Disease-modifying therapies in MS: long-term data


REPORT FROM AAN – BOSTON, MA, APRIL 22-28, 2017 – The following summarizes data from key studies presented at the American Academy of Neurology annual meeting on the long-term use of disease-modifying therapies in multiple sclerosis.

Dimethyl fumarate
Commentary by Daniel Selchen, MD

Teriflunomide: In the phase III TEMSO trial, teriflunomide 14 mg/day significantly reduced the annualized relapse rate (ARR, 31.5% reduction) and 12-week confirmed disability progression (29.8% reduction) versus placebo (O’Connor et al. N Engl J Med 2011;365:1293-303). Similar reductions in ARR (36.3%) and disability progression (31.5%) versus placebo were reported in the phase III TOWER trial (Confavreux et al. Lancet Neurol 2014;13:247-56). For the extension studies, patients in the placebo group were randomized to teriflunomide 7 or 14 mg/day (TEMSO) or received teriflunomide 14 mg/day (TOWER). Relapse rates remained consistently low at 11-year follow-up in both TEMSO and TOWER (ARR 0.230 and 0.239); severe relapses requiring corticosteroids (TEMSO 0.197; TOWER 0.219) or hospitalization (TEMSO 0.061; TOWER 0.111) were uncommon (Maurer et al. AAN 2017; abstract P6.361).  In the 5-year follow-up of the combined TEMSO/TOWER dataset, there was a low risk of worsening disability: 96.4% of patients with baseline EDSS score < 5 on teriflunomide 14 mg/day had not advanced to EDSS 6; and 97% had no sustained disability worsening (Lublin et al. AAN 2017; abstract P6.341).

Two subgroup analyses from TEMSO/TOWER extensions were also performed. In patients recently diagnosed with relapsing MS (n=587; median baseline EDSS score 2.0), a majority did not reach EDSS 4 (90.1%) or EDSS 6 (96.7%) at 5-year follow-up (Oh et al. AAN 2017; abstract P6.339). In the subgroup with progressive MS (SPMS, n=60; PPMS, n=62), 80.3% did not experience 12-week disability worsening over a 9-year follow-up period. The median EDSS score was unchanged from baseline (Nelson et al. AAN 2017; abstract S33.008).

In the long-term extension of the TOPIC trial in clinically isolated syndrome (CIS)/early MS, 63% of subjects did not experience a relapse determining clinically-definite MS; and 78% did not experience disability worsening over the 7-year follow-up (Miller et al. AAN 2017; P2.103).
Back to top

Dimethyl fumarate: Three analyses were performed using data from the ENDORSE extension of the DEFINE and CONFIRM phase III trials (Gold et al. N Engl J Med 2012;367:1098-107; Fox et al. N Engl J Med 2012;367:1087-97). For the extension, patients in the placebo or glatiramer acetate groups were switched to DMF 240 mg BID or TID; all TID patients were subsequently switched to DMF 240 mg BID. At 7-year follow-up (core study + 5 years of extension), the overall continuation rate was 52% (56% in the continuous 240 mg BID group) (Gold et al. AAN 2017; abstract P6.352). In the newly-diagnosed subgroup on continuous DMF 240 mg BID, the ARR (0.13) and 24-week confirmed disability progression rate (18.0%) remained low; 54.9% were free of clinical disease activity. In the subgroup of patients with MRI follow-up (n=211), 64% overall had no new/enlarging T2 lesions (range 62% in the continuous DMF 240 mg BID group, 73% in the placebo/240 mg BID group) at year 7 (Arnold et al. AAN 2017; abstract S12.002).

In the safety analysis, the proportion of patients receiving continuous DMF 240 mg BID experiencing serious adverse events was 27%, and 4% had serious infections (Pozzilli et al. AAN 2017; abstract P5.383). One case of progressive multifocal leukoencephalopathy (PML) was recorded in the TID group (total drug exposure in ENDORSE to date is 7076 patient-years).
Back to top

Fingolimod: A post-hoc analysis was performed on data from the long-term extension of TRANSFORMS, which compared fingolimod with intramuscular interferon-beta-1a (Cohen et al. N Engl J Med 2010;362:402-15). The endpoints were no evidence of disease activity (NEDA-3), and NEDA-4 (NEDA-3 + no brain volume change >0.4%/year) (Cohen et al. AAN 2017; abstract P4.390). At 1 year, the rate of NEDA-3 was 43.3% with fingolimod vs. 30.3% with IFN-beta; the rate of NEDA-4 was 27.8% vs. 16.4%, respectively. The 1-year NEDA-3 and NEDA-4 rates in the IFN group increased to 50.2% and 35.1%, respectively, after switching to fingolimod. From years 3 to 7, NEDA-3 ranged from 45.6% to 74.2%; NEDA-4 ranged from 24.4% to 53.2%. A previous analysis of data from FREEDOMS reported a two-year NEDA-4 rate of 19.7% with fingolimod vs. 5.3% with placebo (Kappos et al. Mult Scler 2016;22):1297-305).

The interim 5-year results from the PANGAEA observational study (n=300) were also presented (Ziemssen et al. AAN 2017; abstract P6.345). In each year of treatment, 70% were free of relapses and 6-month confirmed disability progress. Over 90% experienced EDSS stability or improvement in each year of follow-up. In the long-term safety analysis, the incidence of serious adverse events was 3.9% (Ziemssen et al. AAN 2017; abstract P5.365). The most common adverse effects included lymphopenia, nasopharyngitis and bronchitis. Overall, 45% of subjects reported no adverse effects. The proportion of patients remaining on treatment at 5 years was 67%.
Back to top

Natalizumab: The long-term rate of natalizumab discontinuations was examined in the Tysabri Observational Program (TOP) database (Trojano et al. AAN 2017; abstract P2.399). The discontinuation rate was 10.1% in year 1, 11.2% in year 2, 15.7% in year 3, and 9.5% in year 4. Median time on natalizumab was 2.9 years. The most common reasons for discontinuation were concerns about PML (45.0%), lack of efficacy (14.0%), non-PML safety concerns (9.7%) and poor tolerability (6.9%). Treatments used after natalizumab discontinuation were fingolimod (58.6%), dimethyl fumarate (10.6%), IFN-beta (10.2%), and glatiramer acetate (9.5%). ARR increased after switching but remained lower than the pre-treatment ARR.

A separate study reported that long-term natalizumab exposure may be associated with an increased pathogenic signature of Th17 cells (Janoschka et al. AAN 2017; abstract P2.408). In the study, Th17 acquired a more pronounced pathogenic profile during long-term natalizumab exposure, with increased expression of proinflammatory molecules (e.g. IL23-R) and cytokines (CM-CSF, TNF-alpha). The authors noted that this phenomenon could help to explain the increased disease activity that can occur when natalizumab is withdrawn.
Back to top

Alemtuzumab: Six-year data were presented from the extensions of the CARE-MS I study in de novo patients, and the CARE-MS II study of previously-treated patients (Cohen et al. Lancet 2012;380:1819-28; Coles et al. Lancet 2012;380:1829-39). In both extensions, alemtuzumab was re-dosed as needed. In CARE-MS I, 63% of patients received no additional treatment at 6-year follow-up (Singer et al. AAN 2017; abstract S24.005). ARR remained low (0.15), and 84-89% of patients were relapse-free in each of Years 3-6. Over the 6-year observation period, 77% of patients had no 6-month confirmed disability worsening; 34% experienced confirmed disability improvement. The mean change from baseline in EDSS score was +0.04 points. The proportion of patients free of new Gd-enhancing T1 and new/enlarging T2 lesions was 87% and 67%, respectively, in Year 6 (Arnold et al. AAN 2017; abstract S10.002). The annual NEDA rates were 62% and 57% in years 5 and 6, respectively. In the IFN group, ARR decreased from 0.39 (Year 2) to 0.15 (Year 6) after switching to alemtuzumab; the proportion of patients relapse-free increased from 79% to 86% over the same time period (Oreja-Guevara et al. AAN 2017; abstract P5.360). At 4 years after switching, 81% had no 6-month confirmed disability progression, and 27% had confirmed disability improvement.

In the CARE-MS II extension, 50% of patients did not require re-dosing at 6-year follow-up (Schippling et al. AAN 2017; abstract P5.339). In year 6, ARR was 0.15, 91% were free of new Gd-enhancing T1 lesions, and 69% had no new/enlarging T2 lesions. The proportion of patients with no 6-month confirmed disability progression was 72%, and 43% had confirmed improvement in disability (Fox et al. AAN 2017; abstract S24.006). The mean change from baseline in EDSS score was 0.10 points.

In a subgroup analysis of patients with highly active MS at entry, MRI activity was consistently low during the long-term follow-up (Rovira et al. AAN 2017; abstract S12.006). The proportion free of MRI disease activity was >65% for each of the extension years 3-6. In Year 6, 87% had no Gd-enhancing lesions, and 66% had no new/enlarging T2 lesions.

In the group originally randomized to IFN-beta-1a SC, 71% required no further treatment after switching to alemtuzumab. ARR decreased from 0.52 in Year 2 of IFN, to 0.15 after two years of alemtuzumab; ARR was 0.17 after four years on alemtuzumab (study year 6) (Boyko et al. AAN 2017; abstract P5.332). The proportion of patients who were relapse-free increased from 70% with IFN, to 86% and 85% with alemtuzumab in study years 4 and 6, respectively. The proportion of patients free of Gd-enhancing lesions increased from 78% to 91% at 24 months after switching from IFN to alemtuzumab. The proportion with new/enlarging T2 lesions increased from 48% to 81% after two years of alemtuzumab. In Year 6 (four years after switching), 89% and 69% of patients were free of Gd-enhancing T1 and new/enlarging T2 lesions, respectively. The NEDA rate was 32% after two years of IFN; the annual NEDA rate was 66%, 63% and 60% after switching to alemtuzumab.

Also noteworthy were the long-term effects of alemtuzumab on the rate of brain volume loss (BVL). In CARE-MS I, the median change in brain parenchymal fraction (BPF) was -0.59% in Year 1, -0.25% in Year 2, -0.19% in Year 3, -0.14% in Year 4, -0.20% in Year 5, and -0.17% in Year 6 (Traboulsee et al. AAN 2017; abstract P2.104). The BVL rate similarly slowed in CARE-MS II (from -0.48% in Year 1, to -0.10 in Year 6). These results suggest that brain atrophy rates are normalized 2-3 years after alemtuzumab initiation, based on a previously-reported BVL rate of -0.27%/year in healthy controls (De Stefano et al. J Neurol Neurosurg Psychiatry 2016;87:93-9).
Back to top

Ocrelizumab: The preliminary results from the long-term extension of the 96-week OPERA I/II trials of ocrelizumab 600 mg q24weeks in relapsing MS were presented (Naismith et al. AAN 2017; abstract S31.004). All patients in the control arm (IFN-beta 1a SC) were switched to ocrelizumab for the extension. A total of 1325 of 1656 subjects (80%) enrolled in the extension. At the time of the interim analysis, 1214 subjects (91.6% of the extension group; 73.3% of the initial sample) had received at least 48 weeks of follow-up. In the continuous ocrelizumab group, ARR improved from 0.16 (core study) to 0.136 (OPERA I) and 0.138 (OPERA II). In the IFN switch group, ARR improved with ocrelizumab from about 0.250 with IFN to 0.092 (OPERA 1) and 0.115 (OPERA II) with ocrelizumab.
Back to top


Dr. Daniel Selchen: A major theme at all recent meetings where MS abstracts have been presented (ECTRIMS, AAN, and EAN) is long-term data for existing treatments, mainly from extensions of phase II and phase III clinical trials. This year’s AAN was no exception. The article above summarizes the highlights of several such studies presented at the AAN. What can be gleaned from the study extensions is very different from what we learn from randomized controlled clinical trials. Extensions no longer have a comparator arm and neither the patients nor the clinicians are blinded to treatment – conditions necessary to tell whether a treatment is superior to a comparator (active treatment or a placebo).

What can we learn from this kind of long-term data?

Durability of treatment: These studies can inform us as to whether patients who initially do well on a treatment continue to do well (those who have failed are generally no longer there), and for how long they do well – something we cannot learn from a two-year clinical trial.

Safety over an extended period of time: While a typical trial can inform us about short-term safety, extension data add years of safety analysis, enabling us to see if a common “smoking gun” emerges.  The trial extensions are of course too small to measure/predict very rare occurrences.

Disability progression: Two-year trials can compare rates of disability progression over a short period.  Extensions, while generally limited to responders and crossover patients, can show us what proportion of participants are likely to show progression over a number of years.

Effects on MRI: These studies can show us if the imaging benefits/patterns seen in the core trials are sustained over time.

Effect on brain atrophy: Clinical trials are generally too short to determine if treatments have a sustained effect on brain atrophy.  The extension trials of some agents now have atrophy data going out to six years.

Combined tolerability and efficacy: It is interesting to note both the proportion of patients who go into an extension study, and the number that remain in it. These are indirect measures both of tolerability and efficacy.

In all of the studies cited above, regardless of agent, responders tended to show ongoing relapse rates similar to those in the initial clinical trials, suggesting durability. Similarly, disability progression rates remained low in all of the studies, suggesting relative stability in responders over a period between five and nine years.

With regard to safety, all of these studies showed no new major safety issues with any of the agents, except for the PML case with DMF.

In several of the studies, the MRI data showed stability in the rates of new T2 lesions and gadolinium-enhancing lesions, which seem to remain consistent over time. Perhaps the most striking piece of data from the studies described above is the effect of alemtuzumab on brain atrophy, which showed improvement over five years, suggesting a real  benefit on a major indicator that correlates with disability progression.

In general, the long-term data summarized above suggest that all of the newer agents introduced for treatment of relapsing multiple sclerosis in the last 10 years have predictable, durable benefits, and safety profiles that are consistent with what was seen in the initial clinical trials. It must be noted that rare occurrences (particularly PML and opportunistic infections) are too infrequent to be adequately assessed in the relatively small sample sizes of these extension studies, and need to be followed with “real world” reporting. In addition, given the differences in trial populations and designs, these studies are not of great value for drug-to-drug comparisons, which are better captured in “real world “data collections such as MSBase.

Back to top

Recommend to a Colleague

Related Posts

Go back to home page