Patients

This was a multicentre, open label extension study in patients from a previous 24 week, double blind RCT of etanercept in AS. Two hundred and fifty seven of the original 277 randomised patients began the open label extension phase (fig 1⇓). The study was conducted at 28 sites in America, Canada, the Netherlands, Germany, and France. All patients gave informed consent before screening and enrolment in the preceding clinical trial.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

 Patient disposition in an open label trial examining sustained durability and tolerability of etanercept in patients with AS. BIW, twice weekly.

At RCT enrolment, patients had active AS and fulfilled the modified New York criteria. Detailed information on the inclusion and exclusion criteria for the RCT have been presented elsewhere.² In this extension study, patients were allowed to continue to take hydroxychloroquine, sulfasalazine, or methotrexate and/or to change the doses at the discretion of the principal investigator at each site.

All patients who completed the RCT were eligible for inclusion in this extension study. Additionally, patients from the RCT who had dropped out at or after 12 weeks because of lack of efficacy were eligible to enrol in the extension study, provided that the patient had returned for all the required follow up visits. Patients who had discontinued because of adverse events were eligible to enrol provided that the event(s) had resolved and that the clinical investigator still considered the patient a candidate for etanercept treatment.

Study design and drug

Previously, patients received either subcutaneous (SC) etanercept 25 mg (n = 138) or placebo (n = 139) twice weekly²; in this extension study, all patients received SC etanercept 25 mg twice weekly for up to 72 weeks (fig 1⇑). For patients who had received etanercept for 24 weeks in the RCT, this represented almost 2 years of continuous etanercept treatment. The study drug was supplied as a sterile lyophilised powder containing 25 mg etanercept (Enbrel; Immunex, subsidiary of Amgen) for reconstitution and self administered SC injection.

Efficacy

The primary outcome of the preceding RCT was a comparison between the proportion of patients attaining an ASAS 20 response at 12 weeks among patients receiving etanercept and patients receiving placebo. To qualify as an ASAS 20 responder, a patient had to show an improvement of 20%, equating to a minimum improvement of 10 units on a 100 mm visual analogue scale in at least three of four measures: patient’s global assessment; pain; physical function; and inflammation, with no deterioration (defined as a 20% worsening and absolute change of at least 10 units) in the fourth criterion.⁸ Physical function was measured by the average of the 10 BASFI questions⁹ (range 0 = easy to perform tasks to 100 = impossible to perform tasks) and inflammation was represented by the average of the final two questions on the BASDAI assessing severity and duration of morning stiffness¹⁰ (range 0 = none to 100 = very severe).

The durability of response in those patients continuing etanercept from the RCT, as assessed by the proportion of ASAS 20 responders at the end of 72 weeks of open label treatment, was determined. Secondary outcome measures included achievement of the ASAS 50 (50% and 20 units improvement in the four domains for ASAS 20 criteria, without worsening in the remaining domain) and ASAS 70 (70% improvement and 30 units in the same four domains, without worsening in the remaining domain). Other outcome measures included the effect of etanercept treatment on nocturnal and total back pain, measurements of spinal mobility (chest expansion, modified Schober’s index, and occiput to wall measurements), peripheral joint counts, acute phase reactants, and assessor’s global assessment.

We also incorporated the two alternative definitions of the ASAS criteria, based on those set forward by the ASAS international working group, ASAS for Disease Controlling Anti-Rheumatic Therapies: ASAS 40 response and ASAS 5/6.¹¹

The ASAS 40 response requires a 40% improvement relative to baseline plus absolute improvement of at least 20 units in any three of four of the ASAS criteria listed above, with no worsening in the remaining criterion.

Our modification of the ASAS 5/6 required a 20% improvement in five of six criteria—the four criteria of the ASAS response, a measure of spinal mobility, and an acute phase reactant—with no worsening in the remaining criterion. In addition, a 20% improvement in patient’s global assessment; pain; physical function; and/or inflammation was required to equate to an absolute improvement of at least 10 units. This definition is slightly more restrictive than the ASAS 5/6 proposed by Brandt et al¹¹ and is likely to yield a slightly smaller proportion of responders than would have been seen with the original definition. The measures of spinal mobility and acute phase reactant prespecified in the statistical analysis plan were chest expansion and C reactive protein (CRP), respectively. Brandt et al used lumbar lateral flexion as the measure of spinal mobility.¹¹

Safety evaluation

All patients who received at least one dose of the study drug in the double blind or open label periods were evaluated for safety. The following variables were assessed: results of a physical examination, vital signs, haematology and chemistry profiles, urine analysis, antibodies against etanercept, premature discontinuation, adverse events, serious adverse events, and deaths on study. Adverse events and abnormal laboratory values were graded on a scale derived from the National Cancer Institute Common Toxicity Criteria (common terminology criteria for adverse events available online at http://ctep.cancer.gov, accessed 18 May 2005). Serious adverse events were defined as any events that resulted in death, a life threatening adverse experience, inpatient hospitalisation or prolongation of existing hospitalisation, a persistent or major disability/incapacity, or a congenital anomaly/birth defect.

Statistical analysis

The percentage of responders as determined by the ASAS response criteria⁸ was determined over time by original treatment group. Efficacy analyses are reported without statistical inference between treatment groups. For both groups of patients, measurements from the baseline visit of the RCT were used as baseline values. Data were combined with the results of the preceding RCT to assess the durability of the ASAS response rates. Efficacy summary tables include all subjects with data available at each given time point.

Patients

Of the 277 patients who were dosed and analysed in the RCT, 257 patients started the extension study (fig 1⇑). Ten patients in each RCT group did not enrol in the open label trial. Twelve patients receiving etanercept had discontinued during the RCT. Of these, three patients re-enrolled in the open label trial: two patients who had discontinued at or after 12 weeks owing to a lack of efficacy and one patient who had discontinued because of an adverse event. One patient in the etanercept group who had completed the double blind trial declined to enrol in the open label trial (patient refusal). Nineteen patients receiving placebo had discontinued during the RCT. Of these, eight patients who had discontinued because of lack of efficacy enrolled in the open label trial, as did one patient who had discontinued because of an adverse event.

Baseline demographic characteristics of the group who enrolled in the open label extension were therefore very similar to those of the RCT, with a predominantly male (75%) and white (93%) population. The mean age of the group was 41.6 years (range 18–70 years), with a mean disease duration of 10.7 years (range 0.5–35.8). About one third (31.5%) of patients were taking concomitant disease modifying antirheumatic drug treatment, including sulfasalazine (21.8%), methotrexate (11.3%), and hydroxychloroquine (1.6%). At baseline, the mean (SD) patient global assessment was 63.3 (17.6) and the mean (SD) BASFI was 54.0 (20.6).

One hundred and twenty eight of the patients who enrolled were from the etanercept arm of the preceding trial; the other 129 began etanercept at week 0 of the open label extension study. Two hundred and nine patients received at least 72 weeks of etanercept treatment (104 patients continuing etanercept from the RCT and 105 patients starting etanercept in the open label extension), and 95 received a full 96 weeks.

Sixty seven patients receiving etanercept discontinued (fig 1⇑). This included 10 patients who received etanercept in the RCT but did not enrol in the open label extension. Figure 1⇑ lists the reasons for discontinuation.

Durability of response

Patients who had received placebo in the preceding RCT switched to etanercept 25 mg twice weekly at the beginning of the open label period. Response in these patients was rapid and sustained. After receiving 24 weeks of etanercept treatment in the open label period, 70% of these patients had achieved an ASAS 20 response, compared with 57% of etanercept patients who had achieved an ASAS 20 response at week 24 of the RCT.² Improvements were seen in all levels of ASAS response in those who entered the open label extension study (fig 2⇓). All mean measures of spinal mobility, including chest expansion, occiput to wall measurement, and modified Schober’s index, also improved in this group by the end of 72 weeks of etanercept treatment (table 1⇓).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

 (A) ASAS 20, ASAS 50, and ASAS 70 responses in patients who received placebo in the preceding RCT and enrolled in the open label extension study, by length of treatment. (B) Durability of response, indicated by ASAS 20, ASAS 50, and ASAS 70 responses in patients who received etanercept in the preceding RCT and enrolled in the open label extension study, by length of treatment. BIW, twice weekly; RCT, randomised clinical trial.

Patients who had received etanercept in the preceding RCT and enrolled in the open label extension study showed continued and improved ASAS response for up to 2 years of treatment with etanercept (fig 2⇑). At week 24 of the RCT, 57% of etanercept patients had achieved an ASAS 20 response²; after an additional 72 weeks of open label etanercept (week 96, fig 2⇑), 74% (70/95) of patients had achieved an ASAS 20 response. After 96 weeks of etanercept treatment, almost half (46%) of these patients had attained an ASAS 70 response. All three mean and two of three median measures of spinal mobility in patients who had received etanercept in the RCT showed persistent improvement from baseline (table 1⇑). In both these patients and patients who had received placebo in the RCT, the subgroup of patients with limited spinal mobility (defined as baseline occiput to wall ≠0; baseline chest expansion score <2.5; or baseline Schober’s index <5) showed positive improvements in all mean and all median measures of spinal mobility (data not shown).

Additional outcomes included the individual components of the ASAS response criteria and CRP (table 2⇓). The group of patients who had received placebo in the RCT improved on all measures. The group of patients who had received etanercept in the preceding RCT demonstrated sustained benefit and, in some cases, continued improvement. In a more stringent definition of response, both groups of patients experienced disease improvement as measured by the ASAS 40 and modified ASAS 5/6 response criteria (fig 3⇓).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

 (A) Percentage of patients who attained the modified ASAS 5/6 (20% improvement in five of six criteria—patient’s global assessment, pain, physical function, inflammation, a measure of spinal mobility, and an acute phase reactant—with no worsening in the remaining criterion. In addition, improvement in patient’s global assessment, pain, physical function, and/or inflammation by at least 10 units). (B) Percentage of patients who attained ASAS 40 (40% improvement relative to baseline plus absolute improvement of at least 20 units in any three of four of patient’s global assessment, pain, physical function, and/or inflammation, with no worsening in the remaining criterion).

Safety

Most adverse events reported were of mild or moderate intensity; table 3⇓ lists those that occurred in ⩾5% of patients. Two events occurred in two patients during the open label extension that were considered to be life threatening: an idiopathic third degree heart block and a pulmonary embolus, both of which the investigators considered unrelated to etanercept.

Six patients withdrew from the open label extension because of non-infectious, non-serious adverse events (fig 1⇑), including one patient for an injection site reaction; one patient with dizziness, shortness of breath, and an increased heart rate; one patient with dermatitis and urticaria; two patients with rashes; and one patient with liver steatosis. Seventeen patients experienced serious adverse events, but only four patients discontinued as a result of those events, including one patient who had a flare of Crohn’s disease, one patient with a narcotic addiction, one patient with asthmatic bronchitis, and one patient who had worsening depression. All of these events, however, were identified as unrelated to etanercept by the investigators.

Twenty one per cent of patients had injection site reactions during the open label period; only one patient discontinued from the open label extension because of this reaction. The 21% rate is somewhat lower than the rate seen in the RCT, where 30% of patients receiving etanercept experienced injection site reactions.² One patient discontinued the RCT because of a fever associated with an injection site reaction.

No deaths were reported in either the RCT or the open label extension.

Some of the extra-articular features of the spondyloarthropathies were also noted as adverse events during the trial; patients with a history of these features were not excluded from the trial. Twenty six patients experienced uveitis or iritis, including 13 patients previously receiving placebo and 13 patients continuing etanercept from the RCT. Four of these patients had reported iritis or uveitis during the RCT and most (14/22) others had a pre-trial history of uveitis or iritis. Colitis was reported in five patients (including one case due to C difficile), ileitis was reported in two patients, and there were single reported cases of psoriasis and balanitis.

Five patients experienced serious infectious episodes after starting etanercept. All of these episodes resolved. One patient discontinued because of a non-serious infectious event (otitis media and a furunculus) and one patient discontinued because of a serious infectious event (a flare of hepatitis B in a foreign-born person).

One patient had a positive purified protein derivative (PPD) test but had none of the signs or symptoms associated with tuberculosis; it was unclear from the case report forms why the PPD test had been performed. Two cases of herpes zoster were reported.

No patients discontinued as a result of laboratory abnormalities. One patient had grade 3 increases in liver enzymes on more than one test date; one patient had grade 3 results from alanine aminotransferase and aspartate aminotransferase determinations on the same date, with all other test dates falling within normal ranges. Two patients had single cases of laboratory abnormalities without any associated adverse events amid consistently normal readings: one patient had a single recorded grade 4 laboratory abnormality, with a platelet count of 3×10⁹/l on study day 421 despite consistently registered platelet counts ⩾200×10⁹/l on all other test dates (including study days 337 and 427), and another patient had a single grade 3 increase in liver enzymes on study day 250 (288 U/l) amid consistent test results ⩽40 U/l. The authors believe that these two abnormalities represent spurious laboratory results. A patient with a grade 3 abnormal lymphocyte count before starting the study received only 12 days of etanercept; on study day 63, the patient’s lymphocyte count was still grade 3. One additional patient had a grade 4 absolute neutrophil count at baseline that resolved to essentially normal values (toxicity grade 1) by study day 82.