ClinicSpeak: how high is your risk of getting PML?

PML risk higher than expected. Why? #ClinicSpeak #MSBlog #MSResearch

“The following commentary suggests that the PML risk on natalizumab is higher than the Biogen figures. Why? I am sure this relates to the denominator and have discussed this before many times. People with MS who are JCV negative should be excluded from the risk ratio. At present the denominator includes all pwMS on natalizumab (JVC+ve and JCV-ve) subjects. As we derisk people by  taking them off natalizumab the number of JCV+ve people in the denominator is going down. Therefore the current methods for calculating PML risk is underestimating the risk. This is why we have made a decision to get all of our JCV+ve patients off natalizumab. What has changed is that we now have other treatment options. In the past we didn’t have other options; therefore why take unnecessary risks? Despite this we have some patients who have done so well on natalizumab is that they are reluctant to switch treatments. What I do in this situation is tell them that if their anti-JCV index goes up by more than 10% it indicates that their virus is active and boosting their immune system and if they want to avoid getting PML they must come off  natalizumab. Please remember the process of developing PML is very complex; the JCV virus has to acquire many mutations to be able to cause PML. It is unlikely that when you stop natalizumab that these processes unwind themselves. This means you will carry some of your PML risk with you when you switch to other therapies. This is important when you move over to any treatment that suppresses your immune system. This is why we are going to see increasing number of ex-natalizumabers developing PML on other follow-on drugs.”

Epub: Berger & Fox. Reassessing the risk of natalizumab-associated PML. J Neurovirol. 2016 Feb.

The risk algorithm for natalizumab-associated PML was first established in 2012 using the observations that JC virus antibody status, prolonged duration of natalizumab therapy (>2 years), and prior exposure to immunosuppressive therapy increased the risk for the disease. Prior to the publication of Biogen’s algorithm, a risk algorithm was created by Fox and Rudick using an Excel spreadsheet in order to address the concerns of their patients. Applying the most recently available data regarding natalizumab-associated PML, the risk assessments for PML were recalculated. The current numbers indicate substantially higher risks for PML in 2015 than in 2012. Our calculations suggest that an individual having all three risk factors has an approximately 1 in 44 chance of developing PML.

CoI: multiple

31 thoughts on “ClinicSpeak: how high is your risk of getting PML?”

  1. I' m been on Tisabry 12 years and I' m JC positive, they take lot of blood and twice a year MRI scan to check. Tisabry helping me a lot.I'm worried now started panick attack 😁

    1. I am somewhat similar to cb. I have been on Tysabri for 10 years and recently seroconverted. It was clear I had recently been infected as my levels were off the charts so to speak. My Neuro put me on a drug holiday and tested my titers 3 months in a row and an MRI to make sure there were no changes. I am now on extended dosing every 8 weeks with no change. I am still awaiting the data for extended dosing.I am in the camp that for all of the JC + Tysabri patients switching to any DMT we are all still at risk of PML. We simply do not have a complete understanding of what this virus does over time as these meds mess with the natural elimination of the virus when not on the meds.I feel safer on Tysabri, the beast we understand, rather than switching to Fingoliod or Tecfidera, the beasts we really do not have enough data on yet.For now I am monitored very closely. Those same protocols are not in place yet for the other DMT.

    2. I have been on Tecfidera over a year, I'm in the UK. I am monitored every three months with blood tests to check lymphocyte levels. The three month monitoring will continue and will change to blood test every month if lymphocyte levels drop below 0.8. Then it will be repeated at one month and reported urgently.

    3. I'm another long-term JCV+ Tysabri patient that feels I'm doing well on the medication and am monitored closely. Although MS has affected many aspects of my life, being on Tysabri has enabled me to maintain a good quality of life.The issues for me in thinking about switching over to another DMT are :1. No-one can give me any idea of the magnitude of the continuing risk of PML once I switch to another DMT. 2. In addition to an unknown continuing risk of PML, the other highly-effective DMTs come with risks of major side effects such as other auto-immune disorders.3. Am I going to have a sub-optimal response to the new DMT ?4. I am very concerned about MS rebound between coming off Tysabri and going onto another DMT. I would be releasing the shredder and who knows what havoc it would wreak this time ?I don't feel I am sticking my head in the sand but rather my evaluation of known and unknown risks vs known and unknown benefits results in the current wish to remain on Tysabri.

    4. My thoughts exactly and none of the other DMT seem to have the same efficacy. I am eager to learn and see more data as it comes out.

  2. The rebound effect (EM back in full force) of Natalizumab interruption is more common in those who are JCV positive or negative? Or is it both?

  3. That is wonderful and to be expected for Tecfidera. Were you on Tysabri prior to this? Now that a few cases of PML have showed up on Tecfidera, I would hope they would add JC antibody testing. That was what I was referring to.

    1. I'm JCV+ and Tec is my first DMD. The hospital tested JCV before starting Tec. I started off with very active MS but things have settled down since being on Tec. I should of been on a DMD a lot sooner. If Tec stops working for me in the future I doubt I will be offered Tysabri due to being JCV+.

  4. I’m not a statistician or number cruncher but it seems to me that to develop a complete picture on the tracking of PML/drug/JCV status there needs to be much more “crossover” data being recorded:- All PwMS on natalizumab – both JCV+ve and JCV-ve – and continue to record data on those who remain JCV-ve.- How many convert to JCV+ve – these should then be added to the database for patients remaining on Nat despite their JCV+ve status- How many go off Nat, what drug they move to and how many subsequently develop PML i.e. collect similar stats to that on those who are still on Nat and are JCV+ve- This should all be tied in with collection of data on previous DMT use prior to patients going onto Nat- Maybe the JCV status of all patients starting one of the newer drugs where PML risk may be involved in any way should be tested before they start on that drug.I’m unlikely to be put on Nat (not currently on any DMD and I think I’m now progressive anyway) but the emerging picture seems to be that once you get on the train with Nat you are effectively hopping into bed with the JCV/PML risks and it’s a train that you can’t completely get off. However, what is starting to emerge now makes me unlikely to consider it, even if I was offered it as a treatment choice.I’m sure much of this data exists in “isolated” records/databases, but this is where Big Pharma needs to work in co-operative ways across corporate boundaries with data collection so that patients (and their neuros) contemplating drug choices can be more accurately informed about the true risks of what their choices mean. It would take time for the data to accumulate, but there has to be a starting point somewhere.

    1. Well said. I agree. The only way to understand the long risk is to collect mounds of data and share the data across DMT.

  5. Isn't this tysabri's 3rd crisis since the completion of its golden clinical trials and release by the FDA in 2004? Just sayin' given the extraordinary value we put on golden standard clinical trials.

    1. It shows the value of long term follow-up that you have the data,so if you go and have an off-label treatment who is collating the data so if you do ensure you are registered with MS register in UK, and other registries elsewhere.

    2. Yes yes, but my point stands. The study also shows that the level of JCV infection is higher than preiouvsly thought. That can't be explained by faulty maths? "They found people converted from being anti-JCV negative to anti-JCV positive at the following annual rates: 10 percent in the German group and nearly 9 percent in the French group. Those rates are much higher than the rate of 1 percent per year for the general population and for people with multiple sclerosis not treated with natalizumab. In the German group, 43 of 339 people who were initially anti-JCV negative tested positive for the antibodies during the study. In the French group, 41 of 243 people who were anti-JCV negative tested positive for the antibodies during the study."

    3. Also, while if my gf came down with PML during her time on Tysabri it would have been reported to the manufacturer, she sat for 6 months on Tysabri and no one collected her data. Kind of like if she had a rebound effect post Tysabri, no one was really worried about collecting her data…..

    4. Bojana,Thank you for referencing the paper that you did. I just emailed Dr. G asking him to talk about those studies. I too will also comment that the Golden phase is just a glimpse of one span early in the lifetime of that medicine. All of the new DMTs need to have reams of data collected for very long times.I am a scientist and fully believe that data is important. I am also a patient with MS. So living in the US how do I make sure either I or my Neuro has me registered with the proper MS registries and is indeed reporting the proper information? I do know that all patients in the US on Tysabri are part of the TOUCH program. Unfortunately, I am not going to assume Biogen is collecting the right data let alone sharing it.Thank you all for sharing.

    5. Hi Debbie,I am not a scientist, though I feel over the past year I've had to play one. I believe the paper referenced by Prof G in this post is the same as my quote. Yes, Tysabri patients here are also part of the touch program. I also know that my partner's infusion related side effects were not reported to the program by the hospital that administered the infusions (it actually took the hospital a month to figure out she was having an infusion reaction). Then my partner's neuro told her antibody testing for tysabri wasn't available until the end of the 6 months… I called Biogen and found out that wasn't true. It's also how i found out her infusion reactions were also not reported. What does Biogen do with this data – wouldn't have a clue. Have been too busy trying to ensure my partner is treated right by her clinicians…

    6. None of this is easy as we are all finding out. The more we learn the more questions we ask. I have found that an open relationship with my Neuro and wonderful sources of information and learning like this Blog go a very long way in making good decisions

  6. It seems this denominator inclusion error is an elementary mistake. Most likely it is not a mistake but was a deliberate attempt to mitigate the risk of using this multi-billion dollar drug.If someone gets PML using Biogens risk analysis (and proudly disseminated by Team G through this marketing blog), a smart lawyer could have a field day.

    1. Re: "It seems this denominator inclusion error is an elementary mistake." I have asked Biogen for the data several times. They are not allowed to use the data collected via routine JCV testing as they don't have ethical approval to use it.

    2. Re: "… marketing blog…."The definition of marketing is 'the action or business of promoting and selling products or services, including market research and advertising'.I will let you decide what we are promoting, or not promoting, selling or not selling. Our mission is not to sell anyone anything, we are started out simply trying to interpret good, bad and other research news (ResearchSpeak). We have added a few other functions such as education (ClinicSpeak), CME (NeuroSpeak), politics (PoliticalSpeak) and crowd-sourcing (CrowdSpeak). We have no intentions of starting SellSpeak or MarketingSpeak.

    3. At Barts-MS it is our current treatment guideline to try and derisk all our patients on natalizumab who are JCV-seropositive. Why take unnecessary risks when we now have alternative treatment options. Despite this there are people with MS who don't want to stop natalizumab. Why? Because it has changed their lives. It is an informed decision and we need to respect that; we live in an era of the empowered patient. What you need to remember is that the risk of MS is sometimes so bad for some pwMS that risks of getting PML or dying from a treatment are worth taking.

    4. I certainly understand this line of thinking. But in my opinion there is just not enough data about the long term risk for PML for pw MS who are JC + who have been on Tysabri who switch to another DMT. This is why extended dosing was my choice.

    5. So far I have not seen the data I should clarify that looks at the cohort of past Tysabri JC+ patients who have switched and their further risk of PML. There simply is not enough data is my understanding and to date simply not enough time for PML to rear its ugly head. I would certainly be happy to read more information in order that I could make an informed decision to switch. My Neuro has looked at the adverse profile of both Tecfidera and Fingolimod and feels that risk does not out weigh my risk of PML on extended dosing.

  7. Prof G have you read Julian Borchardt's analysis of the risk ( He estimates that Biogen’s estimates of the PML incidence are wrong: months 25–48, prior IS Biogen = 11.2❻ (1 : 89) JB = 19.3❻ (1 : 52)months 25–48, no prior IS Biogen = 5.3❻ (1 : 189) JB = 7.4❻ (1 : 135)months 49–72, no prior IS Biogen = 6.1❻ (1 : 164) JB = 10.8❻ (1 : 93)What do you think?

    1. Re: "Julian Borchardt's estimates"Yes, I am aware of this works and downloaded and read the PDF. I sent it to a very eminent MS statistician who informed me that the methods JB was using involved a cumulative score, which was incorrect for epoch analysis. In other words people who develop PML in year 2 should not be included in the numerator in year 3 and then in year 4. The statistical methods in the PDF are easy to follow and I simply did not spend time to get my head around them, hence calling in the help of a statistician. We have subsequently found out that Julian Borchardt is a pseudonym and not a real person. We have recommended that he submit his paper for peer-review and possible publication. The latter is how science works. I am hoping to see his analysis published in a peer-review journal. Please note I am not trying to protect Biogen's estimates. I have been pointing out for sometime now that they are really the only estimates we have and as calculated they are underestimating the risk. By how much they are underestimating the risk I don't know. Therefore you can say I agree with Julian Borchardt when he claims Biogen are underestimating the risk. What I don't know, however, if his figures are correct.

    2. "I sent it to a very eminent MS statistician who informed me that the methods JB was using involved a cumulative score, which was incorrect for epoch analysis. In other words people who develop PML in year 2 should not be included in the numerator in year 3 and then in year 4."You should have this eminent statistician write his or her own explanation of this problem. As stated, the explanation you gave is either false or incoherent. The paper provides both epoch and cumulative analyses, and does not include cases from previous epochs in the numerator of later epochs; see the $q_i$ calculations in Table 11 for a clear illustration of this.By the way, I am an applied statistician myself, with a Ph.D., etc. The math in the paper is absolutely trivial though. You don't need a Ph.D. in the field to understand it, or see how Biogen's previous estimates were flawed. See my comment on the latter issue here:

  8. "I have asked Biogen for the data several times. They are not allowed to use the data collected via routine JCV testing as they don't have ethical approval to use it."So Biogen's position is that can't share data because they don't have permission from patients… they'd love to but just can't, it would be unethical.Well, how many Tysabri patients have been asked for this permission? Zero in our family and among our several acquaintances at the infusion center. Anyone know of any patients who have been asked? No? The faux moral compass fails miserably when we consider Biogen withholds data useful to those making, in some cases, life and death treatment decisions. Sharing anonymous data is unethical but withholding life and death information isn't?

  9. I've read Biogen's peer reviewed paper "Anti–JC Virus Antibody Levels in Serum or Plasma Further Define Risk of Natalizumab-Associated PML" (2014) and they clearly state that they use the number of JC+ patients at the denominator.Could you please clarify whether you mean that this paper understimates the risk or do you refer to another publication?

  10. I do not wish to cause panic in the ranks but how the TOUCH program is implented in the US leaves much to be desired since Biogen does NOT mandate following JCV Ab status q 6 monthly and it allows physicians who do not check or do NOT KNOW to check to prescribe the drug. The unsuspecting patient in those scenarios has no idea about his/her JCV Ab status. To date, 635 PML cases have been reported worldwide (as of March, 2016) and even if some or one of them falls into the category of unchecked JCV Ab status, it will become a monumental headache for Biogen as someone will want to scour all cases accumulated thus far (of PML) to have the data audited. Stratify testing was FDA approved in 2012, so beyond that point, there is zero excuse for any doctor not to check the JCV Ab status and even more riskier for someone to be allowed to be on it, while being in the TOUCH program !Dr Avasarala, MD, PhDGreenville, SC, USA

Leave a Reply

%d bloggers like this: