BOSTON — The first prospective study to look at long-term therapy with β-interferon and glatiramer acetate (Copaxone, Teva Neuroscience) in multiple sclerosis (MS) has concluded that these drugs are cost-effective over 6 years of use.

Results suggested that during this time period, use of β-interferon or glatiramer acetate slowed disability and improved quality of life.

The study was presented at MS Boston 2014, the 2014 Joint Americas and European Committees for Treatment and Research in Multiple Sclerosis (ACTRIMS/ECTRIMS) meeting, by Jackie Palace, MD, Oxford University Hospitals, United Kingdom (UK).

To evaluate long-term effects of β-interferon and glatiramer acetate, the researchers compared a cohort of UK patients taking these drugs to a matched natural history cohort of patients from British Columbia, Canada, on whom data had been collected before the drugs were introduced.

Dr. Jackie Palace

"Our data suggest that use of interferon β or glatiramer acetate is cost-effective over 6 years in relapsing-remitting MS according to the £36,000 per QALY [quality-adjusted life-year] threshold (set by the UK government) modeled over a 20-year period," Dr. Palace concluded. "The treatment effect was shown in all our sensitivity analysis and concordance was seen in independent statistical models. We feel that these results are robust and reliable."

The study received a mixed response, judging from questions during the question period following the presentation. For his part, Aaron Miller, MD, Mount Sinai Hospital, New York, New York, concluded, "It's encouraging to know we haven't been hitting our head against the wall."

But a member of the audience questioned whether this was the same dataset "that proved β-interferon didn't work, and was now being used to prove it did work?"

Explaining this comment for Medscape Medical News, Jerry Wolinsky, MD, who took questions from the media on some of the new research presented, pointed out that the same British Columbia natural history dataset had been compared in other studies with patients exposed to interferon in that same region without showing a benefit on disability.

"Those studies have suggested that while there is an effect on relapse and MRI outcomes, there appears to be very little effect on progression," he said. "The UK data are very striking. They got a very attractive outcome. But the big question is why is there such a difference to the similar Canadian studies." Dr. Wolinsky is the Bartels Family and Opal C. Rankin Professor of Neurology at the University of Texas Medical School at Houston and a member of the MS Boston 2014 steering committee.

Dr. Palace responded to Medscape Medical News that the major Canadian study, which was published in JAMA in 2012, did suggest an effect of β-interferon; however, the effect was not significant.

"It showed a hazard ratio of 0.77, which is nearly significant," she said. She added that that paper looked only at 1 Expanded Disability Status Scale (EDSS) outcome: time to reach an EDSS score of 6. "Whereas in our analysis progression over all EDSS scores will have been measured and this may make our analysis more sensitive to change."

Study Requested by UK Government

Dr. Palace explained that the current study was conducted at the request of the UK Department of Health to ensure the drugs were delivering a cost-effective outcome within the UK National Health Service.

Because randomized studies in MS are generally short in duration — normally 2 years maximum of randomized treatment — they cannot be used to tell what the drug is doing over the long term, Dr. Palace noted.

It was therefore decided to monitor data on more than 5000 UK patients from 72 centers who started these drugs over the long term with another natural history dataset of matched patients with MS who had not taken any drug therapy. "The natural history dataset was used to obtain a 'virtual placebo' trajectory," she said.

The natural history dataset was sourced from the British Columbia cohort before 1995, when the drugs were made available and for whom extensive information had been recorded. From that cohort, patients were selected who would have satisfied the UK eligibility criteria for β-interferon/glatiramer acetate had those agents been available. Patients were matched for entry disability with the UK patients receiving drug therapy.

The study assumed that if the effect size from the randomized trials was continued for 20 years, then the drugs might be cost-effective.

"Because disability progression in MS is usually slow, especially in the early years, it is very difficult to show cost-effectiveness in a 2-year study," Dr. Palace explained to Medscape Medical News. "But over a period of 10 or 20 years, an increase in disability is more likely to be detectable, and therefore there is a greater chance of showing a meaningful effect of the drug. For example, if the disability progression is delayed enough to keep enough people out of wheelchairs, then it is likely to be cost-effective — even if it takes 20 years to show this."

To compare the UK data with the British Columbia patient cohort, the researchers used 2 different statistical methods — the Markov model, which calculates transition probabilities for every individual EDSS score, and a multilevel model, which monitors individual trajectories according to baseline EDSS scores and can also give an average group curve. Both these models were analyzed independently by 2 different groups.

Target: 38% Reduction in Worsening Quality of Life

It was calculated that to hit the £36,000 per QALY target set by the UK, in aggregate, the drugs would need to show a 38% reduction in the worsening of quality of life vs the natural history cohort. Because the primary outcome was quality of life, the EDSS results were converted to utility scores (which measure quality of life).

Dr. Palace noted that this conversion was not linear — for example, if the EDSS score were reduced from 8 to 6, then that would give quite a large reduction in utility score, but if it were reduced from 6 to 3 there would be not such a great effect on quality of life. "So the results expressed in utility are not directly proportional to the EDSS," she stressed.

In her presentation, Dr. Palace reported 6-year data from the study. There were 4137 patients in the UK cohort and 898 in the matched British Columbian natural history dataset. She noted that the 2 patient groups were very similar, with a mean age at entry of 38 years and duration of disease of 8 years.

Patients in both groups had a mean of 3 relapses in the previous 2 years. EDSS score at baseline was slightly higher in the UK drug group (mean, 3.06) than in the natural history dataset (mean, 2.44), but this was adjusted for in the analysis.

The primary quality-of-life outcome at 6 years showed that the UK patients taking β-interferon or Copaxone did better than the natural history cohort. The target for the drug treated group for cost-effectiveness was a hazard ratio of 62% (a utility score reduction of approximately 0.06) or less, but they exceeded this by 10% in the Markov model and by 14% in the multilevel model, Dr. Palace stated.

Another way of expressing this result is that the drug-treated patients had an approximately 42% reduction in the change in utility score (a hazard ratio of around 58% vs the natural history group).

All the many supplementary and sensitive analyses showed better effects than in the untreated group, but not all met the cost-effectiveness targets.

Asked whether it was possible to look at each of the drugs individually, Dr. Palace replied. "This is a pooled analysis of the 3 different β-interferon products and glatiramer acetate. Each company has signed up confidentially with the Department of Health for their own individual targets regarding pricing. But that is confidential data."

Natural History Dataset Switched

During the discussion period, Dr. Palace noted that the researchers had first used a different natural history dataset — from Ontario — the disability results from which had been "smoothed," meaning patients weren't allowed to improve.

"We therefore had to apply the same rules to our patients, but when we did this, we saw a 'worse than placebo' effect, but when we took the smoothing rules out it was much better than predicted. This rule obviously created a bias."

The 2-year results from this comparison were published  in the BMJ in 2009. "We explained in that paper that we needed to find a different natural history dataset that didn't use smoothing rules — thus the British Columbian cohort was subsequently selected at most applicable to our cohort," Dr. Palace said.

"If a drug reduces relapses, patients are much more likely to have recovery from previous relapses," she elaborated. "Some natural recovery does happen. The smoothing of the data blocks that effect. There was a really logical reason to find another dataset."

Dr. Palace does advisory work for Biogen Idec, Merck Serono Ltd, Bayer Schering Pharma, Novartis Pharmaceuticals UK Ltd, Teva Pharmaceutical Industries Ltd, Gilenya, Ono Pharmaceutical Co Ltd, Primary i-research, Chugai Pharma Europe, and CI Consulting. She receives research support from the MS Society, QIDIS, Merck Serono Ltd, Novartis Pharmaceuticals, and Bayer Schering Pharma, plus conference expenses from Novartis and Merck Serono Ltd. Her trust is paid for her role in the RSS.

MS Boston 2014: 2014 Joint Americas and European Committees for Treatment and Research in Multiple Sclerosis (ACTRIMS/ECTRIMS). Abstract LB1.2. Presented September 13, 2014.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.