Abstract
Background: Patients with primary progressive MS have atypical clinical and MRI characteristics and have been excluded from most therapeutic trials. The authors report a randomized, controlled trial restricted to primary progressive MS.
Methods: Fifty subjects were randomized to weekly IM interferon β-1a 30 μg, 60 μg, or placebo for 2 years. The primary endpoint was time to sustained progression in disability. Secondary outcomes included the timed 10-meter walk, nine-hole peg test, and on MRI, T2 and T1 brain lesion loads and brain and spinal cord atrophy.
Results: The 30-μg dose of interferon β-1a was well tolerated, but the 60-μg dose caused severe flulike reactions and raised liver enzymes. No treatment effect was seen on the primary endpoint. Subjects on interferon β-1a 30 μg had a lower rate of accumulation of T2 lesion load than controls (p = 0.025); subjects on 60 μg had a greater rate of ventricular enlargement than controls (p = 0.025).
Conclusions: This study has demonstrated that interferon β-1a 30 μg was well tolerated, identified useful outcome measures, but showed no efficacy on the primary outcome measure or on most of the secondary outcome measures.
Ten to 15% of patients with MS have primary progressive MS (PPMS), which is characterized by the progressive accumulation of neurologic deficit from disease onset without relapse or remission.1 Patients with PPMS have atypical clinical, MRI, and pathologic characteristics1-6⇓⇓⇓⇓⇓ that create difficulties for patient selection and therapeutic monitoring in treatment trials,7 resulting in a paucity of trials in this population.
β-Interferon has been extensively investigated in the other subgroups of MS. It has a partial effect on relapse frequency and severity in relapsing-remitting8-10⇓⇓ and secondary progressive MS.11,12⇓ It reduces the rate of development of clinically definite MS in patients with clinically isolated syndromes suggestive of MS.13,14⇓ The effect of β-interferon on disease progression is less clear. Of two published studies in secondary progressive MS, one reports a delay in the accumulation of disability,11 whereas the other finds no significant clinical effect.12 Preliminary reports of two further studies also present conflicting results.15,16⇓ All studies have demonstrated a positive effect of β-interferon on MRI markers of inflammation. Although there is less inflammation in PPMS compared to other subgroups, inflammation clearly occurs; therefore, a trial of β-interferon in this group was justified.
This article presents the first randomized, controlled trial of β-interferon in PPMS. This is an investigator-led, exploratory trial and has been carried out at a single center with a small sample size. The three main objectives of the trial were as follows: 1) to investigate safety and tolerance, 2) to identify potentially useful outcome measures, and 3) to look for a hint of efficacy.
Subjects and methods.
Subjects.
Fifty subjects with PPMS were enrolled at a single clinical center (The National Hospital, London, UK). All subjects gave informed consent and the study was approved by the local ethics committee. At recruitment, specific diagnostic criteria specific for PPMS had not been published.17 Diagnosis was ascertained on the basis of a progressive history without relapse or remission, at least two typical lesions on MRI brain or spinal cord, and oligoclonal bands in the CSF not present in parallel serum or abnormal visual evoked potentials. Particular emphasis was placed on the CSF findings; oligoclonal bands were positive in the CSF and not serum in all of the 39 patients for whom results were available. Inclusion criteria were as follows: 1) PPMS of at least 2 years’ duration, 2) aged 18 to 60 years, and 3) Expanded Disability Status Scale (EDSS)18 score of 2.0 to 7.0 inclusive. Exclusion criteria were as follows: 1) interferon, immunosuppressant, or chronic steroid therapy within the previous 3 months, 2) pregnancy or lactation, 3) seizure within the previous 3 months, and 4) a history of severe depression.
Study design and treatment.
This was a double-blind, placebo-controlled trial of two doses of interferon β-1a (Avonex, Biogen, Cambridge, MA). Individual subjects were randomized by the block method to IM interferon β-1a 30 μg (IFN30), 60 μg (IFN60), or placebo once a week, with 15 subjects in each active treatment arm and 20 subjects in the placebo arm. The study period was 24 months long. Subjects and study personnel were blinded to treatment status. EDSS assessments were performed by an independent evaluating physician blinded to all clinical information. A questionnaire to test blinding was given to all subjects and the treating physician at the end of the study.
Nonsteroidal anti-inflammatory drugs or paracetamol were recommended for prophylaxis of interferon-associated flulike reactions. In the event of study drug intolerance there was an option to halve the dose.
Efficacy evaluation.
Clinical efficacy was evaluated at baseline and every 3 months for 2 years. The primary clinical endpoint was time to sustained progression in disability defined as a ≥1.0-point increase in EDSS score for subjects with a baseline EDSS score ≤5.0, or a ≥0.5-point increase for subjects with a baseline ≥5.5. Progression was considered sustained if documented at two consecutive visits 3 months apart; the time of the first visit was recorded as the time to progression. Secondary clinical outcome measures included the timed 10-meter walk (TTMW) and nine-hole peg test (NHPT). The time to walk 10 meters, with assistance as required, was recorded on one attempt. The time to perform the NHPT was recorded on two consecutive trials for both right and left arms. If a subject was unable to complete the TTMW or NHPT, a time of 300 seconds was recorded.
Efficacy was also evaluated by MRI. Secondary outcome measures included the following: 1) T2-weighted brain lesion load, 2) new T2 brain lesions, 3) new T2 spinal cord lesions, 4) T1-weighted brain lesion load, 5) new T1 brain lesions, 6) spinal cord cross-sectional area, 7) whole brain volume, and 8) ventricular volume. Tertiary MRI outcome measures, which will be reported subsequently, included magnetization transfer ratio (MTR) and 1H MRS. T2- and T1-weighted imaging were performed at baseline and at 6, 12, 18, and 24 months; spinal cord area, brain volumes, MTR, and MRS were performed at baseline and at 12 and 24 months.
Safety evaluation.
Adverse events, physical examination findings, and hematologic and biochemical parameters were monitored throughout the study. An interim safety review by an independent assessor was performed at the midpoint of the trial.
Neutralizing antibodies.
Serum was collected for interferon β-1a neutralizing antibody assay at baseline and every 3 months. Results were not revealed to study personnel before study completion. An antibody titer >1/10 was reported as positive.
MRI and image analysis.
MRI was carried out on a 1.5-T Signa Horizon Echospeed system (General Electric, Milwaukee, WI). Brain imaging was performed with a standard quadrature head coil and spinal cord imaging with a phased array spinal coil. All image analysis was blinded to subject identity.
T2-weighted brain imaging.
Forty-six contiguous, 3-mm slices were acquired using a dual-echo fast spin-echo sequence (echo time [TE] 15/90 ms, repetition time [TR] 3200 ms, echo train length 8, field of view [FOV] 24 × 24 cm, matrix 256 × 256, 1 signal average). All images were copied to film (hard copy) and electronic disc. Identification and marking of every brain lesion on hard-copy proton density–weighted images, with cross-reference to the T2-weighted images, was performed by two observers, with the same observer marking all scans for a given subject. New lesions were identified and the number documented. The electronic copy was displayed using DispImage software (D.L. Plummer, University College London, UK). All marked lesions were contoured by a single rater using a semiautomated local thresholding technique plus manual editing, with reference to published guidelines.19 Individual lesion volumes were calculated as the lesion area multiplied by the slice thickness, and the total lesion volume was the sum of all the individual lesion volumes.
T1-weighted brain imaging.
Forty-six contiguous, 3-mm slices were acquired using a spin-echo sequence (TE 17 ms, TR 600 ms, FOV 24 × 24 cm, matrix 256 × 256, 1 signal average). Imaging analysis was carried out as above except that the lesions were identified and marked by a single observer.
T2-weighted spinal cord imaging.
Nine contiguous, 3-mm sagittal slices of the whole spinal cord were acquired using a dual-echo fast spin-echo sequence (TE 45/90 ms, TR 2500 ms, echo train length 16, FOV 48 × 48 ms, matrix 512 × 512, 2 signal averages) and copied to film. Every spinal cord lesion was marked at baseline by a single observer and the number of lesions documented. New lesions on the corresponding serial scans were identified and the number documented.
Spinal cord cross-sectional area.
A volume-acquired, inversion-prepared, fast spoiled-gradient echo was performed in the cervical spinal cord (60 1-mm slices acquired sagittally, TE 4.2 ms, inversion time [TI] 450 ms, TR 15.6 ms, flip angle 20°, FOV 25 × 25 cm, matrix 256 × 256, 1 signal average). Five contiguous, 3-mm axial slices from the caudal landmark of the C2/3 intervertebral disc were reformatted from the volume data set and a coil radiofrequency uniformity correction applied.20 Cord area was measured using a semiautomated method previously described.21
Brain/ventricular volume.
A whole-brain, volume-acquired, inversion-prepared, fast spoiled-gradient echo was performed (128 1.5-mm slices acquired coronally, TE 4.2 ms, TI 450 ms, TR 15.6 ms, flip angle 20°, FOV 24 × 18 cm, matrix 256 × 192, 1 signal average). Image analysis was carried out using a semiautomated method previously described.22 Serial scans were accurately registered to the baseline scan. The change in brain volume was measured from the registered scans by integrating the shifts in brain–CSF boundaries. The resulting brain boundary shift integral (BBSI) is a direct measure of change in cerebral volume (a positive value indicating atrophy). Ventricular volume, the absolute volume of the lateral ventricles including the temporal horns, was also measured using the semiautomated contouring method on the baseline and registered serial images.
Because of the sensitivity of atrophy measures to movement artifact, all spinal cord and brain volume images were systematically checked by a blinded observer and rejected if there was significant movement artifact.
MRI reproducibility.
Intraobserver variation for same scan analysis was assessed for T2 and T1 lesion volumes (n = 10), spinal cord area (n = 10), BBSI (n = 6), and ventricular volume (n = 10) using the coefficient of variation.
Statistical analysis.
Statistical analysis was carried out on an intention-to-treat basis. Because this was an exploratory trial, primary and secondary statistical methods were included. The primary statistical method for the primary clinical endpoint was the Mantel Cox log rank test and the secondary method was piecewise exponential modelling.
For the secondary clinical and most of the MRI outcomes, the primary statistical method was nonparametric (rank) analysis of covariance and the secondary method was mixed-model regression analysis; these analyses adjusted for any differences at baseline. The raw data for these analyses were the absolute measurements recorded, but the mixed-model regression analysis was performed using a log (y + constant) transformation. For the NHPT, the times of both trials for each arm were included in the analyses. The nonparametric analysis of covariance results are given for month 24; if a treatment difference was identified, the directionality of the effect was determined by using the group mean residuals and by inspection of the data. The mixed-model regression analysis results are restricted to the rates of progression from the first to the last treatment observation and are given in terms of p values because the log space regression coefficient estimates have little clinical meaning.
For new lesions the cumulative number was calculated and categorical data analysis using the mean score statistic was performed. The BBSI measure of brain atrophy, which provided a bivariate change score, was analyzed with multivariate analysis of variance using the Wilks’ lambda statistic.
In addition to individual treatment group comparisons, the active treatment groups were combined and compared to placebo using all statistical methods for the primary clinical endpoint and the primary statistical method for the secondary outcome measures. p Values <0.05 were considered significant.
Comparisons of baseline data were carried out using the parametric two-sample t-test or the nonparametric Mann-Whitney test. Analysis of adverse event data was performed using Fisher’s exact test.
Results.
Study completion/compliance.
Fifty subjects were enrolled (figure 1). Forty-nine subjects completed 24 months of follow-up. Forty-three subjects completed 24 months on study drug; 7 patients had their dose reduced. The majority of treatment withdrawals and dose reductions (table 1) occurred in the IFN60 group, with only 6 of the 15 patients in this group completing the study on full dose. Patients in the active treatment groups withdrew because of flulike reactions and in the placebo group because of perceived lack of benefit. The most common reason for dose reduction was a rise in liver enzymes.
Figure 1. Study profile (CONSORT Statement).
Table 1 Reasons for treatment withdrawal and dose reduction
Baseline characteristics.
There were no significant differences in age, disease duration, EDSS score, or baseline MRI characteristics between treatment arms (table 2). There were no important prerandomization differences in immunosuppressant treatment.
Table 2 Baseline characteristics
MRI reproducibility.
The mean coefficients of variation for same scan analysis were as follows: T2 lesion volume 2.63%, T1 lesion volume 4.08%, spinal cord area 0.48%, BBSI 2.13%, and ventricular volume 0.15%.
Primary clinical endpoint.
The primary clinical endpoint was reached in 48% of subjects. There was no significant difference in disease progression between the individual or combined treatment arms and placebo (figure 2).
Figure 2. Survival curves for time to sustained disease progression. (A) Placebo (solid line) and combined interferon (IFN) (dashed line) groups. (B) Placebo (solid line), IFN30 (long dashed line), and IFN60 (short dashed line) groups.
Secondary clinical outcomes.
No treatment effect was seen on the TTMW (table 3). On nonparametric analysis of covariance of the NHPT, differences were suggested between the IFN30 group and both the control (p = 0.08) and IFN60 groups (p = 0.09) for the right side. On mixed-model regression analysis there were no significant differences between the active treatment groups compared to controls, but the IFN30 group had a lower rate of deterioration in right-side performance than the IFN60 group (p = 0.02).
Table 3 Secondary clinical outcome measures (performance time in seconds) throughout study
MRI outcomes.
T2 lesion load.
No significant differences in T2 lesion load were seen between individual groups on nonparametric analysis of covariance, but comparing the placebo and the combined interferon group there was a nonsignificant difference (p = 0.075) favoring the interferon group (table 4). On mixed-model regression analysis there was a lower rate of increase in lesion load in the IFN30 group compared to the placebo group (p = 0.025).
Table 4 T2 and T1 brain lesion loads throughout study
T1 lesion load.
No significant treatment differences in T1 lesion load were seen (see table 4).
New lesions.
Few new lesions developed and no significant treatment effect was seen.
Spinal cord area and brain volume.
No significant treatment effect was seen in spinal cord area or BBSI (table 5). On nonparametric analysis of covariance of ventricular volume, the IFN60 group had a greater rate of ventricular enlargement than the placebo group (p = 0.025). However, on mixed-model regression analysis no significant differences were seen in rates of progression.
Table 5 Spinal cord area, BBSI, and ventricular volume throughout study
Adverse events.
The commonest adverse event in all groups was flulike reactions (table 6). Mood disturbance was more commonly associated with interferon, although in most subjects it was mild and did not require treatment. Elevations in alanine transaminase levels occurred more often in the IFN60 group, but they were reversible with no clinical sequelae. Four subjects in the placebo group had acute neurologic relapses; all relapses were mild and resolved without steroid therapy. There were no significant differences in serious adverse events requiring hospital admission between the treatment groups (data not shown).
Table 6 Common adverse events (no. of subjects affected on one or more occasions)
Neutralizing antibodies.
One subject in the IFN30 group became positive for neutralizing antibody.
Blinding.
Fifty-eight percent (28/48) of patients correctly identified whether they received placebo or interferon β-1a. The treating physician correctly identified placebo vs interferon treatment in 86% (43/50) of patients.
Discussion.
This is the first randomized, controlled trial of a β-interferon devoted solely to PPMS. Although only an exploratory study, it has demonstrated that therapeutic trials in PPMS are feasible. Subject compliance was excellent, with 98% of subjects completing the study. Disease progression was clearly evident, with 48% of subjects reaching the primary clinical endpoint. The study has provided information on the safety and tolerance of interferon β-1a and the usefulness of outcome measures in this group.
With regard to safety, there were no life-threatening adverse events and no significant differences in serious adverse events between treatment groups. The interferon β-1a 30-μg dosage was well tolerated, with only one treatment withdrawal and two dose reductions. In contrast, the 60-μg dosage was poorly tolerated, owing to flulike reactions and elevated liver enzyme levels, with only 6 of 15 patients completing the study on full dose; this poor treatment compliance indicates that the data from this group must be interpreted with great care. The recent dose comparison study of IM interferon β-1a in relapsing-remitting MS also reported an increased incidence of flulike reactions and elevated liver enzyme levels in the 60-μg group compared to the 30-μg group, but overall both doses were well tolerated.23 The discrepancy in tolerance of the 60-μg dosage between the studies may partly reflect the younger age and lower disability of patients in the relapsing-remitting study.
In view of the small sample size, any discussion of efficacy must be circumspect. No treatment effect was seen on the primary clinical endpoint. There was a suggestion of a treatment difference on right-side performance on the NHPT. The clinical interpretation of this is unclear because this is not the usual method of analysis of the NHPT. However, the NHPT is a potentially useful measure in PPMS. Patients with PPMS frequently have a paraparesis or quadriparesis, with the legs being much more affected than the arms. A measure of upper limb function therefore may be more sensitive to change than measures of ambulation in patients who have already accrued significant lower limb dysfunction. The NHPT has previously been shown to be more sensitive to change than the EDSS in MS.24 In the study of IM interferon β-1a in secondary progressive MS,16 no effect was seen on the EDSS score and the treatment effect on the MS Functional Composite25 was almost entirely due to change in the NHPT.
A differential treatment effect was seen on two of the eight MRI outcomes, T2 lesion load, and ventricular volume. There was a significant increase in T2 lesion load in the placebo group but not in the IFN30 group. This finding supports the recent suggestion26 that T2 lesion load may be a more responsive measure in PPMS than previously appreciated.27 It also suggests the possibility of an anti-inflammatory effect, despite PPMS being considered a disease characterized by low-grade inflammation,6,28⇓ and a therapeutic role for anti-inflammatory agents in PPMS cannot be ruled out.
The IFN60 group appeared to do worse in terms of ventricular enlargement. One possible explanation is that this group had a higher lesion load at baseline, though not reaching significance, and atrophy may be greater in patients with a higher burden of disease.29 Another possible explanation is that β-interferon may contribute to brain atrophy because of its anti-inflammatory effect, as was suggested in a study in secondary progressive MS.30 However, as previously discussed, it is difficult to draw conclusions regarding the IFN60 group because treatment was so poorly tolerated. What is perhaps more important is that a differential effect was detected and ventricular volume may be a useful outcome measure in PPMS, though further evaluation of its relationship with disease progression is required.
Other studies support the feasibility of therapeutic trials in PPMS. A pilot trial of riluzole has also been completed in this group31 and exploratory studies of interferon β-1b32 and mitoxantrone33 and a large, multicenter, phase III trial of glatiramer acetate34 are under way. The lack of reliable MRI markers of disease progression remains a challenge for therapeutic trials in PPMS, but this study has shown that T2 lesion load and ventricular volume have potential in this group. Finally, although this study has provided some evidence to support a phase III study, the advisability of carrying out such a trial remains uncertain.
Acknowledgments
Supported by Biogen. The NMR Research Unit is supported by the Multiple Sclerosis Society of Great Britain and Northern Ireland.
Acknowledgment
The authors thank M. King for expert statistical advice and G.J. Barker, W.R. Crum, S.P. Dyson, N.C. Fox, C.M. Griffin, S.J. Hickmann, G.T. Ingle, D.G. MacManus, R. Gordon, A.S. Lowe, I.F. Moseley, G.J.M. Parker, P. Rudge, N. Tubridy, and D.J. Werring for their invaluable help.
- Received April 18, 2002.
- Accepted August 29, 2002.
References
Disputes & Debates: Rapid online correspondence
REQUIREMENTS
If you are uploading a letter concerning an article:
You must have updated your disclosures within six months: http://submit.neurology.org
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.