Results from the present study suggest that the risk of having a relapse despite the increase of IFNB dose raised according to the number of clinical attacks occurred during the assumption of the low-dose IFNB, and it is reduced when the switching was based on MRI findings rather than on clinical activity. Moreover, the chance of being progression-free after the switch is related to a lower EDSS score and the absence of clinical and MRI activities. At this regard, we may assume that in some patients an incomplete recovery from relapse occurred, with a substantial influence on their EDSS scores.
Overall, less than 1/3 of our patients remained free from clinical disease activity (i.e. absence of relapse and sustained progression on EDSS score) over the 2-year observational period following the increase of IFNB dose. This could imply that the majority of switchers should be considered poor responder to IFNB therapy regardless the dose and/or frequency of administration, and further supports the hypothesis that a treatment strategy encompassing the increase of IFNB dose should be useful only in some selected cases.
After the increase of IFNB dose, the majority of patients switched only for the evidence of MRI activity had a better clinical outcome than those switched because of the occurrence of relapses. Therefore, we might suggest that monitoring the effect of IFNB treatment with regular MRI scans is recommendable even in absence of clinical relapses.
It has been known that conventional MRI represents a powerful tool to monitor latent disease activity, providing a measurable and sensible marker of response to IFNB therapy [22–25]. However, we cannot exclude that patients switched only on the basis of MRI findings might also have had good outcomes without switching to the high-dose IFNB, as the absence of a control group. At this regard, Rio and colleagues showed that patients with only MRI activity and no relapses in the first year of IFNB treatment did not experience an increase of relapses or disability over a 3-year follow-up .
Although randomized clinical trials demonstrated a more pronounced effect of high-dose, high-frequency IFNB when compared with both the low-dose, equal-frequency [1, 27] and the low-dose, low-frequency regimens [14, 15] in naïve patients, data on the effectiveness of increasing the IFNB dose in patients with breakthrough disease are scarce. The open-label extension phase of the EVIDENCE study, involving 223 patients converted from Avonex to Rebif 44, documented a 50% reduction in the annualised relapse rate after the switch . However, we must consider also that in the EVIDENCE study all patients originally randomized to Avonex were offered to receive Rebif 44, independently from the response status during the blind phase of the study. While the authors suggested that increasing the IFNB-1a dose and frequency could rapidly reduce ongoing disease activity, they cannot discharge the hypothesis that the significant reduction in relapse rate might be due, at least in part, to the regression to the mean phenomenon.
Some open-label studies exploring the usefulness of switching among immunomodulating drugs had different designs and provided conflicting results [28–30]. Two studies reported a decrease in both relapse rate and proportion of relapse-free subjects after the switch [29, 30], whilst the QUASIMS study did not provide any support of more favourable outcomes after switching from an IFNB formulation to another . One possible explanation of this discrepancy is that these observational surveys considered different subtype of switch (i.e. from IFNB-1a to IFNB-1b and viceversa, from IFNB to GA, etc.); also, these studies were not specifically aimed to determine the crude effect of an increase of the IFNB dose. Furthermore, in some studies the patients were switched on the basis of tolerability rather than a persistent disease activity.
In the present study, patients had variable periods of observation before and after the switch, thus precluding any attempt to estimate the effectiveness of an increase of the IFNB dose in suppressing disease activity and slowing disability progression.
Being an observational report, our study suffers from other limits, such as the small sample size, the unavailability of control group, blindness and randomization, as well as the lack of data on neutralising antibodies (NAbs) against IFNB. However, there is evidence that NAbs presence could explain only the 20% of the suboptimal response to IFNB treatment .
Despite these limits, our study might contribute to define a therapeutic algorithm to manage breakthrough disease in patients on treatment with a low-dose IFNB. The identification of patients with subclinical disease activity during the low-dose IFNB treatment, and an early switch to the high-dose IFNB, seem to be effective in achieving a better control of the disease. On the contrary, when relapses occurred during the pre-switch period, especially in combination with MRI activity, patients did not seem benefit from the increase of the IFNB dose in the following years.
Since efficacy of GA has been demonstrated comparable to IFNB [32–34], switching among immunomodulating treatments may represent an interesting approach in case of treatment failure [35, 36]. The scientific rationale for switching to other therapies is strongest for patients on IFNB therapy with persistent high-titre of NAbs . However, a more aggressive approach (Natalizumab or Mitoxantrone) is warranted for patients at high risk of accumulation of fixed disability or with shorter intervals between attacks .