key: cord-0701315-6ba2qnm3
authors: Egbuonu, Kenechukwu; Hyle, Emily P; Hurtado, Rocio M; Alba, George A; Zachary, Kimon C; Branda, John A; Hibbert, Kathryn A; Hooper, David C; Shenoy, Erica S; Turbett, Sarah E; Dugdale, Caitlin M
title: Yield of Severe Acute Respiratory Syndrome Coronavirus 2 Lower Respiratory Tract Testing After a Negative Nasopharyngeal Test Among Hospitalized Persons Under Investigation for Coronavirus Disease 2019
date: 2021-05-19
journal: Open Forum Infect Dis
DOI: 10.1093/ofid/ofab257
sha: c4d198aa4d376b9c49911b79374336a35f7ad6e1
doc_id: 701315
cord_uid: 6ba2qnm3

Among hospitalized persons under investigation for coronavirus disease 2019 (COVID-19), more repeated severe acute respiratory syndrome coronavirus 2 nucleic acid amplification tests (NAATs) after a negative NAAT were positive from lower than from upper respiratory tract specimens (1.9% vs 1.0%, P = .033). Lower respiratory testing should be prioritized among patients displaying respiratory symptoms with moderate-to-high suspicion for COVID-19 after 1 negative upper respiratory NAAT.

Accurate diagnosis of coronavirus disease 2019 (COVID- 19) among hospitalized persons under investigation (PUI) for COVID-19 is vital to ensure appropriate use of transmissionbased precautions and initiation of therapy for infected individuals [1] . Nucleic acid amplification tests (NAATs) are the gold standard for severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) detection and are most commonly performed on upper respiratory tract (URT; eg, nasal, nasopharyngeal, and saliva) specimens due to ease of collection and wide availability of validated testing platforms [1] . URT NAAT sensitivity varies from 70% to 95% based on time from symptom onset and other factors, raising concerns for false-negative tests [2] [3] [4] .

NAAT sensitivity is higher using lower respiratory tract (LRT; eg, sputum, tracheal aspirate, and bronchoalveolar lavage [BAL] ) specimens compared with URT specimens [5] [6] [7] [8] . The Infectious Diseases Society of America (IDSA) recommends repeated SARS-CoV-2 testing from LRT rather than URT specimens when suspicion for COVID-19 remains high despite a negative initial test [1] . However, data on the real-world yield of repeated NAATs using LRT specimens are lacking. We aimed to evaluate the yield of LRT NAATs among hospitalized PUI with moderate-to-high suspicion for COVID-19 despite an initial negative test and describe the characteristics of individuals diagnosed using LRT NAATs.

We conducted a retrospective study of adults >18 years old who underwent a SARS-CoV-2 NAAT for initial COVID-19 diagnosis between 1 March and 31 December 2020 at Massachusetts General Hospital (MGH). Subjects were identified through electronic health records. We included PUI without prior diagnosis of COVID-19 who were hospitalized at MGH for ≥24 hours. From 1 March to 7 April 2020, the PUI definition was limited to patients experiencing symptoms consistent with COVID-19 [1] . As of 8 April 2020, the PUI definition was broadened to include patients with an epidemiologic risk factor (eg, persons experiencing homelessness, exposed to a confirmed COVID-19 case, or residing in congregate settings), irrespective of symptoms. We excluded individuals tested while hospitalized but not meeting PUI criteria.

Repeated NAATs for diagnosis were recommended either by infectious diseases physicians (through 20 May 2020) or by the CORAL (COvid Risk cALculator) diagnostic algorithm (21 May 2020 onward) [9] . Sputum induction was not recommended. URT NAATs were almost exclusively performed on nasopharyngeal specimens using US Food and Drug Administration (FDA) emergency use authorization (EUA) assays [9, 10] . LRT NAATs were performed at the Massachusetts state laboratory, or with an internally validated protocol using the Cepheid Xpert Xpress SARS-CoV-2 assay.

Demographic characteristics of patients who did and did not undergo LRT testing were compared using χ 2 tests for proportions and Mann-Whitney U tests for continuous variables (Stata version 15.1). We considered a P value < .05 to be statistically significant. We grouped NAATs performed within a hospitalization into categories based on chronologic order, that is, first tests, second tests, and all third and subsequent tests combined. We compared URT and LRT NAAT positivity within each group using χ 2 tests. We then conducted medical records review to examine the clinical and radiologic characteristics of patients diagnosed with a positive LRT NAAT after an initial negative URT NAAT during the same hospitalization.

The study was approved by the Massachusetts General Brigham Institutional Review Board with a waiver of written informed consent.

A total of 18 379 SARS-CoV-2 NAATs were performed in 9925 hospitalized PUI. Demographic characteristics were similar between patients who did and did not have LRT testing performed, including sex (male: 59.7% vs 55.8%, P = .055), age (median [interquartile range], 61 [51-71] vs 63 [48-75] years, P = .052), and race (nonwhite: 23.9% vs 27.1%, P = .099).

Of the 18 379 NAATs examined, 17 682 (96.2%) were performed on URT specimens and 697 (3.8%) were performed on LRT specimens (Table 1) . Subjects in the intensive care unit had a higher proportion of NAATs performed on LRT specimens than subjects not in intensive care (39.3% vs 8.9%, P < .01). Among LRT NAATs, 56 of 697 (8.0%) were performed on BAL specimens; no BAL NAATs were positive. Among 641 of 697 (92.0%) LRT NAATs that were performed on sputum or tracheal aspirates, 15 (2.3%) were positive.

Among first NAATs performed during the hospitalization, 1209 of 11 198 (10.8%) URT NAATs were positive, while 2 of 28 (7.1%) of LRT NAATs were positive (P = .534). Among second NAATs, 58 of 5593 (1.0%) URT NAATs were positive compared with 6 of 114 (5.3%) LRT NAATs (P < .001). Positivity on third or later NAATs was similar between specimen types (URT: 9/891 [1.0%] vs LRT: 7/555 [1.3%], P = .657). Considering all repeated tests after the first negative NAAT, a lower proportion of URT NAATs were positive compared to LRT NAATs (67/6484 [1.0%] vs 13/669 [1.9%], P = .033).

Among the 13 subjects with COVID-19 diagnosed on LRT specimens after an initial negative URT NAAT, ages ranged from 22 to 82 years, 8 (61.5%) were male, and 6 (46.2%) selfidentified as non-Hispanic white ( Table 2) . All subjects had symptoms consistent with COVID-19 and abnormal chest radiographs; 9 of 12 (75.0%) subjects had chest computed tomographic findings typical of COVID-19 [11] . One subject with a chronic tracheostomy was diagnosed using a tracheal aspirate; all others were diagnosed using expectorated sputum. Subjects required up to 4 NAATs for diagnosis. Average time from symptom onset to first NAAT performed was 10 days (range, 0-21 days). Average time from symptom onset to diagnosis was 15 days (range, 2-26 days). Among subjects who underwent repeated NAATs within 14 days after diagnosis, 4 of 8 (50%) had positive tests (URT: 2/7 [28.6%]; LRT: 2/2 [100%]). Only 2 of 4 (50%) subjects with SARS-CoV-2 serologies performed had positive serologies. All subjects survived until hospital discharge and remained on transmission-based precautions for at least 10 days following diagnosis.

We found that the yield of repeated SARS-CoV-2 NAATs among hospitalized persons under investigation for COVID-19 was higher when the repeated test was performed on LRT compared with URT specimens. The greatest difference in test positivity between specimen types was observed on the second NAAT (URT: 1.0%, LRT: 5.3%). All subjects diagnosed using LRT NAATs displayed symptoms and/or imaging findings highly concerning for COVID-19. However, most subjects diagnosed using a LRT NAAT had their initial negative URT NAAT performed >7 days after symptom onset and were diagnosed with COVID-19 ≥14 days after symptom onset, Subject was unable to provide history directly, so information on presenting symptoms was obtained from a surrogate.

c Subject was immunosuppressed on rituximab; repeat severe acute respiratory syndrome coronavirus 2 serologies 1 month later also remained negative.

suggesting that LRT NAAT may be most useful for PUI presenting late in disease.

Our findings are consistent with other studies reporting higher sensitivity of NAATs from LRT compared with URT specimens [7, 12] . A meta-analysis involving 3442 NAATs found that sputum NAAT sensitivity (71% [95% confidence interval {CI}, 61%-80%]) was higher than nasopharyngeal NAAT sensitivity (54% [95% CI, 41%-67%]) [8] . Reported BAL NAAT sensitivity is even higher at >85% [7, 13] ; however, no subjects were diagnosed with COVID-19 by BAL in our study, potentially reflecting sampling bias. Furthermore, other than 1 patient with a chronic tracheostomy, no subjects diagnosed on LRT NAAT were intubated when the diagnostic specimen was obtained. We hypothesize that the vast majority of individuals with sufficiently severe COVID-19 to require intubation have high enough viral burden upon admission to detect SARS-CoV-2 on URT NAAT [14] .

Most patients diagnosed by LRT NAAT in our study were initially tested by URT late in disease, when URT viral load may be below the limit of detection of URT testing [3, 4] but high viral burden in the lungs may persist [13, 14] . Patients diagnosed ≥14 days into illness may no longer have transmissible disease and thus may not require transmission-based precautions [15] . However, it is still critical to confirm the diagnosis of COVID-19 to guide targeted treatment [16] , initiate contact tracing, and establish the 90-day recovery period during which reinfection is unlikely [15] . Screening of donor lungs prior to transplantation also necessitates the exclusion of SARS-CoV-2 infection by LRT testing [17] .

Operationally, the benefit of improved sensitivity with LRT NAATs must be weighed against the challenges of obtaining LRT testing. Less than a third of patients presenting with COVID-19 endorse sputum production, and many cannot provide expectorated sputum [18, 19] . BAL and sputum induction are additional means of LRT sampling. However, these procedures are considered aerosol generating and BAL is invasive, so they should be avoided unless clinically indicated [15] . Additionally, LRT NAAT turnaround time is often longer than for URT specimens, as most available FDA EUA testing platforms are not authorized for use with LRT specimens, requiring their referral to laboratories with internally validated SARS-CoV-2 LRT testing [10] . Longer test turnaround time leads to increased duration of transmission-based precautions in patients who ultimately test negative and greater use of personal protective equipment, with potential impact on hospital capacity. Validation and FDA authorization of LRT specimen types would help improve LRT NAAT availability and test turnaround time.

This analysis has several important limitations. First, it is a single-site study and may not reflect URT and LRT testing practices in other populations. Second, our study was nonrandomized; patients with COVID-19 may be more likely to have lower respiratory symptoms and produce sputum than PUI without COVID-19 infection [20] , so LRT NAAT may have been more likely to be performed on PUI with COVID-19. Last, we were unable assess the relative value of expectorated sputum and tracheal aspirate specimens in our full study population; however, all but 1 subject diagnosed on LRT NAAT had testing performed on expectorated sputum.

Our findings support the IDSA recommendation to perform repeated SARS-CoV-2 testing on LRT rather than URT specimens, when needed for diagnosis among patients with lower respiratory symptoms. Validation of SARS-CoV-2 tests on LRT specimen types should be prioritized to increase access to LRT testing.

Notes

Infectious Diseases Society of America guidelines on the diagnosis of COVID-19

Variation in false-negative rate of reverse transcriptase polymerase chain reaction-based SARS-CoV-2 tests by time since exposure

Clinical, laboratory, and radiologic characteristics of patients with initial false-negative SARS-CoV-2 nucleic acid amplification test results

Clinical sensitivity and interpretation of PCR and serological COVID-19 diagnostics for patients presenting to the hospital

Reliability of induced sputum test is greater than that of throat swab test for detecting SARS-CoV-2 in patients with COVID-19: a multi-center cross-sectional study

Comparison of throat swabs and sputum specimens for viral nucleic acid detection in 52 cases of novel coronavirus (SARS-CoV-2)-infected pneumonia (COVID-19)

Detection of SARS-CoV-2 in different types of clinical specimens

SARS-CoV-2 detection in different respiratory sites: a systematic review and meta-analysis

COVID-19 diagnostic clinical decision support: a pre-post implementation study of CORAL (COvid Risk cALculator)

SARS-CoV-2 reference panel comparative data

Radiological Society of North America expert consensus statement on reporting chest CT findings related to COVID-19. endorsed by the Society of Thoracic Radiology, the American College of Radiology, and RSNA-secondary publication

Prospective study comparing deep-throat saliva with other respiratory tract specimens in the diagnosis of novel coronavirus disease (COVID-19)

Evaluating the accuracy of different respiratory specimens in the laboratory diagnosis and monitoring the viral shedding of 2019-nCoV infections. medRxiv

Chronological changes of viral shedding in adult inpatients with COVID-19 in Wuhan, China

Centers for Disease Control and Prevention. Coronavirus (COVID-19). 2021. Available at

Guidelines on the treatment and management of patients with COVID-19

SARS-CoV-2 (coronavirus, 2019-NCoV): recommendations and guidance for organ donor testing

Characterization and clinical course of 1000 patients with coronavirus disease 2019 in New York: retrospective case series

Clinical features of patients infected with 2019 novel coronavirus in Wuhan, China

Sputum characteristics and airway clearance methods in patients with severe COVID-19

We thank the many Massachusetts General Hospital (MGH) infectious disease physicians who volunteered their time to review medical records of persons under investigation during the initial coronavirus disease 2019 (COVID-19) surge; the infection preventionists of the MGH Infection Control Unit; MGH microbiology laboratory staff who helped validate and implement severe acute respiratory syndrome coronavirus 2 test platforms; and our colleagues in pulmonary and critical care and internal medicine who cared for these patients during unprecedented circumstances.Disclaimer. The content is solely the responsibility of the authors and does not necessarily represent the official views of Harvard Catalyst, Harvard University and its affiliated academic health care centers, or the National Institutes of Health (NIH). Potential conflicts of interest. S. E. T. receives funding from the US Centers for Disease Control for COVID-19-related work. J. A. B. has received research support from Zeus Scientific, bioMérieux, Immunetics, Alere, and DiaSorin for unrelated research projects and has received consulting fees from Roche Diagnostics, T2 Biosystems, and DiaSorin. All other authors report no potential conflicts of interest.All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.