Efficacy matters: broadening complement inhibition in COVID-19.

We and others have proposed the use of anti-complement agents for the treatment of COVID-19; thus, we read with great interest the Article by Alexander P J Vlaar and colleagues reporting the results of an exploratory, randomised phase 2 trial of IFX-1, an anti-human C5a monoclonal antibody, in patients with severe COVID-19. Here, we discuss plausible explanations for IFX-1's inefficacy in this study.

One major concern is the choice of the primary endpoint, the percentage change in PaO2/FiO2 from baseline to day 5, which was assessed well before the anticipated pharmacodynamic window of IFX-1. In addition, as acknowledged by the authors, the trial was not powered to show statistically significant differences in clinical endpoints, eventually jeopardising any conclusion, even on secondary endpoints. The possible biological efficacy of IFX-1 was not adequately investigated by extensive assessment of key inflammatory markers (eg, C-reactive protein) related to the effect of C5a blockade on hyperinflammation. In fact, upstream complement inhibition at the C3 or C5 level leads to a rapid decline in the concentration of serum inflammatory markers in patients with COVID-19.3, 4

Monitoring plasma C5a concentrations would enable a reliable assessment of the drug's effective therapeutic concentration. Inclusion of pharmacokinetic and pharmacodynamic measurements would have also been informative (eg, the ability of IFX-1-treated plasma to block C5aR1-dependent responses in appropriate assays), helping to resolve issues related to drug plasma residence, target saturation, dosing, and efficacy.

We believe that the selection of a complement target with a narrow therapeutic scope, such as C5a, is contradictory to mounting evidence indicating that COVID-19 thromboinflammation is fuelled by multiple elements of the complement cascade that remain operative during anti-C5a treatment (eg, C3, C3a–C3aR1, and C5b-9).4, 5 For instance, C3 inhibition offers broader control of thromboinflammation driven by neutrophil extracellular traps in patients with COVID-19 than does C5 inhibition, partly explaining the small impact of IFX-1 on coagulation and indicating that D-dimer analysis might not be a uniformly predictive or reliable marker of coagulation in patients with COVID-19.

Considering that high neutrophil numbers are associated with poor prognosis in COVID-19, the projected non-interference of IFX-1 on neutrophil counts might signify that anti-C5a treatment is not the optimal way to treat COVID-19-associated neutrophilia. In fact, blockade of other complement components, acting upstream of C5a, might be a more robust and favourable clinical approach (eg, blockade of C3-mediated signalling with therapeutics like AMY-101). Thus, even if apparently disappointing, the results of this trial indicate that broader, rather than narrower, complement inhibition might be more beneficial for the treatment of COVID-19.