Friday, 31 July 2015

A hammer to crack a nut, is it nuts to think that immunosupressives do not cause PML

Van Schependom J, Gielen J, Laton J, Nagels G. Assessing PML risk under immunotherapy: if all you have is a hammer, everything looks like a nail. Mult Scler. 2015 pii: 1352458515596458. [Epub ahead of print]
Recently, three progressive multifocal leukoencephalopathy (PML) cases have been reported in multiple sclerosis (MS) patients, two treated with fingolimod (Gilenya, Novartis), the third with dimethyl fumarate (Tecfidera, Biogen). Because our immunotherapeutic arsenal in MS and other diseases is increasing, and because PML is a very serious health risk, it is of interest to the clinical community to show how we can assess this risk in a statistically sound way. The null-hypothesis for this analysis was that there is no elevated risk for PML in patients treated with one of these recent drugs, compared to the incidence in the general population. We conclude that the null hypothesis cannot be refuted.

So this stuff says they think that some oral DMT that are generalised immunosuppressives do not cause PML. Wake up an smell the roses people. Long term leucopeania caused by such drugs is going to leave you at risk of infection. It may not be a high risk but it is a risk.

62 comments:

  1. Then again these drugs have been given the status of "highly effective", and it seems this is the road for new drug development.

    As someone who has to self cath and is at higher risk of UTI's I really don't want to suppress my immune system with these drugs. Hopefully treatments will be developed to address why the immune system is attacking the brain instead of marketing drugs to suppress the immune system.

    ReplyDelete
  2. These approaches are around but have unproven efficacy but it would be years before they are available. Induction therapies have the real risk of infection but once the drug is out of the system immune responses against infection can revenerate. You can also stop cells getting in brain but allow then into other tissues or selective removal. Do nothing is your choice as you bear the consequences of that choice.

    ReplyDelete
    Replies
    1. I never said I was doing nothing, I am actively treating my disease with an approved ms therapy + an add on therapy investigated by some of the most respected people in MS research. All this without suppressing my immune system.

      Yes this is my choice and I take my chances, but it is clear to me MS research is going down the wrong path.

      Delete
    2. "..........but it is clear to me MS research is going down the wrong path."
      Well, I suppose that's one point of view. All welcome!

      Delete
    3. Yes, I guess the name of the game now is develop immunosuppressive drugs looking at the safety and effectiveness of a 2 year trial and proclaim these newere thrapies to be "highly effective"

      Once in the field for a couple of years and problems start arising like PML, you have a fallback to say that nothing showed up in the trials. "We are innocent. We couldn't have known." All view points are welcomed. But it seems the goal such as people like Team G is to push the latest marketable pharmaceutical product.

      But as the point of this post points out, you should have known

      Delete
    4. "But it seems the goal such as people like Team G is to push the latest marketable pharmaceutical product."
      Well, I think the White Knights would disagree with this.
      You're right, us immunologists were afraid that these newer immunosuppressives would have their problems. We were right and we need better/more specific ones and perhpas the anti B cell therapies coming on stream may have lower numbers of adverse events, if we don't try we'll never know.
      Personally I'd love to see Cladribine given approval, no big bucks for pharma there!

      Delete
    5. 'approved ms therapy + an add on therapy investigated by some of the most respected people in MS research. All this without suppressing my immune system'

      OK, I give up, I can't think what this approved MS therapy is, that doesn't to some degree suppress the immune system, well I can think of one - LDN -but I didn't know it was ' approved' by an official body (at least not in the UK). I'm curious, not critical, so any information that you can give me, is very welcome.

      Delete
    6. Of course neuroprotectants for MS are now being studied clinically (simvastatin, phenytoin and perhaps THC) after a lot of hard work, particularly by us if I may be immodest and which will be of benefit to those with progressive disease who have been ignored for far too long. We were well ahead on this over 10 years ago and I think have helped push things forward.
      I think we can hold our heads high.

      Delete
    7. Well this is the drug I take. It's not very effective on its own but in combination with an add on therapy it seems to have dramatic effect on relapse rate as has been demonstrated.

      This seems like it could be a promising treatment that doesn't act through immunosuppression, but is DOA.

      Delete
    8. All MS approved drugs act by some form of immunosuppression, you may call it immunomodulation if you like. They are weak moderate or strong and the stronger they are the more risk there is. The difficult one to understand what it is really doing is glaterimer acetate. Having seen it can do all sorts, its action must be relatively non-specific despite the original intention.

      Delete
    9. But it does seem to work, yes? I'm wondering about switching to this as an interim between Tysabri and Lemtrada. Unorthodox certainly, but strong immune modifiers seem to bring problems for me with infusion reactions and week 4 is like having a relapse all over again, every cycle.

      Delete
    10. Oops I meant does glaterimer work, well, yes I know it does but I'm thinking of it working as a bridging treatment as I just can't take any more heavy immune suppressors.

      Delete
    11. If you are talking about Copaxone alone, it is not that effective. The Copaxone + add on therapy is what I am on. This was investigated by Harvard researchers and it showed a dramatic reduction in relapse rate compared to Copaxone only.

      http://activemsers.wssnoc.net/showthread.php?t=1811

      There is no evidence what it has on progression and this was a small patient trial, but this is what I have decided to try with the go ahead from my docs.

      Delete
    12. Thank you for the information and I hope it helps you. I just remembered that copaxone takes a while to reach full efficacy, so probably wouldn't work as a bridging treatment for me :(.

      Delete
    13. But the argument put forward for the study by the Harvard group s Copaxone and Albuterol was that it was targeting the IL-12 pathway......this is immunosuppression.

      However, the IL-12/IL-23 story does not seem to hold up as anti-IL-12p40/IL-23p40 does not seem to inhibit MS ...or for that matter established relapsing EAE.

      Delete
    14. "dramatic effect on relapse rate"

      I suspect that the drug in question is albuterol. (Please correct me if I'm wrong, but I've seen similar descriptions of albuterol on other forums.) The study supporting its possible efficacy had only 40 people. Have you ever looked at the results closely? The "dramatic effect on relapse rate" was a difference between 2 people relapsing on treatment and 8 relapsing on placebo. That's not even significant, though the study fails to mention that. Even if it were significant, it would not support claims of a large effect, you would need a lot more data to do that. A quick calculation suggests that you would need a sample size of 200 to confidently claim that albuterol decreased the number of people relapsing by half. (And that's under very optimistic assumptions.)

      Delete
    15. Anon 3.29. If you have infusion reactions and the drug is not working for long enough it may be that you have started to make neutralising antibodies. This will mop up the drug to stop it working and can give infusion reactions and be warned anaphylaxis could be a possibility.

      I would suggest having this checked and also perhaps have your white cells checked at 3 weeks post infusion to see if the tysabri is still coating the white blood cells, it may be that it is being neutralised so it stops working properly, suggesting you need more frequent infusions or that you need to change to something else.

      P.S. I am not a Doctor so talk to your neurologist.

      Delete
    16. Annon 4:13:00,

      From the study:

      "Ten subjects experienced relapses during the study,
      2 in the glatiramer acetate plus albuterol group and 8 in
      the glatiramer acetate plus placebo group. Subjects in the
      glatiramer acetate plus placebo group experienced a first
      relapse faster than subjects in the glatiramer acetate plus
      albuterol group (P=.03)".

      P = 0.03. This is significantly significant even with the low sample rate. This is whats amazing. Maybe you need to take a course in statistics.

      Delete
    17. Why this didn't make more of a splash, i'll never know.
      http://multiple-sclerosis-research.blogspot.com/2013/07/restoring-self-tolerance-is-way-to-go.html

      Delete
    18. MD,

      I suspect the Harvard researchers thought it acted on IL-12.

      But salbuterol has been shown to reduce EAE and even prevent EAE after initiated in an activated immune system. I suspect GA useful for activation:

      ---"Interestingly, our data show that salbutamol only potentiated tolerance induction during an ongoing immune response and not when salbutamol was coadministered with the Ag before immunization (Fig. 4⇑). Moreover, coadministration of salbutamol before immunization did not further modulate cytokine production by splenocytes as determined on day 9 after immunization of the animals (Fig. 7⇑). Therefore, we conclude that β2-adrenergic stimulation is only effective in potentiating tolerance induction in the context of an activated immune system. One possible explanation for this phenomenon would be that immunization increases the sensitivity of the β2-AR on immune cells."-----

      ---"Interestingly, our data show that salbutamol only potentiated tolerance induction during an ongoing immune response and not when salbutamol was coadministered with the Ag before immunization (Fig. 4⇑). Moreover, coadministration of salbutamol before immunization did not further modulate cytokine production by splenocytes as determined on day 9 after immunization of the animals (Fig. 7⇑). Therefore, we conclude that β2-adrenergic stimulation is only effective in potentiating tolerance induction in the context of an activated immune system. One possible explanation for this phenomenon would be that immunization increases the sensitivity of the β2-AR on immune cells."-----

      http://m.jimmunol.org/content/169/9/5028.full

      Delete
    19. "Maybe you need to take a course in statistics."
      Believe me, MD is no slouch when it comes to statistics.

      Delete
    20. I wasn't saying MD need to take a course in statistis, it was directed to Ann on a 14:30.

      But please comet if these results are not statistically significant in your view. I think usually you need to have a large patient population to detect a difference because the difference is tiny. But if the difference is huge, the smaller population still produces a statistically significant result.

      Delete
    21. My mistake! The results are intriguing, I'll leave it at that ;-)

      Delete
    22. "This is significantly significant even with the low sample rate. This is whats amazing."

      No, the relapse rate comparison is not significant. The study found that 2 relapsed in the treatment group and 8 relapsed in the placebo group. You can run a binomial test to determine whether this difference is significant, and it's not, the two-tailed p-value is >0.1. You can compute this by hand if you want! The significant result was found via a post-hoc analysis on the "time to relapse," which is a different parameter. It's true, the study claims that this parameter was significant, but if you test enough parameters, you'll come up with one that's significant.

      If you're skeptical of my claims, run the power analysis for yourself. Maybe there is a real effect of albuterol -- I'm not saying that there isn't -- but there is no way to reliably estimate the effect size from a study this small. It's very difficult to estimate even from a study 5 times as large.

      Delete
    23. There has been no follow on study....this makes you wonder

      Delete
    24. Who will pay for it? How about Team G?

      Delete
    25. "The significant result was found via a post-hoc analysis on the "time to relapse," which is a different parameter. It's true, the study claims that this parameter was significant, but if you test enough parameters, you'll come up with one that's significant."

      Its quite amazing the difference on how a patient views this compared to an academic. To the academic unless the statistically significant parameters are identified before the trial is started, any results found in the data after the trial is complete must be neglected.

      A patient would want the researchers to fully investigate the trial data to determine if there were any important significant outcomes found in the trial. To me and I suspect most people, time to first relapse and relapse rate are the same thing.

      A post hoc analysis does not negate the fact that it is a statistically significant effect found during a placebo controlled double blind trial, which is the gold standard.

      To an academic, the trial would need to be re-run with the time to first relapse indicated up front for it to have any validity. What a waist. It's no wonder MS research is at the state it is in now.

      Delete
    26. The thing about the use of a beta2 adenergic agonist such as albuterol for a ms treatment is it has a wide background behind it.

      Some prominent researchers from The University of Chicago did extensive research for this drug and obtained a patent. The patent describes the basis for this drug:

      http://www.google.com/patents/US5264459

      This is in stark contrast to Team G's approach at testing every conceivable drug at radom to see what sticks instead understanding how a drug works.

      Since the drug is generic and the patent when issued years ago was owned by the U S government this is not something that would be of interest to pharma or the pharma sponsored research industry which is a shame.

      Delete
    27. Anon 10 This could be p hacking and indeed if you do enough tests you will get a signifacant result as the result was about a 1 in 35 chance that the result occurred by chance.This is why in trials you have to define your result before you start otherwise you cherry pick.

      The question I would pose would be if the trialists thoughtvthe data was strong why has their been no follow up study.....this may be a money issue but i am sure the group could raise the funds.

      Or it could be a lame duck...the oral tolerance trial with myelin from harvard had lots of phase ii fanfare, but was i anticipated, a dodo in phase iii.

      Delete
    28. Anon 10: I was making two points. The first was the point about lack of significance of the primary endpoint. You should look up MD's reference to "p-hacking." Academics want to avoid p-hacking and post-hoc analyses not because of ritualistic adherence to the gods of statistics, but because they lead to unreliable results. The second point was about the lack of statistical power in the study. This point is independent of the first. Let's suppose that albuterol really does have an effect on the relapse rate. The study was too small to determine what the size of its effect is. You're correct that the results of the study would be predicted if albuterol cut the relapse rate by a factor of 4. But the results would also have reasonably high probability if albuterol only cut the relapse rate by a factor of 2, or a factor of 1.1, etc. In order to reliably determine how strong the effect of albuterol was, you would need *a lot* more subjects.

      I am both an academic and a patient with MS. My well-being depends on finding effective treatments as much as yours. That's why I think sound statistical methods are so important for MS research.

      Delete
    29. Looking at the time to first relapse there we only 2 on the Copaxone + albuterol arm while there were 8 in the Copaxone only trial. This would only occur by chance 1 out of 33 times if the trial was repeated over and over again.

      This is a statistically significant event given the low patient population.

      In other words, you could have a trial with 10,000 people you could detect a slight effect that you could never find in a sample size of 10. But in this case if the effect is large, you can detect this effect in a smaller patient population. The power was there but it would not have been obvious during trial design.

      Given this fact and this result, a phase 3 trial is warranted, but the question is who will pay for it?

      If you are saying Harvard could find the funds, let's look at the estimated costs. For 200 people for a 2 year trial, the cost of Copaxone alone would cost over $21 million.

      The fact that other drugs are on the market and this is a add on therapy, it seems unlikely Harvard would investigate this on their own and I doubt they could have a justification.

      Delete
    30. Yes so my misttake . We could argue that the original trial was completely unethical because if there was not plan on how to develop a positive resilt then the trial should not have been done as this a false hope generation and awaste of time testing drugs going nowhere. The question of what next and whst is the pathway to perscription must be answered before doing such trials.???
      Time and time again I see trials going nowhere. Is the view we can do a study and the world will change after a positive result is a dream world that many people have..
      Im not seeing much evidence this is the case such as the simvaststin trial.
      For and RRMS DMT it is another level of diffficulty

      Delete
    31. Yes, I guess if your goal is to provide a product for pharma to produce this is a trial that never should have been investigated. If you are interested in finding a cause for a disease and testing drugs that may have an effect and adding to knowledge and understanding, this was a trial that would be of value.

      Hopefully this avenue will be pursed by people worworking on Parkinsons and other neurological conditions:

      http://www.hindawi.com/journals/jir/2014/103780/

      As for Team G, keep up the good work on immunosuppresants.

      Delete
    32. Anon 7:45, I know that you're well-intentioned, but you're mistaken about this. There is a combinatorial calculation that allows you to compute the p-value for the number of relapses. It's simple enough that you can write it in WolframAlpha, here is a link to it: http://www.wolframalpha.com/input/?i=%28%28sum+%2810+choose+i%29%2C+i%3D0+to+10%29-%28sum+%2810+choose+i%29%2Ci%3D3+to+7%29%29*0.5%5E10

      You can see that the p-value is 0.109. This means that there was a probability of 0.109 of seeing a difference between the groups as extreme as the one observed, given the hypothesis that the true relapse rate is the same in both groups. This is an instance of a binomial test, which is being used in this case to test whether the rate of a binary event differs between two groups.

      "But in this case if the effect is large, you can detect this effect in a smaller patient population. The power was there but it would not have been obvious during trial design."

      I'm not sure that you understand what a power analysis is. A power analysis is used to determine how many samples are needed to reliably distinguish two different hypotheses. In this case, the calculation above shows that there was not enough power to distinguish the hypothesis that albuterol cuts the relapse rate by a factor of 4, from the hypothesis that it has no effect whatsoever. In other words, even though the relapse rate in the albuterol group was 1/4 of the relapse rate in the control group, there was not enough evidence to conclude that this effect was not due to chance. As I previously said, even if you had 200 subjects, you would not be able to reliably distinguish the hypothesis that albuterol cuts the relapse rate by a factor of 4, from the hypothesis that it cuts the relapse rate by a factor of 2. This would be true *even if* the observed relapse rate in the albuterol group was 1/4 the relapse rate in the control group. Here is the relevant calculation for this last claim: http://www.wolframalpha.com/input/?i=sum+boole%28i%3C%3D11+or+i%3E%3D39%29%2850+choose+i%29*%281%2F3%29%5Ei*%282%2F3%29%5E%2850-i%29%2C+i%3D0+to+50

      If you think that I'm mistaken about this, it would be helpful to provide some explicit calculations, since otherwise the claims you're making are too vague.

      Delete
    33. Anon 9:41
      ProfG works on immunosuppresants because this gives people in his care the option of access to a drug that may be of some use years before they are available to the general population.

      As for our lab work we have had little focus on immunosuppressives for years.

      As for "the wouldn't it be fun experiments going no where" what are your thoughts? and is this properly explained when you are being recruited?.

      Is this worth the risk,,because some of these fun experiments can do harm

      Delete
    34. I think the trial is still of value even though it will not be pursued further as a generic add on to Copaxone.

      It think the time to first relapse was a significant effect. You may say it is p hacking but it is an important outcome for someone with MS on Copaxone who wants to improve the efficacy without having to switch to an immnosupressant.

      Given this result along with the mounting data on beta2 adenergic receptor agonists it seems this class of drugs has a potential.

      To me this could be similar to the Rituximab phase 2. It will never reach phase 3 but there was evidence to pursue this line further. In this case Pharma can tweet the molecule and patent it.

      But the lack of a phase 3 trial is not preventing doctors and patients in the US from using Rituxan for MS currently of label Bacause the phase 2 trial is compelling and Rituximab has been in use for many years and seems to be a safe route.

      Maybe Pharma could come up with a patentable beta 2 receptor agonist to pursue. Their seems to be justification for it.

      Delete
    35. Also, here is a paper describing time to first relapse versus relapse rate as an endpoint:

      http://www.ncbi.nlm.nih.gov/pubmed/22914849

      Looking at the data, the Copaxone + Albuterol group had 14 patients, 2 0f which relapsed (14.2%) while the Copaxone only group had 13 patients, 8 of which relapsed (61.5%). The Harvard group calculated this to be statistically significant (p=0.3, random chance 1 out of 33). The p value considered to be statistically significant in drug trials has been historically been p <0.05 (random chance of 1 in 20).

      Adding more patients would lower the p value, butI'll take the Harvard calculations as the result being statistically significant.

      Delete
    36. Anon 11:23, 1:43: Having MS is not a license for sloppy, wishful thinking. Maybe albuterol has an effect on the disease, maybe it doesn't. I think you realize that you don't have the knowledge necessary to assess this claim, and you're deferring to the authority of the paper's authors. They're from Harvard, as you've consistently reminded us! Well, I'm sorry to say that working at Harvard does not protect you from being wrong. You need to assess the paper on its independent merits.

      You shouldn't just take me at my word either. If you have MS, then learning statistics will benefit you in the long run, as there are going to be many more studies to evaluate. Learning enough statistics to understand the problems with this paper is not very hard -- it would take around a month learn the relevant math if you don't have any background in the area. I would be happy to provide references if you are interested.

      Delete
    37. Anon, 4:12:00,

      Can you tell us in what capacity you work with statistics as well as how you are an "academic".

      Please be specific and don't hide behind your Anonimity. It will be useful to know what institution your are from. My guess is that your connection to academia may relate to your position as a janitor at a university.

      Clearly all of your points are baseless.

      Delete
    38. I am not Anon 11:23, 1:43 - but I am interested in references which may help learn statistics relatively quickly (I have some background in single- and multivariable calculus - but have not seen or used it for decades)

      Delete
    39. Here is what our well respected academic Anon 4:46 originally concluded from the trial result:

      "...was a difference between 2 people relapsing on treatment and 8 relapsing on placebo. That's not even significant, though the study fails to mention that."

      It's amazing to me that you couldn't even understand that a P value of 0.03 is statistically significant.

      One of the toilets in the academic office is overflowing.

      Delete
    40. Sorry but my eyes are glazing over, always happens when people resort to insults. I also use stats -the fluffy marketing type that you can manipulate to mean just about anything, if you so wish! Maybe same goes for medical ones as well?. I'll just go hide now;)

      Delete
    41. Anon 4:44: Even if I were a janitor, my arguments would remain valid (or invalid) independent of that fact. Statements in science don't suddenly become true or false because of the social status of the people making them. You need to be able to work through the math for yourself to know whether to trust the paper.

      Anon 4:46: For free resources, I like the lecture notes available here: http://ocw.mit.edu/courses/mathematics/18-05-introduction-to-probability-and-statistics-spring-2014/readings/
      My memory is that they require some calculus but not much else.

      A standard textbook for intro statistics courses is: http://www.amazon.com/Mathematical-Statistics-Analysis-Available-Enhanced/dp/0534399428/ref=sr_1_1?s=books&ie=UTF8&qid=1438620928&sr=1-1&keywords=rice+statistics I think it's pretty good for that purpose, and it's at a pretty similar level to the lecture notes. You may or may not be able to find a free copy online. As for learning the topic relatively quickly, it depends on what that means. I think a month of hard work (taking the time to solve exercises and work out examples by hand) would be enough.

      Delete
    42. Anon 4:58: Can you please write out the computation that results in that p-value? It would help to clarify the discussion. I showed my computations explicitly above (in the WolframAlpha links), please show yours.

      Delete
    43. Actually "stats" as you say is a pretty broad term for this evaluation. What they are doing is survival analysis which I'm sure you are an expert in.

      At any rate, the group from Harvard (yes I like to point this out) surely has the capacity to perform this analysis with accuracy, since some of the most respected researchers in MS were part of this trial.

      What method are you using in your detailed analysis? Please share your results.

      Delete
    44. As discussed, I did not perform my own analysis I'm relying on the calculation the Harvard group performed. You seem to have difficulty in compression. This is odd for such a prestigious academic statistition.

      At any rate, they used survival analysis to determine the p value. It sounds as though you've never heard of this before. I would contact Dr. David A. Hafler or Dr. Howard L. Weiner at Harvard. Maybe they can help you.

      Delete
    45. Interestingly we showed that in the high impact factor journals about 95% of EAE studies use what we have considered to be inappropriate statistical analysisand this includes authors from Harvard.

      Baker D, Lidster K, Sottomayor A, Amor S. Two years later: journals are not yet enforcing the ARRIVE guidelines on reporting standards for pre-clinical animal studies.
      PLoS Biol. 2014;12:e1001756

      Delete
    46. It would be great to have the Harvard group respond to the allegations being made by Team G. Since this was a double blind study published in JAMA which is a pretty prestigious journal, it would be good to get their point of view. Unless there is something Team G is afraid of. If you think their work is garbage, come out and publicly say so. Otherwise you might want to refrain from commenting.

      Delete
    47. Anon 11:42: Suppose that you flip a coin 10 times, and it lands heads twice. How confident would you be that the coin is unfair? Suppose that you flip it again 100 times. How much would you bet that it will come up heads 20 times or less?

      Delete
    48. Anon 11:42

      Exqueeze me I am not sure what you are saying what allegations about this (JAMA) paper that we have made.... I do not believe we have made any about the validity of the work. Please get your facts straight before making accusations.

      'I am happy to stand by any comments on EAE as I do not believe that you should be making neurological scores become parametric data for analysis. The list of publications are there for anyone to read and you can make your own mind up. This comment addresses a comment. However, this has nothing to do with the copaxone and albuterol paper. As of mechanisms of action these are hypothetical.

      I was indeed thinking of writing to the people at Harvard to see where the albuterol and copaxone story had gone since the publication, which is the important aspect.

      However, it is clear that some readers have convinced themselves this is the way forward (It has appeared in the comments repeatedly recently). We all know the drug development process and so we all know there is insufficient data to make any recommendations (one way or the other), because it needs repeating in a phase III trial.........until that happens I and I suspect ProfG are not going to recommend this or other treatments.

      Delete
    49. P.S. If you want examples of fallibility or reproduction of ideas, they can be found for virtually any lab no matter how big or small.

      Delete
    50. "I would contact Dr. David A. Hafler or Dr. Howard L. Weiner at Harvard".....

      This is the work of Samia Khoury and I am sure she would be able to answer questions without the help of her male colleagues.Furthermore Prof Hafler is now at Yale

      Delete
    51. And Prof. Khoury lists American University of Beirut as her affiliation (or one of her affiliations) in recent publications.

      Delete
    52. The clinical design is here NCT00039988. Outcomes details were apparently added December 2005.

      Clinical trials must be registered before people are first treated otherwise certain journals will not publish the trial results.

      Delete
    53. It'seems odd that your post where you proclaimed the following was deleted

      "The significant result was found via a post-hoc analysis on the "time to relapse," which is a different parameter. It's true, the study claims that this parameter was significant, but if you test enough parameters, you'll come up with one that's significant."

      You gave some of your calculations, so why did you remove it? I suspect someone on Team G realized that if a survival analysis was conducted the results are indeed significant. By the way there is no mention at all that this was a post-hoc analysis.

      Team G seems to have a high opinion of themselves and they can easily point faults at others without having a clue. But when it comes down to it Team G are several rungs down the latter than those who they criticize, especially in this case.

      Delete
    54. "Anon 11:42: Suppose that you flip a coin 10 times, and it lands heads twice. How confident would you be that the coin is unfair? Suppose that you flip it again 100 times. How much would you bet that it will come up heads 20 times or less?"

      You obviously have no clue of what survival analysis is all about.

      Delete
    55. Anon 11:10. You are mistaken again about what was and what wasn't done and what the reasoning behind anything is. Posts do get deleted sometimes on purpose and one wonders whether some of the above abasive comments should have been deleted. They are also easy to delete by mistake. I do not remember deleting any posts on purpose.

      Calculations I can see are from Anon 9.43. I have not got the time to check them. Maybe ProfG is sat on a beach depleting your posts but I doubt it and am sure he has not got time to check either.

      This is an opinion it may be it is not be the right one, just like yours in an opinion too.

      It is true that if you endless pick statistics you will find one that gives a result, this is why you have things like clinical trials. gov and you can see which outcomes were going to be tested. I have not issues about whether the result claimed is significant.

      The fact is the primary clinical endpoint according to clinical trial.gov was

      Change in each participant's disease status, as measured by the Multiple Sclerosis Functional Composite score (MSFC)

      At 2 years this was not different according to the authors and therefore it would be a percieved failure despite differences at early time points. It is what it is

      Time to first exacerbation was a defined secondary outcome this was reported to be different. It is what it is

      This post has gone far enough and I will stop as you are simply being irritating.

      Harvard is the number one University in the World we all know that.

      No one is above criticism, including you or I.

      READ THE PAPER IT IS OPEN ACCESS!!!
      Enough of the mudslinging as

      Delete
  3. http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4407783/

    ReplyDelete
    Replies
    1. This indicates that Low Dose Naltrexone does really do anything useful (Figure 1). There is a delay in the onset of disease by a couple of days but they end up in the same place. It seems like the LDN group may have been getting anaesthetised to get injected whereas placebo group wasn't

      Delete
    2. Similar to DMT? A delay but end in the same place: SPMS

      Delete
    3. bit different in eae experiment the delay is 3 days

      Delete