Compassionate use of JAK1/2 inhibitor ruxolitinib for severe COVID-19: a prospective observational study

Abstract

Overwhelming inflammatory reactions contribute to respiratory distress in patients with COVID-19. Ruxolitinib is a JAK1/JAK2 inhibitor with potent anti-inflammatory properties. We report on a prospective, observational study in 34 patients with COVID-19 who received ruxolitinib on a compassionate-use protocol. Patients had severe pulmonary disease defined by pulmonary infiltrates on imaging and an oxygen saturation ≤ 93% in air and/or PaO2/FiO2 ratio ≤ 300 mmHg. Median age was 80.5 years, and 85.3% had ≥ 2 comorbidities. Median exposure time to ruxolitinib was 13 days, median dose intensity was 20 mg/day. Overall survival by day 28 was 94.1%. Cumulative incidence of clinical improvement of ≥2 points in the ordinal scale was 82.4% (95% confidence interval, 71-93). Clinical improvement was not affected by low-flow versus high-flow oxygen support but was less frequent in patients with PaO2/FiO2 < 200 mmHg. The most frequent adverse events were anemia, urinary tract infections, and thrombocytopenia. Improvement of inflammatory cytokine profile and activated lymphocyte subsets was observed at day 14. In this prospective cohort of aged and high-risk comorbidity patients with severe COVID-19, compassionate-use ruxolitinib was safe and was associated with improvement of pulmonary function and discharge home in 85.3%. Controlled clinical trials are necessary to establish efficacy of ruxolitinib in COVID-19.

Fig. 1. Changes in the category of the ordinal scale in individual patients, and in the full cohort of patients.

Each patient is represented as a colored line, where each color indicates the category of the ordinal scale to which the patient belongs, from baseline (day 0, day of first dose of ruxolitinib) to day 28. The vertical bars indicate the last day of treatment with full dose of Ruxolitinib. A solid diamond indicates that the patient died. Patients were monitored daily while hospitalized, and reached by telephone calls every 2–3 days after being discharged. The day of discharge is indicated by an open diamond (a). The cumulative distribution of patients in the individual categories of the ordinal scale, at weekly intervals, is shown in (b).

Fig. 2. Cumulative incidence of clinical improvement from baseline to day 28.

The data are shown for the full cohort of patients (a), for patients in the full cohort stratified according to the ordinal scale category at baseline (b), and for patients in the full cohort stratified according to the arterial oxygen partial pressure (PaO2)/fraction of inspired oxygen (FiO2) (P/F ratio) at baseline (c).

Fig. 3. Changes in peripheral blood mononuclear cell subsets and serum cytokine levels at day 14 in COVID-19 patients compared to levels at baseline and normal subjects.

The absolute count of peripheral blood cell subsets, analyzed by flow cytometry, was measured at baseline (T0, black columns) and at day 14 (T14, dark gray columns) since initiation of ruxolitinib. Columns represent mean value (±SD) of neutrophils, lymphocytes, monocytes, basophils, eosinophils, plasmacytoid, and myeloid dendritic cells (DC). Data were obtained from 16 COVID-19 patients receiving ruxolitinib, and healthy donors (n = 8) (a, b). The activation markers CD64, CD13, and CD11b (on monocytes), and CD66b (on granulocytes), were analyzed by flow cytometry in the same set of samples; results are expressed as the mean value (±SD) of mean fluorescence intensity (MFI) (c). The frequency of Ki67-positive cells, expressed as the Mean (±SD), was obtained from analysis of isolated peripheral blood mononuclear cells of COVID-19 patients (n = 13), collected at T0 and T14, and healthy donors (n = 6) as control (d). The frequency of IFN-gamma of TNF-alpha positive cells, obtained from analysis of isolated peripheral blood mononuclear cells after in vitro polyclonal stimulation, is expressed as mean (+SD). Data refer to 14 COVID-19 patients, and 12 healthy donors, as control (e, f). g Heatmap of serum concentration (pg/ml) of the indicated cytokines and chemokines in healthy controls (n = 4) and COVID-19 patients (n = 16), who were evaluated at baseline (T0) and at day 14 (T14) since initiation of ruxolitinib. Only two patients, indicated by an asterisk, were receiving corticosteroids concurrently with ruxolitinib, in the first 7 days of treatment. The color scale ranges from blue (lower concentration) to red (higher concentration) for each analyte. *p < 0.05, **p < 0.001, ***p < 0.001, as indicated by the bars.

Fig. 4. Changes in C-reactive protein, D-dimer, and ferritin levels at days 7 and 14 in COVID-19 patients, compared to levels at baseline.

The plasma levels of C-reactive protein and D-dimer, and serum levels of ferritin, were measured at baseline and at days 7 and 14 since initiation of ruxolitinib. Individual values are presented as well as the mean value ± SD. Statistically significant differences are shown on top.