Utility of asymptomatic inpatient testing for COVID-19 in a low-prevalence setting: A multicenter point-prevalence study

Anthony D. Bai MD1 , Xena X. Li MD1, Mohammed AlsalemMD1, Sarah KhanMD2, Marek Smieja MD3, Dominik Mertz MD4 and Zain Chagla MD5 1Division of Infectious Diseases, McMaster University, Hamilton, Ontario, Canada, 2Infection Prevention Control, McMaster Children’s Hospital, Hamilton, Ontario, Canada, 3Laboratory Medicine, St Joseph’s Healthcare Hamilton and Hamilton Health Sciences, Hamilton, Ontario, Canada, 4Infection Control, Hamilton Health Sciences, Hamilton, Ontario, Canada and 5Infection Control, St. Joseph’s Healthcare Hamilton, Hamilton, Ontario, Canada


Methods
We conducted a point-prevalence study across 4 tertiary acute-care hospitals in Hamilton from April 15 to April 21, 2020. The Hamilton Integrated Research Ethics Board approved this study (no. 10894).

COVID-19 testing on admission
According to provincial guidelines, testing was based on the following symptoms: fever, new or worsening acute respiratory illness symptom (ie, cough, dyspnea, sore throat, runny nose or sneezing, nasal congestion, hoarse voice, difficulty swallowing, new olfactory or taste disorder(s), nausea or vomiting, diarrhea, abdominal pain), or clinical or radiological evidence of pneumonia. 6 Atypical presentations included unexplained fatigue or malaise, delirium, falls, acute functional decline, exacerbation of chronic conditions, chills, headache, croup, tachycardia, decrease in blood pressure, hypoxia, and lethargy. 6 At the time of this study, a patient with any of the above symptoms or exposure underwent nasopharyngeal swab testing for SARS-CoV-2 upon admission to the hospital. 6

Patient inclusion
On the point-prevalence testing date, all adult inpatients were tested once if they were admitted for 7-14 days, regardless of symptoms or prior negative SARS-CoV-2 test result. Patients with a known positive SARS-CoV-2 test were excluded.
Testing on days 7-14 was based on the estimated median incubation period of 4 days (interquartile range, 2-7 days). 7 Testing after the median incubation period would have captured most COVID-19 cases, even if the exposure occurred as late as the day of admission.

Testing procedure
The nasopharyngeal swabs were collected, and a polymerase-chain reaction assay for the SARS-CoV-2 envelope and 5'-untranslated region genes was performed at the local virology laboratory in the hospital. This assay was validated against the provincial standard testing.

Data collection
Data were extracted from the patient electronic chart system, which included demographics, admitting diagnosis, hospital location, admitting service, reason for admission, Charlson comorbidity index, 8 prior SARS-CoV-2 test result, chest imaging, and other microbiology test results. On the day of testing, patients were assessed for symptoms, as listed above. 6
Only 1 patient (0.8%) was positive for SARS-CoV-2. This patient presented to hospital C reporting 2 weeks of fever and cough. A chest x-ray showed an ill-defined opacity in the left lower lobe. The patient was initially isolated for acute respiratory illness. Isolation was discontinued on admission day 2 after the SARS-CoV-2 test came back negative, and the patient was treated for a presumptive bacterial pneumonia. The following day, the patient was transferred to hospital D. The positive point-prevalence test occurred on admission day 13 at hospital D. At the time of testing, the patient had no new symptoms and the patient's respiratory status continued to improve.

Discussion
In this point-prevalence study, 125 inpatients were tested, and only 1 patient (0.8%) was positive for SARS-CoV-2. This positive case was symptomatic, and the patient had had a prior SARS-CoV-2 test that was likely a false negative. He was initially isolated for acute respiratory illness on presentation to the hospital. For COVID-19, infectiousness has been estimated to decline quickly within 7 days, 9 so the patient was likely no longer infectious at the time of the second test. Therefore, asymptomatic testing did not add any useful information or change infection control practices compared to symptom-based screening. The hospital sites include a 607-bed hospital that is the regional cardiac surgery and neurosurgery center; a 228-bed hospital that is the regional center for cancer care and bone marrow transplant; a 250-bed hospital that specializes in chronic care and rehabilitation; and a 426-bed hospital that specializes in dialysis and renal transplant. c Immunosuppression includes any of the following: chemotherapy, steroid therapy, neutropenia with absolute neutrophil count <0.5, active hematological malignancy, HIV, primary immunodeficiency, or solid organ or hematopoietic stem cell transplant requiring immunosuppressive therapy. To our knowledge, this is the first study to evaluate the benefit of asymptomatic testing for hospitalized patients in a low-prevalence setting. In a New York hospital, universal testing of women admitted for delivery showed 13.5% asymptomatic positive SARS-CoV-2 results. 10 In contrast, our study found no asymptomatic positive cases. This finding is likely due to differences in local prevalence.
The strengths of our study include the systematic approach to testing. Also, the inclusion of 4 hospitals makes the results more generalizable. Our study has 2 limitations. First, repeated pointprevalence testing would have yielded more precise results, but this method would not have been feasible given the capacity of our virology laboratory. Second, nasopharyngeal swabbing may produce false-negative results, given its estimated sensitivity between 75% and 95%. 4 Although imperfect, nasopharyngeal swabbing is practical and is currently the recommended test for asymptomatic patients. 4 In conclusion, our study suggests the minimal utility of asymptomatic testing in hospitalized patients compared to symptom screening and targeted testing in low-prevalence settings, which supports the current IDSA guidelines. 4