The appraisal committee (appendix A) considered evidence submitted by the manufacturer of natalizumab and a review of this submission by the evidence review group (ERG) (appendix B).
In its submission, the manufacturer compared natalizumab with beta interferon, glatiramer acetate and best supportive care (that is, no active treatment) for both the RES and the suboptimal therapy groups. The two major clinical outcomes examined were disability progression, defined as an increase in the expanded disability status scale (EDSS) score sustained for 12 or 24 weeks at 2 years, and annualised relapse rate.
The manufacturer presented data from the multinational, double-blind, randomised AFFIRM study (n=942), which compared natalizumab with placebo. The study comprised people with relapsing–remitting multiple sclerosis, of which a subgroup had highly active relapsing–remitting multiple sclerosis. A post-hoc subgroup analysis of AFFIRM (n=209) provided clinical data for the RES group. The marketing authorisation for the suboptimal therapy group was based on data from the SENTINEL study (n=1,171), which compared natalizumab and beta interferon with beta interferon alone. However, the combination of natalizumab with beta interferon is not included in the marketing authorisation for natalizumab because of concerns over the risk of PML, and data from the SENTINEL study were not presented by the manufacturer. Instead, the manufacturer assumed that the intention to treat (ITT) population from AFFIRM is a suitable proxy for the suboptimal therapy group. The manufacturer provided additional data from two phase 2 studies. The manufacturer did not identify any studies that compared natalizumab with beta interferon or glatiramer acetate.
The AFFIRM study demonstrated that natalizumab statistically significantly reduces the probability of sustained disability progression compared with placebo in both the ITT and RES populations. The hazard ratios (HRs) varied between 0.46 and 0.58 in the ITT population, depending on the measure of disability progression (increase in EDSS sustained for 24 and 12 weeks respectively), and between 0.36 and 0.47 in the RES group. In addition, natalizumab led to a reduction in relapse rate, with a relative risk reduction of 0.68 in the ITT population and 0.81 in the RES group. The manufacturer presented evidence that showed that, compared with placebo, natalizumab significantly improved health-related quality of life when measured with the SF‑36 instrument, although not when the MSQLI instrument was used.
Given the absence of study data comparing natalizumab directly with beta interferon and glatiramer acetate, the manufacturer carried out an indirect comparison. This adopted an existing method to compare the results of AFFIRM with systematic reviews of beta interferon and glatiramer acetate. The systematic reviews included people with relapsing–remitting multiple sclerosis rather than highly active relapsing–remitting multiple sclerosis and did not specifically examine the clinical effectiveness of the drugs in the RES or suboptimal therapy groups. Therefore, the manufacturer assumed that the treatment effect of beta interferon and glatiramer acetate in relapsing–remitting multiple sclerosis was equivalent to that in the RES and suboptimal therapy groups. The results of the indirect analysis showed that natalizumab was associated with a statistically significant reduction in relapse rates compared with beta interferon and glatiramer acetate, with relative risks of 0.63 and 0.57 respectively for the ITT population and 0.49 and 0.43 respectively for the RES group. The results of the indirect analysis for disability progression were submitted to NICE in confidence.
The AFFIRM study showed that natalizumab is not associated with a higher incidence of adverse events than placebo. The indirect comparison performed by the manufacturer found no statistically significant differences in adverse events between natalizumab and glatiramer acetate. However, compared with beta interferon, natalizumab was found to be associated with a statistically significant reduction in the incidence of influenza-like symptoms and myalgia/arthralgia, with relative risks of 0.47 and 0.68 respectively.
The manufacturer presented a multistate Markov model based on the economic model developed by the School of Health and Related Research (ScHARR) at Sheffield University that was used in NICE's technology appraisal on beta interferon and glatiramer acetate for the treatment of multiple sclerosis. The manufacturer's model predicts disability progression and disease activity over a time horizon of 20 years using a series of 1-year cycles. The model took an NHS perspective for the majority of costs, but included carers' disutility in the base case.
The clinical data that populate the manufacturer's model come from the AFFIRM study and the systematic reviews of beta interferon and glatiramer acetate. Additional data on disability progression were derived from the London Ontario data set (a longitudinal study of more than 1,000 people with relapsing–remitting multiple sclerosis followed for a mean of 25 years). Data on costs and utilities (based on EQ-5D scores) associated with EDSS states were derived from a cross-sectional study (the UK MS survey) commissioned by the manufacturer. This survey included people with relapsing–remitting, secondary progressive and primary progressive multiple sclerosis, and the results were based on 2,048 responses (a 15.8% response rate).
The results of the manufacturer's analysis showed that the incremental cost-effectiveness ratios (ICERs) for the RES group compared with best supportive care, beta interferon and glatiramer acetate were £44,600, £32,000 and £34,600 per quality-adjusted life year (QALY) gained respectively. For the suboptimal therapy group the ICERs were £56,100, £43,400 and £44,300 per QALY gained respectively.
Sensitivity analysis demonstrated that the variables that had the greatest effect on the ICERs were the time horizon over which costs and outcomes are evaluated and changing the source of the disability progression data from AFFIRM to the London dataset. Extending the time horizon to 30 years, for example, reduced the ICERs for natalizumab versus beta interferon to £24,600 and £34,200 per QALY gained in the RES and suboptimal therapy groups respectively. In contrast, changing the source of the disability progression data from AFFIRM to the London Ontario dataset increased the ICERs to £42,300 and £55,300 per QALY gained for natalizumab versus beta interferon in the RES and suboptimal therapy groups respectively.
The ERG expressed a number of concerns about the manufacturer's submission. The ERG recognised the general uncertainty associated with indirect analyses and that the data for the comparators was derived from people with relapsing–remitting multiple sclerosis rather than highly active relapsing–remitting multiple sclerosis. The ERG stated that this might alter the conclusions of the analysis, although the magnitude and direction of any such effect was unknown.
The ERG recognised that the approach adopted by the manufacturer in its economic modelling was pragmatic given the absence of better quality data. However, it expressed concern about the extrapolation of 2-year data from the AFFIRM study to a 20-year time horizon. The ERG also expressed concern that the utility and cost data, which were based on the UK MS survey, were not exclusively derived from people with highly active relapsing–remitting multiple sclerosis; in addition, the survey may not have been representative because of the low response rate.
The ERG commented on the limitations of the EDSS instrument, which suffers from limited responsiveness and inter- and intra-rater variability. In addition, the ERG expressed concern that, although the transition probabilities in the manufacturer's model were based on data from AFFIRM, the model appeared to predict a higher rate of sustained disability progression at 2 years than reported in AFFIRM. The ERG stated that this might overestimate the effectiveness of natalizumab, and might therefore lead to more favourable ICERs in the model. The ERG also highlighted the limited evidence for the assumption in the manufacturer's model that natalizumab reduces progression from relapsing–remitting multiple sclerosis to secondary progressive multiple sclerosis.
Full details of all the evidence are in the manufacturer's submission and the ERG report.