The Rebif New Formulation Versus Betaseron Tolerability Study (REFORMS) (ClinicalTrials.gov identifier: NCT00428584) was a randomized, multicenter, 2-arm, Phase IIIb study conducted at 27 clinical sites in the US. The study consisted of a 12-week randomized comparative phase, which was followed by a safety-extension phase of up to 112 weeks (range 82–112 weeks). The study was open-label, except for blinded assessments of ISRs. The initial central Institutional Review Board (IRB) submission was approved by Coast IRB, Colorado Springs, Colorado and, later, Schulman Associates IRB, Cincinnati, Ohio. For those sites that were not permitted to use a central IRB for study approval, submissions were made to the local IRB. This study was performed in accordance with the study protocol, the Declaration of Helsinki, the International Conference on Harmonization (ICH) Harmonized Tripartite Guideline for Good Clinical Practice (GCP), and all applicable regulatory requirements. Patients provided written informed consent for participation in the study.

Eligible patients were 18–60 years of age, had a primary diagnosis of RRMS as defined by the Poser or 2005 revised McDonald criteria 12 13 , and had not previously received IFNβ treatment. Patients were not eligible if they had used any other approved disease-modifying treatment for MS (e.g. glatiramer acetate) or any cytokine or anti-cytokine treatment within 3 months before study initiation, used any immunomodulatory or immunosuppressive treatment within 12 months before study initiation, used any investigational drug or experimental procedure within 12 weeks before screening, received oral or systemic corticosteroids or adrenocorticotropic hormone within 30 days of study initiation, or used other injectable medications on a regular basis during the week before screening. Other exclusion criteria included having an alternative diagnosis to RRMS and being pregnant or breastfeeding. Women of childbearing potential were required to use appropriate contraception. All patients provided written informed consent.

Patients were randomized 1:1 to receive either the serum-free formulation of IFNβ-1a 44 μg sc tiw or IFNβ-1b 250 μg sc qod for the 12 weeks of the comparative phase. Treatments were allocated using a computer-generated randomization code. The doses of IFNβ-1a and IFNβ-1b were up-titrated at the beginning of the study according to the US prescribing information for each drug (Figure 1) 14 15 . Following the 12-week comparative phase, all patients received the serum-free formulation of IFNβ-1a 44 μg sc tiw during the safety-extension phase. Patients who transitioned from IFNβ-1b to IFNβ-1a could be up-titrated to the full dose of IFNβ-1a at the discretion of the investigator. Patients who did not wish to transition from IFNβ-1b to IFNβ-1a were withdrawn from the study. The length of the extension phase varied between 82 and 112 weeks, depending on the patient’s date of enrollment. The extension phase ended within 14 days of when the last enrolled patient completed the last visit at Week 94.

All patients self-administered IFNβ using the Rebiject II® autoinjector (EMD Serono, Inc., Rockland, MA, USA) with a 29-gauge needle for IFNβ-1a or Betaject® (with a 27-gauge needle), Betaject Lite® (with a 30-gauge needle), or Betaject® 3 (with a 27-gauge needle) (Bayer HealthCare Pharmaceuticals Inc., Montville, NJ, USA). Acetaminophen was given prophylactically at the discretion of the treating physician and dosed as needed to ameliorate constitutional symptoms (e.g. fever, myalgia, and FLS). Nonsteroidal anti-inflammatory drugs were given and dosed as needed at the discretion of the treating physician if acetaminophen failed to alleviate or prevent constitutional symptoms or if patients were allergic to, or unable to tolerate, acetaminophen.

Patient-reported pain was evaluated in a visual analog scale (VAS) diary and the Short-Form McGill Pain Questionnaire (SF-MPQ) 16 . Patients used the VAS diary to record the level of pain on a scale from 0 mm (no pain) to 100 mm (worst possible pain), immediately before, immediately after, 10 min after, and 30 min after the injection. The SF-MPQ also included a VAS for patients to record the level of the maximum amount of pain experienced during the 60 min after injection, from 0 mm (no pain) to 100 mm (worst possible pain). In addition, patients were requested to describe the types of pain that they experienced during the 60 min after injection. Patients completed the VAS diary and the SF-MPQ after every injection during the comparative phase and for the first 4 weeks of the safety-extension phase.

The Multiple Sclerosis Treatment Satisfaction Questionnaire (MSTSQ) adapted from Cramer et al. 17 included patient assessments of mood, treatment satisfaction, FLS, and ISRs. The MSTSQ was issued to patients at Weeks 2, 4, 6, 8, 12, 16, 24, 36, and 48. Mean values are reported for each treatment phase.

ISRs were assessed at each visit during the first 48 weeks by a healthcare professional who was blinded to treatment assignment. ISR measures included the diameter of injection-site redness, injection-site swelling, bruising, and consideration of patient-reported itching, within 72 h of the most recent injection.

Compliance was recorded throughout the study and was defined as the actual number of injections divided by the expected number of injections, expressed as a percentage. Safety assessments included AEs (coded to system organ class and preferred term using the MedDRA dictionary [Version 9.1] and summarized by severity and relationship), vital signs, hematology, and serum chemistry. Analgesic use among patients with and without AEs related to FLS was summarized by treatment group during the comparative phase and by treatment group and overall population during the extension phase.

The primary endpoint was the mean change in the VAS diary pain score from pre-injection to 30 min post-injection over the first 21 injections of full-dose IFNβ-1a and IFNβ-1b treatment (“full-dose period”). Due to the different titration schedules and dose frequencies of each treatment, the first 21 full-dose injections were administered during Weeks 5–11 in the IFNβ-1a group and during Weeks 7–12 in the IFNβ-1b group (Figure 1). Secondary endpoints included mean changes in the VAS diary pain score from pre-injection to immediately post-injection and 10 min post-injection; MSTSQ assessments; rater-blinded assessment of the mean diameter of injection-site redness; and SF-MPQ assessments, including the proportion of patients pain-free as recorded on the SF-MPQ VAS. Types of pain and severity experienced by the patient were also recorded on the SF-MPQ. The number of relapses and severity were secondary efficacy endpoints. Relapses were patient-reported and not objectively assessed; the number and severity of relapses were observational clinical assessments.

The primary analysis population was the intent-to-treat population (all patients randomized to treatment). The safety population consisted of all patients who received at least one injection of study drug. The safety-extension population consisted of all patients who received at least one injection of study drug and had available extension phase data. Baseline characteristics of the two treatment groups were compared using analysis of variance (ANOVA) with effects for treatment group and pooled site for continuous variables and the Cochran–Mantel–Haenszel general association test, adjusted for pooled site, for categorical variables.

The null hypothesis was that there would be no difference in mean change in VAS pain score at 30 min post-injection from pre-injection across the treatments at full dose. The primary endpoint was evaluated with a two-way ANOVA model on signed ranked data, including treatment group and pooled site as main effects. The same method was also used to analyze treatment comparisons of the mean changes in the VAS diary pain score from pre-injection to immediately post-injection and 10 min post-injection, as well as mean SF-MPQ pain score at 60 min post-injection. An ANOVA model was used for between-group comparisons; for MSTQ scores, treatment group and pooled site were main effects; for injection-site redness, treatment group and site were main effects. The proportion of patients pain-free on SF-MPQ VAS was analyzed using the Cochran–Mantel–Haenszel general association test, adjusted for site, or Fisher’s exact test, if appropriate. Injection-site swelling, bruising, and itching were compared between groups using a Cochran–Armitage trend test. In the comparative phase, all statistical tests were two-sided and used a significance level of α = 0.05. No adjustment was made for multiple comparisons.

Patient-reported relapses during the comparative phase were compared using a Poisson regression model with the total number of relapses as the dependent variable and treatment group and pooled site as independent variables.

A total of 100 patients (50 per arm) was calculated to provide at least 90% power to detect the difference between treatment groups for the primary objective, when the expected treatment effect size (the difference between the treatment groups divided by the standard deviation [SD]) was at least 0.735. The effect size was based on a difference between the treatment groups of 0.025 mm and an SD value of 0.034 mm. The difference between the treatment groups was based on a mean change of 0.1 mm in the IFNβ-1a group and 0.125 mm in the IFNβ-1b group, and assumed that the mean VAS diary pain scores at pre-injection in the two treatment groups were similar. The calculation also assumed a two-sided Wilcoxon rank sum test, a common SD of the change of ≤0.034 mm, and a Type I error rate of 5%.

Between May 2006 and July 2009, a total of 129 patients were enrolled: 65 were randomized to IFNβ-1a and 64 to IFNβ-1b (Figure 2). Patient baseline characteristics (Table 1) did not differ significantly between groups. Fifty-six patients in the IFNβ-1a group completed the comparative phase and entered the safety-extension phase, and these were termed the “Always IFNβ-1a” group (Figure 2). Of the 63 patients in the IFNβ-1b group who completed the comparative phase, 60 entered the extension phase and were termed the “Delayed IFNβ-1a” group. During the extension phase, the mean (SD) duration of treatment with IFNβ-1a was longer in the Always IFNβ-1a group (436 [251] days) than in the Delayed IFNβ-1a group (338 [260] days).

^aBased on total number of patients with relapses during the 12 months before informed consent.

BMI, body mass index; IFN, interferon; MS, multiple sclerosis; SD, standard deviation.

Treatment compliance during the comparative phase was high: 64/65 (98.5%) patients receiving IFNβ-1a and 62/64 (96.9%) patients receiving IFNβ-1b adhered to their medication schedule ≥90% of the time. Similarly, of the patients enrolled in the safety-extension phase, 102/116 (87.9%) were ≥90% compliant. Throughout the study, analgesics were used by the majority of patients. The percentage of patients using analgesics did not differ greatly between groups in either phase of the study; during the full-dose period of the safety population, analgesic use in the IFNβ-1a and IFNβ-1b was 61.5% and 57.8%, respectively, and mean total (SD) dose was 16,652 (16,941) mg and 12,862 (10,159) mg, respectively. Post-injection analgesic use for the treatment of an ISR was recorded by 1/64 patients in the IFNβ-1b group. During the full dose period, concomitant analgesic use for treatment of an ISR was recorded by 4/65 patients in the IFNβ-1a group and by 1/64 patients in the IFNβ-1b group. During the extension phase, analgesic use in the Always IFNβ-1a group and Delayed IFNβ-1b group was 80.4% and 78.3%, respectively.

The most commonly used (>3 individuals) analgesics during both study phases included ibuprofen, paracetamol, naproxen, and acetylsalicylic acid.

Injection depth may affect pain and ISRs, although the various injection depth options that patients used were not recorded. This study mirrored general-practice use of auto-injectors, whereby patients choose their individual depth setting.

The utilization of VAS pain scores at smaller-ranging scales, for instance, from 0 to 10 mm, may have been more useful in obtaining a more sensitive picture of injection-site pain with these therapies. Also, except for inspection of injection-site redness, the study was not blinded. Lack of patient blinding may have influenced some of the patient-rated measures such as ratings of relapse, pain, and FLS. The clinical relevance of patient-reported relapse data is unclear and so, with the lack of objective assessment of relapses, results of those analyses should be interpreted with caution. In the statistical analyses of patient-rated measures, no adjustments for multiple comparisons were made. Although a limitation, no statistically significant differences across groups on any of the patient-rated assessments were observed, suggesting that this aspect of the study design was not a serious weakness of the study. However, to ensure certainty in any future analyses, the blinding of patients is recommended to reduce any potential of bias in the patient-rated measures.

Although the safety-extension phase was useful in gaining data on longer-term tolerability to IFNβ-1a, large dropouts and confounding variables such as treatment switch from a high dose of IFNβ-1b to IFNβ-1a did not permit systematic analyses from which valid interpretations and conclusions could be made.