The tumour necrosis factor (TNF)α antagonists infliximab, etanercept and adalimumab have been shown to be effective in controlling symptoms in ankylosing spondylitis (AS) and psoriatic arthritis (PsA), and in retarding radiological progression in PsA. These drugs are currently reserved for patients who are non-responsive to conventional treatments. Although there have been no direct comparison studies, the various TNFα antagonists appear to have similar efficacy, with a clinical response in 50–60% of patients. However, around 30% of patients with AS or PsA treated with these drugs withdraw from treatment due to adverse events (AEs) or loss of efficacy.1 Data from clinical trials or national registers suggest that patients with rheumatoid arthritis (RA) who do not tolerate or do not respond to a TNFα antagonist may be successfully treated with another.2^–5 To date, there have been few data on spondyloarthropathy (SpA). Delaunay et al reported a favourable clinical outcome in 15 patients with SpA who switched from infliximab to etanercept because of inefficacy or intolerance.6 An open-label, prospective study has recently demonstrated the efficacy and tolerability of etanercept in 23 patients with active AS who were resistant or intolerant to infliximab.7

In this paper, we report the results of an ongoing, longitudinal, observational study, and evaluate the clinical response after switching from one TNFα antagonist to another in patients with AS and PsA within a ‘real-life’ clinical setting.

PATIENTS AND METHODS

In this ongoing, longitudinal, observational study, we prospectively collected data since 2000 on efficacy and safety for patients starting biological treatments in our rheumatology division.

The present analysis was restricted to patients with a diagnosis of SpA who switched from one TNFα antagonist to another, with a minimum of 6 months’ follow-up by the end of December 2006 (the first SpA patient started treatment in December 2001). Patients with AS were classified according to the modified New York Criteria,8 and patients with PsA according to the Moll and Wright criteria modified by Helliwell.9 The choice of biological agent was based on clinical considerations only; thus, these patients represented a ‘real-life’ sample of subjects treated with TNFα antagonists. Infliximab 3–5 mg/kg was administered intravenously at weeks 0, 2 and 6, then every 6–8 weeks; etanercept (25 mg twice weekly) and adalimumab (40 mg alternate weekly) were given subcutaneously.

Clinical assessment

Patients were evaluated by the same rheumatologist at baseline (before starting the TNFα antagonist), every 3 months and at the last administration of the drug. Data, including demographics, diagnosis, date of diagnosis, comorbidities, past and present treatments, TNFα antagonist prescribed and date of beginning, concomitant medications, were recorded on a standardised form. Current disease activity in patients with AS was measured by the modified Ritchie Index,10 and the Bath Ankylosing Spondylitis Disease Activity Index (BASDAI).11 Clinical assessment in patients with PsA was performed according to the Psoriatic Arthritis Response Criteria (PsARC).12 Each patient completed the Health Assessment Questionnaire (HAQ). Patient and physician global assessments (on a visual analogue scale, VAS) were also performed.13 Response to TNFα antagonist was defined by a 50% response on the BASDAI (BASDAI 50)14 for patients with AS, and by PsARC for patients with PsA. Drug discontinuation was based on the rheumatologist’s opinion and the reason of withdrawal recorded as lack of efficacy (patients who never reached a satisfactory response), loss of efficacy (patients who relapsed after an initial good response), AEs or other. The wash-out period between TNFα antagonists was 6 weeks.

Statistical analysis

Qualitative differences between subgroups were analysed by the χ² and Fisher exact tests. The Wilcoxon paired test was used to compare quantitative variables in the same group. Statistical significance was set at p<0.05.

RESULTS

The study comprised 589 anti-TNFα-naive patients, of whom 165 had a diagnosis of SpA. In total, 22 patients with SpA (7 patients with AS (mean age 33.5 years, range 22–49 years), and 15 with PsA (mean age 47.9 years, range 24–57 years)) received >1 TNFα antagonist (table 1). Two patients with PsA were consecutively treated with all three TNFα antagonists.

A clinical response was seen in 12/16 (75%) patients who changed from infliximab to etanercept, and in 4/7 (57.1%) who switched from etanercept to adalimumab (figs 1 and 2). Patients who switched because of AEs and those who changed because of inadequate efficacy (ie patients who never reached a satisfactory response, or patients who relapsed after an initial good response) presented a similar clinical response (70% and 61.5%, respectively). The mean duration of etanercept treatment was longer than the previous infliximab treatment (19.7 vs 14.1 months, respectively; NS). The mean duration of treatment was not different in patients treated with adalimumab after etanercept (7.3 vs 8.1 months, respectively; NS) (table 1).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1 Clinical parameters in patients with AS (n = 6) who switched from infliximab to etanercept. Box and whiskers plot (median, quartiles, range and possible extreme values) of (A) Ritchie Index, (B) patient and (C) physician global assessment (on visual analogue scale); (D) percentage of patients who had 50% improvement on the Bath Ankylosing Spondylitis Disease Activity Index (BASDAI 50). Values shown are the mean values at baseline (before infliximab treatment), after 3 months of infliximab treatment, at last visit while on infliximab, at baseline (before etanercept treatment) and after 3 months of etanercept treatment.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2 Clinical parameters in patients with psoriatic arthritis (PsA) who switched from infliximab to etanercept (n = 10, column 1) and from etanercept to adalimumab (n = 7, column 2). Box and whiskers plot (median, quartiles, range and possible extreme values) of (A) tender joint count, (B) swollen joint count, (C) patient and (D) physician global assessment (on visual analogue scale), and (E) percentage of patients who met Psoriatic Arthritis Response Criteria (PsARC). Values shown in column 1 are the mean values at baseline (before infliximab treatment), after 3 months of infliximab treatment, at last visit while on infliximab, at baseline (before etanercept treatment) and after 3 months of etanercept treatment. Values shown in column 2 are the mean values at baseline (before etanercept treatment), after 3 months of etanercept treatment, at last visit while on etanercept, at baseline (before adalimumab treatment) and after 3 months of adalimumab treatment.

Six patients with AS switched from infliximab to etanercept (mean age 31.8 years, range 16–49) (table 1). After 3 months of etanercept, the proportion of responders grew from 0% (baseline before etanercept) to 83.3% (p<0.04). Changes in Ritchie Index, patient and physician global assessment, and BASDAI 50 are shown in fig 1.

Ten patients with PsA switched from infliximab to etanercept (mean age 47 years, range 24–56). After 3 months of etanercept, the proportion of PsARC responders increased from 10% (baseline before etanercept) to 70% (fig 2), and HAQ score decreased significantly (p<0.01). Seven patients with PsA switched from etanercept to adalimumab (mean age 49.8 years, range 39–57). After 3 months of adalimumab, the proportion of PsARC responders increased from 14.3% (baseline before adalimumab) to 57.1% (fig 2).

One patient with AS received etanercept as first TNFα antagonist (table 1). He had a satisfactory response at 3 months, which was subsequently lost, thus he was switched to infliximab. After 3 months, a clinical response was seen.

DISCUSSION

The findings of this longitudinal, observational study on a selected population of patients with SpA indicate that the failure of a first anti-TNFα agent does not preclude the response to another, irrespective of the reason for switching.

The introduction of biological agents has allowed a radical change in the standard of treatment in SpA. Clinical response rates to anti-TNFα agents range from 43% to 71% by BASDAI 50 in patients with AS and from 62% to 87% by PsARC in subjects with PsA.1 However, a significant proportion of patients withdraw from treatment because of failure or poor tolerability. Because the TNFα antagonists differ in chemical structure, mechanism of action and safety profile, there is a strong rationale for switching. Several reports support this possibility and, switching to another TNFα antagonist has now become a common practice in RA.2^–5 Very few data are currently available for patients with SpA. Delaunay et al, in their retrospective evaluation, described 15 patients with various SpAs who switched from infliximab to etanercept because of AEs or inadequate efficacy. A clinical response was observed in 42.8% of the seven patients with AS and in all patients with undifferentiated SpA or PsA.6 A prospective study has recently reported the efficacy and tolerability of etanercept in 23 patients with AS who were resistant or intolerant to previous infliximab treatment. In particular, a 54-week ASsessment in Ankylosing Spondylitis (ASAS) 20, ASAS 50, and ASAS 70 response rate was reported in 74%, 61%, and 39% of patients, respectively.7 In the present study, we evaluated the efficacy of switching between TNFα antagonists for patients with AS and PsA. Of the patients who switched from infliximab to etanercept, a clinical response was seen after 3 months in 83% of patients with AS (BASDAI 50) and in 70% of patients with PsA (PsARC). Of the patients with PsA who switched from etanercept to adalimumab, 57.1% showed a clinical response. Patients who switched because of AEs and those who changed because of inadequate efficacy presented a similar clinical response. Hence, our data concur with previous studies on RA, in which a better response to a second TNFα antagonist was observed in most patients, regardless of the reason of failure of the previous treatment.3^–5 In our population, 81.3% of patients who had switched from infliximab to etanercept continued the treatment. Conversely, only 57.1% of patients who had changed from etanercept to adalimumab maintained the treatment. Interestingly, two of the three patients who stopped adalimumab because of inadequate response had already failed the other two TNFα antagonists. This observation seems to confirm previous data on RA patients, suggesting that the failure of two TNFα inhibitors predicts ineffectiveness to the third.15

There are some limitations to our study. First, it was a prospective, observational design without randomisation of treatment options. Second being a real-life study, the decision to withdraw one TNFα antagonist and to switch to another depended only on the treating doctor’s judgement.

In conclusion, the results of this study suggest that patients with SpA with inadequate response or AEs to one TNFα antagonist may be successfully treated with another, regardless of the reason for switching. Considering the sample size and the design of this study, larger prospective trials are warranted in order to confirm this observation.