Original Research Paper Multiple Sclerosis Journal— The ACROSS study: Long-term efficacy of Experimental, Translational and Clinical fingolimod in patients with relapsing–remitting January–March 2020, 1–10 DOI: 10.1177/ multiple sclerosis 2055217320907951 ! The Author(s), 2020. Article reuse guidelines: sagepub.com/journals- T Derfuss , J Sastre-Garriga , X Montalban, M Rodegher, J Wuerfel, L Gaetano , D Tomic, permissions A Azmon, C Wolf and L Kappos Abstract Background: In chronic diseases such as multiple sclerosis requiring lifelong treatment, studies on long-term outcomes are important. Objective: To assess disability and magnetic resonance imaging-related outcomes in relapsing multiple sclerosis patients from a Phase 2 study of fingolimod 10 or more years after randomization and to compare outcomes in patients who had a higher fingolimod exposure versus those with a lower fingo- limod exposure. Methods: ACROSS was a cross-sectional follow-up study of patients originally enrolled in a Phase 2 fingolimod proof-of-concept study (NCT00333138). Disability and magnetic resonance imaging-related outcomes were assessed in patients grouped according to fingolimod treatment duration, based on an arbitrary cut-off: 8 years (high exposure) and <8 years (low exposure). Results: Overall, 175/281 (62%) patients participated in ACROSS; 104 (59%) of these were classified “high exposure.” At 10 years, patients in the high-exposure group had smaller increases in Expanded Disability Status Scale (þ0.55 vs. þ1.21), and lower frequencies of disability progression (34.7% vs. 56.1%), wheelchair use (4.8% vs. 16.9%), or transition to secondary progressive multiple sclerosis (9.6% vs. 22.5%) than those in the low-exposure group. The high-exposure patients also had less progression in most magnetic resonance imaging-related outcomes. Conclusion: After 10 years of fingolimod treatment, disability progression was lower in the high- exposure group than in the low-exposure group. Keywords: Relapsing–remitting multiple sclerosis, fingolimod, oral therapy, long-term treatment, ambu- latory assistance, disability progression Date received: 3 July 2019; Revised received: 18 December 2019; accepted: 23 December 2019 Correspondence to: Introduction authorization. A systematic cross-sectional examina- T Derfuss, The chronic and unpredictable nature of the multiple tion of as many as possible of the initial participants Neurologic Clinic and Policlinic, Departments of sclerosis (MS) disease course poses a great challenge of a controlled study may provide valuable informa- Medicine and Biomedicine, for clinicians in routine clinical practice; monitoring tion about long-term outcomes. University Hospital, Basel, Switzerland. long-term outcomes is therefore important for treat- Tobias.Derfuss@usb.ch ment optimization.1,2 Observations in patients par- Results from the 6-month controlled Phase 2 proof- ticipating in pre-registration studies are particularly of-concept study in patients with relapsing MS valuable because these patients are well character- (NCT00333138) and its extensions over 1, 3, 5, ized and, by definition, have a longer exposure to the and 73–6 years showed that fingolimod is associated drug than those who start treatment after marketing with an early and sustained effect on disease activity. Creative Commons Non Commercial CC BY-NC: This article is distributed under the terms of the Creative Commons Attribution-NonCommercial 4.0 License (https://creativecommons.org/licenses/by-nc/4.0/) which permits non-commercial use, reproduction and distribution of the work without further permission provided the original work is attributed as specified on the SAGE and Open Access pages (https://us.sagepub.com/en-us/nam/open-access-at-sage). Multiple Sclerosis Journal—Experimental, Translational and Clinical T Derfuss, Three large Phase 3 clinical trials have established as measured by the mean change of the Expanded Neurologic Clinic and Policlinic, University the superior efficacy of fingolimod over intramuscu- Disability Status Scale (EDSS) from baseline to 10 Hospital and University of lar interferon beta-1a (IFN b-1a)7 or placebo8,9 on years, compared with low exposure. All EDSS Basel, Switzerland clinical and magnetic resonance imaging (MRI)- scores were obtained by trained, certified assessors J Sastre-Garriga, Multiple Sclerosis Centre of related outcomes, including brain atrophy. using the Neurostatus scoring tools for a standard- Catalonia (Cemcat), However, continuous extension studies face increas- ized, quantified neurological examination. The tools Department of Neurology/ Neuroimmunology, Hospital ing logistic and attrition problems with increasing for scoring and certification can be found online Universitari Vall d’Hebron, duration. In the ACROSS study, we assessed clinical Universitat Autònoma de (www.neurostatus.net). For patients not able or will- Barcelona, Barcelona, Spain disability- and MRI-related outcomes in participants ing to undergo the EDSS scoring in person, a vali- X Montalban, of the Phase 2 study of fingolimod 10 or more years dated EDSS phone interview was done.13 Multiple Sclerosis Centre of Catalonia (Cemcat), after randomization. Department of Neurology/ Secondary assessments Neuroimmunology, Hospital Universitari Vall d’Hebron, Methods Universitat Autònoma de Disability outcomes. Secondary objectives included Barcelona, Barcelona, Spain; the proportion of patients with confirmed disability Division of Neurology, St Study design and patient population Michael’s Hospital, In a 6-month, double-blind, placebo-controlled progression, the proportion of patients with an EDSS University of Toronto, multi-center Phase 2 study, conducted from May score <6, and the frequency and time to first Toronto, Canada M Rodegher, 2003 to September 2004 (core study), the efficacy required use of an ambulatory device or wheelchair. MS Centre, IRCCS Santa of fingolimod in patients with relapsing MS For patients under continuous care of the respective Maria Nascente, Fondazione Don Carlo Gnocchi, via (McDonald criteria)10 was evaluated.3 ACROSS study centers, information about the time to event Capecelatro, Milan was a cross-sectional follow-up including patients outcomes was extracted from history and file J Wuerfel, originally enrolled in this Phase 2 study who had reviews and, where necessary, from other treating L Gaetano, Medical Image Analysis received 1 fingolimod dose (5.0 mg, 1.25 mg, or medical professionals. The proportion of patients Center Basel and Department 0.5 mg) during the core study and its extensions, for with disability progression was defined as an of Biomedical Engineering, University Hospital, a 10-year assessment regardless of their current increase in EDSS score by 1.5 (from baseline score Switzerland treatment status. 0), or 1 (from baseline score 1–5), or 0.5 (from base- D Tomic, line score >5). Paired comparison of the MS A Azmon, Novartis Pharma AG, Basel, After completing the Phase 2 extension studies, Functional Composite score (MSFC z-score) and Switzerland patients did not receive any protocol-specified treat- its components (9-Hole Peg Test (9-HPT), Timed C Wolf, ment. Original Phase 2 study sites that agreed to 25-Foot Walk Test (T25FWT), and Paced Auditory Lycalis, Belgium participate in the ACROSS study located the study Serial Addition Test (PASAT-3)) were done for L Kappos, Neurologic Clinic and participants from their center and asked them to those patients with assessments at baseline and Policlinic, Departments of year 10. Medicine, Clinical Research, return for a 10-year assessment. As defined in the Biomedicine and Biomedical study protocol, based on the duration of fingolimod Engineering, University Secondary progressive MS conversion. Conversion Hospital and University of treatment, returning patients were divided into high Basel, Basel, Switzerland and low exposure groups (i.e. those who had to secondary progressive MS (SPMS) was deter- received fingolimod for 8 or more years and for mined by the treating physician at the 10-year fewer than 8 years, respectively). We selected this follow-up according to the Lublin and Reingold arbitrary cut-off based on precedence from an earlier 1996 criteria.14 long-term follow-up study.11 MRI assessment. A standard MRI protocol was The study protocol was reviewed and approved by acquired at year 10 for comparison with the baseline the Independent Ethics Committee and Institutional MRI of the Phase 2 study. The MRI protocol includ- Review Board at each center as per local regulations. ed axial two-dimensional (2D) T1-weighted (T1w) All patients provided written informed consent images (3 mm slice thickness, no gap) before and before study entry. The study was conducted in com- after administration of a contrast agent, and 2D dual- pliance with the ethical principles of the Declaration echo T2w images in an identical slice position. of Helsinki and the International Conference on Harmonization Good Clinical Practice Guidelines.12 T2w lesion load was quantified on proton density images at baseline and 10 years, following a Primary assessment highly standardized procedure by a central core The primary objective of the study was to assess lab specializing in analyses of Phase 3 clinical whether high exposure to fingolimod (8 years) trials (Medical Image Analysis Center, Basel, was associated with reduced disability progression, Switzerland), and the number of new/enlarging 2 www.sagepub.com/msjetc Derfuss et al. T2w lesions (i.e. lesions visible at 10 years, but not follow-up, which accounted for an ascertainment at baseline) were determined. T1w lesions were also rate of 84% in the participating sites segmented with the same protocol at baseline and at (Supplementary Table 1). Overall, 104 (59.4%) 10 years. Percentage brain volume change (PBVC) and 71 (40.6%) patients were assigned to the high from core baseline to year 10 was also determined. and low exposure groups, respectively. The 106 patients who did not participate in the cross- Safety. During the site visit, patients were asked to sectional follow-up did not differ from the 175 par- report on any adverse events (AEs) and their files ticipating patients for age, race, or body weight at were reviewed for such events. Further, a physical core baseline; however, there was a trend for female examination was done and vital signs were measured preponderance in non-participants (p ¼ 0.0605). (data not presented). Non-participants had a similar disease duration but a higher EDSS score (p ¼ 0.0008) and a trend for a Statistical analyses higher T2 burden of disease (p ¼ 0.0575) at core The primary variable was analyzed using analysis of baseline (Table 1). The mean fingolimod exposure covariance (ANCOVA), where the change in the in the high exposure group was 4003 days, and 1190 EDSS score from baseline to 10 years was the days in the low exposure group. Details pertaining to response variable and the baseline EDSS score was the use of other disease-modifying therapies (DMTs) a continuous covariate, whereas fingolimod treat- are provided in Supplementary Table 2. Significant ment duration (high vs. low exposure) was a fixed differences in baseline characteristics between the effect. A sensitivity analysis with fingolimod treat- groups were observed for the proportion of women ment duration as a continuous variable was also per- (p < 0.05), mean EDSS scores (p < 0.05), and mean formed. Factors in the core phase predicting MS duration since the first symptom (p < 0.02; disability progression after 10 years were analyzed Table 2). with change in the EDSS score from baseline to 10 years as the response variable and baseline char- Disability outcomes acteristics such as gender, age, EDSS score, T2 At 10-year follow-up, EDSS was assessed in 172 lesion counts and volume, MSFC z-score, and fingo- patients (by phone, n ¼ 5). The mean EDSS scores limod treatment duration (high vs. low exposure cat- in the overall population increased from baseline of egories) as covariates. Adjusted least squares means core study by 0.83. EDSS increase was lower in the (LSMs) and their corresponding standard errors were high exposure group compared to the low exposure reported in addition to the adjusted mean differences group (LSM: 0.58 vs. 1.17; p ¼ 0.0155; Table 3). In and standard errors. Kaplan–Meier (KM) plots and the sensitivity analysis, longer exposure to fingoli- hazard ratios (HRs) with 95% confidence interval mod significantly correlated (p ¼ 0.0143) with a (CIs) using a Cox proportional hazard model were lower EDSS at 10 years. Furthermore, a significant reported for the first time use of an ambulatory correlation was observed between change in EDSS device and wheelchair and for patients who transi- from baseline to 10 years and longer exposure to tioned to SPMS. The ANCOVA model was also fingolimod treatment (p ¼ 0.0162) as well as age at used to analyze change in the volume of T2w lesions baseline (p ¼ 0.0116); no correlations were observed at 10 years. with the other covariates (gender, EDSS score, T2 lesion counts and volume, MSFC z-score). Fewer Correlations between fingolimod treatment duration patients experienced disability progression in the and MS disease outcomes (disability, conversion to high exposure group than in the low exposure SPMS and MRI lesions) were assessed using group at 10 years (34.7% vs. 56.1%; p < 0.01; Pearson correlation coefficients. Table 3). Overall, 144 of 172 (83.7%) patients had EDSS scores of <6 at 10 years, and this proportion Results was greater in the high exposure group than in the low exposure group (89.1% vs. 76.1%; p < 0.05). Patient disposition and baseline characteristics Mean (standard deviation (SD)) time to first use of Of the 32 centers that participated in the core Phase an assistive device for ambulation was longer in the 2 study, 26 recruited patients for ACROSS; of note, high exposure group compared with the low expo- all sites from Finland could not participate (n ¼ 20 sure group (8.3 (2.3) vs. 5.2 (3.1) years; Table 3). randomized patients from three centers). A total of The absolute number of patients who needed an 175 (62.3%) of the 281 patients originally random- assistive device for ambulation were lower in the ized in the Phase 2 study participated in this 10-year high exposure group (12/104 (11.5%) vs. 12/71 www.sagepub.com/msjetc 3 Multiple Sclerosis Journal—Experimental, Translational and Clinical Table 1. Patient demographics and characteristics at baseline of the core study and in patients who partic- ipated, or did not participate, in ACROSS. ACROSS study participation Yes No p value Core study Participants, n (%) 175 (62.3) 106 (37.7) 281 (100.0%) Sex, n (%) Female 117 (66.9) 82 (77.4) 0.0605a 199 (70.8) Age group, years, n (%) 30 47 (26.9) 26 (24.5) 0.2150b 73 (26.0) 31–40 64 (36.6) 36 (34.0) 100 (35.6) 41–50 49 (28.0) 27 (25.5) 76 (27.0) >50 15 (8.6) 17 (16.0) 32 (11.4) Mean (SD) age, years 37.4 (9.3) 38.5 (10.1) 0.4292c 37.8 (9.6) BMI, n 172 105 277 Mean, kg/m2 24.45 (4.3) 24.37 (4.4) 0.7692c 24.42 (4.4) Race, n (%) Caucasian 172 (98.3) 104 (98.1) 0.4061a 276 (98.2) Black 2 (1.1) 0 (0.0) 2 (0.7) Asian 1 (0.6) 0 (0.0) 1 (0.4) Duration of MS since first symptom, years Mean (SD) 8.5 (7.8) 9.0 (7.5) 0.3477c 8.7 (7.6) Median 6.7 7.9 7.0 Baseline EDSS Mean (SD) 2.3 (1.2) 3.0 (1.5) 0.0008c 2.6 (1.4) Median 2.0 2.5 2.0 Mean baseline T2 lesion 8584.4 (12,558.9) 10,598.9 (10,639.5) 0.0575c 9355.1 (11881.0) volume, mm3 a Chi-square test. b Mantel Haenszel chi-square test. c Wilcoxon rank sum test. BMI: body mass index; EDSS: Expanded Disability Status Scale; MS: multiple sclerosis; SD: standard deviation. (16.9%); KM analysis also showed similar SPMS conversion findings (Figure 1). The HR for the first time use A total of 26 (14.9%) patients were classified as of an assistive device for ambulation was 0.86 having SPMS at 10 years. A lower proportion of (95% CI: 0.37–2.01; relative risk reduction, 14%; patients in the high exposure group transitioned to p ¼ 0.7338) by treatment group (high vs. low expo- SPMS compared with the low exposure group (KM sure groups). estimates: 8.1% vs. 18.2%; Table 3). Furthermore, mean (SD) time to first classification as SPMS was Mean (SD) time to first use of a wheelchair was longer in the high versus low exposure group (5.4 slightly longer in the high exposure group than in (4.7) years vs. 3.4 (6.0) years; Table 3). In compari- the low exposure group (7.2 (2.5) vs. 6.2 (2.5) son to the low exposure group, the risk of developing years). The absolute numbers of patients who SPMS was 66% lower in the high exposure group needed a wheelchair were lower in the high exposure (HR (95% CI): 0.34 (0.12–0.92); p < 0.05; Table 3). group (5/104 vs. 12/71 for the low exposure group; Table 3). The corresponding KM curves are pre- MSFC z-score sented in Figure 2. In the high exposure group, the At core baseline, MSFC was assessed in 104 patients risk for need of a wheelchair was reduced by 76% from the high exposure and 69 patients from the low versus low exposure group (HR (95% CI): 0.24 exposure group. At follow-up, it was assessed in 89 (0.07–0.85; p ¼ 0.0276); Table 3). No patient in and 51 patients, respectively. The change from core the ACROSS study was bedridden. baseline in the MSFC z-score (mean (SD): 0.1 4 www.sagepub.com/msjetc Derfuss et al. Table 2. Patient demographics and baseline characteristics of patients with high and low fingolimod exposure at entry into the core study. High exposure Low exposure (N ¼ 104) (N ¼ 71) Sex, n (%) Female* 63 (60.6) 54 (76.1) Age group, years, n (%) 30 25 (24.0) 22 (31.0) 31–40 42 (40.4) 22 (31.0) 41–50 31 (29.8) 18 (25.4) >50 6 (5.8) 9 (12.7) Age, years 37.4 (8.5) 37.4 (10.6) BMI, n 102 70 Mean, kg/m2 24.78 (4.1) 23.96 (4.7) Race, n (%) Caucasian 103 (99.0) 69 (97.2) Black 1 (1.0) 1 (1.4) Asian 0 (0.0) 1 (1.4) Duration of MS since 7.4 (6.3) 10.1 (9.3) first symptom, years* Median 6.0 7.8 EDSS at baseline** 2.2 (1.2) 2.6 (1.2) Median 2.0 2.0 T2 lesion volume at baseline, n 96 67 mm3 7698.6 (8209.4) 9853.6 (16,951.7) *p < 0.05 and **p < 0.02 between the treatment groups. All other values were not significantly different. All values are represented as the mean (SD), unless otherwise specified. BMI: body mass index; DMTs: disease-modifying therapies; EDSS: Expanded Disability Status Scale; MS: multiple sclerosis; SD: standard deviation. (0.5) vs. 0.6 (1.3); p < 0.01) and PASAT-3 score (4216.2 mm3), although the difference did not (0.5 (8.3) vs. 5.4 (12.4); p < 0.001) at year 10 was reach significance. significantly in favor of the high exposure group. No significant differences between the two treatment The PBVC from core baseline at year 10 was signif- groups were observed for the T25FWT (1.3 (11.6) icantly lower in the high exposure group than in the vs. 3.9 (16.1)) or 9-HPT (2.3 (5.8) vs. 5.1 (14.5)) low exposure group (LSM: 9.20% vs. 10.01%). scores. Correlation analyses MRI outcomes Positive correlations were observed between the dura- Total T2w lesion volume increased from baseline in tion of fingolimod treatment and the time to first use both groups (Table 4). Mean increase in total T2w of an ambulatory device (r ¼ 0.46; p ¼ 0.0144) lesion volume was lower in the high exposure group whereas a negative correlation was observed with compared to the low exposure group (LSM increase the EDSS score at year 10 (r ¼ 0.27; p ¼ 0.0004). of 1031.5 mm3 vs. 3636.9 mm3; p < 0.05). No correlations were observed between the duration of fingolimod and other variables such as time to first At 10 years, the number of new or enlarging T2w use of a wheelchair (r ¼ 0.36; p ¼ 0.1573), time to lesions was also significantly lower in the high expo- first SPMS classification (r ¼ 0.31; p ¼ 0.1222), sure group than in the low exposure group (LSM: time to first documentation of an EDSS score of 10.7 vs. 22.2; p < 0.0001; Table 4). A lower T1w 6.0 (r ¼ 0.28; p ¼ 0.1935), changes in MRI meas- hypointense lesion volume was observed with high ures from the Phase 2 baseline to the 10-year assess- exposure (3357.6 mm3) than with low exposure ments, and with MRI parameters at 10 years. www.sagepub.com/msjetc 5 Multiple Sclerosis Journal—Experimental, Translational and Clinical Table 3. Disability outcomes at the 10-year follow-up. Fingolimod treatment duration High exposure Low exposure (N ¼ 104) (N ¼ 71) Primary assessment EDSS change from the core study baseline to the 10-year follow-up Core baseline n 104 71 Mean (SD) 2.16 (1.2) 2.60 (1.2) Year 10 n 101 71 Mean (SD) 2.72 (1.9) 3.81 (2.3) Change from baseline to year 10 n 101 71 Mean (SD) 0.55 (1.5) 1.21 (1.6) ANCOVA LSM (SE) 0.58 (0.154) 1.17 (0.185) Difference in LSMs 0.59 p-value 0.0155 Secondary assessments Disability progression,a n/m (%) 35/101 (34.7) 37/66 (56.1) p < 0.01 Patients with EDSS <6, n/m (%) 90/101 (89.1) 54/71 (76.1) p < 0.05 Time to first use of ambulatory assistive devices, years Mean (SD) 8.3 (2.3) 5.2 (3.1) Median 8.3 4.9 Time to first documented EDSS 6.0, years Mean (SD) 7.7 (2.5) 5.9 (3.6) Median 8.1 5.7 Time to first use of a wheelchair, years Mean (SD) 7.2 (2.5) 6.2 (2.5) Median 7.8 6.7 Proportion of patients in need of a wheelchair Kaplan–Meier estimate 4.9 16.9 HR (95% CI) 0.24 (0.07–0.85) p < 0.0276 Time to first classification as SPMS, years Mean (SD) 5.4 (4.7) 3.4 (6.0) Median 6.5 3.8 Proportion of patients classified as having N¼102 N¼67 SPMS over 10 yearsb Kaplan–Meier estimate 8.1 18.2 HR (95% CI) 0.34 (0.12–0.92) p < 0.05 a Disability progression is defined as an increase in EDSS score of 1.5 (from baseline score 0), or 1 (from baseline score 1–5), or 0.5 (from baseline score >5). b First classification as SPMS. Data represented as mean (SD) unless specified otherwise. ANCOVA: analysis of covariance; CI: confidence interval; EDSS: Expanded Disability Status Scale; HR: hazard ratio; LSM: least squares mean; n: number of patients in the treatment group; m: number of patients with non-missing values; SD: standard deviation; SE: standard error; SPMS: secondary progressive multiple sclerosis. 6 www.sagepub.com/msjetc Derfuss et al. participants and provides information about long- term outcomes of fingolimod treatment. Approximately 60% of patients enrolled in ACROSS were exposed to fingolimod for 8 years (high exposure group). These patients had lower dis- ability progression at the 10-year follow up, as mea- sured by the change in EDSS score. This is further supported by the results of predictor analysis, where longer fingolimod treatment along with age at base- line showed significant correlation with change in EDSS over 10 years. Additionally, patients in the high exposure group had a longer time to first use of an ambulatory device or wheelchair and to con- version to SPMS compared with patients in the low Figure 1. Kaplan–Meier plot of time to first use of an exposure group. The cut-off defined for fingolimod ambulatory assistive device by treatment group. treatment duration in this study is arbitrary and was Fingolimod treatment: — High exposure - - - Low expo- based on the cut-off defined by Ebers et al. in the sure þ Censored. IFN b-1b long-term follow-up study.11 An advan- tage of the chosen cut-off is that it resulted in rough- ly comparable group sizes for the high- and low- exposure groups. However, a potential drawback of this approach is that the groups have an imbalance in baseline co-variates possibly related to disease severity favoring the high exposure group. Therefore, a sensitivity analysis using treatment duration as a continuous parameter was performed. It corroborated the findings for the analysis using the arbitrary cut-off. However, the results should still be interpreted with caution. In a study that evaluated IFN b-1a treatment in 122 of 172 initially enrolled patients with relapsing– remitting MS, the mean EDSS score increased Figure 2. Kaplan–Meier plot of time to first use of a from 2.2 at baseline to 5.1 at the 15-year follow- wheelchair by treatment group. up.15 Similarly, in a cross-sectional study with IFN Fingolimod treatment: — High exposure - - - Low expo- b-1b, the EDSS score changed from 2.9 at baseline sure þ Censored. to a mean of 5.17 at the 16-year follow-up.16 In another long-term follow-up study of 100 patients Safety treated with glatiramer acetate (GA), the mean AEs were only reported if they occurred during the EDSS score increased from 2.5 at baseline to 3.1 one or two visits in this study and not retrospectively after 15 years of follow-up.17 In our study, patients collected for the time since patients left the core in the high exposure group had a mean EDSS change study. No AEs considered serious, severe or drug- from 2.16 at baseline to 2.72 at follow-up. A recent, related were reported. Although effort was made to large patient registry study of MS patients who were obtain information about patients who might have on IFN b-1a or GA treatment showed a median died since enrolling in the core study, no such increase of one EDSS point from baseline to the cases were recorded. Overall, safety findings did date of censoring, and 17.8% of patients reached not reveal any new safety issues associated with the milestone of an EDSS score 6 after prolonged fingolimod treatment. 10 years of follow-up.18 Our findings showed that a similar proportion (15.4%) of patients who interrupted fingolimod treatment reached the EDSS Discussion milestone of 6 whereas fewer patients (11.5%) in This cross-sectional long-term follow-up study the ACROSS study who had high exposure to fingo- included 62% of the original Phase 2 study limod reached EDSS 6. Similarly, mean time to www.sagepub.com/msjetc 7 Multiple Sclerosis Journal—Experimental, Translational and Clinical Table 4. MRI outcomes at year 10 by treatment group. Fingolimod treatment High exposure Low exposure (N ¼ 98) (N ¼ 56) LSM difference p value* Total T2w lesion volume, mm3 n 98 54 LSM (SE) 8703.8 (462.7) 11,310.1 (616.8) 2606.3 <0.01 Number of new/enlarging T2w lesions n 96 53 LSM (SE) 10.7 (1.8) 22.2 (2.4) 11.45 <0.0001 T1w hypointense lesion volume, mm3 n 98 54 LSM (SE) 3357.6 (252.3) 4216.2 (336.2) 858.6 NS *p values are derived from a Rank-ANCOVA test. ANCOVA was performed with MRI parameter at year 10 as response variable, duration of MS disease, baseline T2 lesion volume, and baseline EDSS as continuous covariates, and treatment as fixed effect. In the rank ANCOVA the outcome and all continuous covariates are replaced by their ranks. ANCOVA: analysis of covariance; LSM: least squares mean; MRI: magnetic resonance imaging; NS: nonsignificant; SE: standard error; T1w: T1-weighted; T2w: T2-weighted. the first use of walking assistance was longer in When interpreting the results of the ACROSS study, patients with high exposure to fingolimod compared it is important to note the patients in the high expo- to those who had low exposure. It is worth noting sure group had a lower EDSS score, a shorter disease that the mean duration of MS since onset in patients duration and lower T2w lesion volume at baseline who participated in the ACROSS study was almost compared with the low exposure group. Along the 20 years at follow-up. It is also important to consider same line, one-third of patients in the low exposure the baseline EDSS scores, because a higher baseline group were switched to a high efficacy treatment EDSS score might have led to the early use of ambu- after stopping fingolimod. Because these aspects latory devices. may reflect different disease activity in the two groups, intrinsic evolution of the disease may have At follow-up, the number of new/enlarging T2w contributed to positive efficacy outcomes in the high lesions and accumulation of T2 lesion volume were exposure group. Also, the response to fingolimod reduced in the high exposure group, comparable to treatment might be better in this group because sub- the beneficial effects of fingolimod on MRI endpoints group analyses of Phase 3 studies have shown fingo- in the Phase 3 trials.5,19 However, there was no cor- limod had a slightly stronger effect on relapse rates relation between the treatment duration and the MRI in patients with a lower EDSS score and shorter dis- outcomes at 10 years. Differences in image acquisi- ease duration.20 The beneficial response in the high tion and resolution and other technical changes over exposure group may therefore be attributable to four time might have to be taken into account. different factors: (a) an effect of fingolimod on the disease course independent of the baseline character- Compared to core baseline, patients in the high istics; (b) a stronger treatment response to fingoli- exposure group were stable in MSFC z-score and mod in patients with a shorter disease duration and improved in PASAT-3 score at follow-up whereas lower EDSS score at treatment start; (c) a more patients in the low-exposure group exhibited a wors- benign disease course independent of the treatment; ening. However, no significant difference between and (d) treatment satisfaction or dissatisfaction lead- the two treatment groups was observed for ing to longer or shorter exposure, as non-responders T25FWT or 9-HPT scores. A learning effect on the to fingolimod or patients with severe disease are PASAT-3 test cannot be excluded because patients likely to switch to other drugs. Given the lack of a in the high exposure group enrolled in fingolimod matched control cohort, it is not possible to estimate extension studies had more PASAT tests in the the magnitude of the beneficial contribution of these 10-year period than those in the low exposure group. factors. However, it is plausible that these four 8 www.sagepub.com/msjetc Derfuss et al. factors may have all contributed to the outcomes at Xavier Montalban has been a steering committee member follow-up. It would certainly be worthwhile to com- of clinical trials or participated in advisory boards of clin- pare the high exposure population with a propensity ical trials in the past with Bayer Schering Pharma, Biogen Idec, EMD Merck Serono, Genentech, Genzyme, Novartis, score-matched real-world population to distinguish Sanofi-Aventis, and Teva Pharmaceuticals. the treatment effects from a potential selection bias. Other limitations for consideration are that Mariaemma Rodegher has received honoraria for speaking varied DMTs were used among patients who had and travel grants from Merck Serono and Novartis. discontinued fingolimod, there were varying treat- ment durations in patients who had stopped fingoli- Jens Wuerfel is CEO of MIAC AG Basel, Switzerland. He mod, and the ascertainment rate of 62% for the has served on scientific advisory boards for Actelion, overall study (84.0% for the participating sites). Biogen, Genzyme-Sanofi, Novartis, and Roche. He is or was supported by grants from the European Union In conclusion, disability progression in this study (Horizon2020), Swiss National Research Foundation and cohort is low with over 80% of patients below an the German Federal Ministries of Education and Research (BMBF), and Economic Affairs and Energy (BMWI). EDSS of 6, almost 20 years after onset of disease. This proportion is even further increased in patients Laura Gaetano was employed at MIAC AG, where the with high fingolimod exposure. The effects seen for imaging data were analyzed. She is employed by F. high fingolimod exposure are consistent with the effi- Hoffman-La Roche Ltd., but her current institution was cacy reported previously in 5- and 7-year extension not involved in any way in this study. studies. The high persistence rate among patients willing to undergo another assessment 10 years or Ludwig Kappos’s institution (University Hospital Basel) more after the start of the initial Phase 2 study indi- has received in the last 3 years and used exclusively for cates patients and prescribers perceive fingolimod to research support: steering committee, advisory board, and consultancy fees (Actelion, Addex, Bayer HealthCare, be an effective, safe, and well-tolerated treatment. Biogen Idec, Biotica, Genzyme, Lilly, Merck, Mitsubishi, Novartis, Ono Pharma, Pfizer, Receptos, Acknowledgements Sanofi, Santhera, Siemens, Teva, UCB, and Xenoport); The authors would like to thank the patients who partici- speaker fees (Bayer HealthCare, Biogen Idec, Merck, pated in the study. They would also like to thank Elisabetta Novartis, Sanofi, and Teva); support of educational activ- Verdun di Cantogno and Diego Silva for their contribu- ities (Bayer HealthCare, Biogen, CSL Behring, Genzyme, tions to protocol development and study conduct, and Merck, Novartis, Sanofi, and Teva); license fees for Akhilesh Singh and Avinash Thakur (Novartis Neurostatus products; and grants (Bayer HealthCare, Healthcare Pvt. Ltd., Hyderabad, India) for writing assis- Biogen Idec, European Union, Merck, Novartis, Roche tance, editorial review assistance, and in coordinating Research Foundation, Swiss MS Society, and Swiss author reviews. All authors edited the manuscript for intel- National Research Foundation). lectual content, provided guidance during manuscript development, and approved the final version submitted Christian Wolf’s institution has received honoraria for con- for publication. sultancy services from BBB, Celgene, Desitin, Immunic, Keyrus, Mylan, Novartis, Roche, Synthon, and Teva. Conflict of Interests Amin Azmon and Davorka Tomic are employees of The author(s) declared the following potential conflicts of Novartis. interest with respect to the research, authorship and/or publication of this article: Tobias Derfuss serves on scien- tific advisory boards for Novartis Pharmaceuticals, Merck Funding Serono, Biogen Idec, Genzyme, GeNeuro, Mitsubishi The author(s) disclosed receipt of the following financial Pharma, Roche, and Bayer Schering Pharma; has received support for the research, authorship, and/or publication of funding for travel and/or speaker honoraria from Biogen this article: This study was funded by Novartis Pharma Idec, Genzyme, Novartis, Merck Serono, Roche, and AG, Basel, Switzerland. Bayer Schering Pharma; and receives research support from Biogen Idec, Novartis Pharma, the European ORCID iDs Union, the Swiss National Foundation, and the Swiss T Derfuss https://orcid.org/0000-0001-8431-8769 MS Society. J Sastre-Garriga https://orcid.org/0000-0002-1589- 2254 Jaume Sastre-Garriga has received personal fees from L Gaetano https://orcid.org/0000-0002-5214-2858 Merck Serono, Biogen Idec, Teva, Roche, Sanofi, C Wolf https://orcid.org/0000-0002-4683-0507 Almirall, and Novartis outside of the submitted work. L Kappos https://orcid.org/0000-0003-4175-5509 www.sagepub.com/msjetc 9 Multiple Sclerosis Journal—Experimental, Translational and Clinical Supplemental material 16 years after the pivotal IFNB-1b trial. J Neurol Supplemental material for this article is available online. Neurosurg Psychiatry 2010; 81: 907–912. 12. World Medical Association. World Medical Association Declaration of Helsinki: ethical principles References for medical research involving human subjects. JAMA 1. Goldenberg MM. Multiple sclerosis review. PT 2012; 2013; 310: 2191-294. 37: 175–184. 13. Lechner-Scott J, Kappos L, Hofman M, et al. Can the 2. Ziemssen T and Thomas K. Treatment optimization in Expanded Disability Status Scale be assessed by tele- multiple sclerosis: How do we apply emerging evi- phone? Mult Scler 2003; 9: 154–159. dence? Expert Rev Clin Immunol 2017; 13: 509–511. 14. Lublin FD and Reingold SC. Defining the clinical 3. Kappos L, Antel J, Comi G, et al. Oral fingolimod course of multiple sclerosis: Results of an internation- (FTY720) for relapsing multiple sclerosis. N Engl J al survey. National Multiple Sclerosis Society (USA) Med 2006; 355: 1124–1140. Advisory Committee on Clinical Trials of New Agents 4. Comi G, O’Connor P, Montalban X, et al. Phase II in Multiple Sclerosis. Neurol 1996; 46: 907–911. study of oral fingolimod (FTY720) in multiple sclero- 15. Bermel RA, Weinstock-Guttman B, Bourdette D, et al. sis: 3-year results. Mult Scler 2010; 16: 197–207. Intramuscular interferon beta-1a therapy in patients 5. Izquierdo G, O’Connor P, Montalban X, et al. Five- with relapsing-remitting multiple sclerosis: A 15- year results from a phase 2 study of oral fingolimod in year follow-up study. Mult Scler 2010; 16: 588–596. relapsing multiple sclerosis. Mult Scler 2014; 20: 16. Goodin DS, Traboulsee A, Knappertz V, et al. 877–881. Relationship between early clinical characteristics 6. Montalban X, Comi G, Antel J, et al. Long-term and long term disability outcomes: 16 year cohort results from a phase 2 extension study of fingolimod study (follow-up) of the pivotal interferon beta-1b at high and approved dose in relapsing multiple scle- trial in multiple sclerosis. J Neurol Neurosurg rosis. J Neurol 2015; 262: 2627–2634. Psychiatry 2012; 83: 282–287. 7. Cohen JA, Barkhof F, Comi G, et al. Oral fingolimod 17. Ford C, Goodman AD, Johnson K, et al. Continuous or intramuscular interferon for relapsing multiple scle- long-term immunomodulatory therapy in relapsing rosis. N Engl J Med 2010; 362: 402–415. multiple sclerosis: Results from the 15-year analysis 8. Calabresi PA, Radue EW, Goodin D, et al. Safety and of the US prospective open-label study of glatiramer efficacy of fingolimod in patients with relapsing- acetate. Mult Scler 2010; 16: 342–350. remitting multiple sclerosis (FREEDOMS II): A 18. Jokubaitis VG, Spelman T, Kalincik T, et al. double-blind, randomised, placebo-controlled, phase Predictors of long-term disability accrual in relapse- 3 trial. Lancet Neurol 2014; 13: 545–556. onset multiple sclerosis. Ann Neurol 2016; 80: 9. Kappos L, Radue EW, O’Connor P, et al. A placebo- 89–100. controlled trial of oral fingolimod in relapsing multi- 19. Kappos L, O’Connor P, Radue EW, et al. Long-term ple sclerosis. N Engl J Med 2010; 362: 387–401. effects of fingolimod in multiple sclerosis: the ran- 10. McDonald WI, Compston A, Edan G, et al. domized FREEDOMS extension trial. Neurol 2015; Recommended diagnostic criteria for multiple sclero- 84: 1582–1591. sis: Guidelines from the International Panel on the 20. Derfuss T, Ontaneda D, Nicholas J, et al. Relapse rates diagnosis of multiple sclerosis. Ann Neurol 2001; in patients with multiple sclerosis treated with fingo- 50: 121–127. limod: Subgroup analyses of pooled data from three 11. Ebers GC, Traboulsee A, Li D, et al. Analysis of clin- phase 3 trials. Mult Scler Relat Disord 2016; 8: ical outcomes according to original treatment groups 124–130. 10 www.sagepub.com/msjetc