LAUSR

INTRODUCTION: Predictions of ICU outcome vary by diagnosis, even when adjusted for age, physiology and chronic health We investigated APACHE-IVb adjusted outcomes in adult patients with ICU admission for novel coronavirus versus other viral pneumonias (VP) with and without ARDS METHODS: De-identified patient data from 132 hospitals in the US contributing data on VP patients to the APACHE database between 1/1/2014 and 12/31/2019 was compared with 44 hospitals contributing COVID-19 (C19) data between 3/14 and 6/17/20 RESULTS: 4200 pts had a primary (n=2544) or secondary (n=1666) ICU admitting diagnosis of viral pneumonia and 1494 had C19 infection Mean age was 64 1 for VP vs 63 4 for COVID (nsd) APACHEIII/IV scores were similar at 55 0 and 56 7 but gender and ethnic distributions differed with fewer females and Caucasians having C19 PaO2 / FiO2 ratio and WBC count at admission were lower with C19 Standardized APACHE IVb Mortality (95% CI) was 0 98 (0 88-1 09) for VP versus 1 52 (1 36-1 69) for C19 ICU (8 vs 5d) and hospital length of stay (13 vs 10d) were significantly longer with C19 patients, who had less IPPV or NIV support (47 vs 65%) but had more ventilator days (10 vs 4) A subgroup of VP patients with concurrent ARDS (n=202) were most similar to the C19 patients, with longer ICU LOS (13 d) and high mortality (21 6%), but SMR in this group was 0 89 (0 61-1 16) CONCLUSIONS: Severity-adjusted mortality and LOS are higher for C19 patients than for VP patients admitted to ICU C19 patients also have longer ventilator days and LOS, comparable to a subset of VP patients with concurrent ARDS APACHE-IVb under-predicts mortality across all levels of risk with C19 Admission APS/APACHE III/IV scores are similar, but pre-ICU LOS is longer Mortality and LOS increase with age and higher scores in both groups, but observed to predicted mortality and LOS are higher than expected with C19 patients These findings have implications for benchmarking ICU outcomes during the COVID-19 pandemic