ProfG has put an interesting question about whether DMT are all born equal. Whilst the influence on relapse rates is much clearer in terms of the hierarchy, the effects on brain atrophy has led ProfG to ask a sensible question.
It was suggested that HSCT and alemtuzumab are up there at the top of the pile in the atrophy states, leaving the B cell therapies in their wake.
B cell therapies are suggested to be ocrelizumab and cladribine, which appear to be rapidly becoming best sellers as they lack the side effect risks of the two mentioned above. Are you making the right move?
I don’t know!
I have argued that HSCT, alemtuzumab, natalizumab, cladribine and ocrelizumab will have high efficacy in terms of blocking memory B cells from getting into the CNS and so should slow relapsing MS similarly. We know that relapses cause nerve damage and so blocking them must surely block nerve loss and slow accumulation of disability.
So they should all be more of less the same. However, are they?
If they are not the same, it tells us something. But what it is it?
First is there is a real difference?
Brain Atrophy is perhaps a rather useless outcome, because it is not really detecting, what it is claimed to be detecting, which is a loss of brain nervous tissue.
As such artefacts are made by brain shrinkage (from effective anti-inflmmatory activity) and brain swelling (from inflammation) and space filling of glial cells to replace lost tissue. As a result we have had some trials throwing away potentially neuroprotective agents treatments. It is only seeing the tip of the iceberg when it comes to detecting nerve loss and depending on how they measure it they can get better or worse results as it is not truely standardised. Yet we (i.e. neuros, persist on using it, because it is easy, when things like grey matter atrophy may seem of more value.
If there is a difference is the benefit of HSCT because HSCT and alemtuzumab (and natalizumab) hit the innate immune system like macrophages, which cladribine and ocrelizumab are not going to do as much. This would fit with the idea that macrophages are important in progressive MS.
However, could it be the T cells?. HSCT and Alemtuzumab deplete CD4 T cells for years
We have rehearsed the argument about CD4 T cells are their involvement in MS, but besides a selectivity or dimethyl fumarate for CD8 this T cell subset has not be adequately tested.
Remember cladribine depletes CD4 T cells by about 40-50% in year 0-1 and 50-60% in year 1-2 so if it is CD4 T cells then the depletion needs to be better.
However, what about CD8 T cells? Cladribine is less effective at depleting CD8 T cells. CD8 T cells are the dominant T cells by about 8 to 1 in MS lesions and they are cell killers. (In most EAE models CD4 T cels dominate). Are they at the heart of the nerve loss and the brain atrophy?
If this is important, should we deplete B and CD8 T cells? However, will this give us secondary autoimmunities?
Surely no more than currently with alemtuzumab and you would leave the CD4 T regs intact.
However, we ask do the drug s stop progression?. The simple answer is no!
Disability worsens with all treatments, although they may influence and have been shown to slow the rate of development of assignment to secondary progression. We believe the earlier you start treatment the better it will be. Once you damage a nerve then the dying back of the nerve occurs over months and potentially years. Therefore, the longer disease has developed the more nerves targeted for destruction will occur.
So if we look at the trials. People are getting alemtuzumab on average 2-4 years earlier than ocrelizumab and 5-7 years earlier than cladribine although use of ocrelizumab is in the ball park of some HSCT studies.
Now it is argued by Prof G that it has been shown by de Stefano et al. in 2010 (Neurology) that atrophy occurs at the same rate across MS. However, this is with a previso. In the study indicating the similar rate of atrophy they normalised the data to the size of the brain (Which shrinks) at baseline, but this was not done in many of the MS studies. If you do not do this, then there were big differences.
First sign volume loss was 0.4%, relapsing MS was 0.49% (mean duration 7 years) and SPMS 0.64% (duration 15 years) and PPPMS 0.56% (duration 8 years) . In the cladribine trial the placebo group was much higher than 0.49%.
However we could ask. Why is alemtuzumab is seemingly so good?
People getting alemtuzumab had a short disease duration and should have the best neurological reserve. However, did it shrink the brain during year 1 of treatment causing pseudoatrophy, but once you go treatment-free will the brain re-swell making it look bigger than it i,s especially as 50% of people treated get disease activity re-emerging. The manufacturers have repeated been asked to show this data, so see if pseudoatrophy has been reversed. Yet no response. Are they hiding something?
Now to sorry old cladribine languishing at the back of the queue. I bet they (manufacturers) wish they had something to hide, but in reality the data is probably not great because the cladribine programme was terminated before the studies were properly completed in the extension studies. Will they go back and do a study in early MS? Did they measure atrophy in the ORACLE study, I can’t see it in the protocol at clinical trials.gov..
You also have to rememebr that the killing of immune cells is slow with cladribine and takes 2-3 month for maxium effect not like a few days for alemtuzumab, so pseudoatrophy is going to be maximal later than with alemtuzumab. But the data from 12-24 months does not put it in the alemtuzumab league.
However if we look at the cladribine trial people were much slower at progressing than alemtuzumab (EDSS 2-3 in 9 years verses 2, the HSCT will be high probably because of high disease activity) and disease was much less active (Gadolinium positive = 32%), so they should progress slower. However, how one wonders many of the people with relapsing MS in the cladribine study were really relapsing SPMS? If we include them in the mix the atrophy rates are seemingly not that different to myeoablative HSCT which maybe did wonders in relapsing MS rate -0.14% (Lee et al) atrophy after the initial 3.3% loss presumably due to to neurotoxicity of the drug and pseudoatrophy)
Will the cladribine atrophy extension data arrive?
I suspect there is a disincentive to publish it now as would they want to show their drug in a bad light. However, it will probably be a bit of a dog’s breakfast because of the gap created by the termination of the programme.
ProfG has wondered whether cladribine gets in the brain to target B cells and asked whether this will that make a difference. A theorectical possibility. However, alemtuzumab is not really going to get in the brain and it doesn’t get rid of the oligoclonal bands and so that will not probably not explain the difference in atrophy.
Now, I know this doesn’t help you with decisions but the point is the data is not easily comparable and so having a look at a few papers and there is some confusion. The results from the natalizumab and the HSCT trials (above) are hopelessly underpowered (too small) and need hundreds of people if the data is to be meaningful. I have not included anything from year 1 as the brain volume shrinks with effective immunotherapy (pseudoatrophy).
The solution is to do some head to head studies so you compare apples with apples and pears with pears. ProfG has said he wants to do a head to head of alemtuzumab verses HSCT, but ocrelizumab (and cladribine) may need to go into the mix
Why because although the atrophy data intially in the phase III opera I & II were not that great with an annual lost of over 0.4%. In the extension trial by year 2 the rate of brain lost was within the healthy control range. However they baselined at six months rather than zero so it makes it harder to compare to other studies. For this reason I have ignored the first year when there is shrinkage. But how much shrinkage may influence the later time points.
This is supported by studies with rituximab. The Brain parenchymal fraction (brain volume) was reported as -0.19% assessed in 160 pwMS (Salzer et al. Neurol 2015), so what is the difference?
Will neuros have the courage to do the head to heads and sort the Wheat from the Chaff so you really understand risk benefit? Until then you will have to rely on meta-analysis and real-life data basing which is less than ideal