September 12, 2003
Clinical Drug Investigation 23(9):551-559, 2003
Mohammed K. Sharief, Guy's, King's and St Thomas' School of Medicine, Department of Clinical Neurosciences, Guy's Hospital, London, UK
Abstract and Introduction
Data reviewed in this article demonstrate that both interferon-ß-1a (IFNß-1a) and interferon-ß-1b (IFNß-1b) show a dose-response relationship in multiple sclerosis at current clinical doses. Moreover, the efficacy of these therapies is probably dependent on their frequent administration to maintain optimal levels of biological effect. These results differ from the conclusions of a recent review of data comparing two doses of IFNß-1a, given intramuscularly once weekly, that showed no difference in efficacy between the two doses. This and other results were interpreted to indicate that the current dose of this product (30µg intramuscularly once weekly) is optimal. However, the data reviewed here indicate that the efficacy of IFNß differs depending on the features of the dose regimen, including total dose, route and frequency of administration. Direct comparative data indicate that 44µg given subcutaneously three times weekly is more effective on clinical and magnetic resonance imaging measures, at least up to 48 weeks of therapy, compared with 30µg intramuscularly once weekly; risk-benefit favours 44µg three times weekly over 30µg once weekly.
In a recent review of clinical trials of interferon-ß (IFNß) in multiple sclerosis (MS), the author concluded that there is a dose-response effect for IFNß below a certain threshold dose. Above that threshold, it was suggested, there is a ceiling on the efficacy of these drugs and additional doses do not provide additional clinical benefit. This would be expected from normal drug pharmacokinetics. However, the review, in the absence of head-to-head comparisons, concluded that there are no significant dose-effects among the current clinically available doses of IFN. The review appears to be supported by the recently presented results from a study comparing two doses of IFNß-1a (Avonex®, Biogen, Cambridge, MA, USA), 30µg and 60µg, given intramuscularly once weekly to patients with MS. The data showed no differences in the efficacy of these doses on the primary efficacy parameter the progression of disability. However, these conclusions are not supported by preclinical or clinical data, which demonstrate that once-weekly therapy is suboptimal compared with more frequent administration. In addition, both preclinical and clinical data have provided clear evidence that the effects of IFNß in MS are dose-related. Therefore, it seems likely that this apparent lack of a dose-response effect for IFNß-1a, 30 and 60µg once weekly, is related to the frequency of dosing, rather than to a true ceiling effect. Data on the effect of dose and frequency of administration on the efficacy of IFNß in relapsing-remitting (RR) MS are reviewed here.
All the IFNßs bind to specific cell-surface receptors. Binding produces a cascade of effects, resulting in the production of a number of proteins that have antiviral, antiproliferative and immunomodulatory effects. The mechanism of action of IFNß specific to MS remains unclear. Possible factors include effects on the blood-brain barrier and on T-cell activation, resulting in a net anti-inflammatory effect (by inhibiting Th1 or promoting Th2 cytokine production). However, recent evidence indicates that IFNß can also up-regulate some pro-inflammatory gene products, and this has led to the proposal that an inflammatory response may have neuroprotective effects on MS lesions, or that some of the effects of IFNß depend on a subtle balance between anti- and pro-inflammatory effects.[3, 4] Thus, the situation is a highly complex one. Nevertheless, dose- and frequency-related effects have been demonstrated for IFNß on several markers that are thought to be relevant to its effects in MS.
Effects of Frequency of Administration
Preclinical data have demonstrated that frequent dosing with IFNß produces more pronounced effects than once-weekly administration. In healthy volunteers, single doses produced increases in serum neopterin, MxA protein, ß2-microglobulin and interleukin (IL)-10 for about 2-4 days, which then declined to baseline levels after 7-8 days.[5, 6] IFNß-1b (Betaseron®, Schering, Berlin, Germany), 8 MIU, given subcutaneously every other day, had significantly greater overall effects on biological response markers than IFNß-1a, 30µg intramuscularly once weekly. The changes in serum levels of neopterin, ß2-microglobulin and IL-10 produced by the two treatments were similar over the first 4 days of testing. However, the effects of IFNß-1a 30µg intramuscularly once weekly then declined, whereas those of IFNß-1b 8 MIU subcutaneously every other day were maintained, producing greater overall effects on all these markers.
Additional evidence for an increase in the biological effects of IFNß associated with frequent dosing was produced in a study of the pro-inflammatory cytokines IFNß and tumour necrosis factor-a (TNFa). IFNß-1a given to healthy volunteers at doses of 22µg and 66µg subcutaneously once weekly for 1 month produced similar effects on these markers. However, dividing the 66µg dose into doses of 22µg given three times weekly increased the inhibition of pro-inflammatory cytokine levels (Figure 1). The results of pharmacodynamic testing indicate that dosing several times a week (every other day or three times weekly) produces greater and more sustained effects than once-weekly administration of IFN.
Figure 1. Once-weekly (qw) doses of interferon-ß-1a (IFNß-1a) for 1 month showed little evidence of a dose-response effect in suppressing the immunomodulatory cytokines interferon-g (IFNg) and tumour necrosis factor-a (TNF). However, dividing the dose into three-times-weekly (tiw) doses produced significantly greater effects.
Effects of Interferon-ß Dose
Several of the preclinical studies described in section 1.1 have demonstrated dose-response effects of IFNß on several biological-response markers.[5, 6, 9] Thus, in healthy volunteers, increasing single doses of IFNß-1b, given subcutaneously, and IFNß-1a, given both subcutaneously and intramuscularly, resulted in progressively increased levels of neopterin, ß2-microglobulin, MxA protein and 2', 5'-oligoadenylate synthetase.[5, 6] In addition, IFNß-1a produced dose-dependent increases in the anti-inflammatory cytokine IL-10 in both patients and healthy volunteers.
Clinical Evidence for Effects of Interferon-ß Dose
In RRMS, the early course of the disease is characterised by intermittent relapses, and partial or complete recovery of function between exacerbations. The occurrence of relapses is a key measure of disease activity. In the Prevention of Relapses and disability by IFNß-1a Subcutaneously in Multiple Sclerosis (PRISMS) study, placebo or IFNß-1a (Rebif®, Serono, Geneva, Switzerland) at doses of 22 or 44µg subcutaneously three times weekly was given to 560 patients with RRMS for up to 4 years.[10, 11] In the IFNB Multiple Sclerosis Study Group trial, IFNß-1b (Betaseron®) was administered at doses of 1.6 and 8 MIU subcutaneously every other day to 372 patients for up to 5 years.[12, 13] In both trials, IFNß was more effective than placebo, and there were benefits in efficacy that favoured the higher dose of drug in each study.
In the PRISMS trial, after 2 years, the relapse count was significantly lower in both IFNß-1a treatment groups, compared with the placebo group (1.82 for 22µg, 1.73 for 44µg and 2.56 for placebo, p < 0.005). In addition, disability progression was significantly delayed by active treatment at both doses (time to progression was 18.5 months for 22µg, 21.3 months for 44µg and 11.9 months for placebo, p < 0.05). The difference between doses was not statistically significant on clinical measures but was significant for the active lesions seen on magnetic resonance imaging (MRI).
The PRISMS study was extended for a further 2 years, and patients taking placebo were re-randomised to one of the doses of IFNß-1a in a blinded fashion. Patients already receiving active medication continued receiving the same blinded dose. After 4 years, efficacy was maintained on all measures, and there was increasing evidence of treatment effect differences between the 22µg three-times-weekly and the 44µg three-times-weekly groups. Relapse rate was 0.72 for 44µg three times weekly and 0.80 for 22µg three times weekly. This difference did not achieve statistical significance (p = 0.069), but reflects an additional 10% reduction compared with years 1-2 for high-dose relative to low-dose therapy. Considering disability measures, only the high dose over 4 years was significantly more effective than the placebo regimen of placebo for 2 years followed by active therapy for 2 years, for the time to first progression and area under the curve (AUC) of Expanded Disability Status Scale (EDSS) score. The relative improvement of high-dose compared with low-dose therapy for time to progression was 15% (hazard ratio [HR] = 0.85, p = 0.33), for the proportion of patients remaining progression free was 21% (odds ratio [OR] = 0.79), and for number of EDSS progressions was 24% (OR = 0.76, p = 0.030). Although not all measures showed a statistically significant difference between 44µg three times weekly and 22µg three times weekly, the magnitude of difference was clinically important and, as noted earlier, the study was not powered for such outcomes and particularly not between two doses of an effective therapy.
During the extension period alone, dose-related effects on relapse and disability outcomes, as well as MRI parameters, were demonstrated. The rate ratio for relapse during years 3-4 was 0.79 comparing 44µg three-times-weekly with 22µg three-times-weekly groups (p = 0.025). The time to first progression during years 3-4 using a new baseline at the start of year 3 was also significantly longer for the high-dose than for the low-dose group (HR = 0.64, p = 0.036).
In the IFNB study, rates of exacerbations after 2 years were 1.27, 1.17 and 0.84 for the placebo, low-dose and high-dose groups, respectively. After 5 years, the pooled annual exacerbation rates likewise showed numerical differences between the dose groups: rates were 1.12, 0.96 and 0.78 for placebo, low-dose and high-dose groups, respectively. IFNß-1b has not shown a significant effect on disability.
Magnetic Resonance Imaging Parameters
It has been argued that MRI is a more sensitive marker of the underlying disease process in MS than clinical parameters because it provides a measure of subclinical disease activity.[14-16] A positive treatment effect on MRI lesions may have important prognostic implications for patients, as early inflammatory events are now thought to result, at least in some lesions, in axonal transection, leading to permanent damage and a progressive decline in function.[17, 18]
In the first 2 years of the PRISMS trial, the higher dose of IFNß-1a (44µg three times weekly) was significantly more effective than the 22µg three times weekly dose in reducing the number of active T2 MRI lesions. There were median values of 0.75 lesions per patient per scan for the 22µg three times weekly dose and 0.50 lesions per patient per scan for 44µg three times weekly (p = 0.0003).[19, 20] Over 4 years, both active-treatment groups showed significant differences in MRI activity compared with those groups that had received placebo for the first 2 years. In addition, the 44µg group developed fewer new T2 lesions than the 22µg group (0.5 versus 1.3 lesions per patient per scan, p < 0.0001). Moreover, the 44µg three-times-weekly group showed a reduction in MRI burden of disease, compared with an increase in the 22µg three-times-weekly group (3.4% increase for 22µg three times weekly and a 6.4% decrease for 44µg three times weekly, p = 0.009). Likewise, in the IFNB study of IFNß-1b, the change in MRI disease area favoured high-dose compared with low-dose therapy, although the statistical significance of this difference was not reported.
Effects of Dose and Frequency of Administration
In the Once Weekly Interferon for MS (OWIMS) trial, IFNß-1a was administered at doses of 22 and 44µg subcutaneously once weekly. The primary outcome measure was the number of combined unique active lesions on MRI after 24 weeks of treatment. This showed a significant difference only between IFNß-1a 44µg once weekly and placebo, and a numerical (but not statistically significant) difference between 22µg once weekly and placebo, and between the two doses of IFNß-1a. Likewise, the relative reduction in relapse rate between placebo and active treatment groups after 48 weeks was 0% for 22µg once weekly and 19% for 44µg once weekly. The results were not statistically significant for either dose compared with placebo, or between 44 and 22µg given once weekly, although clinical parameters were a secondary outcome measure and the study was therefore not powered to detect such differences.
The results of the OWIMS study were compared with those from the PRISMS trial and the Mul-tiple Sclerosis Collaborative Research Group (MSCRG) trial of IFNß-1a 30µg once weekly,  notwithstanding the limitations of comparing data across studies. Comparing data on reduction in relapse rates at 1 year showed that the data from the MSCRG trial falls between the results for the 22 and 44µg once-weekly groups in the OWIMS study. These reductions were smaller than those produced by 22 and 44µg three times weekly in the PRISMS trial. Thus, all the doses given once weekly were less effective than those given three times weekly in the PRISMS trial, indicating that once-weekly dosing was suboptimal compared with the higher and/or more frequent dosing regimens.
Further comparison of the effects of Rebif® on MRI parameters in the OWIMS and PRISMS trials adds support to the idea that higher and/or more frequent dosing of IFNß-1a produces greater therapeutic effects. Comparison of the clinical results from the pivotal disease-modifying therapy trials demonstrates significant benefits for the IFNß-1a 44µg three-times-daily dosing regimen (Figure 2).[10, 12, 20, 22-24] In addition to these comparisons across trials, two recent, direct comparative studies of different IFNß regimens have added supporting evidence that higher and/or more frequent doses of the drug produce greater clinical effects.
Figure 2. Plotting the results from key clinical trials of disease-modifying therapies for multiple sclerosis demonstrates good benefits in prevention of both relapses and disease progression for interferon-ß-1a (IFNß-1a) [Rebif®] 44µg three times weekly (tiw), [10, 20] compared with less frequent and/or lower doses of IFNß-1a (Avonex®) 30µg once weekly (qw), [20, 22] interferon-ß-1b (IFNß-1b) [Betaseron®] 8 MIU every other day (eod), [12, 20] and glatiramer acetate (GA; Copaxone®) 20mg once daily (od).[23, 24]
In the INdependent COMparison of INterferon (INCOMIN) trial, IFNß-1b 250µg (8 MIU) given subcutaneously every other day, was compared with IFNß-1a 30µg intramuscularly once weekly in an open-label, randomised trial. Over the 2 years of the study, there was a significant difference in favour of IFNß-1b in the primary outcome (unblinded assessment of the number of patients remaining free of relapses, p = 0.03). There was also a significant difference in favour of IFNß-1b in the primary MRI outcome (blinded assessment of the proportion of patients remaining free from new T2 lesions, p < 0.0003). The authors concluded that it is not possible to differentiate in these results between the effects of cumulative doses of these drugs and of frequency of their administration, and that both may affect the efficacy of these therapies in patients with MS.
Another direct comparison study, the Evidence of Interferon Dose-response: European North American Comparative Efficacy (EVIDENCE) study, compared IFNß-1a (Rebif®) 44µg subcutaneously three times weekly (twice the biological activity of IFNß-1b 8 MIU every other day) with IFNß-1a (Avonex®) 30µg intramuscularly once weekly at 24 weeks (primary endpoint) and 48 weeks (follow-up). Outcomes were assessed by investigators blinded to the patients treatments. This study demonstrated statistically significant benefits on the relapse and MRI outcomes for 44µg three times weekly over 30µg once weekly at both timepoints. There was a 10% absolute increase in the proportion of patients who remained free of relapses at 48 weeks that equated to a 19% reduction in the risk of relapse for 44µg three-times-weekly compared with 30µg once-weekly therapy. The HR for time to relapse was 0.70, a 30% relative reduction for patients treated with 44µg three times weekly compared with 30µg once weekly over 48 weeks. Based on the 24-week primary outcome and supported by the 48-week follow-up data, the US FDA approved 44µg subcutaneously three times weekly (Rebif®) for use in the US based on superior clinical efficacy over 30µg intramuscularly once weekly (Avonex®).
Safety of Interferon-ß
Increases in the efficacy of a therapy are only clinically beneficial if they are associated with acceptable tolerability and safety. It is therefore important to compare the incidence of side effects and the proportion of patients who develop neutralising antibodies (NAbs) during treatment with different doses and different forms of IFN.
The adverse events most directly attributable to treatment with IFNß are 'flu-like symptoms, injection-sitereactions, asymptomatic reduced white blood cell counts and asymptomatic elevation of liver enzymes. 'Flu-like symptoms are associated with all IFNß products, with a reported incidence during 2 years of therapy of 59% in patients receiving IFNß-1a 44µg three times weekly, compared with 51% of patients receiving placebo in the PRISMS study,  and 61% for those receiving IFNß-1a 30µg once weekly and 40% for those receiving placebo in the MSCRG study. The incidence of 'flu-like symptoms was not reported in the 2-year report of the IFNB study, but fever was reported in 58% of patients in the 8 MIU group, compared with 34% of patients in the placebo group; and corresponding incidences of myalgia were 41% and 24%, respectively. In the EVIDENCE study, after 48 weeks, fewer patients taking 44µg three times weekly (42%) than those taking 30µg once weekly (49%) experienced 'flu-like symptoms (p = 0.09), probably due to gradual dose escalation and a more sustained level of IFNß effect with less striking peaks and troughs in three-times-weekly than once-weekly dosing.
Injection-site reactions have various definitions in different studies but are significantly more common following subcutaneous than intramuscular injection. In the placebo-controlled studies, the proportion of patients with injection-site reactions was 35% for the IFNß-1a 44µg three-times-weekly group and 13% for the placebo group in the PRISMS study,  69% for the IFNß-1b, 8 MIU every-other-day group and 6% for the placebo group in the IFNB study,  and 10-15% of both the IFNß-1a 30µg once-weekly and placebo groups in the MSCRG study. In the direct comparison study of IFNß-1b 8 MIU every other day and IFNß-1a 30µg once weekly, the incidence of injection-site reactions was 37% for IFNß-1b and 8% for IFNß-1a. In the adverse event assessment by the unblinded treating physician in the IFNß-1a EVIDENCE study, the proportion of patients with any skin reaction was 83% for those receiving 44µg three times weekly, and 28% for those on 30µg once weekly. For injection-site reaction alone, the comparative values were 35% for those receiving 44µg three times weekly, and 12% for those on 30µg once weekly. Most adverse events related to IFNß are mild in severity, occur usually within the first 3-6 months of therapy, and generally diminish in frequency and severity over time.
Asymptomatic laboratory abnormalities occurred frequently in both treatment arms. Reports of abnormal liver function tests occurred in 18% of the 44µg three-times-weekly group compared with 9% of patients in the 30µg once-weekly group (p = 0.002). However, the majority of events were mild (WHO grade 1), and all either resolved spontaneously or responded to dose reduction. Only three patients from the 44µg three-times-weekly group and one patient from the 30µg once-weekly group withdrew from the study due to raised transaminase levels. White blood cell abnormalities were reported in 11% of the 44µg three-times-weekly group and in 5% of the 30µg once-weekly group (p = 0.002).
The effect of NAbs on the clinical efficacy of IFNß is a very complex issue that is likely to be influenced by several different factors and is not yet fully understood. Indeed, a recent report, presented by the American Association of Neurology (AAN) on disease-modifying therapies in MS, concluded that the biological effect of NAbs is uncertain. However, the US FDA has recently reviewed data on NAbs from the EVIDENCE trial. These data showed that, although there was a significantly greater incidence of NAbs in the IFNß-1a 44µg three-times-weekly group compared with the 30µg once-weekly group, there was no association between patients antibody status and their probability of remaining exacerbation-free. The longest-term data on the effects of NAbs on clinical and MRI outcomes was presented in the 4-year results of the PRISMS study, showing that patients who developed NAbs had significantly greater MRI-active lesion counts, an increase in lesion burden and, for 44µg three-times-weekly patients only, higher relapse rates during years 3 and 4 of therapy. However, NAb-positive patients had better outcomes over 4 years than patients who received placebo for 2 years followed by IFNß-1a for 2 years.
This review demonstrates that the effect of IFNß on various pharmacodynamic markers is both dose- and frequency-related. In addition, several clinical trials have shown that the effect on clinical measures is dose- and/or frequency-related. The mechanism of action of IFNß specific to MS is not clear, therefore the precise relationship between preclinical findings and MS clinical outcomes is also unclear. Nevertheless, both preclinical and clinical data point to the same conclusion with regard to the importance of dose and frequency of IFNß dosing in MS. Clinical data from the several clinical trials showed that there is a dose-response effect for IFN, that a treatment effect compared with placebo is apparent within a few months, and the dose-effect at higher doses becomes apparent over longer durations of treatment.
In contrast to the above, there was no evidence of a dose-response in the study of weekly intramuscular IFNß-1a at doses of 30 and 60µg. One potential caveat is the absence of a placebo group to gauge the actual treatment response of either dose. Furthermore, administering the drug once weekly is an important factor, as other data have shown that this is less effective than more frequent dosing. However, using doses of 22 and 44µg once weekly, some nonsignificant differences were noted,  but no differences, except mild MRI differences, were seen between doses of 30µg and 60µg once weekly. This may suggest that, for weekly dosing, a maximal response is seen between 30µg and 44µg such that increasing the dose to 60µg affords no further benefit if given weekly. The other issue in the recent comparison of once-weekly doses is that the primary outcome measure was disability, a less sensitive measure for treatment effect. PRISMS showed that the time to first disease progression by 1.0 EDSS point was different, although not significantly so, after 4 years between 22µg three times weekly (66µg weekly) and 44µg three times weekly (132µg weekly). Of note is that total disability over 4 years, as assessed by number of EDSS changes, was significantly better for the 44µg three-times-weekly group than the 22µg three-times-weekly group. Relapses are a more sensitive measure of treatment effect, but relapses were not assessed adequately in the 30µg/60µg once-weekly comparative study, as they were assessed by patient report only.
The test of the dose frequency hypothesis involves direct comparative studies of different regimens. Studies comparing IFNß-1a 44µg three times weekly with IFNß-1a 30µg once weekly and IFNß-1b 8 MIU every other day with IFNß-1a 30µg once weekly have both shown clinical and MRI advantages for the comparator drugs over the low dose of IFNß given intramuscularly once weekly.[25, 27] It is not possible in these trials to differentiate between the effects of increased dose and increased frequency of administration of IFN, but it seems probable that both of these factors contributed to the greater clinical efficacy of the high-dose, frequent administration regimens.
The sum of available data indicates that dose frequency more than once weekly gains optimal efficacy from IFNß therapy, and the suggestion that there is no dose-response relationship for the IFNßs at current clinical doses is not supported by available data. It seems likely that the apparent lack of a dose-response curve in the trial of two doses of once-weekly IFNß-1a resulted not from ceiling effects of the dose, but probably from a suboptimal frequency of administration.
The enhanced efficacy of higher, more frequent dosing is not attained at the expense of safety. Despite larger proportions of adverse events with higher doses, in the direct comparative study of IFNß-1a 44µg three times weekly and 30µg once weekly, the difference in withdrawals due to adverse events was minimal (4.7% for 44µg three times weekly and 4.2% for 30µg once weekly at 48 weeks) and the overall adherence to therapy was also similar (7% versus 6% treatment discontinuations, respectively). Optimal efficacy requires frequent dosing and is not associated with clinically meaningful tolerance differences; risk-benefit favours 44µg three times weekly over 30µg once weekly.
The author thanks Mrs Margaret Jones and Dr Muhanad Noori for assistance in the preparation of the manuscript.
The author has received fees for lecturing and consulting from Serono, Biogen, Schering, Teva, as well as from other pharmaceutical companies, and receives grants for clinical trial work from various sources.
The use of tradenames is for product identification purposes only and does not imply endorsement.
Dr. Mohammed K. Sharief, Guy's, King's and St Thomas' School of Medicine, Department of Clinical Neurosciences, Guy's Hospital, Hodgkin Building, London, SE1 1UL, UK.
Copyright © 2003, Adis Data Information BV