Saturday 23 February 1991
`
`No 8739
`
`ORIGINAL ARTICLES
`
`The Canadian cooperative trial of cyclophosphamide
`and plasma exchange in progressive multiple
`sclerosis
`
`THE CANADIAN COOPERATIVE MULTIPLE SCLEROSIS STUDY GROUP*
`
`non-specific
`whether
`To
`find
`out
`immunosuppression is beneficial in multiple sclerosis
`(MS) a randomised, placebo-controlled, single-
`masked trial was carried out in nine university
`centres. 168 patients with clinically or laboratory-
`supported definite MS in progressive phase
`(deterioration by at least 1·0 on the expanded
`disability status scale [EDSS] in the previous year)
`randomised
`receive
`intravenous
`to
`were
`cyclophosphamide and oral prednisone (n=55);
`oral cyclophosphamide, alternate day
`daily
`prednisone (22 weeks), and weekly plasma
`exchange (20 weeks) (n=57); or placebo
`medications and sham plasma exchange (n=56).
`All patients were followed for at least 12 months
`(mean 30·4 months) by a monitoring neurologist,
`who was aware of treatment allocation, and an
`evaluating neurologist, who was not. The primary
`analysis was a comparison of rates of treatment
`failure (worsening of evaluating neurologist’s
`assessment of EDSS by 1·0 or more on two
`consecutive 6-monthly assessments). There were no
`significant differences among the groups in this
`primary analysis (19 [35%] treatment failures with
`cyclophosphamide; 18 [32%] with plasma
`exchange; 16 [29%] with placebo). Nor were there
`any differences in the proportions improved,
`stabilised, or worsened at each 6 month assessment
`or in the mean change in the EDSS at the final
`assessment (0·81 cyclophosphamide; 0·69 plasma
`exchange; 0·69 placebo). A slight trend favouring
`
`the plasma exchange group at 12-24 months of
`follow-up was not sustained at the final assessment.
`This study fails to confirm previous reports that
`immunosuppressive treatments result in stabilisation
`or improvement in progressive MS.
`
`Introduction
`Progress towards an effective treatment for multiple
`sclerosis (MS) has been slow. Although it is widely believed
`pathological changes of inflammation,
`the
`that
`demyelination, and gliosis result from a disturbance of the
`immune response! in genetically susceptible individuals 2
`the aetiology and pathogenesis of MS remain imperfectly
`
`*Executive Committee J. H. Noseworthy, G. C. Ebers, M. Gent, T P.
`Seland, K H Shumak, D W Taylor, M K Vandervoort. External Safety
`Monitoring Committee R S Roberts (Hamilton, Ontario), J P Antel
`(Montreal), J F Kurtzke (Washington, DC). Central Coordinating
`Centre. University of Western Ontario, London, Ontario.
`Participating centres University of Western Ontario, London (J. H.
`Noseworthy, M K Vandervoort, W P Mclnnis, G C Ebers, G. P A.
`Rice, B G Weinshenker, D. Hollomby, N Muirhead), University of
`Calgary, Alberta (T. P Seland, O Suchowersky, C Harris, J. Klassen);
`University of Manitoba, Winnipeg (A Auty, J M. Del Campo, J. Todd,
`H. Rayner); University of British Columbia, Vancouver (D. Paty, S.
`Hashimoto, R. Farquhar, W B Benny), University of Toronto (T Gray,
`G Sawa, K. H. Shumak, A. Royal, E McBride, J. J. Freedman),
`University of Saskatchewan, Saskatoon (D.J. MacFayden, W. Hader, E
`Ashenhurst, C Johnson, M. Schmaltz, D Sheridan, R. Card); University
`of Montreal (P Duquette, L Roy, J Pleines, Y Lapointe); McMaster
`University, Hamilton, Ontario (J. E. Paulseth, R Lo, M Gent, D. W.
`Taylor, B Neufeld, J A McBride), University of Ottawa (R. Nelson, D.
`Preston, G. Rick, L Huebsch).
`Correspondence to Dr J H Noseworthy, Department of Neurology,
`Mayo Clinic, 200 First Street, SW, Rochester, Minnesota 55905, USA
`
`MERCK 2061
`HOPEWELL v MERCK
`IPR2023-00481
`
`

`

`442
`
`understood. It is difficult to predict either the occurrence or
`the rapidity of progression.3 Clinical methods of
`determining disease activity, progression, and response to
`treatment have been imprecise, although advances in
`magnetic resonance imaging may provide a means to
`measure these variables objectively.
`Therapeutic efforts have concentrated on treatments that
`suppress the immune response non-specifically. Although
`no such approach has achieved sustained improvement or
`stabilisation in progressive MS patients, modest benefits
`have been claimed with cyclophosphamide,"O plasma
`exchange,’ total lymphoid irradiation," cyclosporin,12,13 and
`azathioprine.14-17
`to two promising reports on
`In
`response
`cyclophosphamide and plasma exchange physicians at
`nine Canadian MS clinics and members of the Canadian
`Apheresis Study Group joined together for a study to
`determine whether either or these approaches to non-
`specific immunosuppression could benefit patients with
`progressive MS.
`
`Patients and methods
`
`The trial design was approved by the institutional review
`committees at each of the nine university centres. Enrolment took
`place from November, 1985, to December, 1988. All patients were
`followed until December, 1989.
`The inclusion criteria were clinically definitel8 or laboratory
`supported19 definite MS judged to be in a progressive phase
`(evidence of deterioration of at least 1 ’0 point on the expanded
`disability status scale [EDSS]2O over the preceding 12 months); an
`EDSS at entry of 4 0-6’5; and age at least 15 years. Both patients
`with chronic progressive (progressive from onset without relapses)
`and relapsing progressive MS (with occasional relapses from which
`the patient did not recover) were eligible for randomisation to
`cyclophosphamide, plasma exchange, or placebo treatment (table i).
`Exclusion criteria were previous treatment with cyclophosphamide,
`cyclosporin, antilymphocyte globulin, or interferon; treatment with
`azathioprine or plasma exchange in the preceding year or
`corticosteroids in the preceding month; illnesses which might be
`adversely affected by any of the experimental treatments (unstable
`cardiovascular, liver, renal, or bone marrow disorders, autoimmune
`illnesses, previous malignant disorder, uncontrolled infection, or
`decubitus ulcers); substantial cognitive impairment; and
`unwillingness to use contraception throughout the trial and 2 years
`afterwards. We also excluded patients for whom plasma exchange
`personnel thought that weekly venous access would be difficult.
`Two superficial arm veins were required concurrently for plasma
`exchange.21 Access by way of subclavian, internal jugular, or
`femoral veins was not permitted; nor was the use of any form of
`fistula, arterial-venous shunt, venous cutdown, or arterial catheter.
`These access sites and interventions would have been unjustified in
`the event of allocation to sham plasma exchange.
`The trial design was reviewed with the eligible patients and
`family members who were given a letter of explanation and consent
`form. An exclusion form was completed on all patients who satisfied
`the inclusion criteria but either did not enter because of exclusion
`factors or who were eligible but refused randomisation.22 The latter
`were not permitted off-protocol treatment with the investigational
`agents.
`Randomisation was done on a day when it was possible to proceed
`directly with treatment. A randomisation sequence was generated
`separately for each centre. Patients were stratified by centre and
`EDSS score (below 6-0 or 6-0 and above).
`Patients assigned to the cyclophosphamide group were admitted
`to hospital and given 1 g cyclophosphamide intravenously on
`alternate days. Treatment was stopped when the white blood cell
`count fell below 45 x 109/1, since we knew this would result in a
`nadir white blood cell count of 1 O-2’O x 10’/I, or when the patient
`had received 9 g cyclophosphamide. Patients received 40 mg
`prednisone orally for 10 days; the dose was reduced by 10 mg on
`
`TABLE I-PATIENT CHARACTERISTICS ON ENTRY TO TRIAL
`
`alternate days and prednisone discontinued on day 16. To reduce
`the risk of haemorrhagic cystitis, all patients were kept well
`hydrated; they were catheterised and the bladder was continually
`irrigated (3 ml neosporin per 3 litres normal saline). Daily urinalysis
`was done and cyclophosphamide was discontinued if haematuria
`(more than 200 red blood cells per high power field) occurred.
`Drug-induced nausea was reduced by administration of
`cyclophosphamide in the evening, restriction of fluid intake, and use
`of antiemetics. Complete blood counts were done daily. The white
`blood count reached the nadir 5-7 days after the last dose of
`cyclophosphamide and began to recover 7 days later. Patients were
`nursed in isolation if the white cell count fell below 1-0 x 109/1 and
`were discharged from hospital when it rose above 2-5 x 109/1.
`plasma exchange
`Patients
`in
`the
`group received
`cyclophosphamide by mouth (1 -5-2-0 mg/kg daily) for 22 weeks
`and alternate day oral prednisone (20 mg every other day tapered
`over 22 weeks). Plasma exchange of one plasma volume (40 ml/kg)
`was done weekly for 20 weeks with either intermittent (5 centres) or
`continuous (4 centres) flow-type centrifuges. 5% serum albumin
`was used as the replacement fluid. Urinalysis and complete blood
`counts were done weekly. The dose of cyclophosphamide was
`adjusted weekly to achieve a target white blood count of 4a-
`5-0 x z1.
`Placebo group patients were given cyclophosphamide placebo by
`mouth daily and prednisone placebo every alternate day for 22
`weeks. Sham plasma exchange of one plasma volume was done
`weekly. These patients received their own plasma as replacement
`fluid. In both the plasma exchange and placebo groups, the
`apparatus and replacement fluids were shielded from the patient’s
`view by a curtain. Adjustments were made in the dose of
`cyclophosphamide placebo to mimic the dose adjustments in the
`plasma exchange group.
`Each patient was followed by both a monitoring neurologist who
`was aware of treatment allocation and an evaluating neurologist who
`was not. The monitoring neurologist supervised the experimental
`treatments. At entry and every 6 months, patients were examined
`consecutively by both neurologists. Typically they observed the
`patient’s gait simultaneously but were instructed not to discuss their
`findings nor consult their previous records. Several steps were taken
`to ensure blinding of the evaluating neurologist. He or she was not
`involved with the patient’s ongoing care, since most evaluating
`neurologists did not work in the hospital where the inpatient and
`outpatient treatments were given, and he or she avoided asking
`questions which would be informative about the treatment. At the 6
`month assessment, all patients wore scalp coverings and gauze
`bandages around the antecubital fossae since the drug-induced
`alopecia and plasma-exchange-induced antecubital bruising would
`have indicated the treatment. Patients with exacerbations or
`progression were seen by the monitoring neurologist who was
`permitted to prescribe corticotropin or corticosteroids (prediusone,
`methylprednisolone). The central coordinating centre (London,
`
`

`

`443
`
`outcome was assigned to the group to which they were randomised
`(intention to treat analysis). All patients were followed and
`assessments continued until the end of the trial whether or not they
`met the definition of clinical failure. The sample size was calculated
`to detect a 30% difference in failure rate between one of the active
`treatments and the control group at an alpha of 0-05 (two-tailed) and
`a power of 90%. Secondary analyses involved the numbers of
`patients improved (reduction of EDSS S of 1.0 or more), stabilised
`(change of 0-5 EDSS points or less), or worsened (increase of EDSS S
`of 10 or more), the mean and median changes in the EDSS, the
`number of patients requiring cointervention, and the time to
`cointervention with cortocosteriods or corticotropin. The
`evaluating neurologist’s judgment was used for all but nine outcome
`assessments. 6 patients could not return to their participating centre
`for the final examination. Their EDSS and functional systems (FS)
`scores were calculated from information supplied by their treating
`physicians. 3 other patients were only seen by the monitoring
`neurologist at the final assessment.
`
`Results
`168 patients entered the study. The three treatment
`groups were well matched for age, marital status,
`comorbidity, duration of disease, and EDSS scores at entry
`(table I). All patients were followed until death or the end of
`the study period.
`In the primary analysis there were no statistically
`important differences in the cumulative proportion of
`treatment failures over time among the treatment groups
`(cyclophosphamide 19 [35%]; plasma exchange 18 [32%];
`placebo 16 [29%]). Although treatment failure tended to
`occur earlier in the placebo group, this difference was not
`significant (mean time to failure 24-8 [SD 76] mo
`cyclophosphamide; 29-3 [10-9] mo plasma exchange; 20-6
`[9-5] mo placebo; fig 1). Neither active treatment
`significantly prolonged the time to treatment failure
`(p=0-78 for cyclophosphamide, p=0’26 for plasma
`exchange, compared with placebo; fig 1).
`The secondary analyses of efficacy (tables n and ill)
`included all assessments irrespective of clinical failure status.
`The reducing numbers at each time point are due to variable
`amounts of follow-up because of sequential intake spread
`over the 36 month enrolment period. Both tables show that
`the active treatment groups did slightly better than the
`placebo group early in the follow-up period. The small
`observed advantage of plasma exchange was lost after about
`2 years; for cyclophosphamide the gain persisted for only 1
`year, after which cyclophosphamide-treated patients fared
`somewhat less well than the control group. None of these
`differences in EDSS between the control group and either
`active treatment group was clinically important or
`statistically significant after allowance for multiple
`comparisons over time points.
`Very few patients showed improvement in their condition
`after treatment in this trial (table II). Of the 16 patients
`
`TABLE III-CHANGES IN EDSSAT EACH ASSESSMENT
`
`Fig 1-Time to treatment failure.
`
`Ontario) was notified of all such cointerventions. At the end of the
`observation period, the patients were examined twice within 8
`weeks by both the evaluating and the monitoring neurologist (final
`assessment and confirmatory examination).
`The external safety monitoring committee monitored the
`progress of the trial every 6 months (severe adverse experiences,
`deaths, clinical status).
`At the start of the trial we decided that the primary analysis would
`involve a comparison over time of the cumulative treatment failure
`rates (by means of Breslow’s test23) in each of the three treatment
`groups by the techniques of survival analysis.24 A treatment failure
`was defined as a worsening of the evaluating neurologist’s score by
`1-0 points or more (2 step change) on the EDSS on two consecutive
`examinations separated by at least 6 months. The first of these two
`examinations was taken as the time of the treatment failure. All
`randomised patients contributed to the survival curve until the
`point of treatment failure, death, or end of follow-up. Subjects who
`did not complete the allocated treatment were followed and their
`
`TABLE II-PROPORTION OF PATIENTS IMPROVED, STABLE, OR
`WORSE ON EDSS AT EACH ASSESSMENT*
`
`*Evaluating neurologist’s assessment improved = 1 1 0 fall; stable= 05 5 change;
`worsened = 1 0 rise
`
`

`

`444
`
`TABLE IV-USE OF COINTERVENTION* BEFORE CLINICAL
`FAILURE
`
`*Corticotropm or corticosteroids
`
`diagnosed
`pulmonary
`embolism
`in
`was
`a
`cyclophosphamide-treated patient. Depression requiring
`psychiatric treatment developed in 3 plasma exchange
`patients, and angina developed in 1 patient in each of the
`cyclophosphamide and placebo groups. Severe alopecia
`occurred in all the cyclophosphamide-treated patients who
`received more than 2 g intravenous cyclophosphamide and
`in 51 % of the plasma exchange group (compared with 16%
`of placebo patients). Amenorrhoea was reported in 42% of
`women in the cyclophosphamide group (permanent in
`24%), 77% in the plasma exchange group (permanent in
`54%), and 11 % in the placebo group (permanent in 7%).
`More than 85% of patients in all three groups took more
`than 80% of their medications (pill count), and 90% of
`patients in the plasma exchange and placebo groups
`completed at least 90% of the planned plasma exchange
`treatments. 44 of 55 cyclophosphamide-treated patients
`achieved a target white blood count of less than 2-0 x 10’L
`The degree of physician and patient masking was
`determined at the final assessment. The evaluating
`neurologist was able to identify the treatment assignment in
`only 5% of cases. 68% of plasma-exchange-treated patients
`and 49% of placebo-treated patients accurately identified
`the treatment they received.
`
`Discussion
`The primary analysis showed no clinically or statistically
`important difference between either of the active treatments
`and placebo treatment. The slight trend in EDSS favouring
`the two active regimens at the 6 month and 12 month
`assessments must be balanced against the inconvenience,
`costs, and potential for serious adverse effects with these
`treatments and by a similar apparent net worsening in the
`cyclophosphamide group (compared with the placebo
`group) at 24 months, 30 months, and the final follow-up
`assessment. The EDSS is an inherently ordinal scale and
`unit changes may not be of equal importance over its whole
`range. It may therefore be inappropriate to analyse changes
`in the EDSS by parametric tests of significance,25 although
`such tests have been used in other studies of MS,26 The
`average difference in EDSS between the two active
`regimens and the placebo group at 6 months and 12 months
`(0-25 EDSS points or less) is half the smallest increment that
`the EDSS recognises (0-5 points). By any standard, changes
`of this size must be of limited clinical importance. Because of
`the limitations of the trial design, we cannot tell whether this
`short-lived apparent minor benefit was from one of the
`components of the active treatment protocols, a
`combination of these components, or a brief "placebo"
`response in the patients who became aware of their
`treatment (all of the cyclophosphamide group and 68% of
`the plasma exchange group). The extent of deterioration in
`the patients whose EDSS worsened by at least 1 0 points at
`
`Fig 2-Extent of deterioration in EDSS.
`Every patient whose EDSS increased by 1 ’0 (eva!uatmgneuro!og!st’s
`assessment) at anytime during trial is represented by a Ime connecting the
`EDSS score at entry with the highest (worst) EDSS score recorded during
`the tnal. Numbers m parentheses= numbers m each group who showed
`an increase of at least 1 0 on the EDSS.
`
`judged to have improved at the 6 month assessment, all but 4
`improved by only 1 EDSS point, the minimum needed for
`classification as improved. By 12 months only 8 patients met
`the criteria for improvement, 5 by the 1 point minimum.
`The degree of clinical change in those patients judged as
`having deteriorated was similar in all treatment groups
`(fig 2).
`The analysis of cointerventions by the monitoring
`neurologists suggested that steroids were used earlier and
`more often in the placebo group than in either of the two
`active treatment groups (table IV).
`2 patients died. 6 months after randomisation 1
`apparently stable cyclophosphamide-treated patient (EDSS
`6-0) died within 12 h of the development of symptoms of
`acute bronchopneumonia. The other death occurred 33
`months after randomisation in a placebo-treated patient
`with advanced liver disease. Two cyclophosphamide-
`treated patients had haemorrhagic cystitis, and 3 became
`septic during their hospital stays. 1 patient in the plasma
`exchange group was treated for vascular collapse during the
`3rd month of plasma exchange and hypertension developed
`in 1 at 24 months. Diabetes developed in 1 patient in each of
`the active treatment groups and herpes zoster infection
`occurred in 1 patient in each of the active groups. 1 non-fatal
`
`

`

`any time after randomisation was similar for all three
`treatment groups (fig 2).
`Was any trend seen? At the 12, 18, and 24 month
`assessments, the proportion of patients stabilised or
`improved was greater in the plasma exchange group than in
`the other groups. The difference in the mean EDSS
`between the plasma exchange and placebo groups was
`always less than 0-5 EDSS points, however, and the trend
`was no longer evident at the final follow-up. In Khatri and
`colleagues’ study,’ 11 of 26 patients treated by plasma
`exchange still had improved by 10 or more EDSS points
`11 months after the start of treatment, whereas only 4 of our
`48 plasma exchange patients had improved (all by 1 -0 EDSS
`points). Although plasma exchange is generally well
`tolerated,z7 90% of our patients had some adverse effect21
`and the procedural expense must be balanced by a
`meaningful clinical response. It remains possible that a more
`aggressive or longer course of cyclophosphamide,
`prednisone, and plasma exchange could produce a clinically
`and statistically significant benefit.
`Because of the theoretical risk that cyclophosphamide
`could potentiate the immune response28 and thereby worsen
`the course of MS, the drug’s known serious toxic effects,z9
`and the negative results of the only previously masked,
`placebo-controlled study,3O we opted for a two-tailed
`analysis of the data (table 11). More patients were stable or
`first
`after
`improved in
`intravenous
`the
`year
`cyclophosphamide than placebo but this difference was not
`significant (table 11). At all subsequent assessments, the
`number of patients judged to be worse was consistently
`greatest in the cyclophosphamide group. Patients in that
`group were certain they had been assigned an active
`treatment (single masked). Furthermore, the placebo
`patients did not receive prednisone. Despite these two
`factors, which might be expected to bias the results in favour
`of the cyclophosphamide group, treatment did not offer a
`significant advantage over placebo.
`There were more cointerventions with corticotropin and
`steroids in placebo-treated patients. In addition, the mean
`time to cointervention and the mean number of
`cointerventions per patient suggested an advantage for the
`active treatments. The decision to use steroids was made by
`the unmasked monitoring neurologist, however. We believe
`that his or her knowledge of the treatment assignment may
`have introduced bias into the decision to intervene. It is
`conceivable that the monitoring neurologist might have
`been more inclined to use steroids earlier and more
`frequently in placebo-treated patients than in actively
`treated patients with similar degrees of deterioration. The
`primary analysis suggests that the combination of sham
`plasma exchange, placebo drugs, and the occasional use of
`corticotropin or steroids as needed, is as effective as the
`experimental treatment regimens.
`Most patients in all three groups did much better in the
`year after entry to the trial than they had immediately before
`(the EDSS had fallen by 1-0 point in the preceding year).
`This finding clearly shows that patients cannot be used as
`their own controls in MS clinical trials.
`Why did the study fail to confirm a benefit with
`immunosuppressive treatments? Although significant
`improvement was rare in this study, the proportion of
`patients stable or improved 12 months after each of the
`active treatments was comparable or superior to previous
`series6.79,1O (table n). This comparability suggests that the
`
`445
`
`minor differences in the active treatment protocols tested in
`this study did not account for the disappointing results.
`Of greater importance was the difference in the behaviour
`of the comparison group. In Hauser and colleagues’ study,6
`only 4 (20%) corticotropin-treated patients (controls) were
`stable or improved at 12 months compared with 36 (75%) of
`our placebo-treated patients and their respective mean
`changes (worsening) on the disability status scale (DSS)
`were 0-70 (0-03) and 0-39 (0-09) at 12 months. The other
`favourable study7 had no untreated controls. The 75 %
`"apparent" stabilisation rate at 1 year in our placebo group is
`similar to that found in natural history studies3’ and
`treatment trialsy,32 Indeed, our placebo-treated patients
`were just as likely, if not more so, to be stable at 12 months as
`patients in series treated with cyclophosphamide, plasma
`exchange 7 and courses of cyclophosphamide with
`"boosters".9no This remarkable "response rate" has been
`seen in compliant placebo-treated patients in other
`settings.33 It is not clear why the corticotropin-treated
`patients in Hauser et al’s study6 did so much more poorly
`than would be expected from our knowledge of the natural
`history of progressive MS. Perhaps awareness of
`corticotropin treatment, which many might have received
`previously with little or no sustained benefit, had a negative
`effect on their disease course. If that control group had not
`done so poorly, that study7 would also have yielded negative
`results.
`Our entry criteria were more demanding than those used
`in the previous studies.6,7 We aimed to minimise the
`numbers of patients who stabilised spontaneously during
`the observation period. These strict entry requirements
`severely restricted the number of patients eligible for
`randomisation and forced us to extend our enrolment
`period, despite our large MS clinic population (more than
`7000 patients).
`The degree of inter-observer variability in the use of the
`EDSS and FS34 and differences in the time spent at each
`level of the DSS35 suggest that the outcome measures
`commonly used to estimate treatment effects (eg, disability
`scales, time to cointervention, time to being wheelchair
`bound) may not be sufficiently precise to detect small but
`clinically relevant slowing of disease progression. More
`sensitive clinical and imaging indices might have shown a
`difference with the active protocols tested in this study.
`Reliable, valid, reproducible, and clinically relevant
`measures of treatment effect will be needed to identify when
`experimental treatments have changed the natural history of
`active, progressive MS. From our study we cannot
`recommend that either of the active treatments tested in this
`trial should be offered to patients with MS. The slight
`confirmed clinical worsening during this trial in the control
`group underscores the requirement that experimental MS
`treatments must be compared with a convincing placebo
`treatment.
`
`This study was supported by grants from the Medical Research Council of
`Canada. J. H. N. is the recipient of a career development award from the
`Multiple Sclerosis Society of Canada. G. C. E. and G. P. A. R. are supported
`by career scientists awards from the Ontario Ministry of Health. We thank the
`Bristol Myers Pharmaceutical Group for cyclophosphamide and
`cyclophosphamide placebo and the Upjohn Company of Canada for
`prednisone and prednisone placebo; and Catherine Marsh and Linda
`A. Goldbeck for secretarial help.
`
`

`

`446
`
`REFERENCES
`
`1. Rationale for immunomodulating therapies of multiple sclerosis.
`Neurology 1988; 38 (suppl 2): 1-89.
`2. Ebers GC, Bulman DE, Sadovnick AD, et al. A population-based study
`of multiple sclerosis in twins. N Engl J Med 1986; 315: 1638-42.
`3. Matthews WB, Acheson ED, Batchelor JR, Weller RO. McAlpine’s
`multiple sclerosis. Edinburgh: Churchill Livingstone, 1985.
`4. Gonsette RE, Demonty L, Delmotte P. Intensive immunosuppression
`with cyclophosphamide in multiple sclerosis: a follow-up of 110
`patients for two to six years. J Neurol 1977; 214: 173-81.
`5. Hommes OR, Lamers KJB, Reeks P. Effect of intensive
`immunosuppression on the course of chronic progressive multiple
`sclerosis. J Neurol 1980; 223: 177-90.
`6. Hauser SL, Dawson DM, Lehnch JR, et
`Intensive
`al.
`immunosuppression in progressive multiple sclerosis. A randomized,
`three-arm study of high-dose intravenous cyclophosphamide, plasma
`exchange, and ACTH. N Engl J Med 1983; 308: 173-80.
`7. Khatri BO, McQuillen MP, Harrington GJ, Schmoll D, Hoffmann RG.
`Chronic progressive multiple sclerosis: double-blind controlled study
`of plasmapheresis in patients taking immunosuppressive drugs.
`Neurology 1985; 35: 312-19.
`8. Goodkin DE, Plencner S, Palmer-Saxerud J, Teetzen M, Hertsgaard D.
`Cyclophosphamide in chronic progressive multiple sclerosis
`maintenance vs nonmaintenance therapy. Arch Neurol 1987; 44:
`823-27.
`9. Carter JL, Hafler DA, Dawson DM, Orav J, Weiner HL.
`Immunosuppression with high-dose i.v. cyclophosphamide and
`ACTH in progressive multiple sclerosis: cumulative 6-year experience
`in 164 patients. Neurology 1988; 38: 9-14.
`10. Mackm GA, Weiner HL, Orav JA, et al. IV cyclophosphamide/ACTH
`plus maintenance cyclophosphamide boosters in progressive MS:
`intenm report of the Northeast Cooperative MS Treatment Group.
`Neurology 1990; 40 (suppl 1): 260.
`11. Cook SD, Devereux C, Troiano R, et al. Effect of total lymphoid
`irradiation in chronic progressive multiple sclerosis. Lancet 1986; i:
`1405-09.
`12. The Multiple Sclerosis Study Group. The efficacy and toxicity of
`cyclosporine A in chronic progressive multiple sclerosis: a randomized,
`double-blinded, placebo-controlled clinical trial. Ann Neurol 1990; 27:
`591-605.
`13. Rudge P, Koetsier JC, Mertin J, et al. Randomized double blind
`controlled trial of cyclosporin in multiple sclerosis. J Neurol Neurosurg
`Psychiatry 1989; 52: 559-65.
`14. Mertin J, Rudge P, Kremer M, et al. Double-blind controlled trial of
`immunosuppression in the treatment of multiple sclerosis: final report.
`Lancet 1982; ii: 351-54.
`15. Hughes RAC. Immunological treatment of multiple sclerosis I. J Neurol
`1983; 230: 73-80.
`16. British and Dutch Multiple Sclerosis Azathioprine Trial Group.
`Double-masked trial of azathioprine in multiple sclerosis. Lancet 1988;
`ii: 179-83.
`17. Ellison GW, Myers LW, Mickey MR, et al. A placebo-controlled,
`randomized, double-masked, variable dosage, clinical trial of
`azathioprine with and without methylprednisolone in multiple
`sclerosis. Neurology 1989; 39: 1018-26.
`18. Rose AS, Ellison GW, Myers LW, Tourtellotte WW. Criteria for the
`clinical diagnosis of multiple sclerosis. Neurology 1976; 26: 20-22.
`19. Poser CM, Paty DW, Scheinberg L, et al. New diagnostic criteria for
`multiple sclerosis: guidelines for research protocols. Ann Neurol 1983;
`13: 227-31.
`20. Kurtzke JF. Rating neurological impairment in multiple sclerosis: an
`expanded disability status scale (EDSS). Neurology 1983; 33: 1444-52.
`21. Noseworthy JH, Shumak KH, Vandervoort MK. Long-term use of
`antecubital veins for plasma exchange. The Canadian Cooperative
`Multiple Sclerosis Study Group. Transfusion 1989; 29: 610-13.
`22. Noseworthy JH, Vandervoort MK, Ebers GC. Acceptance of placebo-
`control trial design by progressive multiple sclerosis patients. The
`Canadian Cooperative Multiple Sclerosis Study Group. Neurology
`1989; 39: 606-07.
`23. Breslow NE. A generalized Kruskal-Wallis test for comparing k samples
`subject to unequal patterns of censorship. Biometrika 1970; 57: 579-94.
`24. Kaplan EL, Meier P. Non-parametric estimation from incomplete
`observations. J Am Stat Assoc 1958; 53: 457-81.
`25. Kurtzke JF. Neuroepidemiology. Part II: assessment of therapeutic trials.
`Ann Neurol 1986; 19: 311-19.
`
`26. Noseworthy JH, Vandervoort MK, Hopkins B, Ebers GC. A referendum
`on clinical trial research in multiple sclerosis: the opinion of the
`participants at the Jekyll Island workshop. Neurology 1989; 39: 977-81
`27. Shumak KH, Rock GA. Therapeutic plasma exchange. N Engl J Med
`1984; 310: 762-71.
`28. Ehrke MJ, Mihich E, Berd D, Mastrangelo MJ. Effects of anticancer
`drugs on the immune system in humans. Semin Oncol 1989; 16: 230-53
`29. Chabner BA, Myers CE. Clinical pharmacology of cancer chemotherapy.
`In: DeVita VT Jr, Hellman S, Rosenberg SA, eds. Cancer. Principles
`and practice of oncology. Philadelphia: JB Lippmcott, 1989: 369-71.
`30. Likosky WH. Experience with cyclophosphamide m multiple sclerosis:
`the cons. Neurology 1988; 38 (suppl 2): 14-18.
`31. Weinshenker BG, Bass B, Rice GPA, et al. The natural history of
`multiple sclerosis: a geographically based study. I. Clinical course and
`disability. Brain 1989; 112: 133-46.
`32. Miller A, Drexler E, Keilson M, et al. Spontaneous stabilization in
`patients with progressive MS. Neurology 1988; 38 (suppl 1): 194.
`33. Coronary Drug Project Research Group. Influence of adherence to
`treatment and response of cholesterol on mortality in the coronary drug
`project. N Engl J Med 1980; 303: 1038-41.
`34. Noseworthy JH, Vandervoort MK, Wong CJ, Ebers GC and the
`Canadian Cooperative MS Study Group. Interrater variability with
`the expanded disability status scale (EDSS) and functional systems
`(FS) in a multiple sclerosis clinical trial. Neurology 1990, 40: 971-75
`35. Weinschenker BG, Rice GPA, Noseworthy JH, et al. The natural history
`of multiple sclerosis. A geographically based study. IV. Applications to
`planning and interpretation o