Comparative Effectiveness of Cycling Versus Swapping Strategies After Advanced Therapy Failure in Axial Spondyloarthritis: A Real-World Retrospective Study

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This study titled “Comparative effectiveness of cycling versus swapping strategies after advanced therapy failure in axial spondyloarthritis: a real-world retrospective study” presents an interesting subject, but there are some inquiries:

Major Comments

1. Clarify the Definition of Treatment Failure
   The authors define failure as discontinuation due to primary/secondary inefficacy or adverse events, but exclude patient choice or pregnancy. While this is reasonable, it introduces potential selection bias. Please clarify how many patients were excluded for these reasons and discuss the impact on generalizability.
2. Address the Imbalance Between Groups
   The Cycling Group (CG) and Swapping Group (SG) differ significantly in disease duration, psoriasis prevalence, IBD prevalence, and prior lines of therapy. Although multivariable adjustment was performed, residual confounding remains a concern. Consider adding propensity score matching as a sensitivity analysis to strengthen causal inference.
3. Clarify the Definition of "Line of Therapy"
   Line of therapy is determined chronologically, but it is unclear whether lines are defined by drug class changes or by sequential treatments regardless of mechanism. Please specify how lines were counted, particularly when patients switched between agents of the same class.
4. Explain the Handling of Multiple Treatment Courses
   The analysis includes 343 treatment courses from 156 patients, meaning some patients contributed multiple observations. Please clarify how within-patient correlation was accounted for in the statistical analysis (e.g., robust standard errors or frailty models).
5. Justify the Use of Treatment Persistence as an Effectiveness Proxy
   Drug survival is influenced by many factors beyond efficacy, including tolerability, access, and provider prescribing habits. Please discuss the limitations of using persistence as a surrogate for effectiveness and consider adding a sensitivity analysis restricted to discontinuations due to inefficacy.
6. Clarify the Distribution of Drug Classes in the Swapping Group
   The SG includes 40 TNFi courses (30.8%), which is notable given that SG is defined as switching to a different mechanism of action. Please clarify whether these represent patients switching from a non-TNFi to a TNFi, and discuss the clinical rationale.
7. Address the High Proportion of Missing Failure Reasons
   Reasons for failure were documented for only 58.3% of treatment courses. This missingness could bias the interpretation of strategy effectiveness. Please discuss how this limitation was addressed and consider multiple imputation for this variable.
8. Interpret the Prescription Year Finding with Caution
   The finding that more recent prescription year is associated with higher discontinuation risk is intriguing but could reflect channeling of sicker patients to newer agents or changes in clinical practice. Please discuss potential explanations and whether this finding persisted after adjusting for line of therapy.
9. Clarify the Handling of Biosimilars
   Biosimilars are grouped with originator biologics for MOA classification. This is acceptable, but please clarify whether biosimilars were considered separate treatment courses or whether switching between an originator and its biosimilar was counted as a new treatment line.
10. Include Subgroup Analyses by Reason for Failure
    The effectiveness of cycling versus swapping may differ depending on whether the prior failure was due to primary inefficacy, secondary inefficacy, or adverse events. Please consider adding subgroup analyses stratified by reason for failure to inform clinical decision-making.

Minor Comments

1. Correct the Table Title
   Table 1 is labeled "Baseline characteristics of PsA patients," but the study focuses on axSpA. Please correct this typographical error.
2. Clarify the Inclusion of Non-radiographic axSpA
   The manuscript includes both radiographic and non-radiographic axSpA. Please confirm that the ASAS classification criteria were applied consistently and consider presenting subgroup analyses by axSpA subtype.
3. Improve Figure Quality
   Figures 1 and 2 are difficult to interpret in the current format. Please increase resolution, ensure axis labels are legible, and consider adding confidence bands to the survival curves.
4. Standardize Terminology
   The manuscript uses "cycling" and "swapping" consistently, but also occasionally uses "switching." Please standardize terminology throughout for clarity.
5. Clarify the Follow-up Duration
   The median follow-up is reported as 4.7 years, but the survival analysis is truncated at 3 years. Please explain why 3 years was chosen as the cutoff and whether longer-term outcomes differed.
6. Add Baseline Disease Activity Data
   BASDAI scores are reported for a subset of patients but not included in the multivariable model due to missing data. Please discuss how this limitation affects the interpretation of findings and consider using multiple imputation for this variable as well.
7. Clarify the Handling of Concomitant Medications
   The manuscript mentions that only the b/tsDMARD component was considered for persistence analysis. Please clarify whether concomitant methotrexate or other csDMARDs were documented and whether their use differed between groups.
8. Expand the Discussion of IL-23i Use
   The manuscript includes IL-23i in the analysis but notes they are not formally indicated for axSpA without psoriasis. Please discuss the rationale for including these agents and whether results differed when they were excluded.
9. Add a Limitations Section Subheading
   The Discussion contains a thorough limitations paragraph but would benefit from a dedicated subheading to improve readability and emphasize transparency.
10. Correct Reference Formatting
    The reference list contains some inconsistencies in formatting (e.g., missing issue numbers, journal name abbreviations). Please ensure all references conform to the journal’s style guidelines.

Comments on the Quality of English Language

The English could be improved to more clearly express the research.

Author Response

Major Comments

 

Clarify the Definition of Treatment Failure

The authors define failure as discontinuation due to primary/secondary inefficacy or adverse events, but exclude patient choice or pregnancy. While this is reasonable, it introduces potential selection bias. Please clarify how many patients were excluded for these reasons and discuss the impact on generalizability.

 

-  We reported in Table 1 also the other drug discontination reasons (patient choice, pregnancy, loss to follow‑up, or switch to a biosimilar), with the majority being switches from originator to biosimilar. These cases were not excluded from the analysis but were treated as censored observations, consistent with standard practice in drug‑survival studies in rheumatology, where non–drug‑related discontinuations are commonly censored. Because these discontinuations are unrelated to treatment performance, their censoring is unlikely to introduce meaningful selection bias. Nonetheless, we have added a statement in the Methods/Limitations to clarify this point and to acknowledge the potential impact on generalizability.

 

Address the Imbalance Between Groups

The Cycling Group (CG) and Swapping Group (SG) differ significantly in disease duration, psoriasis prevalence, IBD prevalence, and prior lines of therapy. Although multivariable adjustment was performed, residual confounding remains a concern. Consider adding propensity score matching as a sensitivity analysis to strengthen causal inference.

 

- To address the potential for residual confounding arising from baseline differences between the CG and SG, we performed an additional propensity score–based sensitivity analysis (Figure 3). The propensity score was estimated using all relevant baseline covariates, and the resulting model confirmed that the risk of discontinuation did not differ significantly between the two groups (HR 1.09, 95% CI 0.66–1.78; p=0.742). These findings are consistent with the multivariable Cox models and support the robustness of our results.

 

Clarify the Definition of "Line of Therapy"

Line of therapy is determined chronologically, but it is unclear whether lines are defined by drug class changes or by sequential treatments regardless of mechanism. Please specify how lines were counted, particularly when patients switched between agents of the same class.

 

-We have clarified the definition of line of therapy in the Methods section. Specifically, we now state that each new b/tsDMARD—regardless of its mechanism of action—was counted as a new line of therapy. Consequently, switches between agents within the same drug class (e.g., between different TNF inhibitors) were considered subsequent lines. We also specify that the first‑ever b/tsDMARD was designated as line 1, irrespective of whether it had been initiated at our center or elsewhere, and that prior treatment records were retrieved whenever possible for patients referred after starting therapy at another institution.

 

 

Explain the Handling of Multiple Treatment Courses

The analysis includes 343 treatment courses from 156 patients, meaning some patients contributed multiple observations. Please clarify how within-patient correlation was accounted for in the statistical analysis (e.g., robust standard errors or frailty models).

 

- As several patients contributed more than one treatment course, we accounted for within‑patient correlation in all time‑to‑event analyses by using Cox proportional hazards models with robust standard errors clustered at the patient level. This approach appropriately adjusts variance estimates for repeated observations contributed by the same individual. We have now clarified this in the Methods section.

 

Justify the Use of Treatment Persistence as an Effectiveness Proxy

Drug survival is influenced by many factors beyond efficacy, including tolerability, access, and provider prescribing habits. Please discuss the limitations of using persistence as a surrogate for effectiveness and consider adding a sensitivity analysis restricted to discontinuations due to inefficacy.

 

- We agree that treatment persistence is influenced by factors beyond efficacy, including tolerability, and therefore represents an imperfect surrogate for effectiveness. Nonetheless, drug survival is widely used in real‑world rheumatology studies as a composite measure reflecting both efficacy and tolerability. To address the reviewer’s concern, we performed an additional sensitivity analysis restricted to discontinuations due to inefficacy. The results were consistent with the main analysis, showing no significant difference in discontinuation risk between the CG and SG. These findings support the robustness of our conclusions despite the known limitations of persistence as an effectiveness proxy.

 

Clarify the Distribution of Drug Classes in the Swapping Group

The SG includes 40 TNFi courses (30.8%), which is notable given that SG is defined as switching to a different mechanism of action. Please clarify whether these represent patients switching from a non-TNFi to a TNFi, and discuss the clinical rationale.

 

- The TNFi courses included in the SG represent patients who switched from a non‑TNFi MOA (e.g., IL‑17i, IL‑12/23i, or JAKi) to a TNFi. These cases were therefore classified as “swapping” because the switch involved a change in MOA, even though the subsequent agent belonged to the TNFi class. Clinically, switching from a non‑TNFi to a TNFi is a common strategy in axial spondyloarthritis. We have clarified this point in the Methods section to avoid confusion.

 

Address the High Proportion of Missing Failure Reasons

Reasons for failure were documented for only 58.3% of treatment courses. This missingness could bias the interpretation of strategy effectiveness. Please discuss how this limitation was addressed and consider multiple imputation for this variable.

 

- We would like to clarify that no data are missing regarding reasons for failure. A reason for discontinuation is recorded only when a treatment course is actually interrupted. The remaining courses were still ongoing at the last observation and were therefore censored, which is standard practice in survival analysis.

 

Interpret the Prescription Year Finding with Caution

The finding that more recent prescription year is associated with higher discontinuation risk is intriguing but could reflect channeling of sicker patients to newer agents or changes in clinical practice. Please discuss potential explanations and whether this finding persisted after adjusting for line of therapy.

 

- We agree that the association between more recent prescription year and higher discontinuation risk should be interpreted with caution. We have expanded the Discussion to address these potential explanations and to acknowledge the limitations inherent in interpreting temporal effects in observational data.

 

Clarify the Handling of Biosimilars

Biosimilars are grouped with originator biologics for MOA classification. This is acceptable, but please clarify whether biosimilars were considered separate treatment courses or whether switching between an originator and its biosimilar was counted as a new treatment line.

 

- Patient lines switching from originator to biosimilar or vice versa were excluded from the study. We have clarified this point in the methods section.

 

Include Subgroup Analyses by Reason for Failure

The effectiveness of cycling versus swapping may differ depending on whether the prior failure was due to primary inefficacy, secondary inefficacy, or adverse events. Please consider adding subgroup analyses stratified by reason for failure to inform clinical decision-making.

 

- We performed an additional sensitivity analysis restricted to discontinuations due to inefficacy and to adverse events in the results section.  

 

 

Minor Comments

 

Correct the Table Title

Table 1 is labeled "Baseline characteristics of PsA patients," but the study focuses on axSpA. Please correct this typographical error.

 

- We have corrected the error.

 

Clarify the Inclusion of Non-radiographic axSpA

The manuscript includes both radiographic and non-radiographic axSpA. Please confirm that the ASAS classification criteria were applied consistently and consider presenting subgroup analyses by axSpA subtype.

 

- We have clarified this point in the Methods section 

 

Improve Figure Quality

Figures 1 and 2 are difficult to interpret in the current format. Please increase resolution, ensure axis labels are legible, and consider adding confidence bands to the survival curves.

 

- We have improved the figures.

 

Standardize Terminology

The manuscript uses "cycling" and "swapping" consistently, but also occasionally uses "switching." Please standardize terminology throughout for clarity.

 

- We have corrected the term “switching”, where appropriate.

 

Clarify the Follow-up Duration

The median follow-up is reported as 4.7 years, but the survival analysis is truncated at 3 years. Please explain why 3 years was chosen as the cutoff and whether longer-term outcomes differed.

- There were no differences in the longer-term outcomes as reported by the multivariate analysis. However the KM curve, due to the low number at risk, may be misleading to the inexerienced readers.

 

Add Baseline Disease Activity Data

BASDAI scores are reported for a subset of patients but not included in the multivariable model due to missing data. Please discuss how this limitation affects the interpretation of findings and consider using multiple imputation for this variable as well.

 

- As reported in Table 1, BASDAI scores were missing for a large proportion of treatment courses. Because BASDAI was not routinely collected in clinical practice during the earlier years of the study period, the missingness is substantial and structurally related to calendar time. Under these conditions, multiple imputation would be unreliable. We reported this limitation in the discussion sectrion. 

 

Clarify the Handling of Concomitant Medications

The manuscript mentions that only the b/tsDMARD component was considered for persistence analysis. Please clarify whether concomitant methotrexate or other csDMARDs were documented and whether their use differed between groups.

 

- Unfortunately, concomitant csDMARD use was not consistently documented across the entire study period, particularly in earlier years, and therefore complete data were not available for analysis. For this reason, csDMARDs could not be reliably included as covariates.

 

Expand the Discussion of IL-23i Use

The manuscript includes IL-23i in the analysis but notes they are not formally indicated for axSpA without psoriasis. Please discuss the rationale for including these agents and whether results differed when they were excluded.

 

-As reported in the discussion IL23i were indlcuded because they were in this subset of patients primarily to address relevant extra-articular manifestations, particularly psoriasis and inflammatory bowel disease. 

 

 

Add a Limitations Section Subheading

The Discussion contains a thorough limitations paragraph but would benefit from a dedicated subheading to improve readability and emphasize transparency.

 

-A Limitantions section su heading has been added. 

 

Correct Reference Formatting

The reference list contains some inconsistencies in formatting (e.g., missing issue numbers, journal name abbreviations). Please ensure all references conform to the journal’s style guidelines.

 

-We checked alla the references  (including journal numbers and journal name abbreviations)

 

Reviewer 2 Report

Comments and Suggestions for Authors

1. “We included all patients aged >17 years” – why 17 and not 18 years (age of legal maturity)? Please justify in the manuscript.
2. “We included all patients aged >17 years with a diagnosis of axSpA who commenced treatment with a bDMARD or tsDMARD at our Rheumatology Unit between January 2004 and July 2025” – please add the country (Italy). Future reviews and meta-analyses will look for it.
3. “We included all patients aged >17 years with a diagnosis of axSpA who commenced treatment with a bDMARD or tsDMARD at our Rheumatology Unit between January 2004 and July 2025” – since not all axSpA patients receive b/tsDMARDs, the reader needs to know what are the criteria for commencing them. Is there a set of national criteria or is it a decision left for the center or for each physician? Please either briefly add to the manuscript initiation criteria or at least reference them to another previous peer-reviewed publication. They act as a proper set of inclusion criteria for your study. Also, mention if these criteria have changed over your observation period. If so, a sensitivity analysis is in order.
4. “The diagnosis of axSpA was made by experienced rheumatologists based on the Assessment of SpondyloArthritis international Society (ASAS) classification criteria.” – please note that the diagnosis of axSpA is not based on any set of criteria. So please rephrase to state that your axSpA diagnoses also fulfilled the ASAS classification criteria.
5. “treatment duration (interval between first and last administration)” – does your center record injection/IV administrations per patient by date in order for you to retrieve it from EMRs? Or did you just have prescription dates? If you only had prescription dates, the definition of treatment duration needs to be changed and further dependent text, if any, should be rephrased.
6. “We analyzed 156 axSpA patients (59 radiographic axSpA, 97 non-radiographic axSpA) who failed at least one line of advanced therapy” – in more than 20 years of observation you only had 156 axSpA patients who failed at least one line of advanced therapy? This number seems low. Please comment/justify (not necessarily in the manuscript).
7. In the statistic section your write: “Descriptive statistics are presented as medians with 95% confidence intervals (CI).” In the results you write: “The median follow-up duration was 4.7 years (IQR 2.3-8.1 years)”. Please be consistent in the manuscript text.
8. The crude overlap in persistence curves may simply reflect limited power. Since the Cox model found only year of prescription significant, I would worry about one of 4 possibilities: insufficient event count, overfitting, collinearity between treatment strategy and drug class/calendar year, or residual confounding. You should report the number of discontinuation events and the full multivariable model and discuss these concerns as appropriate.
9. “These differences reflect real-world clinical practice, where patients with more complex or refractory disease may be preferentially channeled towards swapping strategies, while those with specific comorbidities such as IBD may be maintained on TNFi, which have established efficacy for both axial and intestinal manifestations.” – this text from the Result section is actually a Discussion.
10. “This distribution reflects the evolving treatment landscape and the increasing availability of non-TNFi options over the study period.” - this text from the Result section is actually a Discussion.
11. “The higher proportion of secondary failure in the SG likely reflects the more heavily pre-treated nature of this population.” - this text from the Result section is actually a Discussion. Also, this modular pattern of a result section followed by a short explanation/conclusion/comment hints to AI use. If authors used AI, this should be disclosed.
12. “This consistency suggests the finding is robust and unlikely to be explained by methodological artifacts.” - this text from the Result section is actually a Discussion.
13. The 2 groups appear substantially imbalanced at baseline. Swappers were older, had longer disease duration, different extra-musculoskeletal phenotype frequencies, and a markedly different prior/current mechanism distribution. That raises concern for confounding by indication: the treatment strategy may reflect underlying disease complexity, phenotype, calendar period, drug availability, or prescribing preferences rather than a directly comparable clinical decision. This should be discussed.
14. Table 1, Figure 1, Figure 2 – please use standard layouts and at write at the bottom of the Table/Figure your brief notes and the alphabetical list of all table/figure abbreviations in order for the table/figure to be able to be read alone.
15. “The finding that more recent treatment initiation predicts discontinuation warrants further investigation into evolving therapeutic paradigms. Future research should explore whether this reflects improved treatment standards, changing patient expectations, or other factors. Longitudinal studies with detailed capture of disease activity, treatment response, and reasons for discontinuation could help elucidate the mechanisms underlying this temporal trend. Comparative effectiveness research in this area would benefit from larger, prospective, multicenter cohorts with standardized data collection, including validated disease activity measures, patient-reported outcomes, and biomarker data. Such studies could enable more granular analyses, including agent-level comparisons, identification of predictors of differential response, and evaluation of longer-term outcomes beyond three years.” – these are not conclusions of your study. These are a specific area of the Discussion section, a “further research” section. Please conclude upon your results through the lens of your discussion.
16. Please note that Reference 1 is missing or the entire reference list is skewed by one.

Author Response

Comments and Suggestions for Authors

“We included all patients aged >17 years” – why 17 and not 18 years (age of legal maturity)? Please justify in the manuscript.

 

- We corrected the manuscript

 

“We included all patients aged >17 years with a diagnosis of axSpA who commenced treatment with a bDMARD or tsDMARD at our Rheumatology Unit between January 2004 and July 2025” – please add the country (Italy). Future reviews and meta-analyses will look for it.

 

- We corrected the manuscript

 

“We included all patients aged >17 years with a diagnosis of axSpA who commenced treatment with a bDMARD or tsDMARD at our Rheumatology Unit between January 2004 and July 2025” – since not all axSpA patients receive b/tsDMARDs, the reader needs to know what are the criteria for commencing them. Is there a set of national criteria or is it a decision left for the center or for each physician? Please either briefly add to the manuscript initiation criteria or at least reference them to another previous peer-reviewed publication. They act as a proper set of inclusion criteria for your study. Also, mention if these criteria have changed over your observation period. If so, a sensitivity analysis is in order.

 

- Giveng the long time frame of our study, we reported that the patients were treated with b/tsDMARDs according to local guidelines and the treating physician’s assessment.

 

 

“The diagnosis of axSpA was made by experienced rheumatologists based on the Assessment of SpondyloArthritis international Society (ASAS) classification criteria.” – please note that the diagnosis of axSpA is not based on any set of criteria. So please rephrase to state that your axSpA diagnoses also fulfilled the ASAS classification criteria.

 

- We have modified the statements in the methods section

 

“treatment duration (interval between first and last administration)” – does your center record injection/IV administrations per patient by date in order for you to retrieve it from EMRs? Or did you just have prescription dates? If you only had prescription dates, the definition of treatment duration needs to be changed and further dependent text, if any, should be rephrased.

 

- Our center record every IV administration per patients and every prescription dates of injections. If the treatment was interrupted between two prescription, the date of the last injection was recorded as the one of discontinuation. If the patients was lost on follow-up the date of last visit was recorded the one of discontinuation.

 

“We analyzed 156 axSpA patients (59 radiographic axSpA, 97 non-radiographic axSpA) who failed at least one line of advanced therapy” – in more than 20 years of observation you only had 156 axSpA patients who failed at least one line of advanced therapy? This number seems low. Please comment/justify (not necessarily in the manuscript).

 

- Given that we analyzed only first line failure and given that the axSpA has a very good response to b/tsDMARDs our numbers are in line with the ones of other monocentric studies. 

In the statistic section your write: “Descriptive statistics are presented as medians with 95% confidence intervals (CI).” In the results you write: “The median follow-up duration was 4.7 years (IQR 2.3-8.1 years)”. Please be consistent in the manuscript text.

 

- We corrected the methods section.

 

The crude overlap in persistence curves may simply reflect limited power. Since the Cox model found only year of prescription significant, I would worry about one of 4 possibilities: insufficient event count, overfitting, collinearity between treatment strategy and drug class/calendar year, or residual confounding. You should report the number of discontinuation events and the full multivariable model and discuss these concerns as appropriate.

 

- The events were reported in the Table 1 and in the results (184 events). These events are sufficient for the presented model. 

 

“These differences reflect real-world clinical practice, where patients with more complex or refractory disease may be preferentially channeled towards swapping strategies, while those with specific comorbidities such as IBD may be maintained on TNFi, which have established efficacy for both axial and intestinal manifestations.” – this text from the Result section is actually a Discussion.

“This distribution reflects the evolving treatment landscape and the increasing availability of non-TNFi options over the study period.” - this text from the Result section is actually a Discussion.

“The higher proportion of secondary failure in the SG likely reflects the more heavily pre-treated nature of this population.” - this text from the Result section is actually a Discussion. Also, this modular pattern of a result section followed by a short explanation/conclusion/comment hints to AI use. If authors used AI, this should be disclosed.

“This consistency suggests the finding is robust and unlikely to be explained by methodological artifacts.” - this text from the Result section is actually a Discussion.

 

-These sentences have been removed from the results, 

 

The 2 groups appear substantially imbalanced at baseline. Swappers were older, had longer disease duration, different extra-musculoskeletal phenotype frequencies, and a markedly different prior/current mechanism distribution. That raises concern for confounding by indication: the treatment strategy may reflect underlying disease complexity, phenotype, calendar period, drug availability, or prescribing preferences rather than a directly comparable clinical decision. This should be discussed.

 

- In otder to address the potential for residual confounding arising from baseline differences between the CG and SG, we performed an additional propensity score–based sensitivity analysis (Figure 3). 

 

Table 1, Figure 1, Figure 2 – please use standard layouts and at write at the bottom of the Table/Figure your brief notes and the alphabetical list of all table/figure abbreviations in order for the table/figure to be able to be read alone.

 

- The Table and the figures have been improved.

 

“The finding that more recent treatment initiation predicts discontinuation warrants further investigation into evolving therapeutic paradigms. Future research should explore whether this reflects improved treatment standards, changing patient expectations, or other factors. Longitudinal studies with detailed capture of disease activity, treatment response, and reasons for discontinuation could help elucidate the mechanisms underlying this temporal trend. Comparative effectiveness research in this area would benefit from larger, prospective, multicenter cohorts with standardized data collection, including validated disease activity measures, patient-reported outcomes, and biomarker data. Such studies could enable more granular analyses, including agent-level comparisons, identification of predictors of differential response, and evaluation of longer-term outcomes beyond three years.” – these are not conclusions of your study. These are a specific area of the Discussion section, a “further research” section. Please conclude upon your results through the lens of your discussion.

 

- We have modified the discussion according to this suggestion.

 

Please note that Reference 1 is missing or the entire reference list is skewed by one.

- Reference 1 in reported in line 4 of the introduction. 

Reviewer 3 Report

Comments and Suggestions for Authors

Abstract: is in structural form very well written. Background, objectives and methods are adequately described. Results could be better presented in example (suggestion) sentence: Retention rates at 1, 2, and 3 years were 62.7%, 49.3%, and 24 39.2%, respectively for cycling group comparing to 69.8%, 47.8%, and 31.8% respectively for swapping group (HR 1.13, 95% CI 0.83-1.53; p=0.442). Conclusion is clearly stating study results but by authors opinion they could suggest direction for further investigations and possible improvement of treatment.

Keywords: are well written.

Introduction:

Comment: Introduction is generally well written. I suggest to add short definition and disease description at beginning of chapter in few sentences. Evolution of therapy, and emerging problems with new biologic therapy like side effects, failure are well described, gradually introducing problems and this study objectives. Two considering therapeutic strategies, cycling or swapping are adequately described with easy to understand differences and principal strategies which one to choose after initial biological agent therapy failure. Authors have well explained that RCT comparing those strategies are lacking and only real-world observation studies offer evidence of those therapeutic strategies advantages/disadvantages as well as those studies limitations. Objectives of this study are well defined and described.

Materials and Methods

It consist:

1. Study design (in last part of introduction) is defined as retrospective observational study, suggest to move this in Materials and Methods chapter.
2. Study population and data sources are precisely defined. Cohort are defined at the end of chapter (cycling or swapping treatment strategy), I suggest that they be more clearly defined at the study design description as well main difference between groups clearly stated in the beginning.
3. Inclusion and exclusion criteria – are precisely described, criteria for therapeutic failure (primary and secondary) is adequately described.
4. Implemented procedures are precisely described in details regarding drug implementation, drug categorization by mechanism of action (and all data related to therapy)
5. Statistical analyzes and analyzed variables are appropriate, used statistical tests for analyzes I consider appropriate, what is considered statistical significant is defined (p value) name of software used is listed.
6. Study period is precisely defined.

Results

Comment: Presented data should be limited only to study results without comment or comparison with other studies, that is reserved for discussion chapter. (in example: The study cohort was predominantly middle-aged with a slight  male predominance, reflecting the typical epidemiology of axSpA; These differences reflect real-world clinical practice, where patients with more complex or refractory disease may be preferentially channeled towards swapping strategies, while those with specific comorbidities such as IBD may be maintained on TNFi, which have established efficacy for both axial and intestinal manifestations.) Please correct all comments regarding study results, and move them to discussion chapter.

Comment: main results and subgroup analyzes results and Cox regression results  are well presented.

Discussion

Comment: authors have clearly explained study results, compared to existing cotemporary  data,  and advantages of this study compared to previous at the beginning of chapter. They asserted differences in results of this study and other studies as well as provides possible explanation for that and possible pitfalls in other studies in understandable way. Explanation in changes of therapeutic strategies and introducing of novel strategies is well done as well as defining he main problem of treatment of those patients with described strategies in sentence: “The comparable persistence between cycling and swapping, coupled with the absence of identified clinical predictors of differential response, supports an individualized approach to treatment selection following advanced therapy failure.” Although authors assert in some conditions advantage of certain drugs clear clinical universal decision-making algorithm is missing and authors recommend individual approach. I do recommend to add small graphical representation here regarding described if it is possible.  

Study strengths and limitations are adequately defined.

Conclusion. Well written, direction for further investigations and possible improvement of treatment is included here.

 

Table 1. is generally ok, but I suggest not using dots for empty spaces- just leave it empty, also some symbols / abbreviations missing explanation: M – male gender, F – female gender, n- absolute number etc, p value explanation of statistical significance, please correct this in table explanation, also suggest to add vertical borders.

Figure 1. is poorly done. Letters are too small, graphical representation is not clear do to overlapping blurred borders beaten groups, it shows correct data but representation has to be more clear and easy to read on in this way or do it in another in example table

Figure 2. is generally ok, except letters are too small, and some type errors, please correct.   

Literature: is written according to journal instructions.

Author Response

Comment: Introduction is generally well written. I suggest to add short definition and disease description at beginning of chapter in few sentences. Evolution of therapy, and emerging problems with new biologic therapy like side effects, failure are well described, gradually introducing problems and this study objectives. Two considering therapeutic strategies, cycling or swapping are adequately described with easy to understand differences and principal strategies which one to choose after initial biological agent therapy failure. Authors have well explained that RCT comparing those strategies are lacking and only real-world observation studies offer evidence of those therapeutic strategies advantages/disadvantages as well as those studies limitations. Objectives of this study are well defined and described.

 

- Thank you.

 

Materials and Methods

It consist:

Study design (in last part of introduction) is defined as retrospective observational study, suggest to move this in Materials and Methods chapter.

Study population and data sources are precisely defined. Cohort are defined at the end of chapter (cycling or swapping treatment strategy), I suggest that they be more clearly defined at the study design description as well main difference between groups clearly stated in the beginning.

Inclusion and exclusion criteria – are precisely described, criteria for therapeutic failure (primary and secondary) is adequately described.

Implemented procedures are precisely described in details regarding drug implementation, drug categorization by mechanism of action (and all data related to therapy)

Statistical analyzes and analyzed variables are appropriate, used statistical tests for analyzes I consider appropriate, what is considered statistical significant is defined (p value) name of software used is listed.

Study period is precisely defined.

 

- Few implementation have been done as suggested,

 

Results

 

Comment: Presented data should be limited only to study results without comment or comparison with other studies, that is reserved for discussion chapter. (in example: The study cohort was predominantly middle-aged with a slight  male predominance, reflecting the typical epidemiology of axSpA; These differences reflect real-world clinical practice, where patients with more complex or refractory disease may be preferentially channeled towards swapping strategies, while those with specific comorbidities such as IBD may be maintained on TNFi, which have established efficacy for both axial and intestinal manifestations.) Please correct all comments regarding study results, and move them to discussion chapter.

 

- Comments were removed from tbe results section. 

 

Comment: main results and subgroup analyzes results and Cox regression results  are well presented.

 

- Thank you

 

Discussion

 

Comment: authors have clearly explained study results, compared to existing cotemporary  data,  and advantages of this study compared to previous at the beginning of chapter. They asserted differences in results of this study and other studies as well as provides possible explanation for that and possible pitfalls in other studies in understandable way. Explanation in changes of therapeutic strategies and introducing of novel strategies is well done as well as defining he main problem of treatment of those patients with described strategies in sentence: “The comparable persistence between cycling and swapping, coupled with the absence of identified clinical predictors of differential response, supports an individualized approach to treatment selection following advanced therapy failure.” Although authors assert in some conditions advantage of certain drugs clear clinical universal decision-making algorithm is missing and authors recommend individual approach. I do recommend to add small graphical representation here regarding described if it is possible.  

 

-The aim of our study was not to propose a universal decision‑making algorithm, but rather to present the findings of our monocentric cohort. Further multicenter studies with larger and more heterogeneous populations will be needed to validate our results and to support the development of evidence‑based treatment algorithms.

 

Study strengths and limitations are adequately defined.

 

-Thank you

 

Conclusion. Well written, direction for further investigations and possible improvement of treatment is included here.

 

-Thank you

 

 

Table 1. is generally ok, but I suggest not using dots for empty spaces- just leave it empty, also some symbols / abbreviations missing explanation: M – male gender, F – female gender, n- absolute number etc, p value explanation of statistical significance, please correct this in table explanation, also suggest to add vertical borders.

Figure 1. is poorly done. Letters are too small, graphical representation is not clear do to overlapping blurred borders beaten groups, it shows correct data but representation has to be more clear and easy to read on in this way or do it in another in example table

Figure 2. is generally ok, except letters are too small, and some type errors, please correct.   

 

-Tables and figures were improved. 

 

Literature: is written according to journal instructions.

 

-Thank you

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The authors have performed all needed changed 

Comments on the Quality of English Language

The English could be improved to more clearly express the research.