key: cord-0729849-83re9176
authors: Jefferson, T; Spencer, E A; Brassey, J; Heneghan, C
title: Viral cultures for COVID-19 infectious potential assessment – a systematic review
date: 2020-12-03
journal: Clin Infect Dis
DOI: 10.1093/cid/ciaa1764
sha: 60fd7edb5984890fe5dff2ebfbf79cde59964810
doc_id: 729849
cord_uid: 83re9176

OBJECTIVE: to review the evidence from studies relating SARS-CoV-2 culture with the results of reverse transcriptase polymerase chain reaction (RT-PCR) and other variables which may influence the interpretation of the test, such as time from symptom onset METHODS: We searched LitCovid, medRxiv, Google Scholar and the WHO Covid-19 database for Covid-19 to 10 September 2020. We included studies attempting to culture or observe SARS-CoV-2 in specimens with RT-PCR positivity. Studies were dual extracted and the data summarised narratively by specimen type. Where necessary we contacted corresponding authors of included papers for additional information. We assessed quality using a modified QUADAS 2 risk of bias tool. RESULTS: We included 29 studies reporting attempts at culturing, or observing tissue infection by, SARS-CoV-2 in sputum, nasopharyngeal or oropharyngeal, urine, stool, blood and environmental specimens. The quality of the studies was moderate with lack of standardised reporting. The data suggest a relationship between the time from onset of symptom to the timing of the specimen test, cycle threshold (Ct) and symptom severity. Twelve studies reported that Ct values were significantly lower and log copies higher in specimens producing live virus culture. Two studies reported the odds of live virus culture reduced by approximately 33% for every one unit increase in Ct. Six of eight studies reported detectable RNA for longer than 14 days but infectious potential declined after day 8 even among cases with ongoing high viral loads. Four studies reported viral culture from stool specimens. CONCLUSION: Complete live viruses are necessary for transmission, not the fragments identified by PCR. Prospective routine testing of reference and culture specimens and their relationship to symptoms, signs and patient co-factors should be used to define the reliability of PCR for assessing infectious potential. Those with high cycle threshold are unlikely to have infectious potential.

Effective prevention and management of SARS-CoV-2 infections relies on our capacity to identify those who are infected or potentially infectious. In the absence of predictive clinical signs or symptoms, the major means of detection is testing using Reverse Transcriptase quantitative Polymerase Chain Reaction (RT-qPCR) 1, 2, 3 The test amplifies genomic sequences identified in specimens, and is highly sensitive, being capable of generating observable signals from specimens containing minute amounts of matching genomic sequence. Amplification of genomic sequence is measured in cycle thresholds (Ct), each cycle being a cut off for positive detection. There may be a correlation between Ct values from respiratory specimens, symptom onset to test (STT) date and positive viral culture. Evidence suggests the lower the Ct value and the shorter the STT, the higher the infectious potential. 4 If this is so, we should be able to identify those with the highest infectious potential.

Identification of a whole virion (as opposed to fragments) and proof that the isolate is capable of replicating its progeny in culture cells is the closest we are likely to get to a gold standard. 5 RT-qPCR cannot distinguish between the shedding of live virus or of viral fragments with no infectious potential, and it cannot measure the quantity of live virus present in a person's excreta. Although viral culture is difficult, time consuming and requires specialised facilities it potentially represents the best indicator of infection and infectious potential. We, therefore, set out to review those studies attempting viral culture, regardless of specimen type tested. We investigated the probability of successful culture with time from symptom onset to test and cycle threshold. We also examined the relationship between specimen cycle threshold and infectious potential.

We searched four databases: LitCovid, medRxiv, Google Scholar and the WHO Covid-19 database, using the terms 'viral culture' or 'viral replication' and associated synonyms on 10 September 2020. For relevant articles, citation matching was undertaken and relevant results identified.

We included studies reporting attempts to culture SARS-CoV-2 and those which also estimated the potential infectivity of the isolates or observed tissue infection by SARS CoV-2 and related them to other clinical variables such as date of symptom onset to test and patient characteristics.

Isothermal methods of detection are not included in our review, as they do not provide a Ct value One reviewer extracted data for each study and a second reviewer checked the extraction. Heterogeneity and lack of detail of some of the reported data in the included studies prevented pooling. We tabulated data and summarised it descriptively by specimen: fecal, respiratory, M a n u s c r i p t 4 environment or mixed. Where possible, we also reported the duration of detectable RNA and the relationship of PCR cycle threshold and log 10 copies to positive viral culture.

Where necessary we contacted corresponding authors of the cited papers for additional information. We assessed quality using the QUADAS 2 risk of bias tool, simplified because the included studies were not designed as primary diagnostic accuracy studies. 6 Our methods are more fully described in our protocol (published on the 4th of July and updated on 5 th of October 2020). 7

We identified 145 possible articles for inclusion and after screening, 29 full texts were read and included (see PRISMA 8 flow chart - Figure 1 ). One unpublished study was not included as no permission was given by the authors. The included studies were published in 30 articles (see web appendix references w1-w29), four of which were in pre-print servers. The characteristics of each study are shown in Table 1 . All included studies were case series of moderate quality (Table 2 . Quality of included studies). We could not identify a protocol for any of the studies. All had been made public in 2020. We received five author responses regarding clarifying information (see Acknowledgments).

Nine studies assessed viral viability from fecal specimens positive for SARS-CoV-2 based on RT-PCR result W10, W11, W13, W17, W22, W23, W25-W27 One study reported infecting ferrets with stool supernatant; [w10] two reported visual growth in tissue [w19, w22[ and four reported achieving viral replication [w13, w23, w24, w26] . In one further study, methods were unclear. W28

Seventeen studies reported attempting viral isolation and culture from respiratory specimens [W3, W4, W6-10, W13-16, W18, W21-23, W26, W27] One study successfully cultured 26/90 nasopharyngeal specimens: positive cultures were observed only up to day eight post-symptom onset; [w7] another study obtained cultures from 31/46 nasopharyngeal and oropharyngeal specimens. [w3] The largest study came from the La Scola group publications [w15] with positive cultures of 1,941 from 3,790 specimens. Another study of UK health care workers during a period of low viral circulation isolated SARS Cov-2 from 1/19 specimens. [w5] Two more studies reported a clear correlation between symptoms onset, date of sampling, Ct and likelihood of viral culture. [w18, w21] A c c e p t e d M a n u s c r i p t 5 One study [w14] of nasopharyngeal specimens from 638 patients aged <16 years reported achieving culture from 12 (52%) of the 23 who tested positive for SARS CoV-2 with a Ct of around 28. Gniazdowski [w8] assessed RNA and infectious virus detection in 161 nasopharyngeal specimens from hospitalised Covid-19 patients. Positive culture was associated with Ct values of 18.8 ± 3.4 (median 18.7); negative culture was associated with mean Ct values 27.1 ± 5.7 (median 27.5). Over 90% of the virus isolates were obtained from specimens with a Ct value below 23

Basile [w4] reported 24% culture positivity, with specimens significantly more likely to be positive from ICU. A report by the Korean Centres for Disease Control failed to grow live viruses from 108 respiratory specimens from "re-positives" i.e. people who had tested positive after previously testing negative. [w12] Ladhani [w16] and colleagues reported a successful culture rate of 87/158 RT-PCR positive nasopharyngeal specimens from six nursing homes in London.

Two possible (the text is unclear) positive cultures were obtained from 95 environmental specimens in one study that assessed aerosol and surface transmission potential of SARS-CoV-2 [w20] . No viruses could be grown from specimens from seven areas of a large London hospital from specimens with a cut-off RT-PCR Ct > 30. [w29] Ahn and colleagues [w1] failed to grow live virus from an unspecified number of air specimens from isolation rooms of patients with severe Covid-19, but were able to grow virus from swabs of handrails, and the external surfaces of intubation cannulae.

Some studies labelled as mixed source specimens are also reported by indvidual specimen in this text.

Eight studies reported viral culture from mixed sources: 12 oropharyngeal, nine nasopharyngeal and two sputum specimens [w9] , one stool specimen and an unreported number of other specimens [w10] , from saliva, nasal swabs, urine, blood and stool collected from nine Covid-19 and a possible M a n u s c r i p t 6 specimen stool culture [[w23] , nine nasopharyngeal, oropharyngeal, stool, serum and urine specimens [w13] , seven sputum specimens, three stool specimens and one nasopharyngeal specimen of 11 patients. [w26] . In this study all specimens had been taken within 5 days of symptom onset and there was a relationship between copy thresholds and cytopathic effect observed in infected culture cells.

Kim and colleagues reported no viral growth from an unclear number of serum, urine and stool specimens, despite these specimens being collected soon after admission [w11] . Lu and colleagues also reported no viral growth, however their specimens were from 87 cases tested "re-positive". [w17] One study [w27] reported 21 positive cultures from from naso-pharyngeal specimens of 19 hospitalised patients in Singapore but no growth from specimens with a Ct value >30, or collected >14 days after symptoms onset. No culture was achieved from the urine or stool specimens.

In one study by Andersson [w2] et al 20 RT-PCR positive serum specimens from 12 individual patients were selected at random from a Covid-19 specimen bank at 3 to 20 days following onset of symptoms. None of the 20 serum specimens produced a viral culture.

One study on alveolar specimens from 68 elderly deceased reported postmortem studies on lung tissues from six cases were available for viral isolation. The evaluation showed viable SARS-CoV-2 in all six cases -in one case on day 26 from symptom onset. [w6] Duration of RNA viral detection Table 3 shows that nine studies report on the duration of viral RNA detection as assessed by PCR for SARS-CoV-2 RNA. [w7, w8, w10, w12, w13, w21, w24, w25, w27] All nine studies reported RNA detection for longer than 7 days. Young et al [w27] reported that SARS-CoV-2 was detectable from nasopharyngeal swabs by PCR up to 48 days after symptom onset.

The live viral culture time window was much shorter than for viral RNA identification, ranging from less than 8 days from symptom onset to test [w23] and Ct < 24 [w7] . Median duration of viral RNA identification in culture was 4 days (InterQuartile Range: 1 to 8) [w21] .

A c c e p t e d M a n u s c r i p t 7 The relationship between RT-PCR results and viral culture of SARS-CoV-2 Table 4 shows that ten studies analysed the relationship between Ct values and the possibility of culturing live virus [w4, w5, w7, w8, w9 w15, w16, w21, w23, w27] and three quantified the mean log copies of detected virus and live culture [w9, w14, w18] . All reported that Ct were significantly lower and log copies were significantly higher in those with live virus culture. Five studies reported no growth in specimens based on a Ct cut-off value [w5, w7, w9, w16, w27] ranging from CT > 24 [w7] to 35 [w15] .

The estimated probability of recovery of virus from specimens with Ct > 35 was 8.3% (95% CI: 2.8% to 18.4%) [w21 . All donors above the Ct threshold of 35 (n=5) producing live culture were symptomatic.

In six London nursing homes there was no correlation between Ct values and symptoms in either residents or staff, [w16] although nearly 50% of both categories were asymptomatic.

One study [w9] reported different cut-off thresholds depending on the gene fragment analysed 34 . No growth was found for the NSP 12 fragment at Ct > 31.5, whereas the value was higher for the N gene fragment (>35.2).

The odds for culturing live virus decreased by 0.64 for every one unit increase in Ct (95%CI 0.49 to 0.84, p<0.001) [w7] ; another study [w21] reported similar results in line with empirical evidence of an increased Ct of 0.58 per day since symptoms started. 9

The studies in this review attempted, and some successfully achieved, culture of SARS-CoV-2 in the laboratory, using a range of different specimens. There is evidence of a positive relationship between lower cycle count threshold, likelihood of positive viral culture and date of symptom onset. 10 This is seen clearly in the two studies assessing the infectious potential of "re-positives", i.e. COVID-19 cases who had been discharged from hospital after testing negative repeatedly and who then tested positive again after discharge: Lu 2020 [w17] , Korean CDC [w12] .

Lu and colleagues considered four hypotheses for the origin of "re-positives" [w17] . On the basis of their evidence they discarded re-infection and latency as explanations, and concluded that the most plausible explanations were either contamination of the specimen by extraneous material or M a n u s c r i p t 8 identification in the specimen of minute and irrelevant particles of dead SARS-CoV-2 representing virus long neutralised by the immune system.

Rapid expansion in testing capability requires training protocols and precautions to avoid poor laboratory practice which may not be possible in the time pressure of a pandemic. The evidence in this review shows that those with high cycle threshold are unlikely to have infectious potential.

Interpreting the results of RT-PCR requires consideration of patient characteristics such as symptoms and their severity, contacts history, presence of pre-existing morbidities and drug history, the cycle threshold value, the number of days from symptom onset to test and the specimen donor's age. 11 12 Several of our included studies assessed the relationship of these variables and there appears to be a time window during which RNA detection is at its highest with low cycle threshold and higher possibility of culturing a live virus, with viral load and probability of growing live virus of SARS-CoV2 peaking much sooner than that of SARS CoV-1 or MERS-CoV. 11 We propose that further work should be done on this with the aim of constructing an algorithm for integrating the results of PCR with other variables, to increase the effectiveness of detecting infectious patients.

PCR should be continuously calibrated against a reference culture in Vero E6 cells in which cytopathic effect has been observed [w6] . Confirmation of visual identification using methods, such as an immunofluorescence assay may also be needed to aid diagnosis. 13 Henderson and colleagues have called for a multicentre study of all currently manufactured SARS-CoV-2 nucleic acid amplification tests to correlate the cycle threshold values on each platform for patients who have positive and negative viral cultures. Calibration of assays could then be done to estimate virus viability from the cycle threshold with some certainty. 14 Ascertainment of infectious potential is all the more important as there is good evidence of viral RNA persistence across a whole range of different viral diseases with little or no infectious potential in the post infectious phase of MERS, 15 measles, 16 other coronoviridae, HCV and a variety of animal RNA viruses. 17 In one COVID-19 (former) case, viral RNA was detectable until day 78 from symptoms onset with a very high Ct 18 but no culture growth, implying a lack of infectious potential. SARS CoV-2 methods of cell culture vary and to our knowledge have not been standardised. Methods vary depending upon the selection of the cell lines; the collection, transport, and handling of and the maintenance of viable and healthy inoculated cells. 19 We therefore urgently recommend the development of standard culture methods and external quality assessment schemes for laboratories offering testing for SARS CoV2. 20 21 If identification of viral infectious potential relies on A c c e p t e d M a n u s c r i p t 9 visual inspection of cytopathogenic effect, then a reference culture of cells must also be developed to test recognition against infected cells. Viral culture may not be appropriate for routine daily results, but specialized laboratories should use viruses as controls, perform complete investigations when needed, and store representative clinical strains whenever possible. 22 Current evidence is too limited to establish the feasibility of generating a universal cycle threshold value as this may change with circumstances (e.g. hospital, community, cluster and symptom level), laboratory methods, so more information is urgently needed 23 .

We suggest the WHO produce a protocol to standardise the use and interpretation of PCR and routine use of culture or animal model to continuously calibrate PCR testing, coordinated by designated Biosafety Level III laboratory facilities with inward directional airflow. 24 Further studies with standardised methods 25 and reporting are needed to establish the magnitude and reliability of this association.

The results of our review agree with the scoping review by Byrne and colleagues on infectious potential periods 26 and those of the living review by Cevick and colleagues 11 . The authors reviewed 79 studies on the dynamics, load and RNA detection for SARS CoV-1, MERS and SARS CoV-2 from symptoms onset. They concluded that although SARS-CoV-2 RNA identification in respiratory (up to 83 days) and stool (35 days) can be prolonged, duration of viable virus is relatively short-lived (up to a maximum of 8 days from symptoms onset). Those results are consistent with Bullard et al who found no growth in specimens with a cycle threshold greater than 24 [w7] or when symptom onset was greater than 8 days, and Wölfel et al [w23] who reported that virus could not be isolated from specimens taken after day 8 even among cases with ongoing high viral loads. The review by Rhee and colleagues reaches conclusion similar to ours. 10 The importance of symptom onset and reported PCR threshold is shown in a study that collected test data during a prospective household transmission study. The authors found that Ct values were lowest soon after symptom onset and correlated with time elapsed since symptom onset (within 7 days after symptom onset, the median Ct value was 26.5 compared with a median of 35.0 21 days after onset). Ct values were significantly higher among those participants reporting no symptoms, and lower in those reporting upper respiratory symptoms at the time of specimen collection. 28 The evidence is increasingly pointing to the probability of culturing live virus being related to the amount of viral RNA in the specimen and, therefore, inversely related to the cycle threshold. Thus, detection of viral RNA per se cannot be used to infer infectiousness. Duration of excretion may also be linked to age, male gender and possibly use of steroids and severity of illness. M a n u s c r i p t 10 Our review is limited by the lack of standardised reporting and lack of standard testing methods amongst the included studies 20 . Ct threshold reporting was inconsistent, preventing pooling or further in-depth analysis of the data, and insufficient clinical details were reported to define the possible role of asymptomatics or pre-symptomatics in transmission. The included studies were case reports or case series with a mixture of laboratory and clinical data, and variable in reporting the relation between donor characteristics and PCR results.

We may have missed some studies or new studies as they are published and we aim to update this review with emerging evidence.

The evidence gathered in this review points to a relationship between the time from collection of a specimen to test, cycle threshold, and symptom severity. We recommend that a uniform international standard for reporting of comparative SARS-CoV-2 culture with index test studies be produced. Particular attention should be paid to the relationship between the results of testing, clinical conditions and the characteristics of the source patients, description of flow of specimens and testing methods. Defining cut off levels predictive of infectious potential 27 should be feasible and is necessary for diagnosing viral respiratory infections using molecular tests. M a n u s c r i p t 11

Drs Susan Amirian, Siyuan Ding, Long Rong, Sravanthi Parasato and Bernard La Scola provided additional information for this review. Dr Maryanne DeMasi helped with reference identification. All data included in the review are from publications or preprints.

The review was partly funded by NIHR Evidence Synthesis Working Group project 380.

The views expressed in this commentary represent the views of the authors and not necessarily those of the host institution, the UK NHS, the NIHR, or the Department of Health and Social Care. The views are not a substitute for professional medical advice. It will be regularly updated see the evidence explorer at https://www.cebm.net/evidence-synthesis/transmissiondynamics-of-covid-19/ for regular updates to the evidence summaries and briefs.

Tom Jefferson is a Senior Associate Tutor and Honorary Research Fellow, Centre for Evidence-Based Medicine, University of Oxford. Disclosure statement is here TJ was in receipt of a Cochrane Methods Innovations Fund grant to develop guidance on the use of regulatory data in Cochrane reviews (2015-018). In 2014-2016, TJ was a member of three advisory boards for Boehringer Ingelheim. TJ was a member of an independent data monitoring committee for a Sanofi Pasteur clinical trial on an influenza vaccine. TJ is occasionally interviewed by market research companies about phase I or II pharmaceutical products for which he receives fees (current). TJ was a member of three advisory boards for Boehringer Ingelheim (2014-16). TJ was a member of an independent data monitoring committee for a Sanofi Pasteur clinical trial on an influenza vaccine Table 1 . Characteristics of included studies Figure 1 PRISMA flow chart.

A c c e p t e d M a n u s c r i p t M a n u s c r i p t 24 subsequent Jaafar et al letter.

between 4 and 10 h after sampling and kept at + 4 °C before processing. After centrifugation they were incubated at 37 °C. They were observed daily for evidence of cytopathogenic effect. Two subcultures were performed weekly and scanned by electron microscope and then confirmed by specific RT-PCR targeting E gene. 137 swabs and 59 serum specimens from 70 "re-positive" cases to assess the immunological and virologic characteristics of the SARS-CoV-2 "re-positive" cases. From 23 January, hospital

No cultures were positive "Re-positive" cases are unlikely to be infectious as no intact RNA single helix was detected or viral M a n u s c r i p t 25 137 swabs (51 nasopharyngeal, 18 throat and 68 anal) discharge dischargees followed a strict isolation protocol living (for example) in single dedicated hotel rooms and went home only when nucleic acid tests were negative on both respiratory tract and digestive tract specimens. Specimens (nasopharyngeal, throat and anal swabs), were collected for RT-PCR diagnosis at 7 and 14 days after discharge. Culture was carried out by inoculating Vero E6 cells with patient specimen. CPE were observed daily at 7 days with a second round of passage.

RT-PCR diagnosis was carried out on RNA using three RT-PCR kits to conduct nucleic acid testing, in an attempt to avoid false negatives.

Ct varied from 29 to 39 depending on gene and kit isolated grew. Specimens included in this study included those positive for SARS-CoV-2 by RT-PCR from day of symptom onset (Day 0) up to 21 days post symptom onset.

SARS-CoV-2 Vero cell infectivity of respiratory specimens from SARS-CoV-2 positive individuals was only observed for RT-PCR Ct < 24 and symptom onset to test of < 8 days.

[w8]

Patients that received repeated testing with longitudinal positive results were tested within a time frame that ranged from less than one day to more than 45 days On average, it took 45 days (range: 8 to 82 days) from the initial symptom onset date to testing re positive after discharge. (Based on 226 cases symptomatic at the time of initial confirmation)

This may indicate duration of viral RNA detection over a long period of time and inconsistently.

These data may not be comparable with information from studies specifically observing the duration of viral RNA detection as an outcome.

Time to retesting positive via PCR is reported, among this specific group of individuals who retested positive by PCR.

Duration of SARS-CoV-2 detection by RT-PCR was 7 to 22 days First 12 identified patients in the US. Respiratory specimens collected between illness days 1 to 9 (median, day 4) All patients had SARS-CoV-2 RNA detected in respiratory specimens, typically for 2 to 3 weeks after illness onset.

Mean duration of fever was 9 days. Two patients received a short course of corticosteroids.

[w21]

SARS-CoV-2 viral load identified that the level of SARS-CoV-2 RNA in the URT was greatest around symptom onset, steadily decreased during the first 10 days after illness onset and then plateaued up to day 21

Probability of culturing virus declined to 8% in specimens with Ct > 35 and to 6% 10 days after onset;

The viral load was higher in feces than in respiratory specimens collected at multiple time points ( 

Prediction models for diagnosis and prognosis of covid-19: systematic review and critical appraisal

Report of the WHO-China Joint Mission on Coronavirus Disease

Predicting infectious SARS-CoV-2 from diagnostic s

Traditional and Modern Cell Culture in Virus Diagnosis

QUADAS-2: a revised tool for the quality assessment of diagnostic accuracy studies

Analysis of the evidence of transmission dynamics of COVID-19 Protocol for a scoping evidence review. Jefferson T, Plüddemann A

Preferred reporting items for systematic review and metaanalysis protocols (PRISMA-P) 2015 statement

Temporal, Spatial, and Epidemiologic Relationships of SARS-CoV-2 Gene Cycle Thresholds: A Pragmatic Ambi-directional Observation

Duration of SARS-CoV-2 Infectivity: When is it Safe to Discontinue Isolation?

SARS-CoV-2, SARS-CoV-1 and MERS-CoV viral load dynamics, duration of viral shedding and infectiousness: a living systematic review and meta-analysis

To Interpret the SARS-CoV-2 Test, Consider the Cycle Threshold Value

Traditional and Modern Cell Culture in Virus Diagnosis

The perplexing problem of persistently PCR-positive personnel

Environmental Contamination and Viral Shedding in MERS Patients During MERS-CoV Outbreak in South Korea

Prolonged persistence of measles virus RNA is characteristic of primary infection dynamics

Human Coronaviruses Associated with Upper Respiratory Tract Infections in Three Rural Areas of Ghana

Temporal, Spatial, and Epidemiologic Relationships of SARS-CoV-2 Gene Cycle Thresholds: A Pragmatic Ambi-directional Observation

Point: is the era of viral culture over in the clinical microbiology laboratory?

College of American Pathologists (CAP) Microbiology Committee Perspective: Caution must be used in interpreting the Cycle Threshold (Ct) value

International external quality assessment for SARS-CoV-2 molecular detection and survey on clinical laboratory preparedness during the COVID-19 pandemic

Point: is the era of viral culture over in the clinical microbiology laboratory?

Challenges and Controversies Related to Testing for COVID-19

A c c e p t e d M a n u s c r i p t 31 specimens using a method adapted from one previously used to culture influenza virus. On day 0 and after 5-7 days, cell supernatants were collected, and RT-qPCR to detect SARS-CoV-2 performed as described above. Specimens with at least one log increase in copy numbers for the E gene (reduced Ct values relative to the original specimens) after 5-7 days propagation in cells compared with the starting value were considered positive by viral culture.Key: STT = symptom onset to test date.A c c e p t e d M a n u s c r i p t 32 A c c e p t e d M a n u s c r i p t 37 

A c c e p t e d M a n u s c r i p t Figure 1