The association of interleukin-6 value, interleukin inhibitors, and outcomes of patients with COVID-19 in New York City.

Since cytokine release syndrome with elevation of interleukin-6 (IL-6) is considered to be associated with severe cases of coronavirus disease 2019 (COVID-19); IL-6 inhibitors, such as tocilizumab, are expected to be effective for its treatment. This was a retrospective study using a consecutive cohort of 224 patients hospitalized with COVID-19 in March 2020. Patients were divided into those admitted to the intensive care unit (ICU group) and those not (no ICU group), and clinical data including usage of tocilizumab were compared. Correlation between IL-6 value at admission and at peak, and tocilizumab use, as well as clinical outcomes were also investigated. The ICU group had higher rates of pre-existing comorbidities such as hypertension, diabetes, and coronary disease, and higher IL-6 than no ICU group (all P < .05). Age, peak IL-6, and peak d-dimer were significant predictors of in-hospital mortality (1.05 [1.01-1.09], P = .012; 1.001 [1.000-1.002], P = .002; 1.10 [1.03-1.18], P = .008). Receiver operating characteristics curve showed higher predictability of in-hospital mortality with IL-6 at peak than others (area under curve; IL-6 at peak: 0.875 [0.87-0.942], IL-6 at admission: 0.794 [0.699-0.889], d-dimer at peak 0.787 [0.690-0.883], d-dimer at admission 0.726 [0.625-0.827]). Incidence of fungal infections was significantly higher in patients who were given tocilizumab than those who were not (13.0% vs 1.1%, P < .001). Notably, tocilizumab did not affect in-hospital mortality after adjustment including IL-6 (odds ratio [95% confidential interval]: 1.00 [0.27-3.72, P = .998]). Age, peak IL-6, and peak d-dimer levels were significant predictors of in-hospital mortality. Tocilizumab did not decrease in-hospital mortality in our cohort.