Diagnostic yield of repeat testing for SARS-CoV-2: Experience from a large health system in Los Angeles.

OBJECTIVE: To determine the diagnostic yield of repeat testing for SARS-CoV-2.
METHODS: A retrospective analysis was performed of all SARS-CoV-2 test results within the UCLA Health System between March 9th and April 29th, 2020. All patients with repeat test results were identified and those with discordant results were reviewed.
RESULTS: Between March 9th and April 29th there were 10,165 SARS-CoV-2 test results, of which 630 (6.2%) were positive. Among the 904 patients with repeat test results, 808 (89.4%) were initially negative and 96 (10.6%) were initially positive. Among the 808 patients with an initial negative test, 15 (1.9%) subsequently tested positive. Eleven cases with an initial negative SARS-CoV-2 test and without a known prior positive SARS-CoV-2 test were reviewed; 6 were employed as healthcare workers and 10 were positive on the second test.
CONCLUSIONS: We found a low diagnostic yield of repeat testing for SARS-CoV-2 in our health system. Repeat testing might prove useful in certain clinical scenarios, such as in healthcare workers, when symptoms develop after a negative test, and in hospitalized patients with a high clinical suspicion for COVID-19.

Background

In the United States, diagnostic testing capacity for SARS-CoV-2, the virus causing coronavirus disease 2019 (COVID-19), has increased (Adalja et al., 2020). A key public health intervention to mitigate the spread of COVID-19 is timely testing for early diagnosis of infections (Pan et al., 2020). Early in the epidemic, SARS-CoV-2 PCR testing was prioritized in clinical settings for symptomatic people who were: hospitalized, healthcare workers, or at high risk for complications (Centers for Disease Control and Prevention). Many diagnostic tests for SARS-CoV-2 have been developed and received Emergency Use Authorizations (EUA) by the U.S. Food and Drug Administration (FDA). Yet, concerns exist about the clinical sensitivity and specificity of those tests, particularly with regard to the frequency of false negative test results, which might lead missed infections and ongoing transmission (Woloshin et al., 2020).

Although repeat testing has been suggested as a way to increase diagnostic yield, there are sparse data on the yield of repeat testing (Lee et al., 2020, Omer et al., 2020). Moreover, repeat testing further limits supply and, among hospitalized patients, necessitates continued isolation and use of personal protective equipment, a limited commodity in the US (Livingston et al., 2020). Therefore, it is of high importance for clinicians to understand the diagnostic yield of repeat testing for SARS-CoV-2. We aimed to provide insight into the diagnostic yield of repeat testing for SARS-CoV-2 within a large health system in Los Angeles, and highlight the clinical scenarios of discordant results in inidividuals with an initial negative test.

Methods

We reviewed all SARS-CoV-2 tests within the UCLA Health System collected between March 9th and April 29th, 2020. Nasopharyngeal specimens were the preferred specimen collection method at our institution. During the time period of the study, nasal and oropharyngeal swabs were not used. For inpatients with concern for lower respiratory tract disease, clinicians could order lower respiratory tract specimens (e.g. – expectorated sputum, bronchoalveolar lavage (BAL), tracheal aspirate).

Testing was either performed at the UCLA clinical microbiology laboratory or was sent to a commercial laboratory (Quest Diagnostics, 2020). Testing at UCLA was done using three tests: 1) the U.S. Centers for Disease Control and Prevention’s (CDC’s) 2019-nCoV Real-Time Reverse Transcriptase (RT-) PCR Diagnostic Panel protocol (Atlanta, GA), which uses probes targeting the nucleocapsid gene (N1 & N2) of SARS-CoV-2 (Centers for Disease Control and Prevention); 2) the Diasorin Simplexa COVID-19 Direct RT-PCR (Diasorin Molecular LLC, Cypress, CA), which has two different targets, the S gene encoding for the spike glycoprotein of SARS-CoV-2 and the ORF1ab region encoding well-conserved non-structural proteins (DiaSorin Molecular, 2020); or 3) the TaqPath COVID-19 Combo Kit (Thermo Fisher Scientific Inc., Waltham, MA), which uses probes targeting ORF1ab, N, and S genes (ThermoFisher Scientific, 2020). Quest Diagnostics utilized the Quest SARS-CoV-2 rRT-PCR test (Quest Diagnostics, San Juan Capistrano, CA), which targets two regions of the N gene (N1 & N3) (Quest Diagnostics, 2020). All assays received EUA by the FDA for the qualitative detection of SARS-CoV-2 RNA in upper and lower respiratory specimens. Each specimen submitted for testing was tested using only one of the aforementioned diagnostic assays.

We defined multiple tests performed within the same patient as repeat tests. We excluded cases where the initial test was inconclusive. We report total number of tests, overall test positivity, and the number of patients with repeat tests and discordant results.

Results

In total, there were 10,165 SARS-CoV-2 tests performed and 6.2% (630/10,165) were positive. There were 906 patients with repeat test results, 2 were excluded due to inconclusive initial results. Among the 904 total cases, 808 (89.4%) were initially negative and 96 (10.6%) were initially positive. Those 904 patients accounted for 2,108 tests with an overall positivity of 9.9% (n = 209). The vast majority (93.7%) of those tests were performed on nasopharyngeal specimens and 65% were performed on the Diasorin platform. The number of tests and SARS-CoV-2 positivity by diagnostic assay and specimen type are shown in Table 1 .

Table 1

SARS-CoV-2 tests performed among 904 patients with repeat tests from March 9th to April 29th, 2020 at UCLA Health System in Los Angeles, CA, USA. Total tests and SARS-CoV-2 positivity are reported by diagnostic assay and specimen type.

		SARS-Cov-2 Test Positivityn (%)	Total Tests
All Tests		209 (9.9)	2108
Diagnostic Assay
	Diasorin Simplexa COVID-19 Direct RT-PCR	118 (8.6)	1370
	CDC 2019-nCoV RT-PCR Diagnostic Panel	56 (15.6)	358
	TaqPath COVID-19 Combo Kit	35 (9.2)	380
Specimen Type
	Nasopharyngeal	195 (9.9)	1977
	Bronchoalveolar lavage	9 (9.3)	97
	Sputum	2 (10.5)	19
	Tracheal specimen	3 (30.0)	10
	Other	0 (0)	5

Among the 96 patients with initial positive tests, 56 (58.3%) were repeated once, 24 (25.0%) were repeated twice, eight (8.3%) were repeated thrice, and eight (8.3%) were repeated four or more times. Repeat testing among positive cases was often done for infection control practices and for discharge planning. Among those with an initial positive test, 50% had a subsequent negative result (48/96); the median time between positive and negative tests was 16 days (interquartile range: 9 – 27 days). (Figure 1 , Panel A)

Figure 1

Panel (A) Repeat SARS-CoV-2 testing results among patients with an initial positive test result for SARS-CoV-2. Panel (B). Repeat SARS-CoV-2 testing results among patients with an initial negative test result for SARS-CoV-2.

Among the 808 patients with an initial negative test, 646 (80.0%) were repeated once, 110 (13.6%) repeated twice, 39 (4.8%) were repeated thrice, and 13 (1.6%) were repeated four times; fifteen (1.9%) subsequently tested positive. (Figure 1, Panel B) Clinical information was not available for one case. In three instances, patients were transferred from an outside facility where they had a positive SARS-COV-2 test. All transferred patients had an negative test on arrival, but a subsequent positive test within 24 hours.

There were 11 cases with an initial negative SARS-CoV-2 test, without a known prior positive SARS-CoV-2 test, and with reasons for repeat testing available for review. All of the initial tests were done on nasopharyngeal specimens and ten of the cases tested positive the second time. Six individuals with an initial negative test and a subsequent positive test were employed as healthcare workers (HCWs). The clinical scenarios for testing, as well as the specimen type, setting, diagnostic assay, and possible reasons for the initial negative result are presented in Table 2 . Testing prior to infection was likely in one case, as there were 20 days between the negative test and the development of symptoms. In three cases, pre-symptomatic testing was the likely explanation, as symptoms of acute illness developed after the initial test. In the remaining cases, possible reasons for the initial negative test result were inadequate specimen collection or inconsistent viral shedding.

Table 2

Review of clinical and testing data for patients within a large health system in Los Angeles, CA who underwent repeat testing for SARS-CoV-2, had an initial negative test and a subsequent positive test, and were without a known prior positive SARS-CoV-2 test result.

Initial Negative Test			Subsequent Positive Test
Clinical Scenario	Setting/ Specimen Type	Assay*	Scenario for repeat testing	Days after initial test	Setting / Specimen Type	Assay*	Possible reason for negative test
Presented to the hospital with 10-14 days of fever, fatigue, dyspnea, cough and imaging showing bilateral infiltrates	Emergency Department / NP	A	Worsening hypoxemia and high clinical suspicion for infection	2	Inpatient / NP + Sputum	A + B	Poor sampling or Inconsistent viral shedding
Patient with history of liver transplantation and end-stage renal disease who presented with 1 day of fever, myalgias, cough, and diarrhea	Outpatient / NP	A	Worsening fever and dyspnea	2	Inpatient / NP	A	Poor sampling or Inconsistent viral shedding
HCW with history of asthma, presented with 2 days of chest tightness and dyspnea, both relieved by inhalers; reported having a roommate with COVID-19	Emergency Department / NP	B	New symptoms - 3 days of fever and progressive cough with dyspnea	7	Emergency Department / NP	A	Pre-symptomatic
HCW who presented with 6 days of dry cough, post-nasal drip, and headache.	Outpatient / NP	A	Developed one day of fevers and chills	21	Outpatient / NP	A	Prior to infection
HCW reporting 1-day of rhinorrhea, cough, and chest tightness	Outpatient/NP	A	Developed symptoms of fatigue, arthralgias, fever, cough, and nausea for 4 days	6	Outpatient / NP	A	Pre-symptomatic
HCW, reporting 4-days of sinus congestion and sore throat	Outpatient/NP	B	Developed 1 day of fever, chills, diarrhea, and cough; domestic partner diagnosed with COVID19 after initial testing	4	Outpatient / NP	A	Pre-symptomatic
HCW reporting 2 days of sore throat and work exposure 3 days prior	Outpatient/NP	A	Developed 2 days of fever, myalgias, and arthralgias	3	Outpatient / NP	A	Poor sampling or Inconsistent viral shedding
HCW reporting 1-day of fever, chills, myalgias	Outpatient/NP	C	Ongoing fevers, cough, and fatigue	8	Outpatient / NP	C	Poor sampling or Inconsistent viral shedding
Patient was admitted to the hospital with fever and found to have septic arthritis of the knee. Initial testing done on admission	Inpatient/NP	A	Testing performed prior to surgery for epidural abscess	4	Inpatient / NP	A	Poor sampling or Inconsistent viral shedding
Patient admitted to the hospital after being found unconscious at home. Cardiac arrest in Emergency Room. Chest imaging showed bilateral infiltrates.	Inpatient/NP	A	Repeat testing done per infection control policy	0	Inpatient / Sputum	B	Poor sampling or Inconsistent viral shedding
Patient residing in a skilled nursing facility who presented with 1-day history of fever	Inpatient / NP	A	Persistent fevers; repeat tests done on hospital days 1, 5, and 9	9**	Inpatient / NP	A	Poor sampling or Inconsistent viral shedding

HCW – healthcare worker; NP – nasopharyngeal.

Assay A – Diasorin Simplexa COVID-19 Direct RT-PCR (Diasorin Molecular LLC, Cypress, CA), Assay B – CDC 2019-nCoV Real-Time Reverse Transcriptase (RT-) PCR Diagnostic Panel (CDC, Atlanta, GA), Assay C - TaqPath COVID-19 Combo Kit (Thermo Fisher Scientific Inc., Waltham, MA).

Positive result occurred on the fourth test, the second and third test results were negative.

Discussion

We found that repeat testing for SARS-CoV-2 had low diagnostic yield. Discordant results occurred in only 1.3% of patients with an initial negative test and without a prior diagnosis of SARS-CoV-2 infection. Consideration of the pretest probability alongside the estimated test performance can be used to guide repeat testing (Woloshin et al., 2020). For example, our report highlighted that repeat testing might prove useful in HCWs with ongoing exposures, as over half of those with initial negative tests and a subsequent positive test were employed as HCWs.

Other scenarios where repeat testing after a negative test might be useful include when symptoms of COVID-19 develop after a negative test, if inadequate specimen collection is suspected, and in hospitalized patients with high clinical suspicion for COVID-19 (Lee et al., 2020). Moreover, public health interventions to mitigate the spread of COVID-19 rely on timely testing and early diagnosis of infections, in order to inform isolation, contact tracing, and quarantine efforts, and repeat testing will be needed in these situations (Pan et al., 2020).

SARS-CoV-2 viral shedding can be variable and the timing related to exposure and symptom onset is important to test result interpretation (Wölfel et al., 2020, Zou et al., 2020). Available data suggest SARS-CoV-2 is typically detectable by RT-PCR in the nasopharynx by symptom onset and that viral loads are highest in the nasopharynx in the first week of the infectious course (Sethuraman et al., 2020, Wölfel et al., 2020, Zou et al., 2020). In our report, testing prior to infection was the likely explanation in one case and testing prior to symptom onset appeared to be the reason for three cases. In most cases, it was not possible to differentiate between variability in viral shedding and poor specimen collection. While the timing and quality of specimen collection are key, a false negative test result is still possible. Data regarding the clinical sensitivity remain quite limited. Preliminary reports suggest that nasopharyngeal specimens might have up to 27% false negativity and a systematic review estimated false negatives in 2% to 29% of specimens, although the quality of evidence was low (Arevalo-Rodriguez et al., 2020, Yang et al., 2020).

Retesting positive cases was recommended as a way to document viral clearance and remove isolation precautions (Centers for Disease Control and Prevention). In our report, approximately half of those with an initial positive test had a subsequent negative test. Prior reports have documented prolonged duration of SARS-CoV-2 positivity by RT-PCR, up to 6 weeks in some cases (Lan et al., 2020, Qi et al., 2020, Zhou et al., 2020). RT-PCR tests are characteristically sensitive, but are unable to discriminate between the presence of replicating virus and non-infectious remnants that contain nucleic acid targets. However, emerging evidence now suggests there is very low likelihood of infectivity, as measured by culturing viable SARS-CoV-2, if more than 8 days has elapsed since time of symptom onset (Bullard et al., 2020, Wölfel et al., 2020).

Our results should be considered in light of the following limitations. First, our study was performed within one health system and might not be generalizable to other settings. The report highlighted instances where repeat testing yielded discordant results in a clinical setting and might not be generalizable to repeat testing as part of a broader public health response. Second, we could not assess differences in tests by anatomic site, as we did not have paired specimens. Nevertheless, we provided data on repeat testing for SARS-CoV-2 within a large health system and highlighted cases where repeat testing improved diagnostic yield. In a clinical setting, repeat testing can be considered based on changes to clinical status and the pre-test probability of infection.

Funding

This work was supported by the National Institutes for Health (P.C.A. is supported by grant number T32MH080634 and D. G. is supported by grant number K08DA048163).

Ethical approval

The retrospective analysis was undertaken as a quality assurance project and did not meet research criteria for review by our institutional review board.

Declaration of interests

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

.

Centers for Disease Control and Prevention, 2020, Centers for Disease Control and Prevention, 2020 and Centers for Disease Control and Prevention (2020).