Impact of Equation Choice on Models Assessing the Association Between Lipoprotein(a) and Estimated Glomerular Filtration Rate in Adult Patients Without Chronic Kidney Disease

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors The study is well designed but the authors only analyze for the age indicator. Other indicators such as sex, weight, associated diseases, lipid profile are not mentioned even if their influence on renal function is known. Moreover, the age range chosen 60-80 years being those "affected" is a normal range. It is known that eGFR decreases with age. Some interpretations such as old formulas "mask the truth" are too categorical for an observational study. Regarding growth, there is no external validation, sensitivity analyses, evaluation of model performance. The authors mention the selection problem, missing other factors in the model and sample size. The conclusion that CKD-EPI 2021 is the best is forced, there is not enough data to demonstrate this. The article needs to have serious methodological and interpretation.

Author Response

Comments 1: The study is well designed but the authors only analyze for the age indicator. Other indicators such as sex, weight, associated diseases, lipid profile are not mentioned even if their influence on renal function is known.

Response 1: Thank you for pointing this out. We agree with this comment. We have expanded our statistical analysis to include additional clinical confounding factors. We constructed progressive multivariable logistic regression models adjusting for sex, hypertension, diabetes, and a complete lipid profile (Total Cholesterol, LDL, HDL, and Triglycerides). We have also acknowledged the limitation regarding the lack of BMI data due to data constraints. These additions can be found in Section 3.6 (Assessment of additional clinical confounding factors) and Table 4(page number 8, paragraph 3, lines 247-268), as well as the Study Limitations in Section 4 (page number 10, paragraph 5, lines 342-346). "[To isolate the independent effect of Lp(a) on eGFR in the main group... four progressive multivariable logistic regression models were constructed (Table 4)... Certain potential confounding factors, such as body mass index (BMI) were not included in the final regression models due to data constraints, which represents another limitation of our assessment.]"

Comments 2: Moreover, the age range chosen 60-80 years being those "affected" is a normal range. It is known that eGFR decreases with age.

Response 2: Thank you for pointing this out. We agree with this observation and appreciate the opportunity to clarify our methodology. Our approach was predicated on the clinical consensus that an eGFR ≥ 60 mL/min/1.73 m² is considered normal in adults, provided there is no established structural, functional, laboratory, or pathological evidence of kidney damage - in other words, the absence of chronic kidney disease (CKD), which was an explicit inclusion criterion for our cohort (page number 2, paragraph 5, lines 76-77).

Furthermore, we would like to clarify that the range "60-80" in our study strictly denotes the eGFR values (in mL/min/1.73 m²) defining our target group with mildly reduced filtration, rather than the chronological age of the patients.

In the newly added Section 3.2 (Definition of Target and Control Groups and Cohort Distribution According to eGFR Equations) and Table 1, we stratified this normative range into three subgroups (60-80, 80-90, and > 90 mL/min/1.73 m²) based on calculations utilizing the standardized CKD-EPI 2021 equation. Through a series of analyses utilizing the Cochran-Mantel-Haenszel (CMH) test, we statistically defined and validated this core main group. We have revised the text throughout the manuscript to ensure that the appropriate unit of measurement consistently follows these numerical values to prevent any future misinterpretation. These clarifications and methodological additions can be found in Section  Section 3.2 (page number 4, paragraph 2, lines 126-145).

"[Additional inclusion criteria required an eGFR ≥ 60 mL/min/1.73 m2 and the absence of CKD... Initially, the overall cohort was divided into three subgroups (I: 60-80 mL/min/1.73 m²; II: 80-90 mL/min/1.73 m²; III: >90 mL/min/1.73 m²), upon which the CMH method was applied to verify the relationship across Lp(a) quartiles... The dependent variable eGFR was stratified into two groups: a main group with eGFR (60-80) mL/min/1.73 m2 and a control group with eGFR > 80 mL/min/1.73 m2.]"

 

Comments 3: Some interpretations such as old formulas "mask the truth" are too categorical for an observational study.

Response 3: We agree with this comment. We have revised our language throughout the manuscript to ensure a more objective and appropriate tone. The phrase "mask the truth" was replaced with more measured terminology. This change can be found in the Abstract (page number 1, paragraph 1, lines 30-31) and Section 5 (Conclusions) (page number 10, paragraph 6, lines 350-352). "[Abstract: Older equations and the age-embedded EKFC equation may influence regression models, potentially obscuring the true biomarker associations. / Section 5: The underestimation of eGFR by the older equations and by the European consortium's CKD-EKFC [32] may obscure the true pathological association of the investigated parameters.]"

Comments 4: Regarding growth, there is no external validation, sensitivity analyses, evaluation of model performance. The authors mention the selection problem, missing other factors in the model and sample size.

Response 4: Thank you for pointing this out. We agree. To evaluate model performance and test the sensitivity/robustness of our findings against various clinical confounders, we constructed progressive multivariable models (Table 4) (page number 8, paragraph 4, lines 257-263) and performed a Receiver Operating Characteristic (ROC) analysis. We calculated the Area Under the Curve (AUC) for both the baseline model and the fully adjusted model. We also explicitly stated the absence of external validation as a limitation. These additions can be found in Section 3.6 (Figure 5) (page number 9, paragraph 2, lines 269-285) and Section 4 (Study Limitations) (page number 10, paragraph 5, lines 342-343). "[To evaluate the diagnostic performance and discriminatory power... we calculated the Area Under the Receiver Operating Characteristic Curve (AUC)... The baseline model... achieved an AUC of 0.727... the fully adjusted model demonstrated an improved and robust predictive performance, yielding an AUC of 0.732 (Figure 5). / Additionally, the relatively modest sample size and the absence of external validation necessitate cautious interpretation of the findings.]"

Comments 5: The conclusion that CKD-EPI 2021 is the best is forced, there is not enough data to demonstrate this. The article needs to have serious methodological and interpretation.

Response 5: We agree with this comment. We have modified our conclusions to prevent overstating our findings. Instead of claiming it is "the best," we specify that within our investigated cohort, the equation demonstrated the highest precision and reliability when covariates were included. This change can be found in the Abstract (page number 1, paragraph 1, lines 31-34)and Section 5 (Conclusions) (page number 11, paragraph 2, lines 368-371). "[Abstract: CKD-EPI 2021 demonstrated the highest precision, successfully isolating physiological aging and suggesting an independent association between high Lp(a) and mildly reduced eGFR, irrespective of metabolic and vascular confounders. / Section 5: Our comparative analysis demonstrated that the new CKD-EPI 2021 equation provided the most stable and reliable results in patients with a low-normal eGFR within our cohort. It effectively isolates the independent effect of Lp(a) Q4 from age, sex, and a large majority of the established clinical confounders related to kidney function.]"

Comments 6 Figures and tables must be improved

Response 6: We agree with this comment. All figures and tables have been extensively revised and formatted according to the journal's template to enhance their clarity and overall presentation.

Reviewer 2 Report

Comments and Suggestions for Authors

This paper investigates the association between lipoprotein(a) [Lp(a)] and estimated glomerular filtration rate (eGFR) using four eGFR equations (CKD-EPI 2021, CKD-EPI 2009, CKD-MDRD, and CKD-EKFC). The authors report that, compared with CKD-EPI 2021, the other three equations tend to underestimate eGFR. After adjusting for age (dichotomized as >63 years), the association between Lp(a) and eGFR remained statistically significant only for the CKD-EPI 2021. Based on these findings, the authors conclude that the CKD-EPI 2021 equation provides the best precision. However, the presentation, writing, and statistical methodology require improvement.

 

 

1. In the abstract, you do provide any reference for this statement “Previously, we established a strong and statistically significant association between the highest quartile of lipoprotein(a)”.
2. The primary analysis evaluates the association between categorized eGFR (60–80 vs. >80 mL/min/1.73 m²) and categorized Lp(a) (quartiles). Since both variables are categorical, a contingency table analysis (using a chi-square test) would provide a more straightforward assessment of their association. The authors should justify why multivariable logistic regression was chosen as the primary analysis instead of a contingency table analysis, particularly for the unadjusted results.
3. In Table 2, you adjust the age to > 63. Why did you select 63? You did not provide any explanation.
4. lines 42-43, line 65. “glomerular filtration” to “glomerular filtration rate”.
5. The inclusion criteria were age over 40 years. Why is age 40 used as the threshold?
6. At the beginning of the paper, you use Lp(a), whereas in Lines 80–81 you use lipoprotein(a). Please use consistent terminology throughout the manuscript.
7. The abbreviation estimated glomerular filtration rate (eGFR) is defined in Line 38 and then defined again in Line 85. Please define it only once at the first time you used and use eGFR thereafter.
8. The number of patients should be mentioned in Section 2.1. rather than Section 3.2.

 

Author Response

Comments 1: In the abstract, you do provide any reference for this statement “Previously, we established a strong and statistically significant association between the highest quartile of lipoprotein(a)”.

Response 1: Thank you for pointing this out. We agree that the initial phrasing was confusing. We have revised this sentence in the Abstract to read: "In our recent research..." to clarify that it refers to the foundational premise of our current dataset, rather than an external published study requiring a citation. Furthermore, since these foundational findings were not previously published, we have now fully incorporated the baseline data into the current manuscript to provide a complete context (specifically in the newly added Section 3.3) (page number 5, paragraph 3, lines 162-172). As detailed in our response to your following comment (Comment 2), the initial unadjusted analysis establishing this association is now explicitly presented in the text.

[In our recent research, we established a strong and statistically significant association between the highest quartile of lipoprotein(a) [Lp(a)] and mildly reduced estimated glomerular filtration rate (eGFR).]

Comments 2: The primary analysis evaluates the association between categorized eGFR (60–80 vs. >80 mL/min/1.73 m²) and categorized Lp(a) (quartiles). Since both variables are categorical, a contingency table analysis (using a chi-square test) would provide a more straightforward assessment of their association. The authors should justify why multivariable logistic regression was chosen as the primary analysis instead of a contingency table analysis, particularly for the unadjusted results.

Response 2: Thank you for pointing this out. We agree with this comment and have updated the manuscript to include the results of the initial unadjusted non-parametric evaluation using a Chi-square test. We maintained the multivariable logistic regression as the primary tool subsequently to quantify the risk (Odds Ratios and 95% CIs) and to evaluate the independent predictive value of Lp(a) by adjusting for major confounders. This addition can be found in Section 3.3 (page number 5, paragraph 3, lines 162-172). [In the initial unadjusted analysis based on the CKD-EPI 2021 equation, the Chi-square test demonstrated a statistically significant association between Lp(a) quartiles and eGFR categories (Chi-square = 12.5887, P = 0.0056). Specifically, we observed a distinct variation in the distribution of patients with reduced filtration across the ascending Lp(a) groups. Cramer's V of 0.20 indicates a weak to moderate strength of this association.]

Comments 3: In Table 2, you adjust the age to > 63. Why did you select 63? You did not provide any explanation.

Response 3: We agree with this comment. We have added the necessary explanation to clarify that 63 years represents the median age of our study cohort. This clarification can be found in Section 3.5 (page number 7, paragraph 2, lines 222-223). [Age was dichotomized at ≥ 63 years for the regression analysis as this represents the median age of our study cohort (age range 40-87)]

Comments 4: lines 42-43, line 65. “glomerular filtration” to “glomerular filtration rate”.

Response 4: Thank you for pointing this out. We have accordingly revised the text to replace instances of "glomerular filtration" with "glomerular filtration rate". Furthermore, to ensure correct and consistent medical terminology throughout the entire manuscript, we have ensured that after its first definition, the term is strictly replaced with the abbreviation "eGFR".

Comments 5: The inclusion criteria were age over 40 years. Why is age 40 used as the threshold?

Response 5: Thank you for pointing this out. We have added an explanation detailing that this specific biological threshold was selected because the physiological decline in renal function becomes clinically measurable after the fourth decade of life. This change can be found in Section 2.1 (Subjects) (page number 2, paragraph 5, lines 72-76). [Тhis threshold was selected because the physiological, non-linear decline in renal function typically becomes clinically measurable after the fourth decade of life. Beyond this biological threshold, structurally normal kidneys experience a progressive reduction in nephron mass...]

Comments 6: At the beginning of the paper, you use Lp(a), whereas in Lines 80–81 you use lipoprotein(a). Please use consistent terminology throughout the manuscript.

Response 6: We agree with this comment. We have systematically reviewed and revised the text to consistently use the abbreviation "Lp(a)" following its initial definition.

Comments 7: The abbreviation estimated glomerular filtration rate (eGFR) is defined in Line 38 and then defined again in Line 85. Please define it only once at the first time you used and use eGFR thereafter.

Response 7: Thank you for pointing this out. We have removed the duplicate definitions and ensured the abbreviation eGFR is used consistently after being defined for the first time in the Introduction.

Comments 8: The number of patients should be mentioned in Section 2.1. rather than Section 3.2.

Response 8: We agree. We have moved the sample size information to the Subjects section. This change can be found in Section 2.1(page number 2, paragraph 5, lines 69-71). [The participant selection technique involved consecutive sampling of individuals who visited the laboratory and met the inclusion criteria, until the target sample size of 310 volunteers was reached.]

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

The authors have addressed my comments.