Any drug that ends in the letters “mab” is a monoclonal antibody. These are manufactured antibodies designed to specifically target certain proteins and block or damage them. Tocilizumab is a monoclonal antibody that binds to the receptor for interleukin-6 (IL-6) preventing IL-6 from binding and having an effect on cells. IL-6 is a chemical known as a cytokine which promotes inflammation in the body; blocking IL-6 reduces inflammation, and also immunosuppresses the patient. Patients with severe Covid-19 have uncontrolled excessive inflammation which is causing damage to their bodies, especially their lungs, and it is believed that blocking IL-6 with Tocilizumab will reduce the amount of inflammation and save more lives. Great, a game-changing treatment!
Tocilizumab has been around for a while and is normally used to treat patients with rheumatoid arthritis. It costs about £500 per dose, and patients with Covid-19 can be given 1 or 2 doses depending on if they improve or not. This may seem expensive but current estimates are that an intensive care unit bed costs about £2200 per night!
It is hoped that Tocilizumab will provide additional benefit on top of corticosteroids which have been shown to reduce mortality in severe Covid-19 by about 30%.
So, what’s new about Tocilizumab?
On the 11th February the Randomised Evaluation of Covid-19 Therapy (RECOVERY) trial released the preliminary results of the Tocilizumab arm of their study. Up until this point we had 6 small trials of Tocilizumab use in Covid-19 which showed no benefit and 1 moderate sized study called the Randomised, Embedded, Multi-factorial, Adaptive Platform Trial for Community-Acquired Pneumonia (REMAP-CAP) study which showed benefit. This caused confusion as to what was the real result, some in the Tocilizumab corner and some saying not so fast.
The RECOVERY study is much bigger than all of the previous studies combined and so its results have been eagerly awaited.
RECOVERY
In RECOVERY, patients with Covid-19 who had low oxygenation (oxygen saturation < 92% on air) and evidence of inflammation (C-reactive protein ≥ 75mg/L) were randomised to receive either Tocilizumab PLUS standard care (corticosteroids) or standard care alone (just corticosteroids), all patients get respiratory support. Patients were then followed up for at least 28 days to see how many died at 28 days, how many went on to need mechanical ventilation and how quickly they were discharged if they got better.
Results were compared to see if Tocilizumab gave additional benefit over standard care (corticosteroids).
The headlines
So, we know the media and politicians love a good press release; the headline results of the RECOVERY trial Tocilizumab data was that “Tocilizumab reduced mortality by a third in patients with severe Covid-19, reduced the need for mechanical ventilation and reduced time to discharge”.
Wow! This is amazing! At least that was my first impression on hearing the news… but then I dug out the paper and read the detail…. Why do I do this, it only ever makes me disappointed!
What did the study show?
The media’s headline figure of a third is actually a misquote; the study’s figure was that for severe Covid-19 the addition of Tocilizumab to corticosteroids reduced mortality at 28 days from 33% to 29%. This isn’t that Tocilizumab alone reduced mortality by a third but that BOTH dugs together reduced mortality by a third; Tocilizumab alone reduced mortality by only 4% over corticosteroids alone. Don’t get me wrong this is still good, but not “game changing” AND I feel the authors have massaged this figure a bit as the data should actually be 29.5% versus 33.1% i.e. 3.6%... the authors have either rounded down both figures rather than rounding up the 29.5%... why do that?? Maybe to make the difference bigger and the result look better?!
If you then look at the data in more detail, the cause of death is only available for a proportion of the patients, and when this is taken into account the additional benefit of Tocilizumab in reducing mortality from Covid-19 as recorded on the death certificate is only 2.2%... and that’s about half of the headline figure of 4%! Some of these patients in this study have died from their underlying cancer or heart disease, they just had Covid-19 to deal with too but this was not the reason for their demise.
Tocilizumab is also reported as reducing the length of stay in hospital for Covid-19 by 7%, although this data is only known for 1093 of 2022 Tocilizumab patients and 990 of 2094 standard care patients. But as the data is not complete the actual result of the whole data may in fact be better or worse than the study suggests.
Subliminal messages
When you get stuck into the nitty gritty of the data again for the effects on mortality, length of stay and need for ventilation there is a clear message; if you don’t give corticosteroids WITH the Tocilizumab then all of the benefits disappear. In fact, the paper suggests in the supplementary data that giving Tocilizumab on its own may even adversely affect outcome. If this IS correct then this is really important, and the authors should give the clear message that if you can’t give corticosteroids for some reason then don’t just give the Tocilizumab on its own.
The other message in the RECOVERY study is that corticosteroids give most of the benefit in treating severe Covid-19. But we already know this! Don’t we?
Limitations
Patients were only eligible for Tocilizumab if they were getting worse within 21 days of previous treatments including corticosteroids therefore the study was automatically selecting patients who were failing corticosteroid treatment. As a result, this isn’t actually a study of Tocilizumab PLUS corticosteroids versus just corticosteroids; it is Tocilizumab PLUS corticosteroids versus patients failing corticosteroids. This selection of patients may under value the normal effects of corticosteroids as anyone who does well on corticosteroids alone is excluded from the trial, therefore this would over value the effects of Tocilizumab.
Patients were also excluded at the discretion of their consultant, so if a secondary infection was suspected the patient was excluded. On the face of it this is acceptable as immunosuppressing someone (the main desired effect of Tocilizumab e.g. suppressing inflammation in rheumatoid arthritis) who has an active bacterial or fungal infection will usually make them sicker. Yet by doing this it also means that the study may not represent a real world situation; where a number of the sickest Covid-19 patients who have been in hospital for a while may also have developed secondary infections. This again would potentially over value the benefits of Tocilizumab in all patients, then when it is used widely after the trial, in all patients, it may result in higher numbers of other infections and potentially higher mortality.
The data for the RECOVERY trial is also incomplete and a bit weird. Of the 2022 patients randomised to receive Tocilizumab only 1602 had data available for analysis and of these only 1333 actually received the drug! What happened to the 269 who were meant to get the drug? Why are they in the Tocilizumab group if they don’t get the drug? The RECOVERY team even say they don’t know why…. Also 44 patients who shouldn’t have got Tocilizumab actually did… this is inaccurate at best and I believe a bit weird. This may all be fine but the study seems to oscillate between the values 2022, 1602 and 1333 depending on what they are “reporting”…. Is this done to massage the figures to improve the effects of Tocilizumab or is it all above board? I couldn’t tell and that makes me a bit suspicious. Remember I am a cynic, I believe if it’s “as-good-as-all-that” you’d present it clearly, consistently, loudly and proudly!
Another problem I have is the study design; in particular the way “cause of death” data was collected for this study raises some concerns with me about the validity of saying what the patient died from. For starters the data is incomplete with the cause of death only being available for 476/596 Tocilizumab recipients, and 539/694 standard care patients. Also, the cause of death was taken from what was written in part 1a of the death certificate.
Part 1a of the death certificate is the disease or event that the doctor thinks was the immediate cause of death e.g. acute myocardial infarction (heart attack) or pneumonia etc. I suspect that during the pandemic anyone who died with a diagnosis of Covid-19 had this recorded as 1a, e.g. the patient may have had a heart attack from the strain placed upon their body from Covid-19 but “Covid-19” did not kill the patient the heart attack “did”, likewise a patient may die of a ventilator associated pneumonia due to a bacterium, or sepsis from a central venous catheter infection while having a positive Covid-19 result.
I think the fact that Covid-19 was recorded as 1a on the majority of certificates could therefore be an overstatement, as many patients will have died weeks after admission when the original Covid-19 infection had resolved. In this case Covid-19 should not be part 1a but should be 1b or 1c i.e. it contributed to death but wasn’t the final cause of death, in which case Covid-19 shouldn’t come up as the cause of death in the RECOVERY trial method. They are saying Tocilizumab reduces death from Covid-19 but we can’t be sure that Covid-19 was the final cause of death in these groups. Normally studies would collect data directly from the medical teams and more detailed data would be sought when designing a research study, not just the scanty death certificate details as in the RECOVERY study.
Potential pitfall
There is one major question that the RECOVERY trial hasn’t addressed. To be fair the study wasn’t set up to answer the question as it was done in a hurry… but in my opinion it should have been. My question: is it reasonable to expect a clinical trial using immunosuppression as a treatment to collect data on infections as a result of that immunosuppression? I would have thought so… I think immunosuppressing a patient and NOT following up any resulting infections is a design flaw but RECOVERY didn’t do this. That’s a significant oversight in my eyes.
Tocilizumab and corticosteroids profoundly immunosuppress patients leaving them vulnerable to severe bacterial and fungal infections. Not only are these patients at risk of infection, but blood markers such as C-reactive protein, which we use to monitor for developing infections, are suppressed by Tocilizumab for months. That makes infections more likely AND makes them more difficult to diagnose. I am really disappointed that RECOVERY didn’t look for infections as a potential serious adverse consequence of using Tocilizumab and corticosteroids in these patients. What’s the point of treating Covid-19 if the patient dies from bacterial sepsis 2 months later?! My anecdotal experience of Tocilizumab is that it has high associated rates of severe bacterial infections when used to treat Covid-19 but then I may be biased by only being involved in treating the sickest Covid-19 patients and not seeing those who get better without any complications. I tend to only meddle, I mean give advice, in the sickest patients.
Okay, the current RECOVERY data shows a reduced mortality, but remember that data is incomplete and the effects of Tocilizumab last months, not just 28 days. We need the promised 6 month data to know if Tocilizumab is really a benefit in the treatment of Covid-19 or not… and that won’t be until towards the end of July this year… the same time that Boris Johnson has us all out of lockdown and the pandemic is over… let’s hope hindsight is kind when we look back on our usage of Tocilizumab now!