Expert Perspectives: Q&A on Actemra’s phase 3 EMPACTA study

by Bob Dickson

In the first of an ongoing series of “Expert Perspectives” with members of inDemic Foundation’s scientific advisory board, we discuss data from EMPACTA, a randomized, double-blind, placebo-controlled phase 3 trial evaluating the efficacy and safety of Roche’s Actemra (tocilizumab) in hospitalized patients with COVID-19 pneumonia (NCT04372186). Actemra is one of several IL-6 inhibitors that’s been evaluated for patients with COVID-19.  At the time of this writing, data from EMPACTA has not been published in a peer-reviewed journal, but some of its key findings have been released by Roche/Genentech.  

This Q&A is with Bob Dickson, MD, associate professor of pulmonary and critical care medicine at University of Michigan, and Hrishikesh Kulkarni, MD, MSCI, assistant professor of pulmonary and critical care medicine at Washington University in St. Louis.  

On September 18, 2020, Roche announced that the phase 3 EMPACTA study met its primary endpoint, showing that patients with COVID-19 associated pneumonia who received Actemra plus standard of care were 44% less likely to progress to mechanical ventilation or death compared to patients who received placebo plus standard of care.  However, there was no statistical difference in 28-day mortality between patients who received Actemra (10.4%) versus placebo (8.6%).  

What do you think of the EMPACTA study’s use of proportion of patients progressing to mechanical ventilation as its primary endpoint? 

Dr. Dickson: Obviously in a vacuum we would all agree that lowering rates of need for mechanical ventilation would be a good thing. But as you point out, that’s of highly questionable significance if the point estimate for mortality is actually in the other direction! Imagine this from a patient’s perspective: “The good news is this drug may keep you off the ventilator, the bad news is you’re just as likely – if not more likely – to die in the next month.” That’s hardly a patient-centered outcome. 

In that case, is less frequent use of mechanical ventilation, without any improvement in survival, a relevant clinical endpoint?  It’s also worth clarifying – the decision of if and when to put patients on mechanical ventilation – how much does that vary across providers and hospital centers? 

Dr. Dickson: I’m really skeptical of any endpoint “softer” than, for example, 28-day mortality, especially since we’ve got multiple studies now for IL-6 inhibitors that should be sufficiently powered to answer the mortality question. There are so many problems with using mechanical ventilation as an endpoint. Initiating mechanical ventilation is ultimately a subjective clinical decisionnot a physiologic outcome, that’s highly variable across providers and centers. The appropriate timing of intubation is utterly undeterminedwe’ve seen similar mortality outcomes in centers who intubate “early” and “late,” generally reflecting how long they will trial patients on heated high-flow nasal cannula [a form of noninvasive ventilation]We assume that initiation of mechanical ventilation reflects a link in the causal chain of how COVID-19 kills patients – that is, that it’s a measure of the severity of lung injury  but we know that there’s more than one way to die from COVID-19, including the occasional sudden cardiac death, multiorgan failure with kidney injury, or the patient’s family withdrawing care because of perceived prognosis, etc. In order to get me to change my practice, I really need a “harder” endpoint than initiation of mechanical ventilation. 

Dr. Kulkarni: I would entirely agree with Dr. Dickson’s analysis.  

The one point he may have missed, but implied, is that composite outcomes always make me cautious. There are unfortunately many people dying from COVID-19, so 28-day mortality should still be a valid endpoint.  

Another issue is that multiple secondary infections have been reported in the context of IL-6 inhibitors.  This has been reported both within [1] and outside of the United States [2]. Our center, Washington University, has also reported this observation (albeit in a non-randomized study) in a paper that has been provisionally accepted. 

Right, IL-6 inhibitors suppress the immune system, so secondary infections aren’t too surprising, and definitely a cause for concern. Actemra is FDA-approved to treat rheumatoid arthritis and a couple other autoimmune conditions, and it has a black-box warning. Something like: “Patients treated with Actemra (tocilizumab) are at an increased risk for developing serious infections that may lead to hospitalization or death.”  But despite this risk, Actemra is still being evaluated for hospitalized COVID-19 patients because it may blunt the supposed cytokine storm or hyperinflammation experienced by some patients with severe COVID-19.  Is that the story? 

Dr. Dickson: To my mind, a major problem with IL-6 blockade is that it SEEMS really effective because it so potently blunts inflammatory indices. The patient’s temperature will drop, the C-reactive protein and other inflammatory biomarkers will go down, the patient may even feel better.  Fewer rigors, decreased O2 demand, etc.  And this apparent effect on the “overall clinical picture” and trajectory could absolutely inform clinician decision-making regarding whether or not to intubate them. But we are now smarter about this than we were in March 2020, and it’s clear that “blunting the cytokine storm” is an extremely simplistic approach to treating COVID-19 [3-4]. By my read, none of the IL-6 blockade data to date suggest that it meaningfully alters the ultimate severity of lung injury or 28-day mortality. For this reason, I’m even more skeptical than usual about imperfect “proxy outcomes” like decisions to intubate.  

So, if I’m understanding Dr. Sinha’s papers correctly [3-4], ARDS due to COVID-19 may not be associated with higher systemic inflammation, and in fact, the hyperinflammatory phenotype is less prevalent in COVID-19 ARDS than ARDS from other causes. In Dr. Sinha’s paper [3], IL-6 levels in patients with severe COVID-19 are actually quite low compared to historical data from patients with the hyperinflammatory phenotype of ARDS. 

When the biopharma companies started clinical trials for IL-6 inhibition in COVID-19 several months ago, pulmonologists thought that mortality from severe COVID-19 was a result of cytokine storm, hyperinflammation, elevated IL-6, etc.  Now, we’re realizing this model of severe COVID-19 pathophysiology may not be valid, so I guess it’s not too surprising that Actemra and other IL-6 inhibitors haven’t worked as well as we’d hoped.  

Now, let’s get to the bottom line.  Are you going to use Actemra for COVID-19 patients if the FDA grants Emergency Use Authorization (EUA)? 

Dr. Kulkarni: To be convinced to use Actemra in the setting of an EUA, I would still like to see more convincing data regarding mortality endpoints, partly because of the risk of secondary infection. Personally, at the current time, I would be skeptical to use any immunomodulatory therapy outside of a clinical trial setting in COVID-19, with the possible exception of steroids. I am a strong proponent of stratifying therapy and endotyping, while acknowledging that our current methods for endotyping are imperfect. But I do believe there are signals that indicate which patients are likely to benefit from these targeted therapies. So, I would hold off on using them empirically unless: 

  • They demonstrate a clear mortality benefit, OR 
  • In a post-hoc analysis, a group of patients shows a clear improvement in outcomes. We can then use this knowledge to decide who is most likely to benefit from the drug. 

Are there any patient populations that you may or may not give Actemra to? 

Dr. Dickson: I suspect the EMPACTA study may have excluded DNI patients, or patients with an advance directive of “Do Not Intubate”, although I can’t confirm this from their webpage.  But, since they looked at mechanical ventilation as the primary endpoint, including DNI patients would be a source of bias.  At University of Michigan, we’ve had plenty of DNI patients supported with high levels of HHFNC who never needed intubation Obviously, there’s no evidence in this study to suggest Actemra might help DNI patients. 

That’s all super interesting, Dr. Dickson and Dr. Kulkarni.  Of course, we’ll wait for the peer-reviewed publication on the EMPACTA study, and perhaps revisit the topic of IL-6 inhibitors at that time, but for now, this is all very interesting food for thought.  


[1] Kimmig, Lucas M., et al. “IL6 inhibition in critically ill COVID-19 patients is associated with increased secondary infections.” medRxiv (2020). 

[2] Busani, Stefano, et al. “Two fatal cases of acute liver failure due to HSV-1 infection in COVID-19 patients following immunomodulatory therapies.” Clinical Infectious Diseases(2020). 

[3] Sinha, Pratik, Michael A. Matthay, and Carolyn S. Calfee. “Is a “cytokine storm” relevant to COVID-19?.” JAMA internal medicine (2020). 

[4] Sinha, Pratik, et al. “Prevalence of phenotypes of acute respiratory distress syndrome in critically ill patients with COVID-19: a prospective observational study.” The Lancet Respiratory Medicine (2020). 

© 2022 inDemic Foundation. All Rights Reserved.