Monitoring Water Balance to Predict Hospitalization in Patients with Chronic Heart Failure: A Retrospective Study

Round 1

Reviewer 1 Report (Previous Reviewer 1)

The authors have made appropriate corrections.

Author Response

Response to Reviewer 1

We appreciate constructive criticisms and appropriate comments on our manuscript.

 

Reviewer 2 Report (Previous Reviewer 2)

The authors have largely satisfactorily addressed the issues raised in my original review. A few points remain.

Line 60 also line 91.

InBody infers all body components only from impedance, height and weight,	61
and no statistical correction is performed.

Frankly, I don't understand or believe the company's comment of no statistical correction. I would delete this phrase.

Line 66. Correlation is NOT agreement; variables may be highly correlated but totally different.

 

Line 288 I think that this should be ANCOVA

Author Response

 

 

Author Response File: Author Response.docx

This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.

Round 1

Reviewer 1 Report

This retrospective study assessed that water balance was significantly correlated to the history of hospitalization, 19 and age was a confounding bias in this correlation, regardless of whether patients have been hospitalized with chronic heart failure.

I have concerns about the method, especially selection bias and measurement bias.

Some specific comments that need to be addressed are below:

1. The limitation of the measurement method should include the variability in timing of bioelectrical impedance analysis.

2. What is the effect of diuretics on bioelectrical impedance analysis?

3. Please describe the percentage of HFrEF and HFpEF.

4. The novelty of this research is unclear.

Reviewer 2 Report

The paper is generally well written and clearly describes the study. The following points require consideration.

Line 48. "If a standard value is known," What does this mean? Standard value of what?

Line 49. "BIA can estimate both intracellular water content (ICW) and extra- 49 cellular water content (ECW) separately [11]" This is not strictly correct. BIA is a generic term to cover the general technique. Within BIA, there is SFBIA for single frequency methods, MFBIA for multiple frequency and BIS for spectroscopic methods. SFBIA  is not suitable to estimate water compartments, MFBIA may although BIA theory notes that BIS is required since ECW is predicted at zero frequency and TBW at infinite frequency. This is covered well in many publications particularly those more recent than Kyle (ref 11). The authors must amend the text to reflect this point, particularly since they use an MFBIA device.

Sections 2.3 and 2.4 and Discussion The authors used the InBody device. This device uses proprietary prediction equations of unknown provenance to estimate ECW, ICW etc. Although the precise algorithms are not known they will inevitably invoke either directly or indirectly an assumption that body fluid resistivity is not changed in heart failure or, indeed any condition where there may be differential expansion of a fluid compartment. This is not guaranteed to be the case and will introduce inaccuracy in predictions. The assessment of excess fluid (edema) in HF is analogous to the assessment of excess fluid in lymphedema where impedance has long been used as an assessment method.  BUT in this case it is recognized that it is preferable to assess change  in low frequency (R0) resistance as a proportional index of fluid accumulation rather than to predict volume with its inherent assumptions. ( see 10.1089/lrb.2020.0085). 

The authors should review this literature and include in the Discussion the inherent errors associated with considering volume rather than raw impedance data.

The authors cite Cha et al (ref 18) in support of their use of MFBIA and the InBody in particular. I question this. This publication was before InBody existed. In addition this paper showed good correlations between particular frequencies and water compartments but was silent on accuracy of prediction  limits of agreement analysis or cross validation was not undertaken. Inspection of Fig 1 in this publication shows a very wide scatter of data, again questioning accuracy. There are more recent publications actually attempting to validate the InBody device. Review the literature and source more appropriate references. Also Discuss accuracy of  BIA methods. At present this accepted uncritically.

BIA is a method that requires careful standardization of measurement protocols. No details are provided.

 

Section 3.2. The authors have correlated many variables but where significance was observed the correlations were low.  What were the effect sizes? Notably there were very weak correlations with CHF.

 

Fig 3.3 Although visually there appear to be differences in distribution this has not been analyzed statistically. This must be undertaken. Similarly in Fig 2 the authors found a minor difference in the regression lines but again no statistical analysis. This must be performed, procedures exist for the comparison of regressions - correlation coefficients, intercepts and slopes. The authors use the noted small difference as justification for the ANCOVA. This should only be used if the regressions are indeed significantly different.

Discussion

Statistics. This is likely to require amendment in the light of the new analyses requested above.

 

Line 192. The authors claim significance but I cannot see the supporting analysis.

The authors have failed to consider relevant recent literature, e.g. 10.3904/kjim.2016.358 ; 10.1253/circj.CJ-15-0723; 10.1002/ehf2.13641; and recent reviews  10.1177/10998004221132838 and 10.1016/j.cardfail.2022.08.005

A thorough literature search should be undertaken and the current data placed in the context of existing knowledge.