Glatiramer acetate for multiple sclerosis (Review)

Transcription

1 La Mantia L, Munari LM, Lovati R This is a reprint of a Cochrane review, prepared and maintained by The Cochrane Collaboration and published in The Cochrane Library 2010, Issue 5

2 T A B L E O F C O N T E N T S HEADER ABSTRACT PLAIN LANGUAGE SUMMARY SUMMARY OF FINDINGS FOR THE MAIN COMPARISON BACKGROUND OBJECTIVES METHODS RESULTS Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure DISCUSSION AUTHORS CONCLUSIONS ACKNOWLEDGEMENTS REFERENCES CHARACTERISTICS OF STUDIES DATA AND ANALYSES Analysis 1.1. Comparison 1 Glatiramer acetate versus placebo in relapsing remitting patient, Outcome 1 Patients who progressed Analysis 1.2. Comparison 1 Glatiramer acetate versus placebo in relapsing remitting patient, Outcome 2 Change in disability score at the end of follow-up Analysis 1.3. Comparison 1 Glatiramer acetate versus placebo in relapsing remitting patient, Outcome 3 Patients relapse free Analysis 1.4. Comparison 1 Glatiramer acetate versus placebo in relapsing remitting patient, Outcome 4 Mean number of relapses Analysis 2.1. Comparison 2 Glatiramer acetate versus placebo: secondary outcomes, Outcome 1 Number of hospitalisations at the end of follow-up Analysis 2.2. Comparison 2 Glatiramer acetate versus placebo: secondary outcomes, Outcome 2 Number of steroid courses at the end of follow-up Analysis 3.1. Comparison 3 Glatiramer acetate versus placebo: adverse effects, Outcome 1 Localised to the injection site. 53 Analysis 3.2. Comparison 3 Glatiramer acetate versus placebo: adverse effects, Outcome 2 Systemic adverse effects.. 54 Analysis 3.3. Comparison 3 Glatiramer acetate versus placebo: adverse effects, Outcome 3 Adverse effects causing treatment withdrawal Analysis 4.1. Comparison 4 Glatiramer acetate versus placebo in progressive patients, Outcome 1 progression of disability. 58 ADDITIONAL TABLES APPENDICES FEEDBACK WHAT S NEW HISTORY CONTRIBUTIONS OF AUTHORS DECLARATIONS OF INTEREST INDEX TERMS i

3 [Intervention Review] Glatiramer acetate for multiple sclerosis Loredana La Mantia 1, Luca M Munari 2, Roberta Lovati 3 1 Department of Neuroscience, Fondazione I.R.C.C.S. - Istituto Neurologico C. Besta, Milano, Italy. 2 Sapio Life, Monza, Italy. 3 U.O. Oncologia, Ospedale San Paolo, Milano, Italy Contact address: Loredana La Mantia, Department of Neuroscience, Fondazione I.R.C.C.S. - Istituto Neurologico C. Besta, Via Celoria, 11, Milano, 20133, Italy. lamantia@istituto-besta.it. Editorial group: Cochrane Multiple Sclerosis Group. Publication status and date: New search for studies and content updated (no change to conclusions), published in Issue 5, Review content assessed as up-to-date: 14 September Citation: La Mantia L, Munari LM, Lovati R. Glatiramer acetate for multiple sclerosis. Cochrane Database of Systematic Reviews 2010, Issue 5. Art. No.: CD DOI: / CD pub2. Background A B S T R A C T This is an updated Cochrane review of the previous version published (Cochrane Database of Systematic Reviews 2004, Issue 1. Art. No.: CD DOI: / CD004678) Previous studies have shown that glatiramer acetate (Copaxone ), a synthetic amino acid polymer is effective in experimental allergic encephalomyelitis (EAE), and improve the outcome of patients with multiple sclerosis (MS). Objectives To verify the clinical efficacy of glatiramer acetate in the treatment of MS patients with relapsing remitting (RR) and progressive (P) course. Search methods We searched the Cochrane MS Group Trials Register (26 March 2009), the Cochrane Central Register of Controlled Trials (The Cochrane Library, Issue 1, 2009), MEDLINE (PubMed) (January 1966 to 26 March 2009), EMBASE (January 1988 to 26 March 2009) and hand searching of symposia reports ( ). Selection criteria All randomised controlled trials (RCTs) comparing glatiramer acetate and placebo in patients with definite MS, whatever the administration schedule and disease course, were eligible for this review. Data collection and analysis Both patients with RR and P MS were analysed. Study protocols were comparable across trials. No major flaws were found in methodological quality. However, efficacy of blinding should be balanced against side effects, including injection-site reactions. Main results Among 409 retrieved references, we identified 16 RCTs; six of them, published between 1987 and 2007, met the selection criteria and were included in this review. Five hundred and forty RR patients and 1049 PMS contributed to the analysis. In RR MS, a decrease in the mean EDSS score (-0.33 and -0.45), was found respectively at 2 years and 35 months without any significant effect on sustained disease progression. The reduction of mean number of relapse was evident at 1 year (-0.35 ) 2 years (-0.51 ) and 35 months (-0.64), 1

4 but significant studies heterogeneity was found. The number of hospitalisations and steroid courses were significantly reduced. No benefit was shown in P MS patients. No major toxicity was found. The most common systemic adverse event was a transient and selflimiting patterned reaction of flushing, chest tightness, sweating, palpitations, anxiety. Local injection-site reactions were observed in up to a half of patients treated with glatiramer acetate, thus making a blind assessment of outcomes questionable. Authors conclusions Glatiramer acetate did show a partial efficacy in RR MS in term of relapse -related clinical outcomes, without any significant effect on clinical progression of disease measured as sustained disability. The drug is not effective in progressive MS patients. P L A I N L A N G U A G E S U M M A R Y The use of glatiramer acetate (Copaxone ) in people with multiple sclerosis This is an updated Cochrane review of the previous version published (Cochrane Database of Systematic Reviews 2004, Issue 1. Art. No.: CD DOI: / CD004678) Treatment with glatiramer acetate (Copaxone ) of patients with Relapsing-Remitting (RRMS) and with Progressive Multiple Sclerosis (PMS) seems to have few beneficial effects in RRMS, while the drug is not effective in PMS patients Previous studies indicate that glatiramer acetate, a synthetic drug, is effective in animal models of MS, and shows some benefits in MS patients. The objective of this review was to assess the efficacy of glatiramer acetate in RRMS and PMS patients. Among the pertinent medical literature six studies met the criteria of the methodological quality necessary for their inclusion in this review. 540 RRMS patients and 1049 PMS patients contributed to this analysis. The data showed no beneficial effects on disease progression in both MS forms, a slight beneficial effect in the reduction of risk of relapses in RRMS patients and no benefits in PMS patients. Adverse events such as flushing, chest tightness, sweating, palpitations, anxiety and local injection-site reactions occurred quite frequently, but no major adverse effects were observed. 2

5 S U M M A R Y O F F I N D I N G S F O R T H E M A I N C O M P A R I S O N [Explanation] Glatiramer acetate versus placebo in relapsing remitting patient for multiple sclerosis Patient or population: patients with multiple sclerosis Settings: Intervention: Glatiramer acetate versus placebo in relapsing remitting patient Outcomes Illustrative comparative risks*(95% CI) Relative effect (95% CI) No of Participants (studies) Assumed risk Corresponding risk Control Glatiramer acetate versus placebo in relapsing remitting patient Patients who progressed -at2years Study population RR 0.75 (0.51 to 1.12) 282per per1000 (144to316) 299 (2) Medium risk population 362per per1000 (185to405) Patients who progressed -at35months Study population RR 0.81 (0.5 to 1.29) 288per per1000 (144to372) 203 (1) Medium risk population 289per per1000 (144to373) Quality of the evidence (GRADE) See comment Comments 3

6 Change in disability score at the end of follow-up - at 2 years of follow-up ThemeanChange indisabilityscoreattheendof follow-up - at 2 years of follow-up in the intervention groups was 0.33 lower (0.58 to 0.08 lower) 301 (2) Change in disability score at the end of follow-up-at35monthsof follow-up ThemeanChange indisabilityscoreattheendof follow-up -at 35months of follow-up in the intervention groups was 0.45 lower (0.77 to 0.13 lower) 203 (1) See comment Mean number of relapses-within1yearof follow-up The mean Mean number ofrelapses-within1year of follow-up in the intervention groups was 0.35 lower (0.53 to 0.16 lower) 287 (2) Mean number of relapses-at2yearsoffollow-up The mean Mean number ofrelapses-at2yearsof follow-up in the intervention groups was 0.51 lower (0.81 to 0.22 lower) 298 (2) *The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention(and its 95% CI). CI: Confidence interval; RR: Risk ratio; 4

7 GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Lowquality:Furtherresearchisverylikelytohaveanimportantimpactonourconfidenceintheestimateofeffectandislikelytochangetheestimate. Very low quality: We are very uncertain about the estimate. xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx 5

8 B A C K G R O U N D Multiple sclerosis (MS) is a chronic inflammatory disease of the central nervous system (CNS) with either relapsing/remitting or progressive course. The pathology is characterized by random foci of demyelination and axonal loss throughout the CNS. Despite a better knowledge of these pathologic findings in the last decade, little is known about their underlying etiology. Based on experimental data, an autoimmune damage of the myelin sheath has been postulated as a mechanism of CNS inflammation. Susceptible animals inoculated with myelin components are known to develop experimental allergic encephalomyelitis (EAE), which is considered a laboratory model of MS (Wisniewski 1977). Glatiramer acetate (Copaxone ) is a synthetic amino acid polymer empirically found to suppress EAE. In animal models, the development of EAE can be prevented by glatiramer acetate administration (Teitelbaum 1997), possibly due to a displacement of immune cells targeted at native myelin components. Clinical results consistent with this rationale have also been shown in humans, leading to regulatory authorization of MS treatment from 1997 in the US and 2000 in Europe. Furthermore glatiramer acetate has been recently (June 2009) approved in Italy also for the treatment of clinically isolated syndrome with MRI parameters compatible with MS. Given the expectations raised by this agent and its worldwide use, we believe that updating of this systematic review of all randomised controlled trials (RCTs) evaluating glatiramer acetate (Munari 2004) needs to be undertaken in order to provide both clinicians and consumers with the most comprehensive information. O B J E C T I V E S This review is aimed at determining clinical efficacy and safety of glatiramer acetate in patients with MS. The main outcomes of interest were: (1) Clinical progression of disease in terms of sustained disability (2) Mean changes in EDSS disability score (3) Frequency of clinical relapses (4) Number of patients relapse free (5) Incidence of any adverse events (6) Patient s quality of life Secondary questions to be answered concern: 7) Number of patients treated with steroids and number of steroid courses administered during acute relapses or active disease progression (8) Impact of treatment on hospital admissions and length of stay, in order to detect potential savings both in terms of healthcare resources and patient s time M E T H O D S Criteria for considering studies for this review Types of studies All randomised or quasi-randomised controlled trials (RCTs) comparing glatiramer acetate and placebo in patients with definite MS were eligible for the review. Uncontrolled trials and studies where glatiramer acetate has been compared with interventions other than placebo were not included. Both double-blind and singleblind studies were eligible. Types of participants Patients of any age and either gender with definite MS according to Poser criteria (Poser 1983), whatever disease severity, were eligible for the review. Any patterns of MS course (relapsing/remitting (RR) relapsing/progressive, secondary progressive or primary progressive (P) have been considered. MS patients receiving cytostatics, immuno modulators or immunosuppressants in the 6 months prior to study enrolment were excluded from the analysis. Therefore, information on patient treatment regimens before entering the trial has been sought. Types of interventions All therapeutic schedules involving glatiramer acetate administration, whatever the administration route, dosage, treatment duration and the interval between symptom onset and randomisation were considered as test treatment. Courses of steroids were permitted, provided they were administered without any restriction in both arms. Types of outcome measures We sought the following measures in both treatment groups at 12 and 24 months and at the end of the scheduled followup period: Patients who progressed. Whenever unspecified, progression has been defined as a persistent worsening of at least one point in EDSS (Kurtzke 1983), recorded out of relapse and confirmed by a follow-up assessment at six months (Rio 2002). However, other definitions of progression given in the original paper could be accepted, including a persistent half-point increase starting from EDSS score 5.5 (Rio 2006). Mean changes in EDSS disability score We considered different relapse-related clinical outcomes and in particular : Frequency of clinical relapses, number of patients relapse free and number of patients relapse free over time. Secondary questions to be answered concern: 6

9 Number of patients treated with steroids and number of steroid courses administered during acute relapses or active disease progression and impact of treatment on hospital admissions and length of stay, in order to detect potential savings both in terms of healthcare resources and patient s time. Safety outcomes were assessed among primary endpoints by unique measures cumulating all events occurred throughout the trial: Number of both local and systemic side effects. Number of patients with severe side effects. If not otherwise specified, side effects have been defined as severe when leading to one of the following: death, hospitalisation, treatment discontinuation. Search methods for identification of studies A systematic search without language restrictions was conducted using the optimally sensitive strategy developed for the Cochrane Collaboration to identify all relevant published and unpublished randomised controlled trials (Lefebvre 2008). For additional information about the Group s search strategy please see: Cochrane Multiple Sclerosis Group Electronic searches We searched the following databases: 1. The Cochrane Multiple Sclerosis Group Trials Register (26 March 2009) 2. The Cochrane Central Register of Controlled Trials (CENTRAL) The Cochrane Library (issue 1, 2009) (Appendix 1) 3. MEDLINE (PubMed) (January 1966 to 26 March 2009) (Appendix 2) 4. EMBASE (EMBASE.com) (1974 to 26 March 2009) (Appendix 3) Searching other resources 1. Handsearched references quoted in the identified trials 2. Handsearched symposia reports ( ) from the most important neurological associations and MS Societies in Europe and America; 3. Contacted researchers who were participating in trials on GA. Contacts with the owner pharmaceutical company (Teva Pharmaceutical, Ltd.) were attempted without reply. So, we established reliable contacts with researchers involved in GA development. Data collection and analysis DATA EXTRACTION Selection of eligible studies and data extraction have been carried out independently by three reviewers (LM, LLM, RL). Results were then compared in order to rule out any misunderstandings, mistakes or biases possibly arising from data evaluation. Details on treatment administration schedule, patient enrolment criteria, diagnostic criteria, randomisation methods, blinding, outcome analysis, follow-up length, dropouts, side effects were also recorded for each study, in order to evaluate quality profiles (see Methodological quality). All data were entered in a collection form. Disagreements were resolved by discussion amongst reviewers. Trialists were asked to provide further details on study characteristics if they were unclear in the article. TRIAL QUALITY ASSESSMENT The methodological quality of each trial was assessed independently by reviewers. We used the recommended methods outlined in the Cochrane Reviewers Handbook version (Higgins 2008). All studies were given a quality score ranging from 0 to 5 (Jadad 1996), based on the following criteria: randomisation, allocation concealment, blinding, decisions about dropouts and withdrawals. Relevant information was collected using a data extraction form developed by the Multiple Sclerosis Cochrane Review Group. Randomisation has been defined as either telephone calls to a randomisation centre, reference to computer-generated random lists or tables of random numbers. Quasi-randomised trials without properly concealed allocation (e.g.: patient alternation, open random list, date of birth, day of the week or hospital admission number) have been included in the review. Allocation concealment and blinding have been scored in the risk of bias tables for each included study. Disagreements were resolved by discussion among the authors in order to achieve a unique score for each considered item. In case of significant differences between treatment and placebo, the effect of blinding could be tested in sensitivity analysis, since knowledge of treatment allocation may affect the assessment of study endpoints. Trial quality scores are listed in the additional Table 1. STATISTICAL ANALYSIS Data have been analysed according to an intention-to-treat approach. Relative risks, risk difference and their 95% confidence intervals (CI) have been calculated for binary outcomes. Continuous outcomes have been evaluated as weighted mean differences in treatment effects and their standard deviation (SD). The weighted treatment effect was calculated across trials for each outcome. Combined results were expressed as weighted estimates of relative risks with their 95% CI when binary variables were considered. Continuous outcomes were combined using weighted mean differences and their 95% CI. Basically, data were analysed in a fixed-effect model (Yusuf 1985). Homogeneity across trials have been tested in a chi square test with alpha=0.10. When significant heterogeneity was found, results were checked in a random-effects model (Brocke 1996). Characteristics of trials have been listed in the corresponding Characteristics of Included/excluded studies. All results have been organised and processed by the Review Manager 5.0 (RevMan 2008) developed by the Cochrane Collaboration. 7

10 The effects of potential sources of heterogeneity have been explored by subgroup analysis where appropriate (see results). Sensitivity analysis on trial quality and missing data was not needed. R E S U L T S Description of studies See: Characteristics of included studies; Characteristics of excluded studies; Characteristics of ongoing studies. Out of 409 references identified by the search strategy up to 26 March 2009, 133 abstracts were provisionally selected to be read as full published papers. Ninety three papers were then excluded for the following reasons: 53 were uncontrolled open-label studies (Abramsky 1977; Baumhefner 1988; Boiko 2006; Bornstein 1982;Brochet 2008;Caon 2006; Capobianco 2008; Carra 2008; Daugherty 2005; De Seze 2000; De Stefano 2008; De Stefano 2009; Debouverie 2007 ; Duda 2000; Flechter 2002b;Ford 2006; Fusco 2001; Gajofatto 2009 ;Garcia-Barragan 2009; Ghezzi 2005; Ghezzi b 2005; Haas 2005; Johnson 2000; Johnson 2003; Johnson 2005; Khan 2001; Kott 1997; Lage 2006 ; Le Page 2008; Mancardi 1998; Meiner 1997; Milanese 2005 ; Miller 1998; Miller 2006; Miller 2008; Ollendorf 2008; Orlova 2005; ;Ramtahal 2006; Rio 2005 ; Rovaris 2007 ;Schwid 2007 ; Sindic 2005; Tilbery 2006; Torkildsen 2007;Twork 2007 ; Valenzuela 2007; Vallittu 2005; Weder 2005; Wolinsky 2001;Ytterberg 2007 ;Zavalishin 2005; Ziemssen 2008; Zwibel 2006). Five studies were controlled not randomised studies evaluating the efficacy of GA and other immunomodulating agents without placebo group (Castelli-Haley 2008;Deen 2008; Flechter 2002a; Khan 2005; Zavalishin 2006), 7 studies restricted the analysis to MRI parameters (Cohen 1995; Mesaros, 2008; Rovaris 2005; Shipova 2009; Sormani 2002; Sormani 2005; Sormani 2007) 7 studies reported on experimental investigations where only laboratory endpoints have been assessed (lymphocyte activity, cytokine outburst, uric acid increase) or clinical immunological studies ( Blanco 2006; Brenner 2001; Chen 2001; Constantinescu 2000; Farina 2001; Karandikar 2002; Qin 2000), 21 studies aimed to evaluate adverse events during treatment with GA ( Achiron 2005; Ball 2008 ; Bosca 2006; Charach 2008; Cicek 2008; Feigin 2005; Fiore 2005 ; Harde 2007; khan 2008 ; La Mantia 2006; Madray 2008; ; Neumann 2007; Nolden 2005; Patten 2008;Pöllmann 2006; Rauschka 2005; Sidoti 2007;Soares 2006; Then, Bergh F 2006; Thouvenot 2007; Tremlett 2007) (See table of excluded studies). The remaining papers were related to 16 RCTs, nine RCTs were excluded because comparative trials evaluating the efficacy of two dosages of GA (Cohen 2007; Wynn 2008), of GA versus IFN beta (Cadavid 2009;Mikol 2008 ) of natalizumab versus placebo in Ga -treated MS patients (Goodman 2009 ) of GA after induction with mitoxantrone vs GA alone (Vollmer 2008;Arnold 2008) or cognitive function in GA versus placebo ( Weinstein 1999) or treatment of local reaction (Jolly 2008 ). One study was excluded because evaluating the efficacy of GA in isolated central nervous system syndrome ( Comi 2008). Six RCTs contributing to this review (29 related references) published between 1987 and 2007 (Bornstein 1987; Bornstein 1991; Johnson 1995; Comi 2001:Filippi 2006 ;Wolinsky 2007). These studies account for a total of 3233 patients, 2043 of whom allocated to glatiramer acetate and 1190 to placebo. Four studies enrolled patients with relapsing-remitting (RR) disease (Bornstein 1987; Johnson 1995; Comi 2001; Filippi 2006;). Two RCTs investigated the effect of glatiramer acetate in progressive MS (Bornstein 1991 Wolinsky 2007). Therapeutic schedules were homogeneous, except for Filippi 2006 study evaluating oral administration of GA. This trial was separately analyzed for concerns about the comparability of parenteral and oral administration. Therefore, the following treatments have been compared with placebo: glatiramer acetate: 20 mg subcutaneously, self-administered daily in RR MS glatiramer acetate 50-5 mg oral-administered daily in RRMS glatiramer acetate: 30 mg-20 mg subcutaneously, selfadministered daily in P MS The treatment has been given for 9 (Comi 2001), 14 (Filippi 2006 ) 24 (Bornstein 1987; Bornstein 1991) or 35 months (Johnson 1995). and 36 months (Wolinsky 2007). The characteristics of the studies are reported in the corresponding tables. All trials on RR MS enrolled patients with definite disease (Poser 1983). Bornstein et al. (Bornstein 1987) randomised patients within an age range of 20 to 35 years, with at least two exacerbations in the two years before admission, provided they were not severely disabled (EDSS score below 6) and/or emotionally unstable. Fifty-eight percent of study population were female and 64% of initially screened patients were excluded due to any of the following: age, low frequency of exacerbations, lack of documentation, impaired psychological profile, transition to CP MS, distance from the clinic or pregnancy. The US phase III pivotal trial (Johnson 1995) was a multicentre study involving 11 centres in the US. Eligible patients had an EDSS 5 and at least two documented relapses in the two years prior to entry, the last one occurring at least one year before randomisation; they should also be neurologically stable and free from corticosteroid therapy for at least 30 days prior to entry. Patients could be enrolled within a larger age range (18 to 45) and the final proportion of female subjects was 73%. Only 12% of candidate participants were excluded based on the following criteria: treatment with glatiramer acetate or previous immunosuppression with cytotoxic therapy or lymphoid irradiation, pregnancy or lactation, diabetes mellitus, positive HIV/HTLV-1 serology, Lyme disease, need of aspirin or chronic non-steroidal anti-inflammatory drugs 8

11 throughout the trial, unwillingness to undergo adequate contraception. Only EDSS modifying attacks confirmed by current neurological examination were accepted as relapses. Out of 215 patients who completed the first 24-month follow-up, 203 entered an additional 11-month treatment schedule (Johnson 1995), reproducing the same trial design. The investigators also carried out a further open-label follow-up up to six years from randomisation in 208 patients (Johnson 2000;Johnson 2003), to 8 years in 142 patients (Johnson 2005 ) to 10 years in 108 patients (Ford 2006) from the original cohort of 251, not included in this review. The European-Canadian MRI study (Comi 2001) applied the following inclusion criteria: patients aged 18 to 50 with an EDSS 5 with MS from at least one year. One documented relapse in the preceding two years was deemed sufficient to enter the study, but at least 1 enhancing lesion was essential in the screening brain MRI. Moreover, all randomised patients were clinically relapsefree and steroids-free in the 30 days before entry. A total of 29 centres participated in the study and 51% of screened patients were excluded due to any of the following: previous use of glatiramer acetate or oral myelin, prior lymphoid irradiation, use of immunosuppressant or cytotoxic agents in the past two years, use of azathioprine and/or other immunosuppressant including steroids during the previous six months, concomitant therapy with an experimental drug for either MS or another disease, serious intercurrent systemic or psychiatric illnesses; unwillingness to practice reliable contraception during study and known hypersensitivity to gadolinium, unavailability to repeat MRI studies. We excluded from the review the 9-month open-label extension phase of this trial. Flippi study (Filippi 2006) was separately evaluated. This study assessed whether two doses of glatiramer acetate given orally could improve clinical and MRI measures of inflammation and neurodegeneration in a large cohort of patients with relapsing-remitting multiple sclerosis. One thousand nine hundred and twelve patients with relapsing-remitting multiple sclerosis were screened and 1651 were randomised to receive 50 mg or 5 mg of glatiramer acetate or placebo by daily oral administration over 14 months. The intention-to-treat cohort consisted of 1644 patients who took at least one dose of study medication (50 mg glatiramer acetate [n=543], 5 mg glatiramer acetate [n=553], placebo [n=548]). After baseline investigation, clinical assessments were done every 2 months and MRI was obtained for all patients at baseline and at study exit. The main clinical data of the patients are reported in Table 2. Briefliy RR showed a disease duration ranging from 5.5 to 8.1 years, low disability and active clinical disease. Patients enrolled in the European-Canadian MRI study may represent a less severe subset, since they were eligible after a single relapse in the two previous years; however in this study an active MRI scan was needed. Patients enrolled had to be free of any steroid treatment for at least 30 days (Bornstein 1987; Johnson 1995; Comi 2001; Filippi 2006) and clinically stable for at least 30 days (Johnson 1995; Comi 2001). Minimum time elapsed from the last relapse was not specified in one study (Bornstein 1987). The study of Bornstein 1991 randomised patients between the age of 20 and 60, with a chronic-progressive course for at least 18 months, less than two exacerbations in the previous 24 months, disability on EDSS, emotional stability and a favourable psychosocial profile. These criteria were assessed in a pre-trial observation period lasting no more than 15 months and led to exclude 47% of candidate participants. The inclusion criteria may suggest that patients were affected by secondary progressive or progressive relapsing course.the primary outcome was confirmed progression (worsening of 1 EDSS or 1.5 according to basal EDSS ( 5 or less) maintained at 3 months. The Wolinsky 2007 study included primary progressive multiple sclerosis randomized to GA or placebo (PBO) in a 3-year, doubleblind trial. 37 patients out of 943 have been confirmed relapses during the follow-up, suggesting that a small proportion of patients exhibited the progressive relapsing phenotype. The primary end point was an intention-to-treat analysis of time to 1- (entry EDSS ) or 0.5-point expanded disability status scale change (entry EDSS ) sustained for 3 months. The trial was stopped after an interim analysis by an independent data safety monitoring board indicated no discernible treatment effect on the primary outcome. The main clinical data of the Progressive patients are reported in the Table 3.; the patients were more disable than RR MS and had a longer disease duration. CLINICAL OUTCOMES The studies on RR MS reported as primary outcome measures: Proportion of relapse-free patients at the end of follow-up (Bornstein 1987), mean number of relapses (Johnson 1995), total number of enhancing lesions on T1-weighted MRI images (Comi 2001), the total number of confirmed relapses (Filippi 2006). Studies on RR MS also evaluated the following secondary (and tertiary) endpoints: time to progression (Bornstein 1987), proportion of patients with sustained disease progression (Johnson 1995),change in EDSS scores from baseline (Johnson 1995; Bornstein 1987; Filippi 2006) and area under curve for the EDSS change (Filippi 2006) time to walk and ambulation index (Filippi 2006) relapse rate (Bornstein 1987; Comi 2001), number of relapses (Comi 2001), proportion of relapse-free patients (Johnson 1995; Comi 2001,Filippi 2006 ), time to first relapse after randomisation ( Comi 2001,Filippi 2006 ) the proportion of patients with two or more relapses (Comi 2001, ), steroid courses (Comi 2001 ;Filippi 2006 ) and relapse-related hospitalizations (Comi 2001,Filippi 2006 ) and other MRI measures (Comi 2001: Filippi 2006). MRI data of Johnson 1995 s study were reported in 135 out of the 251 patients of the original cohort in the open -label extension trial (Wolinsky 2001). Progression was defined in all studies as an increase of at least 1 point EDSS maintained for at least 3 months (Bornstein 1987; Johnson 1995). It is noteworthy that the review protocol was 9

12 more conservative, requiring at least 6 months of sustained 1-point EDSS worsening to be classified as progression, even if other definitions could be accepted. As a separate endpoint from progression, 2 trials analysed the proportion of patients worsened by at least 1 point in disability score at the end of follow-up as compared to baseline (Bornstein 1987; Johnson 1995). It assumed that this endpoint does not take into account if a sustained increase in EDSS score has occurred, and it is open to misinterpretations as to the final patient outcome. Therefore, we have chosen not to analyse clinical worsening as reported by these studies in order to avoid misleading results when inconsistent with those obtained in disease progression (see Discussion). Consistently, clinical improvement based on a 1 point decrease in EDSS score versus baseline was not analysed. Relapse was defined as the appearance or reappearance of one or more neurologic symptoms, with signs persisting for at least 48 hours and immediately preceded by a relatively stable or improving neurologic state of at least 30 days (Johnson 1995; Comi 2001; Filippi 2006 ). Another trial protocol required that patient symptoms were associated with changes in the neurologic exam involving an increase of at least 1 point in any of the 8 Kurtzke functional groups. Sensory symptoms alone were not considered (Bornstein 1987).The relapse was confirmed when the symptoms were accompanied by objectives changes corresponding to an increase of 0.5 EDSS or 1 grade in the two or more functional systems (Comi 2001; Filippi 2006), The studies on Progressive MS reported as primary outcome measures: time to sustained confirmed at 3 months, of 1 point of EDSS increase (according to baseline EDSS of 5.0 or more) (Bornstein 1991), of 1.5 EDSS increase ( Baseline EDSS less than 5) (Bornstein 1991) or 1 (basal EDSS 3-5) and 0.5 (basal EDSS 5.5 or more) ( Wolinsky 2007). as secondary outcome measures unconfirmed progression and progression of 0.5 EDSS units (Bornstein 1991) and proportion of progression free, changes from baseline in mean EDSS score and mean MSFC scores and MRI measures (Wolinsky 2007) SIDE EFFECTS AND ADVERSE EVENTS The number of patients experiencing side effects of treatment have been counted, by event, in all studies. However, information on how many patients reported at least one adverse event whatever was unavailable, so that the overall incidence of side effects could not be calculated. The number of patients who dropped out because of adverse effects could be extracted from studies (Bornstein 1987; Johnson 1995; Comi 2001: Wolinsky 2007). SECONDARY ENDPOINTS Two studies have compared the number of hospitalisations observed at the end of follow-up between glatiramer acetate and placebo arms (Johnson 1995; Comi 2001). Number of relapses requiring hospitalisation was also evaluated in Filippi s study (Filippi 2006) but that data were not shown. Data on the number of steroid courses administered were also available from two studies (Bornstein 1991; Comi 2001). MRI PARAMETERS One study (Comi 2001) evaluated the total number of enhancing lesions on MRI as the primary endpoint, clinical outcomes being analysed as tertiary endpoints. Secondary outcomes of this trial were: total volume of enhancing lesions, number of new enhancing lesions, number of new lesions on T2-weighted images, percentage change of lesion volume on T2-weighted images, change in the volume of hypointense lesions on T1-weighted images. MRI parameters were also analysed in secondary reports from the US phase III pivotal study, both for a small subset of the main trial (Ge 2000) and the open-label extension phase (Wolinsky 2001). CONCOMITANT MEDICATION In two studies, standard steroid treatment could be administered during relapses, without restrictions (Bornstein 1987; Johnson 1995). Moreover, symptomatic medications (Bornstein 1987) or conventional therapy received at the time of randomisation (Johnson 1995) could be maintained throughout the study. A standard treatment schedule for relapses was specified in one trial protocol as 1.0 g i.v. methylprednisolone for three consecutive days (Comi 2001). Limitations to the use of steroids were introduced in the CP study (Bornstein 1991), where the maximum dose should not exceed 100 mg prednisone or 80 UI ACTH daily during exacerbations, lasting no more than four weeks. Risk of bias in included studies (summary data are reported in Figure 1 and Figure 2) 10

13 Figure 1. Methodological quality summary: review authors judgements about each methodological quality item for each included study. 11

14 Figure 2. Methodological quality graph: review authors judgements about each methodological quality item presented as percentages across all included studies. RANDOMISATION Method of randomization are reported in risk of bias tables (see tables of characteristics of included studies).allocation concealment was adequate in four studies Bornstein 1991; Johnson 1995; Comi 2001; Filippi 2006 ) and not reported in one study (Wolinsky 2007) In another study (Bornstein 1987) patients were randomised within matched pairs, but the method to obtain treatment allocation was not clearly specified. Allocation concealment was therefore defined as unclear for this report. BLINDING All trials were double-blind in design. However, the occurrence of peculiar side effects of glatiramer acetate (e.g.: injection site and skin reactions) casts doubts on the possibility to ensure a reliable masking. In the attempt to reduce this flaw, all study protocols introduced a separate evaluation by two independent physicians: an examining neurologist was responsible for the scheduled monitoring of clinical endpoints, while a treating physician was in charge of managing side effects and concomitant therapy. The latter physician could be either aware (Bornstein 1987; Bornstein 1991;Filippi 2006 ;Wolinsky 2007) or unaware (Johnson 1995) of patient allocation. In another study, blinding of physicians was not formally assessed because clinical endpoints were only considered as tertiary outcomes (Comi 2001). Independently of investigators accuracy, it can be assumed that all trials failed to carry out a fully blind assessment. In one study claimed to be double blind (Bornstein 1987), both patients and physicians correctly identified 70 to 80% of treatment allocations. Surprisingly, however, investigators stated that the ability to guess treatment correctly was influenced by the effect of treatment rather than by side effects. WITHDRAWALS AND LOST TO FOLLOW-UP Bornstein et al. (Bornstein 1987) report that two patients out of 25 allocated to placebo discontinued the study and were excluded from the analysis because of unreliable data due to an altered psychological profile. This was considered as a violation of the intention-to-treat analysis. Therefore, we had to count 23 participants in the placebo arm when data were extracted from either percentages or means in the original paper. Data from other five patients who dropped out were analysed, two in the placebo arm and three allocated to glatiramer acetate. One exacerbation and two adverse events were counted in this group. The US pivotal trial (Johnson 1995) counted 19 withdrawals in glatiramer acetate-treated patients and 17 among those taking placebo. Causes of discontinuation were not reported in 10 glatiramer acetate-allocated patients and 14 controls, representing 9.6% of the randomised sample altogether. Out of 215 patients who completed the first 24-month follow-up, 12 refused to enter the 11-month extension having opted to receive the newly emerging beta-interferon therapy. The two-year clinical profiles exhibited by these patients and those enrolled in the extension trial were comparable. A further nine subjects dropped out at the end of the 35-month follow-up (three in the treatment arm, seven allocated to placebo). All data related to this group were included in the analysis, although causes of dropout are not reported in detail. The European/Canadian trial (Comi 2001) had 14 dropouts, equally balanced between treatment and placebo. All of them where included in the analysis. The oral study (Filippi 2006) had 14,12,13% of withdrawn in the three experimental groups. 12

15 The CP MS study also reported a balanced withdrawal pattern (Bornstein 1991), with 10 glatiramer acetate treated patients and 10 controls discontinuing medication. Early withdrawals were all included in the analysis: 17 were censored at the time of discontinuation, the other 3 (glatiramer acetate=2, placebo=1) being counted as confirmed progression. In the Wolinsky 2007 study 188/627 GA and 98/316 Placebo treated patients withdrew for various reasons before sponsor decision for trial termination. At the end of follow-up only 114/621 (GA) and 46/314 (P) were available for the analysis of the main outcome (See Fig 2 of the paper). Four GA and 7 death Placebo - treated were also reported. VALIDITY SCORE The Jadad score was calculated as a measure of internal validity. The Jadad score is reported in the additional table (Table 1). One study was given three because of unclear allocation concealment and insufficient details on withdrawn patients and unsuccessful blinding (Bornstein 1987).One study was given three because of unclear allocation concealment and insufficient details on blindness (Wolinsky 2007) The others studies obtained a full score. Effects of interventions See: Summary of findings for the main comparison Glatiramer acetate versus placebo in relapsing remitting patient for multiple sclerosis PRIMARY OUTCOMES The efficacy of GA versus placebo was evaluated separately in RR and Progressive MS patients A total of 3233 patients, 2184 affected by RR (1365 actively and 819 placebo treated) and 1049 by Progressive MS (678 actively and 371 placebo treated) were included in these trials, although only 540 RR patients and 1049 PMS contributed to the analysis of treatment efficacy. Relapsing Remitting MS PATIENTS WHO PROGRESSED Information about progression of disability was available from two trials and 226 patients (Bornstein 1987; Johnson 1995).The risk of progression was not significantly modified by the therapy at 2 years : 0.75 (95% CI [0.51, 1.12] p=0.16) and at 35 months 0.81 (95% CI [0.50 to 1.29] (Figure 3). Figure 3. Forest plot of comparison: 1 Glatiramer acetate versus placebo in relapsing remitting patient, outcome: 1.1 Patients who progressed. 13

16 CHANGE IN DISABILITY SCORE Mean changes in EDSS disability score were calculated in two trials (Bornstein 1987; Johnson 1995). As different follow-up durations are available from the US phase III trial, both 24- and 35-month data are shown although results are not pooled. A slight decrease in EDSS score favouring glatiramer acetate is observed at two years (WMD= -0.33, 95% CI [-0.58 to -0.08], p = 0.009) and at 35 months (WMD= -0.45, 95% [-0.77 to -0.13], p = 0.006) (Figure 4). Figure 4. Forest plot of comparison: 1 Glatiramer acetate versus placebo in relapsing remitting patient, outcome: 1.2 Change in disability score at the end of follow-up. PATIENTS RELAPSE-FREE This information was available in three studies and 255 subjects with RR MS, evaluated at different follow-up lengths (Bornstein 1987; Johnson 1995; Comi 2001). Results have been split into three time windows: within 1 year (which includes the 9-month assessment reported in the European/Canadian study), at 2 years and at 35 months. Relative risks of experiencing no exacerbation were, respectively: 1.28 (95% CI[1.02, 1.62] p= 0.03) within 1 year of treatment, and 1.39 (95%C I[0.99, 1.94], p=0-06 at 2 years, and 1.33 (95% CI [0.86, 2.06] at 35 months ( Figure 5). Since the same study appears in more than one stratum (Johnson 1995), no pooled analysis is provided for this outcome. Significant heterogeneity was found between Bornstein s pilot trial and the European/Canadian study (p=0.03), possibly related to different trial duration. Then we tested pooled relative risk of relapse within 1 year of randomisation in a random-effect model, without any significant difference between glatiramer acetate and placebo: relative risk = 0.64 (95% CI [0.31 to 1.34], p= 0.2). MEAN NUMBER OF RELAPSES 14

17 Figure 5. Forest plot of comparison: 1 Glatiramer acetate versus placebo in relapsing remitting patient, outcome: 1.3 Patients relapse free. A significant reduction was found at 1 year (-0.35) at 2 years (-0.51) and at 35 months (-0.64). However a significant heterogeneity was found between the studies( Figure 6). 15

18 Figure 6. Forest plot of comparison: 1 Glatiramer acetate versus placebo in relapsing remitting patient, outcome: 1.4 Mean number of relapses. RELAPSE-FREE SURVIVAL Median time to first relapse was analysed in one study (Johnson 1995), with a median time of 287 days in patients treated with glatiramer acetate and 198 days in controls (Weibull regression model, p =0.097). Our elaboration on individual patient data extracted from the pilot trial paper (Bornstein 1987) point to a median of 5 months (95% CI [2 to 8]) in the placebo arm, while the median of glatiramer acetate-treated group could not be calculated as more than 50% of those subjects were censored without relapse at 24 months (log-rank chi-square = 6.68, p = ). These results could not be combined. ORAL TREAMENT WITH GA This treatment was considered only by one study (Filippi 2006 ): the available data did not allowed a meta-analysis according to the predefined protocol. The cumulative number of confirmed relapses did not differ between the two active treatment groups and the placebo group. Relative to placebo, the rate ratio for the 50 mg glatiramer acetate treated group was 0.92 (95% CI , p=0.30) and for the 5 mg glatiramer acetate treated group was 0.98 ( , p=0.76). No drug effect was seen for any of the secondary and tertiary endpoints. Progressive MS PATIENTS WHO PROGRESSED This information was available in two studies (Bornstein 1991; Wolinsky 2007) including 832 patients. Risk of progression was not reduced by GA at 1 year (0.88 (95% CI 0.60,1,27) at 2 years ( 0.84 ( 0.60,1.19) and 3 years 0.75 (0.38,1.50) (Figure 7).The data must be considered with caution since they were obtained from the survival curve, because not clearly reported in the paper. 16

19 Figure 7. Forest plot of comparison: 4 glatiramer acetate versus placebo in progressive patients, outcome: 4.1 progression of disability. CHANGE IN DISABILITY SCORE This information was available only from one study (Wolinsky 2007) including 943 cases. Mean EDSS scores increased from baseline by 0.61+/-1.13 in the placebo group and by 0.58+/-1.00 point in the GA group (not statistically different) (data unshown). CHANGES IN QUALITY OF LIFE SCORES No study planned to analyse patient quality of life as an outcome measure. ADVERSE EFFECTS All trials evaluated adverse events, accounting for 407 to 646 patients. Two studies (Johnson 1995; Comi 2001) mainly focused on injection-site changes and patterned transient systemic reactions, while the other two (Bornstein 1987; Bornstein 1991) reported a more analytical list of all observed side effects. Patterned reactions were most commonly reported, consisting of a transient self-limiting combination of: flushing, chest tightness, sweating, palpitations, anxiety. These symptoms unpredictably occurred within minutes of injection and spontaneously resolved before 30 minutes. Patterned reactions were more often observed in glatiramer acetate treated patients, with a relative risk of 3.27 (95% CI[2.07, 5.16]p < ]). Other systemic side effects significantly related to glatiramer acetate administration were: palpitations (relative risk = 3.58 [1.16, 11.06], p =0.03)., dyspnoea 3.58 [1.16, 11.06] p The incidence of headache, anxiety, faintness, drowsiness, cramps, joint pain, appetite loss, constipation, abdominal discomfort, nausea and vomiting was not significantly different between groups. Rash was more common in placebo treated patients. Local injection-site reactions included any of the following: itching (relative risk = 8.28 [4.99, 13.73] p < ]), swelling (relative risk = 4.01 [2.67, 6.03], p < ]), redness or erythema (relative risk = 4.58 [3.58, 5.88], p <0.0001]) and pain (relative risk = 2.46 [2.05, 2.95], p < ]). No adverse events leading to patient s death or major toxicity were reported. One study (Comi 2001) mentioned the occurrence of serious adverse experiences in 10 glatiramer acetate treated and 6 placebo patients, respectively, but these unspecified events were classified as unrelated to treatment. Side effects causing treatment discontinuation were observed in three trials (Bornstein 1987, Johnson 1995, Comi 2001), but their relation with glatiramer acetate is not definitely established (relative risk = 1.44 [0.94, 2.23], p=0.10] (Figure 8). 17

20 Figure 8. Forest plot of comparison: 3 Glatiramer acetate versus placebo: adverse effects, outcome: 3.1 Localised to the injection site. Side effects were similar in oral GA -treated and placebo patients, mainly involving the gastrointestinal and nervous system: headache,asthenia, pain, depression, accidental injury,paraesthesia, nausea,abdominal pain, arthralgia, back pain, diarrhoea, constipation, anxiety and dyspepsia (Filippi 2006). SECONDARY OUTCOMES HOSPITALISATIONS AT THE END OF FOLLOW-UP Data from hospital admission rates at nine or 35 months were extracted from two studies and 449 patients [Comi 2001; Johnson 1995]. Hospitalisations were significantly decreased in the glatiramer acetate group: relative risk = 0.60 (95% CI [0.40 to 0.91, p = 0.02]) ( Figure 9). 18