#MSCOVID19: anti-CD20 nuances


Barts-MS rose-tinted-odometer: ★★★

I had to have a detailed discussion with a patient this week about starting ocrelizumab during the COVID-19 pandemic. This patient was concerned about (1) the data showing an increased risk of COVID-19 in anti-CD20 treated patients, (2) an increasing number of cases not seroconverting when being infected with the coronavirus and (3) she will not be able to have a coronavirus vaccine when one emerges. 

All of these three concerns are not black-and-white, but very grey, and need explaining. 

1. The increased risk of COVID-19 and severe COVID-19 on anti-CD20 therapy

This finding has been reported by the Italian register, the Iranians, the Swedes and the MSIF’s Data Alliance initiative. There is also a similar signal in the US data and French data. However, despite the increased risk of COVID-19 it is not associated with increased mortality. In the Italian data there is a weak signal that the longer you have been on an anti-CD20 therapy, particularly more than 3 years, the greater the risk of COVID-19. 

One issue that is not adequately addressed is reporting bias, i.e. more severe COVID-19 cases get reported and the less severe ones don’t because they are not registered as having COVID-19. Reporting bias is likely to affect DMTs that require patients to attend hospitals, i.e. treatments that bring them into contact with HCPs and hence they have a chance to report symptoms, and those DMTs that are the most widely prescribed. It is clear that both of these factors play a role in anti-CD20 therapies. One clue in support of this is the fact that there is a similar signal of a greater COVID-19 and severe COVID-19 signal emerging with natalizumab, i.e. another DMT that requires frequent hospital visits (reporting bias) but is less prescribed than anti-CD20 so the signal is not quite significant yet. 

One way to address this issue is to study non-biased trial populations, which has to be the data gold-standard. Roche presented 51 COVID-19 cases in over 4,000 ocrelizumab exposed trial patients (1.3%) (see presentation below). There was clearly no link between COVID-19 and treatment duration (slide 6). Three patients out of 51 patients died (5.9%). The numbers are too small to make a call on whether this represents a higher mortality than the background rate. Another important factor is that COVID-19 was not linked to hypogammaglobulinaemia. This ‘gold-standard’ data challenges some of the dogma that has emerged around CD20 therapies and COVID-19 and alters my interpretation of the Italian data.   

(2) An increasing number of ocrelizumab-treated cases not seroconverting when being infected with the coronavirus

Although there are several cases of SARS-CoV-2 positive COVID-19 cases on ocrelizumab who have been reported that have not seroconverted (detectable antibodies) there are many more normal people who have have COVID-19 who haven’t seroconverted as well. Whether this observation is assay dependent, i.e. insensitive assays that don’t detect low level antibody, or true biology needs further investigation. The fact that these people, ocrelizumab-treated or normal people, recover from COVID-19 is telling us that an antibody response is not necessary to clear the virus and for recovery from COVID-19. It is likely these patients have very good cellular immune responses that will protect them from reinfection in the future.

These observations have implications for vaccine responses and hence we may have had the wool pulled over our eyes focusing on the easier to measure antibody responses. I suspect, as do many others, that it is not humoral, but cellular, immunity that will be important for  protective immunity against SARS-CoV-2. I am not saying antibody responses won’t be important, but it is likely the dominant protection will come from cellular immunity. 

(3) Not being able to have a coronavirus vaccine when one emerges

This is clearly not correct. Patients on ocrelizumab will be able to have DNA, RNA and component coronavirus vaccines. The only vaccines they may not be able to have are live attenuated vaccines and potentially vaccines using a live viral vector to deliver the immunogen. The question is whether or not ocrelizumab-treated patients will be able to mount an adequate protective immune response to these vaccines is a moot point, which is why I have been urging Roche to plan and set-up registry studies to see if ocrelizumab-treated patients develop adequate immune responses to these vaccines. It is important that these studies are well designed and include both antibody or humoral and cellular components. 

Another thing to remember is that no vaccine is likely to be 100% effective. Even if the vaccine is only 60% to 70% effective it will be sufficient to create herd immunity and stop the spread of coronavirus. Vaccines are about population health and not necessarily about individual health, which is why regulators are fixated on safety.

I also told this patient that there are many other factors at play. For one the death rate or mortality from COVID-19 is falling. This is happening for several reasons. Firstly, we now have approved treatments for COVID-19 and the circulating strains are likely less virulent than the initial strains. Based on simple evolutionary principles the more benign strains are out competing the virulent strains. Another factor that I have commented on before is that we may be nearing herd immunity in some areas of the country, for example in London. This is based on the observation that many people have cross-reactive cellular and humoral immunity, presumably from other coronaviruses, that are protecting them from getting COVID-19. Therefore the risks of getting COVID-19 are falling. Add to this sensible personal hygiene and social distancing and the risks remain low. The second surge is really happening in areas of the country with low herd immunity and amongst care-free populations who are not being adherent to the government’s guidelines; for example, University students. 

The bottom line is that if you are low risk of getting COVID-19 and you double that risk the risk remains low. I would also only cross the vaccination bridge when it arises. Trying to preempt when and what vaccine will emerge first is really a guessing game. What is not a guess is that this patient has active MS, with a relatively poor prognostic profile (spinal cord disease) and needs treatment. She does not have highly-active MS therefore the only high-efficacy DMT available to her first-line on the NHS is ocrelizumab. She could select a platform therapy, which has also been offered to her, and to then see how she does, but as she is very well informed and understands the concept of ‘flipping the pyramid’ and that ‘time is brain’ doesn’t want to take a chance on a lower efficacy DMT. The outcome from our discussion is that she has decided to go ahead with ocrelizumab after she has had her pneumococcal and seasonal influenza vaccines. 

As you can see the COVID-19 anti-CD20 data is quite complex with a lot of nuances, which makes it difficult to communicate to patients. But if you take time to explain it to patients, not only do you allay their fears, but you end-up with a well informed patient who knows what they are signing up for. 

CoI: multiple

Twitter: @gavinGiovannoni             Medium: @gavin_24211

About the author

Prof G

Professor of Neurology, Barts & The London. MS & Preventive Neurology thinker, blogger, runner, vegetable gardener, husband, father, cook and wine & food lover.


  • I think the key phrase here is neuros “take the time to explain”. If only.,. If I didn’t have the internet I wouldn’t know what MS was.

    Anyway, thank YOU for taking the time to explain. This is all good stuff to know and is going in my email folder of stuff to discuss with the doctor in Round 2 of Me vs Maintenance-Escalation therapy.

    Could you tell us how you define “Active” vs “Highly Active” MS though?

    Thank you!

    • Active naive = relapses in the last two years

      Active on DMT = relapses or MRI activity in the last 2 years

      Highly active naive patients = two relapses in the last 12 months with Gd-enhancing lesions at baseline or an increase in T2 lesion load on MRI in the last 12 months

      Highly active on a DMT = one or more relapses in the last 12 months with Gd-enhancing lesions at baseline or an increase in T2 lesion load on MRI in the last 12 months

      Rapidly evolving severe = two disabling relapses in the last 12 months with Gd-enhancing lesions at baseline or an increase in T2 lesion load on MRI in the last 12 months

      • Okay, so I am “highly active on a DMT”. My problem, where do I go next? I had adverse reactions and breakthrough activity on my first line anti-CD20, plus additional lesions in most recent spine and brain MRI (24 months since initiating anti-cd20). Your paraplegia comment is very concerning to an individual like me, particularly given my age

  • Thanks Prof, very useful (and slightly reassuring) information. I have a question about your patients ‘relatively poor prognosis profile’. I understand that spinal cord lesions are one of the prognostic risk factors in your tally chart but my understanding was that a low score on the other factors could still place you in the green zone (30% – good prognosis). I know that its not an exact science and that spinal activity is ‘relatively’ worse than activity in the brain but is it sufficient enough to lead to poor prognosis?

    • You have to understand that the prognostic factors are only really relevant at baseline and refer to the pre-DMT era, i.e. what would happen to this person in the future if they are not treated. With treatment, we think we are changing the prognosis for all people. This patient had a few cervical spine lesions and hence this was the dominant factor in terms of her prognosis. She has bladder problems and reduced walking distance due to fatiguable weakness in one of her legs. In other words, she has reduced reserve capacity in her motor pathway. One more large cord relapse may be all that is necessary to cause major problems, e.g. paraplegia. This is why I recommended flipping the pyramid. It is a pity she didn’t fulfil our criteria for having highly active or rapidly evolving severe MS then the treatment options could have been extended to natalizumab, cladribine and alemtuzumab.

      • Prof G,

        As you can imagine, the thought of a relapse causing paraplegia is a nightmare scenario and differentiates this disease from other so called auto immune diseases.

        For patients who have had treatment with a highly effective induction therapy, yet still have deficits caused by earlier spinal lesions (as your patient above), is there any research looking at how to stimulate repair / regrowth of axons etc.? I recall some time ago a lot of research work being undertaken looking at repairing damaged spinal cord (some funding coming from the late Christopher Reeve – the actor paralysed after a horse riding accident). However, like a lot of neurological research, nothing seemed to come of it.

        • Yes, there are several agents in quite late-stage development for remyelination and repair. The most exciting are opicinumab (Biogen) and elezanumab (Abbvie). The other appealing option is add-on neuroprotection and this is where there is a hole in late-stage development despite quite promising agents that have been identified.

          • Didn’t opicinumab pretty much fail Phase 2? It might have been up to study design and the implied lack of power, but my take way from Synergy was less than positive… A slam dunk it is not.

            I really wish they would start doing these trials as combo with high efficacy DMTs.

      • Thank you…. I expect that i am in a similar boat, albeit lower down the cord. My follow up question has already been asked (by Jo) so i will leave it there but where is MD with his questions of the month 😁

  • Hi,
    All you guys have been talking about since this pandemic started is AntiCD20 therapies and their effects on COVID-19.
    What about rest of us on Mavenclad therapy? Where are we in all that? For instance, I am in my depletion on base now, having finished my 1st year/2nd dose two months ago.

  • MD here
    In the interest of balance I just want to add abit of complexity to that presented.

    It has been reproted that people on long-term use of ocrelizumab and rituximab can develop reduced levesl of IgG. This is associated with a small increased risk of inection (Derfuss et al. Mult Scler 2019; 25 (Suppl 2):20)

    In the data held by Roche, none of the people who got COVID-19 had a reduced level of IgG (Hughes et al. SS02.05. MSVirtual 2020 ECTRIMS/ACTRIMS 2020)

    However I would add here that increased risk of severe COVID-19 has been associated with lack of an effective IgM response (Shen et al. Emer Microbe Infect 2020; 9:1096)

    Ocrelizumab can be associated with a rapid and marked reduction in IgM levels, even when IgG levels are normal (Baker et al. Mult Scler Rel Disord 2020. doi:10.1016/j.msard.2020.102174).

    Therefore it will be important to show that IgM was not reduced in the COVID-19 individuals (I bet it will be for most people), as it often drops within months of treatment onset as this could contribute to the infection risk….I do not think it will influence infection…but if the infection becomes symptomatic, remembering that 40-80% of people get asymptomic disease. If you don’t get rid of the virus the viral load increases increasing chance of symptomatic disease.

    ProfG is correct a large number of people who deal with the virus and exhibit no or mild disease often have undetectable antibody levels, indicating that the innate (macrophages + neutrophils) and T cells can do a good job of getting rid of the virus. Boosting this with a vaccine may be all that you need

    • The unanswered question, that can be answered by doing some detailed immunology studies, is do pwMS on ocrelizumab have lower pre-existing cross-reactive immunity or is the risk associated with a blunted primary or initial response to the virus? I have difficulty accepting that pwMS on ocrelizumab are more exposed to the virus hence greater infection. I prefer the hypothesis that they have less cross-reactive immunity therefore less likely to have asymptomatic and/or milder infections.

  • If patients on Ocrelizumab showed a decrease in post-COVID antibody production (point 2), was this not also a potential bias for some Registers (point 1) in the hypothesis that patients with a severe form were more easily tested by PCR in the first wave due to the limitation of testing. On what basis (PCR or serology) were COVID-19 diagnoses confirmed in registers?

  • Re:Cellular immunity to SARS CoV-2 may be more important than humoral immunity. In order to estimate the prevalence of the virus in the population are there effective assays to measure cellular immunity if antibody response is difficult to detect at the lower end?

    “Cross reactivity to other coronaviruses is thought to protect against severe illness.”Can these common cold viruses be used in an attenuated form to provide a level of cellular immunity to protect people from developing severe illness?

  • Do we have any data that filters out the effect of the steroids administered during the infusion every 6 months? How long do the steroids administered during the infusion continue to have effects on the immune system? Obviously both the steroids and anti-CD20 therapy are immunosuppressants, but the data around a lower death rate but higher everything else sounds a lot like the effect of steroids.

    • Good point, but the effects must wear off quiet quickly otherwise they would give so often to people with allergies. In addition alemtuzumab would be associated this the same steroid risk, but the data does not show it is like rituximab/ocrelizumab risk

  • All things being equal allowing CD19 repopulation to return to 1-3% levels is understandable.

    Haven’t seen a discussion about how long to defer Rituximab after receiving the Pfizer vaccine. 1-2 weeks? Longer?

By Prof G



Recent Posts

Recent Comments