COVID-19: Clean up on IL-6

To the Editor:

The emergence of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) represents the latest threat to global health security, and the pressure to identify effective therapeutics during this pandemic is immense. This stress has led to the use of unproven therapies with greater than minimal risk. One example is the use of IL-6 receptor antagonists. After an early report of a “cytokine storm” in patients with coronavirus disease (COVID-19), there is increased interest in anti–IL-6 therapy as a treatment option, with ill-defined criteria for use (1).

The prevailing theory is that SARS-CoV-2 induces the production of cytokines, in particular IL-6, and that these cytokines are a key driver of both lung damage and mortality. However, it is crucial to recognize the uncertainty surrounding the role of IL-6 in viral infections. In experimental viral infection models, IL-6 is a pleiotropic cytokine with complex interactions with multiple signaling cascades, at times producing either proinflammatory or antiinflammatory effects (2). As such, it remains unknown whether elevated IL-6 in viral infections represents a therapeutic target or part of a functioning adaptive immune response.

There has been a long-standing interest in cytokine levels in diverse diseases, including acute respiratory disease syndrome (ARDS) (3). Comparing COVID-19 with our historical knowledge of IL-6 concentrations in other disease states and healthy individuals may be useful. From a study of healthy adult volunteers, IL-6 levels ranged from 1 to 79 pg/ml, with 95% of patients having an IL-6 concentration of less than 42 pg/ml (4). An analysis of patients with ARDS from the SAILS (Statins for Acutely Injured Lungs from Sepsis) study parsed patients into the following two subphenotypes: hyperinflammatory and hypoinflammatory. The hyperinflammatory subphenotype had a median IL-6 concentration of 1,618 pg/ml, whereas the hypoinflammatory group had a median IL-6 concentration of 282 pg/ml (3). Another analysis of serum IL-6 levels reported mean values of 712 pg/ml in patients with ARDS and of 834 pg/ml in patients with severe bacterial pneumonia (5). A COVID-19 cohort in Shanghai had peak elevated IL-6 concentrations ranging from 100 to 200 pg/ml at 20–30 days after symptom onset (6). For comparison, patients who have received chimeric antigen receptor modified T-cell therapy (CAR-T) and developed cytokine release syndrome (CRS) had median IL-6 levels of close to 1,000 pg/ml for grade 4 CRS and above 100 pg/ml for milder grades (7). In light of these data, it appears that patients with SARS-CoV-2 infection have elevated IL-6 levels; however, these are markedly lower than those seen in other severe respiratory diseases or those seen in CAR-T patients with CRS. Although use of blood biomarker levels to direct therapeutic choice seems tempting, the different biology of CAR-T CRS, influenza, and SARS-CoV-2 suggests that a “one cutoff value fits all” approach is inappropriate. In addition, we speculate that the optimal timing of blockade and prolonged IL-6 inhibition is likely disease and IL-6 concentration specific and should be carefully considered for a cytokine with proinflammatory and antiinflammatory effects.

The uncertainties about the use of anti–IL-6 therapy in COVID-19 do not end there. Patients with severe COVID-19 who were administered anti–IL-6 treatments have been reported to have increased secondary infections (8, 9). The side effects of immunomodulatory agents should give pause to off-label use, and they should ideally be administered only as part of a clinical trial. Further data are necessary to determine whether IL-6 is a maladaptive or an adaptive inflammatory factor in COVID-19 pathogenesis. After our initial submission of this letter, Sinha and colleagues published an editorial that the evidence for cytokine storm in COVID-19 pathogenesis is insufficient (3)—we agree and also conclude that the early focus on IL-6 may have cluttered the aisle in our understanding of COVID-19 ARDS pathogenesis. Although a subgroup of patients may benefit from IL-6 blockade, the current data argue against its routine use until more is known. Fortunately, several randomized control trials are underway to determine the efficacy of anti–IL-6 therapy (10).

This pandemic has challenged many of the tenets of good medical practice and intensive care—to first do no harm, to provide safe and evidence-based therapies, and in the absence of other treatments, to provide the best supportive care possible. The threshold for treatment with an unproven therapy should not change based on the novelty of disease and any potential therapies rigorously studied. Otherwise, we will lose the opportunity to learn the most about the disease we are trying so desperately to defeat.