Anakinra treatment in critically ill COVID-19 patients: a prospective cohort study

Abstract

Background: A subset of critically ill COVID-19 patients develop a hyperinflammatory state. Anakinra, a recombinant interleukin-1 receptor antagonist, is known to be effective in several hyperinflammatory diseases. We investigated the effects of anakinra on inflammatory parameters and clinical outcomes in critically ill, mechanically ventilated COVID-19 patients with clinical features of hyperinflammation.

Methods: In this prospective cohort study, 21 critically ill COVID-19 patients treated with anakinra were compared to a group of standard care. Serial data of clinical inflammatory parameters and concentrations of multiple circulating cytokines were determined and aligned on start day of anakinra in the treatment group, and median start day of anakinra in the control group. Analysis was performed for day - 10 to + 10 relative to alignment day. Clinical outcomes were analyzed during 28 days. Additionally, three sensitivity analyses were performed: (1) using propensity score-matched groups, (2) selecting patients who did not receive corticosteroids, and (3) using a subset of the control group aimed to match the criteria (fever, elevated ferritin) for starting anakinra treatment.

Results: Baseline patient characteristics and clinical parameters on ICU admission were similar between groups. As a consequence of bias by indication, plasma levels of aspartate aminotransferase (ASAT) (p = 0.0002), ferritin (p = 0.009), and temperature (p = 0.001) were significantly higher in the anakinra group on alignment day. Following treatment, no relevant differences in kinetics of circulating cytokines were observed between both groups. Decreases of clinical parameters, including temperature (p = 0.03), white blood cell counts (p = 0.02), and plasma levels of ferritin (p = 0.003), procalcitonin (p = 0.001), creatinine (p = 0.01), and bilirubin (p = 0.007), were more pronounced in the anakinra group. No differences in duration of mechanical ventilation or ICU length of stay were observed between groups. Sensitivity analyses confirmed these results.

Conclusions: Anakinra is effective in reducing clinical signs of hyperinflammation in critically ill COVID-19 patients. A randomized controlled trial is warranted to draw conclusion about the effects of anakinra on clinical outcomes.

Keywords: Coronavirus disease 2019; Critical care; Cytokines; Immunity; Interleukin-1 receptor antagonist protein; SARS-CoV-2.

Fig. 1

Circulating cytokine concentrations. Concentrations of circulating a tumor necrosis factor (TNF)-α, b interleukin (IL)-6, c IL-8, d IL-10, e interferon gamma-induced protein (IP)-10, f monocyte chemoattractant protein (MCP)-1, and g IL-1 receptor antagonist (IL-1RA) on day of intensive care unit (ICU) admission and serial data within six days pre- and post-alignment day (day 0). Data are presented as geometric mean with 95% confidence intervals and were analyzed using mixed-models analysis (time*group interaction factor) to evaluate differences between groups over time. p values under graph titles reflect overall between-group differences (day − 6 until day 6). Between-group p values for day − 6 until day 0 and day 0 until day 6 are shown on the left and right of each panel, respectively

Fig. 2

Proteomics inflammation panel and kinetics of inflammatory parameters. a Volcano plot of proteomics inflammation panel for 75 proteins in both groups. Mean fold change and − ¹⁰log(p value) were shown on the x-axis and y-axis, respectively. p values were calculated using t tests. A mean fold change of ≤ -2 or ≥ 2 was considered relevant. A − ¹⁰log(p value) > 3.204 was considered significant (p < 0.000625). b Body temperature and plasma levels of c ferritin, d procalcitonin, e white blood cell counts, and f C-reactive protein (CRP) over time within 10 days pre- and post-start anakinra alignment day (day 0). Data are presented as geometric mean with 95% confidence intervals and were analyzed using mixed-models analysis (time*group interaction factor) to evaluate differences between groups over time. p values under graph titles reflect overall between-group differences (day − 10 until day 10). Between-group p values for day − 10 until day 0 and day 0 until day 10 are shown on the left and right of each panel, respectively

Fig. 3

Individual parameters of sequential organ failure assessment (SOFA) score and total SOFA score. Plasma levels of a creatinine, b bilirubin, and c thrombocytes and d PaO₂/FiO₂ (P/F) ratio, e infusion rate of norepinephrine, and f SOFA score over time within 10 days pre- and post-alignment day (day 0). PaO₂/FiO₂ ratio and SOFA score were presented until day 6. Data are presented as geometric mean with 95% confidence intervals and were analyzed using mixed-models analysis (time*group interaction factor) to evaluate differences between groups over time. p values under graph titles reflect overall between-group differences (day − 10 until day 6 or 10). Between-group p values for day − 10 until day 0 and day 0 until day 6 or 10 are shown on the left and right of each panel, respectively

Fig. 4

Clinical outcomes. Kaplan–Meier graphs of a time on mechanical ventilator, b length of stay in the intensive care unit (ICU), and c mortality. Data are presented for the first 28 days after anakinra alignment day. Patients who were no longer mechanically ventilated on alignment day were not included in time on ventilator graph. p values were calculated using log-rank tests. Numbers at risk on each timepoint per group are shown below graphs