Genetic variants mimicking therapeutic inhibition of IL-6 receptor signaling and risk of COVID-19

Few effective therapeutic options are available for the treatment of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection. IL-6 receptor blockade has been proposed as one potential therapeutic strategy, and more than 40 clinical trials of anti-IL-6 receptor antibodies (including tocilizumab and sarilumab) in the setting of SARS-CoV-2 infection are underway (appendix p 2). Early evidence from observational studies and open-label, uncontrolled trials has suggested that IL-6 receptor blockers might confer benefit, particularly in patients with severe COVID-19.¹

Human genetics enables the investigation of potential opportunities for drug repurposing. We leveraged large-scale human genetic data to investigate whether IL-6 receptor blockade might confer therapeutic benefit in COVID-19. A genetic instrument consisting of seven genetic variants in or close to IL6R (pairwise r ²≤0·1; appendix pp 8–9) was recently shown to be associated with altered concentrations of C-reactive protein, fibrinogen, circulating IL-6, and soluble IL-6 receptor, concordant with known effects of pharmacological IL-6 receptor blockade (appendix p 4).² We investigated the effect of these IL6R variants on risk of hospitalisation for COVID-19 and other SARS-CoV-2-related outcomes using data from the COVID-19 Host Genetics Initiative (appendix pp 2–3, 10).³

Based on data from UK Biobank, the IL6R variants were confirmed to be associated with lower serum C-reactive protein concentrations (appendix pp 5–6). Meta-analysis of scaled estimates identified a lower risk of rheumatoid arthritis (odds ratio 0·93 per 0·1 SD lower C-reactive protein; 95% CI 0·90–0·96, p<0·0001; appendix pp 5–6), supporting this established indication for IL-6 receptor inhibitors (eg, tocilizumab and sarilumab), as well as a lower risk of coronary heart disease (0·96; 0·95–0·98, p<0·0001), which has previously been linked to genetic variation in IL6R.⁴ The IL6R instrument was also associated with a lower risk of hospitalisation for COVID-19 (0·88; 0·78–0·99, p=0·03; appendix pp 5–6). We found a consistent association when using a population-based control group (ie, all non-cases; 0·91; 0·87–0·96, p=0·0005; appendix pp 5–6).

On evaluation of further SARS-CoV-2-related outcomes, we detected an association of the IL6R instrument with risk of SARS-CoV-2 infection (0·92; 0·89–0·95, p<0·0001 using a population-based control group; appendix pp 5–6), but no evidence of association with death or need for respiratory support. The results of the main analysis were robust to various sensitivity analyses (appendix p 7).

Our findings show that IL6R variants mimicking therapeutic inhibition of IL-6 receptor are associated with a lower risk of hospitalisation for COVID-19, a phenotype that is associated with disease severity (eg, requiring supplemental oxygen is a typical reason for hospitalisation). This result suggests that pharmacological blockade of IL-6 receptor might be expected to lead to reduced COVID-19 severity. We also found an association with lower risk of SARS-CoV-2 infection; although this finding might suggest that IL-6 receptor blockade lowers susceptibility to SARS-CoV-2 infection, these phenotypes could be biased by symptom severity (ie, individuals with more severe symptoms might be more likely to present for testing, to be offered testing, or to have a positive test). The lack of association with very severe COVID-19 requiring respiratory support or leading to death might be relevant in the context of the announced failure of sarilumab in a phase 3 randomised controlled trial in patients with COVID-19 requiring mechanical ventilation.⁵ However, the genetic analysis with very severe COVID-19 was based on relatively few cases (n=536), limiting robust inference, and the EMPACTA randomised controlled trial reported a lower risk of progressing to mechanical ventilation or death in patients with COVID-19 pneumonia who received tocilizumab compared with patients who received placebo (hazard ratio 0·56; 95% CI 0·32–0·97, p=0·035).⁶

Our results serve as genetic evidence for the potential efficacy of IL-6 receptor blockade in COVID-19 and support the study of IL-6 receptor inhibitors in randomised controlled trials. On their own, these data should not be used to influence clinical care. Ongoing, large-scale randomised controlled trials of IL-6 receptor inhibitors will be instrumental in identifying the potential settings, including stage of disease and potential pharmacogenetic subgroups, in which these agents might be effective.