Discussion
This nationwide cohort followed a population of 852 patients with relapsing–remitting MS who initiated treatment with GA over a period of 5 years. Around one-third of these patients were still under GA treatment at the end of the 5-year period. At the last study visit, confirmed worsening of disability was documented in 31.9% of patients, evolution to secondary progressive MS in 13.4%, whereas 37.2% had no documented exacerbation during the study. Clinical outcome was somewhat better in the subgroup of patients continually treated with GA for 5 years (27.6% with confirmed worsening of disability, 8.1% with evolution to SPMS, but 47.5% with no documented exacerbation).
The strengths and limitations of this study largely reflect the naturalistic nature of the study design. The absence of protocol-specified study procedures enabled a large number of patients to be included in a range of settings from specialist MS centres to community neurology practices, and followed up for 5-year. To the extent that the practice of participating neurologists is representative of all MS care in France, which cannot be evaluated, the study provides an accurate picture of actual treatment practices in France. This naturalistic design also brings a certain number of disadvantages for the interpretation of the study. In particular, there is no specified follow-up schedule, leading to a loss of precision for determining the disease course (exacerbation rate, rate of change of EDSS), with a potential risk of bias occurring if patients decide to consult their neurologist only in case of new manifestations of disease. For the same reason, adverse events may be under reported. Given these limitations, the findings should be interpreted conservatively. Apart from infrequent neurologist consultations, another deviation from current practice standards is the absence of information on MRI. At the time the study was designed, there were no consensus standards for MRI in France, and routine use of MRI was essentially restricted to specialist MRI centres. It should be noted that the pivotal clinical trial of GA did not measure MRI outcomes either [
10]. The lack of MRI data is a limitation of the study, since it precludes a comprehensive evaluation of residual disease activity. Nonetheless, it is probable that clinical standards have improved over the decade since recruitment into this patient cohort, with more frequent, systematic and comprehensive follow-up of patients. As such, the findings will be of interest as a benchmark for the evolution of standards of care in future long-term prospective naturalistic studies of patients with MS treated with other DMTs.
Clinical outcome following initiation of GA observed in the registry is very close to what has been described previously in the long-term follow-up of the phase III pivotal trial [
11]. For example, the annualised exacerbation rate in the fifth year after starting treatment was 0.28 exacerbation/year in our cohort and 0.25 exacerbations/year in the clinical trial extension. The proportion of patients without confirmed worsening of disability was 68% in our study and 58% in the clinical trial extension [
11]. These findings suggest that the benefits of GA treatment observed in the context of an interventional clinical trial with a structured patient follow-up protocol in well-selected patients can also be achieved under conditions of everyday care in unselected patients. Similar findings of relatively stable disease (75% of patients without EDSS worsening) have also been reported from a retrospective Spanish cohort of patients treated with GA for at least 5 years [
12]. The adverse drug reactions documented over the course of the study were not unexpected for GA [
10,
13], corresponding principally to local injection site reactions or systemic immediate post-injection reactions.
A relatively high proportion of patients were lost to follow-up before the end of the 5-year study period (30%). This principally occurred due to the patient no longer consulting the participating neurologist and suggests that measures to ensure continuity of care for these patients with a chronic disease taking long-term DMT therapies would be useful. This relatively high attrition rate does, however, compromise the precision with which we can determine outcome in the overall cohort due to the risk of attrition bias. Another source of missing information concerns the patients for whom the EDSS score at inclusion was measured close to an exacerbation, and thus were excluded from the analysis of evolution of disability.
Thirty percent of patients enrolled into the cohort were treated with GA for the full 5-year period of follow-up. The median treatment duration was 3.4 years. Published reports on-treatment persistence with GA have reported a very wide range of findings, with median treatment durations ranging from 1.7 years [
14] to 9.2 years [
4], and our study falls within this range. A recent study of over 15,000 patients in the French national MS registry reported a 2-year persistence rate of around 60% for all injectable first-line DMTs [
15].
Since 2006, care of patients with MS has been orientated towards multidisciplinary MS Centres, of which there are 18 in France. All these centres participated in our study, although they only recruited one-third of the patients enrolled. This may reflect the relatively recent establishment of the reference studies when the study started. To evaluate the impact of potential differences in standards of care or in patient evaluation (for example, scoring of the EDSS), we performed a subgroup analysis comparing outcome in patients treated in reference centres and in these treated elsewhere. Exacerbation-related outcomes were similar in all three groups. No statistically significant differences in outcome were observed. The largest difference observed between the three groups was the difference between subjective (physician judgement) and objective (decision rules) methods of identifying patients who had evolved to SPMS. Compared to the period of the study, most patients with MS are now expected to be followed in MS reference treatment centres.
The demographic and clinical characteristics of the included population are close to those reported in other relevant populations. Notably, the age at diagnosis and gender ratio are similar to those reported in recent epidemiological surveys of MS in France [
16,
17]. Moreover, age and disease duration at start of treatment, gender ratio, EDSS score and pre-treatment exacerbation rates were generally similar to those reported in other cohorts of patients starting GA or interferon-β in Spain [
6], Britain [
18] and Canada [
4]. The rather long disease duration when GA was initiated (8 years) may reflect the fact that the study was initiated shortly after GA was first made available in France, and there was a reservoir of patients who could not be treated by interferon-β awaiting a new treatment. It would be expected that the time between MS onset and treatment initiation would now be much shorter, in accordance with current treatment guidelines.
The principal reasons for treatment discontinuation were inadequate efficacy, local tolerability issues and personal convenience. For patients who discontinued for inadequate efficacy, around half escalated to a second-line treatment, principally natalizumab. The relatively limited recourse to natalizumab (which only became available after the cohort had begun, in 2007) may reflect the requirement that this treatment be provided in MS reference centres only and the fact that most enrolled patients were managed by community-based neurologists. In addition, the study covered the period between the identification of progressive multifocal leucoencephalopathy as an adverse drug reaction to natalizumab and the time when the risk stratification programme for natalizumab had demonstrated its utility; during this period, physicians and patients may have been reluctant to use natalizumab. Patients discontinuing GA due to a tolerability issue were most frequently switched to an interferon-β. Switches to treatments other than natalizumab or interferon-β accounted for < 12% of GA discontinuations. An unanticipated finding was that around half of patients who discontinued GA received no alternative DMT. In part, this may be explained by patients evolving to secondary progressive MS for whom no established DMT existed. These patients did indeed discontinue GA more frequently than patients who remained in a relapsing–remitting phase, but only account for 20% of all discontinuations. It should also be noted that no further treatment options existed for patients who had been treated with an interferon-β before GA, who constituted the majority of patients stopping GA, and for whom natalizumab was not available. It is possible that the imminent arrival of oral therapies encouraged patients to stop GA and await an oral treatment. Fingolimod was the only oral treatment introduced onto the French market during the course of the study (in 2011) as a second-line therapy for patients whose disease remains highly active in spite of adequate treatment with a first-line DMT. Nonetheless, between 2010 and 2014, after the introduction of oral DMTs, around one-third of patients in the national OFSEP patient registry for MS were not prescribed any DMT [
15]. A high proportion of untreated patients who discontinue GA or interferon-β without another DMT being introduced has also been reported in a large Canadian cohort [
4]. In general, clinical outcomes were worse in patients who discontinued GA than in those taking GA continuously, although it is not possible to determine the direction of any causality.
A number of variables were identified as being associated with a higher probability of treatment discontinuation. These include greater disability (EDSS score) at inclusion, more exacerbations prior to inclusion, younger age at inclusion and not in employment (not working due to MS or being a housewife). Younger age and disability have also been observed to be associated with a greater risk of discontinuation in the Canadian [
4] and Catalan [
19] prospective cohorts, as well as, for age, in retrospective studies of prescription claims databases [
20,
21]. With respect to the association with pre-treatment disease activity, this may be explained by the poorer prognosis of patients with more active or advanced disease [
22], leading to a higher probability of discontinuation for poor treatment response. For the association with age, it has been suggested by others that treatment failure may be easier to detect in younger patients [
4].
Variables independently associated with clinical response were previous exacerbation history (patients with more exacerbations prior to inclusion had a less favourable prognosis) and educational level. The finding that patients with more active disease prior to inclusion respond less well to DMTs has been observed in a number of other cohorts of patients treated with interferon-β [
22‐
24]. The relationship between response and educational level is more surprising, and we can only hypothesise that lower education levels may be associated with poor adherence to therapy, which in turn, is associated with a lower probability of response.
In conclusion, this study demonstrated that clinical outcome in MS patients treated with GA in everyday clinical care in France was close to that previously demonstrated in interventional clinical trials. Around one-third of patients took GA continuously for 5 years and in general had a favourable outcome. Many patients discontinuing GA did not receive any alternative treatment. This suggests that more specific practice guidelines are needed to guide decisions about discontinuing and switching DMTs; this would be particularly timely given the availability of oral treatments in France and the consequent increase in treatment choice.
Acknowledgements
This study was funded by TEVA. H. Abboud, Paris; F. Abdul Samad, Châteauroux; A. Abdulnayef, Villemombe; A. Al Khedr, Amiens; H. Alchaar, Nice; J. Amevigbe, Beauvais; G. Angibaud, Montauban; O. Anne, La Rochelle; M.-S. Artaud-Uriot, Angoulême; G. Ast, Manosque; J. Augustin, Bois Guillaume; M. Aupy, Bordeaux; C. Azais-Vuillemin, Toulouse; M. Bailbe, Perpignan; P. Barbaste, Douai; P. Barres, Nice; A. Belhadia, Calais; R. Benrabah, Paris; O. Berets, Clamart; F.-X. Bergouignan, Bayonne; T. Bierme, Meyzieu; F. Bille-Turc, Marseille; S. Blanc, Lyon; C. Boisselier, Calais; I. Bonnet, Amiens; J.-P. Borsotti, Dijon; C. Bossu Van Nieuwenhuyse, Trélazé; C. Bouchard, Boulogne-Billancourt; A. Bouchareine, Paris; G. Boudouresques, Marseille; J.-M. Boulesteix, Cahors; P. Boulu, Clichy; B. Bourghol, Saint-Quentin; D. Brassat, Toulouse; A. Bredin, Blaye; G. Breteau, Lens; W. Camu, Montpellier; B. Carlander, Montpellier; W. Casseron, Aix-en-Provence; G. Castelnovo, Nîmes; S. Chapuis, Montluçon; J.-P. Chartier, Rodez; G. Chauplannaz, Lyon; F. Chaury, Illkirch-Graffenstaden; C. Cherlet, Berck; M. Clanet, Toulouse; P. Clavelou, Clermont-Ferrand; F. Clerc, Montbéliard; R. Colamarino, Vichy; P. Contis, Ramonville; C. Couratier, Aix-en-Provence; M. Coustans, Quimper; C. Crauser, Saint-Quentin; N. Daluzeau, Lisieux; J.-M. De Bray, Angers; T. De Broucker, Saint-Denis; V. De Burghgraeve, Rennes; P. De Facq, Armentières; J. De Sèze, Strasbourg; M. Debouverie, Nancy; G. Defer, Caen; I. Degaey, Dunkerque; S. Delassaux, Epinal; J. Delorme, Lyon; B. Denis, Marseille; N. Derache-Belpame, Caen; O. Dereeper, Calais; S.-A. Dereux, Saint-Pol-sur-Mer; F. Derouiche, Mulhouse; P. Devos, Boulogne-sur-Mer; A.-M. Deyrolle, Saint Palais; J. Dib, Metz; M. Dib, Paris; M. C. Dourneau-Lethiecq, Tours; M. Doux, La Roche -sur-Yon; S. Dufourd-Delalande, Tourcoing; M. Dujardin, Evreux; A. Dunac, Nice; G. Durand, Nice; P. Dussaux, Cergy; P. Eck, Le Havre; A. Engles, Roubaix; M.-P. Feltin, Echirolles; A. Ferrier, Clermont-Ferrand; M.-C. Fleury, Strasbourg; P. Gaida, La Teste-de-Buch; O. Gal, Thionville; A. Gayou-Joyeux, Dax; A. Gentil, Dijon; G. Geraud, Toulouse; J. Gere, Chambéry; J.-C. Getenet, Saint-Etienne; C. Giannesini, Paris; E. Giraud, Chambéry; P. Giraud, Pringy; P. Gras, Dijon; G. Griffie, Tourcoing; C. Gueganno-Roche, Moulins; A.-M. Guennoc, Tours; L. Guilloton, Lyon; H. Guinot, Marseille; C. Guiraud-Chaumeil, Albi; C. Guiziou, Lannion; C. Gurau-Vasilescu, Dreux; N. Guy-Renouil, Clermont-Ferrand; J.-B. Hamon, Brest; O. Heinzlef, Poissy; C. Hemet-François, Rouen; J.-L. Henlin, Dole; P. Homeyer, Aubenas; P. Hourmant, Morlaix; J.-P. Hurtevent, Lille; O. Ille, Mantes-la-Jolie; E. Josien, Béthune; C. Juhel, Rennes; M. Kalafat, Paris; S. Kelfa-Testut, Elbeuf; A. Kopf, Paris ; C.-P. Kpade, Pontivy; M. Lalisse, Houilles; D. Laplaud, Nantes; J. Lapras, Saint-Etienne; J.-L. Larrieu, Agen; G. Lavernhe, Gap; A. Layet, Montivilliers; A. Le Bayon, Nîmes; P. Le Coz, Arras; C. Lebrun-Frenay, Nice; J. Leche, Vendôme; A. Legout, Le Mans; P. Lejeune, La Roche-sur-Yon; P. Lemarquis, Toulon; G. Level, Verdun; J. Lizeretti, Voiron; D. Locuratolo, Gap; M. Logak, Creil; P. Louchart, Douai; P. Louiset, Bordeaux; M. Lubeau, Brive-la-Gaillarde; A. Mackowiak, Lomme; L. Magy, Limoges; M. Maillet-Vioud, Montluçon; J. Mallecourt, Dreux; C. Mallecourt, Toulon; S. Marcel, Chambéry; I. Mari, Paris; P. Marrel, Freyming-Merlebach; J. Maupetit, Libourne; S. Medjbeur, Champigny-sur-Marne; D. Menard, Rennes; P. Meynieu, Lille; M.-C. Minot-Myhie, Rennes; B. Montagne, Roubaix; T. Moreau, Dijon; C. Moreau, Angers; A. Moulignier, Paris; P. Muh, Lorient; J.-B. Nkendjuo, Dunkerque; A. Nègre, Boulogne-sur-Mer; P. Neuschwander, Lyon; V. Neuville, Maubeuge; A. Nibbio, Roubaix; J. Ollier, Néris-les-Bains; B. Ondze, Mont De Marsan; J.- C. Ouallet, Bordeaux; N. Patte-Karsenti, Boulogne-Billancourt; J. Pelletier, Marseille; P. Perrotte, Le Havre; D. Pez, Paris; C. Pierrot-Deseilligny, Paris; J.-C. Pin, Saint-Michel; S. Pittion-Vouyovitch, Nancy; C. Popescu, Fécamp; A. Poujois, Saint-Etienne; A. Pouliquen, Saint-Aubin-sur-Scie; A. Pouyet, Saint-Brieuc; C. Rémy, Romans; C. Renglewicz-Destuynder, Colmar; G. Riche, Villeurbanne; L. Rieu, Clermont-Ferrand; C. Robin, Roanne; A. Robinson, Carcassonne; N. Rosey-Dufosse, Boulogne-sur-Mer; B. Roualdes, Champigny-sur-Marne; M.-H. Rougie, Toulouse; F. Rouhart, Brest; I. Ruggieri, Marseille; S. Saad, Chaumont; J. Saintarailles, Mérignac; A.-M. Salandini-Roque, Toulouse; M. Salzmann, Roanne; N. Schmidt, Rueil-Malmaison; J. Senant, Rouen; T. Sergeant, Quimper; V. Seyeux, Pontarlier; J. Siboni, Carcassonne; O. Simon, Paris; S. Stoquart-El Sankari, Amiens; G. Taurin, Saint-Malo; S. Thibault-Tanchou, Lille; J.-N. Tillier, Argenteuil; A. Tougeron, Poissy; L. Toureille-Borlet, Annemasse; P. Tourniaire, Avignon; H. Touzani, Dunkerque; S. Trefouret, La Seyne-sur-Mer; X. Vandamme, La Rochelle; J. Vaunaize, Reims; F. Viala, Toulouse; U.-C. Viola, Tours; J.-M. Visy, Reims; M. Vlaicu-Bustuchina, Paris; F. Vuillemet, Colmar; V. Wattier, Saint-Julien-en-Genevois; A. Wavreille, Cambrai; K. Wegener, Granville; S. Wiertlewski, Nantes; C. Zaenker, Colmar; H. Zephir-Thi, Lille; F. Ziegler, Belfort.