Participants were all Korean, aged 18 to 80 years, and taking NSAIDs for the treatment of RA. All satisfied the 1987 American College of Rheumatology (ACR) classification criteria of RA , and had a disease duration of over 3 months and an ACR functional class of I, II or III . Furthermore, all had been taking stable doses of DMARDs, such as, methotrexate, sulfasalazine, hydroxychloroquine, or leflunomide for over 3 months at the screening visit and maintained these doses during the study period. A stable low dose of prednisolone at 10 mg or less per day over 4 weeks was permitted. On randomization, patients were required to have worsened pain, defined as an increase in pain visual analog scale (VAS) score at least 10 mm or of 20% from baseline during the washout period, and to have moderate or severe arthritis, which was defined as a disease activity score of 28 (DAS28) ≥3.2.
Patients were excluded if they had a history of hypersensitivity to NSAIDs or a serious cardiovascular, liver, kidney or blood disease or another autoimmune disease. Those with a peptic ulcer or GI bleeding confirmed by endoscopy or radiography within 6 months of enrollment or who could not discontinue their GI medication, such as, H2 blocker, misoprostol or proton pump inhibitor were excluded. Patients were also excluded if they had been treated with intra-articular corticosteroid within 4 weeks prior to screening or had previously been administered biologic DMARDs, such as, infliximab, adalimumab, etanercept, anakinra, or abatacept within 6 months or rituximab within 1 year of enrollment. Pregnant women, breastfeeding women, and women of childbearing potential not using an appropriate method of contraception, such as, condoms, intrauterine devices, and oral contraceptives, were excluded.
This was a 6-week, multicenter, randomized, double-blind, double-dummy, parallel-group, phase III, non-inferiority clinical trial conducted at 14 medical centers in Korea from October 2010 to October 2011. The study protocol was approved by the Institutional Review Boards at each of the participating institutes, which were; Seoul National University College of Medicine-Seoul National University Hospital Institutional Review Board (IRB), Gachon University Gil Medical Center IRB, The Catholic University of Korea, Yeouido St. Mary's Hospital Catholic IRB, Kyung-Hee University Hospital IRB, Dong-A University Hospital IRB, Pusan National University Hospital IRB, Severance Hospital IRB, Ewha Womans University Medical Center IRB, Chonnam National University Hospital IRB, Chonbuk National University Hospital IRB, Chung-Ang University Hospital IRB, Chungnam National University Hospital IRB, and Hallym University Sacred Heart Hospital IRB. Written informed consent was obtained from each patient before study enrollment. No important change to study methods was made after trial commencement.
Study medications and comparators were packed in identical appearance and consecutively numbered according to the allocation sequence. Random allocation sequence generated using SAS 9.1v software (SAS Institute, Cary, NC) by an independent statistician and was stratified by center with a 1:1 allocation using random block sizes of 4 and 6. After a 3–14 day washout period depending on the half-lives of NSAIDs, patients who experience pain worsening were assigned order numbers by investigators at each center and received the corresponding medication from pharmacists. A double-dummy design was implemented because the pelubiprofen tablet and celecoxib capsule differed in appearance and dosage. Patients in pelubiprofen group received a pelubiprofen 30 mg tablet three times daily and a celecoxib placebo capsule twice daily. Patients in celecoxib group received a 200 mg Celebrex® capsule twice daily and pelubiprofen placebo tablet three times daily. The allocation sequence was concealed from the researcher enrolling and assessing participants and who kept randomization envelopes. The sponsor and the principal investigators at each center were in charge of the envelopes, which were opaque, sealed, and not opened until trial completion. Blinding could only be broken in emergency situations for reasons of patient safety.
The drug administration period was 6 weeks. During weeks 0, 2 and 4, participants were given sufficient study medication to last until their next scheduled visit plus an additional amount to accommodate visits scheduled within the allowed visit variance (±4 days), and were asked to return all unused medication during visits on weeks 2, 4, and 6. Compliance with study medication was monitored by noting returned medication on case report forms. Rescue medication was allowed in the form of an acetaminophen extended-release (ER) 650 mg tablet. The total amount of rescue medication used was tracked using an accountability procedure.
Efficacy and safety profiles
The primary end point was pain decrease from baseline to week 6 as determined using a 100 mm pain VAS. Patients specified general pain experienced during the previous 48 hours by indicating a position along a continuous line between the two end-points (0 mm = no pain and 100 mm = severest pain imaginable) under the supervision of the study investigator . Changes in VAS pain from week 0 to 6 in the two groups were compared using the t-test and are presented with 97.5% confidence intervals (CI). We hypothesized that pelubiprofen was non-inferior to celecoxib if the lower limit of the 97.5% confidence interval for the treatment difference [(pain reduction in pelubiprofen group) – (pain reduction in celecoxib group)] was more than −10 mm. Secondary end points were as follows: (1) reduction of Korean health assessment questionnaire (KHAQ) score; (2) decreased duration of morning stiffness; and (3) decrease in the frequency and total dose of rescue drugs after 6 weeks of treatment.
Differences in the duration of morning stiffness between baseline and week 6 were calculated and presented in median. Duration of morning stiffness is reported in minutes and durations of more than 360 minutes are reported as 360 minutes. The KHAQ consisted of 20 items in 8 categories that addressed; (1) dressing and grooming, (2) arising, (3) eating, (4) walking, (5) hygiene, (6) reach, (7) grip, and (8) common daily activities. For each of these categories, patients reported the amount of difficulty experienced when performing 2 or 3 items using a 4 score: 0 = without any difficulty, 1 = with some difficulty, 2 = with much difficulty, and 3 = unable to do. A category score was determined using the highest score of the items in that category. The sum of all category scores was then divided by the number of categories answered to produce a single disability index score [10, 11]. Usages of rescue medication were assessed at baseline and on treatment weeks 2, 4, and 6 and analyzed for frequencies and amounts of acetaminophen ER 650 mg tablets used during the previous 2 weeks by repeated measure analysis of variance (ANOVA). All the other efficacy assessments were performed at baseline and week 6.
The investigators, who were blinded to treatment group identities, described and assessed all clinical and laboratory adverse events (AEs). Clinical AEs were evaluated by physical examination and general questioning at every visit, and laboratory AEs were evaluated by complete blood count, serum chemistry and urinalysis at screening and last visit (week 6). All AEs were categorized according to the likelihood of a causal relationship with the study drug, that is, as definitely related, probably related, possibly related, probably not related, definitely not related, or unknown . Causal relationships of AEs were determined based on clinical judgment by the investigators. Adverse drug reactions (ADRs) were defined as AEs that were at least possibly related to study medications. Serious AEs and ADRs were defined as those associated with any of the following: death; an event associated with a high risk of mortality; an event requiring hospitalization; or the development of a permanent disability or congenital malformation.
Baseline patient characteristics in the pelubiprofen and celecoxib study groups were analyzed. Continuous variables are expressed as means ± standard deviations (SD), and categorical variables as numbers and percentages.
For the sample size calculation, we referred to the report of Song et al. . In this non-inferiority trial using celecoxib, reduction in 100 mm VAS pain was 17.87 mm and its SD was 19.06 mm. Based on the non-inferiority margin of −10 mm reported, pelubiprofen was considered non-inferior to celecoxib if the lower limit of the 97.5% CI of treatment difference [(pain reduction in the pelubiprofen group) – (pain reduction in the celecoxib group)] was more than −10 mm. With a 1-sided significance level of 0.025 and assuming a SD of 19.06 mm, a sample size of 58 patients per group was determined to provide a study power of 80%, and thus, 73 patients were deemed to be required given a 20% dropout rate. Secondary outcomes were also tested for non-inferiority using one-sided tests.
For the efficacy profile analysis, per protocol (PP) populations were used for the main analysis and intent-to-treat (ITT) populations were used for supplementary analysis with the input of any missed observations according to the last observation-carried-forward method.
The safety population included all patients who had received ≥1 dose of study medication after randomization and incidences of AEs and of ADRs in the study groups were compared using the χ2 test or Fisher’s exact test. The ITT population included all patients who had received ≥1 dose of study medication after randomization and were available for analysis of the primary end point. The PP population included only those patients who had completed the protocol at the end of the study period. Protocol violators and patients with <80% drug compliance were excluded from the PP analysis. Statistical analysis was conducted using SAS version 9.1 (SAS Institute Inc., Cary, North Carolina). Statistical significance was accepted for p values <0.05.