- Email: [email protected]
Lessons from randomised direct comparative trials
Journal of the Neurological Sciences 277, S1 (2009) S19 S24
Lessons from randomised direct comparative trials Anat Achirona, *, Sten Fredriksonb a b
Multiple Sclerosis Center, Sheba Medical Center, Tel-Hashomer, and Sackler School of Medicine, Tel-Aviv University, Israel Department of Clinical Neuroscience, Karolinska Institute Stockholm, Sweden
ARTICLE
INFO
Keywords: Multiple sclerosis Randomised clinical trials Glatiramer acetate Interferon-beta Evidence-based medicine
ABSTRACT For over a decade, four immunomodulatory therapies have been available for the treatment of relapsing remitting multiple sclerosis. However, few direct comparative data were available to facilitate the choice of treatment. This choice has been influenced by the perception that interferon-b preparations have greater efficacy than glatiramer acetate, due to apparently more rapid and robust reduction of gadolinium-enhancing lesions seen on magnetic resonance imaging in the pivotal trials of these agents. This situation has changed in the last year, with the outcomes of three randomised clinical trials comparing the efficacy and safety of glatiramer acetate with that of a high-dose interferon-b in relapsing remitting multiple sclerosis. These are the REGARD, BEYOND and BECOME trials. In the REGARD trial, 764 patients were randomised to treatment with either interferon-b 1a sc 44 mg or glatiramer acetate for 96 weeks; no significant difference in the time to first relapse was observed. The largest of the three comparative studies, the BEYOND trial, compared treatment with interferon-b 1b sc 500 mg, interferon-b 1b sc 250 mg or glatiramer acetate for two years in 2,244 patients. The hazard ratio for multiple relapses was close to unity for comparisons between all groups, indicating equivalent efficacy in all three treatment arms. Relapse rates (around 0.3 relapses/year) in all these studies were much lower than anticipated and lower than those reported a decade previously in the pivotal trials of b-interferons and glatiramer acetate. No unexpected safety issues were identified in any of these studies. The completion of these direct comparative studies has considerably enriched the clinical evidence database by contributing large numbers of patients. This provides an invaluable contribution for helping the physician make an informed choice about treatment. The results of the direct comparative studies provide evidence that glatiramer acetate and high-dose interferon-b preparations have comparable clinical efficacy. © 2009 Elsevier B.V. All rights reserved.
1. Introduction For nearly ten years, a choice of four different immunomodulatory treatments has been available to treat patients with relapsing remitting multiple sclerosis (RRMS). These are glatiramer acetate (GA) and three interferon-b preparations (interferon-b 1a im [Avonex® ], interferon-b 1b sc [Betaferon® ] and interferon-b 1a sc [Rebif® ]). Over this period, there has been much debate about the relative efficacy of these treatments. Attempts to compare outcome measures across different clinical trials are fraught with difficulties due, amongst others, to differences in the study populations, differences in disease activity prior to treatment, and the lack of a common definition of outcome variables. Nonetheless, the interferon-b preparations have benefited from a perception by certain * Corresponding author. E-mail address: [email protected] (A. Achiron). 0022-510X/ $
see front matter © 2009 Elsevier B.V. All rights reserved.
neurologists that they may present greater clinical efficacy than GA. This perception has arisen from their ability to decrease robustly the number of gadolinium-enhancing lesions on magnetic resonance imaging (MRI) and from the fact that collectively there have been more pivotal trials with these agents that have included a greater number of patients overall [1]. Until 2007, only two clinical studies had directly compared the efficacy and tolerability of different immunomodulatory treatments. These were the EVIDENCE study [2], performed in the USA, which compared interferon-b 1a sc 44 mg to interferon-b 1a im, and the INCOMIN study [3], performed in Italy, comparing interferon-b 1b sc to interferon-b 1a im. Both these studies demonstrated that high-dose subcutaneous interferon-b was more effective than intramuscular interferon-b. Recently, three additional studies compared the efficacy of glatiramer acetate (GA) with that of high-dose subcutaneous
S20
A. Achiron, S. Fredrikson / Journal of the Neurological Sciences 277 (2009) S19 S24
interferon-b. These were the REGARD trial [4], which compared GA to interferon-b 1a sc 44 mg (Rebif® ) and the BECOME [5, 6] and BEYOND [7] trials, which both compared GA to interferon-b 1b sc (Betaferon® ). The BECOME trial was a small, single-centre, investigator-driven study using MRI outcome measures, whereas the BEYOND and REGARD trials were large, multicentre, industry-sponsored studies using clinical outcome measures. 2. The BECOME study The BECOME study was an investigator-initiated study comparing high-dose interferon-b 1b sc 250 mg every other day to daily GA sc 20 mg in 75 patients with RRMS [5,6]. The hypothesis evaluated in the BECOME study was that high-dose interferon-b 1b would inhibit the development of new brain lesions more strongly than GA treatment. To this end, the study used an ‘optimised’ MRI protocol aimed at improving detection of lesions within the brain, involving high-field MRI (3 T) and triple-dose gadolinium enhancement. The primary outcome measure was the mean number of combined active lesions per scan per patient. Combined active lesions were defined as either gadolinium-enhancing T1 lesions or new non-enhancing T2/FLAIR lesions. Given the relatively large number of lesions anticipated, it was considered that the inclusion of 36 patients per group would be sufficient to demonstrate a significant difference in this outcome measure over a treatment period of one year. In fact, 36 patients received interferon-b and 39 GA. The final analysis of the data demonstrated that the mean number of combined active lesions declined over one year in both treatment groups, with no significant inter-group difference. Similarly, no differences in any of the secondary MRI outcome measures were observed between the two treatment groups [5]. Although insufficiently powered to detect differences in clinical outcome, annualised relapse rates and disability progression were similar in the interferon-b and GA treatment groups [6]. No information on tolerability was reported. Consequently, it was concluded that the assumed superiority of interferon-b over GA for reducing the incidence of active inflammation in RRMS may have been overestimated [5]. 3. The REGARD study 3.1. Objectives and design The REGARD study was a randomised, prospective, raterblinded, multicentre, comparative study initiated and implemented by Ares Serono (now Merck Serono), manufacturers of interferon-b 1a sc (Rebif® ) [4]. Within one month of a screening visit at which a complete neurological assessment was performed, patients fulfilling the inclusion criteria were randomised to treatment with either interferon-b 1a sc 44 mg three times weekly or daily GA sc 20 mg for 96 weeks. Patients were evaluated with a comprehensive neurological examination at weeks 24, 48, 72 and 96. A subset of 460 patients (60.2%) also underwent MRI at these time points, and all patients provided a serum sample for measurement of neutralising antibodies. The inclusion criteria specified that patients should fulfil McDonald criteria [8] for relapsing remitting multiple sclerosis, should have experienced at least one relapse in the year prior to the study, and have neurological disability as evaluated with the Expanded Disability Status Score (EDSS [9]) of 5.5. The study was designed to have 80% power to detect
Table 1 Comparison of relapse outcomes in the pivotal randomised clinical trials of interferon-b 1a sc 44 mg and glatiramer acetate sc 20 mg
Annualised relapse rate Placebo Active treatment Difference Time to first relapse Placebo Active treatment Difference
Interferon-b 1a 44 mg PRISMS trial [10]
Glatiramer acetate Johnson et al., 1995 [11]
1.28 0.87 33%
0.84 0.59 30%
135 days 288 days +113%
198 days 287 days +45%
a 30% difference in time to first relapse over the study period, which would necessitate the inclusion of 736 patients of whom 460 would be expected to experience a relapse. Accordingly, 764 RRMS patients were included. However, disease activity in these patients was relatively low as, after a mean duration of six years since diagnosis, the mean EDSS score was only 2.3 and the mean number of gadoliniumenhancing lesions at inclusion was 1.56. Thirty-five percent of patients had experienced only a single relapse in the two years preceding inclusion. No differences were observed between the two treatment groups for most baseline variables, with the exception of T1 hypointense lesion volume, the proportion of patients experiencing multiple relapses in the previous two years, and the proportion of patients having received steroids in the previous six months, which were all significantly higher in patients assigned to the GA treatment group. The primary outcome measure in the REGARD study was time to first relapse, determined using a Cox proportional hazard model. In the pivotal trials of these two drugs [10,11], which both used annualised relapse rate as their primary outcome measure, the reduction in relapse rate compared to placebo was very similar for the two treatments (Table 1). In contrast, the time to first relapse was prolonged compared to placebo more in the interferon-b 1a trial (+113% for the 44 mg dose) than in the GA trial (+45%). Nonetheless, the absolute median time to first relapse was very similar in the active treatment groups in both trials (288 days and 287 days respectively; Table 1). The secondary endpoints of the REGARD study were related to MRI measures and were restricted to a sub-population of patients who underwent serial MRI scans. These were the number of new or enlarged T2 lesions accumulating over the course of the study, the mean number of gadolinium-enhancing lesions per patient and per scan, and the change in volume for both these lesion types. The annualised relapse rate was considered a tertiary endpoint, together with appearance of T1 hypointense lesions, change in brain volume and accumulation of disability. 3.2. Clinical outcomes With regard to relapses, the number of patients experiencing at least one relapse was rather low, namely 126 patients in the interferon-b group and 132 in the GA group, corresponding to 34% of the included patients. This was considerably lower than the 63% anticipated in the sample size estimation. No difference in time to first relapse was demonstrated between the two treatment groups; the Kaplan Meier survival curves for this variable were practically superimposable and crossed at a number of points (Fig. 1). The hazard ratio, determined by the Cox proportional hazards method, was 0.94 [95% confidence
A. Achiron, S. Fredrikson / Journal of the Neurological Sciences 277 (2009) S19 S24
672 days (96 weeks)
Survival distribution function
1.00
0.75
42% in the number of gadolinium-enhancing lesions/patient/ scan in favour of the interferon-b treatment group was noted. The reduction in brain volume was significantly lower in the GA treatment group (Table 2), suggesting that GA is more effective in preventing brain atrophy, as has been proposed previously [12]. 3.4. Tolerability
0.50
0.25
0.00 0
100
200
300
400
500
600
700
Time to first relapse (days) Fig. 1 Time to first relapse in the REGARD study. Kaplan Meier survival curves for first relapse in 386 patients treated with interferon-b 1a (upper curve) and in 378 patients treated with glatiramer acetate (lower curve). Reproduced from reference [4] with permission.
interval: 0.74, 1.21] ( p = 0.64). It was not possible to determine the median time to first relapse, since less than 50% of subjects experienced a relapse during the course of the study. The study investigators reported that the time for 30% of the patients to experience a first relapse was 495 days in the interferon-b 1a group and 432 days in the GA group. However, it should be noted that the choice of the 30% threshold is entirely arbitrary and, since the survival curves for the two groups were not parallel but intertwined, the use of a different threshold would have yielded a different relative time-to-event interval for the two groups. The mean annualised relapse rate over the 96-week study duration was 0.30 relapses/year in the interferon-b group and 0.29 relapses/year in the GA group. This represented a decrease of around 70% compared to pre-treatment relapse rates in both treatment groups. The proportion of patients who remained relapse-free throughout the study was 62% in both treatment groups. Similarly, the proportion of patients with confirmed disability progression was comparable between the two groups, 11.7 in the interferon-b arm and 8.7% in the GA arm. 3.3. MRI outcomes With respect to the subpopulation of patients with serial MRI evaluations, no difference was observed in the number of new or enlarging T2 lesions per patient per scan or in the number of new T1 hypointense lesions between the two groups (Table 2). A statistically significant difference ( p < 0.001) of Table 2 Magnetic resonance imaging outcomes in the REGARD study a Interferon-b 1a
Glatiramer acetate
p
T2 active lesions
0.67
0.82
0.18
Gd-enhancing lesions
0.24
0.41
<0.001
New T1 hypointense lesions
0.23
0.24
0.15
Change in brain volume a
S21
1.24%
1.07%
0.018
Lesion data correspond to mean numbers of lesions per patient per scan. Data are presented for the subset of 460 patients (230 per group) who underwent serial MRI evaluations. Data are taken from reference [4].
The overall proportion of patients reporting adverse events was similar in the two treatment groups (1,880 events reported in 90.8% of patients receiving interferon-b and 1,917 events reported in 85.3% of patients receiving GA). However, the nature of the adverse events differed between the two groups. Influenza-like illness, headache, myalgia and alanine amino transferase elevation were reported more frequently in the interferon-b group, while injection-site reactions, dyspnoea and immediate systemic post-injection reactions occurred more frequently in the GA group (Table 3). Overall, 17% of patients discontinued treatment during the course of the study, with a higher proportion of patients discontinuing in the interferon-b group (20.9%) compared to the GA group (13.6%). Table 3 Adverse events reported in the REGARD study a Adverse event
Interferon-b N = 386
Influenza-like illness
Glatiramer acetate N = 378
119 (31.2%)
5 (1.3%)
Injection site pain
47 (12.2%)
53 (14.0%)
Headache
74 (19.4%)
35 (9.3%)
Injection-site pruritus
8 (2.1%)
78 (20.0%)
Injection site bruising
33 (8.5%)
38 (10.1%)
Nasopharyngitis
34 (8.8%)
44 (11.6%)
4 (1.0%)
42 (11.2%)
Injection-site swelling Myalgia
22 (5.8%)
8 (2.1%)
Alanine amino transferase elevation
21 (5.5%)
19 (1.3%)
Injection-site induration
9 (2.4%)
25 (6.7%)
Dyspnoea
4 (1.0%)
22 (5.9%)
Immediate post-injection reaction a
None
19 (5.1%)
Only events reported in more than 10% of patients in either group or whose incidence differed significantly between the two treatment groups are listed. Adverse events in bold were significantly more frequent in the interferon-b group and those in italic significantly more frequent in the glatiramer acetate (GA) group. Data are taken from reference [4].
4. The BEYOND study 4.1. Objectives and design The BEYOND study was a much larger study initiated and implemented by Bayer-Schering, the manufacturers of interferon-b 1b sc 250 mg (Betaseron® ). Since the principal goal of the study was to use the data, in case of success, to secure registration of the higher 500 mg dose of interferon-b 1b, it was conducted according to current regulatory standards, thus ensuring a high quality of design and data monitoring and sufficient statistical power to test the study hypotheses. The BEYOND study attempted to test two hypotheses, firstly whether the higher dose of 500 mg of interferon-b 1b would have superior efficacy to the currently recommended dose of 250 mg, and secondly whether either dose would have
S22
A. Achiron, S. Fredrikson / Journal of the Neurological Sciences 277 (2009) S19 S24
superior efficacy to GA. The study included 2,244 treatmentnaive patients with RRMS fulfilling the McDonald criteria [8] for clinically definite disease from 198 study centres worldwide. These were randomised in a ratio of 2:2:1 to one of three treatment arms, namely interferon-b 500 mg (n = 899), interferon-b 250 mg (n = 897) or GA (n = 448). This is thus a very large study, in fact the largest randomised study in RRMS reported to date. As expected, the mean baseline characteristics of the patients included were comparable between all three treatment groups. The treatment period was twenty-four months, followed by an eighteen-month extension during which the doubleblind was maintained and systematic follow-up continued for patients completing the 24-month core study until all patients had completed the latter. Patients returned quarterly to the study centre for a complete neurological evaluation and were also contacted between these study visits by telephone. Yearly MRI scans were performed. Particular attention was paid to concealment of the randomisation. The two interferon-b groups were compared under double-blind conditions. Due to the difference in administration frequency, it was not possible to compare the interferon-b groups with the GA group under double-blind conditions. However, the rater who performed the quarterly neurological assessment was blinded as to treatment and the MRI scans were evaluated centrally and anonymously. The primary outcome measure was the hazard ratio for multiple relapses determined with the Anderson Gill extension of the Cox proportional hazards model. This rigorous statistical method allows differences in relapse frequency between individual patients to be taken into account in the analysis, and represents another strength of the study. The statistical testing followed a hierarchical null-hypothesis approach, in which one-to-one group comparisons were tested in the following order: interferon-b 500 mg with interferon-b 250 mg, then interferon-b 500 mg with GA and finally interferon-b 250 mg with GA. Secondary outcomes included time to confirmed EDSS progression and the development of T1 hypointense lesions (‘black holes’). Supportive outcome variables measured included the proportion of patients free from disease activity, a composite measure defined as the proportion of patients who experienced no relapses and who developed no new T2 lesions or Gd-enhancing lesions following initiation of treatment, the time to first relapse, the number and volume of T2 lesions, and changes in total brain volume. 4.2. Clinical outcomes Clinical outcomes from the BEYOND study have been reported in abstract form at two recent neurology congresses [7,13]. In the primary outcome evaluation, the hazard ratio for multiple relapses in all three pair-wise comparisons was not significantly different from unity, indicating no difference in relapse risk between any of the three treatments [7]. During the two-year treatment period, the mean annualised relapse rate decreased by approximately 80% in all three treatment groups. No significant inter-group difference was found in the proportion of relapse-free patients, in time to first relapse or in accumulation of disability. With respect to the composite endpoint of the proportion of patients free from all disease activity, this represented 31.5% in the interferon-b 500 mg group, 25.6% in the
interferon-b 250 mg group and 25.7% in the GA group. A significant difference in favour of interferon-b 500 mg with respect to both interferon-b 250 mg ( p = 0.007) and GA ( p = 0.031) was reported [13]. This difference was attributed to a more pronounced effect of the higher interferon-b dose on gadolinium-enhancing lesions and on new or enlarging T2 lesions and was not translated into concrete clinical differences during the observation period of this study. 4.3. MRI outcomes With respect to MRI outcomes [14], no significant differences were observed for most measures. In particular, no significant differences between the groups in the number or volume of T1 Gd-enhancing lesions at study end, in the accumulation of T1 hypointense lesions (‘black holes’) or in changes in total brain volume between the three treatment groups. Exceptions were the cumulative number of T2 lesions up to the last scan and the relative increase in T2 lesion volume, which were both significantly higher in the GA group than in either of the interferon-b groups. However, the differences observed were relatively small. 4.4. Tolerability Tolerability was assessed systematically in the BEYOND study by monitoring the incidence of adverse events and by regular blood tests. Adherence to treatment was also reported. All three treatments were relatively well tolerated [15]. The adverse events reported were consistent with the known safety profiles of interferon-b and GA. During the first year of treatment, influenza-like symptoms were reported significantly more frequently in the interferon-b 250 mg group compared to the GA group. Although initially high, the frequency of influenza-like symptoms in interferon-b-treated patients decreased rapidly over time. Injection-site reactions, notably pain and pruritus, were significantly more frequent in the GA treatment group than in the interferon-b 250 mg group but again, the incidence of these reactions diminished over the course of the study. The incidence of other adverse events such as fatigue, depression, arthralgia and paraesthesia was comparable between treatment groups. Concerning laboratory safety evaluation, elevations in liver enzymes, anomalies in blood lipids and markers of thyroid function, as well as leukocytopenia, were more frequently detected in patients treated with interferon-b 1b than in those treated with GA. Adherence to treatment was high in all three treatment groups. The proportion of patients completing the anticipated treatment period ranged from 73% in the interferon-b 500 mg group through 78% in the GA group to 82% in the interferon-b 250 mg group. No significant inter-group differences in discontinuation rates were observed. 5. Immunogenicity Data on neutralising antibodies (NAbs) are only available for the REGARD study. In this study, neutralising antibodies to interferon-b 1a were detected in 34% of patients, and these persisted until the end of the study in 27% of patients. Antibody titres were rather high in these patients, being over 1000 NU/ml in around one-third of the patients with persistent antibodies. Seroprevalence rates for NAbs in the REGARD study were unexpectedly high, being, for example,
A. Achiron, S. Fredrikson / Journal of the Neurological Sciences 277 (2009) S19 S24
twice as high as those reported in the PRISMS study (12.5% of patients) [10]. Time to first relapse was not significantly different in interferon-b-treated patients who developed NAbs compared to those who did not, suggesting that there was no correlation between the emergence of NAbs and clinical efficacy. However, it should be noted that the study duration (96 weeks) was short compared to the time necessary to observe a loss of clinical efficacy in patients developing NAbs in response to interferon-b [16]. Binding antibodies to GA were identified in 98% of patients at the first assessment at 24 weeks of treatment, as expected. GA generates an immune response resulting in the production of binding antibodies that do not compromise therapeutic efficacy and may contribute to efficacy [17]. A very surprising observation was that binding antibodies to GA were detected in around one quarter of patients before treatment was initiated. In principle, anti-GA antibodies are only detected in patients who have been exposed to GA. This finding suggests that either the patients were not treatment-naive contrary to what was specified in the inclusion criteria or that there was a problem with the assay used to detect the presence of antibodies. 6. Strengths and weaknesses of the directive comparative trials A major strength of the REGARD and BEYOND studies was the large number of patients included; these direct head-to-head comparisons evaluated over 2.5 times as many patients as the pivotal trials (Table 4). The BEYOND trial alone included twice as many patients as all pivotal trials of interferon-b combined. Further strengths include the randomised design, rater blinding, the relatively low drop-out rates, and the use of rigorously defined primary outcome measures. Table 4 Patient numbers in pivotal trials and direct comparative studies of immunomodulatory treatments in multiple sclerosis Trial(s)
No. of patients
Pivotal trials Interferon-b Interferon-b 1a im
301
Interferon-b 1b sc
372
Interferon-b 1a sc All Interferon-b Glatiramer acetate All pivotal trials
560 1233 251
7. Conclusions
Interferon 677
INCOMIN
188
All Interferon interferon comparisons
865
Interferon and Glatiramer acetate BECOME
75
REGARD
764
BEYOND
2244
All interferon glatiramer acetate comparisons
3083
All direct comparisons
Among the weaknesses, we may cite the open-label design of the studies. Performing an accurately blinded study with injectable treatments that differ in frequency of administration, form of application device and side-effect profile is obviously difficult. To some extent these problems could have been avoided by the use of a double-dummy design, although that would have exposed patients to additional injections, possibly compromising recruitment. Measurement biases should have been avoided by the requirement for the raters to be blinded to treatment assignment, and expectation biases, if they existed, cannot have played a major role since the study assumption that high-dose b-interferons would be superior to GA was actually not confirmed by the data. A further limitation relates to the low number of relapses observed in the REGARD and BEYOND trials, which may have compromised the statistical power of the studies to demonstrate between-group differences. In fact, the number of patients experiencing an on-treatment relapse was nearly half that anticipated in the power calculations of the REGARD study. However, there was no hint from the relapse data that any qualitative or quantitative difference between treatment groups existed. Disease activity at the time of inclusion was much lower in the study sample than in previous trials of these agents, which may explain the unexpectedly low number of relapses reported during the study. This is indicative of a change in the population of patients recruited into current clinical trials in multiple sclerosis compared with those performed over a decade ago. Indeed, the on-treatment relapse rate observed in the REGARD and BEYOND trials was considerably lower than those observed in the pivotal trials of interferon-b 1b, interferon-b 1a sc 44 mg and GA, which were 0.84, 0.87 and 0.59, respectively [10,11,18]. However, a relapse rate of 0.30 is very similar to that observed for interferon-b 1a sc 44 mg in another recent trial comparing this agent to alemtuzumab [19], as well as in recent trials of other active treatments such as natalizumab [20] and fingolimod [21]. This trend of a declining on-treatment relapse rate has several important consequences for the design of future trials. Firstly, future trials will require inclusion of a larger number of patients than was the case in the 1990s in order to generate sufficient power to detect a treatment effect. It also makes comparison between outcome in different clinical trials conducted in different populations at different times extremely difficult. However, on a positive note, it suggests that immunomodulatory treatment of patients with less active disease, who have been diagnosed earlier through the use of MRI, may provide better relapse control than had been anticipated from the pivotal trials.
1484
Direct comparative studies EVIDENCE
S23
3948
The conclusion from three well-conducted direct comparative studies is that the efficacy of GA is comparable to that of interferon-b 1b on a wide range of clinical and MRI outcomes [22,23]. Both interferon-b and GA were well tolerated, and adverse events reported were consistent with their known safety profiles. Conflict of interest statement S. Fredrikson declares having received speaker fees from Teva, Sanofi-Aventis, Bayer Schering, Biogen Idec, and Serono. No funding applicable.
S24
A. Achiron, S. Fredrikson / Journal of the Neurological Sciences 277 (2009) S19 S24
A. Achiron declares having received speaker fees from Teva, and research grants from Teva, Serono, and Bayer-Schering. References 1. Galetta SL, Markowitz C, Lee AG. Immunomodulatory agents for the treatment of relapsing multiple sclerosis: a systematic review. Arch Intern Med 2002;162:2161 9. 2. Panitch H, Goodin DS, Francis G, Chang P, Coyle PK, O’Connor P, et al. Randomized, comparative study of interferon beta-1a treatment regimens in MS: The EVIDENCE Trial. Neurology 2002;59:1496 506. 3. Durelli L, Verdun E, Barbero P, Bergui M, Versino E, Ghezzi A, et al. Everyother-day interferon beta-1b versus once-weekly interferon beta-1a for multiple sclerosis: results of a 2-year prospective randomised multicentre study (INCOMIN). Lancet 2002;359:1453 60. 4. Mikol DD, Barkhof F, Chang P, Coyle PK, Jeffery DR, Schwid SR, et al. Comparison of subcutaneous interferon beta-1a with glatiramer acetate in patients with relapsing multiple sclerosis (the REbif vs Glatiramer Acetate in Relapsing MS Disease [REGARD] study): a multicentre, randomised, parallel, open-label trial. Lancet Neurol 2008;7:903 14. 5. Wolansky L, Cook S, Skurnick J, Lincoln J, Tulloch K, Franco P, et al. Betaseron vs. Copaxone in MS with triple-dose gadolinium and 3-T MRI Endpoints (BECOME): announcement of final primary study outcome. Mult Scler 2007;13:P207. 6. Cadavid D, Wolansky L, Cook S, Halper J, Hill J, Younes M, et al. Betaseron versus copaxone in multiple sclerosis with triple-dose gadolinium and 3 Tesla MRI Endpoints (BECOME): announcement of secondary clinical outcomes. Mult Scler 2007;13:P207. 7. O’Connor P, Arnason B, Comi G, Filippi M, Cook S, Goodin D, et al. Interferon beta-1b 500 mcg, interferon beta-1b 250 mcg and glatiramer acetate: primary outcomes of the BEYOND (Betaferon® /Betaseron® Efficacy Yielding Outcomes of a New Dose) study. Neurology 2008;70(11; Late Breaking Science Supplement):S004. 8. McDonald WI, Compston A, Edan G, Goodkin D, Hartung HP, Lublin FD, et al. Recommended diagnostic criteria for multiple sclerosis: guidelines from the International Panel on the diagnosis of multiple sclerosis. Ann Neurol 2001; 50:121 7. 9. Kurtzke JF. Rating neurologic impairment in multiple sclerosis: an expanded disability status scale (EDSS). Neurology 1983;33:1444 52. 10. PRISMS (Prevention of Relapses and Disability by Interferon beta-1a Subcutaneously in Multiple Sclerosis) Study Group. Randomised doubleblind placebo-controlled study of interferon beta-1a in relapsing/remitting multiple sclerosis. Lancet 1998;352:1498 504.
11. Johnson KP, Brooks BR, Cohen JA, Ford CC, Goldstein J, Lisak RP, et al. The Copolymer 1 Multiple Sclerosis Study Group. Copolymer 1 reduces relapse rate and improves disability in relapsing remitting multiple sclerosis: results of a phase III multicenter, double-blind placebo-controlled trial. Neurology 1995;45:1268 76. 12. Zivadinov R, Reder AT, Filippi M, Minagar A, Stuve O, Lassmann H, et al. Mechanisms of action of disease-modifying agents and brain volume changes in multiple sclerosis. Neurology 2008;71:136 44. 13. Kappos L, Arnason B, Comi G, Cook S, Filippi M, Goodin D, et al. High proportion of patients free from disease activity in all 3 arms of the highdose Betaferon® trial. J Neurol 2008;255(Suppl 2):36. 14. Filippi M, Arnason B, Comi G, Cook S, Goodin D, Hartung H, et al. Magnetic resonance imaging findings of a phase III trial comparing Betaferon® with Copaxone® treatments in relapsing remitting multiple sclerosis. J Neurol 2008;255(Suppl 2):9. 15. Comi G, Arnason B, Cook S, Filippi M, Goodin D, Hartung H, et al. Tolerability and adherence to Betaferon® and Copaxone® in a phase III clinical trial. J Neurol 2008;255(Suppl 2):111. 16. Giovannoni G, Goodman A. Neutralizing anti-IFN-beta antibodies: how much more evidence do we need to use them in practice? Neurology 2005;65:6 8. 17. Brenner T, Arnon R, Sela M, Abramsky O, Meiner Z, Riven-Kreitman R, et al. Humoral and cellular immune responses to Copolymer 1 in multiple sclerosis patients treated with Copaxone. J Neuroimmunol 2001;115:152 60. 18. The IFNB Multiple Sclerosis Study Group and The University of British Columbia MS/MRI Analysis Group. Interferon beta-1b in the treatment of multiple sclerosis: final outcome of the randomized controlled trial. Neurology 1995;45:1277 85. 19. Coles A, CAMMS 223 Study Group. Alemtuzumab compared with subcutaneous high-dose IFNB-1a in treatment-naive relapsing remitting multiple sclerosis: primary efficacy outcomes of CAMMS223 at 3 years. Neurology 2008:70(Suppl 1): Abstract S22.006. 20. Polman CH, O’Connor PW, Havrdova E, Hutchinson M, Kappos L, Miller DH, et al. A randomized, placebo-controlled trial of natalizumab for relapsing multiple sclerosis. N Engl J Med 2006;354:899 910. 21. Kappos L, Antel J, Comi G, Montalban X, O’Connor P, Polman CH, et al. Oral fingolimod (FTY720) for relapsing multiple sclerosis. N Engl J Med 2006;355: 1124 40. 22. Sorensen PS. REGARD: what can we learn from randomised, open-label, head-to-head studies? Lancet Neurol 2008;7:864 6. 23. Goodin D. Comparative studies of glatiramer acetate and interferon beta. Int MS J 2008;15:39 41.
Lessons from randomised direct comparative trials Anat Achirona, *, Sten Fredriksonb a b
Multiple Sclerosis Center, Sheba Medical Center, Tel-Hashomer, and Sackler School of Medicine, Tel-Aviv University, Israel Department of Clinical Neuroscience, Karolinska Institute Stockholm, Sweden
ARTICLE
INFO
Keywords: Multiple sclerosis Randomised clinical trials Glatiramer acetate Interferon-beta Evidence-based medicine
ABSTRACT For over a decade, four immunomodulatory therapies have been available for the treatment of relapsing remitting multiple sclerosis. However, few direct comparative data were available to facilitate the choice of treatment. This choice has been influenced by the perception that interferon-b preparations have greater efficacy than glatiramer acetate, due to apparently more rapid and robust reduction of gadolinium-enhancing lesions seen on magnetic resonance imaging in the pivotal trials of these agents. This situation has changed in the last year, with the outcomes of three randomised clinical trials comparing the efficacy and safety of glatiramer acetate with that of a high-dose interferon-b in relapsing remitting multiple sclerosis. These are the REGARD, BEYOND and BECOME trials. In the REGARD trial, 764 patients were randomised to treatment with either interferon-b 1a sc 44 mg or glatiramer acetate for 96 weeks; no significant difference in the time to first relapse was observed. The largest of the three comparative studies, the BEYOND trial, compared treatment with interferon-b 1b sc 500 mg, interferon-b 1b sc 250 mg or glatiramer acetate for two years in 2,244 patients. The hazard ratio for multiple relapses was close to unity for comparisons between all groups, indicating equivalent efficacy in all three treatment arms. Relapse rates (around 0.3 relapses/year) in all these studies were much lower than anticipated and lower than those reported a decade previously in the pivotal trials of b-interferons and glatiramer acetate. No unexpected safety issues were identified in any of these studies. The completion of these direct comparative studies has considerably enriched the clinical evidence database by contributing large numbers of patients. This provides an invaluable contribution for helping the physician make an informed choice about treatment. The results of the direct comparative studies provide evidence that glatiramer acetate and high-dose interferon-b preparations have comparable clinical efficacy. © 2009 Elsevier B.V. All rights reserved.
1. Introduction For nearly ten years, a choice of four different immunomodulatory treatments has been available to treat patients with relapsing remitting multiple sclerosis (RRMS). These are glatiramer acetate (GA) and three interferon-b preparations (interferon-b 1a im [Avonex® ], interferon-b 1b sc [Betaferon® ] and interferon-b 1a sc [Rebif® ]). Over this period, there has been much debate about the relative efficacy of these treatments. Attempts to compare outcome measures across different clinical trials are fraught with difficulties due, amongst others, to differences in the study populations, differences in disease activity prior to treatment, and the lack of a common definition of outcome variables. Nonetheless, the interferon-b preparations have benefited from a perception by certain * Corresponding author. E-mail address: [email protected] (A. Achiron). 0022-510X/ $
see front matter © 2009 Elsevier B.V. All rights reserved.
neurologists that they may present greater clinical efficacy than GA. This perception has arisen from their ability to decrease robustly the number of gadolinium-enhancing lesions on magnetic resonance imaging (MRI) and from the fact that collectively there have been more pivotal trials with these agents that have included a greater number of patients overall [1]. Until 2007, only two clinical studies had directly compared the efficacy and tolerability of different immunomodulatory treatments. These were the EVIDENCE study [2], performed in the USA, which compared interferon-b 1a sc 44 mg to interferon-b 1a im, and the INCOMIN study [3], performed in Italy, comparing interferon-b 1b sc to interferon-b 1a im. Both these studies demonstrated that high-dose subcutaneous interferon-b was more effective than intramuscular interferon-b. Recently, three additional studies compared the efficacy of glatiramer acetate (GA) with that of high-dose subcutaneous
S20
A. Achiron, S. Fredrikson / Journal of the Neurological Sciences 277 (2009) S19 S24
interferon-b. These were the REGARD trial [4], which compared GA to interferon-b 1a sc 44 mg (Rebif® ) and the BECOME [5, 6] and BEYOND [7] trials, which both compared GA to interferon-b 1b sc (Betaferon® ). The BECOME trial was a small, single-centre, investigator-driven study using MRI outcome measures, whereas the BEYOND and REGARD trials were large, multicentre, industry-sponsored studies using clinical outcome measures. 2. The BECOME study The BECOME study was an investigator-initiated study comparing high-dose interferon-b 1b sc 250 mg every other day to daily GA sc 20 mg in 75 patients with RRMS [5,6]. The hypothesis evaluated in the BECOME study was that high-dose interferon-b 1b would inhibit the development of new brain lesions more strongly than GA treatment. To this end, the study used an ‘optimised’ MRI protocol aimed at improving detection of lesions within the brain, involving high-field MRI (3 T) and triple-dose gadolinium enhancement. The primary outcome measure was the mean number of combined active lesions per scan per patient. Combined active lesions were defined as either gadolinium-enhancing T1 lesions or new non-enhancing T2/FLAIR lesions. Given the relatively large number of lesions anticipated, it was considered that the inclusion of 36 patients per group would be sufficient to demonstrate a significant difference in this outcome measure over a treatment period of one year. In fact, 36 patients received interferon-b and 39 GA. The final analysis of the data demonstrated that the mean number of combined active lesions declined over one year in both treatment groups, with no significant inter-group difference. Similarly, no differences in any of the secondary MRI outcome measures were observed between the two treatment groups [5]. Although insufficiently powered to detect differences in clinical outcome, annualised relapse rates and disability progression were similar in the interferon-b and GA treatment groups [6]. No information on tolerability was reported. Consequently, it was concluded that the assumed superiority of interferon-b over GA for reducing the incidence of active inflammation in RRMS may have been overestimated [5]. 3. The REGARD study 3.1. Objectives and design The REGARD study was a randomised, prospective, raterblinded, multicentre, comparative study initiated and implemented by Ares Serono (now Merck Serono), manufacturers of interferon-b 1a sc (Rebif® ) [4]. Within one month of a screening visit at which a complete neurological assessment was performed, patients fulfilling the inclusion criteria were randomised to treatment with either interferon-b 1a sc 44 mg three times weekly or daily GA sc 20 mg for 96 weeks. Patients were evaluated with a comprehensive neurological examination at weeks 24, 48, 72 and 96. A subset of 460 patients (60.2%) also underwent MRI at these time points, and all patients provided a serum sample for measurement of neutralising antibodies. The inclusion criteria specified that patients should fulfil McDonald criteria [8] for relapsing remitting multiple sclerosis, should have experienced at least one relapse in the year prior to the study, and have neurological disability as evaluated with the Expanded Disability Status Score (EDSS [9]) of 5.5. The study was designed to have 80% power to detect
Table 1 Comparison of relapse outcomes in the pivotal randomised clinical trials of interferon-b 1a sc 44 mg and glatiramer acetate sc 20 mg
Annualised relapse rate Placebo Active treatment Difference Time to first relapse Placebo Active treatment Difference
Interferon-b 1a 44 mg PRISMS trial [10]
Glatiramer acetate Johnson et al., 1995 [11]
1.28 0.87 33%
0.84 0.59 30%
135 days 288 days +113%
198 days 287 days +45%
a 30% difference in time to first relapse over the study period, which would necessitate the inclusion of 736 patients of whom 460 would be expected to experience a relapse. Accordingly, 764 RRMS patients were included. However, disease activity in these patients was relatively low as, after a mean duration of six years since diagnosis, the mean EDSS score was only 2.3 and the mean number of gadoliniumenhancing lesions at inclusion was 1.56. Thirty-five percent of patients had experienced only a single relapse in the two years preceding inclusion. No differences were observed between the two treatment groups for most baseline variables, with the exception of T1 hypointense lesion volume, the proportion of patients experiencing multiple relapses in the previous two years, and the proportion of patients having received steroids in the previous six months, which were all significantly higher in patients assigned to the GA treatment group. The primary outcome measure in the REGARD study was time to first relapse, determined using a Cox proportional hazard model. In the pivotal trials of these two drugs [10,11], which both used annualised relapse rate as their primary outcome measure, the reduction in relapse rate compared to placebo was very similar for the two treatments (Table 1). In contrast, the time to first relapse was prolonged compared to placebo more in the interferon-b 1a trial (+113% for the 44 mg dose) than in the GA trial (+45%). Nonetheless, the absolute median time to first relapse was very similar in the active treatment groups in both trials (288 days and 287 days respectively; Table 1). The secondary endpoints of the REGARD study were related to MRI measures and were restricted to a sub-population of patients who underwent serial MRI scans. These were the number of new or enlarged T2 lesions accumulating over the course of the study, the mean number of gadolinium-enhancing lesions per patient and per scan, and the change in volume for both these lesion types. The annualised relapse rate was considered a tertiary endpoint, together with appearance of T1 hypointense lesions, change in brain volume and accumulation of disability. 3.2. Clinical outcomes With regard to relapses, the number of patients experiencing at least one relapse was rather low, namely 126 patients in the interferon-b group and 132 in the GA group, corresponding to 34% of the included patients. This was considerably lower than the 63% anticipated in the sample size estimation. No difference in time to first relapse was demonstrated between the two treatment groups; the Kaplan Meier survival curves for this variable were practically superimposable and crossed at a number of points (Fig. 1). The hazard ratio, determined by the Cox proportional hazards method, was 0.94 [95% confidence
A. Achiron, S. Fredrikson / Journal of the Neurological Sciences 277 (2009) S19 S24
672 days (96 weeks)
Survival distribution function
1.00
0.75
42% in the number of gadolinium-enhancing lesions/patient/ scan in favour of the interferon-b treatment group was noted. The reduction in brain volume was significantly lower in the GA treatment group (Table 2), suggesting that GA is more effective in preventing brain atrophy, as has been proposed previously [12]. 3.4. Tolerability
0.50
0.25
0.00 0
100
200
300
400
500
600
700
Time to first relapse (days) Fig. 1 Time to first relapse in the REGARD study. Kaplan Meier survival curves for first relapse in 386 patients treated with interferon-b 1a (upper curve) and in 378 patients treated with glatiramer acetate (lower curve). Reproduced from reference [4] with permission.
interval: 0.74, 1.21] ( p = 0.64). It was not possible to determine the median time to first relapse, since less than 50% of subjects experienced a relapse during the course of the study. The study investigators reported that the time for 30% of the patients to experience a first relapse was 495 days in the interferon-b 1a group and 432 days in the GA group. However, it should be noted that the choice of the 30% threshold is entirely arbitrary and, since the survival curves for the two groups were not parallel but intertwined, the use of a different threshold would have yielded a different relative time-to-event interval for the two groups. The mean annualised relapse rate over the 96-week study duration was 0.30 relapses/year in the interferon-b group and 0.29 relapses/year in the GA group. This represented a decrease of around 70% compared to pre-treatment relapse rates in both treatment groups. The proportion of patients who remained relapse-free throughout the study was 62% in both treatment groups. Similarly, the proportion of patients with confirmed disability progression was comparable between the two groups, 11.7 in the interferon-b arm and 8.7% in the GA arm. 3.3. MRI outcomes With respect to the subpopulation of patients with serial MRI evaluations, no difference was observed in the number of new or enlarging T2 lesions per patient per scan or in the number of new T1 hypointense lesions between the two groups (Table 2). A statistically significant difference ( p < 0.001) of Table 2 Magnetic resonance imaging outcomes in the REGARD study a Interferon-b 1a
Glatiramer acetate
p
T2 active lesions
0.67
0.82
0.18
Gd-enhancing lesions
0.24
0.41
<0.001
New T1 hypointense lesions
0.23
0.24
0.15
Change in brain volume a
S21
1.24%
1.07%
0.018
Lesion data correspond to mean numbers of lesions per patient per scan. Data are presented for the subset of 460 patients (230 per group) who underwent serial MRI evaluations. Data are taken from reference [4].
The overall proportion of patients reporting adverse events was similar in the two treatment groups (1,880 events reported in 90.8% of patients receiving interferon-b and 1,917 events reported in 85.3% of patients receiving GA). However, the nature of the adverse events differed between the two groups. Influenza-like illness, headache, myalgia and alanine amino transferase elevation were reported more frequently in the interferon-b group, while injection-site reactions, dyspnoea and immediate systemic post-injection reactions occurred more frequently in the GA group (Table 3). Overall, 17% of patients discontinued treatment during the course of the study, with a higher proportion of patients discontinuing in the interferon-b group (20.9%) compared to the GA group (13.6%). Table 3 Adverse events reported in the REGARD study a Adverse event
Interferon-b N = 386
Influenza-like illness
Glatiramer acetate N = 378
119 (31.2%)
5 (1.3%)
Injection site pain
47 (12.2%)
53 (14.0%)
Headache
74 (19.4%)
35 (9.3%)
Injection-site pruritus
8 (2.1%)
78 (20.0%)
Injection site bruising
33 (8.5%)
38 (10.1%)
Nasopharyngitis
34 (8.8%)
44 (11.6%)
4 (1.0%)
42 (11.2%)
Injection-site swelling Myalgia
22 (5.8%)
8 (2.1%)
Alanine amino transferase elevation
21 (5.5%)
19 (1.3%)
Injection-site induration
9 (2.4%)
25 (6.7%)
Dyspnoea
4 (1.0%)
22 (5.9%)
Immediate post-injection reaction a
None
19 (5.1%)
Only events reported in more than 10% of patients in either group or whose incidence differed significantly between the two treatment groups are listed. Adverse events in bold were significantly more frequent in the interferon-b group and those in italic significantly more frequent in the glatiramer acetate (GA) group. Data are taken from reference [4].
4. The BEYOND study 4.1. Objectives and design The BEYOND study was a much larger study initiated and implemented by Bayer-Schering, the manufacturers of interferon-b 1b sc 250 mg (Betaseron® ). Since the principal goal of the study was to use the data, in case of success, to secure registration of the higher 500 mg dose of interferon-b 1b, it was conducted according to current regulatory standards, thus ensuring a high quality of design and data monitoring and sufficient statistical power to test the study hypotheses. The BEYOND study attempted to test two hypotheses, firstly whether the higher dose of 500 mg of interferon-b 1b would have superior efficacy to the currently recommended dose of 250 mg, and secondly whether either dose would have
S22
A. Achiron, S. Fredrikson / Journal of the Neurological Sciences 277 (2009) S19 S24
superior efficacy to GA. The study included 2,244 treatmentnaive patients with RRMS fulfilling the McDonald criteria [8] for clinically definite disease from 198 study centres worldwide. These were randomised in a ratio of 2:2:1 to one of three treatment arms, namely interferon-b 500 mg (n = 899), interferon-b 250 mg (n = 897) or GA (n = 448). This is thus a very large study, in fact the largest randomised study in RRMS reported to date. As expected, the mean baseline characteristics of the patients included were comparable between all three treatment groups. The treatment period was twenty-four months, followed by an eighteen-month extension during which the doubleblind was maintained and systematic follow-up continued for patients completing the 24-month core study until all patients had completed the latter. Patients returned quarterly to the study centre for a complete neurological evaluation and were also contacted between these study visits by telephone. Yearly MRI scans were performed. Particular attention was paid to concealment of the randomisation. The two interferon-b groups were compared under double-blind conditions. Due to the difference in administration frequency, it was not possible to compare the interferon-b groups with the GA group under double-blind conditions. However, the rater who performed the quarterly neurological assessment was blinded as to treatment and the MRI scans were evaluated centrally and anonymously. The primary outcome measure was the hazard ratio for multiple relapses determined with the Anderson Gill extension of the Cox proportional hazards model. This rigorous statistical method allows differences in relapse frequency between individual patients to be taken into account in the analysis, and represents another strength of the study. The statistical testing followed a hierarchical null-hypothesis approach, in which one-to-one group comparisons were tested in the following order: interferon-b 500 mg with interferon-b 250 mg, then interferon-b 500 mg with GA and finally interferon-b 250 mg with GA. Secondary outcomes included time to confirmed EDSS progression and the development of T1 hypointense lesions (‘black holes’). Supportive outcome variables measured included the proportion of patients free from disease activity, a composite measure defined as the proportion of patients who experienced no relapses and who developed no new T2 lesions or Gd-enhancing lesions following initiation of treatment, the time to first relapse, the number and volume of T2 lesions, and changes in total brain volume. 4.2. Clinical outcomes Clinical outcomes from the BEYOND study have been reported in abstract form at two recent neurology congresses [7,13]. In the primary outcome evaluation, the hazard ratio for multiple relapses in all three pair-wise comparisons was not significantly different from unity, indicating no difference in relapse risk between any of the three treatments [7]. During the two-year treatment period, the mean annualised relapse rate decreased by approximately 80% in all three treatment groups. No significant inter-group difference was found in the proportion of relapse-free patients, in time to first relapse or in accumulation of disability. With respect to the composite endpoint of the proportion of patients free from all disease activity, this represented 31.5% in the interferon-b 500 mg group, 25.6% in the
interferon-b 250 mg group and 25.7% in the GA group. A significant difference in favour of interferon-b 500 mg with respect to both interferon-b 250 mg ( p = 0.007) and GA ( p = 0.031) was reported [13]. This difference was attributed to a more pronounced effect of the higher interferon-b dose on gadolinium-enhancing lesions and on new or enlarging T2 lesions and was not translated into concrete clinical differences during the observation period of this study. 4.3. MRI outcomes With respect to MRI outcomes [14], no significant differences were observed for most measures. In particular, no significant differences between the groups in the number or volume of T1 Gd-enhancing lesions at study end, in the accumulation of T1 hypointense lesions (‘black holes’) or in changes in total brain volume between the three treatment groups. Exceptions were the cumulative number of T2 lesions up to the last scan and the relative increase in T2 lesion volume, which were both significantly higher in the GA group than in either of the interferon-b groups. However, the differences observed were relatively small. 4.4. Tolerability Tolerability was assessed systematically in the BEYOND study by monitoring the incidence of adverse events and by regular blood tests. Adherence to treatment was also reported. All three treatments were relatively well tolerated [15]. The adverse events reported were consistent with the known safety profiles of interferon-b and GA. During the first year of treatment, influenza-like symptoms were reported significantly more frequently in the interferon-b 250 mg group compared to the GA group. Although initially high, the frequency of influenza-like symptoms in interferon-b-treated patients decreased rapidly over time. Injection-site reactions, notably pain and pruritus, were significantly more frequent in the GA treatment group than in the interferon-b 250 mg group but again, the incidence of these reactions diminished over the course of the study. The incidence of other adverse events such as fatigue, depression, arthralgia and paraesthesia was comparable between treatment groups. Concerning laboratory safety evaluation, elevations in liver enzymes, anomalies in blood lipids and markers of thyroid function, as well as leukocytopenia, were more frequently detected in patients treated with interferon-b 1b than in those treated with GA. Adherence to treatment was high in all three treatment groups. The proportion of patients completing the anticipated treatment period ranged from 73% in the interferon-b 500 mg group through 78% in the GA group to 82% in the interferon-b 250 mg group. No significant inter-group differences in discontinuation rates were observed. 5. Immunogenicity Data on neutralising antibodies (NAbs) are only available for the REGARD study. In this study, neutralising antibodies to interferon-b 1a were detected in 34% of patients, and these persisted until the end of the study in 27% of patients. Antibody titres were rather high in these patients, being over 1000 NU/ml in around one-third of the patients with persistent antibodies. Seroprevalence rates for NAbs in the REGARD study were unexpectedly high, being, for example,
A. Achiron, S. Fredrikson / Journal of the Neurological Sciences 277 (2009) S19 S24
twice as high as those reported in the PRISMS study (12.5% of patients) [10]. Time to first relapse was not significantly different in interferon-b-treated patients who developed NAbs compared to those who did not, suggesting that there was no correlation between the emergence of NAbs and clinical efficacy. However, it should be noted that the study duration (96 weeks) was short compared to the time necessary to observe a loss of clinical efficacy in patients developing NAbs in response to interferon-b [16]. Binding antibodies to GA were identified in 98% of patients at the first assessment at 24 weeks of treatment, as expected. GA generates an immune response resulting in the production of binding antibodies that do not compromise therapeutic efficacy and may contribute to efficacy [17]. A very surprising observation was that binding antibodies to GA were detected in around one quarter of patients before treatment was initiated. In principle, anti-GA antibodies are only detected in patients who have been exposed to GA. This finding suggests that either the patients were not treatment-naive contrary to what was specified in the inclusion criteria or that there was a problem with the assay used to detect the presence of antibodies. 6. Strengths and weaknesses of the directive comparative trials A major strength of the REGARD and BEYOND studies was the large number of patients included; these direct head-to-head comparisons evaluated over 2.5 times as many patients as the pivotal trials (Table 4). The BEYOND trial alone included twice as many patients as all pivotal trials of interferon-b combined. Further strengths include the randomised design, rater blinding, the relatively low drop-out rates, and the use of rigorously defined primary outcome measures. Table 4 Patient numbers in pivotal trials and direct comparative studies of immunomodulatory treatments in multiple sclerosis Trial(s)
No. of patients
Pivotal trials Interferon-b Interferon-b 1a im
301
Interferon-b 1b sc
372
Interferon-b 1a sc All Interferon-b Glatiramer acetate All pivotal trials
560 1233 251
7. Conclusions
Interferon 677
INCOMIN
188
All Interferon interferon comparisons
865
Interferon and Glatiramer acetate BECOME
75
REGARD
764
BEYOND
2244
All interferon glatiramer acetate comparisons
3083
All direct comparisons
Among the weaknesses, we may cite the open-label design of the studies. Performing an accurately blinded study with injectable treatments that differ in frequency of administration, form of application device and side-effect profile is obviously difficult. To some extent these problems could have been avoided by the use of a double-dummy design, although that would have exposed patients to additional injections, possibly compromising recruitment. Measurement biases should have been avoided by the requirement for the raters to be blinded to treatment assignment, and expectation biases, if they existed, cannot have played a major role since the study assumption that high-dose b-interferons would be superior to GA was actually not confirmed by the data. A further limitation relates to the low number of relapses observed in the REGARD and BEYOND trials, which may have compromised the statistical power of the studies to demonstrate between-group differences. In fact, the number of patients experiencing an on-treatment relapse was nearly half that anticipated in the power calculations of the REGARD study. However, there was no hint from the relapse data that any qualitative or quantitative difference between treatment groups existed. Disease activity at the time of inclusion was much lower in the study sample than in previous trials of these agents, which may explain the unexpectedly low number of relapses reported during the study. This is indicative of a change in the population of patients recruited into current clinical trials in multiple sclerosis compared with those performed over a decade ago. Indeed, the on-treatment relapse rate observed in the REGARD and BEYOND trials was considerably lower than those observed in the pivotal trials of interferon-b 1b, interferon-b 1a sc 44 mg and GA, which were 0.84, 0.87 and 0.59, respectively [10,11,18]. However, a relapse rate of 0.30 is very similar to that observed for interferon-b 1a sc 44 mg in another recent trial comparing this agent to alemtuzumab [19], as well as in recent trials of other active treatments such as natalizumab [20] and fingolimod [21]. This trend of a declining on-treatment relapse rate has several important consequences for the design of future trials. Firstly, future trials will require inclusion of a larger number of patients than was the case in the 1990s in order to generate sufficient power to detect a treatment effect. It also makes comparison between outcome in different clinical trials conducted in different populations at different times extremely difficult. However, on a positive note, it suggests that immunomodulatory treatment of patients with less active disease, who have been diagnosed earlier through the use of MRI, may provide better relapse control than had been anticipated from the pivotal trials.
1484
Direct comparative studies EVIDENCE
S23
3948
The conclusion from three well-conducted direct comparative studies is that the efficacy of GA is comparable to that of interferon-b 1b on a wide range of clinical and MRI outcomes [22,23]. Both interferon-b and GA were well tolerated, and adverse events reported were consistent with their known safety profiles. Conflict of interest statement S. Fredrikson declares having received speaker fees from Teva, Sanofi-Aventis, Bayer Schering, Biogen Idec, and Serono. No funding applicable.
S24
A. Achiron, S. Fredrikson / Journal of the Neurological Sciences 277 (2009) S19 S24
A. Achiron declares having received speaker fees from Teva, and research grants from Teva, Serono, and Bayer-Schering. References 1. Galetta SL, Markowitz C, Lee AG. Immunomodulatory agents for the treatment of relapsing multiple sclerosis: a systematic review. Arch Intern Med 2002;162:2161 9. 2. Panitch H, Goodin DS, Francis G, Chang P, Coyle PK, O’Connor P, et al. Randomized, comparative study of interferon beta-1a treatment regimens in MS: The EVIDENCE Trial. Neurology 2002;59:1496 506. 3. Durelli L, Verdun E, Barbero P, Bergui M, Versino E, Ghezzi A, et al. Everyother-day interferon beta-1b versus once-weekly interferon beta-1a for multiple sclerosis: results of a 2-year prospective randomised multicentre study (INCOMIN). Lancet 2002;359:1453 60. 4. Mikol DD, Barkhof F, Chang P, Coyle PK, Jeffery DR, Schwid SR, et al. Comparison of subcutaneous interferon beta-1a with glatiramer acetate in patients with relapsing multiple sclerosis (the REbif vs Glatiramer Acetate in Relapsing MS Disease [REGARD] study): a multicentre, randomised, parallel, open-label trial. Lancet Neurol 2008;7:903 14. 5. Wolansky L, Cook S, Skurnick J, Lincoln J, Tulloch K, Franco P, et al. Betaseron vs. Copaxone in MS with triple-dose gadolinium and 3-T MRI Endpoints (BECOME): announcement of final primary study outcome. Mult Scler 2007;13:P207. 6. Cadavid D, Wolansky L, Cook S, Halper J, Hill J, Younes M, et al. Betaseron versus copaxone in multiple sclerosis with triple-dose gadolinium and 3 Tesla MRI Endpoints (BECOME): announcement of secondary clinical outcomes. Mult Scler 2007;13:P207. 7. O’Connor P, Arnason B, Comi G, Filippi M, Cook S, Goodin D, et al. Interferon beta-1b 500 mcg, interferon beta-1b 250 mcg and glatiramer acetate: primary outcomes of the BEYOND (Betaferon® /Betaseron® Efficacy Yielding Outcomes of a New Dose) study. Neurology 2008;70(11; Late Breaking Science Supplement):S004. 8. McDonald WI, Compston A, Edan G, Goodkin D, Hartung HP, Lublin FD, et al. Recommended diagnostic criteria for multiple sclerosis: guidelines from the International Panel on the diagnosis of multiple sclerosis. Ann Neurol 2001; 50:121 7. 9. Kurtzke JF. Rating neurologic impairment in multiple sclerosis: an expanded disability status scale (EDSS). Neurology 1983;33:1444 52. 10. PRISMS (Prevention of Relapses and Disability by Interferon beta-1a Subcutaneously in Multiple Sclerosis) Study Group. Randomised doubleblind placebo-controlled study of interferon beta-1a in relapsing/remitting multiple sclerosis. Lancet 1998;352:1498 504.
11. Johnson KP, Brooks BR, Cohen JA, Ford CC, Goldstein J, Lisak RP, et al. The Copolymer 1 Multiple Sclerosis Study Group. Copolymer 1 reduces relapse rate and improves disability in relapsing remitting multiple sclerosis: results of a phase III multicenter, double-blind placebo-controlled trial. Neurology 1995;45:1268 76. 12. Zivadinov R, Reder AT, Filippi M, Minagar A, Stuve O, Lassmann H, et al. Mechanisms of action of disease-modifying agents and brain volume changes in multiple sclerosis. Neurology 2008;71:136 44. 13. Kappos L, Arnason B, Comi G, Cook S, Filippi M, Goodin D, et al. High proportion of patients free from disease activity in all 3 arms of the highdose Betaferon® trial. J Neurol 2008;255(Suppl 2):36. 14. Filippi M, Arnason B, Comi G, Cook S, Goodin D, Hartung H, et al. Magnetic resonance imaging findings of a phase III trial comparing Betaferon® with Copaxone® treatments in relapsing remitting multiple sclerosis. J Neurol 2008;255(Suppl 2):9. 15. Comi G, Arnason B, Cook S, Filippi M, Goodin D, Hartung H, et al. Tolerability and adherence to Betaferon® and Copaxone® in a phase III clinical trial. J Neurol 2008;255(Suppl 2):111. 16. Giovannoni G, Goodman A. Neutralizing anti-IFN-beta antibodies: how much more evidence do we need to use them in practice? Neurology 2005;65:6 8. 17. Brenner T, Arnon R, Sela M, Abramsky O, Meiner Z, Riven-Kreitman R, et al. Humoral and cellular immune responses to Copolymer 1 in multiple sclerosis patients treated with Copaxone. J Neuroimmunol 2001;115:152 60. 18. The IFNB Multiple Sclerosis Study Group and The University of British Columbia MS/MRI Analysis Group. Interferon beta-1b in the treatment of multiple sclerosis: final outcome of the randomized controlled trial. Neurology 1995;45:1277 85. 19. Coles A, CAMMS 223 Study Group. Alemtuzumab compared with subcutaneous high-dose IFNB-1a in treatment-naive relapsing remitting multiple sclerosis: primary efficacy outcomes of CAMMS223 at 3 years. Neurology 2008:70(Suppl 1): Abstract S22.006. 20. Polman CH, O’Connor PW, Havrdova E, Hutchinson M, Kappos L, Miller DH, et al. A randomized, placebo-controlled trial of natalizumab for relapsing multiple sclerosis. N Engl J Med 2006;354:899 910. 21. Kappos L, Antel J, Comi G, Montalban X, O’Connor P, Polman CH, et al. Oral fingolimod (FTY720) for relapsing multiple sclerosis. N Engl J Med 2006;355: 1124 40. 22. Sorensen PS. REGARD: what can we learn from randomised, open-label, head-to-head studies? Lancet Neurol 2008;7:864 6. 23. Goodin D. Comparative studies of glatiramer acetate and interferon beta. Int MS J 2008;15:39 41.