Abstract
Background: The beneficial effect of interferon beta on exacerbations in relapsing-remitting MS has been demonstrated repeatedly, but results concerning disability vary. Objective: This multicenter, randomized, parallel-group, placebo-controlled study tested two doses of interferon beta-1a in patients with secondary progressive MS, which may include relapses but is dominated by accumulating disability. Methods: A total of 618 patients received subcutaneous placebo or interferon beta-1a, 22 or 44 μg three times weekly for 3 years. Patients were assessed every 3 months. Results: The primary outcome, time to confirmed progression in disability, was not significantly affected by treatment (hazard ratio, 0.83; 95% CI, 0.65 to 1.07; p = 0.146 for 44 μg versus placebo). Relapse rate was reduced from 0.71 per year with placebo to 0.50 per year with treatment (p < 0.001 for both doses). Significant treatment effects were seen on other exacerbation-related outcomes and on a composite measure incorporating five separate clinical and MRI outcomes. The hazard ratio for time to progression for the combined interferon beta-1a groups compared with placebo was 0.74 among patients reporting relapses in the 2 years before study (p = 0.055), and 1.01 for those without prestudy relapses (p = 0.934). An unexpected treatment-by-sex interaction favored women. The drug was well tolerated. Conclusions: Treatment with interferon beta-1a did not significantly affect disability progression in this cohort, although significant treatment benefit was observed on exacerbation-related outcomes. Exploratory post hoc analyses suggested greater benefit in women and in patients who had reported at least one relapse in the 2 years before the study.
MS is a chronic disabling disease of the CNS that presents in either relapsing or progressive forms. The most common form at onset, relapsing-remitting MS (RRMS), affects 70% to 80% of patients. The remainder of patients present with primary progressive or progressive relapsing disease.1 Most patients with RRMS eventually develop the secondary progressive (SP) form, in which disability accumulates steadily with or without superimposed relapses.
Interferon beta (IFNβ) has demonstrated consistent reductions in exacerbation frequency in both RRMS2-4⇓⇓ and SPMS,5 ranging from 18% to 32% based on intent-to-treat analysis. Dose and frequency of administration appear to have an important effect, particularly in the first year. Weekly doses between 66 and 132 μg are associated with reductions in relapse rate of 33% to 37% compared with placebo during the first year,2,4⇓ whereas reductions of 0% to 19% are seen with weekly doses of 22 to 44 μg over the same period.2,3,6⇓⇓ Reduction in exacerbations in these trials were accompanied by highly significant effects on MRI variables, including proton density–T2-weighted, T1-weighted–gadolinium-enhanced, and combined unique (proton density–T2 plus T1–gadolinium) lesion activity, proportions of scans showing activity, and changes in lesion burden.4,7,8⇓⇓
Effect on disability has been less consistent: two IFNβ-1a trials in patients with RRMS that examined slightly different populations showed significant increases in time to sustained progression over 2 years,3,4⇓ whereas a trial of IFNβ-1b in RRMS did not show this benefit.2
Recent data from MRS9,10⇓ and pathologic studies11 have indicated that axonal dysfunction and damage are more common in MS than previously believed, and occur early in the disease course.12 These findings support prompt initiation of therapy to delay or limit the development of a more disabling phase of MS.
A recently published trial of a single dose level of IFNβ-1b in SPMS showed significant reduction (by 1 point) in time to sustained progression on the Expanded Disability Status Scale (EDSS), with beneficial effects on exacerbation rate and MRI variables.5 The IFNβ-1b study was terminated after a 2-year interim analysis demonstrated efficacy. However, preliminary results of a second study of the same product in SPMS showed no disability benefit.13 This report of 3 years of therapy provides the first information on the effects of two different doses of IFNβ-1a in SPMS.
Patients and methods.
Design.
This randomized, double-blind, placebo-controlled study enrolled 618 patients in 22 centers in Europe, Canada, and Australia.
An independent monitoring panel consisting of a statistician, an interferon biologist, and two MS specialists who were otherwise uninvolved in the study reviewed safety data and supervised analyses. MRI analysis was performed centrally by the University of British Columbia MS/MRI Analysis Group, Vancouver.
Patient population.
Eligible patients had clinically definite SPMS, defined as progressive deterioration of disability for at least 6 months with an increase of at least 1 EDSS point over the last 2 years (or 0.5 point between EDSS score of 6.0 and 6.5), with or without superimposed exacerbations, following an initial RR course. At study entry, patients were between 18 and 55 years old, with EDSS scores from 3.0 to 6.5 and pyramidal functional score of at least 2. Exclusion criteria included immunosuppressive or immunomodulatory treatments during the previous 3 to 12 months depending on the drug, prior treatment with interferon or total lymphoid irradiation, corticosteroid use or a disease exacerbation in the previous 8 weeks, severe concurrent illness, and pregnancy or lactation. Potentially fertile women were required to use effective contraception.
Treatments and randomization.
After screening, patients were randomized to receive 22 or 44 μg IFNβ-1a (Rebif; Serono International, Geneva, Switzerland) or matching placebo subcutaneously three times a week for 3 years. Treatment was assigned using a computer-generated randomization list provided by Serono, stratified by center; treatments were equally allocated with a block size of six. The block size was not revealed to the investigators. Study medications were supplied as solutions. To allow patients to adjust to the medication, the volume administered was increased gradually over the first 4 to 8 weeks. In case of toxicity, the dose could be reduced or treatment interrupted according to guidelines in the protocol.
Treatment was discontinued in case of severe toxicity or pregnancy, and could be discontinued for protocol violations, adverse events, or noncompliance. Patients who discontinued treatment and who consented were observed until the end of the planned 3-year period. Patients who withdrew were not replaced.
Acetaminophen was recommended for prevention or relief of IFN side effects such as fever and influenza-like symptoms. Steroids were to be given only for acute exacerbations, using a standard regimen of 1.0 g/d IV methylprednisolone for 3 consecutive days.
Blinding.
Solutions of IFNβ-1a and placebo were physically indistinguishable, and packaging and labeling were prepared to preserve blinding. The manufacturer labeled containers of study medication with patient identification numbers based on the randomization list, and patients received the medication labeled with their numbers. Treatment assignments were provided to investigators in sealed envelopes for emergency use: two envelopes were opened at the request of patients who withdrew due to adverse events.
Because IFN side effects are well recognized, a treating physician supervised drug administration, monitored safety, and managed adverse events, and a separate evaluating physician conducted neurologic assessments and followed-up exacerbations. Patients were instructed to cover injection sites and to discuss only neurologic matters during neurologic evaluations. Clinical and neurologic data were recorded in separate binders. At the end of the study a questionnaire was administered to patients and evaluating physicians to determine the success of blinding.
Study assessments.
Patients underwent clinical evaluations at 3-month intervals, or more frequently in case of exacerbations or high-frequency MRI scanning. Cranial MRI scans were performed at baseline and every 6 months; subsets of patients underwent monthly scans. MRI methods and results are included in this issue of Neurology.14
Treatment effect on disability was assessed in terms of time to confirmed progression, defined as increase from baseline by at least 1 EDSS point (or 0.5 point if baseline EDSS was ≥5.5), confirmed 3 months later with no intervening score lower than the minimum required level. The EDSS has known interrater variability: accordingly, each patient was to be examined by the same evaluating physician throughout the study, and guidelines for evaluation of the EDSS and other neurologic rating scales were provided in the protocol. Confirmation of EDSS change at a second visit was required to reduce the chance of intrarater and intrapatient variability producing nonsustained EDSS changes. Scoring of the EDSS was based on Kurtzke functional scores up to and including an EDSS score 5.0. Above this level, ambulation took precedence.
During the trial, an exacerbation was defined as the appearance of a new symptom or worsening of an old symptom attributable to MS, accompanied by an appropriate new neurologic abnormality or focal neurologic dysfunction lasting at least 24 hours in the absence of fever and preceded by stability or improvement for at least 30 days.15 Evaluating physicians described each exacerbation and its worst observed severity in terms of change in Scripps Neurologic Rating Scale (SNRS) score. A mild relapse was defined by SNRS change of 0 to 7 points, a moderate relapse by change of 8 to 15 points, and a severe relapse by change of more than 15 points. Where SNRS evaluation was not possible, severity was described in terms of effects on the activities of daily living (ADL): mild attacks did not interfere with ADL, moderate attacks impaired but did not prevent ADL, and severe attacks required hospitalization. Adverse events and changes in concomitant medications were followed throughout the study, and clinical laboratory evaluation was performed at the 3-month evaluation visits or as needed.
Patients were tested for neutralizing antibodies to IFNβ at 6-month intervals using a viral cytopathic bioassay (data on file); positivity was defined as a titer of at least 20 neutralizing units per milliliter.
Because of concerns about a possible association of IFNβ with depression, psychological instruments were administered to the 337 patients in English-speaking centers (the General Health Questionnaire,16 the Center for Epidemiologic Studies Depression Mood Scale,17 and the Beck Hopelessness Scale18).
Analyses and outcomes.
The primary efficacy outcome was time to confirmed progression in disability, as defined above. This outcome included all study time including month 36; for patients progressing for the first time at 36 months, confirmation used the EDSS score from the month 39 visit (during the study’s extension phase). Time to confirmed progression was analyzed using the Cox proportional hazards (PH) model. Hazard ratios (HR, with 95% CI and p values) smaller than 1.0 indicate a reduced risk of the event (progression) compared with a reference group (placebo). The assumption of proportional hazards was assessed by the inclusion of a time-dependent covariate. Time-to-progression curves were constructed using the Kaplan–Meier method. Because the primary comparison was between the 44-μg dose of IFNβ-1a and placebo, no adjustment was made for multiple comparisons. Differences between the low-dose and placebo groups were examined for information and assessment of dose effect.
A significant treatment-by-sex interaction in the primary outcome prompted further analyses. A list of potentially important covariates was defined by an investigator liaison committee after study unblinding but before analyses were undertaken. The association of these factors with time to progression was analyzed using the Cox PH model in conjunction with the likelihood ratio test.
Secondary clinical outcomes (with corresponding regression methods) included proportion of patients progressing (logistic), exacerbation count (Poisson), time to first exacerbation (Cox PH), time between first and second exacerbations (Cox PH), number of moderate and severe exacerbations (Poisson), number of steroid courses for MS (Poisson), number of hospitalizations for MS (Poisson), and Integrated Disability Status Score (IDSS, defined by area under an EDSS time-curve adjusted for baseline19; analysis of variance on ranks). All models included center as a covariate, and all Poisson and logistic regression models included time on study as an offset. The likelihood ratio test was used to evaluate the significance of interaction terms incorporated in models. Proportions of patients experiencing adverse events were compared across treatments using either χ2 or Fisher’s exact tests, as appropriate.
Composite measures permit summarization of results from multiple outcomes in complex diseases like MS, allowing assessment of overall treatment efficacy. In this study, a prospectively defined composite score was determined using rank values for five major outcomes: time to progression, exacerbation rate, MRI lesion burden, MRI T2 activity, and IDSS.20 Log rank scores were used for time to progression in calculating the composite score.
Serono Biometrics (Geneva) performed statistical analysis, using SAS (SAS Institute, Cary, NC) and S-Plus software (Math Soft, Cambridge, MA). All p values were derived from two-sided significance tests. Analyses were performed on an intent-to-treat basis, including all randomly assigned patients. Patients who dropped out were considered as censored at the time of dropout for time-to-event outcomes, and their time on study was used for event count analyses. No imputation was used for any outcome.
Two subgroup analyses were undertaken at the request of regulatory authorities: patients were grouped by sex and by presence or absence of relapses in the 2 years preceding the study. In the subgroup analysis by presence or absence of before-study relapses, active treatment groups were combined to maintain statistical power. It should be noted that p values obtained from subgroup and exploratory analyses do not have their nominal type I error value and should be interpreted with caution.
Sample size and interim analysis.
Natural history studies suggested that 60% of placebo-treated patients would progress by 1 EDSS point over 3 years. On the assumption that 40% of patients treated with high-dose IFN would progress, it was determined that for 80% power and a significance level of 0.05 (allowing for one interim analysis), 98 evaluable patients per group were needed to detect significant difference between placebo and high-dose IFN in the proportion of patients progressing (log-rank test). Based on an estimated 20% dropout rate, 118 patients per group would be enrolled. Centers recruited much faster than expected, leading to over-enrollment. The decision to continue accrual was not based on interim results; the interim analysis was not performed until accrual had been completed. The increased number of patients allowed for 80% power to determine a smaller reduction (from 60% for placebo to 46% for high-dose IFN) in the proportion of patients who would progress by 1 EDSS point over 3 years.
An interim analysis was performed when 300 patients had completed 18 months of treatment to allow for early stopping of the study or termination of the placebo arm. Only the primary outcome was analyzed. The study was to be stopped for significance in favor of placebo (p = 0.005, two-sided). If the treatment were significantly better (p = 0.005) than placebo, patients receiving placebo would be randomly reassigned to one of the two IFN doses. Results were presented to the independent monitoring panel, which decided that the study should continue to its planned termination. The final analysis of the primary outcome was performed with a nominal significance level of 0.048.
Ethics.
The study was conducted in accordance with the Declaration of Helsinki. Consent was obtained from the ethics committees of all participating institutions before study initiation. Written informed consent was obtained from all patients before beginning prestudy assessments.
Results.
Study population and patient disposition.
Of 618 patients enrolled, 506 (82%) completed 3 years of treatment, and an additional 65 who stopped therapy were followed for the remainder of the 3 years, providing full data for 92.4% of patients (figure 1). One hundred twelve patients discontinued therapy prematurely (18%). Reasons for discontinuation included adverse events (38 patients: 5 placebo, 15 low-dose and 18 high-dose IFN), disease progression (20 patients: nine placebo, seven low-dose and four high-dose IFN), death (four patients: one placebo, one low-dose and two high-dose IFN; a second placebo-treated patient died during follow-up), protocol deviation (four patients: one low-dose and three high-dose IFN) and patient decision (44 patients: 17 placebo, 12 low-dose and 15 high-dose IFN). Forty-seven (47) patients were lost to follow-up, 14 each in the high-dose and low-dose IFN groups and 19 in the placebo group. Baseline characteristics were similar among treatment groups (table 1).
Figure 1. Patient enrollment and disposition.
Baseline characteristics by treatment group
Disability progression (primary endpoint).
The difference in time to sustained progression in disability between patients receiving 44 μg IFNβ-1a and placebo was not significant (p = 0.146; figure 2 andtable 2). A transient effect of the high dose was observed, lasting approximately 1 year. No significant difference was noted between patients receiving low-dose IFN and placebo (HR = 0.88; p = 0.305). The assumption of PH in this model was examined by incorporation of a time-varying component in the Cox PH model, and was not found to be violated. Subgroup analyses required by regulatory authorities highlighted a difference in treatment effect between male and female patients (figure 3). Detection of a treatment-by-sex interaction (p = 0.035) confirmed this difference, which had not been anticipated. Women, comprising 62% of the study population, showed delay in progression compared with placebo at both doses (p = 0.006 for 44 μg and p = 0.038 for 22 μg), whereas men did not. Unexpectedly, a difference was seen in placebo group behavior as well: male patients had consistently better outcomes than female patients receiving placebo, although only time to progression differed significantly (HR = 0.64; 95% CI, 0.43 to 0.95; p = 0.016). In an attempt to explain these differences, baseline characteristics (age, EDSS, before-study EDSS change, duration of MS or SPMS, SNRS, Ambulation Index, before-study relapse rate, and baseline MRI lesion burden) were examined with respect to sex. No significant imbalances were found.
Figure 2. Kaplan–Meier curves for time to confirmed Expanded Disability Status Scale (EDSS) progression for all patients.
Hazard ratios for time to progression based on protocol-defined planned analysis (adjusted for center), covariate-adjusted analysis (center, SNRS, duration of SPMS disease, rate of progression before study), and subgroup analyses based on presence or absence of prestudy relapses and sex
Figure 3. Kaplan–Meier curves for time to confirmed Expanded Disability Status Scale (EDSS) progression for male and female patients.
In a further attempt to explain the treatment-by-sex interaction for the primary outcome, the investigator liaison committee was asked to identify other factors that might directly influence time to progression. Age, before-study relapse rate and change in EDSS, disease duration, duration of the SP phase, body mass index, baseline EDSS, SNRS, Ambulation Index, and pyramidal scores were prospectively identified. Their association with time to progression was analyzed using the Cox PH model, and three factors were found to be joint significant predictors: baseline SNRS (low score unfavorable), duration of SP disease (long duration unfavorable), and rate of EDSS change preceding the study (rapid change favorable). Duration of SP disease, prestudy exacerbation rates, and prestudy changes in EDSS were determined from medical records or from retrospective review. Although these numbers may be somewhat inaccurate, there is no reason to believe that this inaccuracy would affect treatment comparisons. When these three factors were added to the Cox PH model, a stronger treatment effect on time to progression was detected (for 44 μg IFNβ-1a compared with placebo, HR = 0.78; 95% CI, 0.60 to 1.00; p = 0.046). However, adjustment for these covariates failed to lessen the treatment-by-sex interaction.
Patients with SPMS may or may not continue to experience relapses. Therefore, the main study outcome measures were analyzed separately for patients who did (n = 293) and did not report (n = 325) relapses during the 2 years preceding the study. The subgroup with relapses before the study was significantly younger, had a shorter duration of disease, and had deteriorated slightly faster than the nonrelapsing group at baseline (data not shown). Patients with before-study relapses were more likely to benefit from therapy with respect to time to confirmed progression (figure 4 and table 2). It should be noted, however, that the interaction between treatment and the presence or absence of before-study relapses was not significant (p = 0.289, Cox PH). OR for the proportion of patients progressing in the combined IFNβ-1a groups compared with placebo was 0.52 for patients with (95% CI, 0.29 to 0.93; p = 0.027) and 1.07 for patients without prestudy relapses (95% CI: 0.64–1.78, p = 0.802). HR for time to first progression for IFN-treated relapsing patients was 0.74 compared with placebo (p = 0.055), whereas HR for IFN-treated patients without relapse was 1.01 compared with placebo (p = 0.934). The HR for time to progression for patients treated with 44 μg IFNβ-1a compared with placebo was 0.76 for relapsing (p = 0.14) and 0.93 for nonrelapsing patients (p = 0.69).
Figure 4. Kaplan–Meier curves for time to confirmed Expanded Disability Status Scale (EDSS) progression for patients with and without relapses during the 2 years before the study.
Exacerbations.
A highly significant benefit was seen on exacerbation rate for both doses. Treatment benefit was also observed with respect to time to first exacerbation, time between first and second exacerbations, numbers of moderate and severe exacerbations, steroid use, and hospitalization (table 3). Relapse rates in men were 0.49 per year for 44 μg IFN, 0.43 for 22 μg IFN, and 0.66 for placebo (p = 0.007 for 44 μg versus placebo and p = 0.025 for 22 μg versus placebo). In women, corresponding rates were 0.51 for 44 μg IFN, 0.53 for 22 μg IFN, and 0.75 for placebo (p = 0.009 for 44 μg versus placebo and p = 0.012 for 22 μg versus placebo). There was no treatment-by-sex interaction for this measure (p = 0.888). As with EDSS progression, treatment effect was greater in the subgroup with than the subgroup without prestudy relapse (treatment by relapse history interaction: p = 0.004). With treatment, relapse rates among patients with prestudy relapse were 0.57 per year for the 22 μg group and 0.67 for 44 μg IFN groups versus 1.08 per year for placebo (p < 0.001 for both doses compared with placebo). Corresponding rates among nonrelapsing patients were 0.43 and 0.36 per year (p value for both nonsignificant compared with the placebo rate of 0.39 per year).
Exacerbation-related outcomes (all patients)
Composite score.
The composite score showed a marked treatment benefit for each dose (p < 0.001). A treatment-by-sex interaction (p = 0.014) was observed for this outcome.
Safety.
Adverse event data are presented in table 4. IFNβ-1a was well tolerated. The most frequent adverse events clearly associated with treatment were application site disorders, flulike symptoms, and lymphopenia. Treatment was interrupted or the dose reduced because of adverse events in 22 patients receiving placebo, 40 receiving 22 μg IFN, and 66 receiving 44 μg IFN, of whom only three patients receiving placebo, eight receiving 22 μg, and seven receiving 44 μg permanently discontinued treatment. Liver function abnormalities were more common with active treatment; however, these were generally mild or moderate and either resolved with treatment interruption or were not severe enough to warrant intervention. Injection site necrosis was reported nine times in seven patients receiving 22 μg IFN and 22 times in 18 patients receiving 44 μg IFN, yielding rates of one episode of necrosis per 9,600 injections for 22 μg and one per 3,800 injections for 44 μg. One episode of necrosis required surgical intervention, and three patients stopped treatment because of this event. Five patients died: one patient receiving placebo had a presumed subarachnoid hemorrhage, one patient in each treatment group (including placebo) committed suicide, and one patient receiving 44 μg had an intracerebral hemorrhage. Three patients receiving placebo, three receiving low-dose IFN, and two receiving high-dose IFN attempted suicide. Depression was reported in 29% of patients receiving placebo, 32% receiving low-dose IFN, and 35% receiving high-dose IFN. One patient receiving placebo, five receiving 22 μg, and six receiving 44 μg discontinued therapy due to depression (overall treatment difference: p = 0.038, Fisher’s exact test). There were, however, no significant differences between groups in results of the General Health Questionnaire, Center for Epidemiologic Studies Depression scale or Beck Hopelessness Scale, either at baseline or during the study.
Reported adverse events of interest with interferon therapy (all values are percentages)
Results of neutralizing antibody (NAb) testing are shown in table 5. The rate of NAb development was higher in the group receiving 22 μg IFN than those receiving 44 μg. The majority of NAb developed during the first 18 months of treatment. Development of NAb did not affect the primary outcome, but there was a suggestion of reduced effect on relapses for NAb-positive patients in the 44 μg group, in which the difference between NAb-positive patients and those receiving placebo was no longer significant (patients with at least one positive titer > 20 NU/L were considered NAb-positive). However, the numbers of patients involved were small.
Neutralizing antibody formation
Blinding.
Evaluating physicians guessed treatment assignments correctly for 29% of patients on active treatment and 26% of patients on placebo. They were incorrect for 16% of patients on active treatment and 18% of patients on placebo, and gave no opinion for the rest. Fifty-nine percent of patients on active treatment and 52% on placebo guessed their treatment correctly. In the high-dose group 64% guessed correctly, whereas in the low-dose group 54% guessed correctly. Patients were incorrect in 21% of active and 27% of placebo cases, and gave no opinion in 20% of cases in the active group and 21% in the placebo group.
Discussion.
This study provides important new information about IFN use in MS, and in particular helps to clarify the role of IFN in SPMS. The analysis of the primary outcome, time to sustained progression, showed minimal evidence of treatment effect (HR = 0.83) which did not reach statistical significance. Despite the lack of significant benefit on the primary outcome, all secondary outcomes showed a response to treatment, and the composite score provided evidence of treatment benefit in this patient group when multiple endpoints were considered. The lack of benefit on disability progression was disappointing given the clinical importance of this outcome, and initially appeared to contradict the findings of the European study of IFNβ-1b in SPMS. However, the results are consistent with those of a more recent study using the same medication (see below).
Clearly IFN does not provide the same spectrum of benefits in SPMS as in RRMS. However, absence of efficacy in this population would be an erroneous conclusion.
Results of this study were affected by an unexpected treatment-by-sex interaction. Additional analyses prompted by this finding and by requests from health authorities suggested that different subgroups of SPMS patients may experience different responses to treatment.
Analysis by sex showed a significant benefit on disability progression with both doses in women, but no benefit in men for this outcome. IFN therapy showed efficacy on all secondary measures in both sexes. There are several possible explanations for this finding. First, women may indeed respond better to IFN than men, but this finding would be new and there is no supporting evidence in the literature. However, additional findings from the European study of IFNβ-1b in patients with SPMS allude to a slightly worse treatment outcome in men.21 Second, baseline characteristics could have been unbalanced, but close examination showed that there were no significant differences between groups. Third, variability in the assessment of disability may have affected the results. However, the finding of similar treatment-by-sex interactions for MRI measures14 argues against this explanation. Finally, sex may be a proxy factor for one or more unmeasured clinical features. As a definitive explanation for this interaction is not evident after extensive review of the data, the effect of chance must be considered. The combination of an unexpected difference in time to progression between male and female patients receiving placebo and a difference in male and female response to IFN in SPMS could produce the interaction found. It will be important to evaluate this finding prospectively in subsequent studies.
An additional post hoc analysis requested by health authorities examined the impact of relapses in the 2 years before study entry on subsequent relapses, disease progression, and response to therapy. Patients who had reported at least one relapse during this period had a generally less-advanced disease course. In the combined treatment groups, those with at least one prestudy relapse showed better treatment response compared with placebo than those without prestudy relapses, with respect to both progression and relapse rate. These data suggest that IFN therapy has clinical benefit in SPMS, predominantly affecting relapses, but has only a modest effect on disability in those patients still experiencing relapses during the few years before treatment initiation. This finding suggests that treatment early in the disease course, when relapses are usually more frequent, may be more effective than later therapy.
The results on disability in this study differ from those of the European trial of IFNβ-1b in SPMS, in which there was a significant increase in time to sustained progression in disability with IFNβ-1b compared with placebo.5 Neither the level of disability nor the presence of relapses was reported to affect outcome. However, the recently reported North American study of IFNβ-1b in SPMS13 found no effect on time to confirmed progression (p = 0.71), although as in the present study (Secondary Progressive Efficacy Clinical Trial of Recombinant Interferon-beta-1a in MS [SPECTRIMS]) significant benefit was shown on all other outcomes. These discrepant results require consideration.
To put the findings of these three studies into perspective, it is important to consider possible explanations for the apparent difference in the main outcome. These differences include a shorter observation period in the European IFNβ-1b study due to early termination (2 years compared with 3), shorter duration of the prerandomization SP phase in the European IFNβ-1b study (2.2 years compared with 4.0 in the other trials, although definitions may vary), a higher rate of treatment discontinuation in the European IFNβ-1b study despite the shorter observation period, and most importantly, a higher proportion of patients with prestudy relapses in the European IFNβ-1b study (70% versus 48% in SPECTRIMS and 45% in the North American study). As discussed previously, patients in SPECTRIMS who experienced prestudy relapses responded best to IFN according to all outcome measures, including the primary outcome; this would partially explain the difference in results between the European IFNβ-1b study and the SPECTRIMS and North American studies.
The CI of the OR for the proportion of treated patients who progressed on treatment compared with placebo (adjusted for center and baseline EDSS) from SPECTRIMS and the original report of the European IFNβ-1b study overlap: 0.74 (95% CI, 0.46 to 1.20) for SPECTRIMS and 0.65 (95% CI, 0.52 to 0.83) for the European study. The differences in patient characteristics and study duration, coupled with review of the OR for progression, suggest that any difference in results between these two SPMS studies is due to patient variability and differences in the populations studied and the duration of observation, or to chance. The data further support the hypothesis that IFNβ has beneficial effects predominantly on the presumed inflammatory components of MS (such as relapses and MRI activity), with at best a modest effect on continuous progression in disability. The OR for disability progression in the Prevention of Relapses and Disability by Interferon-beta-1a Subcutaneously in Multiple Sclerosis (PRISMS) study of RRMS was 0.58 (95% CI, 0.35 to 0.96), which is lower than that in either of the SPMS trials. As one moves from pure RRMS into more progressive SPMS, less effect is seen on disability progression: this finding argues further in favor of treatment for MS patients still experiencing relapses.
As reported in the companion paper to this article,14 both doses significantly reduced MRI T2 activity, lesion burden accumulation, and combined unique lesion activity.
The lack of significant impact on disability despite dramatic effects on relapses supports a recent revival of interest in the contribution of axonal damage to clinical disability. Relapses and MRI measures are thought to reflect mostly inflammatory changes, whereas persistent disability probably also reflects axonal loss. The observed difference in efficacy between relapse and disability outcomes suggests that the axonal damage believed to underlie progressive accumulation of disability may result from other pathogenetic mechanisms besides the inflammation associated with relapsing disease,11 or that there may be a time lag between inflammation and consequent axonal destruction. The putative additional mechanisms contributing to axonal dysfunction and destruction remain unknown, but may involve direct injury, bystander effects, or a loss of trophic factors.
The adverse event profile of IFNβ-1a in the present trial was similar to that in PRISMS.4 Overall, treatment was well tolerated: 82% of patients receiving low-dose and 84% of patients receiving high-dose IFN remained on treatment at 36 months, compared with 79% of patients receiving placebo.
In conclusion, this study failed to show significant impact on time to sustained progression of disability. However, the trial confirmed that IFNβ-1a is as well tolerated and efficacious with respect to exacerbation-related and MRI outcomes in SPMS as it is in RRMS. Subgroup analyses suggest that the maximal benefit of treatment in SPMS is seen in women and in patients still experiencing relapses when treatment is started, and it would appear reasonable to consider treatment with IFNβ-1a for such patients. Further detailed data regarding the use of IFNβ in SPMS are eagerly awaited to clarify the efficacy of this therapy in more advanced MS.
Appendix
The SPECTRIMS investigator liaison committee consists of R.A.C. Hughes (chair), O. Hommes, D. Paty, and M. Sandberg-Wollheim. The writing committee consists of R.A.C. Hughes (chair), O. Hommes, D. Li, D. Paty, M. Sandberg-Wollheim, P. Soelberg Sørensen, and G. Francis. Members of these committees have received honoraria for lecturing, travel expenses for attending meetings, and financial support for their departments from Serono.
The independent monitoring panel consisted of J. Noseworthy (chair) and P. O’Brien, Mayo Clinic, Rochester, MN; E. Borden, Cleveland Clinic, Cleveland, OH; and J. Wolinsky, University of Texas, Houston, TX.
Acknowledgments
Funded by Serono International SA, Geneva, Switzerland.
Acknowledgment
In addition to the many site personnel instrumental in the successful conduct of the study, the authors thank A. Abdul-Ahad, J. Alsop, K. Diab, M. Dubourgeat, F. Dupont, R. Furcha, S. Gerin, A. Jaccottet, T. Mattioni, M. Olson, M.-O. Pernin, M. Stam Moraga, and B. Vayssier-Lemaire of Serono for their contributions to the study, A. Dubois (Serono medical writer) for manuscript assistance, and the patients who participated in this clinical trial. Study monitoring was performed by Serono Clinical Research Associates.
Footnotes
-
See also page 1505
-
*A complete list of the participants in the Secondary Progressive Efficacy Clinical Trial of Recombinant Interferon-beta-1a in MS (SPECTRIMS) Study Group in provided in the Appendix on page 1503.
- Received June 21, 2000.
- Accepted February 15, 2001.
References
Letters: Rapid online correspondence
REQUIREMENTS
You must ensure that your Disclosures have been updated within the previous six months. Please go to our Submission Site to add or update your Disclosure information.
Your co-authors must send a completed Publishing Agreement Form to Neurology Staff (not necessary for the lead/corresponding author as the form below will suffice) before you upload your comment.
If you are responding to a comment that was written about an article you originally authored:
You (and co-authors) do not need to fill out forms or check disclosures as author forms are still valid
and apply to letter.
Submission specifications:
- Submissions must be < 200 words with < 5 references. Reference 1 must be the article on which you are commenting.
- Submissions should not have more than 5 authors. (Exception: original author replies can include all original authors of the article)
- Submit only on articles published within 6 months of issue date.
- Do not be redundant. Read any comments already posted on the article prior to submission.
- Submitted comments are subject to editing and editor review prior to posting.
You May Also be Interested in
Hastening the Diagnosis of Amyotrophic Lateral Sclerosis
Dr. Brian Callaghan and Dr. Kellen Quigg