Failure is constructive feedback that tells you to try a different approach to accomplish what you want. – Idowu Koyenikan1
Secondary failure of a biologic disease-modifying antirheumatic drug (bDMARD) describes loss of therapeutic response over time after initial response, in contrast to the less commonly observed primary failure, where there has been an inability to achieve an initial response given an adequate period of treatment. Whereas it is well studied in patients with rheumatoid arthritis (RA), this issue has been less well examined in prospective psoriatic arthritis (PsA) patient cohorts.
Possible etiological factors include (1) immunogenicity with the development of neutralizing antidrug antibodies; (2) disease progression despite adequate drug levels; (3) flares due to intercurrent contributors like intercurrent infection; (4) breaks in therapy including cessation of concomitant therapy like methotrexate, and stopping and starting therapy for intercurrent factors like pregnancy, surgery, and/or infection; (5) adverse effects from therapy leading to reduced doses or cessation of concomitant therapy; (6) patient factors, from central sensitization/chronic widespread pain to uncontrolled symptoms from joint damage; (7) less commonly, the development of potentially confusing alternate diagnoses such as tumor necrosis factor (TNF)-induced systemic lupus erythematosus (SLE)-like syndromes in patients with RA or PsA, or Janus kinase (JAK) inhibitor–induced atopic dermatitis or acneiform rash incorrectly considered as a flare of psoriasis; (8) therapeutic nonadherence; and (9) socioeconomic issues affecting access and reimbursement.
What is known about the incidence of secondary failure in this patient population?
Haddad et al2 reported a retrospective cohort of 3851 patients with PsA, divided into 2005-2013 and 2014-2023 groups, given the increase in treatment options in latter years. The majority of patients commenced a TNF inhibitor (TNFi) as the first bDMARD. In the more recent period, secondary failure led to 52% switching bDMARD once, followed by 23% twice, 12% 3 times, and 6% switching mode of therapy 5 or more times. Predictors for secondary failure were female sex, obesity, smoking, and lower socioeconomic background.
Drug survival and persistence on bDMARDs tend to be higher for first-line TNFi, and is shorter with second-line and subsequent therapies. Switching in mode of action leads to increases in use of interleukin-17 inhibitors (IL-17i) and IL-23i.
The study by Haddad et al2 was limited by the absence of treatment indications and the reasons for secondary failure and discontinuation, nor did it include disease activity measures.
Lumetti et al,3 in a monocentric retrospective observational study of 122 patients with PsA in 2016-2022, compared secondary failure leading to a switch within TNFi class with switch to different mechanism of action, specifically TNFi to and from IL-17i: 67% of patients switched within TNFi class, 72% from TNFi to IL-17i, and 46% from IL-17i to TNFi. The reason for switching was predominantly for secondary failure. The retention rate of TNFi to IL-17 switch was highest, but not statistically significant, at 58% at 2 years. Predictors for failure leading to a switch was high Disease Activity Index for Psoriatic Arthritis (DAPSA) and year of commencing treatment; the latter speaks to the ease of changing therapy as new options arrive. Again, the reasons for change of therapy were not detailed.
A systematic literature review of 7 registries and 7 observational studies showed that first-line therapy discontinuation, including for secondary failure, ranged from 17% to 34% at 1 year, and was higher as time passed. Reasons for discontinuation included adverse effects (7-11%) and patient preference (7-11%). The degree of nonadherence was difficult to determine. Predictors of failure and/or discontinuation were not described.4
Haberman et al described a cohort of 960 patients followed for an average of 30 months.5 TNFi were the first-line bDMARDs in 83% of patients and the persistence rate was 72% at 1 year. Female patients were more likely to fail therapy and discontinue. Persistence decreased with each subsequent exposure and the predictors were depression, female sex, obesity, and axial disease. Multiple failures were associated with more severe skin disease, older age, and longer disease duration.5
Further, in a retrospective cohort with 264 patients, Lorenzin et al found 44% switched, including due to secondary failure, and the risk factors were female sex, polyarticular disease, worse function as measured by Health Assessment Questionnaire, extent of psoriasis, presence of dactylitis, and poorer Charles Comorbidity Index.6 Reasons for discontinuation, which included secondary failure, were inefficacy—mainly in joint disease but to a lesser extent skin disease—as well as adverse effects. Other reasons included infection, infusion reactions, extent of psoriasis, and dactylitis. Regression analysis found polypharmacy, insurance status, and corticosteroid use as risk factors.
Finally, Sewerin et al detailed 348 patients and found a persistence rate of 58% at 1 year.7 Other studies have suggested concurrent conventional synthetic DMARD (csDMARD) use, especially methotrexate (MTX), with TNFi was associated with treatment continuation.8
New prospective data
In this issue of The Journal of Rheumatology, Kharouf and colleagues describe a prospective monocentric observational PsA cohort followed over the years 2000-2023 who had remained on a bDMARD for > 1 year from enrollment.9 Response to therapy was assessed not by a validated disease activity index, but rather by a cutoff of ≥ 40% reduction in swollen joint count and ≥ 50% reduction in Psoriasis Area and Severity Index (PASI) score or PASI change to ≤ 2. Another problematic issue is that the authors defined secondary failure as the inability to maintain response or based it on the assessment by “clinician judgment.” Of 482 patients, predominantly treated with TNFi, 55% were responders at 1 year, with 36% developing secondary failure at a median of 1.6 years. Risk factors were higher swollen joint counts and higher PASI score at baseline, accepting that these outcomes were already reduced at the 1-year starting point, as medication had been commenced for up to 12 months prior to enrollment. Concomitant MTX was not a significant risk factor.9
So, a number of consistent themes are apparent from these studies: (1) TNFi dominates first bDMARD use and the benefits of concomitant csDMARDs, in particular MTX, are variable. (2) Cohorts are generally monocentric and retrospective. (3) The incidence of discontinuation or switching is concerningly frequent at 1 year in the 20-40% range, and dissecting out secondary failure as a cause for discontinuation is often unclear. (4) Consistent risk factors across studies include female sex, obesity, and smoking; less consistent but still relevant factors include increasing age, extent of psoriatic rash, polyarthritis/dactylitis, and poor function.
The implications from these studies are clear. Too many patients have churned off medication by 12 months, and the reasons are often not examined in detail or addressed. Perhaps it is easier and often more timely to simply switch drugs, as we are spoiled for choice.
The reason for sex differences in disease activity and response to therapy needs further elucidation (the Group for Research and Assessment of Psoriasis and Psoriatic Arthritis [GRAPPA] has a large international study underway) and future clinical trials should be stratified for sex at baseline.
Obesity, which similarly drives disease activity and impairs response, must be more seriously tackled, and the advent of the glucagon-like peptide-1 (GLP-1) agonists has opened the door to effective therapy. Despite their costs and given that the PsA population has increased risks of metabolic syndrome including gout and premature atherosclerosis, as well as being prone to excess alcohol and smoking, significant improvements in major adverse cardiac events, nonalcoholic steato-hepatitis, reduction in gout flares, improved recovery from knee replacement surgery, and improvements in lupus nephritis have all been demonstrated with GLP-1 therapy.10 It is time that rheumatologists either involved obesity specialists or preferably learned to use these drugs appropriately themselves. It is also disappointing that, in general, rheumatologists make little effort to insist our patients stop smoking on one hand, while happily deploying expensive drugs on the other.
Other identified risk factors tend be features of aggressive disease and many groups such as the European Alliance of Associations for Rheumatology (EULAR), Assessment of SpondyloArthritis international Society (ASAS), and GRAPPA are trying to define and validate “treatment-refractory disease” (patients with objective signs of ongoing inflammation who have failed an adequate trial of bDMARDs of ≥ 2 different mechanisms of action) and “complex-to-manage disease” (patients with significant symptoms who have no objective signs of inflammation; eg, on imaging or with normal acute-phase reactants who have also failed ≥ 2 bDMARDs of different mechanisms of action). Strategy trials yet to be performed, perhaps of combination bDMARD therapy, may improve outcomes in the former; in the latter, managing central sensitization and chronic widespread pain with a variety of measures that do not include unsuccessful repeated cycling through expensive bDMARDs would be worthwhile.
Conclusion
The study by Kharouf et al9 adds useful prospective information about the incidence and predictors of secondary failure of bDMARD therapy in PsA. Perhaps rheumatologists should take heed of this “constructive feedback that tells us to try a different approach”1 for the benefit of our patients.
Footnotes
See Secondary failure to biologics in PsA, page 162
- Copyright © 2026 by the Journal of Rheumatology
