ABSTRACT
INTRODUCTION: Severe coronavirus disease-19 (COVID-19) is characterized by progressive hypoxemia and patients may require advanced oxygen therapy, including high-flow nasal cannula (HFNC) therapy and mechanical ventilation. Previous data has suggested that the ROX index, IL-6 levels, thrombocytopenia, and kidney injury may predict failure of high-flow nasal cannula therapy. Our study aims to evaluate risk factors that predict HFNC failure in our patient population. METHOD(S): Retrospective cohort study of patients treated for COVID-19 across 4 hospitals in Atlanta, Georgia between February 2020 and February 2021. Patients placed on high-flow nasal cannula within the first 24 hours of admission and who remained on high-flow nasal cannula for at least 6 consecutive hours were identified. Patients that met our cohort criteria were followed for the first seven days of admission and transition across oxygen therapy modalities were examined. Demographic and comorbidity data of patients who failed high-flow nasal cannula therapy within the first 7 days, defined as need for mechanical ventilation or death, were compared to patients who did not fail. RESULT(S): There were 1205 patients placed on high-flow nasal cannula oxygen therapy in our hospitals between February 2020 and February 2021. In total, 465 patients met inclusion criteria. Of the cohort, 35.9% remained on highflow, 32.0 % transitioned to low-flow or room air, and 31.6% failed high-flow nasal cannula therapy within the first week of hospitalization (26.2% failed due to requiring intubation and 5.4% failed due to death). When comparing demographics and comorbidities, patients who failed were older (median age 67.5 vs 62 years, p=0.01) and more frequently had renal disease (28.8% vs 18.5%, p=0.02). There were no significant differences in sex, race, congestive heart failure, pulmonary disease, hypertension, diabetes mellitus, liver disease, or metastatic cancer. CONCLUSION(S): In our patient population, 31.6% of patients failed high-flow nasal cannula therapy within the first week of admission due to mechanical ventilation or death. Age and renal disease were significant risk factors for highflow nasal cannula therapy failure in COVID-19 patients.
ABSTRACT
In the current situation of a pandemic caused by COVID-19 developing models accurately predicting the dynamics of the outbreaks in time and space became extremely important. Individual-based models (IBM) simulating the spread of infection in a population have a few advantages compared to classical equation-based approach. First, they use individuals as units, which represent the population, and reflect the local variations happening in real life. Second, the simplicity of modelling the interactions between the individuals, which may not be the case when using differential equations. We propose to use freely available population density maps to simulate the infection spread in the human population on the scale of an individual country or a city. We explore the effect of social distancing and show that it can reduce the outbreak when applied before or during peak time, but it can also inflict the second wave when relaxed after the peak. This can be explained by a large proportion of susceptible individuals, even in the large cities, after the first wave. The model can be adapted to any spatial scale from a single hospital to multiple countries.