Prevalence and Clinical Correlates of Iron Deficiency in Vietnamese Patients with Chronic Heart Failure

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This study aimed to determine the prevalence of iron deficiency and identify its independent clinical correlates among patients affected by chronic heart failure.

I have some statistical concerns:

1. Do not use "multivariate" since it refers to multiple outcomes, not multiple covariates; use "multivariable" instead.
2. For the multivariable model, it is necessary to use methods for automated variable selection, such as Akaike Information Criteria (AIC) minimization or LASSO regularization; do not use stepwise, since it is no more a reccomended method which identify a high rate of false positive (Smith G. Step away from stepwise. J of Big Data, vol 5:32, 2018); include in the models all baseline variables; it is unclear how variables for multivariable model were selected, since a lot of other variables are presented in table 1; account for possible overfitting
3. For the multivariable model report c-index and calibration plots or at least hosmer and lemeshow test, in order to assess discrimination and calibration of the model
4. in my opinion, some variables in tables have superfluous number of decimals, ie age and BMI (1 is enough)
5. report all p values with 3 decimals, also in the abstract (some have 4, some 3); in general, be consistent with number of decimals

Author Response

1. Use “multivariable” instead of “multivariate”:

• Response: All instances of “multivariate” have been replaced with “multivariable” throughout the text.

2. Automated variable selection (AIC/LASSO) and avoiding stepwise:

• Response: We have revised our variable selection process. Instead of stepwise methods, we employed Akaike Information Criteria (AIC) minimization to select the final multivariable model, ensuring a more robust selection of covariates.

3. Inclusion of baseline variables and overfitting:

• Response: We have clarified that variables were selected based on clinical relevance (Table 1) and AIC values to prevent overfitting while maintaining a sufficient ratio of events per variable.

4. Model performance (c-index and Hosmer-Lemeshow test):

• Response: We have included the C-index (0.81) and the Hosmer-Lemeshow test (p = 0.42) in the Results section to demonstrate the discrimination and calibration of our model.

5. Decimal consistency (Age, BMI, p-values):

• Response: Age and BMI are now reported with one decimal place. All p-values have been standardized to three decimal places.

Reviewer 2 Report

Comments and Suggestions for Authors

I enjoied reading this paper for several reasons including the quality of english, the clarity of presentation of the results based upon few but still comprehensive parameters, the soundness of statistical analysis, the depth and conciseness of discussion and the delineation of the unavoidable limits the study as all studies do.

At line 160 a typo awaits amendement (ptext)  

Author Response

1. Typo at line 160 (ptext):

• Response: The typo has been corrected to “p-value”.

Reviewer 3 Report

Comments and Suggestions for Authors

The study is very simplistic, starting from a subject already known and studied.

It does not bring any new element, it is purely statistical.

The correlations are limited to NT pro BNP and CRP

It would have been interesting to have correlations with the etiology of heart failure, with associated comorbidities, with renal function and especially with certain treatments.

Also, the impact of iron deficiency on the evolution of patients with HF, respectively the more frequently associated complications are not mentioned.

Many phrases need to be reformulated especially the title because they cannot provide the essential ideea.

I have marked some innaccuracies into the text

The article has a low scientific value.

 

Comments for author File: Comments.pdf

Comments on the Quality of English Language

english language must be improved

Author Response

1. Novelty and clinical value:

• Response: While ID in CHF is known, data from Vietnamese tertiary care settings is scarce. Our finding of a 75.4% prevalence—nearly ubiquitous—highlights a major gap in current Southeast Asian diagnostic protocols.

2. Correlations with etiology and comorbidities:

• Response: We acknowledge this. While the primary focus was on hemodynamic/inflammatory markers, we have added a brief discussion regarding the high prevalence of anemia (87.7%) and its relationship with ID in this specific population.

3. Impact on evolution/complications:

• Response: As a cross-sectional study, we are limited in assessing “evolution”. We have added this as a clear limitation and a call for future longitudinal research in Vietnam.

Round 2

Reviewer 3 Report

Comments and Suggestions for Authors

I have reanalyzed the new version of the submitted manuscript and compared the 2 versions.
I have not noticed any significant differences in terms of content and especially the discussions related to this topic.
The changes made are more of a form than a substance
I do not believe that the current form has any extra things that would allow the previous decision to be changed.

Comments on the Quality of English Language

english language must be improved