The course of action for effective anti-cytokine treatment in COVID-19

Even in the aftermath of global anti-SARS-CoV-2 vaccination campaigns, safe and effective treatments to inhibit inflammation and reduce mortality continue to be needed for the substantial proportion of unvaccinated individuals at risk of developing severe COVID-19. The COV-AID trial, the results of which were reported by Jozefien Declercq and colleagues in The Lancet Respiratory Medicine, was a factorial, randomised controlled trial investigating interleukin (IL)-1 blockade (anakinra) and IL-6 blockade (tocilizumab or siltuximab) in patients with COVID-19, respiratory failure, and cytokine release syndrome.¹ This trial was done between April 4 and Dec 6, 2020 in 16 Belgian hospitals and enrolled 342 patients. The primary outcome was time to clinical improvement (increase in 2 points from baseline status on a 6-point ordinal scale). Patients were randomly assigned in a 1:2 ratio to anakinra or standard-of-care treatment, and then 1:1:1 to siltuximab, tocilizumab, or standard of care. The trial showed near-identical times to clinical improvements across groups, indicating marginal or no added benefit for cytokine blockade in this setting.

The use of a factorial design adds complexity to the interpretation of this data. The advantages of this design are the possibility to test multiple treatments simultaneously, while minimising the number of patients exclusively receiving standard of care. However, the dual randomisation strategy resulted in four treatment groups (no cytokine inhibitors; IL-1 inhibition, IL-6 inhibition; IL-1 and IL-6 inhibition combined) with disparities in allocation: notably, only 34 patients received anakinra alone, compared with 129 patients receiving IL-6 inhibitors alone. In addition, the power of factorial designs is influenced by potential interactions between treatments: it is increased by synergistic interaction and decreased by detrimental interaction.² The authors assumed a priori no interaction between IL-1 and IL-6 blockade; however, whether there is no interaction is debatable from a biologic standpoint as IL-1 is found upstream of IL-6 in inflammatory cascades: inhibiting IL-1 results in IL-6 inhibition, hence the two treatments are partially redundant.³ In addition, co-administration of anticytokine treatments is typically avoided on the basis of safety concerns. It is interesting to note that co-administration of IL-1 and IL-6 inhibitors was not associated with increased adverse events in COV-AID. However, it is possible that a safety signal did not emerge owing to the relatively small sample size, and the rationale for co-administration of IL-1 and IL-6 inhibitors remains a questionable one.

The results of the COV-AID trial are partially at odds with available evidence. Previously, only observational studies had compared IL-1 and IL-6 inhibition in COVID-19 and found that anakinra was more effective.⁴ Two controlled trials evaluated anakinra. The CORIMUNO trial enrolled patients with relatively mild disease and was prematurely interrupted because of assumed futility.5, 6 Conversely, the large SAVE-MORE trial enrolled patients on the basis of biomarker profiling: using soluble urokinase plasminogen activator receptor (suPAR) as a proxy for IL-1 bioactivity, the investigators selected patients at high risk for clinical deterioration and observed dramatic reductions in mortality in patients receiving anakinra in addition to the standard of care (including dexamethasone).⁷ In contrast, IL-6 inhibition received much more attention: 27 trials were completed and included in a meta-analysis, which suggested an overall benefit of IL-6 blockade, probably driven by the large RECOVERY and REMAP-CAP trials.⁸ However, a clearly notable factor differentiating these trials from the COV-AID trial is the 28-day mortality in the standard-of-care group, which exceeded 25% compared with only 10% in the COV-AID trial.⁸

This low mortality in the COV-AID trial reflects historical trends, which make it ever more difficult to reveal incremental benefits in patients with COVID-19 treated with biologic agents in randomised trials. Of note, the COV-AID study was designed to reveal a large treatment effect; it was not designed to show moderate treatment effects. It is worth remembering that the moderate clinical effectiveness of dexamethasone, which contributed to saving innumerable lives, required a massive nationwide investigation to be revealed.⁹ Thereby, writing off potential therapeutic options (ie, cytokine inhibitors) on the basis of the negative results of relatively small trials only powered to detect large clinical effects is not advisable.

Two important messages can be taken from this investigation and instruct further use of cytokine inhibitors in COVID-19. First, this study adds to the zeitgeist that universal treatment of COVID-19 with cytokine inhibitors is neither a suitable nor a sustainable option. Consequently, patient selection is crucial, and the severity of the individual's inflammatory response is the main criterion for eligibility to treatment. Indeed, a patient-level meta-analysis indicated that the benefit of anakinra treatment in COVID-19 is maximum for patients with serum C-reactive protein concentrations greater than 100 mg/L,¹⁰ whereas the SAVE-MORE trial found that early identification of candidate patients with suPAR followed by treatment with anakinra resulted in a 55% relative decrease in mortality, which reached 80% for patients with cytokine storm. Screening of individuals on the basis of molecular taxonomy of inflammatory responses is not always practical. Still, it is a necessary and feasible leap to transition from blanket therapeutic attempts, which are bound to yield mixed or disappointing results, to refined and appropriate management strategies.

© 2021 Tim Vernon/Science Photo Library

GC and LD have received consulting fees and honoraria from manufacturers of IL-1 and IL-6 inhibitors Swedish Orphan Biovitrum, Novartis, Sanofi, Roche, outside of the submitted work.