Do we really have the ’best’ tests?

A regular feature of the daily Covid-19 briefings is the declaration—most often made by President Donald J. Trump—that the U.S. has the “best tests” and is testing more often than any other country.

But what does the evidence reveal about Covid-19 assays currently available?

The first American Covid-19 test, developed by the CDC, was approved by the FDA Feb. 4. By early April, there were 30 approved commercial assays, as well as five with the okay for use by the institutional labs that developed them.

“We’ve come a long way in the last month,” said Kimberly Hanson, MD, of the University of Utah School of Medicine in Salt Lake City. But testing for Covid-19, the disease caused by the novel coronavirus SARS-CoV-2, remains a challenge.

“Our goal is to test everybody who has symptoms,” Hanson told reporters in a briefing organized by the Infectious Diseases Society of America. But shortages — of the tests themselves, as well as in the supplies needed to administer them — mean not all patients are getting tested as often as physicians would like.

Grading the Tests

“We still have backlogs,” said Angela Caliendo, MD, PhD, of Brown University’s Alpert Medical School in Providence, Rhode Island. On any given day, she told reporters, physicians have to set testing priorities based on what supplies are available; if some necessities aren’t in stock, some normally desirable tests will go by the board.

Medical assays generally get approval after they go through analytical study using mock specimens for such things as sensitivity, specificity, and reproducibility, followed by clinical testing to show that the lab analysis translates well to the real world.

But all of the Covid-19 tests so far have been given what’s called an Emergency Use Authorization (EUA), Caliendo noted, which means that the second stage of the process is simply skipped.

So how good are the tests? “We actually don’t know that,” she said. “That’s the trade-off… if you need it quickly, you don’t have the time to run the clinical trial.”

The two key measures of a test are its sensitivity and specificity — how well it distinguishes between those who are infected and those who aren’t. Put another way, an ideal test would be positive for every infected person and negative for every uninfected person; there would be no false-negatives and no false-positives.

The new assays seem to be pretty accurate, at least when they’re positive, commented Rochelle Walensky, MD, of Boston’s Massachusetts General Hospital, speaking at an earlier IDSA briefing. But in clinical practice so far, they appear to yield falsely negative results in about three of 10 patients with a SARS-CoV-2 infection.

That’s not necessarily a problem for acute care, since most people now being tested are symptomatic with fever, cough, and respiratory distress, so physicians simply assume they are infected. “We don’t take a patient off precautions” just because of a negative test, Walensky said.

She added there are number of reasons a test might be a false negative, including the simple mechanical difficulty of taking a nasopharyngeal sample. “You really have to jam that (swab) pretty far back,” she said.

Indeed, “a lot goes into getting a good result,” Caliendo said, including sample collection, the biology of the virus, and the quality of the test itself.

“A good specimen is absolutely critical,” she said. The best specimens come from patients on ventilators, where doctors can collect sputum from deep in the lung, but nasopharyngeal or throat swabs are standard for most people, she noted.

Hospital personnel are very used to taking nasopharyngeal swabs for such things as influenza tests, but as Covid-19 testing is expanded into the community — with drive-by testing sites, for example — less experienced people will inevitably be taking samples and that might lead to a proportion of false negatives, she said.

Timing can also be important: “If you test me the very first day that I have symptoms, the amount of virus in my respiratory tract is going to be low, compared with what it will be four or five days later,” she said.

And, the last aspect is the sensitivity of the test itself. “I would say that what we have on our hands now are very good tests, but no test is perfect,” she said.

There are two types of tests that can be useful in an outbreak, Caliendo said — tests for the virus itself using its genetic material, and serological tests for the aftermath of infection that look for antibodies. Of the tests approved by the FDA, all but one — a blood test developed by Cellex, Inc., of Research Triangle Park, North Carolina — use polymerase chain reactions (PCR) to test for the presence of viral genes.

A complete list of approved tests — a list that is changing day by day — is on the FDA website.

SARS-CoV-2 is a positive-sense, single-stranded RNA virus, meaning its genetic material is mRNA that can be directly translated into proteins within the target cell.

The PCR tests look for various RNA segments, including regions of:

  • The nucleocapsid (N) gene.
  • The spike (S) gene.
  • The envelope (E) gene.
  • Genes for non-structural proteins contain in open reading frame 1 (ORF1).

Whether all those targets are equally good is an open question, according to Hanson. “That could affect the specificity of the test,” she told BreakingMED, adding that the analytic studies and computer modeling needed for the EUA are meant to ensure that the chosen targets are valid.

But, “in clinical practice we don’t know if there are differences,” she said.

Caliendo said data are emerging about the value of different test targets, but it’s too early to make any definite statements. “Once we have a lot of specimens in the lab, these are the types of things that we’ll do a deep dive and have a much better understanding which targets are more specific and which are more sensitive,” she said.

Hanson added that researchers are periodically sequencing viruses isolated from patients to see if there are changes that might affect how well tests work.

Is Process the Problem?

One issue that has limited availability of testing, Caliendo said, is that the EUA process is to some extent skewed toward companies, which have the expertise to navigate the process, and against hospital and academic labs that are able to quickly design and build tests systems but are less familiar with FDA rules.

Many academic labs, she said, were designing and validating their own tests early on, when commercial tests were not yet available. “If you took that away from us, it would have a dramatic effect on clinical care,” she said.

Hanson said it might be a good idea in the future to have some academic labs “pre-certified” so that, in a new pandemic, they could quickly get tests up and running without having to go through the full FDA approval process.

Given limited availability, most tests are being used on patients who have some of the symptoms of Covid-19 and show up at hospital. But should others be tested, Caliendo said, especially those that are displaying symptoms but remain well enough to stay home?

In Rhode Island, she said, there aren’t yet enough tests to do that, “but we are really ramping up very well.” But the cautionary note is that a “negative test isn’t a license to stop social distancing.”

It would also be nice to test recovered patients before discharge, but that’s again not routine, unless the patient is being sent back to a long-term care facility.

It is also not routine to test close contacts of known patients unless they are also showing symptoms, Hanson noted.

Part of the problem, she said, is that it’s not clear how the tests perform in asymptomatic people, although studies of the issue are under way.

Again, timing is an issue: “If I test a close contact today and they’re negative, it doesn’t necessarily mean they might not be positive tomorrow, or the next day,” she said. “Would you need to do serial testing of contacts? What’s the optimal time to test a contact who hasn’t yet displayed symptoms? We’re still trying to figure those things out.”

The advent of serological tests will allow larger community studies and will be “very informative to understand the extent of the population who might have been exposed and potentially immune,” she added.

Sallie Glomb Reinmund, PhD, Sr. Scientific Content Director, @Point of Care, contributed insights and research during production of this series. This is the second part of a special BreakingMED series examining the state of the science regarding SARS-CoV-2 and Covid-19.

  1. The development of tests for Covid-19 is evolving rapidly, with 30 approved commercial assays, as well as five with the okay for use by the institutional labs that developed them.

  2. The initial goal is to test everyone who has symptoms of Covid-19.

Michael Smith, Contributing Writer, BreakingMED™

Caliendo had nothing to disclose.

Hanson disclosed relevant relationships with Biofire and T2 Biosystems.

Walensky disclosed relevant relationships with the National Institutes of Health and the U.S. Department of Health and Human Services panel on Antiretroviral Guidelines for Adults and Adolescents.

Cat ID: 125

Topic ID: 79,125,730,933,125,190,520,926,192,927,151,928