A randomised, double-blind, placebo-controlled, multicentre clinical trial of AZD1656 in diabetic patients hospitalised with COVID-19: The ARCADIA Trial - implications for therapeutic immune modulation

Abstract

Background: A potential immunotherapeutic role for AZD1656 (a glucokinase activator) in the treatment of COVID-19 was hypothesized. The ARCADIA trial investigated the safety and efficacy of AZD1656 in diabetic patients admitted to hospital with COVID-19.

Methods: The ARCADIA trial was a Phase II randomised, double-blind, placebo-controlled clinical trial. Adult diabetic patients, admitted with COVID-19, were recruited at 28 hospitals in the UK, Romania and Czech Republic and randomly assigned (1:1) to receive AZD1656 tablets (100mg twice a day), or matched placebo, for up to 21 days, in addition to usual care. All involved were masked to treatment allocation. The primary endpoint was clinical improvement measured at Day 14. The Full Analysis Set (FAS) included all patients who received at least one dose of assigned treatment. ARCADIA is complete and registered with ClinicalTrials.gov (NCT04516759).

Findings: Between 29 September 2020 to 16 April 2021, 170 patients were screened and 156 patients were randomised, three of whom did not commence treatment. Of the remaining 153, 80 were assigned to AZD1656 and 73 were assigned to placebo and included in the Full Analysis Set (FAS). The primary analysis showed no statistically significant difference between groups (AZD1656: 76·3%; Placebo: 69·9%, p=0·19). There was no difference in the number of adverse events between groups (AZD1656: 35·7%; Placebo: 33·3%). Mortality was lower in the AZD1656 group compared to the placebo group (AZD1656: four (5%); Placebo: nine (12·3%), p=0·090)). At Day 7 there were zero deaths in the AZD1656 group compared to six deaths in the placebo group (p=0·011, post hoc). A difference between groups in time to hospital discharge was also seen (p=0·16). Immunophenotyping data suggested that AZD1656-treated patients had a less pro-inflammatory immune response and a better adaptive immune response than those treated with placebo.

Interpretation: Although the trial did not achieve its primary endpoint, AZD1656 was associated with a decrease in deaths and a reduction in the duration of hospitalisation, as compared to Placebo. Immunophenotyping and immunochemistry indicated an immunomodulatory effect of AZD1656. The trial suggests a beneficial therapeutic effect of AZD1656 and identifies a new therapeutic concept: small molecule activation of endogenous homeostatic immune cells which themselves become the therapeutic agent within the body. Phase 2 trials of this size carry the risk of false positive results and confirmation of these results in a larger clinical trial is now required.

Funding: UK Research and Innovation (UKRI) 'Innovate UK' programme and Excalibur Medicines Ltd.

Keywords: ARCADIA; AZD1656; COVID-19; Clinical trial; Cytokine; Diabetes; Glucokinase activator; Immune; Immunophenotyping; Regulatory T cells; SARs-COV-2; Tregs.

Conflict of interest statement

Dr. McCafferty reports grants and personal fees from AstraZeneca (who donated Drug Product for this trial) for activities outside of the submitted work. The authors declare no other competing interests.

© 2022 The Authors.

Figures

Figure 1

Trial profile. 170 patients consented and were screened for study eligibility. 156 eligible patients were subsequently randomised (UK: 74 patients; Czech Republic: 49 patients; Romania: 33 patients). 153 of the randomised patients received at least one dose of either AZD1656 or placebo and were included in the Full Analysis Set (FAS) (the Intent-to-Treat population). The FAS included 80 patients in the AZD1656 group and 73 patients in the Placebo group.

Figure 2

Time from admission to hospital discharge (analysis set: FAS). Kaplan-Meier estimates for time to hospital discharge showed that the probability of staying in hospital curve remained higher in the placebo group than in the AZD1656 group from approximately Day 7 onwards (p=0·16).

Figure 3

Clinical improvement at day 14 – odds ratios by overall and subgroups – forest plot (analysis set: FAS). Subgroup analyses of clinical improvement at Day 14 indicated a difference between groups in favour of AZD1656 in male patients (OR 1·78; 95% CI 0·72-4·4), patients aged ≥ 65 to

Figure 4

Time from randomisation to Death - showing deaths only - post hoc (analysis set: FAS). More patients died in the Placebo group than in the AZD1656 group (Placebo: 9 AZD1656: 4) and they did so earlier in the course of their hospitalization than those in the AZD1656 group (p-value = 0·074). Only patients who died are presented in this plot. All patients, with and without an event (death), were used in the log-rank test.

Figure 5

Time from randomisation to hospital discharge, subgroup: Patients with baseline IL6 >= 13 pg/mL - post hoc (Analysis Set: FAS) Patients with baseline IL-6 values ≥ 13 pg/mL had a lower probability of staying in hospital in the AZD1656 group than in the Placebo group (p-value = 0·021). In this sub-group, the mean hospitalization time was approximately 3 days shorter in the AZD1656 group than the Placebo group (AZD1656: 191 hours; Placebo: 271 hours (p=0·056).

Figure 6

Immune effects of AZD1656 observed in the ARCADIA trial.