ABSTRACT
Background: COVID-19 is associated with acute respiratory distress and cytokine release syndrome. The Janus kinase (JAK)1/JAK2 inhibitor ruxolitinib reduces inflammatory cytokine concentrations in disorders characterised by cytokine dysregulation, including graft-versus-host disease, myelofibrosis, and secondary hemophagocytic lymphohistiocytosis. We assessed whether treatment with the JAK1/JAK2 inhibitor ruxolitinib would be beneficial in patients with COVID-19 admitted to hospital. Methods: RUXCOVID was an international, randomised, double-blind, phase 3 trial of ruxolitinib plus standard of care versus placebo plus standard of care in patients with COVID-19. Patients who were hospitalised but not on mechanical ventilation or in the intensive care unit [ICU] were randomly assigned (2:1) to oral ruxolitinib 5 mg twice per day or placebo for 14 days (14 additional days were allowed if no improvement). The primary endpoint was a composite of death, respiratory failure (invasive ventilation), or ICU care by day 29, analysed by logistic regression including region, treatment, baseline clinical status, age, and sex as covariates. This trial is registered with ClinicalTrials.gov, NCT04362137. Findings: Between May 4 and Sept 19, 2020, 432 patients were randomly assigned to ruxolitinib (n=287) or placebo (n=145) plus standard of care; the mean age was 56·5 years (SD 13·3), 197 (46%) were female, and 235 (54%) were male. The primary objective was not met: the composite endpoint occurred in 34 (12%) of 284 ruxolitinib-treated patients versus 17 (12%) of 144 placebo-treated patients (odds ratio 0·91, 95% CI 0·48-1·73; p=0·77). By day 29, nine (3%) of 286 ruxolitinib-treated patients had died compared with three (2%) of 145 placebo-treated patients; 22 (8%) of 286 ruxolitinib-treated patients had received invasive ventilation compared with ten (7%) of 145 placebo-treated patients; and 30 (11%) of 284 ruxolitinib-treated patients had received ICU care compared with 17 (12%) of 144 placebo-treated patients. In an exploratory analysis, median time to recovery was 1 day faster with ruxolitinib versus placebo (8 days vs 9 days; hazard ratio 1·10, 95% CI 0·89-1·36). Adverse events included headache (23 [8%] of 281 on ruxolitinib vs 11 [8%] of 143 on placebo) and diarrhoea (21 [7%] vs 12 [8%]). Interpretation: Ruxolitinib 5 mg twice per day showed no benefit in the overall study population. A larger sample is required to determine the clinical importance of trends for increased efficacy in patient subgroups. Funding: Novartis and Incyte.
ABSTRACT
Emotion-aware music recommendations has gained increasing attention in recent years, as music comes with the ability to regulate human emotions. Exploiting emotional information has the potential to improve recommendation performances. However, conventional studies identified emotion as discrete representations, and could not predict users’ emotional states at time points when no user activity data exists, let alone the awareness of the influences posed by social events. In this study, we proposed an emotion-aware music recommendation method using deep neural networks (emoMR). We modeled a representation of music emotion using low-level audio features and music metadata, model the users’ emotion states using an artificial emotion generation model with endogenous factors exogenous factors capable of expressing the influences posed by events on emotions. The two models were trained using a designed deep neural network architecture (emoDNN) to predict the music emotions for the music and the music emotion preferences for the users in a continuous form. Based on the models, we proposed a hybrid approach of combining content-based and collaborative filtering for generating emotion-aware music recommendations. Experiment results show that emoMR performs better in the metrics of Precision, Recall, F1, and HitRate than the other baseline algorithms. We also tested the performance of emoMR on two major events (the death of Yuan Longping and the Coronavirus Disease 2019 (COVID-19) cases in Zhejiang). Results show that emoMR takes advantage of event information and outperforms other baseline algorithms.