458. A Machine Learning Approach Identifies Distinct Early-Symptom Cluster Phenotypes Which Correlate with Severe SARS-CoV-2 Outcomes

Abstract Background The novel coronavirus disease 2019 (COVID-19) pandemic remains a global challenge. Accurate COVID-19 prognosis remains an important aspect of clinical management. While many prognostic systems have been proposed, most are derived from analyses of individual symptoms or biomarkers. Here, we take a machine learning approach to first identify discrete clusters of early stage-symptoms which may delineate groups with distinct symptom phenotypes. We then sought to identify whether these groups correlate with subsequent disease severity. Methods The Epidemiology, Immunology, and Clinical Characteristics of Emerging Infectious Diseases with Pandemic Potential (EPICC) study is a longitudinal cohort study with data and biospecimens collected from nine military treatment facilities over 1 year of follow-up. Demographic and clinical characteristics were measured with interviews and electronic medical record review. Early symptoms by organ-domain were measured by FLU-PRO-plus surveys collected for 14 days post-enrollment, with surveys completed a median 14.5 (Interquartile Range, IQR = 13) days post-symptom onset. Using these FLU-PRO-plus responses, we applied principal component analysis followed by unsupervised machine learning algorithm k-means to identify groups with distinct clusters of symptoms. We then fit multivariate logistic regression models to determine how these early-symptom clusters correlated with hospitalization risk after controlling for age, sex, race, and obesity. Results Using SARS-CoV-2 positive participants (n = 1137) from the EPICC cohort (Figure 1), we transformed reported symptoms into domains and identified three groups of participants with distinct clusters of symptoms. Logistic regression demonstrated that cluster-2 was associated with an approximately three-fold increased odds [3.01 (95% CI: 2-4.52); P < 0.001] of hospitalization which remained significant after controlling for other factors [2.97 (95% CI: 1.88-4.69); P < 0.001]. (A) Baseline characteristics of SARS-CoV-2 positive participants. (B) Heatmap comparing FLU-PRO response in each participant. (C) Principal component analysis followed by k-means clustering identified three groups of participants. (D) Crude and adjusted association of identified cluster with hospitalization. Conclusion Our findings have identified three distinct groups with early-symptom phenotypes. With further validation of the clusters’ significance, this tool could be used to improve COVID-19 prognosis in a precision medicine framework and may assist in patient triaging and clinical decision-making. Disclaimer Disclosures David A. Lindholm, MD, American Board of Internal Medicine (Individual(s) Involved: Self): Member of Auxiliary R&D Infectious Disease Item-Writer Task Force. No financial support received. No exam questions will be disclosed ., Other Financial or Material Support Ryan C. Maves, MD, EMD Serono (Advisor or Review Panel member)Heron Therapeutics (Advisor or Review Panel member) Simon Pollett, MBBS, Astra Zeneca (Other Financial or Material Support, HJF, in support of USU IDCRP, funded under a CRADA to augment the conduct of an unrelated Phase III COVID-19 vaccine trial sponsored by AstraZeneca as part of USG response (unrelated work))

Background. Up until this day, over 3.5 million fatalities related to coronavirus disease 2019 (COVID-19) have been registered worldwide by the World Health Organization. Healthcare professionals require prognostic tools for COVID-19 patients in order to guide treatment strategies. Elevated troponin levels, a biomarker of cardiac injury, have been detected among patients with COVID-19, hence associating it with cardiac injury. Although several studies have mentioned it, the role of troponin as a prognosis biomarker is unclear. Elevation in troponin levels has been observed in patients with community-acquired pneumonia (CAP). However, its association with mortality is scarcely mentioned in literature. Thus, we sought to determine the utility of serum troponin I levels as a mortality predictor for patients with COVID-19 and CAP.
Methods. A prospective observational study was carried out at Clinica Universidad de La Sabana, Colombia, with patients hospitalized due to CAP and COVID-19. Troponin biomarker was quantified in serum samples using the PATHFAST system within the first 24 hours of hospital admission. Serum concentrations of troponin were compared among study groups. To assess the biomarker´s capacity to predict mortality, ROC curves were used, quantifying their differences through the DeLong´s test.
Results. A total of 88 patients with CAP and 152 with COVID-19 were included in the study. In all cohort the median [IQR] serum concentration of troponin (ng/ml) was higher in those who died (34.2, [9.74-384] vs 5.89, [2.44-27.9] p< 0.001). Furthermore, troponin was higher in deceased patients with COVID-19 vs those who survived (77.35 [11.9-346.5] vs. 4.88 [2.10-13.02], p< 0.001). However, there was no significant difference between CAP deceased and not deceased patients (18.1 [8.52-398]  Background. The novel coronavirus disease 2019 (COVID-19) pandemic remains a global challenge. Accurate COVID-19 prognosis remains an important aspect of clinical management. While many prognostic systems have been proposed, most are derived from analyses of individual symptoms or biomarkers. Here, we take a machine learning approach to first identify discrete clusters of early stage-symptoms which may delineate groups with distinct symptom phenotypes. We then sought to identify whether these groups correlate with subsequent disease severity.
Methods. The Epidemiology, Immunology, and Clinical Characteristics of Emerging Infectious Diseases with Pandemic Potential (EPICC) study is a longitudinal cohort study with data and biospecimens collected from nine military treatment facilities over 1 year of follow-up. Demographic and clinical characteristics were measured with interviews and electronic medical record review. Early symptoms by organ-domain were measured by FLU-PRO-plus surveys collected for 14 days post-enrollment, with surveys completed a median 14.5 (Interquartile Range, IQR = 13) days post-symptom onset. Using these FLU-PRO-plus responses, we applied principal component analysis followed by unsupervised machine learning algorithm k-means to identify groups with distinct clusters of symptoms. We then fit multivariate logistic regression models to determine how these early-symptom clusters correlated with hospitalization risk after controlling for age, sex, race, and obesity.
(A) Baseline characteristics of SARS-CoV-2 positive participants. (B) Heatmap comparing FLU-PRO response in each participant. (C) Principal component analysis followed by k-means clustering identified three groups of participants. (D) Crude and adjusted association of identified cluster with hospitalization.
Conclusion. Our findings have identified three distinct groups with early-symptom phenotypes. With further validation of the clusters' significance, this tool could be used to improve COVID-19 prognosis in a precision medicine framework and may assist in patient triaging and clinical decision-making. Disclaimer.

Background. Evidence on outcomes after COVID-19 hospitalization is limited.
This study aimed to characterize 30-day readmission beyond the initial COVID-19 hospitalization.
Methods. This descriptive retrospective cohort study included adult patients admitted between 07/01/2020 and 01/31/2021 with a discharge diagnosis of COVID-19 (ICD-10-CM: U07.1), using a large hospital inpatient chargemaster with a linked open claims dataset. The first COVID-19 hospitalization was considered index hospitalization; baseline was defined as first 2 days of index hospitalization; readmission was assessed within 30 days of discharge from index hospitalization. We describe the demographics, treatments and outcomes of the index hospitalization and readmission.

Conclusion.
In a large, geographically diverse cohort of hospitalized COVID-19 patients, 16% required readmission, especially in those with greater age and comorbidities. Over the study period, all-cause readmission remained stable and was lower in RDV treated patients.
Disclosures Background. To combat higher rates of COVID-19 infection, hospitalization, and death among minorities, it is crucial to identify safe, efficacious, and generalizable treatments. Therefore, the purpose of this systematic literature review was to assess the demographic characteristics of COVID-19 clinical trial participants.
Methods. A literature search was performed according to the PRISMA checklist using PubMed from December 1, 2019 to November 24, 2020 with the following search terms: 2019-nCoV, COVID-19, SARS-CoV-2, clinical trial, randomized controlled trial, observational study, and veterinary. To capture additional results, keyword searches were performed using various versions and plural endings with the title/ abstract field tag. Randomized controlled trials evaluating a pharmacologic treatment for COVID-19 patients from one or more U.S site written in the English language were eligible for inclusion. Descriptive statistics were calculated to characterize age, gender, race, and ethnicity of patients enrolled in the included COVID-19 clinical trials, as well as for comparison with national COVID-19 data.
Results. A total of 4472 records were identified, of which 16 were included. Most were placebo-controlled (69%) and included hospitalized patients with COVID-19 (69%). Demographic data were reported for each study arm in 81% of studies. Median number of participants was higher in studies of nonhospitalized patients (n=452 [range 20-1062] vs n=243 [range 152-2795]). Nine (56%) studies reported mean or median ages of 50 years or older amongst all study arms. Males comprised more than half of the study cohort in 50% of studies. Race and ethnicity were reported separately in five (31%) studies, reported in combination in four (25%), while six (38%) reported only race or ethnicity. White or Caucasian patients made up most participants across all arms in 75% of studies. Based on national COVID-19 data, hospitalizations were similar between White persons and African American persons, but higher than other race or ethnic groups, and evenly distributed among males and females.
Conclusion. Lack of heterogeneously reporting demographic characteristics of COVID-19 clinical trial participants limits the ability to assess the generalizability of pharmacologic treatments for COVID-19.
Disclosures. All Authors: No reported disclosures