Hi i’m rick baron and thanks for joining us again for another segment of our ef5 today dr. gary grant zath and i want to talk about the randomized controlled trials of methotrexate and mycophenolate for myasthenia gravis interest in mycophenolate began in the late 1990s and there were several large series published of anecdotal reports which all seemed to indicate
That michael finley was helping my senior gravis patients then there was a small randomized control trial by dr. mary jo lee with 14 patients that also gave a hint that mycophenolate might be effective this was followed by two rather large placebo-controlled randomized trials one that was investigator initiated and led by dr. sanders and it was federally funded
And the other one that was led by a pharmaceutical company as spiriva the sanders study which i was involved in designing was just a three-month trial where as the asst free betrayal was nine months and to everyone’s surprise at the end of the day both studies were negative for showing a beneficial effect of mycophenolate and myasthenia gravis here is a graph of on
The top of the sander study showing the qmg score it was this essentially the same in both groups both groups improved over time but there was no difference between placebo and mycophenolate and and under that the similar data from the apsara spiriva study so why were these trials negative this is continues to be hotly debated in the neurology neuromuscular world
And i think there are several possibilities one is that the drug just doesn’t work for my senior gravis the other is that all the patients and the the sander study were on put on the same prednisone does 20 milligrams a day and prednisone may just be so effective in myasthenia gravis that it masked the my ventilate effect may be the studies weren’t long enough
And you can say that about the sanders three months study but i don’t but the spree even nine months study you would think was long enough but some argue it wasn’t where our endpoints not good enough does a qm g is it not as reliable as we thought and some of the other end points we use or was there something wrong with the population that was enrolled because
Of the negative mycophenolate study we then launched a methotrexate in myasthenia gravis study dr. past nor and i in the muscle study group were able to get another federal grant to do this investigator initiated multi-site as trial and we enrolled 25 patients in a methotrexate group 25 patients in the placebo group patients were all on prednisone and instead of
Looking at qmg zorp ryan endpoint as a primary endpoint we decided to look at prednisone dose using prednisone as a primary endpoint to try to mimic what was done in the old days the thigh print study and thus the results of this study were recently published in neurology and they were presented a couple years ago and as the microphone as was similar in the michael
Family study this study was also negative this just came out neurology last month and it showed that if you look at the pregnant had a forced prednisone taper but using a standardized protocol and at the end of the day there was no difference between the prednisone dose in the patients that were taking methotrexate versus placebo 13 verses 15 milligrams which is
Not statistically significant and we looked at secondary endpoints such as q mg mg atl and other scores and also there was no statistically significant difference between both groups so once again you could ask the question why was the methotrexate mg trial negative what and again i come up with similar type of answers prednisone may work too well and it’s just
Very hard to do an mg study with patients on prednisone again maybe like michael fennelly to conclusions that the drug just doesn’t work in myasthenia gravis maybe we used a too low dose of methotrexate but we did use 20 milligrams a day or a week orally we did have a lot of dropouts in the placebo group could that have been a problem where we underpowered and
And did we do some did we did we just not handle the dropouts well so we don’t i don’t have the answer but i think i’m going to have my colleague dr. grant seth come up and try to give his insight into why studies like this are so difficult to do intend to prove to be positive doctor grants a–the thanks dr. barron for having me back so some of these issues rick
Is already touched on but they’re basically three issues that you should ask and you shouldn’t just ask these issues because the study disagrees with what you expected you should ask these studies regardless we commonly do that unfortunately so the first question is was there bad study design and execution in other words poor internal validity and the answer for
The randomized control trials is no now dr. mary jo rosie rowley study is rated class to buy a and criteria still a pretty good study although severely underpowered but the other studies were all class one so a very low risk of bias relative to design and execution of the study the second question in a common cause for a study to be quote- is in precision lack of
Power and it’s straightforward to tell if a study had enough power you just look at the width of the confidence intervals and so you know these are supposed to be steroid sparing agents so i looked at the difference in the daily prednisone youths used that don sanders article measured that as an outcome and mamatha mamatha past north that was their primary outcome
And if you look at the difference in daily dose of prednisone the point estimates is about two milligrams less if you were on mycophenolate or on methotrexate but the confidence intervals are consistent with in a mamatha study with it being almost seven milligrams less is that important probably not important probably not but that would tell you whether if this
Includes something important or more importantly excludes an important effect that the study was adequately precise what about clinical incomes looking at the q mg of that was an outcome for all the mycophenolate studies and if you looked at the difference in q mg from baseline for patients on the steroid sparing agent versus placebo the those confidence intervals
Are very narrow and were to exclude a change in qmg favoring the steroids bearing agents you know of less than minus two so that study was sufficiently precise or all those studies together were sufficiently precise to exclude a benefit what about methotrexate not so much it’s still possible that it was only because of random error in precision that we didn’t
See a benefit because the study is it inconsistent with a with a difference of five qm g’s which would be a moderately important effect so i would argue that we can’t really be that confident about that outcome but certainly you can be more confident about the steroid outcome and then the last one is poor external validity which is a synonym for generalizability
Or sometimes called directness or indirectness this is the one where people sort of take advantage and just how i can reject any study if it does if it’s not consistent with what i want because my patients are different so the studies show that for these patients using the drug the way it was used compared to steroids alone in the way that they were used that the
Outcomes were not different for that time they they show that now you can argue while the that you know my patients are different or there’s a subgroup of patients that are different but you don’t know that maybe the weight steroids are used is different but certainly the time horizon may be an issue but the bottom line is you know the studies are never definitive
But if if you believed that mycophenolate or methotrexate worked once these studies were published your confidence in that belief should be lowered now does it mean that it doesn’t work but your confidence should be lowered thank you so that’s that’s the conclusion you’re all should come out from that the confidence is lower this little possibility might work boom
We couldn’t prove them right using traditional study design right it’s less likely that the drugs are effective now if your confidence hasn’t changed at all or looking at all the articles that were published after because some of our colleagues at your confidence is that she’s all right it happen exactly i was going to make that point based on all the articles
And we call that cynicism oh okay then then there was no point in doing the studies in the first place re fi
Transcribed from video
RRNMF – Randomized Controlled Trials of Methotrexate & Mycophenolate in MG By University of Kansas Medical Center
