Wednesday, 15 February 2017

#ClinicSpeak & #NeuroSpeak: alternative facts and PML

Pharma vs. the EMA: whose PML figures are more reliable? #ClinicSpeak #NeuroSpeak #MSBlog

A person with MS from Spain emailed me last week asking for advice. I assume he is on natalizumab and is JCV seropositive and is at risk of developing PML (progressive multifocal leukoencephalopathy) as a complication of his treatment. His neurologist has offered him a choice of dimethyl fumarate (DMF, Tecfidera) or fingolimod (Gilenya). He is however very concerned about the risk of PML on both of these options; the PML figures his neurologist gave him in relation to the number of cases of PML on both of these DMTs are very different to figures he found when searching online. He emailed me to find out which figures are correct. 


Biogen informed me a few weeks ago that a 5th case of PML had been reported on Tecfidera and that last case had had lymphopaenia with only one total lymphocyte count below 500/mm3 (0.5x109/L). Similarly, in December Novartis informed the MS community of the 9th case of PML on Gilenya unrelated to previous natalizumab exposure. I was therefore surprised when this patient referred me to the European database of suspected adverse drug reaction reports.





It is clear that in this database, as at 31st January 2017, there are at least 19 cases of PML reported on Tecfidera and at least 56 cases of PML on Gilenya. Which figures do we believe? I have asked both Biogen and Novartis to clarify, which figures are correct, i.e. their official figures of 5 and 9 cases of PML, respectively, or the EudraVigilance figures of 19 or 56 cases of PML, respectively. 

Please note that European database of suspected adverse drug reaction reports is an official website. All the data displayed in the web reports is taken from EudraVigilance, a system designed for collecting reports of suspected side effects, used for evaluating the benefits and risks of medicines during their development and monitoring their safety following their authorisation in the European Economic Area (EEA).

The following is a list of important facts about the source of the data you can view on this website for each web report:
  1. Each individual case in EudraVigilance refers generally to a single patient; an individual case is composed of at least one report, called the initial report, which might be complemented by follow-up reports.
  2. A web report shows serious spontaneous cases held in EudraVigilance since the medicine or active substance was authorised for use in the EEA. A case is classified as 'serious' by the reporter when a side effect is one that (i) results in death, (ii) is life-threatening, (iii) requires hospitalisation or prolongation of existing hospitalisation, (iv) results in persistent or significant disability/incapacity (as per reporter's opinion), (v) is a congenital anomaly/birth defect, or (vi) results in some other medically important conditions.
  3. A web report shows serious spontaneous cases where an authorised medicine or active substance is suspected by the reporter to have caused or contributed (e.g. by interacting with one or more other medicines) to a serious side effect. Reports where an authorised medicine or active substance is reported as a concomitant medicine are excluded. 
  4. The figure displayed is always the running total of serious spontaneous cases reported up to the end of the previous month. The figures are updated online on the 15th of the current month.
  5. Pharmaceutical companies that hold the marketing authorisation of a medicine, as well as national medicines regulatory authorities, are legally required to submit reports of suspected side effects that occurred in the EEA to EudraVigilance. This includes reports received from healthcare professionals and patients.
  6. Pharmaceutical companies that hold the marketing authorisation for a medicine in the EEA are also legally required to submit to EudraVigilance all reports of suspected unexpected adverse reactions that are serious and that occurred in a third country (non-EEA) where they hold a marketing authorisation.
  7. The web report does not include reports from studies (e.g. clinical trial, non-interventional study) or other types of reports (i.e. only spontaneous reports).
I was as confused as this Spanish person with MS who simply wanted to know the risk of PML on DMF and/or Fingolimod. The clarifications below are very helpful. The EudraVigilance figures are a crude indicator of an adverse event on a particular drug, but the numbers cannot be used to calculate incidences. 

Addendum Added 16h20, GMT, 15-Feb-2017:

I have now had a formal response from Novartis. The EMA, or EudraVigilance, figures are essentially from two sources. One from EU members states and the other from Novartis. The Novartis figures are global figures. The numbers in the 
EudraVigilance database are both suspected, and confirmed, cases of PML and as expected some of them will turn-out not to have PML and some will be duplicate entries. Novartis provide the EMA with information about whether or not the cases are associated with, or without, the previous use of natalizumab. 

The following are Novartis’ official and latest figures:

The overall rate of PML under fingolimod therapy not attributed to previous natalizumab treatment is very rare and estimated to be less than 1:10,000 patients.

10 PML cases in ~184,000 fingolimod treated patients (~397,000 patient-years) as of November 2016.

(Estimated risk (95% CI) is 0.054 (0.026, 0.1)/1,000 patients and incidence rate (95% CI) is 2.52 (1.21, 4.63)/100,000 patient-years exposure.)

Two cases presented with confounding factors, 1 case had previous natalizumab exposure for 10 months (3 years 9 months before PML diagnosis).

One additional case, PML occurred during 3-month natalizumab exposure, after 4.5 years fingolimod treatment, history of recent exposure to steroids.

Nine patients were within the age range of 49 to 63 years, while one patient was 32 years old.

In nine cases, fingolimod exposure ranged between 30 and 54 months, while one received fingolimod for 18 months.

None of the patients had sustained grade 4 lymphopenia.


Addendum Added 16h36, GMT, 15-Feb-2017:

The Biogen figures are 5 confirmed PML cases in ~230,000 dimethylfumarate treated patients (~330,000 patient-years of exposure) as of October 2016. 


CoI: multiple

27 comments:

  1. Prof G why would you believe Pharma? They have been caught lying so many times in the past that it is clear that they are the real masters when it comes to fake news. They are more concerned about the bottom line than patient safety.

    ReplyDelete
  2. Wow, this is hot! An expose worthy of a BBC Panorama programme?

    ReplyDelete
  3. Is there any way of finding out how many of the EMA's PML cases on Tec and Gilenya are related to prior treatment with Tysabri? This is important for this patient's query.

    ReplyDelete
  4. lol. what more can i say?

    I'm not laughing at the pain of it all. I'm just trying to hold back tears of frustration and injustice.

    lol also at the comment that it may be worth a of BBC Panorama expo (that's genuine laughter)

    thank you for the blog and all the effort that goes into it.

    ReplyDelete
  5. Why do you call it "Pharma Fact vs. EMA Fact" when 90% of EMA funding comes from industry: http://www.ema.europa.eu/ema/index.jsp?curl=pages/about_us/general/general_content_000130.jsp&mid=WC0b01ac0580029336

    ReplyDelete
  6. Fake news everywhere, news networks are on the entertainment business nowadays.

    ReplyDelete
  7. what you are looking at is adverse event reporting vs clinically confirmed cases. In the U.K. This is called the yellow card scheme

    In the age of internet self diagnosis etc, I would expect you to know better than speculate in such fashion. People listen to you.

    ReplyDelete
    Replies
    1. Yellow card reporting is confirmed cases; they follow-up with requests for more information.

      Delete
    2. how is rebound speculation? you may speculate in some cases on whether the current or a previous dmt is the cause (and i agree wiht the question posed above - Is there any way of finding out how many of the EMA's PML cases on Tec and Gilenya are related to prior treatment with Tysabri? This is important for this patient's query. - but it is otherwise pretty hard to speculate on the rebound effect.

      Delete
  8. Anon 10.33 has a valid point, it is the European Database of SUSPECTED ADRs not confirmed cases. That said, it's an excellent database and very useful for pwMS choosing a dmt. Thanks for posting this, I did not know of its existence. Out of curiosity I checked Tysabri for PML, 926 reported of which 93 fatal. I think it is probably safe to say the fatalities were confirmed cases, unless of course it was actually something else... (Tecfidera 1, Gilenya 0)

    ReplyDelete
    Replies
    1. what mimicks pml but is not pml? lol. a bad ms attack?

      i don't necessarily disagree with you, but to me sometimes it comes to the duck test (if it looks like a duck and it walks like a duck and it talks like a duck, perhaps it's a duck ;)

      Delete
    2. ps. https://en.wikipedia.org/wiki/Duck_test

      Delete
    3. "....if it looks like a duck and it walks like a duck and it talks like a duck, perhaps it's a duck"

      But pharma spins it's a swan ;-)

      Delete
    4. lol so do my clients, but it's my professional responsibility to give them advice to the contrary. if i don't and it ends up in court the lady or man wiht a funny robe wearing a horses' tail on their head gives me a good speaking to and quotes that very phrase :)

      Delete
    5. Haha! Let's call it a duck then ;-) but if the EMA database is like the UK yellow card scheme it is very much 'if in doubt, report anyway' especially with serious reactions. Maybe some reports have been submitted prior to confirmation tests and not yet removed? Maybe multiple reports have been submitted for same event by different people and not yet merged? Or as anon 11.25 said maybe some of the reports should have been put down to Tysabri?

      Delete
    6. i think Prof G's "black swan" is just a big duck in swan's feathers:-)

      Delete
  9. I have a horrible feeling that this is based a definition of on what is PML and what is not PML. Pharma are necessary and I do not believe it is in their interests to be shown to be telling an out and out lie.

    Who do EudraVigilence report to and where does their funding come from?

    It is like believing everything you read in the British press. None the less it is very worrying when people who know what they are talking about are confused by the figures are this affects the advice that can be given to a patient

    ReplyDelete
    Replies
    1. Eudravigilence is funded by the European medicines agency, which is itself funded by the EU and also by fees charged to industry for processing medicine applications etc.

      Delete
  10. Prof G is using the term fake news as a general term in relation to a social trend. What he refers to is alternative facts, which is the better term. I assume Pharma base their facts on different criteria to the EMA. In reality the real answer will probably be somewhere between the two.

    ReplyDelete
  11. Fascinating. If you look-up Lemtrada and add-up all the Lisiteria cases (meningitis, sepsis, encephalitis, Listeriosis) you get 20. Any comment?

    ReplyDelete
    Replies
    1. We have reported on the issues of lemtrada listeria in the past have a search through the blog

      Delete
  12. Surely it's just like trial data, two similar studies yield very different complication rates because of their methods. DAPs has high sensitivity, the Spc has high specificity.

    From memory DAPs (MHRA equivalent https://yellowcard.mhra.gov.uk/idap) was always perceived to be a bit of a junk yard. I used to be reluctant to refer HCPs to the data as it doesn't mean much.

    The Pharma data weeds out inaccuracies, which results in lower numbers due to loss to follow up. The barriers to pharma collecting accurate data are numerous. (Getting A HCP to complete a drug safety form is like trying to catch fog). There are strict quality standards which all companies comply with (GVP guidelines/abpi) but most pharma employees find the whole thing very frustrating as a huge amount of money goes into PV and we never seem to provide anything close to a satisfactory service...

    ReplyDelete
  13. I have now had a formal response from Novartis. The EMA, or EudraVigilance, figures are essentially from two sources. One from EU members states and the other from Novartis. The Novartis figures are global figures. The numbers in the database are both suspected, and confirmed, cases of PML and as expected some of them will turn-out not to have PML and some will be duplicate entries. Novartis provide the EMA with information about whether or not the cases are associated with previous use of natalizumab.

    The following are Novartis’ official and latest figures:

    The overall rate of PML under fingolimod therapy not attributed to previous natalizumab treatment is very rare and estimated to be less than 1:10,000 patients.

    10 PML cases in ~184,000 fingolimod treated patients (~397,000 patient-years) as of November 2016.
    (Estimated risk (95% CI) is 0.054 (0.026, 0.1)/1,000 patients and incidence rate (95% CI) is 2.52 (1.21, 4.63)/100,000 patient-years exposure.)

    Two cases presented with confounding factors, 1 case had previous natalizumab exposure for 10 months (3 years 9 months before PML diagnosis).

    One additional case, PML occurred during 3-month natalizumab exposure, after 4.5 years fingolimod treatment, history of recent exposure to steroids.

    Nine patients were within the age range of 49 to 63 years, while one patient was 32 years old.

    In nine cases, fingolimod exposure ranged between 30 and 54 months, while one received fingolimod for 18 months.

    None of the patients had sustained grade 4 lymphopenia.

    ReplyDelete
  14. The Biogen figures are 5 confirmed PML cases in ~230,000 dimethylfumarate treated patients (~330,000 patient-years of exposure) as of October 2016.

    ReplyDelete
  15. I agree that HCP do not report as much as they should, so the levels of suspicion would have to be high to work on it.

    Prof G, if you find out how many PML cases for natalizumab there are according to Biogen, maybe it shows if the PML cases are over or under reported?

    ReplyDelete
  16. Any update on the PML numbers? Ive heard 8 for Tecfidera and 13 for Novartis.

    ReplyDelete

Please note that all comments are moderated and any personal or marketing-related submissions will not be shown.