Special Article
`
`Disease modifying therapies in
`multiple sclerosis
`Report of the Therapeutics and Technology Assessment
`Subcommittee of the American Academy of Neurology and
`the MS Council for Clinical Practice Guidelines
`
`D.S. Goodin, MD; E.M. Frohman, MD; G.P. Garmany, Jr., MD; J. Halper, MSN, ANP, FAAN;
`W.H. Likosky, MD; F.D. Lublin, MD; D.H. Silberberg, MD; W.H. Stuart, MD; and S. van den Noort, MD
`
`progressive–relapsing MS (PRMS), also begins with
`a progressive course although these patients also ex-
`perience occasional attacks.
`
`Outcome measures in MS clinical trials. Evalua-
`tion of the relative effectiveness of different thera-
`pies requires consideration of which outcome
`measure or measures are relevant to the goals of
`therapy. Clearly, the most important therapeutic
`aim of any disease-modifying treatment of MS is to
`
`
`
`
`
`
`prevent or postpone long-term disability. However,p p p g y However,
`
`
`long-term disability in MS often evolves slowly overlong-term disability in MS often evolves slowly overg
`
`many years.1-3 Clinical trials, by contrast, study pa-many years.
`tients for only short periods of time (2 or 3 years)
`and, therefore, use only short-term outcome mea-
`sures to assess efficacy. As a result, it is important to
`y
`
`s a result, it is important top
`validate any short-term measure by its correlation
`
`validate any short-term measure by its correlationy y
`
`with the actual patient outcome many years later.
`with the actual patient outcome many years later.
`For a discussion of these issues, interested readers
`should consult the full-length assessment on the
`Neurology Web site at www.neurology.org.
`
`yA
`
`Scope of this guideline. The purpose of this as-
`sessment is to consider the clinical utility of these
`disease-modifying agents including the anti-in-
`flammatory, immunomodulatory, and immunosup-
`pressive treatments that are currently available.
`Symptomatic and reparative therapies will not be
`considered.
`Before considering the evidence from individual
`
`Overview. Clinical types of MS. MS is a chronic
`recurrent inflammatory disorder of the CNS. The
`disease results in injury to the myelin sheaths, the
`oligodendrocytes, and, to a lesser extent, the axons
`and nerve cells themselves.1-5 The symptoms of MS
`vary, depending in part on the location of plaques
`within the CNS. Common symptoms include sensory
`disturbances in the limbs, optic nerve dysfunction,
`pyramidal tract dysfunction, bladder or bowel dys-
`function, sexual dysfunction, ataxia, and diplopia.5
`Four different clinical courses of MS have been de-
`fined.6 The first, relapsing–remitting MS (RRMS), is
`characterized by self-limited attacks of neurologic
`dysfunction. These attacks develop acutely, evolving
`over days to weeks. Over the next several weeks to
`months, most patients experience a recovery of func-
`tion that is often (but not always) complete. Between
`attacks the patient is neurologically and symptomat-
`ically stable. The second clinical course, secondary
`progressive MS (SPMS), begins as RRMS, but at
`some point the attack rate is reduced and the course
`becomes characterized by a steady deterioration in
`function unrelated to acute attacks. The third clini-
`cal type, primary progressive MS (PPMS), is charac-
`terized by a steady decline in function from the
`beginning without acute attacks. The fourth type,
`
`Additional material related to this article can be found on the Neurology
`Web site. Go to www.neurology.org and scroll down the Table of Con-
`tents for the January 22 issue to find the title link for this article.
`
`Approved by the TTA Subcommittee August 3, 2001. Approved by the Practice Committee August 4, 2001. Approved by the AAN Board of Directors October
`20, 2001.
`Address correspondence and reprint requests to the American Academy of Neurology, 1080 Montreal Ave., St. Paul, MN 55116.
`
`Copyright © 2002 by AAN Enterprises, Inc. 169
`
`Merck 2014
`TWi v Merck
`IPR2023-00049
`
`

`

`trials, however, a few statistical and interpretational
`points are worth bearing in mind. First, although a p
`value of 0.05 is commonly taken as evidence of a
`therapeutic benefit to treatment, there is concern
`that this may be too liberal a standard. For example,
`the Type I error rate (i.e., the so-called ␣-error) re-
`flects the likelihood of concluding incorrectly that a
`useless treatment is of value. Surprisingly, however,
`for an experimental observation with a p value of
`0.05, the calculated (i.e., theoretically expected) min-
`imum Type I error rate, for a two-tailed comparison,
`is actually 13%.7-10 For a one-tailed comparison, this
`minimum Type I error rate is actually 21%.7-10 Thus,
`if the aim is to reduce the Type I error rate to the
`nominal value of 5% for statistical significance (for
`a single comparison), using this type of analysis,
`the observed p value would need to be ⱕ0.01.7-10
`Consequently, when evaluating the results from a
`particular trial, statistical observations between p ⫽
`0.01 and p ⫽ 0.05 should be regarded as marginal.
`This is especially true when the study under consid-
`eration reports multiple between-group statistical
`comparisons, because multiple comparisons mark-
`edly inflate the actual Type I error rate and require
`a much more stringent statistical adjustment.11-15
`There is also concern about the Type II error rate of
`clinical trials (i.e., the so-called ␤ error), which re-
`flects the likelihood of concluding incorrectly that a
`useful treatment is of no value.16 For example, one
`recent trial17 found that after 2 years of treatment,
`sustained disability progression was nonsignificantly
`reduced by 12%. Clearly, such a result cannot be
`used to reject a true 12% reduction in this measure,
`and, in fact, this nonsignificant observation is still
`compatible with an even more robust treatment ef-
`fect.16 The issue is the statistical power (i.e., 1-␤) of
`the clinical trial to detect group differences and this,
`in turn, is related to the number of subjects stud-
`ied.16 In this particular trial,17 the number of subjects
`studied (i.e., 251) provided insufficient power to de-
`tect a 12% change on this outcome. If a much larger
`number of subjects had been entered into the trial,
`and if the same magnitude and variability of the
`treatment effect had been obtained, this change
`would have been statistically significant. As a conse-
`quence of such difficulties, it is important to recog-
`nize that negative results from small clinical trials
`generally provide little assurance that a true treat-
`ment effect has not been missed. Second, because it
`is uncertain which outcome measures correlate best
`with future function, clinical trials that use a combi-
`nation of outcome measures, including both clinical
`and confirmatory MRI measures, should be judged
`as stronger evidence than those that rely on only a
`single measure, especially when that measure is a
`subjective clinical score. Third, it is important to rec-
`ognize that both the statistical significance of a find-
`ing and the magnitude of the treatment effect (i.e.,
`the effect-size) provide important complementary in-
`formation about the quality of the evidence. The sta-
`tistical significance relates to the believability of a
`170 NEUROLOGY 58 January (2 of 2) 2002
`
`result, whereas the effect size relates to its clinical
`importance. Trials with large effects of marginal sig-
`nificance and trials with significant effects of mar-
`ginal importance should both be judged as providing
`equivocal evidence. Fourth, it should be noted that
`treatments aimed at limiting future CNS injury
`would not be expected to cause an already disabled
`patient to improve dramatically, even though some
`patients may experience some clinical improvement
`based on intrinsic self-repair mechanisms. Conse-
`quently, reports of substantial improvement follow-
`ing the use of such agents should be viewed with
`caution.
`A synopsis of the conclusions and recommenda-
`tions for all the treatments considered is provided in
`the Summary. The actual analysis of the evidence
`(table), however, is provided here only for the immu-
`nomodulatory treatments. Readers interested in the
`analysis of the evidence for other therapies should
`consult the full-length assessment on the Neurology
`Web site at www.neurology.org.
`
`Analysis of the evidence.
`Immunomodulatory
`Interferon beta. Clinical trial results.
`treatments.
`The multicenter study of IFN␤-1b (Betaseron; Berlex
`Laboratories, Montville, NJ) in RRMS18-20 was ran-
`domized, double-blind, and placebo-controlled (Class
`I evidence). It included 372 patients with RRMS who
`had scores on the extended disability status scale
`(EDSS) ⱕ5.5 and who had experienced at least two
`attacks in the prior 2 years. Patients were random-
`ized to receive placebo, low-dosage (1.6 million of
`International Units [MIU]; 50 ␮g), or high-dosage (8
`MIU; 250 ␮g) IFN␤-1b subcutaneously (SC) every
`other day for 2 years. After 2 years, compared with
`placebo, treatment with high-dosage IFN␤-1b re-
`duced the clinical relapse rate (⫺34%; p ⬍ 0.0001),
`which was the primary endpoint of the study. In
`addition, the MRI attack rate as measured by me-
`dian number of T2 active lesions (⫺83%; p ⬍ 0.009)
`and the median volume of MRI T2 disease burden
`(⫺17.3%; p ⫽ 0.001) were reduced in the IFN␤-1b
`arm compared with placebo-treated patients. The
`high dosage also resulted in a reduction in the con-
`firmed 1-point EDSS progression rate, but this was
`not statistically significant (⫺29%; p ⫽ 0.16). This
`trial, however, did report a reduction in the uncon-
`firmed 1-point EDSS worsening over 3 years of study
`(⫺31%; p ⫽ 0.043).
`In summary, this trial provides (Class I) evidence
`that IFN␤ reduces the relapse rate (measured either
`clinically or by MRI) in patients with RRMS. The
`effect of treatment on measures of disease severity
`(i.e., MRI disease burden and disability progression)
`is less consistent. There was a robust effect of treat-
`ment on the MRI disease burden but no statistically
`significant effect on the measure of confirmed 1-point
`EDSS progression.
`The IFN␤-1a (Avonex; Biogen, Cambridge, MA)
`trial21-23 also was multicenter, randomized, and
`placebo-controlled (Class I evidence). It included 301
`
`

`

`Table Rating of evidence classification scheme
`
`Rating of recommendation
`
`Translation of evidence to
`recommendations
`
`A—Established as effective,
`ineffective, or harmful for the
`given condition in the specified
`population.
`
`Level A rating requires at least
`one convincing Class I study or
`at least two consistent,
`convincing Class II studies.
`
`Level B rating requires at least
`one convincing Class II study or
`at least three consistent Class
`III studies.
`Level C rating requires at least
`two convincing and consistent
`Class III studies.
`
`B—Probably effective, ineffective,
`or harmful for the given
`condition in the specified
`population.
`C—Possibly effective, ineffective,
`or harmful for the given
`condition in the specified
`population.
`
`U—Data inadequate or conflicting.
`Given current knowledge,
`treatment is unproven.
`
`Rating of therapeutic article
`
`Class I: Prospective, randomized, controlled clinical trial
`with masked outcome assessment, in a representative
`population. The following are required:
`a. primary outcome(s) is/are clearly defined,
`b. exclusion/inclusion criteria are clearly defined,
`c. adequate accounting for dropouts and crossovers
`with numbers sufficiently low to have minimal
`potential for bias,
`d. relevant baseline characteristics are presented and
`substantially equivalent among treatment groups
`or there is appropriate statistical adjustment for
`differences.
`Class II: Prospective matched group cohort study in a
`representative population with masked outcome
`assessment that meets a–d above or a RCT in a
`representative population that lacks one criteria a–d.
`Class III: All other controlled trials (including well-
`defined natural history controls or patients serving as
`own controls) in a representative population, where
`outcome assessment is independent of patient
`treatment.
`Class IV: Evidence from uncontrolled studies, case
`series, case reports, or expert opinion.
`
`patients with RRMS who had an EDSS score of 1.0 –
`3.5, and who had experienced at least two attacks in
`the 3 years prior to entering the study. Patients were
`treated either with placebo or IFN␤-1a, 6 MIU/wk
`(30 ␮g/wk), intramuscularly (IM) for 2 years. This
`trial was stopped earlier than originally designed, so
`only 57% (172 patients) completed the full 2 years on
`study medication. Compared with placebo, treatment
`with Avonex for 2 years produced a reduction in the
`confirmed 1-point EDSS progression rate (⫺37%; p ⫽
`0.02), which was the primary endpoint of the trial. In
`addition, the clinical attack rate (⫺18%; p ⫽ 0.04)
`and the MRI attack rate as measured by the median
`number of gadolinium enhancing lesions (⫺33%; p ⫽
`0.05) were reduced in the IFN␤-1a arm compared
`with placebo-treated patients. The total volume of T2
`disease burden seen on MRI also was reduced com-
`pared with placebo, but this was not statistically
`significant (⫺6.7%; p ⫽ 0.36). This trial also found
`that the reduction in attack rate in the first year of
`therapy (⫺9.6%, not significant) was less than the
`reduction in patients who had completed 2 years of
`therapy (⫺32%; p ⫽ 0.002), suggesting that the full
`clinical benefits of IFN␤-1a therapy might be de-
`layed for a year or more after the initiation of treat-
`ment.21,24,25 Nevertheless, the authors provide no
`statistical evidence of a difference between the
`1-year and 2-year data, and, in addition, the other
`IFN␤ trials in RRMS did not observe such a delay in
`therapeutic benefit.18-20,24,26,27 Most important, how-
`ever, this subgroup of patients (who had a 32% re-
`duction in attack rate over 2 years) had a similar
`reduction in attack rate (⫺29%) at the 1-year mark.25
`
`Such an observation indicates that this particular
`subgroup of patients (i.e., the 2-year completers) is
`not representative of the study cohort as a whole. As
`a result of this anticipated bias, the validity of any
`separate analysis on this subgroup of patients is
`questionable. A re-analysis of the trial data (for the
`subgroup of 2-year completers only) using the “brain
`parenchymal fraction” to measure brain atrophy28
`showed no statistically significant reduction in brain
`atrophy after 2 years of treatment (p ⫽ 0.30). A
`subgroup analysis did show a reduction of accumu-
`lated atrophy in the second year of treatment (p ⫽
`0.03). This latter observation, however, was only
`marginally significant and was the result of a post
`hoc analysis on a biased subset of the study popula-
`tion, and the reported p value was not adjusted for
`the three between-group statistical comparisons of
`brain parenchymal fraction presented in the article’s
`figure.28 Therefore, the validity of this observation is
`uncertain.
`In summary, this trial provides (Class I) evidence
`that IFN␤-1a reduces the biologic activity of RRMS.
`Importantly, the results of this trial replicate, in
`general, the earlier IFN␤-1b trial for both clinical
`and MRI outcomes, although again the effect of
`treatment on attack rate measures was more consis-
`tent than for measures of disease severity. Thus,
`both clinical and MRI measures of attack rate were
`similarly improved at 2 years. In addition, there was
`a reduction in the confirmed 1-point EDSS progres-
`sion rate, although there was no statistically signifi-
`cant concomitant benefit on either MRI disease
`burden or brain atrophy during the 2 years of study.
`January (2 of 2) 2002 NEUROLOGY 58
`171
`
`

`

`The IFN␤-1a (Rebif; Serono International SA, Ge-
`neva, Switzerland) trial26,27 was a similarly random-
`ized, multicenter, double-blind, and placebo-
`controlled study (Class I evidence). A total of 560
`patients with RRMS with an EDSS score ⱕ5.0 were
`entered. Only patients who had experienced 2 or
`more relapses in the prior 2 years were included.
`Patients were treated for 2 years with placebo or
`IFN␤-1a at dosages of either 22 ␮g (6 MIU) or 44 ␮g
`(12 MIU) SC three times weekly. After 2 years, there
`was a significant beneficial effect of treatment with
`either dose on both clinical and MRI outcome mea-
`sures. Thus, compared with placebo, treatment with
`IFN␤-1a, 132 ␮g/wk (36 MIU/wk), reduced the clini-
`cal attack rate (⫺32%; p ⬍ 0.005), which was the
`primary endpoint of the trial. In addition, the MRI
`attack rate as measured by median number of T2
`active lesions (⫺78%; p ⬍ 0.0001), the volume of
`white matter disease seen on T2-weighted MRI
`(⫺14.7%; p ⬍ 0.0001), and the confirmed 1-point
`EDSS progression rate (⫺30%; p ⬍ 0.05) also were
`reduced in the IFN␤-1a arm compared with placebo.
`In summary, this trial provides (Class I) evidence
`that IFN␤-1a reduces the biologic activity of RRMS.
`As in other IFN␤ trials, this trial demonstrated a
`benefit to treatment on both clinical and MRI mea-
`sures of attack rate. Also, this was the first trial of
`IFN␤ in RRMS to show both a reduction in the con-
`firmed 1-point EDSS progression and a highly signif-
`icant reduction in the T2 disease burden.
`The IFN␤-1b (Betaferon; Schering AG, Berlin,
`in SPMS29 was a randomized,
`Germany)
`trial
`placebo-controlled, double-blinded study conducted
`among 32 European centers (Class I evidence). In-
`cluded were 718 patients with an EDSS of 3.0 – 6.5.
`Patients had to have either two relapses or more
`than a 1.0 point increase in EDSS in the prior 2
`years. Those included were randomized to receive
`either placebo or IFN␤-1b, 250 ␮g (8 MIU) SC, every
`other day for up to 3 years. Compared with treat-
`ment with placebo, treatment with 28 MIU/wk
`Betaferon reduced the confirmed 1-point EDSS pro-
`gression rate (⫺22%; p ⫽ 0.0008), the primary end-
`point of the study. In addition, the clinical attack
`rate (⫺31%; p ⫽ 0.0002), the MRI attack rate (⫺78%;
`p ⫽ 0.0008), and the volume of white matter disease
`seen on MRI (⫺13%; p ⫽ 0.0001) all were signifi-
`cantly reduced in the IFN␤-1b arm compared with
`placebo. This study also demonstrated that treat-
`ment with IFN␤-1b reduced the likelihood of becom-
`ing wheelchair bound during the study (⫺33%; p ⫽
`0.01). After dividing patients into those who had ex-
`perienced clinical attacks in the 2 years before enter-
`ing the study and those who only experienced steady
`clinical deterioration, the benefit of treatment was
`comparable in both subgroups. After dividing pa-
`tients into those who did and those who did not expe-
`rience attacks during the trial, the benefit of
`treatment was again found to be similar in the two
`subgroups. After dividing patients into three groups
`based on their baseline EDSS scores (Group 1 ⫽
`NEUROLOGY 58 January (2 of 2) 2002
`172
`
`3.0 –3.5; Group 2 ⫽ 4.0 –5.5; and Group 3 ⫽ 6.0 – 6.5),
`IFN␤-1b was found to be similarly beneficial in all
`three groups. However, when the full 3-year data are
`analyzed, the benefit of treatment in patients with
`an EDSS ⱖ 6.0 is not apparent.
`In summary, this trial provides (Class I) evidence
`that treatment with IFN␤-1b favorably impacts both
`clinical and MRI outcomes for attack rate and dis-
`ease severity in patients with SPMS.
`The results of another recently completed (Class I)
`trial of IFN␤-1b (Betaseron) in SPMS also has been
`reported in preliminary form.30 This trial failed to
`find a statistically significant reduction in the con-
`firmed 1-point EDSS progression rate (the primary
`endpoint of the trial), although it did report signifi-
`cant reductions in the clinical attack rate, the MRI
`attack rate, and the volume of white matter disease
`found on T2-weighted MRI. Publication of the final
`results from this trial is pending. The reason for the
`apparently discrepant findings between these two
`trials of IFN␤-1b is not clear. Some observers have
`noted that the North American cohort of patients
`had significantly fewer attacks than their European
`counterparts, and that perhaps IFN␤ is most effec-
`tive in the relapsing phase of the illness. At the
`moment, however, such a notion is speculative.
`The recently published trial of IFN␤-1a (Rebif) in
`SPMS31,32 also failed to find a statistically significant
`reduction in the confirmed 1-point EDSS progression
`rate (the primary endpoint of the trial). Like the
`IFNb-1b (Betaseron) trial, however, this trial also
`found significant reductions in the clinical attack
`rate, the MRI attack rate, and the volume of white
`matter disease found on T2-weighted MRI. Also,
`when the results of this trial were reanalyzed by
`separating patients into those with and those with-
`out attacks, a benefit to treatment on the confirmed
`1-point EDSS progression rate was noted (p ⫽ 0.027)
`in patients with relapses. The validity of such a re-
`analysis of the data is clearly open to question, but
`nevertheless might be taken as weak support for the
`speculation (noted above) that IFN␤ is more effective
`in patients with SPMS who continue to experience
`relapses.
`IFN␤-1a
`(Class I) study of
`Another recent
`(Avonex) in the treatment of SPMS has been re-
`ported in preliminary form.33 Using the MS func-
`tional composite as the primary outcome, this trial
`found that, compared with placebo, treatment with
`IFN␤-1a, 60 ␮g/wk, IM was beneficial over a 2-year
`period (p ⫽ 0.03). This study, however, did not find
`any concomitant benefit on the outcome of confirmed
`1-point EDSS progression. Moreover, the benefit
`seen on the MS functional composite outcome was
`due primarily to the results from the Nine-Hole Peg
`Test portion of the composite score. The reported
`benefit of therapy in this trial, therefore, is of uncer-
`tain reliability.
`Two recently completed trials of IFN␤-1a (Avonex
`and Rebif) in patients at high risk of developing MS
`have shown that early treatment significantly slows
`
`

`

`the subsequent rate of conversion to clinically defi-
`nite MS (CDMS).34,35 The IFN␤-1a (Avonex) trial34
`was a multicenter, randomized, placebo-controlled
`trial involving 383 patients who were followed for up
`to 3 years (Class I evidence). Patients needed to have
`just experienced their first clinically isolated (mono-
`symptomatic) CNS event consisting of an optic neu-
`ritis, a spinal cord syndrome, or a brainstem/
`cerebellar syndrome. Patients also had to have an
`abnormal brain MRI defined as two or more clini-
`cally silent lesions (ⱖ3 mm) on T2-weighted MRI
`scans, at least one of which needed to be ovoid in
`appearance or periventricular in location. Patients
`initially were treated with intravenous methylpred-
`nisolone, 1 g/d for 3 days, followed by a course of oral
`prednisone, 1 mg/kg/d for 15 days. Patients subse-
`quently received either IFN␤-1a (30 ␮g/wk, IM) or
`placebo throughout the study. Using a Cox propor-
`tional hazard model, the relative risk of developing
`CDMS in the treated group was 0.56 (p ⫽ 0.002),
`indicating a 44% decrease in the rate of conversion to
`MS after administration of IFN␤-1a, which was the
`primary endpoint of the trial. MRI measures also
`demonstrated a robust treatment effect. Thus, at 18
`months, the number of new lesions (⫺57%; p ⬍
`0.0001), the percentage change in the T2 lesion vol-
`ume (⫺14%; p ⫽ 0.0004), and the number of enhanc-
`ing lesions (⫺67%; p ⬍ 0.0001) all were reduced
`using IFN␤-1a when compared with placebo. The
`IFN␤-1a (Rebif) trial35 also was a multicenter ran-
`domized trial (Class I evidence) involving 309 pa-
`tients who had experienced their first clinical
`episode suggestive of demyelinating disease (either
`mono- or polysymptomatic) and who were followed
`for 2 years thereafter. Patients received either
`IFN␤-1a (22 ␮g/wk, SC) or placebo throughout the
`study. The proportion of patients converting to
`CDMS was less in the treated group compared with
`placebo (⫺24%; p ⫽ 0.047). In addition, the median
`number of T2 active lesions seen on MRI also was
`reduced in the treated compared with placebo pa-
`tients (p ⬍ 0.001). The T2 disease burden also was
`reduced in the treated arm compared with placebo in
`both year 1 and year 2 of the trial (p ⫽ 0.006 and p ⫽
`0.002, respectively).
`These trials, therefore, provide (Class I) evidence
`that treatment with IFN␤-1a delays the develop-
`ment of CDMS in patients at high risk for this out-
`come. Such a result is hardly surprising. Indeed, any
`treatment for RRMS that can delay the time be-
`tween attacks 2 and 3 or between attacks 3 and 4
`(i.e., any treatment that reduces the attack rate) also
`would be expected to delay the time between attacks
`1 and 2. These studies do not, however, provide evi-
`dence that the ultimate development of CDMS is
`prevented by such treatment. Neither do they pro-
`vide any evidence that early treatment affects long-
`term disability outcome.
`Effects of IFN␤ type, route of administration, and
`dose on clinical outcome. The total dosage of IFN␤
`used in the different clinical trials of both RRMS and
`
`SPMS has varied considerably between studies and
`it is important to consider the evidence that there
`may be a dose-response curve in the use of IFN␤ for
`the management of patients with MS. Because the
`pharmaceutical
`companies
`that manufacture
`Avonex, Betaseron, and Rebif use slightly different
`assays to measure IFN␤ activity, the MIU scales
`reported in the different papers are not directly com-
`parable between publications. Nevertheless, because
`Avonex and Rebif are both forms of IFN␤-1a, they
`can be compared on a microgram for microgram ba-
`sis. Also, the conversion of IFN␤-1a to IFN␤-1b doses
`can be calculated using published data,36 with the
`result that 6 MIU Avonex (30 ␮g) is equivalent to
`approximately 7-9 MIU Betaseron (220-280 ␮g).
`IFN␤ induces the expression of many gene prod-
`ucts and interferon-specific markers, including 2',5'-
`oligoadenylate synthetase (2',5'-OAS), neopterin,
`tryptophan, ␤2-microglobulin, and human Mx pro-
`tein.37 These markers reflect a range of biologic activ-
`IFN␤,
`ities
`of
`including MHC Class-I gene
`expression, antiviral and antiproliferative actions,
`and monocyte activation. These markers have been
`used as indicators of the biologic activity of IFN␤.
`The relative dose of the different preparations also
`can be assessed from another recent publication38 in
`which antiviral protein (MxA) stimulation was stud-
`ied in the untreated blood from 10 healthy volunteer
`subjects. In this study, in vitro stimulation of periph-
`eral blood with all three agents (Avonex, Betaseron,
`and Rebif) resulted in a dose-dependant increase in
`MxA levels that was roughly equivalent for each
`agent on a MIU for MIU basis using the published
`MIU values.
`One study39 initially suggested that IM adminis-
`tration of IFN␤-1a caused a substantially greater
`area under the concentration-time curve for IFN␤
`activity in the serum compared with SC administra-
`tion. By contrast, a different study36 compared the
`effects of IFN␤-1a given SC and IM and IFN␤-1b
`given SC on neopterin, human Mx protein, and 2',5'-
`OAS in 75 healthy volunteer subjects. IFN␤-1a was
`administered at doses of 1, 3, 6, 9, and 12 MIU and
`IFN␤-1b at doses of 2, 4, 8, 12, and 16 MIU; each
`patient in the study received a single dose. The re-
`sults showed that the production of all three mark-
`ers was induced in a dose-dependent manner for
`both IFN␤-1a and IFN␤-1b. Moreover, this study
`found no differences in any of these biologic effects
`between the two types of IFN␤ or between the differ-
`ent routes of administration. Similar results have
`been found by other investigators.40,41 Thus, the bal-
`ance of the evidence favors the view that the route of
`IFN␤ administration is not of clinical importance.
`The previously cited study38 also examined the
`levels of MxA in the peripheral blood in 237 patients
`with CDMS after administration of IFN␤. There
`were 78 patients receiving IFN␤-1b (Betaseron) at a
`dosage of 8 MIU (250 ␮g) every other day; 71 pa-
`tients receiving IFN␤-1a (Rebif) at a dosage of 6 MIU
`(22 ␮g) SC either weekly or three times weekly; and
`January (2 of 2) 2002 NEUROLOGY 58
`173
`
`

`

`21 patients receiving IFN␤-1a (Avonex) at a dosage
`of 6 MIU (30 ␮g) IM once weekly. The level of MxA
`was 2.29 ng/105 peripheral blood lymphocytes (PBL)
`in the Betaseron-treated patients, 1.00 ng/105 pe-
`ripheral blood lymphocytes in the Rebif-treated pa-
`tients, and 0.57 ng/105 peripheral blood lymphocytes
`in the Avonex-treated patients. In summary, the re-
`sults of this trial suggest that increasing the total
`weekly IFN␤ dose is associated with an increasing
`biologic effect (Class II evidence). However, whether
`the measured biologic effect (on MxA levels) is rele-
`vant to the effect of IFN␤ on disease activity, cannot
`be assessed from this trial.
`The results of the pivotal clinical trials of IFN␤ in
`RRMS also suggest a dose-response curve.18-27 Thus,
`in general, when comparing the different findings of
`these trials, both the magnitude of the reported ef-
`fects on clinical and MRI outcomes, as well as their
`statistical significance, seem to be greater with in-
`creasing dosages of IFN␤. Nevertheless, because of
`differences in trial design, differences in the MS pop-
`ulations studied, and the fact that the results were
`obtained in independent clinical trials, this observa-
`tion can only be considered as weak (Class III) evi-
`dence of a dose response.
`The findings from the two placebo-controlled Class
`I IFN␤ studies that investigated different doses of
`IFN␤ provide mixed results.18-20,26,27 Thus, in the Be-
`taseron trial,18-20 treatment with low-dose IFN␤-1b
`(5.6 MIU/wk) was significantly better than placebo
`(p ⬍ 0.01) on the measure of clinical attack rate over
`the first 2 years, although it was significantly less
`effective on this measure (p ⬍ 0.0086) than the
`higher dose of 28 MIU/wk. Trends in favor of the
`higher dose also were seen on other outcome mea-
`sures, although no other statistically significant dose
`effects were noted. In the Rebif trial,22,26 both doses
`were highly effective, although the high-dose arm
`did better on each clinical and MRI outcome measure
`than the low-dose (18 MIU/wk) arm. With the excep-
`tion of the outcome of T2 active lesions (p ⫽ 0.0003
`comparing low dose to high dose), however, there
`were no statistical differences between the two doses
`at the 2-year time point. Thus, although based on
`high-quality (Class I) studies, the evidence in favor
`of a dose response provided by these trials is only
`equivocal.
`The Rebif trial was continued for an additional 2
`years.42 Placebo-treated patients during the first 2
`years were re-randomized in a double-blind fashion
`to receive IFN␤-1a, either 66 ␮g or 132 ␮g weekly, in
`divided doses. After 4 years, a dose-response rela-
`tionship was seen for some clinical and MRI out-
`comes but not for others. Thus, the high dose was
`more effective than the low dose (p ⬍ 0.05) at reduc-
`ing the relapse rate during years 3 and 4, prolonging
`the time to second relapse, and increasing the per-
`centage of relapse free patients. Similarly, treatment
`with high dose IFN␤-1a reduced the MRI disease
`burden and T2 lesion activity (p ⬍ 0.001) compared
`with low dose (Class I evidence). By contrast, the
`NEUROLOGY 58 January (2 of 2) 2002
`174
`
`high-dose group was not statistically better than the
`low-dose group on the outcomes of attack rate mea-
`sured over years 1 to 4 (⫺12%; p ⫽ 0.069), or the
`time to confirmed 1-point EDSS progression (⫹17%;
`p ⫽ 0.33). In addition, an analysis (Class III evi-
`dence) of the combined results of the Avonex and
`Rebif trials suggested that IFN␤-1a has increasing
`clinical efficacy (as measured by the clinical attack
`rate at 1 year) between the doses of 22 and 132 ␮g
`weekly.24 By contrast, the results of the SPECTRIMS
`trial of IFN␤-1a in SPMS demonstrated no difference
`between 66 and 132 ␮g weekly with respect to any
`clinical outcome measure relating to relapse rate.31
`The results of a multicenter, double-blind, dose-
`comparison trial of IFN␤-1a (Avonex) recently has
`been reported.43 This trial included 678 patients with
`RRMS who received IFN␤-1a, either 30 ␮g/wk or 60
`␮g/wk, IM once weekly for a period of at least 3 years
`(Class I evidence). There was no difference in out-
`come between the two dosage groups with respect to
`EDSS progression, relapse rate, gadolinium (Gd)-
`enhancing lesions, T2 lesion burden, or brain atro-
`phy over the course of the trial.43 This trial thus
`provides Class I evidence that 60 ␮g IFN␤-1a (IM)
`once weekly provides no additional benefit over 3
`years of therapy compared with 30 ␮g (IM) once
`weekly over the same period.
`Recently, the preliminary results of two head-to-
`head comparison trials of different IFN␤ prepara-
`tions have been reported.44,45 The first44 was a 2-year,
`open-label, randomized trial of IFN␤-1b (Betaseron;
`28 MIU/wk, SC) compared with IFN␤-1a (Avonex; 30
`␮g/wk, IM) in 188 patients with RRMS. Only the
`data after 1 year of therapy have been presented.
`This trial found a greater clinical benefit in the
`higher dose (more frequently administered)