Brain, Vol. 122, No. 4, 641-647,
April 1999
© 1999 Oxford University Press
The natural history of multiple sclerosis: a geographically based study
6. Applications to planning and interpretation of clinical therapeutic trials in primary progressive multiple sclerosis
Department of Clinical Neurological Sciences, University of Western Ontario, London, Ontario, Canada
Correspondence to:
Dr G. C. Ebers, Department of Clinical Neurological Sciences, London Health Sciences Centre-University Campus, London, Ontario, Canada N6A 5A5 E-mail: gebers{at}lhsc.on.ca
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionReferences |
|---|
The natural history of primary progressive multiple sclerosis (PP-multiple sclerosis) recently has been defined in a geographically based multiple sclerosis population. For a series of prognostically defined hypothetical entry criteria based upon current trends in presentation to the London Multiple Sclerosis Clinic, we determined the number of patients who would have been trial eligible. Using 23 year mean longitudinal natural history data, we identified the observed rate of deterioration for frequently used trial endpoints. Hypothetical entry criteria were based on the practical considerations which would attend the execution of clinical trials in progressive multiple sclerosis. We then developed a series of sample size tables giving the number of patients with PP-multiple sclerosis and the length of observation that would be required to detect a significant result (P = 0.05) for a 25, 50 and 75% decrease in the median time to progression with 80 or 90% power, with treatment efficacy based upon the ability to slow progression on the disability status score. It is expected that the considerations outlined here will prove useful for both trial design and interpretation of trials in PP-multiple sclerosis which will require multi-centre collaborative efforts.
primary progressive-multiple sclerosis; natural history; progression rate; treatment; clinical trials
DSS = disability status score; EDSS = extended disability status score; PP-multiple sclerosis = primary progressive multiple sclerosis
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Primary progressive-multiple sclerosis (PP-multiple sclerosis) has recently been identified as a possible distinct clinical phenotype of multiple sclerosis (Thompson et al., 1997
). PP-multiple sclerosis comprises ~20% of the total multiple sclerosis population (Cottrell et al., 1999) and has a prognosis for disability and mortality which is considered less favourable than the more common relapsing–remitting/secondary progressive types. The PP-multiple sclerosis population has received little attention in previous clinical trials. This is related in part to the lower frequency of this form of the disease, a reportedly less active MRI scanning profile, a pathology which may be less inflammatory, some uncertainty about the relationship of this subtype to the more common relapsing–remitting and secondary progressive forms, and finally to incomplete knowledge of the natural history of the disease. With the advent of effective therapies in the more common relapsing–remitting/secondary progressive forms, trials in PP-multiple sclerosis will undoubtedly be contemplated. Knowledge of the natural history of PP-multiple sclerosis may be a helpful prerequisite for such studies. We recently have described the natural history of a series of PP-multiple sclerosis patients from a clinic-based population followed longitudinally for a mean of 23 years (Cottrell et al., 1999) and we apply this information here to the practical problem of designing trials in PP-multiple sclerosis. These studies have demonstrated that definition and assessment are potentially easier, that predictable progression is largely independent of gender and age at onset and anatomical site of initial involvement, and that there is some dependence on the number of systems involved at onset. Fluctuations caused by relapses are uncommon, although plateaus in the DSS (disability status score) do occur. In order to take into account changes in time to diagnosis of PP-multiple sclerosis since the original studies which may bear on patient accrual, we have incorporated data from a second series of PP-multiple sclerosis patients seen over the last 5 years. The practical applications of this natural history data to the planning and interpretation of clinical therapeutic trials in PP-multiple sclerosis are outlined and the challenges of such potential studies are discussed.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Database
Our database has been detailed in previous publications (Weinshenker et al., 1989a
, b, 1991a, b; Cottrell et al., 1999). In brief, 216 PP-multiple sclerosis patients were followed-up over a mean of 23 years from onset of PP-multiple sclerosis at the Multiple Sclerosis Clinic in London, Ontario, Canada. This included follow up to death for 49.5% of this population. Levels of disability were evaluated longitudinally and, because of the characteristic relentless progression of PP-multiple sclerosis, little ambiguity in time taken to reach specific DSS levels was encountered. The clinical features and course, demographic prognostic factors, mortality data and comparison of progression with other progressive types of multiple sclerosis are detailed in companion papers (Cottrell et al., 1999; M. Kremenchutzky et al., in preparation). Although the long duration of follow-up carries the vast majority of patients through relevant endpoints for clinical trials, it was our impression that the clinical characteristics of patients at presentation have shifted in recent years to a shorter duration of disease and to a lower disability level. Accordingly, for confirmation and validation of hypotheses derived from the original cohort, a second series of 165 PP-multiple sclerosis patients was identified from a total of 1099 multiple sclerosis patients seen during the period between June 1992 and June 1997 (the London Multiple Sclerosis Clinic sees ~250–300 new multiple sclerosis cases per year). The new PP-multiple sclerosis patients were identified using the same definition as in the accompanying paper. Information on time to presentation, DSS at time of presentation, years to each DSS level, systems involved at onset and MRI were recorded. This second series enabled changes in time to diagnosis occurring between the two series to be incorporated into sample size calculations for specified entry criteria.
Design of clinical trials
We considered several hypothetical entry criteria and based our analyses upon both practical and statistical considerations. The demographic and disability data at presentation from the second series PP-multiple sclerosis cohort and the distribution of DSS levels seen in a cross-section of the clinic's total PP-multiple sclerosis population were employed to make the projected calculations current and relevant. Therefore, the level of disability for the criteria was derived from the more predictable and consistent early stages of the DSS. At these lower levels, median time for progression to the next Kurtzke level is similar. However, to avoid unrealistic goals in the practical clinical setting, this had to be balanced against the fact that the DSS at presentation varies widely and can be surprisingly high. This can disqualify many patients for trials requiring independent ambulation. The level of disability at first presentation had changed in the time between the first and the second series, with a trend towards earlier presentation, a favourable development. There needed to be allowance made for the fact that clinic populations contain a concentration of patients at DSS 6, reflecting the relative stability of this Kurtzke stratum (Ebers and Paty, 1998; Willoughby and Paty, 1988). In addition, we set a time restriction on early progression to DSS 3 and 4 to eliminate the most benign PP-multiple sclerosis cases. We aimed to eliminate ~20% of the slowest progressing patients to improve trial efficiency. This modification may be considered to incorporate a measure of safety for the sample sizes given in the tabulated calculations to follow, since the projected progression rates are calculated from the total population including the relatively slow progressors.
We considered the commonly used endpoint of deterioration of 1 point on the DSS, except for progression from DSS 6 where the endpoint was half a point, i.e. 6.5 [extended DSS (EDSS)]. It has become commonplace in multiple sclerosis trials to treat progression from any level to the next on the DSS as equivalent (in terms of treatment failure or progression of disability) despite the scale being ordinal. This will be examined and remedies suggested in another paper in this series. Since DSS and not EDSS scores had been recorded, the assumption was made that the progression from EDSS 6.0 to 6.5 is equivalent to 6.5 to 7.0. This variation in endpoint was made based upon the observation that the mean time spent at DSS 6.0 is roughly double that for other lower DSS levels (Table 2) which are relevant to PP-multiple sclerosis trial entry.
|
We explored two sets of entry criteria for a trial duration of 3 years: (A) patients at DSS 3, 4 and 5 who had progressed one level in the last year and had reached DSS 3 within 5 years and (B) patients at DSS 4, 5 and 6 who had progressed one level in the last year and had reached DSS 4 by 10 years. The requirement for progression in the past year was dictated by the practical constraint characterizing potential trial patients
Statistical analysis
All analysis was performed using SAS (SAS Institute Inc., Cary, NC, USA). Progression probabilities and median staying times were estimated by survival analysis performed by SAS/LIFE TEST using the life table (actuarial) method with intervals of 1 year. In this analysis, patients who had not yet reached the given DSS level but who had been followed for a known period of time were included as right censored.
In order to estimate a single distribution of progression probabilities for patients entering a trial at a variety of different levels, entry of eligible patients into a trial and their performance was simulated for each entry criterion. For patients who were eligible to enter the trial on more than one occasion during their history, the simulation program randomly selected a level of disability at which each patient entered the trial. Thus, each eligible patient supplied a single outcome that consisted of the time to reach the next DSS level. Since only DSS levels 6 and 7 were recorded in the data, the time to level 6.5 was assumed to be half the time from DSS 6 to DSS 7. In the case of criterion B, patients who entered at DSS 6 were considered to have progressed in half the time it took them to reach DSS 7. Ten simulation runs were done for each set of entry criteria. Each run produced a survival distribution of time to the next level of DSS (progression) and a median staying time. These distributions and medians were averaged and the results used as a basis for sample size computations. The sample size computations were based on formulas given by Donner (1984).
Total number of multiple sclerosis patients needing to be screened
Using the 1992–1997 cohort, we determined the number of newly diagnosed PP-multiple sclerosis cases who would be eligible for entry according to either set of entry criteria over a 2 year accrual in 1996 and 1997. Using the total multiple sclerosis cases as the numerator and the number of suitable candidates for each criterion as the denominator, we multiplied this by the number of PP-multiple sclerosis patients required for the power and efficacy in order to create Table 8.
|
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Data from the June 1992 to June 1997 5 year cohort
From a total of 1099 patients newly presenting to the London Multiple Sclerosis Clinic in the last 5 years, 51 had either uncertain, ambiguous or insufficient historical information about the onset of the disease for confidence about phenotypic type. Of the 1048 remaining, we found 165 (15.7%) to have PP-multiple sclerosis. The demographics of this population are shown for comparison with the original cohort in Table 1
. Of these 165 patients, 96.9% had the diagnosis supported by MRI. The median time to presentation to the London Multiple Sclerosis Clinic from onset of symptoms was still 4 years, which is less than that of 8 years found in the original PP-multiple sclerosis cohort. The lower rate of PP-multiple sclerosis in this second series was undoubtedly influenced by a bias towards seeing more relapsing patients interested in relapse-preventing therapies and was less evident in a strictly defined epidemiological sample (data not shown).
|
Both the mean and median DSS scores were 4 at presentation to the London Multiple Sclerosis Clinic for PP-multiple sclerosis patients in the period of June 1992 to June 1997. In this cohort, 43% of patients presented at DSS 3 or less, 68% at DSS 4 or less and 80% at DSS 5 or less. If a trial entry criterion was for PP-multiple sclerosis patients at DSS 3 or less, 57% of new referrals to this clinic would be ineligible on this basis alone. However, if the bar were set for patients at DSS 5 or less, only 20% would be ineligible.
There were 59 PP-multiple sclerosis patients out of 300 multiple sclerosis cases seen in the 2 year accrual period 1996–1997, of which 43 (73%) fulfilled criteria A and 49 (83%) criteria B.
Significant prognostic factors and DSS at presentation
We examined whether having three or more systems at onset affected the DSS at presentation since this group may have had an unusually bad prognosis (Cottrell et al., accompanying paper). If so, they could be selected more justifiably as a subgroup for pilot studies of risky therapies. In our cohort of 165 new cases, eight (4.8%) had three or more systems affected at onset; their median DSS at presentation was 4, just as for patients with less than three systems involved.
Probability of progression
Table 2
shows estimates of the cumulative proportion of patients from the total PP-multiple sclerosis population who would progress from each DSS level to the next in t years along with the median of each distribution. As in previous reports, patients at DSS levels 6–9 have substantially longer staying times. The extremely high median staying time for DSS level 8 in Table 2 is probably artefactual, inflated by the relatively frequent lack of data for time to DSS 9 compared with that for DSS 8 and DSS 10 in our database.
Tables 3 and 4 show progression probabilities and median staying times for all patients eligible under criteria A and B, respectively, as well as the simulation mean and standard error results averaged from 10 computer-generated simulation runs.
|
|
Estimates of number of patients required in trials
Tables 5, 6 and 7
are derived from the simulation results of the probabilities of progression. The control rate in Tables 5 and 6 is one minus the mean progression rate. It is the estimated proportion (0.1, 0.2 or 0.3) of untreated patients who would remain stable (i.e. not progress) for the given number of years. The sample sizes are based upon absolute percentage increases in stability; for example, in a 2-year trial with entry criteria A, 368 patients per group would be required for a trial in order to have an 80% chance of detecting an increase in the stability rate from 33% in the untreated control group to 43% in the treated group. That is, 10% more of the total treatment population will remain at their original entry DSS at the end of the trial compared with the control group.
|
|
|
The results in Table 7
are based on a survival analysis approach. The assumption here is that the median staying time for untreated patients admitted by either entry criterion would be ~1.5 years, as suggested by the simulation results. As an example, in a 2 year trial with a 1 year recruitment period, 543 patients per group would be necessary to have an 80% chance of detecting an increase in median staying time from 1.5 to 1.875 years. It should be noted that in a trial of short duration (i.e. 1 year), only ~30–35% of the untreated patients would be expected to progress so that the medians could not be estimated. Here one could still test for a difference in survival curves based on months or days to survival. The calculations in Table 7 are based on the assumption that these survival curves would have the assumed medians even though they may not be estimable from the trial results.
In Table 8, we give the number of all multiple sclerosis patients needed to generate sufficient numbers of trial-eligible individuals using the entry criteria A and B, the numbers of years and the powers of 80 or 90%.
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
The standard for success in a clinical trial for PP-multiple sclerosis must be clinical response, and the DSS/EDSS have been until now the standard measurement for disability in multiple sclerosis. We address herein the number of patients (sample size calculation) that would be required for a trial with different degrees of hypothetical drug effectiveness, using entry criteria which would not exclude many PP-multiple sclerosis patients based on our Multiple Sclerosis Clinic's accession pattern. These arbitrary criteria were designed to accommodate as much as possible the practical problems which will attend recruitment for a multiple sclerosis subtype.
Although there are factors which serve to facilitate trials in PP-multiple sclerosis using the predominantly spinal-influenced EDSS, several drawbacks characterize prospective trials in PP-multiple sclerosis patients when compared with studies of patients with a relapsing–remitting phase. These patients are generally older and, therefore, will have more concomitant diseases which may cause a higher rate of exclusion, independently affect their mobility and confound the DSS, or even affect survival and preclude follow-up. For example, the walking distance in older patients can be influenced by co-morbid respiratory, cardiac and muscular skeletal/arthritic disorders. Some such conditions may be exclusionary for other reasons, such as influencing thresholds for side effects. A proportion of PP-multiple sclerosis patients may have had previous treatments which could disqualify them from randomized trials. Furthermore, for patients having higher levels of disability and a poorer social network of care giving, regular visits as part of an intensive follow-up regime are practically more difficult. Finally, since these patients are older they may be less able and/or less willing to tolerate potential adverse events. We have not taken these factors into account and would suggest including a correction factor for dropouts, a widely variable phenomenon with substantial site specificity.
The median referral time for patients with PP-multiple sclerosis has fallen from 8.0 years in the original cohort entering the clinic between 1972 and 1984 to 4.0 years for the new cohort entering between 1992 and 1997. We believe this is due mainly to the increased availability of diagnostic imaging, although the problem of non-specificity of scanning results increases with age. As reported above, some 97% of new PP-multiple sclerosis patients seen now have already had MRI scans. Nevertheless, we see more patients with PP-multiple sclerosis at an earlier stage of the disease, commonly at DSS levels as low as 2 or 3. Some of these patients would be relatively resistant to the idea of potentially harmful experimental therapies when their symptoms and disability are not greatly interfering with their quality of life.
It may well be that recent activity is more a practical matter than one bearing on predicted outcome as a determinant of sample size in trial design. In our experience, stable patients often are reluctant to entertain experimental therapies. Using prognostic outcome as an argument to enhance their participation, especially if this information has not been requested, will be considered by some to be invasive.
|
Acknowledgments |
|---|
We wish to thank Lamia Abid, Mary Hader, Cheryl Johnson, Debra King, Wilma Koopman, Jennifer Pritchard and Rita Schuhmacher for their enthusiastic help. This study was supported by the Multiple Sclerosis Society of Canada, its Ontario division, by Schering AG and by the Multiple Sclerosis Foundation of Canada.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Donner A. Approaches to sample size estimation in the design of clinical trials—a review [published erratum appears in Stat Med 1990; 9: 1228] [see comments]. [Review]. Stat Med 1984; 3: 199–214. Comment in: Stat Med 1993; 12: 1643.[Web of Science][Medline]
Ebers GC, Paty DW. Natural history studies and applications to clinical trials. In: Paty DW, Ebers GC, editors. Multiple sclerosis. Philadelphia: F. A. Davis; 1998. p. 192–228.
Thompson AJ, Polman CH, Miller DH, McDonald WI, Brochet B, Filippi M, et al. Primary progressive multiple sclerosis. [Review]. Brain 1997;120: 1085–96.
Weinshenker BG, Bass B, Rice GP, Noseworthy J, Carriere W, Baskerville J, et al. The natural history of multiple sclerosis: a geographically based study. 1. Clinical course and disability. Brain 1989a; 112: 133–46.
Weinshenker BG, Bass B, Rice GP, Noseworthy J, Carriere W, Baskerville J, et al. The natural history of multiple sclerosis: a geographically based study. 2. Predictive value of the early clinical course. Brain 1989b; 112: 1419–28.
Weinshenker BG, Rice GP, Noseworthy JH, Carriere W, Baskerville J, Ebers GC. The natural history of multiple sclerosis: a geographically based study. 3. Multivariate analysis of predictive factors and models of outcome. Brain 1991a; 114: 1045–56.
Weinshenker BG, Rice GP, Noseworthy JH, Carriere W, Baskerville J, Ebers GC. The natural history of multiple sclerosis: a geographically based study. 4. Applications to planning and interpretation of clinical therapeutic trials. Brain 1991b; 114: 1057–67.
Willoughby EW, Paty DW. Scales for rating impairment in multiple sclerosis: a critique. Neurology 1988; 38: 1793–8.
Received May 28, 1998. Revised October 27, 1998. Accepted November 17, 1998.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  L. Bosma, J. Kragt, L Brieva, Z Khaleeli, X Montalban, C. Polman, A. Thompson, M Tintore, and B. Uitdehaag The search for responsive clinical endpoints in primary progressive multiple sclerosis Multiple Sclerosis, June 1, 2009; 15(6): 715 - 720. [Abstract] [PDF]
|
|
|
|
|
|
H Butzkueven, J Chapman, E Cristiano, F Grand'Maison, M Hoffmann, G Izquierdo, D Jolley, L Kappos, T Leist, D Pohlau, et al. MSBase: an international, online registry and platform for collaborative outcomes research in multiple sclerosis Multiple Sclerosis, November 1, 2006; 12(6): 769 - 774. [Abstract] [PDF]
|
|
|
|
 |
|
 M. Kremenchutzky, G. P. A. Rice, J. Baskerville, D. M. Wingerchuk, and G. C. Ebers The natural history of multiple sclerosis: a geographically based study 9: Observations on the progressive phase of the disease Brain, March 1, 2006; 129(3): 584 - 594. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S L Minden, D Frankel, L Hadden, J Perloff, K P Srinath, and D C Hoaglin The Sonya Slifka Longitudinal Multiple Sclerosis Study: methods and sample characteristics Multiple Sclerosis, February 1, 2006; 12(1): 24 - 38. [Abstract] [PDF]
|
|
|
|
 |
|
 J Oh, R G Henry, C Genain, S J Nelson, and D Pelletier Mechanisms of normal appearing corpus callosum injury related to pericallosal T1 lesions in multiple sclerosis using directional diffusion tensor and 1H MRS imaging J. Neurol. Neurosurg. Psychiatry, September 1, 2004; 75(9): 1281 - 1286. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. C Ebers Natural history of primary progressive multiple sclerosis Multiple Sclerosis, June 1, 2004; 10(1_suppl): S8 - S15. [Abstract] [PDF]
|
|
|
|
|
|
G. C Ebers Natural history of primary progressive multiple sclerosis Multiple Sclerosis, May 1, 2004; 10(3_suppl): S8 - S15. [Abstract] [PDF]
|
|
|
|
|
|
G. T. Ingle, V. L. Stevenson, D. H. Miller, and A. J. Thompson Primary progressive multiple sclerosis: a 5-year clinical and MR study Brain, November 1, 2003; 126(11): 2528 - 2536. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G T Ingle, V L Stevenson, D H Miller, S M Leary, M Rovaris, F Barkhof, B Brochet, V Dousset, M Filippi, X Montalban, et al. Two-year follow-up study of primary and transitional progressive multiple sclerosis Multiple Sclerosis, April 1, 2002; 8(2): 108 - 114. [Abstract] [PDF]
|
|
|
|
|
|
X Montalban and A J Thompson Workshop on primary progressive multiple sclerosis: meeting summary Multiple Sclerosis, April 1, 2002; 8(2): 177 - 178. [PDF]
|
|
|
|
 |
|
 J. H. Noseworthy, C. Lucchinetti, M. Rodriguez, and B. G. Weinshenker Multiple Sclerosis N. Engl. J. Med., September 28, 2000; 343(13): 938 - 952. [Full Text] [PDF]
|
|
|
|
|
|
V L Stevenson, D H Miller, S M Leary, M Rovaris, F Barkhof, B Brochet, V Dousset, V Dousset, M Filippi, R Hintzen, et al. One year follow up study of primary and transitional progressive multiple sclerosis J. Neurol. Neurosurg. Psychiatry, June 1, 2000; 68(6): 713 - 718. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M Kremenchutzky, D Lee, G P A Rice, and G C Ebers Diagnostic brain MRI findings in primary progressive multiple sclerosis Multiple Sclerosis, April 1, 2000; 6(2): 81 - 85. [Abstract] [PDF]
|
|
|
|
|
|
G. C. Ebers, W. J. Koopman, W. Hader, A. D. Sadovnick, M. Kremenchutzky, P. Mandalfino, D. M. Wingerchuk, J. Baskerville, and G. P. A. Rice The natural history of multiple sclerosis: a geographically based study: 8: Familial multiple sclerosis Brain, March 1, 2000; 123(3): 641 - 649. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. Kremenchutzky, D. Cottrell, G. Rice, W. Hader, J. Baskerville, W. Koopman, and G. C. Ebers The natural history of multiple sclerosis: a geographically based study: 7. Progressive-relapsing and relapsing-progressive multiple sclerosis: a re-evaluation Brain, October 1, 1999; 122(10): 1941 - 1950. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. Rudge The value of natural history studies of multiple sclerosis Brain, April 1, 1999; 122(4): 591 - 592. [Full Text] [PDF]
|
|
|
|
|
|
D. A. Cottrell, M. Kremenchutzky, G. P. A. Rice, W. J. Koopman, W. Hader, J. Baskerville, and G. C. Ebers The natural history of multiple sclerosis:a geographically based study: 5. The clinical features and natural history of primary progressive multiple sclerosis Brain, April 1, 1999; 122(4): 625 - 639. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||