Latest News

Testing Caveats Learned (And Relearned) During the Pandemic

By Deborah Borfitz

September 8, 2021 | A long list of diagnostic adages have held true during the COVID-19 pandemic: a proactive and organized scientific approach yields excellent results, serology is not useful for detecting respiratory viruses, not all nucleic acid amplification tests perform equivalently, different specimens yield different results, pooling will miss low-level positives, RT-PCR is more sensitive than antigen tests, pretest probability is directly related to post-test-probability, and errors occur when implementations are rushed.

So began a presentation at the recent Next Generation Dx Summit by Gary Procop, M.D., director of molecular microbiology, virology, mycology, and parasitology at the Cleveland Clinic. As recently detailed in the Journal of Hospital Management and Health Policy (DOI: 10.21037/jhmhp-21-9), the Cleveland Clinic adopted a systemwide approach to dealing with the pandemic that included development of an Incident Command Center to serve as the central hub for information and delegation. 

Lab medicine staff came out of the shadows and were “brought to the fore” during twice-daily meetings alongside their pathology colleagues, he says. Since antibody tests are not recommended for the diagnosis of COVID, the question was what sort of test to use and how best to apply it. 

Initially, the federal government was saying “you can’t help,” Procop says. But since the U.S. lacks the infrastructure to test all Americans, labs across the country soon “came to the rescue” of the public health service. “We had to scramble.”

Unfolding Situation 

The first question asked—what the sensitivity and specificity of a test needed to be—was unanswerable since there weren’t yet enough COVID-19 cases, he continues. Requirements for Emergency Use Authorization from the Food and Drug Administration were burdensome, and it was unclear what was and wasn’t a “good” limit of detection. 

A five-test comparison has since become possible, revealing a high false-positive rate among rapid molecular and antigen tests relative to the slower, gold-standard PCR test, says Procop.

At the start of the pandemic, people weren’t allowed to go back to work until they had two consecutive negative test results, he continues. No one was yet thinking about viral load, which is low early in the disease process and during convalescence, which might not be enough to elicit a positive test result and may or may not matter in terms of personal health or transmissibility. 

Multiple types of specimens—including bronchoalveolar lavage (BAL) as well as nasopharyngeal, oropharyngeal, and nasal mid-turbinate swabs—have been tried, Procop notes. An assessment of the options, published in JAMA (DOI: 10.1001/jama.2020.3786) early last year, found BAL fluid specimens showed the highest positive rates (93%), followed by sputum (72%), nasal swabs (63%), fibrobronchoscope brush biopsy (46%), pharyngeal swabs (32%), feces (29%), and blood (1%). None of the 72 urine specimens tested positive. 

The Centers for Disease Control and Prevention (CDC) lists the specimen types that can be used, he says, “but not even all swabs are created equal;” rather, they come in shades of good, better, and best. The smaller ones are needed to reach the area where the virus replicates. The Cleveland Clinic switched to mid-turbinate swabs when nasal swabs became unavailable. 

When reagents were in short supply, the Cleveland Clinic also began pooled sample testing, as it has done for years in other disease areas, including HIV infection, Procop says. But samples at the test’s detection limit will get missed since they would be diluted into pools of negative specimens. 

Procop led a study, published earlier this year in the American Journal of Clinical Pathology (DOI: 10.1093/ajcp/aqaa273), testing a 10:1 pooling technique for asymptomatic patient testing. All 10 specimens with midrange viral load were detected by RT-PCR assay, in contrast to 7 of 10 among specimens with a low viral load—which may be good enough in the outpatient setting if patients are advised of the risk of a false-negative result and instructed to maintain standard handwashing and social distancing mitigation strategies. 

Pooling can save on reagent use but is “taxing” on labs since they must deconstruct any pool that tests positive, Procop says. “You lose money when prevalence is high. We don’t do it now reagents are more available.” 

Much To Learn 

Testing without a plan is “chaos,” says Procop, referencing pop-up sites with long lines of cars filled with the worried well. Counterintuitively, the same good test also performs differently in a screening versus a diagnostic scenario, although they are frequently viewed interchangeably.

In the example Procop gives, a patient with respiratory symptoms with a 50-50 chance of having COVID would have a 93% positive test probability whereas pre-surgical patients with a 0.5% positivity rate would have a 7% positive test probability using the same assay.

At Yale, he shares, researchers have demonstrated the potential of saliva-based testing as an alternative to deep nasal swabs. As reported by the CDC, the Yale team also showed the stability of SARS-CoV-2 RNA detection in saliva stored for prolonged periods in a variety of settings, making it a potentially viable and cost-effective alternative for large-scale testing without the need for personal protective equipment and healthcare personnel to collect patient samples.

Load more comments