Comment on: Clinical outcomes in psoriatic arthritis patients with rheumatoid factor positivity: Reply

Dear Editor, We thank Soysal et al. for their thoughtful and methodologically rigorous comments [1] regarding our study evaluating the clinical impact of rheumatoid factor (RF) positivity in psoriatic arthritis (PsA) [2].

First, we agree that transient, low-titre RF positivity requires careful interpretation. This point was central to our manuscript rather than an unrecognized limitation. We explicitly reported that RF positivity was observed in 16.1% of patients, with persistent RF positivity in only 3.4%. Thus, transient RF positivity may be related to age, smoking, or other factors. However, low-titre, intermittent RF positivity is a common clinical scenario in PsA and may influence diagnostic and treatment decisions in routine practice. To address the possibility that baseline RF positivity or diagnostic overlap explained the findings, we also evaluated newly developed RF positivity, which showed a similar direction of association. We agree that future studies with lagged exposure definitions may provide more refined estimates.

Second, our use of generalized estimating equations (GEE) was intended to estimate population-averaged longitudinal associations from correlated repeated-measures data and not to establish causal effects. We agree that treatment decisions in PsA are dynamic, and time-varying confounding may be relevant to estimate causal effects of RF status. Marginal structural models with inverse probability weighting may be useful in studies specifically designed to address such questions. However, our objective was to evaluate whether RF positivity was associated with clinical outcomes and treatment discontinuation in a longitudinal PsA cohort, for which GEE represents an appropriate analytical approach. The absence of marginal structural modelling limits causal interpretation, but does not invalidate the observed longitudinal associations.

Third, the authors correctly note that the association between RF positivity and lower odds for minimal disease activity was driven predominantly by patient-reported outcomes (PROs). This was explicitly reported and was central to our interpretation. We do not interpret RF positivity as a marker of more severe objective inflammatory disease activity in PsA and agree that this pattern may reflect non-inflammatory pain mechanisms, including nociplastic pain [3, 4]. Nevertheless, PROs are clinically meaningful outcomes that influence both composite disease activity states and treatment decisions, even when not fully explained by objective inflammation.

Fourth, we agree that diagnostic overlap between PsA and seropositive rheumatoid arthritis (RA) is an important consideration. However, the risk of major disease misclassification is mitigated by the setting of a dedicated longitudinal PsA cohort and standardized rheumatologist assessments. Moreover, most RF positivity was low-titre and transient. If a substantial proportion of RF-positive patients had represented misclassified RA, one might expect stronger associations with objective peripheral inflammatory activity, which was not observed. We agree that the absence of anti-cyclic citrullinated peptide (CCP) antibody data limits our ability to fully exclude RA-like overlap phenotypes. Notably, anti-CCP positivity was observed in 7.2% of patients in the CASPAR study and these patients were confirmed to have PsA by a review committee [5].

Fifth, we agree that bDMARD discontinuation is a heterogeneous end point encompassing primary inefficacy, secondary loss of response, adverse events, patient preference, and other factors. As a pragmatic step to reduce the capture of early non-response or intolerance, discontinuations occurring within the first 6 months of treatment were excluded. Cause-specific or competing-risk analyses for different reasons for discontinuation would provide greater granularity and should be incorporated in future studies. Our proposed hypothesis regarding Fc-mediated immune complex formation was offered as a biologically plausible mechanism rather than a confirmed finding. Pharmacokinetic data, anti-drug antibody measurements, RF isotype characterization, and drug-specific analyses will be necessary to evaluate this further [6]. Nevertheless, treatment discontinuation is a clinically meaningful real-world outcome, and the observed association with RF positivity, particularly newly developed RF positivity, remains relevant and hypothesis-generating.

Finally, regarding directed acyclic graphs and covariate selection, adjustment for disease activity and PROs in the discontinuation analysis was intended to evaluate whether RF positivity remained associated with treatment discontinuation beyond contemporaneous clinical disease burden. We agree that PROs may plausibly lie on the causal pathway between RF positivity and discontinuation. We therefore also ran a GEE model with ever and new RF as exposure and bDMARD discontinuation beyond 12 months as the outcome, without adjusting for PROs. Both ever RF positivity [OR 2.57, 95% CI, 1.33–4.94] and new RF positivity [OR 2.66, 95% CI, 1.49–4.75] remained associated with discontinuation, supporting the robustness of the signal.

In summary, we agree that our findings should be interpreted within an associative rather than causal framework. However, the methodological considerations raised do not invalidate the main observations of our study: RF positivity in PsA was more frequent than historically expected, was often low-titre and transient, and was associated with greater patient-reported disease burden and increased likelihood of bDMARD discontinuation. Future studies should incorporate persistent vs transient RF status, anti-CCP antibody profiling, systematic assessment of non-inflammatory pain mechanisms, drug-level and anti-drug antibody measurements, and causal modelling approaches tailored to specific estimands.

Data availability

Data are available upon reasonable request by any qualified researchers who engage in rigorous, independent scientific research, and will be provided following review and approval of a research proposal and Statistical Analysis Plan (SAP) and execution of a Data Sharing Agreement (DSA). All data relevant to the study are included in the article.

Funding

No specific funding was received from any bodies in the public, commercial or not-for-profit sectors to carry out the work described in this article.

Disclosure statement: Pankti Mehta, Virginia Carrizo Abarza, Fadi Kharouf, and Shangyi Gao have no conflicts of interest. Dafna Gladman has received grants and/or consulting fees from Abbvie, Amgen, AstraZeneca, BMS, Eli Lilly, Janssen, Novartis, Pfizer, UCB. Vinod Chandran has received research grants from AbbVie, Amgen, and Eli Lilly and has received honoraria for advisory board member roles from AbbVie, BMS, Eli Lilly, Fresenius Kabi, Janssen, Novartis, and UCB. His spouse is an employee of AstraZeneca. Denis Poddubnyy has received research grants for AbbVie, Eli Lilly, Janssen, Novartis, Pfizer, UCB and has received honoraria for advisory board member roles from AbbVie, Biocad, Bristol-Myers Squibb, Eli Lilly, Janssen, Moonlake, Novartis, Pfizer, and UCB.