The time to confirmed progression was significantly different between groups: the placebo group reached the 40% quantile at 549 days and the IFNB-1b group at 893 days (p  =  0.0008). The probability of remaining progression free was significant after 12 months of treatment and remained significant thereafter. The proportion of patients with confirmed progression over the study period was 49.7% for the placebo group and 38.9% for the IFNB-1b group, a relative reduction of 21.7% (p  =  0.0048) and an absolute risk reduction of 9.8%. The probability of progression for patients on placebo was 1.6 times that of a patient on IFNB-1b (OR  =  0.63, 95% CI: 0.47–0.93). The development of confirmed progression was not influenced by the occurrence or absence of a relapse within 2 years prior to study entry.

During the trial, annualized relapse rates (ARR) were significantly different between IFNB-1b-treated patients (ARR 0.44 over 3 years) and placebo-treated patients (ARR 0.64 over 3 years; p  =  0.0002); the median time to first relapse was 644 and 403 days, respectively (p  =  0.003). As with other IFNB-1b cohorts, 28% of patients developed neutralizing antibodies (NAbs). The presence of NAbs reduced the effectiveness of IFNB-1b on relapse rates (as in the RRMS RCT) but is reported not to have altered disability outcomes. It should be noted that others have convincingly shown a loss of the therapeutic effect of IFNB on relapse activity and disability progression when NAbs are detected [19].

The MRI results were published in detail separately to the clinical outcome study in four papers [20–23].

MRI measures of inflammation: Overall, T2 lesion load increased over the 3 years of the study in the placebo group by 15% and decreased by 2% in the IFNB-1b-treated group (p  <  0.0001) [20]. New T2 lesions and enhancing lesions decreased by a mean of 57% and a median of 70% in the treated group compared to placebo. No active lesions were seen in 16% of placebo patients scans compared to 38% of IFNB-1b patients (p  <  0.0001). In a subgroup of patients having frequent monthly gadolinium-enhanced MRI scans, there was a decrease of 65% in the mean number of new active lesions between months 1 and 6 and a decrease of 78% between months 19 and 24 in the IFNB-1b group compared to placebo. IFNB-1b, compared to placebo, reduced the rate of small new enhancing lesions by 70% and of large enhancing lesions by 46%. The proportion of enhancing lesions developing into T1 hypointensities did not differ between the groups and were more likely to form from large than small enhancing lesions [21].

MRI measures of cerebral atrophy: A subgroup of 95 patients (13% of the total cohort of 718) had 6-monthly T1 scans in five centers [22]. Overall, there was no difference between the placebo- and the IFNB-1b-treated patients in relation to atrophy. This disappointing finding was partly ascribed to the small number of patients scanned, but also might relate to the dynamics of cerebral volume change in patients with SPMS. Atrophy may reflect an on-going loss of tissue consequent to inflammatory disease, which occurred in the years prior to entry to the study. Also, paradoxically, a drug with a potent anti-inflammatory effect will cause ­apparent atrophy (pseudoatrophy) compared to the placebo group because of loss of volume caused by reduced edema in the cerebrum consequent upon reduced inflammation. This will be most apparent in the first 6–12 months of treatment. This was borne out when a post hoc analysis was performed. Patients with no gadolinium enhancement at baseline (less inflammatory activity) demonstrated a treatment effect on cerebral atrophy with a 5.1% loss of volume in the placebo group and 1.8% loss in the IFNB-1b arm at 36 months (p  =  0.0026; see Fig. 9.1).

MRI–clinical correlations: MRI findings in relation to clinical outcomes were assessed in the subgroup of 125 patients with monthly scans [23]. This study emphasized the weak correlations in SPMS between MRI measures (baseline T2 lesion volume, increase in T2 lesion load, increase in new and enhancing lesions) and clinical measures of disability progression. The authors emphasized that MRI measures of inflammation do not quantify axonal injury and demyelination in the brain and this may explain the lack of correlation between MRI markers of inflammation and short-term changes in disability over 2 years.

This was the first pivotal study to assess HrQoL in MS and used the generic instrument: the Sickness Impact Profile (SIP) [24]. The changes were weakly positive, indicating a significant positive effect on the SIP-physical at 6 months, 12 months, and at the last visit.

Subgroup analyses of the European SPMS study emphasized that overall there was a 16% treatment effect in the actively treated cohort [25]. The treatment effect was most marked for patients with active disease; those who had two or more relapses in the 2 years before study entry or who had 1.0-point change in EDSS in the 2 years before study entry. Also (surprisingly), a more marked effect was seen both in patients who had longer disease duration and those with a longer time of progressive deterioration.

The differences between the two trials and the reason why the European study, and not the North American, attained the primary endpoint have been addressed in detail [27]. Differences in the baseline characteristics of the two study populations explain the reason for the divergent results. Patients in the European study at entry were younger, had a shorter disease duration, had a higher pre-study ARR, had more relapses in the preceding 2 years, and had more gadolinium-enhancing lesions at baseline brain MRI than those in the North American study (Table 9.1). Seventy percent of European patients but only 45% of North American patients had relapses in the preceding 2 years. Although there were several reasons for the biological heterogeneity of the two study populations, the most important determining factor was the difference in the entry criteria. In both studies, a diagnosis of SPMS was required but, whereas the North American study stipulated an increase in the EDSS by 1.0 point or more, the European study required either a one point or more increase in the EDSS or two relapses in the preceding 2 years. How European neurologists differentiated between patients having relapses with resulting accumulating residual disability and those patients with slowly progressive disability and superimposed relapses was left to them. Remarkably, 285 (40%) of the 718 patients entered the European study because of relapse history alone without any required documented pre-study increase in disability. European patients had thus more evidence of an active inflammatory component to their illness than the North American patients and would be more likely to respond to an anti-inflammatory drug.

The primary endpoint, time to confirmed progression in the EDSS, was not significant for either dose of IFNB-1a.

There was a significant effect on all measures of relapse activity at both doses of IFNB-1a compared to placebo. Subgroup analysis showed that women had a delay in confirmed progression in both treatment arms, 22 μg (p  =  0.038) and 44 μg (p  =  0.006) compared to placebo. In the placebo group, it was found unexpectedly that men performed better than women in relation to the disability outcome (HR  =  0.64; 95% CI, 0.43–0.95, p  =  0.016). Further subgroup analysis showed that patients who had progressed more rapidly in the previous 2 years and those with pre-study relapses tended to show more treatment effect. NAbs were more common in the 22 μg dose arm (21% positive) than the 44 μg dose arm (15% positive) and did not seem to affect time to disability progression. In the 44 μg dose cohort, NAbs appeared to negate the effect on relapse suppression.

In a further paper, analysis of various ratings of depression in 365 patients during the SPECTRIMS study showed no evidence of an increased risk for depression in patients treated with either dose of IFNB-1a [29].

In the companion paper on MRI outcomes, a significant therapeutic effect on MRI evidence of inflammatory activity was noted with a 70–75% reduction in the T2 active lesions [30]. At baseline, there was a difference in the burden of disease (BOD; total T2 lesion area) with men having a significantly higher BOD than women. As with the clinical outcome measures, a gender difference on therapeutic effect was noted in relation to the percentage change in BOD, which was greater and evident in both 22 and 44 μg doses in women than in men.

The MRI study also provided evidence of the importance of NAb testing. At both the 22 and 44 μg doses, patients who were Nab-positive at any stage (once positive, always positive), lost evidence of an MRI therapeutic effect (Fig. 9.2). Once-off NAbs were found in 24% of the 44 μg dose and 29% of the 22 μg dose. Persistent NAbs were found in 15% of the 44 μg dose patients and 21% of the 22 μg dose patients.

The primary endpoint was the change in the MSFC from baseline to month 24. In both groups, the mean MSFC worsened over 24 months but the degree of worsening was significantly less in the active treatment group (−0.495) in comparison to the placebo group (−0.362) (p  <  0.033). The difference between the two groups in the degree of MSFC worsening related mainly to reduced worsening in the 9-HPT (p  <  0.024) and the PASAT3 (p  <  0.061) in the group treated with IFNB-1a. The T25FWT showed no significant change.

The usual disability measure, the EDSS, did not show any significant treatment effect, either assessed as time to confirmed progression or mean change in the EDSS. The annualized relapse rate during the study was 0.3 in the placebo group and 0.2 in the IFNB-1a group (p  <  0.008). On a measure of health-related quality of life, the active treatment group showed significant improvement compared to placebo.

There was a significant effect of IFNB-1a, compared to placebo, on all MRI endpoints, including a 53% reduction of new or enlarging T2 lesions at 12 months (46% at 24 months) and a reduction of T2 lesion volume by 78% at 12 months and 69% at 24 months.

This was the first study to use the MSFC and the only SPMS study, other than the European SPMS study, to demonstrate an effect on the primary endpoint. However, the MSFC difference in the two study arms was minimal, and although statistically significant was certainly not clinically important or relevant, and therefore this was essentially a negative study. The problem with the MSFC is that, although it is a continuous scale, it is difficult to define, in terms of impact on function, the effect of less worsening by 0.133 on the mean MSFC over 2 years. In the European SPMS study, the effect of IFNB-1b could be described as a delay in disability progression by almost 12 months during the study period, an observation that is understandable in terms of function, even if one might debate whether it is or is not clinically important. For the MSFC change, the functional effect of reduced worsening is difficult to judge [32].

There was no difference in the two groups in time to sustained disability progression.

Confirmed disability progression was found in 41% of the IFNB-1a group and 38% of the placebo patients. There was no effect on annualized relapse rates, time to first relapse, proportion relapse free, or hospitalizations for MS relapses. Subgroup analyses in relation to effects of gender or prior disease history showed no significant differences.

Although trials in early multiple sclerosis using low doses of IFNB had shown therapeutic effects in the clinically isolated syndrome [34, 35], it is clear from the Nordic SPMS study that in established SPMS, such low doses neither affect measures of inflammation nor of disability.

The EDSS worsened by 0.23 in the placebo group over the 2 years of study and improved by a mean of 0.13 in the mitoxantrone 12 mg/m² group (p  =  0.0194). The number of treated relapses was significantly less in the mitoxantrone 12 mg/m² group than placebo (p  =  0.0002). First quartile time to first treated relapse was 6.7 months in the placebo group and 20.4 months in the mitoxantrone 12 mg/m² group (p  =  0.0004). Significant treatment effects were also seen in the Ambulation index change and the standardized neurological status.

The annualized relapse rate(ARR) was significantly lower in the 12 mg/m² mitoxantrone group than in the placebo group for year 1 (0.42 vs. 1.15, p  <  0.0001) and year 2 (0.27 vs. 0.85, p  =  0.0001; differences 63% and 68%). Confirmed progression of disability by the EDSS by one point differed between the groups: patients in the 12 mg/m² group showed a significant advantage in the analysis of time to confirmed EDSS deterioration at 3 months (p  =  0.03) and 6 months (p  =  0.03).

Significantly fewer patients on 12 mg/m² mitoxantrone had enhancing lesions at 24 months compared to those on placebo (none vs. 16%, p  =  0.02). The mean increase in the number of T2-weighted lesions was 0.29 in the mitoxantrone group and 1.94 in the placebo group (p  =  0.03).

The authors acknowledged that significant treatment effects may relate to relapse suppression; 140 (74%) patients had prior relapses in the year prior to study entry. There was a favorable, but nonsignificant, effect of treatment on EDSS progression in patients without relapses in the year prior to study entry.

A subgroup of 110 patients of the 194 who entered the study had MRI evaluation of treatment effect in the three arms: mitoxantrone 5 mg/m² (40 patients), mitoxantrone 12 mg/m² (34 patients), and placebo (40 patients) [39]. The results were disappointing and the authors have explained this by the non-randomized nature of patients who were selected for scanning. Only the number of new T2 lesions at 24 months differed significantly in the mitoxantrone 12 mg/m² from placebo (p  <  0.027).

On the basis of the results of the MIMS trial, the US Food and Drug Administration approved mitoxantrone 12 mg/m² administered intravenously every third month for treatment of worsening relapsing– remitting, progressive–relapsing, and secondary progressive MS. The entry criteria and the baseline characteristics of the patients in this study were very similar to the characteristics of patients who were entered in the European SPMS study of IFNB-1b (Table 9.1) [18]. Although the authors suggest in their discussion that SPMS patients without relapses might also benefit from treatment with mitoxantrone, they have no evidence to support this (and there has been none since). Other smaller trials and subsequent experience in clinical use would indicate that unless a patient with SPMS has associated relapses and/or gadolinium-enhancing lesions on brain MRI, they are unlikely to benefit from mitoxantrone [38, 39].

A report of the Therapeutics and Technology Assessment Subcommittee of the American Academy of Neurology in 2003 was justifiably cautious in their assessment and gave only a type B recommendation: mitoxantrone may have a beneficial effect on disease progression in patients with MS whose clinical condition is deteriorating. In general, however, this agent is of limited use and of potentially great toxicity. Therefore, it should be reserved for patients with rapidly advancing disease who have failed other therapies [40]. They made a strong recommendation for a further phase III study of mitoxantrone in SPMS.

Two toxic side effects, cardiomyopathy and, in particular, treatment-related acute leukemia (TRAL), add concern in the use of mitoxantrone in clinical practice. Cardiomyopathy may develop early in the treatment course; patients require regular 3-monthly echocardiography to detect changes in left ventricular ejection fraction (LVEF). An expert review in 2010 suggested a 12% incidence of reduced LVEF and 0.4% of congestive cardiac failure [41].

Although initial studies suggested that TRAL was rare with an incidence of 0.07% [42], the reported incidence of TRAL has been rising; 0.3% in a UK review [43], 0.69% in an Italian series [39], and 0.81% in a recent review by the Therapeutics and Technology Assessment Subcommittee of the American Academy of Neurology [41]. The increasing incidence of TRAL, personal experience of one case, and the availability of other less toxic therapies have made the author wary of using mitoxantrone in recent years. A Cochrane Collaboration review in 2005 concluded that “mitoxantrone has a partial efficacy, but, due to its unclear long-term safety profile, it should be used to treat patients with worsening RR and SPMS with evidence of worsening disability [44].”

Experience of the use of natalizumab has added to concerns about the prior use of mitoxantrone (or other immunosuppressants). The risk of progressive multifocal leucoencephalopathy (PML) in patients on natalizumab therapy is increased if (a) duration of therapy is greater than 24 months, (b) previous immunosuppressive therapy has been given, and (c) if the patient is seropositive for anti-JCV antibodies. Each factor independently increases the risk for PML. Patients who have anti-JCV antibodies being treated with natalizumab and who have had previous immunosuppressive therapy are at much higher risk of progressive multifocal leukoencephalopathy than those patients who had not received previous immunosuppression. The risk of natalizumab-associated PML after 2 years in JCV antibody positive patients who were not previously treated with immunosuppressive agents is 2.8/1,000 (95% CI: 2.0–3.8) and 8.1/1,000 (5.4–11.6) in those who have received previous immunosuppressives [45]. Given the efficacy of natalizumab in aggressive–relapsing MS with accumulating disability, one may be loathe to treat patients initially with immunosuppressive therapy if the increased risk for PML might prevent subsequent use of natalizumab.

There was no difference in time to confirmed progression between the two treatment arms. More patients deteriorated in the IVIG arm (77/159) than the placebo arm (70/159).

There were no differences between the two treatment arms in relation to performance on the 9-HPT, median increase in the EDSS scores, or on-study relapse rates.

Detailed MRI results were published separately and reported no differences in inflammatory disease activity (% enhancing scans, number of enhancing lesions, number of new or enhancing lesions between the two groups) [47]. There was significantly more cerebral volume loss in the placebo group than the IVIG arm (IVIG arm: −0.62% vs. placebo: −0.88%, p  =  0.009). Given the extent of inflammatory disease in both groups, this is unlikely evidence of a protective effect of IVIG. The authors postulated that the effect of IVIG on brain atrophy development of these patients might relate to an effect of IVIG on remyelination. However, magnetization transfer (MT) magnetic resonance imaging in a study subgroup found no significant differences between the two arms [48].

This was a completely negative study in that not only was the primary endpoint not achieved but also, in contrast to many of the phase III studies in SPMS, none of the secondary endpoints of inflammatory activity were positive. The authors commented that their study population was more similar to the SPECTRIMS than the European IFNB-1b SPMS study in relation to having fewer preceding relapses and at a more advanced stage of SPMS. There is no evidence that IVIG has any therapeutic effect in SPMS. A Cochrane Review in 2009 came to the same conclusion [49].