Risk stratification does not decrease the risk of PML

PML risk stratification is not working. Why? #MSBlog #MSResearch

“The following editorial claims that the risk stratification for natalizumab-associated PML risk is not working and challenges the evidence. The authors provide several reasons why it is not working.”

“My gut feeling is it is working as the number of new cases of PML reported each month seems to be dropping despite the increasing use of natalizumab. How could this be possible? May be the reporting or notification system is not working and a large number of PML cases are being missed?”

“What you need to remember that the risk groups predict your risk over the previous epoch and are linear in the next epoch. So if you are JCV positive, have had no prior exposure to immunosuppression and have been on the drug 25 months the risk of you having developed PML in the last 24 months was 1 in 1429; your risk in the next 24 months is 1 in 189. What is not clear is that the risk is not spread evenly across this next epoch and goes up with time. What is also not captured in these figures is the fact that the denominator is changing; the number of MSers at risk of PML staying on the drug goes down with time therefore the number of cases (numerator) are being divided by a changing denominator. For this reason I suspect that the actual risk after 2 years of treatment is much higher than that reported. I tell all my patients it is simply a numbers game if we leave enough MSers at risk of PML on this drug someone at our centre will develop PML. With new treatment alternatives why should we gamble? I simply don’t want anyone to develop PML at our centre. Despite this some MSers choose to stay on the drug as they say the risk is acceptable. I therefore negotiate with them and put them on 3-monthly MRI monitoring and monitor their anti-JCV antibody levels every 6 months; any change on MRI or a significant rise in antibody level triggers a switch. So far all of my reluctant-switchers have signed-up to this strategy.”

Epub: Cutter GR, Stüve O. Does risk stratification decrease the risk of natalizumab-associated PML? Where is the evidence? Mult Scler. 2014 May 8.

The use of natalizumab has likely been limited by its association with progressive multifocal leukoencephalopathy (PML), an infection caused by the human polyomavirus John Cunningham (JC). Three factors were recently identified that contribute to the overall risk of natalizumab-associated PML: (1) Positive serostatus for anti-JCV antibodies, (2) prior use of immunosuppressants, and (3) duration of natalizumab therapy. This risk stratification algorithm has not led to a reduction in the incidence of PML in natalizumab-treated patients with multiple sclerosis between April 2010 and February 2014. This observation may appear perplexing, as treatment duration and JCV serostatus are modifiable risk factors. Potential reasons for the lack of success of companion diagnostics that determine the overall risk of natalizumab-associated PML are discussed.


…… it appears that the risk stratification algorithm may not have been a success. The reason for this failure is unknown. Possible explanations are:
  1. It is possible that neurologists have difficulties interpreting the risk stratification algorithm. Also, the fact that natalizumab treatment cessation is associated with MS disease reactivation makes the risk assessment of PML versus that or MS quite complex in some instances.
  2. Any protein detection assay has a defined lower limit of detection, and there are numerous cases of natalizumab-associated PML that had previously tested negative for JCV with the above-mentioned ELISA.Thus, these cases either represent de-novo infections, or the test results were false negative.
  3. Neurologists may primarily focus on the JCV serostatus when assessing the overall risk of natalizumab-associated PML, and neglect other risk factors. As stated above, prior pharmacological immunosuppression increases PML susceptibility in patients who at the time of natalizumab initiation are JCV-negative, but subsequently seroconvert undetected.
  4. It may simply be too soon to detect a possible impact of risk stratification due to the fact that a high number of person years of patients on natalizumab treatment is dominated by the first year of exposure, during which potential risk changes are more difficult to detect due to the lower incidence.
  5. With incident rates in the range of 1–10 per thousand, it theoretically only takes a single case to maintain the current event rates.
  6. It could be that patients are willing to forego the risks in favor of the drug despite the excess risk
  7. Examining rates by initial JCV-positive and negative status would help clinicians be sure that this risk mitigation strategy is benefiting their patients, and indeed worth the cost.
CoI: multiple

Leave a Reply

%d bloggers like this: