I have been watching this forum for some time. After 7 years of betaseron and a bit of brain shrinkage and a few more lesions, my neuro switched me to rebif. With tysabri being more effective (apparently) than anything else, I wanted to know how others were doing which led me here. I was curious about how the NMSS would respond to this study. So I sent an email and have copied the response below. How can any of us know what is really effective? Is tysabri really this effective:
At the end of a 2-year study, TYSABRI reduced how often flare-ups occurred by 67% compared with placebo. The average number of flare-ups a patient had each year was 0.22 for TYSABRI and 0.67 for placebo
Significantly more people were relapse free at 2 years with Rebif 44mcg vs placebo (32% vs 15%) http://www.mslifelines.com/rebif/guide/effectiveness.jsp
I guess relapse and flareup are different? So if both are comparable, what are they comparing?
This was the response I got from the NMSS:
“There is abundant proof, published in peer-reviewed journals, that the approved disease-modifying therapies are significantly better than the inactive placebos against which they were tested. Building evidence also suggests that they not only impact relapses but also may slow the progression of disease. There is no substitute for well-designed, placebo-controlled studies that are done for extended periods of time. Most statisticians believe that it is very difficult to combine results from studies, such as was done for this article. The problem with mushing the results together is that each study used different patient populations (with differences in their recent relapse experience, age, time from disease onset, and many other variables) and different outcome measures and different timeframes of study. At the recent American Academy of Neurology meeting in Chicago this spring, a statistician (Dr. Gary Cutter, U Alabama) looked at this very issue and concluded, among other things, that head-to-head studies was the only way to compare results. “
Here is Dr. Cutter's abstract:
S02.001] Statistical Analysis of Clinical Endpoints in Studies of Disease-Modifying Therapies for Multiple Sclerosis
Gary Cutter, Birmingham, AL, Maria Sormani, Milan, Italy, Amy Pace, Hao Zhang, Cambridge, MA
OBJECTIVE: To evaluate statistical methodologies used to analyze clinical efficacy endpoints in placebo-controlled studies of multiple sclerosis (MS) disease-modifying therapies (DMTs). BACKGROUND: Interpretation of MS clinical study outcomes are complicated by variations in patient characteristics, study designs, and definitions/analyses of clinical outcomes. As new clinical trial results from studies of DMTs become available, the treating neurologist faces the challenge of interpreting these results to optimize patient care. DESIGN/METHODS: This review evaluates common definitions of relapse and disability endpoints and statistical methods used to analyze them in clinical studies of MS DMTs. Advantages and limitations of each statistical method are considered. RESULTS: Relapses have not been operationally defined, and the definition of disability progression based on Expanded Disability Status Scale score (which is often a noisy variable) differs among clinical studies. Treatment effects on relapse or disability endpoints may be estimated via relative risk reductions (RelRR), a measure of the difference in outcome on active treatment relative to placebo, or from rate ratios estimated from statistical models. RelRRs exhibit less variability across populations with different baseline risks, but do not discriminate between small and large treatment effects. Treatment effects on clinical endpoints may also be estimated via absolute risk reductions (AbsRR) and number needed to treat (NNT) to prevent one undesired outcome, a derivative of AbsRR. AbsRRs are derived from absolute differences between active treatment and placebo outcomes, and thus are highly dependent on placebo rates or proportions. The AbsRR lends itself to interpretability, although differences in patients baseline activities may inflate or diminish AbsRR outcomes. CONCLUSIONS/RELEVANCE: Differences in inclusion criteria, study patient populations, and placebo-group behavior make direct comparisons of treatment effects across studies difficult to interpret. AbsRRs/NNTs are not recommended for making comparisons between treatment effects among studies with different baseline risks or populations. Supported by: Biogen Idec. Editorial support provided by Scientific Connexions.
Category - MS and Related Diseases
SubCategory - Clinical Science
Additionally, I have included a link to our expert opinion paper on the disease modifying drugs.