ABSTRACT
BACKGROUND: Predicting outcomes of critically ill intensive care unit (ICU) patients with coronavirus-19 disease (COVID-19) is a major challenge to avoid futile, and prolonged ICU stays. METHODS: The objective was to develop predictive survival models for patients with COVID-19 after 1-to-2 weeks in ICU. Based on the COVID-ICU cohort, which prospectively collected characteristics, management, and outcomes of critically ill patients with COVID-19. Machine learning was used to develop dynamic, clinically useful models able to predict 90-day mortality using ICU data collected on day (D) 1, D7 or D14. RESULTS: Survival of Severely Ill COVID (SOSIC)-1, SOSIC-7, and SOSIC-14 scores were constructed with 4244, 2877, and 1349 patients, respectively, randomly assigned to development or test datasets. The three models selected 15 ICU-entry variables recorded on D1, D7, or D14. Cardiovascular, renal, and pulmonary functions on prediction D7 or D14 were among the most heavily weighted inputs for both models. For the test dataset, SOSIC-7's area under the ROC curve was slightly higher (0.80 [0.74-0.86]) than those for SOSIC-1 (0.76 [0.71-0.81]) and SOSIC-14 (0.76 [0.68-0.83]). Similarly, SOSIC-1 and SOSIC-7 had excellent calibration curves, with similar Brier scores for the three models. CONCLUSION: The SOSIC scores showed that entering 15 to 27 baseline and dynamic clinical parameters into an automatable XGBoost algorithm can potentially accurately predict the likely 90-day mortality post-ICU admission (sosic.shinyapps.io/shiny). Although external SOSIC-score validation is still needed, it is an additional tool to strengthen decisions about life-sustaining treatments and informing family members of likely prognosis.
ABSTRACT
BACKGROUND: In the Île-de-France region (henceforth termed Greater Paris), extracorporeal membrane oxygenation (ECMO) for severe acute respiratory distress syndrome (ARDS) was considered early in the COVID-19 pandemic. We report ECMO network organisation and outcomes during the first wave of the pandemic. METHODS: In this multicentre cohort study, we present an analysis of all adult patients with laboratory-confirmed SARS-CoV-2 infection and severe ARDS requiring ECMO who were admitted to 17 Greater Paris intensive care units between March 8 and June 3, 2020. Central regulation for ECMO indications and pooling of resources were organised for the Greater Paris intensive care units, with six mobile ECMO teams available for the region. Details of complications (including ECMO-related complications, renal replacement therapy, and pulmonary embolism), clinical outcomes, survival status at 90 days after ECMO initiation, and causes of death are reported. Multivariable analysis was used to identify pre-ECMO variables independently associated with 90-day survival after ECMO. FINDINGS: The 302 patients included who underwent ECMO had a median age of 52 years (IQR 45-58) and Simplified Acute Physiology Score-II of 40 (31-56), and 235 (78%) of whom were men. 165 (55%) were transferred after cannulation by a mobile ECMO team. Before ECMO, 285 (94%) patients were prone positioned, median driving pressure was 18 cm H2O (14-21), and median ratio of the partial pressure of arterial oxygen to the fraction of inspired oxygen was 61 mm Hg (IQR 54-70). During ECMO, 115 (43%) of 270 patients had a major bleeding event, 27 of whom had intracranial haemorrhage; 130 (43%) of 301 patients received renal replacement therapy; and 53 (18%) of 294 had a pulmonary embolism. 138 (46%) patients were alive 90 days after ECMO. The most common causes of death were multiorgan failure (53 [18%] patients) and septic shock (47 [16%] patients). Shorter time between intubation and ECMO (odds ratio 0·91 [95% CI 0·84-0·99] per day decrease), younger age (2·89 [1·41-5·93] for ≤48 years and 2·01 [1·01-3·99] for 49-56 years vs ≥57 years), lower pre-ECMO renal component of the Sequential Organ Failure Assessment score (0·67, 0·55-0·83 per point increase), and treatment in centres managing at least 30 venovenous ECMO cases annually (2·98 [1·46-6·04]) were independently associated with improved 90-day survival. There was no significant difference in survival between patients who had mobile and on-site ECMO initiation. INTERPRETATION: Beyond associations with similar factors to those reported on ECMO for non-COVID-19 ARDS, 90-day survival among ECMO-assisted patients with COVID-19 was strongly associated with a centre's experience in venovenous ECMO during the previous year. Early ECMO management in centres with a high venovenous ECMO case volume should be advocated, by applying centralisation and regulation of ECMO indications, which should also help to prevent a shortage of resources. FUNDING: None.