Objective: To investigate the response to biologic drugs in psoriatic arthritis and to quantify non-response and outcome from switching agents.
Methods: 60 patients (33 men and 27 women, mean age 46 years, median disease duration 16 years) prescribed biologic drugs for psoriatic arthritis between 2001 and 2006 were studied. Response was evaluated using joint counts, C-reactive protein levels and disease activity scores (using 28 joints; DAS28).
Results: The mean percentage improvements seen were 56% in tender joint count, 70% in swollen joint count, 64% in C-reactive protein level and 36% in the overall disease activity score. Improvements were sustained beyond 24 months with no loss of effect. Side-effects leading to cessation or switching of first-line therapy were only seen in 5% of patients and non-response occurred in 20% long term. Overall, 90% of patients achieved a significant response, using switching in 20% of cases. Outcomes were similar regardless of drug used, duration of disease and subtype of arthritis.
Conclusions: Treatment of active psoriatic arthritis with anti-tumour necrosis factor agents leads to a sustained response over 3 years with most patients tolerating these drugs well. The rate of non-response is low with the majority of patients responding to second- and third-line therapies.
Statistics from Altmetric.com
The advent of tumour necrosis factor (TNF) α inhibitors has offered new therapeutic options in rheumatology, and their efficacy in psoriatic arthritis (PsA) has been proven in randomised controlled trials.1–5 Data, however, are sparse on the longer-term efficacy of these drugs6 and there are few data on real life experience of patients treated in the clinical setting.
It has been shown in rheumatoid arthritis (RA) that up to 40% of patients fail to respond to biological therapies.7 This has been classified further into primary non-response and secondary loss of efficacy,8 although the mechanisms for this are not understood.9 Previous research in RA has provided some evidence of response to a second- or third-line agent when switching anti-TNF agents.10 However, evidence to quantify the rate of non-response or who will respond to “switching” in PsA is limited to a few case series.11
The use of anti-TNF agents in the UK has previously been restricted in PsA. Following the publication of guidelines for the use of anti-TNF drugs in PsA by the British Society for Rheumatology (BSR) in 200511 and National Institute for Clinical Excellence (NICE) in 2006,12 their use has been standardised. Below we summarise our experience of patients with PsA treated with biologic drugs in the last 5 years.
A retrospective analysis of patients with PsA who received any biologic drugs through the Leeds Biologics and Spondyloarthropathy clinics between 2001 and 2006 was performed. All patients who had received treatment for 12 weeks or more and any patients who stopped therapy earlier than 12 weeks due to adverse events were included. As the majority of our data precede the above-mentioned guidelines11 12 treatment response will be presented using disease activity scoring (using 28 joints; DAS28).
All patients treated in our clinic are invited to be registered in the Leeds Spondyloarthropathy and the Resistant Arthritis databases that have been approved by the local ethics committee. Diagnosis of PsA in this cohort was based on the Moll and Wright criteria13 as the study was started before the publication of the CASPAR criteria.14 Following informed written consent, clinical and laboratory data are collected prospectively. These include demographics, disease duration, pattern of arthritis, tender joint count, swollen joint count, C-reactive protein (CRP) levels and visual analogue scores for global health to allow for the calculation of a disease activity score (DAS28). In addition, any reason for discontinuing therapy is carefully documented.
Data are presented as percentage of improvement. Significance testing was done using a one-way ANOVA test. A p-value of less than 0.05 was considered as statistically significant.
A total of 60 PsA patients receiving 82 prescriptions for TNF blockers were identified. All of the patients prescribed TNF blockers consented to join the Resistant Arthritis database and their data were collected prospectively. The demographics of the study cohort are detailed in table 1.
Full data regarding response are available for 67 of the 82 prescriptions of TNF blocker therapies. The mean percentage improvements seen were 56% in tender joint count, 70% in swollen joint count, 64% in CRP level and 36% in the overall DAS score. Patients treated with one drug for over 24 or 36 months (n = 30, 45%) showed a consistent reduction these parameters with no significant loss of effect (fig 1A). Interestingly, a significant improvement in swollen joint count was observed with a longer duration of treatment (p = 0.039).
When looking at response as measured by the BSR RA criteria (reduction in DAS28 of 1.2), data were available for 52 patients, all of whom reported a subjective clinical response (based on the opinion of the patient and their physician). The majority (71%) met the BSR RA criteria. Of those who did not, six patients had a secondary loss of efficacy and their disease activity had increased again.
Serious side-effects causing cessation or switching of therapy were only seen in three patients (5%). One patient developed symptoms of a possible myelitis. Two patients developed raised liver function tests on infliximab and switched to etanercept. One patient continues on this, the other switched again due to a skin rash.
Ninety-seven per cent of the patients (n = 58) had a clinical response to the first biologic at 12 weeks. The patients suffering non-response and side-effects are detailed in table 2.
Overall, 90% (54 of 60) of patients in our cohort achieved a significant response to an anti-TNF agent, using switching if required. When considering only those patients who switched for non-response, 40% (four of 10) responded to a second-line agent and half responded (three of six) to a third therapy.
The data were analysed further looking for possible predictors of response. The percentage improvement in joint counts, CRP and DAS28 scores were analysed with respect to the biologic used, disease duration (from time of diagnosis) and disease subtype. There was no significant differences (p>0.22) between responses to treatment seen in any of these sub-analyses (fig 1).
Our data confirm previous observations on the efficacy of TNF blockers in PsA and that the reductions in disease activity are maintained over time. Here, this is measured by the DAS28, which has been shown to be a responsive outcome measure for peripheral joint disease in PsA.15 Perhaps the most important limitation of the DAS28 in PsA is the constraint to a 28 joint count, which excludes disease in the feet. Indeed, this may account for the nine patients who were classified as having made a satisfactory clinical response to treatment but did not achieve an improvement of 1.2 in their DAS28 score.
An important observation from the data reported here is that the incidence of non-response appears to be significantly lower than that in RA.7 Both the rate of primary non-response at 12 weeks and the rate of secondary loss of effect after the 12-week assessment are low. Furthermore, although the response rate after switching seems to be reduced following an initial non-response, the majority of patients (70%) who switched due to non-response obtained a sustained response to either a second- or third-line agent.
When looking at possible predictors of response, similar rates were found regardless of disease duration, drug used and subtype of disease. Nevertheless there are some limitations on interpreting these data as groups become smaller. There are also significant possible biases involved in these analyses. Earlier on, infliximab was most commonly used as a first-line agent as a result of clinical trials. Since the NICE appraisal, etanercept is now used as the first-line therapy in the UK unless contraindicated. Therefore, it is not fair to compare the results of the different agents without taking into account previous loss of efficacy.
An interesting observation is that patients had a similar response rate regardless of disease duration. This suggests that treatment should always be considered in patients with active uncontrolled disease regardless of duration. Another relevant finding from this cohort is that there was no difference in outcome seen within the different disease subtypes, suggesting that anti-TNF treatment is efficacious in polyarticular, oligoarticular and mutilans-type PsA. The caveat for this statement is that the proportion of different PsA subtypes in our cohort is unlikely to be representative of the general PsA population. The proportion of patients with arthritis mutilans was high, related to their recognised poor prognosis.16 The application of the BSR and NICE guidelines for biologic treatment (which requires at least three tender and three swollen joints) also alters the proportion of different subtypes as many patients with an oligoarticular-type pattern will fail to qualify for anti-TNF treatment. Our data suggest nevertheless that this represents a promising treatment option for patients with resistant oligoarticular disease.
In conclusion, these results show that treatment of active PsA with anti-TNF agents leads to a prompt and sustained response with the majority of patients tolerating these drugs well. In addition, non-response is rare with most patients responding to the first agent. The majority of patients who fail to respond to a first agent will respond to a second- or third-line drug, so switching should always be considered in PsA. Further research into this subgroup of patients may allow for the identification of predictors of response.
We thank Barbara Padwell and David Pickles for their help with data collection.
Competing interests: None.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.