clarified/revised, and the conclusions should be toned down so that they remain closer to the statistically supported contrasts and the remaining imprecision/heterogeneity.

R/ The manuscript has been revised to clarify that the ranking framework is derived exclusively from the two outcomes that generated connected networks (marginal adaptation and retention loss). In addition, the definition of clinical importance has been revised to avoid interpreting any deviation from the null value as clinically meaningful. Treatment effects are now interpreted in conjunction with their confidence intervals and the clinical plausibility of the observed differences.

The Conclusions section has been revised to better align with the statistically supported contrasts and to avoid overinterpretation of the network ranking results. The revised text now emphasizes that significant differences were limited to a subset of comparisons and that the observed patterns should be interpreted cautiously given the remaining imprecision and heterogeneity across studies.

This clarification also addresses the reviewer’s concern regarding potential overinterpretation of P-score rankings.

Reviewer 2 Report

Comments and Suggestions for Authors

This manuscript presents a comprehensive and methodologically rigorous systematic review and frequentist random-effects network meta-analysis evaluating the clinical performance of restorative strategies for non-carious cervical lesions. The study looks to be clearly reported, and adheres to the major Prisma guidelines. The use of CINeMA to assess confidence in the network estimates further strengthens the transparency and interpretability of the findings.

Overall, the manuscript represents a valuable contribution to the literature on adhesive dentistry and NCCL management. The conclusions are generally supported by the data. However, several issues related to clarity, methodological justification, and presentation have to be addressed to improve the manuscript’s robustness and readability.

• The choice of resin-modified glass ionomer cement (RMGI) as the reference comparator throughout the network meta-analyses is clinically reasonable, but the rationale should be more explicitly justified in the methods section. Please clarify whether RMGI was selected due to historical use, frequency within the included trials, or its distinct bonding mechanism. Additionally, discuss how the relatively smaller number of RMGI arms may influence indirect comparisons and interpretation of effect estimates.
• As acknowledged in the Limitations section, most included trials analyzed restorations rather than patients, and intra-patient correlation was not accounted for. Please expand this discussion slightly to explain how this limitation may differentially affect outcomes such as retention loss versus marginal adaptation. If possible, indicate whether sensitivity analyses or design considerations could mitigate this issue in future research.
• While the forest plots and relative risks are clearly presented, the manuscript would benefit from a more explicit discussion of clinical relevance. For example, are the observed reductions in risk (as RR values below 1.0 for selective-etch strategies) likely to translate into meaningful differences in daily clinical practice? Consider adding brief absolute-risk context or qualitative interpretation in the Discussion.
• The manuscript appropriately cautions against overinterpretation of P-scores; however, the ranking figure (Figure 5) is visually prominent. Consider reinforcing in the Results or Discussion that rankings do not imply superiority unless supported by statistically significant and precise estimates.
• Terminology consistency: please, ensure consistent use of abbreviations (CR-NANO vs. CR_NANO) throughout the text, figures, and tables.
• Check for minor typographical errors

 

The manuscript is scientifically sound and clinically relevant and some minor adjustments will further enhance its quality and impact.

 

Author Response

We would like to sincerely thank the Editor and the reviewers for their careful evaluation of our manuscript and for their insightful and constructive comments. We greatly appreciate the time and expertise invested in reviewing our work. The suggestions provided were highly pertinent and have significantly contributed to improving the clarity and presentation of the results and the overall quality of our manuscript. All comments have been carefully addressed and the manuscript has been revised accordingly. For ease of identification, all modifications introduced in the revised manuscript have been highlighted in yellow.

The authors

Reviewer 2

This manuscript presents a comprehensive and methodologically rigorous systematic review and frequentist random-effects network meta-analysis evaluating the clinical performance of restorative strategies for non-carious cervical lesions. The study looks to be clearly reported, and adheres to the major Prisma guidelines. The use of CINeMA to assess confidence in the network estimates further strengthens the transparency and interpretability of the findings.

Overall, the manuscript represents a valuable contribution to the literature on adhesive dentistry and NCCL management. The conclusions are generally supported by the data. However, several issues related to clarity, methodological justification, and presentation have to be addressed to improve the manuscript’s robustness and readability.

The choice of resin-modified glass ionomer cement (RMGI) as the reference comparator throughout the network meta-analyses is clinically reasonable, but the rationale should be more explicitly justified in the methods section. Please clarify whether RMGI was selected due to historical use, frequency within the included trials, or its distinct bonding mechanism. Additionally, discuss how the relatively smaller number of RMGI arms may influence indirect comparisons and interpretation of effect estimates.

R/ We appreciate this comment and have expanded the Methods section to clarify the rationale for selecting resin-modified glass ionomer cement (RMGI) as the reference comparator. RMGI represents a clinically distinct restorative approach commonly used in NCCL management due to its chemical adhesion to dentin and fluoride release. Its consistent presence across the included trials also made it a suitable and clinically interpretable anchor for the network meta-analysis.

To highlight these aspects, a paragraph has been added in section 2.5 after the node definition.

 

As acknowledged in the Limitations section, most included trials analyzed restorations rather than patients, and intra-patient correlation was not accounted for. Please expand this discussion slightly to explain how this limitation may differentially affect outcomes such as retention loss versus marginal adaptation. If possible, indicate whether sensitivity analyses or design considerations could mitigate this issue in future research.

R/ We agree and have expanded the Limitations section to clarify how the use of restorations rather than patients as the unit of analysis may differentially affect outcomes such as retention loss and marginal adaptation.

A paragraph was added at the end of the limitations expanding them and incorporating suggested elements.

 

While the forest plots and relative risks are clearly presented, the manuscript would benefit from a more explicit discussion of clinical relevance. For example, are the observed reductions in risk (as RR values below 1.0 for selective-etch strategies) likely to translate into meaningful differences in daily clinical practice? Consider adding brief absolute-risk context or qualitative interpretation in the Discussion.

R/ Paragraph has been added to the Discussion emphasizing that relative risk reductions should be interpreted in the context of absolute event rates and clinical decision-making. The revised text clarifies that statistically significant differences may correspond to modest absolute differences in clinical practice.

 

The manuscript appropriately cautions against overinterpretation of P-scores; however, the ranking figure (Figure 5) is visually prominent. Consider reinforcing in the Results or Discussion that rankings do not imply superiority unless supported by statistically significant and precise estimates.

R/ Additional text has been incorporated in both the Results and Discussion sections reinforcing that P-score rankings should be interpreted cautiously and should not be interpreted as evidence of superiority unless supported by statistically significant and precise effect estimates.

 

Terminology consistency: please, ensure consistent use of abbreviations (CR-NANO vs. CR_NANO) throughout the text, figures, and tables.

R/ Todas las inconsistencias en la terminología fueron corregidas en el texto y tablas, ajustándolas a la manera en que aparecen en las figuras.