Predictors and outcomes of high-flow nasal cannula failure following extubation: A multicentre observational study

INTRODUCTION@#Despite adhering to criteria for extubation, up to 20% of intensive care patients require re-intubation, even with use of post-extubation high-flow nasal cannula (HFNC). This study aims to identify independent predictors and outcomes of extubation failure in patients who failed post-extubation HFNC.@*METHODS@#We conducted a multicentre observational study involving 9 adult intensive care units (ICUs) across 5 public hospitals in Singapore. We included patients extubated to HFNC following spontaneous breathing trials. We compared patients who were successfully weaned off HFNC with those who failed HFNC (defined as re-intubation ≤7 days following extubation). Generalised additive logistic regression analysis was used to identify independent risk factors for failed HFNC.@*RESULTS@#Among 244 patients (mean age: 63.92±15.51 years, 65.2% male, median APACHE II score 23.55±7.35), 41 (16.8%) failed HFNC; hypoxia, hypercapnia and excessive secretions were primary reasons. Stroke was an independent predictor of HFNC failure (odds ratio 2.48, 95% confidence interval 1.83-3.37). Failed HFNC, as compared to successful HFNC, was associated with increased median ICU length of stay (14 versus 7 days, @*CONCLUSION@#Post-extubation HFNC failure, especially in patients with stroke as a comorbidity, remains a clinical challenge and predicts poorer clinical outcomes. Our observational study highlights the need for future prospective trials to better identify patients at high risk of post-extubation HFNC failure.

Severe COVID-19 and coagulopathy: A systematic review and meta-analysis

INTRODUCTION@#Coronavirus disease 2019 (COVID-19)-induced coagulopathy (CIC) has been widely reported in the literature. However, the spectrum of abnormalities associated with CIC has been highly variable.@*METHODS@#We conducted a systematic review of the literature (until 1 June 2020) to assess CIC and disease severity during the early COVID-19 pandemic. Primary outcomes were pooled mean differences in platelet count, D-dimer level, prothrombin time, activated partial thromboplastin time (aPTT) and fibrinogen level between non-severe and severe patients, stratified by degree of hypoxaemia or those who died. The risk factors for CIC were analysed. Random-effects meta-analyses and meta-regression were performed using R version 3.6.1, and certainty of evidence was rated using the Grading of Recommendation, Assessment, Development, and Evaluation approach.@*RESULTS@#Of the included 5,243 adult COVID-19 patients, patients with severe COVID-19 had a significantly lower platelet count, and higher D-dimer level, prothrombin time and fibrinogen level than non-severe patients. Pooled mean differences in platelet count (-19.7×109/L, 95% confidence interval [CI] -31.7 to -7.6), D-dimer level (0.8μg/mL, 95% CI 0.5-1.1), prothrombin time (0.4 second, 95% CI 0.2-0.6) and fibrinogen level (0.6g/L, 95% CI 0.3-0.8) were significant between the groups. Platelet count and D-dimer level were significant predictors of disease severity on meta-regression analysis. Older men had higher risks of severe coagulopathic disease.@*CONCLUSION@#Significant variability in CIC exists between non-severe and severe patients, with platelet count and D-dimer level correlating with disease severity. Routine monitoring of all coagulation parameters may help to assess CIC and decide on the appropriate management.