Clinical Determinants of Urinary Podocyte Biomarkers and Their Feasibility in Paraprotein-Related Kidney Disease

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Dear authors,

First of all, I want to congratulate you for this interesting subject. 

Podocyte injury is an established feature of renal disease progression. Podocyte loss is a widely supported hypothesis to explain glomerular damage. Podocyte detachment (podocytopathy) and their presence in the urine (podocyturia) are hallmarks of glomerular disease progression. The detection of podocytes in urine as a biomarker of disease progression is a major advance in the diagnosis and monitoring of glomerular nephropathies. Clinical determinants of urinary podocyte biomarkers (such as podocin and nephrin mRNA/protein) in paraprotein-related renal diseases—including monoclonal gammopathy of renal significance (MGRS) and multiple myeloma (MM)—are linked to structural podocyte injury, with urinary tract infections (UTIs) acting as a significant confounder.

Your study makes an important step to improve knowledge in this subject.

I hope this kit for podocin and nephrin determination will be soon available for the glomerular diseases.

I wish you succes with this paper!

 

Author Response

We express our sincere gratitude to the reviewer for their kind comments.

Reviewer 2 Report

Comments and Suggestions for Authors

This is a clinically relevant pilot study on an understudied topic, and the biopsy-linked subset is a genuine strength.

1. The most important issue is the internal inconsistency of the statistical reporting. In Table 2, the 95% CIs for podocin ELISA versus uPCR and uACR both cross 0, yet the corresponding p-values are reported as significant. In Table 5, the podocin ELISA CI appears entirely above 0, while the p-value is 0.097. These combinations cannot all be correct as written. Please recheck the original outputs, correct the tables, and then harmonize the abstract, Results, and Discussion with the corrected values.

2. The ROC analysis is insufficiently defined and should be technically rechecked. The Methods state that ROC analysis was performed in subjects with plasma cell dyscrasia and albuminuria, whereas Figure 1 is labeled as prediction of “kidney involvement in Multiple Myeloma.” It is therefore unclear what the binary outcome actually was, whether healthy controls were included, and how many subjects entered the ROC model. In addition, AUCs of 0.350 for podocin ELISA and nephrin mRNA raise the possibility of reversed outcome or predictor coding. Please define the positive class explicitly, report the exact sample size used for ROC analysis, and verify whether the low AUCs reflect inverse biomarker direction rather than true worse-than-random discrimination.

3. The UTI finding is potentially important but currently overstated. The text says that both podocin and nephrin mRNA were significantly increased in the presence of UTI, but Table 3 shows p=0.007 for podocin mRNA, p=0.080 for nephrin mRNA, and p=0.293 for podocin ELISA. Based on the table, only podocin mRNA is statistically significant. Also, Table 3 accounts for 84 subjects (65 UTI-negative and 19 UTI-positive), not the full cohort of 86, so missingness should be clarified. Please correct the wording and, ideally, provide a sensitivity analysis excluding UTI-positive samples from the main association and ROC analyses.

4. The statistical plan is difficult to follow in its current form. The Methods state that Pearson correlations with Sidak-corrected p-values were used for associations with uPCR and uACR, but Table 2 presents beta coefficients from nonparametric regression rather than correlation coefficients. It is therefore unclear which analysis is primary and where the reported p-values originate. Please clarify which model generated the values shown in Table 2, and consider simplifying the analysis to a more consistent framework that clinical readers can easily interpret.

5. The cohort is heterogeneous and underpowered for subgroup comparisons. The study combines MM, MGRS, LC amyloidosis, MGUS, and healthy controls, with very small numbers in some categories, especially MGRS. Under these conditions, statements that biomarkers “did not differ” across disease groups are too strong. It would be more accurate to state that no differences were detected in this exploratory cohort. Please also clarify which analyses included healthy controls and which were restricted to diseased participants only.

6. The framing of the manuscript should be better aligned with the data. The title emphasizes “Diagnostic Accuracy,” yet the ROC performance is poor, and the study is explicitly presented as a pilot without a formal sample size calculation. In addition, the abstract and conclusion refer to “prognostic biomarkers,” although the study is retrospective and essentially cross-sectional. I recommend revising the title, abstract, and conclusion to emphasize exploratory biomarker evaluation rather than diagnostic accuracy or prognostic utility.

Author Response

1. The most important issue is the internal inconsistency of the statistical reporting. In Table 2, the 95% CIs for podocin ELISA versus uPCR and uACR both cross 0, yet the corresponding p-values are reported as significant. In Table 5, the podocin ELISA CI appears entirely above 0, while the p-value is 0.097. These combinations cannot all be correct as written. Please recheck the original outputs, correct the tables, and then harmonize the abstract, Results, and Discussion with the corrected values.
   We thank the reviewer for their keen observation and apologize for the confusion, which resulted from insufficient transparency in the original manuscript regarding the number of observations included in each reported statistical test. Table 2 included observations from the full cohort, while table 5 reflects only findings from a subset of individuals with biopsy proven podocytopathia. The fact that the 95%CI in table 5 excluded 0 despite the non-significant p-value is a result of the very low sample size and the discrepancy between the symmetrical "normal approximation" for calculation of the p-value and the non-symmetrical distribution of the 95% bootstrapped percentile CIs. Following the recommendation of the reviewer, we have now clarified this discrepancy in the analysis. However, we can confirm the correctness of the reported results.
2. The ROC analysis is insufficiently defined and should be technically rechecked. The Methods state that ROC analysis was performed in subjects with plasma cell dyscrasia and albuminuria, whereas Figure 1 is labeled as prediction of “kidney involvement in Multiple Myeloma.” It is therefore unclear what the binary outcome actually was, whether healthy controls were included, and how many subjects entered the ROC model. In addition, AUCs of 0.350 for podocin ELISA and nephrin mRNA raise the possibility of reversed outcome or predictor coding. Please define the positive class explicitly, report the exact sample size used for ROC analysis, and verify whether the low AUCs reflect inverse biomarker direction rather than true worse-than-random discrimination.
   We thank the reviewer for this suggestion. In line with this comment we have clarified the title. It now accurately reflects the scope of the analysis. Additionally, we can confirm that both outcome and predictor were coded accurately and the findings truly reflect worse-than random discrimination. The methods section has been adjusted to improve clarity on the way the analysis was performed.
3. The UTI finding is potentially important but currently overstated. The text says that both podocin and nephrin mRNA were significantly increased in the presence of UTI, but Table 3 shows p=0.007 for podocin mRNA, p=0.080 for nephrin mRNA, and p=0.293 for podocin ELISA. Based on the table, only podocin mRNA is statistically significant. Also, Table 3 accounts for 84 subjects (65 UTI-negative and 19 UTI-positive), not the full cohort of 86, so missingness should be clarified. Please correct the wording and, ideally, provide a sensitivity analysis excluding UTI-positive samples from the main association and ROC analyses.
   Once again, we thank the reviewer for their keen observation - the table in the original manuscript contained a numerical error regarding the sample size included in this analysis as the data does indeed represent the whole patient cohort. This error has now been amended. Similarly, the table contained a numerical error for the p-value of nephrin-mRNA which has now been fixed (p=0.080 vs p=0.008). Following the suggestion of the reviewer we have now included a sensitivity analysis excluding subjects with UTIs in the supplement.
4. 4. The statistical plan is difficult to follow in its current form. The Methods state that Pearson correlations with Sidak-corrected p-values were used for associations with uPCR and uACR, but Table 2 presents beta coefficients from nonparametric regression rather than correlation coefficients. It is therefore unclear which analysis is primary and where the reported p-values originate. Please clarify which model generated the values shown in Table 2, and consider simplifying the analysis to a more consistent framework that clinical readers can easily interpret.
   We apologize for the confusion introduced by the methods section in the original manuscript. The paragraph the reviewer is referring to reflects early-stage analysis which was intended to guide more comprehensive statistical analysis. The methods section has been corrected accordingly.

5. The cohort is heterogeneous and underpowered for subgroup comparisons. The study combines MM, MGRS, LC amyloidosis, MGUS, and healthy controls, with very small numbers in some categories, especially MGRS. Under these conditions, statements that biomarkers “did not differ” across disease groups are too strong. It would be more accurate to state that no differences were detected in this exploratory cohort. Please also clarify which analyses included healthy controls and which were restricted to diseased participants only.
   The reviewer is correct in stating that the wording of the corresponding sections in the original manuscript was misleading. These sections have been adjusted in line with the recommendations provided by the reviewer. Additionally, we have improved the clarity of figure captions and introduced n-numbers to the tables where appropriate to improve transparency regarding the way the analyses had been performed.

6. The framing of the manuscript should be better aligned with the data. The title emphasizes “Diagnostic Accuracy,” yet the ROC performance is poor, and the study is explicitly presented as a pilot without a formal sample size calculation. In addition, the abstract and conclusion refer to “prognostic biomarkers,” although the study is retrospective and essentially cross-sectional. I recommend revising the title, abstract, and conclusion to emphasize exploratory biomarker evaluation rather than diagnostic accuracy or prognostic utility.
   We agree with the assessment of the reviewer that the introduction of this wording in the original manuscript can be misleading and the respective sections of the manuscript were corrected accordingly.