Why have annualized relapse rates in RRMS trials decreased?
Take a look at our latest data visualization. It appears that there's been a substantial decrease in annualized relapse rates (ARRs) among patients in the placebo arm of RRMS trials over the last 2 decades. I'm interested in what may account for this change, and how those hypotheses may be tested. Please add your ideas below.
It seems to me from the plot that the larger effect may be that trials with larger enrollment have smaller mean ARR. Because those trials are more recent, your model (which does not specify a term for trial size) captures that in the time term.
Also, I could not replicate your fit line from the equation in the post. Was there some pre-processing step in the data, like a normalization, or subtracting out the overall mean first?
I am the author of the data visualization.
1/ Actually, it is the decreasing ARR over time in placebo patients which is actually pressing the sponsors (pharma companies) in designing larger trials. As ARR is getting smaller, it takes more patients to conclude at a significant difference between placebo and active groups.
Also the model is taking trial size into consideration as a weight.
2/ I checked the model again and it is correct. Perhaps didn't you take the log10 scale into consideration.
Thanks for the reply. I think I found the nuance I was missing in the model specification. It seems your model is actually
log10(ARR) = 0.104055 + [ -0.02449 x (Year-1993) ].
I was missing that year offset when I tried to replicate your fit line.
You are right, I did forget to mention that.
Thank you for your interest in our data visualization !
Interesting idea about vitamin D. But is there any evidence that vitamin D levels have increased over time in people with MS? Did the studies included in the visualization measure vitamin D levels in participants?
Aside from having MS what did all placebo patient had in common...probably the vitamin D and the placebo treatment !
Maybe part of the answer is within the composition of the placebo that was used within that pahse, unless things have changed the FDA does not regulate what goes into placebos.
Could the placebo have a modulating effect on drug bioavailability on MS treatment Ex: Copaxone or others ?
Thank you for your amazing work.
Have you taken age into account?
As far as I know younger MS patients have more ARRs than older MS patients.
Thank you for your feedbacks!
We did not take the age into account in this data visualization but it could be interesting to include this parameter in a future graphic
It seems likely that the increasing availability of MRIs and MS awareness would lead to patients being diagnosed at younger ages and earlier in the disease course. CIS wasn't even defined when many of these studies were started.
So as mentioned previously, adding average age of the participants to the visualization would help illustrate if that's a factor. EDSS scores, or years since diagnosis might also be informative.
If I am not mistaken, in 2006 was the discovery of anti-AQP4 for NMO. In the graphic it seems that there are two areas with different behavior, before and after.
It could be just happening that some NMO patients were included in the control groups before 2006.
One of the problems of 'relapses' in MS is how the 'experts' define acute relapses or breakthrough disease. We have built a huge armamentarium of NEDA-related material on the one hand while on the other hand, we have not addressed what a relapse is. For instance, patients often (some of them, anyway) want steroids when they feel like it and so the common ploy is to get to the ED and tell the doctor that their MS is 'acting up'. Most ED docs have not heard about a pseudo-relapse but even if that is excluded, how does one decide on what a relapse is ? Symptoms last for > 24 hrs, or so the patient says (remember the patients read the same material or at least they have access to, most scientific and non-scientific literature) and the doctor (even if that doc is a neurologist) does not know what to do. So, why not evaluate breach of BBB by doing an MR of brain or cord, if feasible, and decide objectively ? Why isn't that the standard of care ? What is this 'clinical decision making tree' and why is 'clinical evaluation' required (and most neurologists now do not even carry a patellar hammer let alone an ophthalmoscope or a pen light) that is riddled with errors ? Sure, perhaps even the MRI is not 100% in detection rate of BBB disruption but then, what is ?
Jagannadha Avasarala, MD, PhD
Greenville Health System