Randomized, comparative study of interferon β-1a treatment regimens in MS

Patients.

In this study, 677 IFN-naïve patients with definite RRMS16 and Expanded Disability Status Scale (EDSS) scores of 0 to 5.5 were enrolled at 56 centers (15 in Europe, 5 in Canada, and 36 in the United States). All patients had experienced at least two exacerbations of MS in the prior 2 years. Principal exclusion criteria were as follows: previous use of IFN, cladribine, or total lymphoid irradiation; use of glatiramer acetate or cytokine therapy in the prior 3 months; use of IV immunoglobulin in the prior 6 months; and use of other immunomodulatory agents in the prior 12 months.

Treatment.

Patients were randomized to receive either Rebif® 44 μg SC tiw, or Avonex® 30 μg IM qw for 24 weeks. For the purpose of this report, “44 μg tiw” denotes the Rebif® regimen and “30 μg qw” denotes the Avonex® regimen. Avonex® was obtained through commercial suppliers and was administered according to the manufacturer’s recommendations. Rebif® was supplied as a liquid formulation in prefilled syringes. Guidelines were provided for the treating physician based on the World Health Organization (WHO) side-effect severity scale for either dose reduction or interruption. Acetaminophen or nonsteroidal anti-inflammatory drugs (NSAIDs) could be used either as prophylaxis or to treat constitutional influenzalike symptoms associated with the study drugs.

Blinding.

Separate treating and evaluating physicians were designated prior to randomization. Patients were instructed not to disclose their treatment assignment or symptoms related to their treatment regimen, to the blinded evaluating physician, and to cover injection sites before scheduled and relapse-related neurologic examinations. Evaluating physicians communicated with patients only as needed to complete the neurologic examinations and to calculate the functional systems (FS) and EDSS scores.17 MRI evaluation was performed centrally (University of British Columbia MS/MRI Research Group, Vancouver, Canada) by blinded examiners who had no knowledge of a patient’s treatment or outcome. A forced choice questionnaire relating to study blinding was circulated to evaluating physicians after week 48.

Study assessments.

Screening took place four weeks prior to randomization and initiation of therapy. It included a complete neurologic and medical history, physical and neurologic examination, and MRI scanning with proton-density/T2-weighted and pre- and postgadolinium (Gd) T1-weighted sequences. The neurologic examination and MRI scans were repeated on study day 1. Patients returned to the study center for scheduled follow-up every 4 weeks during the initial 24-week treatment period and also at 36 and 48 weeks. Between visits, patients were contacted by telephone to inquire if symptoms of a possible relapse had occurred, and if so, they were referred to the clinical center for evaluation. Detailed neurologic assessments by the evaluating physician, including FS and EDSS scoring, were performed at baseline, 12, 24, 36, and 48 weeks, and as needed for relapse assessment. Blood samples were obtained serially for hematologic, biochemical, and thyroid function testing and for determination of neutralizing antibody (NAb) titers. NAb were measured as previously described.18

A relapse was defined as the appearance of a new symptom or worsening of an old symptom, accompanied by an appropriate objective finding on neurologic examination by the blinded evaluator, lasting at least 24 hours in the absence of fever and preceded by at least 30 days of clinical stability or improvement. An objective finding was defined as an abnormality on examination that was consistent with the reported neurologic symptom. A relapse was recorded only if the blinded evaluator described new findings consistent with the patient’s reported symptoms, and if the treating physician had excluded the possibility of a pseudorelapse.

Relapse severity was based on changes in FS and EDSS scores. A mild relapse was defined as an EDSS increase of one-half point, or a one-point FS change in one to three FS; a moderate relapse as an EDSS increase of one or two points, a two-point FS change in one or two systems, or a one-point change in four or more systems; and a severe relapse as one exceeding the prior criteria. For documented relapses, the unblinded treating physician decided on a course of treatment. Corticosteroids were permitted (methylprednisolone, 1.0 grams IV daily for 3 days). Study-related MRI scans were performed either before beginning methylprednisolone or at least 7 days after the last dose. Therapy for spasticity, depression, pain, bladder control, fatigue, and other MS symptoms was permitted, although the initiation of new medication was discouraged during the study.

MRI scans were performed according to a specific protocol under the direction of the University of British Columbia MS/MRI Research Group. They were done with and without Gd at screening, on study day 1, and every 4 weeks up to week 24. At week 48, a final scan was performed without Gd enhancement. Workshops were conducted to standardize scanning procedures and all sites performed qualifying scans before study initiation. Rejected scans (i.e., those not meeting quality criteria and therefore not acceptable for analysis) were repeated at screening but not at subsequent visits.

Outcome measures.

The primary endpoint was the proportion of patients remaining free of relapses during the 24 weeks, with the primary outcome measure being an odds ratio (OR), adjusted for study center. Relapse rate (relapses per patient per time on study), relapse severity, use of steroids for relapses, and time to first relapse were considered secondary and tertiary clinical outcome measures related to the attack rate. Disability was defined as progression by one point on the EDSS scale confirmed at a visit 3 or 6 months later without an intervening EDSS value that would not meet the criteria for progression. Safety evaluations involved the recording of adverse events, withdrawals owing to adverse events, serious adverse events, and monitoring of laboratory abnormalities. These observations were then carried out for the remainder of the study up to week 48.

The principal MRI outcome measure was the number of combined unique (CU) active lesions per patient per scan, a measure of both active T2 and T1 Gd-enhancing lesions.19 An active T2 lesion was defined as a new or enlarging lesion or a lesion reappearing at a site of previous lesion resolution. A CU lesion was defined as an active lesion on T1 post-Gd or T2 sequences, or both, avoiding double counting. Additional secondary MRI outcome measures included the number of T2 and T1 lesions per patient per scan, the proportion of active scans (T2, T1, and CU) per patient, and the proportion of patients in whom active scans (T2, T1, and CU) either occurred or did not occur during the initial 24 weeks of the trial. At week 48, because Gd was not administered, only T2 lesions were counted and measured.

Statistical considerations.

Ethics.

The study was conducted in accordance with the Declaration of Helsinki. Approval was obtained from the institutional review boards or ethics committees of all participating institutions before study initiation, and written informed consent was obtained from all patients before any study-related procedures were performed.

Patient disposition and baseline characteristics.

Of 767 patients screened at the start of the study, 339 were randomized to receive IFNβ-1a 44 μg SC tiw and 338 to receive IFNβ-1a 30 μg IM qw (figure 1). Ninety patients failed screening by not meeting the inclusion/exclusion criteria. The mean time between screening and the first dose of therapy was 29.3 days. One patient randomized to 30 μg qw did not receive treatment. The distribution of patients who completed 24 weeks and 48 weeks of treatment was similar in both groups (95% for 44 μg tiw and 96% for 30 μg qw at 24 weeks; 93% for 44 μg tiw and 94% for 30 μg qw at 48 weeks).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1. Patient disposition over 48 weeks.

There were no significant differences in any baseline variable between treatment groups (table 1). Approximately 91% of patients were Caucasian and 75% were women. The baseline EDSS scores ranged from 0 to 5.5, but overall, the level of disability was mild (see table 1). All baseline MRI variables were also similar for the two treatment groups (see table 1). The mean numbers of CU and T1 Gd-enhancing lesions were slightly different because of a few outliers that skewed the mean value; however, median values were identical and the distribution of lesions by group was very similar (see table 1, figure 2). The statistical analysis included baseline MRI activity as a covariate in the regression model so that differences in baseline MRI measures would not bias the results.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2. Numbers of patients according to the number of combined unique (CU) active lesions between screening and the first dose of drug (4 weeks). Black bars = interferon (IFN) β-1a 44 μg three times weekly; white bars = IFN β-1a 30 μg once weekly.

Blinding.

Evaluating physicians guessed treatment allocation correctly in 52% of patients (47% correct in the 30 μg qw group and 57% in the 44 μg tiw group). Protocol violations wherein treating physicians served as evaluators occurred in 1% of patients in both groups.

Efficacy.

Over the initial 24 weeks of treatment, 75% (254/339) of patients in the 44 μg tiw group and 63% (214/338) of those in the 30 μg qw group remained relapse free (figure 3). The OR, adjusted for center, was 1.9 (95% CI, 1.3 to 2.6; p = 0.0005), indicating a relative increase of 90% in the odds of remaining relapse free during the first 24 weeks of therapy for patients receiving 44 μg tiw compared with those receiving 30 μg qw. A similar conclusion (favoring high-dose treatment) was reached when the proportion of relapse-free patients was analyzed using either Fisher’s exact test (p = 0.002) or the CMH test (p = 0.001). Over 48 weeks of treatment, 62% (209/339) of patients in the 44 μg tiw group and 52% (177/338) of those in the 30 μg qw group remained relapse free (see figure 3). The OR, adjusted for center, was 1.5 (95% CI, 1.1 to 2.1; p = 0.009), indicating a relative increase of 50% in the odds of remaining relapse free for patients receiving 44 μg tiw compared with those given 30 μg qw.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3. Kaplan–Meier curves illustrating the cumulative probability of patients experiencing a first relapse during the study. There was a 32% relative reduction in the proportion of patients treated with 44 μg three times weekly (tiw) who experienced a relapse during the initial 24 weeks and a 21% relative reduction after 48 weeks. Hazard ratio = 0.70 (p = 0.003). Solid line = interferon (IFN) β-1a 44 μg tiw; dotted line = IFNβ-1a 30 μg once weekly.

Relapse status (“yes” or “no”) was imputed for patients who dropped out before experiencing a relapse. Of such withdrawals, 2 of the 7 patients treated with 44 μg tiw and 3 of the 9 patients treated with 30 μg qw had imputed relapses. Although such imputation of data could alter results, sensitivity analyses showed that even if all patients who withdrew in the 44 μg tiw group were assumed to have had relapses and all those in the 30 μg qw group were assumed to have been relapse free over the pivotal 24 weeks, the between-group comparison (p = 0.006) still favored the 44 μg tiw arm. Forty-one percent (40/98) of relapses in the 44 μg tiw group were detected at unscheduled visits, compared with 36% (48/132) of relapses in the 30 μg qw group over 24 weeks. This difference was not significant. Over 48 weeks, 45% of relapses in both groups were detected at unscheduled visits.

The time to first relapse was prolonged over the course of the study for patients treated with 44 μg tiw (hazard ratio = 0.70; 95% CI, 0.55 to 0.88; p = 0.003). Outcome data on other relapse-related measures also favored 44 μg tiw treatment. Relapse rates (mean number of relapses per patient) were 0.29 for the 44 μg tiw group and 0.40 for the 30 μg qw group at 24 weeks, a 27% relative difference (p = 0.022). At 48 weeks, the difference was less pronounced: 0.54 in the 44 μg tiw group compared with 0.64 in the 30 μg qw group, a 16% relative reduction (p = 0.093). The mean rate of steroid use for relapses was 0.12 courses per relapse in the 44 μg tiw group and 0.19 in the 30 μg qw group (p = 0.017). The number of relapses was fewer for the 44 μg tiw group than for the 30 μg qw group at each severity level, although the proportions of relapses in each group that were mild, moderate, or severe were not different.

Although the study duration was short, disability based on EDSS scores was assessed every 12 weeks for 48 weeks. There were 43 patients in the 44 μg tiw group and 49 patients in the 30 μg qw group with confirmed EDSS progression at 2 consecutive visits 3 months apart, indicating a trend toward reduction of progression risk in the 44 μg tiw group compared with the 30 μg qw group for this small number of patients (hazard ratio = 0.87; 95% CI, 0.58 to 1.31; p = 0.51). Using a more stringent criterion of confirmation after 6 months, 20 patients receiving 44 μg tiw and 28 patients receiving 30 μg qw showed disease progression (hazard ratio = 0.70; 95% CI, 0.39 to 1.25; p = 0.23).

Patients treated with IFNβ-1a 44 μg tiw had fewer CU, T1, and T2 active lesions per MRI scan compared with those treated with 30 μg qw at week 24 (nonparametric ANCOVA, p < 0.0001 for all activity measures; table 2). To examine the time to onset and persistence of effect, the monthly mean CU active lesion count was assessed. Figure 4 demonstrates a maximal treatment effect within 2 to 3 months of starting therapy, with a persistent reduction in active lesions favoring 44 μg tiw each month thereafter. The mean number of active scans per patient was also reduced for patients receiving 44 μg tiw compared with those receiving 30 μg qw. At baseline, 45% of patients in each group had no CU active lesions on MRI. After 24 weeks of therapy, however, 48% (157/325) of patients treated with 44 μg tiw had experienced no new MRI activity, compared with 33% (108/325) of the patients treated with 30 μg qw (p < 0.0001). Because Gd was not administered with the 48-week MRI scan, CU data are not available, and only the number of T2-active lesions could be calculated by comparing the baseline and week 48 scans. As shown in table 2, the differences between treatment groups were maintained in favor of the 44 μg tiw patients (p < 0.001 for all comparisons) for mean number of T2 lesions per patient per scan (36% relative reduction), proportion of T2-active scans (38% relative reduction), and proportion of patients with no T2 active lesions over 48 weeks (32% relative increase).

• Download figure
• Open in new tab
• Download powerpoint

Figure 4. Mean number of combined unique (CU) active lesions per patient per scan from baseline to week 24. Black bars = interferon (IFN) β-1a 44 μg three times weekly; white bars = IFN β-1a 30 μg once weekly.

Neutralizing antibodies.

Sera obtained at 48 weeks were tested for NAb. Considering any antibody value ≥20 neutralizing units (NU)/mL as positive, 84/335 (25%) of 44 μg tiw patient sera and 7/330 (2%) of 30 μg qw patient sera had neutralizing activity (p < 0.001). Of the 84 patients with NAb receiving 44 μg tiw, 49 (58%) were positive at week 24, whereas 1 of 7 patients (14%) who became NAb+ on 30 μg qw did so by week 24. The mean value of NU/mL at week 48 was 537 (median 174) for 44 μg tiw and 165 (median 34) for 30 μg qw. Despite the difference in antigenicity of the two treatment regimens, there was no apparent loss of clinical efficacy on relapse activity over the 48-week course of the trial. The proportion of patients remaining relapse free and the probability of first relapse were nearly identical for the patients receiving 44 μg tiw whether they were NAb+ or NAb− (figure 5). Relapse rates were also identical for NAb+ and NAb− patients in the 44 μg tiw group. The patients receiving 44 μg tiw who had NAb titers ≥20 NU/mL had fewer relapses throughout the study than those in the 30 μg qw group as a whole or the subgroup of 30 μg qw patients who remained NAb−.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5. Kaplan–Meier curves illustrating the cumulative probability of patients treated with interferon (IFN) β-1a 44 μg three times weekly (tiw) experiencing a relapse based on neutralizing antibody status at week 48. Solid line = IFNβ-1a 44 μg tiw NAb⁺; dotted line = IFNβ-1a 44 μg tiw NAb⁻.

The data on the effect of NAb on mean numbers of MRI lesions in the 44 μg tiw group showed that NAb− patients had the fewest T2-active lesions on the week 48 scan (0.6 lesions), whereas those who were NAb+ had more (1.6 lesions, p = 0.0004). Differences in lesion count between groups defined by week 48 NAb results were also seen at week 24, but were not different (0.8 for NAb− patients and 1.4 for NAb+ patients, p = 0.351) when only 58% of those who became positive at week 48 had developed NAb. No baseline differences were seen between groups. Median numbers of T2-active lesions were 0 for NAb subgroups at all time points.

Safety.

The mean time on study (46.9 weeks for the 44 μg tiw group, 47.1 weeks for the 30 μg qw group) and mean time on treatment (46.0 weeks for the 44 μg tiw group, 46.3 weeks for the 30 μg qw group) were similar for both treatment arms. The total amount of drug administered, however, was almost four times higher in the 44 μg tiw group (mean dose = 5,537 μg per patient) than in the 30 μg qw group (mean dose = 1,507 μg per patient).

Treatment was generally well tolerated in both groups (table 3), although some treatment-emergent adverse events were more common in the 44 μg tiw group than in the 30 μg qw group. The most common adverse events were injection-site disorders, which were reported in 83% (282/339) of 44 μg tiw–treated patients and 28% (93/337) of 30 μg qw–treated patients. Eighty-four percent of the high-dose events and 92% of the low-dose events were considered mild by physicians. Four patients on 44 μg tiw and one on 30 μg qw cited injection-site events as a contributing reason for stopping therapy. There were 160 occurrences of systemic influenzalike symptoms in 143 (42%) patients receiving 44 μg tiw and 188 similar events in 165 (49%) patients receiving 30 μg qw (p = 0.089). The incidence of subcomponents of the influenzalike syndrome (fever, myalgia, headache) was similar between groups, except for rigors, which were more common in 30 μg qw–treated patients (7%) than in 44 μg tiw–treated patients (3%, p = 0.051).

Asymptomatic laboratory abnormalities were noted frequently in both treatment arms. Liver function test results reported as adverse events were more common with 44 μg tiw regimen (60 patients or 18%) than with 30 μg qw regimen (32 patients or 9%, p = 0.002). These events were graded as mild (WHO grade 1, <2.5 times the upper limit of normal [ULN]) in 54% of 44 μg tiw–treated patients and 48% of 30 μg qw–treated patients. Severe elevations (WHO grade 3, >5 times but <20 times ULN for transaminases and >1.5 to 3 times ULN for bilirubin) were reported as adverse events in six patients in both groups. Two patients treated with 44 μg tiw and one patient treated with 30 μg qw stopped therapy because of raised transaminase levels. Elevated levels of transaminases decreased spontaneously in the majority of patients despite sustained dosing. However, approximately one-fifth of patients with grade 2 elevation progressed to grade 3 elevation if the drug was continued. Dose reductions or interruptions always led to improvement in enzyme levels.

White blood cell abnormalities were reported as adverse events in 38 (11%) patients treated with 44 μg tiw and in 16 (5%) patients treated with 30 μg qw (p = 0.002), with the events graded as mild in 67% of the patients receiving 44 μg tiw and in 94% of those receiving 30 μg qw. Lymphopenia was reported in 13 patients in the 44 μg tiw group and in 1 patient in the 30 μg qw group (p < 0.001). In 3 patients treated with 44 μg tiw and 1 patient treated with 30 μg qw, the lymphopenia was severe.

The number of patients with serious adverse events was 21 (6%) in the 44 μg tiw group and 18 (5%) in the 30 μg qw group. Nine of these events were considered by the investigators to be related to the study drug: one case each of lymphopenia, spontaneous abortion, depression, and suicidal ideation for the 44 μg tiw group; two cases of depression, one MS relapse, one episode of diplopia (not considered a relapse), and one of chest pain for the 30 μg qw group. Treatment discontinuation because of adverse events occurred in 16 (4.7%) of the 44 μg tiw patients and 14 (4.2%) of the 30 μg qw patients. The adverse events contributing to discontinuation (in some cases, more than one per patient) on IFNβ-1a 44 μg tiw were influenzalike syndrome (five patients), injection-site disorders (four patients), white blood cell abnormalities (two patients), liver enzyme abnormalities (three patients), depression (three patients), and unique miscellaneous events (eight patients). Events cited for discontinuation of 30 μg qw were influenzalike syndrome (six patients), depression (three patients), injection-site reaction (one patient), elevated transaminase level (one patient), and miscellaneous events.11 There was one death (solo pilot airplane crash) of a patient assigned to the 44 μg tiw group, but no other life-threatening serious adverse events in the study.