Critical Care

Efficacy and safety of an inhaled pan-Janus kinase inhibitor, nezulcitinib, in hospitalised patients with COVID-19: results from a phase 2 clinical trial

Abstract

Background The inhaled lung-selective pan-Janus kinase inhibitor nezulcitinib had favourable safety and potential efficacy signals in part 1 of a phase 2 trial in patients with severe COVID-19, supporting progression to part 2.

Methods Part 2 was a randomised, double-blind phase 2 study (NCT04402866). Hospitalised patients aged 18–80 years with confirmed symptomatic COVID-19 requiring supplemental oxygen (excluding baseline invasive mechanical ventilation) were randomised 1:1 to nebulised nezulcitinib 3 mg or placebo for up to 7 days with background standard-of-care therapy (including corticosteroids). Efficacy endpoints included respiratory failure-free (RFF) days through day 28 as the primary endpoint. Secondary endpoints included safety and change from baseline oxygen saturation (SaO2)/fraction of inspired oxygen (FiO2) ratio on day 7, and 28-day mortality rate was a prespecified exploratory endpoint.

Results Between June 2020 and April 2021, 205 patients were treated (nezulcitinib, 103; placebo, 102). There was no statistically significant difference between nezulcitinib versus placebo in the primary endpoint (RFF days; median, 21.0 vs 21.0; p=0.6137) or secondary efficacy endpoints. Nezulcitinib was generally well tolerated with a favourable safety profile.

Conclusions Although the prespecified primary, secondary and exploratory efficacy endpoints, including RFF through day 28, change from baseline SaO2/FiO2 ratio on day 7, and 28-day mortality rate, were not met, nezulcitinib was generally well tolerated and had a favourable safety profile. Further studies are required to determine if treatment with nezulcitinib confers clinical benefit in specific inflammatory biomarker-defined populations of patients with COVID-19.

What is already known on this topic

  • Anti-inflammatory treatment can be beneficial in patients with severe COVID-19. Nezulcitinib is an inhaled lung-selective pan-Janus kinase (JAK) inhibitor that showed favourable safety and potential efficacy signals in part 1 of a phase 2 trial in patients with severe COVID-19.

What this study adds

  • In part 2 of this phase 2 trial, treatment with once-daily nebulised nezulcitinib 3 mg versus placebo for up to 7 days did not significantly improve the number of respiratory failure-free days through day 28, change from baseline oxygen saturation/fraction of inspired oxygen ratio on day 7, or 28-day mortality rate in the overall study population; however, in post hoc analyses, nezulcitinib treatment improved time to recovery, 28-day mortality, and ventilator-free survival among the subgroup of patients with baseline C reactive protein <150 mg/L. Safety was similar between treatment arms.

How this study might affect research, practice or policy

  • The results of this study suggest that lung-targeted JAK inhibition with nezulcitinib may be beneficial in populations of patients with severe COVID-19 defined by specific inflammatory biomarkers.

Introduction

Severe COVID-19 is characterised by a systemic and pulmonary hyperinflammation response, resulting in release of a cascade of proinflammatory cytokines and chemokines.1 This ‘cytokine storm’ produces acute lung injury, which can progress to acute respiratory distress syndrome or death.2 3

Optimal COVID-19 treatment depends on disease severity and progression. Dexamethasone treatment decreases mortality in patients with COVID-19 receiving respiratory support and is considered standard of care (SOC) for patients with severe COVID-19.4 5 The Janus kinase (JAK) family of enzymes transduce inflammatory signals,6 and orally administered JAK inhibitors were investigated in patients with severe COVID-19.7 8 The JAK 1/2 inhibitor baricitinib improved clinical outcomes—including decreased 28-day all-cause mortality, decreased recovery time and reduced progression to mechanical ventilation—when added to the antiviral remdesivir or SOC therapy.7 9 10 Addition of tofacitinib, a JAK 1/3 inhibitor, to SOC treatment reduced the incidence of death or respiratory failure through day 28 relative to placebo.8

Unlike orally administered subtype-specific JAK inhibitors, the inhaled lung-selective pan-JAK inhibitor nezulcitinib was designed to produce high sustained lung exposure to maximise effects in pulmonary tissue while maintaining low plasma levels to minimise potential for systemic adverse events (AEs). Because pulmonary inflammation and consequent diffuse alveolar damage drive COVID-19 morbidity and mortality,1 lung-targeted anti-inflammatory treatment could improve clinical outcomes and reduce mortality. Pan-JAK inhibition (all 4 JAK subtypes) could also provide greater anti-inflammatory effects compared with subtype-specific JAK 1/2 or JAK 1/3 inhibition11 12 and may complement dexamethasone in combination therapy due to their different mechanisms of action.13

We recently reported that nezulcitinib was generally well tolerated and showed potential for efficacy in the part 1 ascending-dose phase of a phase 2 trial in patients with severe COVID-19.14 Based on an assessment of both clinical response at day 7 and available safety from this sentinel part 1 cohort (n=25), as well as phase 1 safety outcomes, nezulcitinib 3 mg was selected for further evaluation in part 2, a larger, phase 2 double-blind, parallel-group, placebo-controlled 28-day outcome study in patients with confirmed COVID-19 hospitalised for symptomatic respiratory insufficiency.

Methods

Study design and patients

This was a randomised, double-blind, parallel-group, multicentre, multinational study. Patients or their legal representatives provided informed consent; in the UK, as a protocol-specified secondary option, proxy consent could be provided by a clinician and a second health professional. Due to the emergent nature of the pandemic and patients’ health status at enrolment, patients and the public were not involved in the study design, recruitment or conduct and will not be involved in disseminating the results; study participants were not asked to assess the burden of the intervention and time required to participate. Part 1 of the study started on 24 June 2020, and part 2 concluded on 21 April 2021.

Patients 18–80 years of age with confirmed symptomatic COVID-19 who were hospitalised or planned to be hospitalised and required supplemental oxygen to maintain saturation >90% (clinical status 5–6 on the National Institute of Allergy and Infectious Diseases 8-point ordinal scale (OS) (online supplemental table 1) were eligible.15 COVID-19 was confirmed by a positive test for SARS-CoV-2 RNA detected by real-time PCR on a swab collected from the upper respiratory tract. COVID-19 symptoms for >2 and ≤14 days before hospital admission were required. Enrolment was planned for approximately 20% of participants to be OS status 6 at baseline. Patients receiving invasive mechanical ventilation at screening and those unlikely to survive for >24 hours in the investigator’s opinion were excluded. Other key exclusion criteria were evidence of serious active infections other than COVID-19, septic shock and body mass index ≥40 kg/m2. Treatment with anti-interleukin (IL)-6, anti-IL-1 or anti-T-cell antibodies; IL-6 receptor antagonists; supplemental interferon therapy, tyrosine kinase inhibitors, or JAK inhibitors within the past 30 days; or planning to receive a JAK inhibitor during the study was not permitted. Full exclusion criteria are listed in online supplemental methods.

Patient and public involvement

None.

Randomisation and masking

Patients were randomised 1:1 to receive nezulcitinib 3 mg or placebo. The randomisation schedule and treatment assignment were via the Randomisation Trial Supply Management System (Suvoda; Conshohocken, Pennsylvania, USA). Randomisation was stratified by baseline age (≤60 vs >60 years) and concurrent use of antiviral medications (eg, remdesivir) at baseline. Patients, study investigators and staff, and sponsor personnel involved in the conduct of the study were blinded to treatment assignment until the end of the study. Nezulcitinib and matching placebo were clear solutions provided in amber glass vials.

Procedures and outcomes

Nezulcitinib 3 mg (single 6 mg loading dose on day 1) or placebo was administered via inhalation using the Aerogen Solo nebuliser system (Galway, Ireland) once daily for up to 7 days or until discharge from hospital, whichever was earlier, with follow-up through day 28. Oxygen saturation (SaO2) was collected via pulse oximetry and fraction of inspired oxygen (FiO2) was recorded daily though day 7 and from chart review on days 14, 21 and 28 and/or at hospital discharge. Clinical status on the 8-point OS was recorded daily through day 7 and through day 28 while patients remained hospitalised; for those discharged earlier, clinical status was obtained by telephone on days 14, 21 and 28. Alive and respiratory failure-free (RFF) status on day 28 was reported. Vital signs were recorded predose on days 1–7 and on days 14, 21 and 28 and/or at hospital discharge. Concomitant medications and AEs were also recorded. Blood was collected for haematology and serum chemistry evaluations on days 1 and 7. Additional assessments are described in online supplemental methods.

The primary efficacy outcome was number of RFF days—defined as number of days the patient was alive and did not require invasive mechanical ventilation, noninvasive positive-pressure ventilation, high-flow oxygen devices or oxygen supplementation—from randomisation through day 28. Secondary endpoints were safety (ie, incidence and severity of treatment-emergent AEs (TEAEs)) and tolerability; change from baseline SaO2/FiO2 ratio on day 7; proportion of patients in each category of the 8-point OS on days 7, 14, 21 and 28; and proportion of patients alive and RFF on day 28. Prespecified exploratory efficacy endpoints included time to recovery (TTR) through day 28 and 28-day all-cause mortality rate. Post hoc endpoints included ventilator-free survival (VFS) through day 28.

Statistical analysis

For part 2, 94 patients per treatment arm (188 total) were estimated to provide ≥80% power to detect a difference between patients receiving nezulcitinib versus placebo in the primary endpoint of number of RFF days through day 28 in the intention-to-treat (ITT) population, assuming an equal variance of 72.25 (SD of 8.5 days), placebo-adjusted improvement of 3.5 days, a two-sided type I error rate of 5%, and use of a two-sample t-test assuming equal variance. Enrolment of 198 patients was planned to provide 188 patients for the primary efficacy analysis assuming a dropout rate of 5%.

The primary analysis set for safety data included all patients who received ≥1 dose of study drug analysed according to treatment received. The primary analysis set for efficacy data, the ITT analysis set, included all randomised patients analysed according to their randomised treatment group. Post hoc analyses of exploratory endpoints were performed in patient subgroups defined by baseline C reactive protein (CRP) values <150 mg/L vs ≥150 mg/L, previously identified as a marker for a high-risk inflammatory phenotype16 and supported by analyses using different CRP thresholds (online supplemental figure 1).

Treatment comparison of nezulcitinib versus placebo for the primary efficacy endpoint of number of RFF days through day 28 was evaluated from the proportional odds (PO) ordinal regression model adjusting for baseline age stratum (≤60 vs >60 years). The p value was based on the Van Elteren test adjusted for baseline age stratum; the difference between nezulcitinib-treated and placebo-treated patients was summarised based on median number of days. Change in SaO2/FiO2 ratio at day 7 was compared between patients receiving nezulcitinib and placebo using a mixed model for repeated measures including fixed effects for treatment, study day, treatment group by study day interaction, baseline SaO2/FiO2 ratio, treatment group by baseline SaO2/FiO2 ratio interaction, and baseline age group, and a random effect for each patient. Least squares means, SEs, treatment differences in least squares means and 95% CIs were estimated from the model, and a two-sided nominal p value was reported. Proportion of patients in each clinical status category at days 7, 14, 21 and 28 was evaluated using a PO ordinal regression model adjusting for baseline age group with p value based on the Van Elteren test. Treatment comparisons for binary endpoints (proportion of patients alive and RFF on day 28 and 28-day mortality rate) was analysed using a Cochran-Mantel-Haenszel χ2 test. For time-to-event data, such as TTR and VFS and overall survival, within-group summaries (median and third-quartile time) were analysed using Kaplan-Meier estimates, and treatments were compared using a Cox proportional hazards model with p value based on a log-rank test. Statistical tests of treatment effects were performed at a two-sided significance level of 0.05. No multiplicity adjustment was performed due to the hypothesis-generating nature of this phase 2 study; hence, all secondary and exploratory endpoint p values are presented as nominal. Additional details are included in online supplemental methods.

Demographics and baseline characteristics were summarised using descriptive statistics. AEs were coded according to Medical Dictionary for Regulatory Activities V.24.0 and summarised as frequency and percentage. Analyses were performed using SAS V.9.4 (SAS Institute).

Results

A total of 249 patients at 19 centres in North America, South America and Europe were screened, 210 were enrolled and randomised (ITT analysis set; nezulcitinib 3 mg, 106; placebo, 104), and 205 were treated (safety analysis set; nezulcitinib 3 mg, 103; placebo, 102). Of these, 92 nezulcitinib-treated patients and 89 placebo-treated patients completed follow-up on day 28 (figure 1). The most frequent reason for withdrawal in both treatment arms was AEs.

Figure 1
Figure 1

Phase 2 part 2 trial profile. ITT, intention to treat.

Baseline demographics and clinical characteristics are shown in table 1. Mean age was 58 years in both nezulcitinib-treated and placebo-treated patients. A majority of patients were White (nezulcitinib, 98.1% and placebo, 98.1%) and male (61.3% and 60.6%). A large proportion of patients were enrolled in Europe. Proportions of patients with baseline OS clinical status of 5 vs 6 were similar between treatment groups. The majority of patients used corticosteroids, specifically dexamethasone; remdesivir was used primarily in the USA and in few patients overall (table 1). The largest proportion of patients had ≥2 comorbidities (nezulcitinib, 46.1%; placebo, 46.6%) at study entry, the most common of which were hypertension, coronary artery disease, diabetes mellitus and sleep apnoea.

Table 1
|
Baseline demographics and clinical characteristics in the ITT analysis set*

There was no difference between nezulcitinib and placebo in the primary endpoint of RFF days from randomisation to day 28: the median (IQR) was 21.0 (17.5–23.0) days in nezulcitinib-treated patients and 21.0 (15.0–23.0) days in placebo-treated patients (p=0.6137) (table 2). No statistically significant difference between treatment groups was observed in the secondary efficacy endpoints of change from baseline SaO2/FiO2 ratio on day 7 (table 2, online supplemental figure 2), proportion of patients alive and RFF on day 28 (table 2), or proportion of patients in each category of the 8-point OS scale at days 7, 14, 21 and 28 (online supplemental figure 3).

Table 2
|
Primary and secondary efficacy endpoints in the ITT analysis set

Trends suggesting numerical improvement in patients treated with nezulcitinib 3 mg versus placebo were observed in some prespecified or post hoc exploratory endpoints, although none reached nominal significance (table 3). Median and third-quartile TTR was 1 and 4 days shorter for nezulcitinib-treated patients (10 and 16 days) versus placebo-treated patients (11 and 20 days) (online supplemental figure 4A). The 28-day all-cause mortality rate was 5.7% (6/106) in patients treated with nezulcitinib compared with 12.5% (13/104) in placebo-treated patients (nominal p=0.0844) (table 3, online supplemental figure 5A). Ninety-eight of 106 (92.5%) patients treated with nezulcitinib vs 89/104 (85.6%) placebo-treated patients were alive and ventilator-free at day 28 (nominal p=0.1099) (online supplemental figure 6A).

Table 3
|
Exploratory and post hoc endpoints of interest in the ITT analysis set and post hoc subgroups defined by baseline CRP

Due to these favourable trends and evidence that baseline CRP <150 mg/L may be a positive prognostic factor,16 post hoc subgroup analyses were performed for TTR, 28-day all-cause mortality rate, and VFS in patient subgroups defined by baseline CRP <150 vs ≥150 mg/L (table 3). Baseline clinical characteristics in the baseline CRP <150 mg/L subgroup were balanced across treatment arms (online supplemental table 2), similar to the ITT population. Among patients with baseline CRP <150 mg/L, nezulcitinib-treated patients had shorter median and third-quartile times to recovery (10 and 15 days) vs placebo-treated patients (11 and 20 days; nominal p=0.0204) (online supplemental figure 4B). In patients with baseline CRP ≥150 mg/L, median TTR was 18 days for nezulcitinib-treated patients vs 14 days for placebo-treated patients (nominal p=0.5493) (online supplemental figure 4C). The 28-day all-cause mortality rate was 1.2% (1/86) in nezulcitinib-treated patients vs 10.6% (9/85) in placebo-treated patients with CRP <150 mg/L (nominal p=0.0089; table 3, online supplemental figure 5B). Among those with baseline CRP ≥150 mg/L, 28-day all-cause mortality rates were similar between treatment arms (29.4% in nezulcitinib-treated patients and 30.8% in placebo-treated patients; table 3, online supplemental figure 5C). The percentage of patients alive and ventilator-free through Day 28 was 97.7% (84/86) in nezulcitinib-treated patients vs 87.1% (74/85) in placebo-treated patients with baseline CRP <150 mg/L (nominal p=0.0069) (online supplemental figure 6B) but 64.7% (11/17) in nezulcitinib-treated patients vs 69.2% (9/13) in placebo-treated patients with baseline CRP ≥150 mg/L (nominal p=0.6656) (online supplemental figure 6C). In supporting analyses, the 150 mg/L cut-off point showed robust and statistically significant treatment effect for TTR, 28-day all-cause mortality rate, and VFS compared with other baseline CRP threshold values assessed (online supplemental figure 1).

The most frequently reported TEAE was respiratory failure, in 9 (8.8%) placebo-treated patients vs 3 (2.9%) nezulcitinib-treated patients (table 4). Elevated alanine aminotransferase occurred in six patients in each treatment arm (nezulcitinib, 5.8%; placebo, 5.9%). One nezulcitinib-treated patient experienced elevations in both alanine transferase and aspartate aminotransferase and discontinued treatment according to predefined study drug discontinuation criteria. Frequencies of other discontinuations due to TEAEs were similar between treatment arms. Treatment-emergent serious AEs (SAEs) occurred in 10 (9.7%) patients treated with nezulcitinib and 16 (15.7%) placebo-treated patients. No SAEs were considered related to study drug. The most frequent SAE was respiratory failure, in 3 (2.9%) patients treated with nezulcitinib and 5 (4.9%) placebo-treated patients. There were 19 deaths during this study, including 6 (5.8%) patients treated with nezulcitinib and 13 (12.7%) treated with placebo. The most frequent causes of death were pulmonary embolism, in 4 (3.9%) placebo-treated patients; respiratory failure, in 3 (2.9%) placebo-treated patients; and multiple organ dysfunction syndrome, in 3 (2.9%) nezulcitinib-treated patients and 1 (1.0%) placebo-treated patient. Two placebo-treated patients died of cardiac arrest. Death was attributed to pneumothorax, bacterial sepsis and vascular shock in 1 (1.0%) patient each in each treatment arm; ‘sudden death’ was reported in one placebo-treated patient. No nezulcitinib-treated patients had venous thromboembolism (VTE), versus five placebo-treated patients (four pulmonary embolisms and one deep vein thrombosis). The frequency of TEAEs considered related to study treatment was similar between patients treated with nezulcitinib versus placebo, and no new safety signals were detected.

Table 4
|
Safety through day 28

Discussion

Treatment with nebulised nezulcitinib 3 mg versus placebo did not significantly reduce number of RFF days or improve change from baseline SaO2/FiO2 ratio on day 7, 8-point clinical status, or the proportion of patients alive and RFF on day 28. Nebulised nezulcitinib 3 mg once daily for 7 days was well tolerated. There were fewer SAEs in nezulcitinib-treated versus placebo-treated patients. SAEs were generally consistent with COVID-19 and not associated with nezulcitinib treatment. AEs associated with systemic JAK inhibitors, such as VTE, were not observed in nezulcitinib-treated patients.14

Because the lung is the primary site of COVID-19, targeting the lung with nebulised nezulcitinib was anticipated to result in clinical benefit as seen for systemic JAK inhibitors. One possibility is that systemic JAK inhibition with targeting of circulating immune cells is required for efficacy in patients with severe COVID-19; however, inhaled JAK inhibitors could still be beneficial in patients with less severe disease. This is seen with glucocorticoids, where inhaled delivery may be effective in patients with mild COVID-19 pneumonia, but systemic treatment is beneficial in hospitalised patients.17 18 Consistent with this hypothesis, favourable trends were observed in patients treated with nezulcitinib versus placebo in exploratory endpoints including TTR, 28-day all-cause mortality rate and VFS. These trends reached nominal significance in a post hoc analysis in patients with baseline CRP <150 mg/L, although caution is required in interpreting this result due to the low number of mortality events and post hoc nature of the analysis; still, the results suggest the overall trends favouring nezulcitinib were driven by this population.

JAK inhibition is a viable mechanism of action for treatment of severe COVID-19, as evidenced by the clinical benefit of oral JAK inhibitors in hospitalised patients.7 19 In COV-BARRIER, treatment with baricitinib plus SOC versus placebo reduced all-cause mortality.7 Addition of baricitinib to dexamethasone and tocilizumab treatment in RECOVERY reduced 28-day all-cause mortality in patients hospitalised with COVID-19.9 Similarly, treatment of patients with COVID-19 with tofacitinib added to SOC—including corticosteroids—decreased risk of clinical events.8 The WHO recommendations for SOC for patients with severe or critical COVID-19 now include both IL-6 receptor blockers and corticosteroids or baricitinib and corticosteroids.4 Both baricitinib and IL-6 receptor blockers included in SOC are systemic treatments and not localised to the organ of interest in COVID-19, the lung.7 20 Targeting the lung—the organ of greatest relevance for reducing COVID-mediated inflammation—with an inhaled pan-JAK inhibitor, like nezulcitinib, has potential to further increase treatment benefit.

Patients who develop severe COVID-19 are notably heterogeneous.21 Higher baseline CRP levels—which are more strongly associated with poor outcomes during hospitalisation for severe COVID-19 than clinical risk factors such as age or comorbidities—potentially indicate risk for a severe COVID-19 clinical phenotype, suggesting patients admitted to the hospital should be routinely screened for inflammatory biomarkers.16 22 In the post hoc subgroup analysis in patients with baseline CRP <150 mg/L—the majority of patients—TTR, 28-day all-cause mortality rate, and VFS were all improved with nominal significance in patients treated with nezulcitinib versus placebo. This is consistent with emerging evidence for oral JAK inhibitors dosed at or above the high end of their approved dose ranges for up to 14 days, where survival benefits on a population level begin to emerge after 7–14 days of dosing.7–9 23 24 Despite improvement in survival, mortality rates compared with SOC continue proportionally out to ~28 days, demonstrating additional unmet need and opportunity to impact outcomes. Therefore, extending and/or augmenting nezulcitinib treatment duration and dose might also be beneficial to maximise the potential benefit of this immune-modulating therapy, especially in patients requiring prolonged organ support.

Limitations of this study include the small population; power calculations were based on the primary endpoint, and the study was underpowered to examine exploratory endpoints and some post hoc comparisons. The effect of concomitant remdesivir treatment could not be assessed as few patients received remdesivir. Most patients were categorised as OS 5, and results may not be generalisable to patients in higher categories. Only a single dose level of 3 mg was assessed, and outcomes after 28 days were not evaluated. The variants of SARS-CoV-2 that were circulating during the study time period are no longer the most prevalent; results may not be generalisable to newer variants. Due to the locations where the study was conducted, only White patients were enrolled, which may limit generalisability to other populations.

Nezulcitinib treatment of hospitalised patients with COVID-19 was not associated with reduction in number of RFF days compared with placebo treatment. However, observed numerical improvements in TTR, 28-day all-cause mortality, and VFS in patients treated with nezulcitinib 3 mg versus placebo suggest that nezulcitinib treatment may be beneficial in certain biomarker-defined subpopulations. Reducing overall mortality in hospitalised patients with severe COVID-19 remains an important goal of clinical research. Therefore, additional clinical studies of efficacy and safety of novel immunomodulator therapies, such as nezulcitinib, which may have additive anti-inflammatory effects when used in combination with systemic corticosteroid treatment are ongoing.