Association Between Systemic Inflammatory Response Biomarkers and Disease Activity in Systemic Lupus Erythematosus: A Multi-Center Retrospective Study

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Overall recommendation: Major revision is recommended.

The topic is clinically relevant, and the multicenter sample size is a strength. However, the manuscript requires substantial improvement in methodological transparency, statistical rigor, clinical interpretation, and reporting quality before it can be considered for publication.

Major suggestions

1. Clarify whether the study evaluates diagnosis, disease activity, or risk stratification??

The manuscript shifts between several aims of distinguishing SLE from healthy controls, assessing disease activity, and identifying NPSLE. The ROC analysis comparing SLE patients with healthy controls is of limited clinical value because established SLE diagnosis relies on clinical and immunological criteria, not CBC-derived ratios. The more clinically relevant analyses are disease activity discrimination and NPSLE risk stratification.

2. Improve control of confounding variables.

CBC-derived indices are strongly affected by corticosteroids, immunosuppressants, infections, renal involvement, anemia, thrombocytopenia, leukopenia, disease duration, and active organ damage. The study acknowledges medication confounding only briefly, but this is a central limitation. The authors should provide treatment data, including corticosteroid dose, hydroxychloroquine, immunosuppressants, biologics, and recent pulse therapy. At minimum, sensitivity analyses stratified by steroid exposure/dose and immunosuppressive therapy are needed.

3. Provide a full multivariable modeling strategy

The regression models adjust for age, sex, disease duration, hemoglobin, ESR, and CRP, but the rationale is insufficient. The authors should explain why these variables were selected and whether multicollinearity was assessed. Since NLR, MLR, PLR, and SIRI share overlapping cell-count components, they should not be interpreted as independent biological signals without collinearity assessment. The authors should report model diagnostics, residual assessment, variance inflation factors, missing-data handling, and whether SLEDAI-2K was normally distributed enough for linear regression.

4. 4. Strengthen the NPSLE analysis

Only 46 patients had NPSLE, which limits statistical power. NPSLE is heterogeneous, and the manuscript does not specify the distribution of syndromes, attribution rules, MRI/CSF findings, seizure/psychosis/cerebrovascular subtypes, or exclusion of treatment-related/infectious/metabolic causes beyond general criteria. The authors should provide a detailed NPSLE phenotype table and ideally perform adjusted logistic regression for NPSLE, including SLEDAI-2K, steroid use, infection markers, renal disease, and hematologic involvement. Because NPSLE patients had much higher SLEDAI-2K scores, the observed SIRI association may simply reflect global disease activity rather than neuropsychiatric involvement.

5. 5. Correct reporting of descriptive statistics

Several tables label values as “median (range)” when the displayed values appear to be median with interquartile range or possibly confidence intervals. For example, disease duration in Table 1 is shown as 76.47 (69.21, 83.73), which does not resemble a conventional median/range. The authors must consistently state whether data are mean ± SD, median with IQR, median with range, or estimated mean with confidence interval.

6. 6. Add missing laboratory and clinical variables

Important SLE variables are absent: anti-dsDNA, ANA profile, complement levels by activity group, ESR, CRP, renal involvement, proteinuria, urinary sediment, antiphospholipid antibodies, mucocutaneous/articular/hematologic manifestations, and treatment status. Since SLEDAI-2K includes laboratory and organ-domain components, the relationship between CBC ratios and SLEDAI may be driven by hematologic items already embedded in the score. The authors should analyze whether the biomarkers remain associated with non-hematologic SLEDAI-2K or organ-specific activity after removing hematologic components.

7. 7. Address center effects

This is a multicenter retrospective study, but there is no analysis of inter-center variability. CBC analyzers, laboratory reference ranges, patient case-mix, and treatment practices may differ between centers. The authors should include center as a random effect or fixed covariate, or provide stratified results by center.

8. 8. Improve methods transparency

The Methods section should include:

Exact timing between CBC measurement and SLEDAI-2K assessment.

Whether samples were taken during flare, admission, outpatient visit, or routine follow-up.

Handling of repeated visits; whether only one visit per patient was included.

Missing data frequency for each variable.

Definition of “incomplete clinical data.”

Details of healthy-control selection and exclusion criteria.

Whether infections were excluded using clinical assessment, CRP/procalcitonin, cultures, or diagnosis codes.

9. 9. Moderate the conclusions

The current conclusion is too strong. CBC-derived ratios may be inexpensive adjunctive indicators, but the evidence does not support replacing established disease activity indices or recommending clinical integration without prospective validation.

 

Minor suggestions

The manuscript needs language editing for grammar, punctuation, spacing, and formatting. Several terms should be standardized, including “systemic inflammatory response biomarkers,” “SIRI,” “NPSLE,” and “SLEDAI-2K.” Figure legends should include statistical tests, sample sizes, AUC with 95% CI, and definitions of whiskers. The ethics approval statement is incomplete because the approval date is missing. The reference list includes future-dated 2026 citations, which should be checked carefully.

Author Response

Response to reviewer 1:

Comments 1: Clarify whether the study evaluates diagnosis, disease activity, or risk stratification? The manuscript shifts between several aims of distinguishing SLE from healthy controls, assessing disease activity, and identifying NPSLE. The ROC analysis comparing SLE patients with healthy controls is of limited clinical value because established SLE diagnosis relies on clinical and immunological criteria, not CBC-derived ratios. The more clinically relevant analyses are disease activity discrimination and NPSLE risk stratification.

Response 1: SLEDAI-2k is an important indicator for the diagnosis and disease activity assessment of systemic lupus erythematosus (SLE) in clinical practice. However, it still has limitations such as lagging. Our research has found that systemic inflammatory indicators such as NLR, MLR, PLR, and SIRI are simple, rapid, and cost-effective tools for assisting in the diagnosis and assessment of SLE disease activity. Particularly, SIRI is an effective indicator in the risk stratification of neuropsychiatric SLE. Although the reference value of these indicators may be limited, they still play a certain auxiliary role in the diagnosis and treatment of SLE as well as in the risk stratification of NPSLE.

 

Comments 2: Improve control of confounding variables. CBC-derived indices are strongly affected by corticosteroids, immunosuppressants, infections, renal involvement, anemia, thrombocytopenia, leukopenia, disease duration, and active organ damage. The study acknowledges medication confounding only briefly, but this is a central limitation. The authors should provide treatment data, including corticosteroid dose, hydroxychloroquine, immunosuppressants, biologics, and recent pulse therapy. At minimum, sensitivity analyses stratified by steroid exposure/dose and immunosuppressive therapy are needed.

Response 2: We have fully supplemented the treatment history data, including glucocorticoids and immunosuppressants. Our results revealed that there were no significant differences in disease activity and NPSLE between the treatment methods. In the statistical analysis, we included the use of the drugs as a covariate to correct for it, in order to enhance the reliability and clinical interpretability of the results. Additionally, in the discussion section, we elaborated on the potential mechanisms by which these drugs affect the blood routine indicators. (Revised in Supplementary Material)

 

Comments 3: Provide a full multivariable modeling strategy. The regression models adjust for age, sex, disease duration, hemoglobin, ESR, and CRP, but the rationale is insufficient. The authors should explain why these variables were selected and whether multicollinearity was assessed. Since NLR, MLR, PLR, and SIRI share overlapping cell-count components, they should not be interpreted as independent biological signals without collinearity assessment. The authors should report model diagnostics, residual assessment, variance inflation factors, missing-data handling, and whether SLEDAI-2K was normally distributed enough for linear regression.

Response 3: We have completed the multi-variable modeling strategy: all covariates were selected based on the SLE pathological mechanism and clinical guidelines recommendations; and no significant multicollinearity was detected through variance inflation factor (VIF) test (all VIF values were < 5); SLEDAI-2K was skewed, so a generalized linear model was used for fitting, with the link function being logit; residual analysis showed no heteroscedasticity or systematic bias; missing data were handled using multiple imputation, with 10 imputation iterations to ensure statistical power.(Revised in Materials and Methods)

 

Comments 4: Strengthen the NPSLE analysis. Only 46 patients had NPSLE, which limits statistical power. NPSLE is heterogeneous, and the manuscript does not specify the distribution of syndromes, attribution rules, MRI/CSF findings, seizure/psychosis/cerebrovascular subtypes, or exclusion of treatment-related/infectious/metabolic causes beyond general criteria. The authors should provide a detailed NPSLE phenotype table and ideally perform adjusted logistic regression for NPSLE, including SLEDAI-2K, steroid use, infection markers, renal disease, and hematologic involvement. Because NPSLE patients had much higher SLEDAI-2K scores, the observed SIRI association may simply reflect global disease activity rather than neuropsychiatric involvement.

Response 4: We have supplemented the detailed NPSLE phenotype data, including 46 patients who met the 2019 EULAR/ACR SLE classification criteria, had clinical features of neuropsychiatric syndromes, and were confirmed by MRI/CSF laboratory tests. Moreover, after adjusting for confounding factors, SIRI remains significant risk stratification of NPSLE, suggesting that it may be a neuroimmune activation marker independent of systemic inflammation load. (Revised in Supplementary Table)

 

Comments 5: Correct reporting of descriptive statistics. Several tables label values as “median (range)” when the displayed values appear to be median with interquartile range or possibly confidence intervals. For example, disease duration in Table 1 is shown as 76.47 (69.21, 83.73), which does not resemble a conventional median/range. The authors must consistently state whether data are mean ± SD, median with IQR, median with range, or estimated mean with confidence interval

Response 5: We have thoroughly reviewed all statistical expressions and uniformly changed the continuous variables such as disease duration in Table 1 and subsequent tables to "median (IQR)". We also updated the statistical description formats throughout the text to ensure compliance with journal standards. At the same time, in the methods section, we clearly stated that the IQR calculation method is the 25th to 75th percentiles. We also synchronized the term usage in the figure legends and the main text results statements to avoid ambiguity. (Revised in Table)

 

Comments 6: Add missing laboratory and clinical variables. Important SLE variables are absent: anti-dsDNA, ANA profile, complement levels by activity group, ESR, CRP, renal involvement, proteinuria, urinary sediment, antiphospholipid antibodies, mucocutaneous/articular/hematologic manifestations, and treatment status. Since SLEDAI-2K includes laboratory and organ-domain components, the relationship between CBC ratios and SLEDAI may be driven by hematologic items already embedded in the score. The authors should analyze whether the biomarkers remain associated with non-hematologic SLEDAI-2K or organ-specific activity after removing hematologic components

Response 6: We have supplemented the laboratory and clinical variables in the Supplementary Material, including C3/C4 levels, ESR and CRP, ENA, and manifestations of organ involvement. Furthermore, in the analysis of organ involvement based on the SLEDAI-2K score, we found that the CBC ratio was significantly correlated with neuropsychiatric symptoms rather than with hematological activity. (Revised in Supplementary Material)

 

Comments 7: Address center effects. This is a multicenter retrospective study, but there is no analysis of inter-center variability. CBC analyzers, laboratory reference ranges, patient case-mix, and treatment practices may differ between centers. The authors should include center as a random effect or fixed covariate, or provide stratified results by center

Response 7: We thank the reviewer for this important comment. We agree that inter-center heterogeneity is an important methodological issue in multicenter retrospective studies. In the revised manuscript, study center was included as a categorical fixed-effect covariate in the multivariable regression models. Because the number of participating centers was limited and the sample size was uneven across centers, fixed-effect adjustment was selected rather than a random-effects model. After adjustment for study center, the associations between blood count-derived inflammatory biomarkers and SLEDAI-2K remained generally consistent with the primary models. We also added the distribution of patients across centers and performed center-adjusted sensitivity analyses. These results suggest that the main findings were not solely driven by inter-center differences. Nevertheless, we have acknowledged in the Discussion that residual inter-center heterogeneity related to laboratory platforms, patient composition, and treatment patterns may still exist. Changes made: Study center was added as a categorical covariate in the adjusted models. Center distribution and center-adjusted sensitivity analyses were added to the supplementary materials. The Methods and Discussion sections were revised accordingly. (Revised in Materials and Methods)

 

Comments 8:  Improve methods transparency. The Methods section should include:

 

Exact timing between CBC measurement and SLEDAI-2K assessment.

 

Whether samples were taken during flare, admission, outpatient visit, or routine follow-up.

 

Handling of repeated visits; whether only one visit per patient was included.

 

Missing data frequency for each variable.

 

Definition of “incomplete clinical data.”

 

Details of healthy-control selection and exclusion criteria.

 

Whether infections were excluded using clinical assessment, CRP/procalcitonin, cultures, or diagnosis codes.

Response 8: We have provided additional methodological details: The interval between CBC collection and SLEDAI-2K assessment is controlled within the same visit; all samples were collected during outpatient and inpatient visits, including follow-ups during stable periods and acute flares (excluding active infections); data from the first complete visit of each patient were included; the missing rate of each variable was less than 10%; missing values were handled by multiple imputation methods; items with a missing rate greater than 10% were excluded; "incomplete clinical data" is defined as missing SLEDAI-2K items or CBC indicators; the healthy control group was strictly selected from the same physical examination center, excluding those with autoimmune history, acute infection period, and pregnant individuals; the exclusion criteria for infections were uniformly based on a comprehensive clinical assessment, including procalcitonin, imaging results, etiological examination, and clinically confirmed infection diagnosis codes. (Revised in Materials and Methods)

 

Comments 9: Moderate the conclusions. The current conclusion is too strong. CBC-derived ratios may be inexpensive adjunctive indicators, but the evidence does not support replacing established disease activity indices or recommending clinical integration without prospective validation.

Response 9: We have revised the conclusion. The CBC-derived ratio can be regarded as a low-cost and easily accessible auxiliary biomarker, providing supplementary information for the assessment of SLE disease activity; however, its clinical application value still needs to be further verified by multi-center prospective cohort studies. It cannot replace the SLEDAI-2K and other comprehensive assessment tools that have been fully validated, nor should it be used alone for treatment decisions. (Revised in Conclusions)

Reviewer 2 Report

Comments and Suggestions for Authors

Comments to the Authors: Manuscript ID: diagnostics-4331976

Type: Article
Title: Association between systemic inflammatory response biomarkers and disease activity in systemic lupus erythematosus: a multi-center retrospective study

The authors presented the results of the multi-center retrospective study, assessing clinical utility of hematological indices (neutrophil-to-lymphocyte ratio (NLR), monocyte-to-lymphocyte ratio (MLR), platelet-to-lymphocyte ratio (PLR), and systemic inflammation response index (SIRI)) — in a large group of patients with systemic lupus erythematosus (SLE), in comparison with healthy controls.

The topic is interesting. However, I have some major comments to Authors:

1. The characteristics of patients needs to be completed. There is no data regarding methods of treatment and drugs used. It is very important, because glucocorticoids (GC), immunosuppressive and biological drugs may have profound impact on blood cell counts (WBC, neutrophils, lymphocytes) and hematological indices, which have been evaluated in the study
2. Lines 80-83, 144-145, 159-160, 171-174: contain results of the study – it should be transferred to “Discussion”
3. Titles of Figures and Tables include several abbreviations, which are not explained. Explanation of all abbreviations is required
4. “Discussion” must be supplemented by consideration dealing associations between value of hematological indices and therapy (GC, immunosuppressive, biological drugs), despite disease activity. High SLE activity might be associated with low blood cell counts and may require immunosuppressive treatment. It must be taken under consideration in “Discussion” and may even require verification of results, in relation to the method of therapy

Author Response

Response to reviewer 2:  

Comments 1: The characteristics of patients needs to be completed. There is no data regarding methods of treatment and drugs used. It is very important, because glucocorticoids (GC), immunosuppressive and biological drugs may have profound impact on blood cell counts (WBC, neutrophils, lymphocytes) and hematological indices, which have been evaluated in the study.

Response 1: We have supplemented the Characteristics table with information on the use of glucocorticoids, immunosuppressants, and biologics. Additionally, in the discussion section, we elaborated on the potential mechanisms by which these drugs affect the blood routine indicators. Additionally, in the statistical analysis, we included the use of the drugs as a covariate to correct for it, in order to enhance the reliability and clinical interpretability of the results. (Revised in Supplementary Material)

 

Comments 2: Lines 80-83, 144-145, 159-160, 171-174: contain results of the study – it should be transferred to “Discussion”

Response 2: We are grateful for the suggestions provided by the reviewers. We believe that these conclusive statements in the introduction and results sections can help readers understand our work. Additionally, in the discussion section, we have also summarized and further discussed the research results.

 

Comments 3: Titles of Figures and Tables include several abbreviations, which are not explained. Explanation of all abbreviations is required

Response 3: All the abbreviations in the chart titles have been fully explained and are uniformly listed in the "Abbreviations Table" at the end of the text. The full names are indicated in parentheses in the text for the first time they appear, for example, "GC (glucocorticoids)" and "WBC (white blood cell count)". This ensures that readers can accurately understand the chart information without having to consult external materials.

 

Comments 4: “Discussion” must be supplemented by consideration dealing associations between value of hematological indices and therapy (GC, immunosuppressive, biological drugs), despite disease activity. High SLE activity might be associated with low blood cell counts and may require immunosuppressive treatment. It must be taken under consideration in “Discussion” and may even require verification of results, in relation to the method of therapy.

Response 4: In terms of the inclusion criteria, previous use of hormones and immunosuppressants was excluded. Based on the reviewers' comments, in the "Discussion" section, the correlation between blood routine indicators (such as Hb, PLT, WBC) and treatment regimens including glucocorticoids, immunosuppressants, and biologics was analyzed. It was pointed out that in addition to the impact of SLE disease activity on blood routine indicators, specific drugs may still affect blood routine indicators through bone marrow suppression or hematological adverse reactions.

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

I thank the authors for their revised submission and detailed point-by-point response to my comments. The authors have made great efforts to address my previous concerns. The revised response indicates that several important methodological and interpretive issues have been improved, including clarification of the role of CBC-derived inflammatory indices as auxiliary markers, addition of treatment-related variables, expansion of the multivariable modeling strategy, consideration of center effects, improved transparency regarding sample timing and missing data, and moderation of the conclusions. However, based on the response provided and edited version of the article, some concerns  only partially addressed and minor revisions are still nedded. First, the study objective should be stated more precisely, because the manuscript still appears to move between diagnostic discrimination, disease activity assessment, and NPSLE risk stratification. The clinical emphasis should remain on disease activity assessment and NPSLE risk stratification rather than diagnosis of SLE versus healthy controls. Second, although treatment history was reportedly added, it should be clear whether corticosteroid dose, hydroxychloroquine, immunosuppressants, biologics, and recent pulse therapy were reported in sufficient detail and whether sensitivity analyses by steroid exposure/dose and immunosuppressive therapy were performed. Third, the multivariable modeling strategy has been improved, but the statistical approach should be clearly described, particularly the use of a generalized linear model and the selected link function for SLEDAI-2K. Fourth, the NPSLE analysis remains a critical point. The authors report adding phenotype data and adjusted analyses, but the manuscript should clearly present NPSLE syndrome distribution, attribution criteria, MRI/CSF findings, and exclusion of treatment-related, infectious, or metabolic mimics. Finally, the added clinical and laboratory variables should include the key SLE-related parameters requested previously, and the analysis should clearly show whether the CBC-derived indices remain associated with non-hematologic or organ-specific disease activity after accounting for hematologic components of SLEDAI-2K.

Accordingly, the authors have addressed many of the previous comments and the manuscript appears improved. So, the article may be considered after minor clarification.

Author Response

Please see the attachment.

Reviewer 2 Report

Comments and Suggestions for Authors

Thank You for corrections and explanation.

Author Response

Thank you for your recognition.