We covered a lot of ground during our Diagnostic for COVID-19 virtual meeting, that is now available on-demand. Here are the questions and answers that we were not able to cover during the live event.
Be sure to Pre-Register for our next event, Therapeutic Pipeline for COVID-19 that is scheduled for September 17, 2020.
Comparing Diagnostic Methods for COVID-19: Practical and Statistical Considerations
Timothy J. O’Leary, MD, PhD, Adjunct Professor, Pathology, University of Maryland School of Medicine; Former Chief Research and Development Officer, Veterans Affairs
If you begin with a very sensitive test, like a qRT-PCR test, and a new test falls outside of the 10% non-inferiority stat, do you recommend not using the new test, even for a new application, like rapid diagnostics?
The decision on test sensitivity and specificity should not, in my opinion, be based on any arbitrary standard, but depends on the context for its use. Consider, for example, a test for covid19 that is applied to a population which will be boarding a ship having 4000 passengers and crew. Everyone has spent the last two weeks in an area with 1% disease prevalence. A-priori, therefore, we would expect 40 individuals to have active virus. If we give a test with 80% sensitivity, we reduce that to 8, on average; that does not seem good enough to me. However, if it were 100 people boarding a passenger plane, then we are down to 0.2% chance of having someone who can transmit on board, which seems a lot more acceptable. I think that we need to consider speed, cost, population prevalence and specific intended use, decide what our tolerance for risk is, and decide on that basis.
Multiplatform Comparisons of Molecular Tests for COVID-19: Predicting Differences in Case Detection
David Hillyard, MD, Medical Director, Molecular Infectious Diseases; Professor, University of Utah School of Medicine
Will multiplex testing for multiple respiratory viruses become the standard of care clinically?
It is already for Flu. It will be the standard of care for COVID/respiratory virus testing in the fall
Have you seen the paper from Justin Lessler (Johns Hopkins) that indicates rtPCR is essentially unable to detect COVID infections until after 5 days post infection? Also why are you not comparing digital PCR for COVID which has at least a LOG better LOD than rtPCR based methods?
Five days has been an estimate for some time; however, it also depends on concentration of the initial inoculum, site of infection, and the host. Digital PCR is not reliably a log more sensitive than standard PCR. We know that based on using it to validate our proficiency materials which have been tested both ways.
Come flu season this fall, being able to distinguish people who have cold/flu symptoms, the speaker mentioned having to do a multiplex or companion tests for flu and RSV in addition to sars coV2. My question was what about metapneumovirus, which can be indistinguishable from RSV but equally damaging in the very young, and other cold viruses who will likely make up the majority of people coming in with cold symptoms. between adeno and rhino viruses there is a lot of them to test for, perhaps too many?
Most Co-Testing will be FluAB/ COVID, FLUAB/RSV will also be used/. Full panels with COVID using tests like BioFire will routinely be used as well. We are unlikely to have a high throughput large panel COVID including assay this year (it’s hard to do).
Is there a multi-Fluor system which may be appropriate, seeing as this will become a major issue?
There are many many many multifluor systems that will address the issues above. More than 5 Fluors gets dicy.
The Critical Role of Antigens in COVID-19 Serological Test Development
Teng Peng, PhD, Technical Application Manager, ACROBiosystems
Is the glycosylation profile of the native Corvid19 virus established?
No, we have not done any glycosylation profile in house of the native Corvid19 virus nor the SARS-Cov-2 recombinant protein products.
Or is the presumption that HEK293 glycosylation is correct only a presumption? Certainly, it is better than the insect expression, but is it the same as in humans? HEK293 is not exactly a “normal cell!”
The data of glycosylation profile in the presentation slide only suggested that the glycosylation profile of the SARS-Cov-2 recombinant protein expressed in HEK293 cells is closer to the native version compared to that expressed in insect cells. HEK293 cell line is not exactly a “normal cell” but still carrying the PTM system similar to normal human cells.
Development, Implementation, and Regulation of Diagnostic Methods at the State Level – New York Perspective
Kirsten St. George, PhD, Director, Virology Laboratory; Chief, Viral Diseases, Wadsworth Center; Clinical Professor, School of Public Health, Biomedical Sciences, SUNY
Are the staff performing testing being tested for exposure to SARS-CoV-2?
Staff at the Wadsworth Center are monitored daily on arrival at work. They are questioned with regard to signs and symptoms of COVID-19, and have their temperature checked. They have also been offered serology testing for SARS-CoV-2 antibodies.
For the selection of a sample, what about the sensitivity test of sputum against other types of sample?
This is an interesting question. I noticed speakers during two of the other presentations referring to limits of detection for tests listed on the FDA web site that have been granted EUA. The New York SARS-CoV-2 real-time RT-PCR assay is noted as having a limit of detection (LOD) of 25gc/reaction. What is not usually noticed unless you read the IFU carefully, is that this refers to the LOD for sputum. At the time the amendment validation data for this assay was submitted (among the earliest of all EUAs), facilities were advised to submit validation data for the most difficult or challenging specimen type they intended to test. The LOD for upper respiratory tract swabs in this assay is about 5gc/reaction. So, in our experience, the LOD for sputum is about 5-fold higher than that for upper respiratory swabs.
The Rapidly Evolving Role of Antibody Testing for SARS CoV-2
Co-presenter: Patricia Slev, D(ABCC), Section Chief Immunology Division, Medical Director Core Immunology; Medical Director Microbial Immunology, Medical Director Serologic Hepatitis and Retrovirus Laboratory; Associate Professor, Pathology, University of Utah
What little information on the antibody reagents used seems hidden under the guise of commercial secrecy, and we are only talking about anti-human-IgG, IgM or IgA antibodies. No data on specificity, selectivity, titre or source of reagents. As this is now considered to be basic information to allow reproducibility of pre-clinical assays, it seems to add to the uncertainty of serotests data that these data are hidden from the community. Especially as we are probably looking at the use of 100s of millions of such tests (I think of gate checks for airlines, as a trivial example).
Do you have any thoughts on this issue?
Commercial protection of antibody and reagents used in an assay is not new for diagnostics tests. Trade secrets associated with assays does not mean you cannot verify the performance characteristics such as sensitivity and specificity of a diagnostic test. Laboratories “validate” commercial assays on a regular basis using established guidelines and rigorous studies. The problem is when these assays are not validated properly, if at all.
What is the reimbursement program for serology tests? reimbursement program(s)?
There are CPT codes established for COVID 19 serology and molecular testing for reimbursement.
Co-presenter: Elitza Theel, PhD, Associate Professor, Laboratory Medicine, and Pathology, Mayo Clinics
Can she elaborate on non-antibody-mediated immunity? Role of immune redundancy when Abs may not offer full protection.
There is quite a bit of work focused on the cellular immune response and its role in COVID-19 recovery and immunity and we’ll be learning much more with respect to this in the coming months.
More of a comment: Non-“useful” binding Abs, divided in two main groups, merely binding and antagonistic Abs, the latter enhance disease. Useful Abs have a variety of different mechanisms, e.g. prevention of infection, aid in clearance and interfering with viral functioning. Macaque study was not statistically significant, only involved low single digit numbers
The study was not statistically significant; however, it was among the first to show what happens after re-infection making it quite relevant in the field.
Is there a view on immunity passports etc?
Given that we are unable to definitively state whether or not the presence of antibodies provides protective immunity, and for how long any protective immunity may last, the utility of ‘immunity passports’ seems quite limited and potentially miss-leading.
Considering most Ab EUA test is qualitative or semi-quantitative, is there a necessity for a quantitative test? And if yes, what is the analytical detection limit is useful?
A quantitative test would be worthwhile once there is a correlate of immunity-related to an antibody level or if these assays were correlated to neutralizing antibody levels. Otherwise, the clinical role of just quantifying antibodies is limited.
For those pts who make no or low levels of nAbs, is the clinical course of disease different?
This is still being looked into. However recent studies suggest that individuals who are asymptomatic may produce lower levels and/or see a more rapid decline in antibody levels over time (https://www.nature.com/articles/s41591-020-0965-6)
Interpretation of COVID-19 Diagnostic Test Results: A Clinician’s Perspective
Ralph Rogers, MD, Assistant Professor of Medicine, Clinician Educator, Infectious Diseases, Warren Alpert Medical School of Brown University
Wouldn’t dexamethasone be possible to be applied in this case as a general inflammation inhibitor?
I’m not sure which case this question is referring to. In a more general sense, though prior NIH consensus guidelines suggested that corticosteroids should not be routinely used for most patient populations with severe SARS-CoV-2 infection (see guidelines here), recent clinical data has emerged which is more supportive of the use of corticosteroids in hospitalized patients with severe SARS-CoV-2 pneumonia who require oxygen supplementation (see preprint here).
How will you approach testing for the differential diagnosis of respiratory infection when the seasonal cold and influenza season begins in the autumn and winter?
Ideally, multiplex PCR panels testing for respiratory pathogens on currently available platforms may be revised to include SARS-CoV-2 targets. This may aid in making a diagnosis in hospitalized patients with a severe respiratory infection, but will likely be too expensive/impractical for use elsewhere, and will likely face similar challenges with clinical sensitivity as do the currently available SARS-CoV-2 molecular assays. The development of public health strategies for efficient diagnostic triage (and isolation) of patients with symptomatic respiratory infection will be challenging, and may vary depending on local prevalence of COVID-19 and resources available for containment or mitigation strategies.
Do we have any comparison of negative covid-molecular tests follow-up with quality (ELISA) serotests so that molecular negative could be compared with true sero-positives?
This has not yet been reported on by any large-scale systematic analysis (to my knowledge). Though I agree that this proposed analysis would likely provide some measure of the clinical sensitivity of molecular assays, the presence of SARS-CoV-2 antibody may not necessarily be a perfect “gold-standard” test. There remains the possibility of a false-positive antibody test (analytic false positive or clinical false positive due to SARS-CoV-2 antibody presence related to a prior COVID-19 infection rather than the symptomatic illness with negative molecular testing in question) or a false negative antibody test (not all patients with COVID-19 will subsequently develop a measurable/persistent antibody response).
Did this patient with persistent positive have COVID antibodies?
Are persistent positives treated as infected? The CDC has mSRNA subgemonic levels with RT PCR with viral cultures.
Most patients with persistent positive PCR testing are presumed not to have a persistent clinical infection since most have had complete resolution of their clinical symptoms (with repeat PCR testing obtained solely for infection control rather than clinical purposes). A smaller number of patients may have persistent/recurrent clinical symptoms (and persistently positive PCR testing) after an initial COVID-19 infection, though it is often unclear if the symptoms in an individual patient are due to a persistent productive viral infection, a post-infectious inflammatory/thrombotic/fibrotic complication, or an unrelated comorbid illness (as may have been the case with the third case discussed). With this last case, I was hoping to emphasize the point that persistently positive qualitative molecular testing is very non-specific with regards to differentiating these potential causes of ongoing/recurrent clinical symptoms.
Home Collection of Nasal Specimens – Lowering the Barrier to SARS-CoV-2 Testing and Expanding Access to Care
Karen A. Heichman, PhD, Senior Program Officer, Diagnostic Development, Innovative Technology Solutions, Global Health Division, Bill & Melinda Gates Foundation
Where would an individual obtain the self-collection kit? For example, what about them being available in pharmacy chains right on the shelf?
Self-collection kits are not yet available off-the-shelf. FDA would like them to be restricted to people with a test prescription. however, that could change in the future as more information becomes available. our grantee, audere, is planning to submit a self-collection kit to the fda for eua approval. that collection kit (or one similar to it with the same types of components) could eventually be used ubiquitously for public health purposes, and could be associated with specific tests and labs. the components of the home collection kit are commercially available to be put together now.
Is Dr Heichman aware of any lateral flow rapid tests for antigens? In the UK, you order online and it comes through the post next day. Interestingly, these self-administered tests without a clinician present are widely used in the UK. We’ve been using them since March and they can just go in the post.
Yes, there are several lateral flow assays that are being developed for sars-cov-2 testing. the performance of most of these are still being determined. our grantee, path, is poised to assess the performance of these using a standard reference method, standards, and characterized patient specimens. the results will be published and made available to the community. there are also efforts underway to assess the newly emerging rapid diagnostic tests by other experts in global health.
Most respiratory illnesses will be caused by more innocuous viruses; how do the panelist suggest these negative tests caused by these other pathogens be handled? How do they expect co-infections would affect the testing and disease transmission?
Yuan-Po Tu: Co-infection with more than one respiratory virus can occur. In adults with a normal immune system, I tend only to worry and test for influenza and COVID. Influenza has a treatment. Influenza and CVOID have the largest impact on hospitalization, morbidity and mortality.
The other respiratory virus in a Biofire panel do not have therapy and are usually not major drivers of hospitalization. Treatment is symptomatic for all viral infections except for influenza.
Karen Heichman: That is a complicated question and a complicated problem. Some tests and platforms may be capable of being applied to a panel of pathogens with common symptoms, such as an upper respiratory panel, which might include combinations such as influenza A/B, RSV and SARS-COV-2. Moving toward the future, I would expect to see several of these tests emerge. there are also algorithmic approaches to testing, which involves testing sequentially, based on temporal and geographic prevalence. for instance, it might make sense to test for influenza first, and if the test is negative, then test with less common viruses. however, there have been reports that co-infection of sars-cov-2 with other upper respiratory pathogens is not uncommon, so taking a parallel approach rather than a sequential approach might make the most sense for certain environments.
Are PCR-based test sensitive enough or do these tests need improvement or a different approach?
Timothy O’Leary: Are PCR-based tests good enough? It would be nice to have tests with higher sensitivity, to confirm cases that look like covid19 but have returned negative results. A test that we created for botulinum toxin would probably be several orders of magnitude more sensitive than RT-PCR if adapted for covid19. However, it might well identify “positives” who would not be capable of transmitting infection. In addition, introducing a more complex algorithm into the testing lab might do more harm than good if it reduced the total number of people that we can test. I personally suspect that our efforts should be directed to using the tests that we have in a strategic way to enable public health decisions that will optimize reduction of disease transmission while maximizing the opportunity for people to retain as much normalcy as possible. If someone reading this is hot to make a more sensitive test, they should contact me and I will point them to three different approaches, any of which will likely help.
Karen Heichman: PCR is a very sensitive method, and is probably the most sensitive of the types of assay methodologies because the target is amplified (not just the signal). there is quite a bit of variability as to the sensitivity of various PCR-based tests. the most sensitive PCR tests can detect a few hundred viral genomes or fewer. Note, however, that newly emerging data suggests that a much higher viral load may be required for a person to be considered infectious.
Yuan-Po Tu: The critical question is how long are individuals contagious and at what RT-PCR cycle time is the patient no longer contagious. The RT-PCR is too sensitive. RT-PCR not distinguish between virus fragments and competent virus than can be recovered by viral culture. We probably need an alternative approach that distinguishes between virus parts (RT-PCR positive that is not culturable) vs virus that is able to be cultured and more likely contagious.
No test is perfect. If you look at repeat RT-PCR testing, there is a false negative rate as well. This is from a literature review and pooled analysis of inpatients and outpatients with SAR-CoV-2 where RT-PCR has a false negative rate.
The other phenomenon that is becoming clearer is that infected individuals with more than 10 days (US CDC data) to 14 days (Korean CDC) of symptoms which are improving, viral cultures are negative. This is despite the persistence of RT-PCR at high cycle times (low levels of viral fragments) which can last for weeks.
Have any of the panelists thought of this fall in terms of possible ADE occurring upon re-infection?
Timothy O’Leary: Antibody-dependent enhancement is certainly a potential concern. People have speculated that a slow immune response has allowed ADE to occur in older victims of covid19, leading to more severe disease. It would be good to get multiple samples for sequencing throughout the course of the disease to see whether this type of viral adaptation is occurring now. ADE is also a concern for vaccine production. My pessimistic outlook is that if rapid reinfection is possible and ADE occurs, then we are above a waterfall in a rickety boat with no paddle.
Karen Heichman: I do not have much knowledge on ADE or the possible impact in COVID. I do not believe that there have been many confirmed cases of re-infection so I think it is an open question.