On August 18, the leaders at the Infectious Diseases Society of America (IDSA), the Arlington, Virginia-based association that is “a community of over 12,000 physicians, scientists and public health experts who specialize in infectious diseases,” released a new set of guidelines on serologic testing for COVID-19. It was the third set of guidelines that the society had published, following guidelines for treatment and management (April 11) and for infection prevention (April 27).
The IDSA’s leaders wrote on their organization’s webpage that “Accurate molecular diagnostic tests are necessary for confirming a diagnosis of coronavirus disease 2019 (COVID-19). Direct detection of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) nucleic acids in respiratory tract specimens informs patient, healthcare institution and public health level decision-making. The numbers of available SARS-CoV-2 nucleic acid detection tests are rapidly increasing, as is the COVID-19 diagnostic literature. Thus, the Infectious Diseases Society of America recognized a significant need for frequently updated systematic reviews of the literature to inform evidence-based best practice guidance.”
As the IDSA leaders noted, “The IDSA’s goal was to develop an evidence-based diagnostic guideline to assists clinicians, clinical laboratorians, patients and policymakers in decisions related to the optimal use of SARS-CoV-2 nucleic acid amplification tests. In addition, we provide a conceptual framework for understanding molecular diagnostic test performance, discuss the nuance of test result interpretation in a variety of practice settings, and highlight important unmet research needs in the COVID-19 diagnostic testing space.”
The IDSA leaders reported that “IDSA convened a multidisciplinary panel of infectious diseases clinicians, clinical microbiologists, and experts in systematic literature review to identify and prioritize clinical questions and outcomes related to the use of SARS-CoV-2 molecular diagnostics. Grading of Recommendations Assessment, Development and Evaluation (GRADE) methodology was used to assess the certainty of evidence and make testing recommendations.”
And, as the IDSA leaders noted, “Universal access to accurate SARS-CoV-2 nucleic acid testing is critical for patient care, hospital infection prevention and the public response to the COVID-19 pandemic. Information on the clinical performance of available tests is rapidly emerging, but the quality of evidence of the current literature is considered low to very low. Recognizing these limitations, the IDSA panel weighed available diagnostic evidence and recommends nucleic acid testing for all symptomatic individuals suspected of having COVID-19. In addition, testing is recommended for asymptomatic individuals with known or suspected contact with a COVID-19 case. Testing asymptomatic individuals without known exposure is suggested when the results will impact isolation/quarantine/personal protective equipment (PPE) usage decisions, dictate eligibility for surgery, or inform administration of immunosuppressive therapy. Ultimately, prioritization of testing will depend on institutional-specific resources and the needs of different patient populations.”
The key recommendations:
Ø Recommendation 1: The IDSA panel recommends a SARS-CoV-2 nucleic acid amplification test (NAAT) in symptomatic individuals in the community suspected of having COVID-19, even when the clinical suspicion for COVID-19 is low (strong recommendation, very low certainty of evidence).
Ø Recommendation 2: The IDSA panel suggests collecting nasopharyngeal, or mid-turbinate or nasal swabs rather than oropharyngeal swabs or saliva alone for SARS-CoV-2 RNA testing in symptomatic individuals with upper respiratory tract infection (URTI) or influenza like illness (ILI) suspected of having COVID-19 (conditional recommendation, very low certainty of evidence).
Ø Recommendation 3. The IDSA panel suggests that nasal and mid-turbinate (MT) swab specimens may be collected for SARS-CoV-2 RNA testing by either patients or healthcare providers, in symptomatic individuals with upper respiratory tract infection (URTI) or influenza like illness (ILI) suspected of having COVID-19 (conditional recommendation, low certainty of evidence).
Ø Recommendation 4: The IDSA panel suggests a strategy of initially obtaining an upper respiratory tract sample (e.g., nasopharyngeal swab) rather than a lower respiratory sample for SARS-CoV-2 RNA testing in hospitalized patients with suspected COVID-19 lower respiratory tract infection. If the initial upper respiratory sample result is negative, and the suspicion for disease remains high, the IDSA panel suggests collecting a lower respiratory tract sample (e.g., sputum, bronchoalveolar lavage fluid, tracheal aspirate) rather than collecting another upper respiratory sample (conditional recommendations, very low certainty of evidence).
Ø Recommendation 5: The IDSA panel suggests performing a single viral RNA test and not repeating testing in symptomatic individuals with a low clinical suspicion of COVID-19 (conditional recommendation, low certainty of evidence).
Ø Recommendation 5: The IDSA panel suggests performing a single viral RNA test and not repeating testing in symptomatic individuals with a low clinical suspicion of COVID-19 (conditional recommendation, low certainty of evidence).
Ø Recommendation 6: The IDSA panel suggests repeating viral RNA testing when the initial test is negative (versus performing a single test) in symptomatic individuals with an intermediate or high clinical suspicion of COVID-19 (conditional recommendation, low certainty of evidence).
Ø Recommendation 7: The IDSA panel makes no recommendations for or against using rapid (i.e., test time ≤ 1hour) versus standard RNA testing in symptomatic individuals suspected of having COVID-19 (knowledge gap).
Ø Recommendation 8: The IDSA panel suggests SARS-CoV-2 RNA testing in asymptomatic individuals who are either known or suspected to have been exposed to COVID-19 (conditional recommendation, very low certainty of evidence).
Ø Recommendation 9: The IDSA panel suggests against SARS-CoV-2 RNA testing in asymptomatic individuals with no known contact with COVID-19 who are being hospitalized in areas with a low prevalence of COVID-19 in the community (conditional recommendation, very low certainty of evidence).
Ø Recommendation 10: The IDSA panel recommends direct SARS-CoV-2 RNA testing in asymptomatic individuals with no known contact with COVID-19 who are being hospitalized in areas with a high prevalence of COVID-19 in the community (i.e., hotspots) (conditional recommendation, very low certainty of evidence).
Ø Recommendation 11: The IDSA panel recommends for SARS-CoV-2 RNA testing in immunocompromised asymptomatic individuals who are being admitted to the hospital regardless of exposure to COVID-19 (strong recommendation, very low certainty of evidence).
Ø Recommendation 12: The IDSA panel recommends SARS-CoV-2 RNA testing (versus no testing) in asymptomatic individuals before immunosuppressive procedures regardless of a known exposure to COVID-19 (strong recommendation, very low certainty of evidence).
Ø Recommendation 13: The IDSA panel suggests for SARS-COV-2 RNA testing in asymptomatic individuals (without known exposure to COVID-19) who are undergoing major time-sensitive surgeries (conditional recommendation, very low certainty of evidence).
Ø Recommendation 14: The IDSA panel suggests against SARS-CoV-2 RNA testing in asymptomatic individuals without a known exposure to COVID-19 who are undergoing a time-sensitive aerosol generating procedure (e.g., bronchoscopy) when PPE is available (conditional recommendation, very low certainty of evidence).
Ø Recommendation 15: The IDSA panel suggests SARS-CoV-2 RNA testing in asymptomatic individuals without a known exposure to COVID-19 who are undergoing a time-sensitive aerosol generating procedure (e.g., bronchoscopy) when PPE is limited, and testing is available (conditional recommendation, very low certainty of evidence).
The IDSA website carries remarks and detailed explanations for all 15 recommendations in this set.
On Aug. 20, IDSA leaders provided a web-based and telephonic press briefing. Christopher Buskey, the IDSA’s CEO, moderated the briefing. Presenting and participating were Kimberley E. Hanson, M.D., M.H.S, an associate professor of internal medicine and an adjunct associate professor of pathology, at the University of Utah School of Medicine in Salt Lake City, and the chair of the IDSA’s COVID-19 Diagnostic Guidelines Expert Panel and a member of the IDSA’s Board of Director; and Angela M. Caliendo, M.D., PH.D., FIDSA, a professor and executive vice chair in the Department of medicine at the Alpert Medical School at Brown University in Providence, Rhode Island, and a member of the IDSA COVID-19 Diagnostic Guidelines Expert Panel, and secretary of the IDSA’s Board of Directors.
Dr. Hanson began, by stating that, “As Chris [Buskey] mentioned, the focus of our briefing is the new guidelines we’ve issued around serologic or antibody testing. It’s in our language now that antibodies are specific proteins the immune system makes in response to infection. We can use those antibodies as a marker for having experienced infection. It’s been incredible to watch the pace at which these diagnostic tests have come to market, the pace of articles in the literature,” Hanson said. “And it’s hard to tell on the surface how these tests perform and how they should be used in clinical practice. So the IDSA convened an expert panel of which Angie and I were a part. Panel included infectious diseases specialists like ourselves, Clinical microbiologists, experts at looking at the medical literature.”
With regard to serologic testing, Dr. Hanson said that “It’s worth looking at antibodies. There are two big general classes: neutralizing and non-neutralizing/binding antibodies. There’s been a lot of interest in neutralizing antibodies: when an antibody binds. Non-neutralizing antibodies can’t do that. The diagnostic tests can’t separate out the neutralizing from the binding antibodies. And that becomes important in predicting who might be protected, who might be immune. The tests don’t differentiate among them. Also, there are different classes: IGM and IGG antibodies. IGM antibodies are typically produced in the body first and can be detected within a week of infection. IG antibodies are often detectable in the blood as long as years later.”
Hanson noted that, “In the U.S., the FDA [Food and Drug Administration] does say that if you’re going to market an antibody test for COVID, you must enter into their process of review to meet their benchmark. It hasn’t always been that way. Early on in the pandemic, the submission for an EUA [emergency use authorization] was voluntary. That turned out to be a problem, and early on in the outbreak, the market was flooded with tests that were pretty poorly performing. So FDA corrected that.”
Hanson reported that “Our panel reviewed at least 9,500 papers; about 47 met our criteria to include in our review, in terms of worthwhile evaluation of tests. And the reference standard was a molecular diagnostic test that detects the virus itself, not just immune response. And even though it was only 47 papers, the data was really complicated. Tests detect different classes of antibodies; the platform can vary from a near-patient point-of-care platform to one that is lab-based; whether EUA-approved or not; and the timing of the testing was variable. So we really tried to focus on timing. And as it turns out, timing was really important.”
Further, Dr. Hanson said, “We found that the IGM antibodies were not detectable in the first week or so after the onset of symptoms. That observation informed our first recommendation. That is, we’re recommending against antibody testing within the first week or so after infection; you can’t exclude the possibility that the patient might still have COVID-19. We also saw that the tests performed at different levels. IGM tests were more variable than IGG tests. So clinicians really need to know the different performance levels of the tests.”
Dr. Caliendo then picked up the discussion, stating that, “If we understand that the first two weeks after the onset of symptoms is not the best timeframe, what is the best timeframe? And detecting IGG or a blend? At three to four weeks after onset, that’s when these tests have the highest sensitivity; it’s still only 95-96 percent, and the specificity is 99 percent; there will still be false negatives and positives; but using an IGG test or a total-antibody test at three to four weeks is best. It was interesting that IGM test isn’t more sensitive; and in fact, with IGM, the sensitivity of the test drops after 4 weeks. And that means, with IGM, you’re going to miss some people. So we did not feel that IGM added a lot beyond IGG. IGA tests have a very low level of specificity, so we did not feel that IGA tests were useful. There are tests that look at IGM or IGG, but they have very low specificity,” she noted.
“We identified two best uses for serologic tests,” Caliendo stated. “One is with patients who come in with symptoms, chest x-rays, lab results, etc., but they get the nasal nucleic acid test and it’s negative once or twice; there, particularly if the person is 3-4 weeks out from onset of symptoms, the serology testing is very useful. After 2 weeks? 50-60 percent may be positive; but if they’re negative at two weeks, you’ll want to test them again. So why would the nucleic acid test be negative?” she asked. “It’s likely that the virus is no longer in your nose at that point, but now in your lungs. And so that’s one good use of a serology test as a diagnostic. For pediatric patients, we recommend both nucleic testing and serology testing. Children can present with the inflammatory syndrome and not show other symptoms; that’s why we recommend both nucleic and IGG testing for children.”
Indeed, Caliendo continued, “Far and away, the main use of serology testing is in sero-surveillance, to determine what percentage of a population shows antibodies. Detecting antibodies gives us an idea of what percentage of people in a population are infected, and you can test again over time. Now if you’re going to use these tests for sero-surveillance, we recommend a very high level of specificity—99.5 percent. If you get a level of 99 percent, but you test a large number of people in a population where the actual number of people infected is 1 percent, you’ll get an equal number of false positives and true positives. That’s why you want a really high level of specificity. If someone gets a false positive, they may never have been infected, and you might be giving them false comfort.”
Further, Caliendo added, “When you do these seroprevalence studies, the higher level of infection rate, false positive results become less important, and false negative results become more important. So doing sero-prevalence testing can be tricky; you can get both false negatives and false positives. That’s why I would recommend highly specific tests. Yet there are not a large number of studies assessing these tests. So we tried to break things out by timeline, by immunoserologic type, etc. But I would reiterate that it is essential that the laboratory and the clinicians using the laboratory understand what types of tests are involved.”
Asked by a member of the press, “Can the panelists discuss how hospitals are using antigen tests compared to PCR tests, to determine COVID-19 cases?” Hanson responded, “There has also been a lot of interest in antigen tests; you’re detecting usually a part of the protein on the virus. They’re less expensive; but they tend to be much less sensitive than molecular tests. To try to decide how antigen tests might fit in, we haven’t been able to evaluate them. Supply has been short. I think it’s a note of caution that we know that they are less sensitive than molecular tests. A negative antigen test doesn’t rule out that the patient has COVID-19. If the test is positive, though, there is good specificity. So with a positive test, do you go on to do a molecular test? We’re sorting out their best use. I’m interested in Angie’s take on that.”
“Our hospital is not using antigen tests at this point,” Caliendo said. “Colleagues are using it in the ED when they can’t get a rapid PCR reagent. Some of us can’t easily get PCR reagents; some colleagues are relying on the antigen tests, realizing they’re less sensitive. We’re considering using antigen tests for symptomatic children in the schools; they’re not sensitive enough for use with asymptomatic people. But they can be used in schools in that context. We’ll be using antigen tests in the school seeing in our state. We’re doing things that I would describe as less than ideal, out of necessity.”
