Review Reports - Vancomycin Exposure Dynamics and Clinical Outcomes in Critically Ill Patients: A Retrospective Cohort Study

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This manuscript addresses the evaluation of vancomycin exposure metrics in ICU patients, a topic of clear clinical relevance. Optimizing vancomycin therapy in critically ill populations remains challenging due to high pharmacokinetic variability, altered physiology, and the narrow therapeutic window. The attempt to explore novel metrics such as Time in Range (TIR) and volatility index in relation to ICU mortality is interesting and potentially valuable. However, several important methodological and interpretative concerns should be addressed:

- The manuscript introduces TIR and a volatility index as exposure metrics; however, their clinical utility remains unclear. These parameters require multiple therapeutic drug monitoring (TDM) measurements, which may not be feasible in many ICU settings. The authors should better justify:

What is the added value of these metrics compared to standard PK/PD indices (e.g., AUC/MIC)?
How could TIR and volatility index be realistically implemented in routine clinical practice?
Are there actionable thresholds that clinicians can use prospectively?

- Although the reported odds ratios suggest potentially meaningful associations with mortality, the confidence intervals are quite wide. This indicates limited precision and raises concerns about sample size adequacy, statistical power, stability of the estimated effects. These issues should be explicitly acknowledged and discussed. Whether the findings should be considered exploratory?

- ICU mortality is inherently multifactorial. Adjustment only for basic covariates (e.g., age and sex) is insufficient. Important confounders that should be considered include severity of illness scores (e.g., SOFA, APACHE II), comorbidities, renal function and its dynamics, infection severity… Without accounting for these variables, the risk of residual confounding is high, limiting the validity of the conclusions.

- The authors should clarify:

Whether steady-state conditions can reasonably be assumed
Whether dosing regimens remained stable during the period of multiple TDM measurements
How changes in renal function and dose adjustments were handled analytically

Author Response

Please see the attachment "Point-by-Point Response Letter_antibiotics_revised - Reviewer-1"

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

Dear Authors,

Thank you for submitting your manuscript entitled “Vancomycin Exposure Dynamics and Clinical Outcomes in Critically Ill Patients: A Retrospective Cohort Study.” The study addresses an important clinical challenge in optimizing vancomycin therapy for critically ill patients, and your exploration of trough-derived metrics such as TIR and volatility index adds valuable perspective. The work is well structured and provides meaningful data, but several points require clarification and refinement to strengthen the manuscript.

Below I provide 25 specific comments and suggestions for correction or modification:

In the abstract, clarify the definition of volatility index to avoid ambiguity.
Ensure consistency in reporting percentages (e.g., ARC 4.6%) with proper decimal formatting.
The phrase “Goldilocks window” in the abstract should be explained more scientifically.
In the introduction, provide clearer justification for using TIR and volatility index as surrogate measures.
Define ARC consistently with units (mL/min/1.73 m²) at first mention.
Clarify whether AUC/MIC targets were considered in this retrospective design.
In patient selection, specify how sepsis diagnosis was confirmed beyond Sepsis-3 criteria.
Provide more detail on vancomycin assay methodology to ensure reproducibility.
Clarify whether trough sampling times were standardized or varied across patients.
In methods, explain why bootstrap resampling was chosen and how iterations were determined.
In Table 1, correct “vacomycin-susceptible” to “vancomycin-susceptible.”
Ensure all p-values are reported with consistent decimal places (e.g., 0.0061 vs 0.006).
Clarify whether mortality was all-cause ICU mortality or infection-related mortality.
In results, explain why ARC was observed only in younger patients.
Figures should include clearer axis labels (e.g., “Mean vancomycin trough, mg/L”).
In Figure 2, clarify that TIR >15 refers to proportion of troughs exceeding 15 mg/L.
Discuss whether volatility index is clinically interpretable for bedside use.
In Figure 4, add quartile ranges directly on the graph for clarity.
Clarify whether AKI staging was based solely on creatinine or also urine output.
In results, explain why restricted cubic spline did not show non-linear associations.
Discuss potential confounding factors such as concomitant nephrotoxic drugs.
In discussion, emphasize that trough-based metrics are exploratory and not validated.
Correct typographical errors such as “TIR <,>10-15” in Table 1.
Ensure consistency in reporting OR values with confidence intervals across text and figures.
Add a limitations section highlighting retrospective design, missing dosing histories, and heterogeneous sampling.

Author Response

Please see the attachment "Point-by-Point Response Letter_antibiotics_revised - Reviewer-2"

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

This manuscript presents a retrospective cohort study exploring alternative vancomycin exposure metrics (TIR and volatility index) and their association with ICU mortality and AKI in critically ill patients. The topic is clinically relevant, especially given ongoing challenges in implementing AUC-guided monitoring in real-world ICU settings. The manuscript is generally well-structured, and the authors demonstrate awareness of current literature and guideline recommendations. However, several methodological and interpretative issues require clarification and refinement before the manuscript can be considered for publication. The study provides interesting hypothesis-generating findings, but the strength of conclusions should be moderated.

The retrospective, single-center nature of the study, combined with the absence of illness severity indicators (as APACHE II, SOFA), weakens causal interpretation. With ICU mortality reaching 47.7% , higher vancomycin exposure may simply mirror reduced renal clearance in more critically ill patients rather than a direct toxic effect. This aspect needs clearer acknowledgment.
The operational definition of TIR raises conceptual concerns. In this analysis, it reflects the proportion of sampled trough values rather than true time-dependent exposure . This distinction introduces potential bias linked to sampling frequency and may lead to misinterpretation; clearer framing or alternative terminology would improve clarity.
The identification of an apparent “Goldilocks” range (8–12 mg/L) is interesting but remains exploratory. The observed U-shaped pattern could be influenced by confounding and does not necessarily align with established therapeutic targets, particularly in severe infections. A more cautious interpretation is warranted.
The inverse relationship between volatility index and AKI appears counterintuitive. A more plausible explanation relates to the absence of sustained supratherapeutic exposure rather than any protective role of variability itself. This nuance deserves stronger emphasis to avoid overinterpretation.

Author Response

Please see the attachment "Point-by-Point Response Letter_antibiotics_revised - Reviewer-3"