Early SGLT2 Inhibitor Therapy in Acute Coronary Syndrome: Mitigating Adverse Remodeling in High-Risk Phenotypes—A Real-World Study

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Congratulations to the authors for their interesting manuscript; I just have some comments about it:

The follow-up period (3–34 months) is excessively wide, potentially introducing significant variability in the remodelling process. Although statistical adjustment was performed, such heterogeneity likely weakens the physiological interpretation of the findings; indeed the authors should provide a clear justification for this broad time range.

It is also recommended that the authors implement a propensity score–matched design and incorporate multivariate models with interaction terms (SGLT2i × diabetes) to explore whether therapeutic effects differ according to diabetic status.

The concept of “phenotype-dependent remodelling” is scientifically appealing but remains largely speculative; the discussion of mechanistic pathways should be streamlined to maintain clarity and focus. Moreover authors are encouraged to include in their discussion the latest evidences of pleiotropic effects of sglt2i drugs (doi: 10.1016/j.hrthm.2025.09.023.)

 

Author Response

Dear Editor, Dear Reviewers,

We would like to express our sincere gratitude for the time and effort dedicated to reviewing our manuscript. We found the comments to be highly constructive, insightful, and instrumental in improving the quality, clarity, and methodological rigor of our study.

We have carefully addressed each point raised by the reviewers. The manuscript has been revised to clarify the statistical approach, temper causal language, and incorporate additional analyses regarding the interaction between diabetes and treatment effects.

Below is a point-by-point response detailing the changes made.

Comment 1: "The follow-up period (3–34 months) is excessively wide, potentially introducing significant variability in the remodeling process. Although statistical adjustment was performed, such heterogeneity likely weakens the physiological interpretation of the findings; indeed the authors should provide a clear justification for this broad time range."

Response: We agree that the wide follow-up range is a significant limitation inherent to the retrospective, real-world nature of our study.

• Action: We have expanded the Limitations section to explicitly acknowledge this heterogeneity and discuss how it might introduce "noise" into the remodeling assessment. We also clarified in the Methods section that the follow-up duration was included as a covariate in all multivariable regression models to statistically adjust for this variability.
• Location in Manuscript: [See Methods]

Comment 2: "It is also recommended that the authors implement a propensity score–matched (PSM) design and incorporate multivariate models with interaction terms (SGLT2i × diabetes) to explore whether therapeutic effects differ according to diabetic status."

Response: We appreciate these excellent methodological suggestions.

1. Regarding Interaction Terms: We have performed the suggested analysis. We introduced interaction terms (SGLT2i × Diabetes) into our multivariable models. The results showed no statistically significant interaction (p > 0.05 for all outcomes), suggesting that the association between SGLT2i and structural stabilization was consistent across both diabetic and non-diabetic subgroups. We have added a new subsection "3.4. Interaction Analysis by Diabetes Status" and a new Table 4 presenting these results.
2. Regarding PSM: We seriously considered implementing PSM. However, given our relatively small sample size in the active treatment group (n=71), a strict matching process would have significantly reduced the sample size and statistical power. As our goal was to analyse the full "real-world" cohort, we opted for multivariable linear regression as the primary adjustment method to preserve the sample size, a decision we have now formally justified in the Methods section.

• Location in Manuscript: [See Methods]

Comment 3: "The concept of “phenotype-dependent remodeling” is scientifically appealing but remains largely speculative; the discussion of mechanistic pathways should be streamlined to maintain clarity and focus."

Response: We agree. We have streamlined the discussion regarding metabolic mechanisms to avoid over-speculation and focused more on the clinical observations and the new interaction analysis results.

Comment 4: "Moreover authors are encouraged to include in their discussion the latest evidence of pleiotropic effects of sglt2i drugs (doi: 10.1016/j.hrthm.2025.09.023)."

Response: Thank you for pointing out this relevant recent evidence. We have included the suggested reference to support the discussion on the pleiotropic effects of SGLT2 inhibitors, specifically regarding structural integrity and electrophysiology.

• Location in Manuscript: [See Discussion 48]

Reviewer 2 Report

Comments and Suggestions for Authors

This study investigates a significant and current clinical issue: the effects of starting SGLT2 inhibitors soon after an acute coronary syndrome (ACS) event and how this relates to changes in the left ventricle in real-world clinical settings.  This study is well-structured, methodically organized, and based on current understandings of disease processes, supported by recent findings from randomized clinical trials. 
The manuscript's strengths include a detailed description of the patients' characteristics, a clear acknowledgment of potential biases related to the reasons for treatment, appropriate use of multivariable statistical methods, and a thorough discussion that avoids making unsupported claims about cause and effect.  The echocardiographic measurements used are clinically important, and the figures and tables are presented in a clear and helpful way. 
However, several aspects could be improved to enhance both interpretability and scientific validity. 
Causal analysis and contextual framing are important for understanding complex situations. 
Even though the authors correctly acknowledge the study's retrospective and observational design, some of the language used in the discussion and conclusions suggests a protective or beneficial effect of SGLT2 inhibitors.  Because the adjusted models didn't show statistical independence, it's better to describe the results in terms of association and risk correlation rather than as evidence of therapeutic effectiveness. 
2. Addressing confounding factors related to indications 
The authors acknowledge that confounding by indication is a key aspect of this study.  However, the manuscript would be improved by providing a clearer explanation of why alternative methods, like propensity score matching or inverse probability weighting, weren't used. Alternatively, a brief discussion of how these methods might have affected the results could be included. 
3. Follow-up Variability 
Even though the average follow-up time was similar for both groups and statistically controlled, the wide range of follow-up periods, from three to thirty-four months, could still contribute to biological differences in how remodeling progresses.  The discussion of changes over time could benefit from a more thorough examination of this limitation. 
4. Structural Endpoints 
While relying on standard 2D echocardiographic measurements is acceptable, it's important to note that the lack of strain imaging or cardiac MRI could be a reason for potentially missing subtle or early signs of heart muscle changes. 
This study offers a thorough and clinically relevant analysis based on real-world data.  With minor changes to improve the clarity of cause-and-effect language and to strengthen the methodological discussion, this manuscript would significantly contribute to the existing research on managing patients after an acute coronary syndrome (ACS) and the use of SGLT2 inhibitors.

Author Response

Comment 1: "Even though the authors correctly acknowledge the study's retrospective and observational design, some of the language used in the discussion and conclusions suggests a protective or beneficial effect of SGLT2 inhibitors. Because the adjusted models didn't show statistical independence, it's better to describe the results in terms of association and risk correlation rather than as evidence of therapeutic effectiveness."

Response: We fully accept this criticism. We have carefully reviewed the entire manuscript (Abstract, Discussion, and Conclusions) and systematically revised the language.

• Action: We replaced causal terms such as "mitigated," "prevented," or "efficacy" with observational phrasing like "was associated with," "correlated with," "closely paralleled," and "structural stabilization." This ensures the conclusions strictly reflect the statistical findings.
• Location in Manuscript: [See Abstract, Discussion, Conclusions]

Comment 2: "Addressing confounding factors related to indications. The manuscript would be improved by providing a clearer explanation of why alternative methods, like propensity score matching or inverse probability weighting, weren't used."

Response: As mentioned in our response to Reviewer 1, we have added a dedicated paragraph in the Methods section explaining that multivariable regression was preferred over PSM to avoid the exclusion of patients and loss of statistical power in this specific cohort size. We also acknowledged this decision in the Limitations section.

• Location in Manuscript: [See Methods]

Comment 3: "Follow-up Variability. The discussion of changes over time could benefit from a more thorough examination of this limitation."

Response: We have expanded the Limitations section to discuss how the variable follow-up (3–34 months) represents a challenge in longitudinal real-world data and may obscure subtle therapeutic effects, despite statistical adjustment.

• Location in Manuscript: [See Limitations]

Comment 4: "Structural Endpoints. It's important to note that the lack of strain imaging or cardiac MRI could be a reason for potentially missing subtle or early signs of heart muscle changes."

Response: We agree that this is a critical limitation. We have revised the limitations section to explicitly state that the reliance on standard 2D echocardiography (LVEF/LVM) rather than more sensitive modalities like GLS or CMR might explain the lack of independent statistical significance in the adjusted models. Unfortunately, this data is not available for this study, our data base being selected from a retrospective cohort reflecting the real-life practice.

• Location in Manuscript: [See Limitations]

Conclusion

We believe that these revisions have significantly strengthened the manuscript. We thank the reviewers again for their guidance and hope the revised version is now suitable for publication in Medicina.

Sincerely,

Teodora Mateoc-Sîrb et al.

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

My congratulations to the authors for their revised version of the manuscript