Epidemiology of Laboratory-identified Late-onset SARS-CoV-2 Positivity in Two Large, Urban, Acute-Care Hospitals: Implications for Surveillance of Hospital-Acquired COVID-19

ABSTRACT

Background. Laboratory identification (Lab-ID) of late-onset SARS-CoV-2 positive tests during a hospital stay is a potential public health surveillance approach for hospital-acquired COVID-19. However, prolonged RNA fragment shedding and intermittent detection of SARS-CoV-2 virus via PCR testing among infected patients may hamper interpretation of laboratory-identified events. We aimed to describe the epidemiology of late-onset SARS-CoV-2 laboratory events using clinical criteria, to evaluate the feasibility of a Lab-ID approach to detection of nosocomial SARS-COV-2 infection. Methods. We evaluated all SARS-CoV-2 RT-PCR positive results recovered from patients at two acute-care hospitals in Chicago, IL, during March 1 - November 30, 2020. Each hospital maintained stringent infection control policies through-out the study period. Through chart review (WT & CS), we categorized all initial SARSCoV-2 positive tests collected > Hospital Day 5 (defined as 'late-onset' based on the 5-day mean incubation period for COVID-19) into the following clinical categories Community Acquired;Unlikely Hospital Acquired;Possible Hospital Acquired;and Probable Hospital Acquired. Categorizations were made using hospital day, symptoms, alternative diagnoses, and clinical notes (Figure 1). Results. Of 2,671 SARS-CoV-2-positive patients, most positive tests (n=2,551;96%) were recovered pre-admit or by Hospital Day 2;first positive tests were uncommon during Hospital Days 6 to 14 (n=40;1.5%);and rare after Hospital Day 14 (n=15;0.6%). By chart review, of the 55 late-onset records reviewed, categorizations in descending order were Prior positive at outside facility (n=23);Possible Hospital Acquired (n=16);Community Acquired (n=12);Probable Hospital Acquired (n=4). Less than half of the late-onset cases were categorized as a possible or probable hospital acquisition (Figure 2). Conclusion. Hospital-acquired SARS-CoV-2 infection was uncommon. Most late-onset episodes of SARS-CoV-2 were explained by detection at an outside healthcare facility or by delayed diagnosis of patients with symptoms at time of presentation. A Lab-ID approach to nosocomial COVID-19 surveillance would potentially misclassify a substantial number of patients.