LONGITUDINAL CHARACTERIZATION OF PRESENTATION CLINICAL TRAITS IN PATIENTS DURING INDEX AND AT LEAST ONE READMISSION POST-ACUTE COVID-19 HOSPITALIZATION

ABSTRACT

BACKGROUND: COVID-19 patients at hospitalization exhibit heterogeneous risk factors putatively associated with hospital mortality. However, stability of risk stratification across readmission (ReAdm) remains unclear. We evaluated longitudinal patterns in presentation clinical traits at index COVID-19 hospitalization and subsequent ReAdm(s). METHODS: Under IRB exemption, discharge electronic medical records underwent extraction of presentation demographics, anthropometrics, laboratory results, and ICD-10 codes. Univariate logistic regression was used to test association (p<.05) of putative clinical traits with COVID-19 hospitalization mortality. Continuous data summarized with median [IQR] were compared using Kruskal-Wallis K statistic. Discrete data summarized as counts or proportions were compared with chi-squared test. Confounders statistically balanced included age, sex, race, comorbidities, and attendant local 4-surges of pandemic. Statistical significance was Bonferroni corrected for multiple contrasts at .017. RESULTS: Among patients discharged alive not to hospice between March 18, 2020 and September 30, 2021, 4430 underwent index COVID-19 hospitalization. Incident with at least one subsequent ReAdm was experienced respectively by 630 and 166 patients. There was no difference in median age 76 [62,85] years among 45% woman distributed across Whites (82%), Blacks (9%) and other races (9%). Time to incident and first subsequent ReAdm respectively was 14 [4,59] vs 21 [6,54] days. Prominent comorbidity prevalence sustained included hypertension (54%), diabetes (34%), chronic pulmonary disease (29%), obesity (24%), and coagulopathy (15%). Significant differential comorbidity prevalence manifested sequentially with deficiency anemias (26%, 32%, 49%) and heart failure (23%, 28%, 38%). Notable at two or more ReAdms and consistent with some comorbidity patterns, is significantly worsening hypoalbuminemia, anemia, neutropenia with increasing creatinine, pBNP and D-dimer. CONCLUSIONS: Patients undergoing ReAdm demonstrated that characterizing trends in SARS-CoV-2 evoked clinical traits may reveal mitigable features of post-acute COVID-19 syndrome. Evaluating linkage between biomarkers and comorbidities across ReAdm patterns can identify those of value for estimating a given outcome.