Review Reports

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The presented article entitled "Hydroxyurea Absorption Phenotype: A Key PK/PD Determinant in Sickle Cell Disease Treatment" is quite interesting, however required improvement following given suggestions.

1. Suggested to expand the discussion on study limitations and generalizability; the manuscript acknowledges the small pediatric cohort (n=22 patients, 44 PK profiles) implicitly through conservative covariate thresholds, but lacks a dedicated limitations section. Explicitly discuss how the limited sample size may limit statistical power for detecting subtle covariate effects (e.g., cystatin C, creatinine) and PK-PD correlations, particularly the non-significant HbF trend (p=0.0835). Address generalizability beyond European pediatric SCD patients with normal renal function to adults, diverse ethnicities, or those with comorbidities, as the cohort includes only 2 adults and normal eGFR.
2. Suggested to strengthen PK-PD relationships with additional analyses; while phenotype-group differences in MCV (p=0.0001), reticulocytes (p=0.01), and HbF trend are reported, individual PK parameters (ka, Cl/F, Vd/F) show no correlations with hematological endpoints (r²<0.16). Perform subgroup simulations using the final popPK model to quantify how absorption phenotype impacts predicted HbF response or MTD attainment. Clarify mechanistic hypotheses for phenotype switch (2 patients rapid-to-slow) and neutrophil lack of difference, perhaps linking to infections or genetics.
3. Suggested to enhance model justification and external validation; alternative absorption models (lag-time, transit) were tested but discarded due to imprecision; however, minor pcVPC deviations persist at early times (10 min) and late tails (>5h for slow profiles). Provide eta-shrinkage plots and VPC binning details to justify the 1-compartment first-order model. Compare parameter estimates quantitatively to prior popPK studies in a summary table, and discuss bootstrap stability (bias <15%) alongside need for external validation in larger, multi-center datasets.
4. Suggested to Improve clinical translation and dosing recommendations; the conclusions advocate phenotype integration into PK-guided dosing, but provide no practical algorithm (e.g., phenotype-based ka adjustment in Bayesian forecasting for target AUC 115 mg·h/L). Derive and simulate dose adjustments for rapid vs. slow absorbers to achieve equivalent Cmax or AUC, emphasizing early Tmax phenotyping via sparse C2h monitoring. Discuss integration with ongoing OPTIMDREP outcomes.
5. Suggested to refine methods and reporting for reproducibility; a detail GC-MS validation (calibration range 0.79-100 mg/L, but precision/accuracy?); specify GraphPad Prism NCA assumptions for λz estimation in short 6h profiles. Report full covariate screening table (OFV drops, LRT p-values) and make anonymized data/model code available (beyond "upon request"). Standardize Tmax cutoff rationale (strictly <1h vs. literature 15-30/60-120 min).

 

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

In the Abstract, aim of the study is not clear and should be written more precisely. Also, briefly specify design elements of the trial (e.g. sample size, age range, dosing scheme). This should also be done in the M&M section (it is currently explained only in the supplementary material). Next, saying that the population is “Predominantly pediatric” is not precise enough. Please, provide median age, range, or percentage of pediatric vs adult participants.

The introduction is too short and lacks a clearly defined knowledge gap. The rationale for the selected PD biomarkers is not introduced. Add clear aim at the end of the Introduction section.

In the M&M section - It is not clear on what basis this definition of MTD was established. There is no explanation or reference provided for the selected neutrophil and reticulocyte thresholds.

Blood sampling - Centrifugation conditions are missing. Please, add details on speed, duration, temperature, and type of centrifuge. It is also not clear what type of collection tubes/anticoagulant was used (EDTA, heparin). Duration of storage before analysis is also not reported.

GC-MS - there is no information on accuracy, precision, recovery values, calibration - the number of calibration points, concentration range distribution. What was the number of replicates? Source of blank plasma is not specified.

In the whole M&M section methods should consistently use past tense (“was added,” “were centrifuged”), but present tense is sometimes used throughout the section (GC-MS subsection).

Statistical analysis section is written too vague and it is currently non-informative. Listing software does not describe what analyses were performed. Also, which software was used for what exactly?

Comments on the Quality of English Language

In the whole M&M section methods should consistently use past tense (“was added,” “were centrifuged”), but present tense is sometimes used throughout the section (GC-MS subsection).

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

Current manuscript presents population pharmacokinetic (PK) analysis of hydroxyurea (HU) in patients with sickle cell disease (SCD), primarily pediatric, using data from the OPTIMDREP clinical trial. The study uses non-compartmental analysis followed by nonlinear mixed-effects modeling to quantify PK parameter variability and identify covariates influencing HU disposition and hematological response. The main findings include the identification of two distinct absorption phenotypes (rapid vs. slow) as a key covariate explaining approximately half of the interindividual variability in the absorption rate constant, with rapid absorbers showing superior erythropoietic response (higher MCV, lower reticulocytes, trend toward higher HbF %) without increased myelosuppression. The study's strengths include the intensive PK sampling design, rigorous modeling approach and novel identification of absorption phenotype as a predictor of hematological response.

However, after examining the manuscript I have following questions and suggestions:

1. The absorption phenotype (rapid vs. slow) was defined based on Tmax from the same concentration-time data used for model building, then tested as a covariate on the absorption rate constant (ka). This creates potential circularity. Please, validate the absorption phenotype as an independent covariate using an alternative approach, or clearly state this as a limitation and propose prospective validation in a future study.
2. Table 3 shows no significant correlations between any PK parameters and any PD endpoints. Yet the authors conclude that rapid absorbers show superior hematological response based on group comparisons. This discrepancy suggests the absorption phenotype may capture something beyond measurable PK parameters. Please, address this paradox directly in the discussion.
3. Small sample size (n=22) for population PK modeling with covariate identification. The risk of overfitting and type I error is existing. Please, provide a power analysis or simulation-based justification that this sample size is adequate.
4. Two patients initially rapid at V0 developed slow profiles at V5. This raises questions about whether absorption phenotype is a stable patient characteristic or varies with time, dose or other factors. Please, discuss whether phenotype instability limits the clinical utility of a single assessment.
5. MCV is interpreted as a marker of HU adherence, but no direct adherence measures (pill counts, electronic monitoring, pharmacy refill records) were collected. Please, acknowledge this limitation explicitly in the discussion.
6. The prediction-corrected visual predictive checks show "slight underestimation of absorption predictions for rapid profiles", indicating the model may not adequately capture rapid absorption. Please, explore alternative structural absorption models, quantify the magnitude of underestimation and discuss its clinical significance.
7. Genetic factors can confuse or interact with absorption phenotype. Please, acknowledge the absence of pharmacogenetic data as a limitation.
8. Lines 75-76. “The study OPTIMDREP was designed to identify the most effective approach to reduce the time to reach the MTD of HU (Nazon et al, submitted).” Please provide the published reference or remove the citation and briefly describe the trial design without citing an unpublished work.
9. The authors screened multiple covariates without explicit adjustment for multiple comparisons. Please, report the number of covariates tested. Provide the base model IIV values to show the magnitude of reduction achieved by each retained covariate.
10. Please, either exclude 2 adult patients and re-run the analysis on the pediatric-only cohort (n=20) to confirm findings, or include age as a categorical covariate (pediatric vs. adult) in the covariate screening and report whether it significantly affects any PK parameter.
11. The target AUC used for PK-guided dosing is cited from Dong et al., but it is unclear whether this target was derived from adult or pediatric data.
12. The conclusion "without increased myelosuppression" is based only on neutrophil counts. Other markers of myelosuppression (platelets, hemoglobin, etc.) or non-hematologic toxicity were not reported.

The manuscript presents interesting and potentially clinically relevant findings regarding absorption phenotype as a determinant of HU response in SCD. However, several methodological concerns must be addressed during major revision before publication. With these revisions, the manuscript would provide a valuable contribution to the literature on HU personalized dosing in SCD. Without addressing the fundamental concern of circular covariate selection, the conclusions regarding absorption phenotype as an independent predictor remain questionable.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The authors have reflected all the said suggestions and comments, which made the manuscript enhanced with improved readability; Thus, I suggest for further consideration with acceptance.

Author Response

Final comment: The authors have reflected all the said suggestions and comments, which made the manuscript enhanced with improved readability; Thus, I suggest for further consideration with acceptance.

Final response: We sincerely thank the reviewer for the thorough and constructive review process and for recommending this manuscript for acceptance. We are grateful for the time and expertise dedicated to evaluating our work, and we believe that the successive rounds of revision have substantially strengthened the manuscript. We hope that this work will make a meaningful contribution to the literature on personalized hydroxyurea dosing in pediatric sickle cell disease.

Reviewer 3 Report

Comments and Suggestions for Authors

The authors have provided a revised manuscript, supplementary materials, and a detailed point-by-point response to my previous comments. The revised manuscript is significantly improved in transparency and scientific rigor. The core findings remain novel and clinically relevant.

However, one critical issue remains unresolved before acceptance.

-- Line in main manuscript: "Supplementary Materials: No supplementary data" is now incorrect, as supplementary materials exist.
- Table 4. The row for "Present model" shows "20 children and 2 adults; 34.9 kg (typical)". The notation "20 children" is fine, but the typical body weight of 34.9 kg is presented without context. Clarify that this is the median weight used for normalization.
- Please, also correct Figure 1 legend from "T1/2" to "Tmax".

Once these corrections are made, the manuscript will be suitable for publication. The scientific content is sound and the work provides a valuable contribution to the literature on personalized HU dosing in SCD.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf