'By the time they send things out, it ought to be foolproof'
Greninger said one reason to include a third reagent in the coronavirus test was to prepare for the possibility that the virus might mutate. An additional primer could offer a safety net to detect the virus if its genetic sequence changed.
"They were sort of trying to make an assay that could detect other SARS-like coronaviruses that circulate in bats," Greninger said. "For whatever reason, it just didn't work. That happens all the time when you make PCR tests — some of the primers just don't play ball together and you just gotta toss 'em."
But health experts told Business Insider that distributing a test with a faulty reagent was somewhat unusual.
"By the time they send things out, it ought to be foolproof," William Schaffner, a CDC adviser who is a professor of infectious diseases at Vanderbilt University, told Business Insider. "As a friend of mine said, it ought to be a gorilla test. Even a gorilla can run it."
"I don't know where the glitch came from with that reagent," he added. "Something went wrong, that's for sure."
On a February 21 press call, Messonnier said problems with the test kits were "a normal part, unfortunately, of these processes."
More than a manufacturing issue
When the FDA first approved the test kits on February 4, the agency did so under an "Emergency Use Authorization," which expedites the process of diagnostic testing during a public-health crisis.
The first EUA was issued 11 days after the CDC announced its test. From there, it took the FDA an additional 22 days to award another EUA for that same test without the one nonessential ingredient.
"The EUA is like a recipe that you have to follow," Greninger said. "It'd be like if I sent you a recipe for a pound cake and I put 10 times more salt in it than I should. And you're like, this pound cake sucks, but you have to follow the recipe because it's the approved recipe. And then I wait 16 days and I change the recipe. You still have the same ingredients, but you're just using them the right way now."
At a press briefing on March 7, the FDA's commissioner, Stephen Hahn, suggested it was manufacturing problems that made it difficult for public-health laboratories to start using the CDC test. The FDA and the CDC, he added, had quickly resolved the issue by using a third-party manufacturer.
"Quality testing, distribution of tests, and verification of the tests in labs can take a few days, although FDA, CDC staff, and the laboratory community are working around the clock to speed up these processes," Hahn said.
But Greninger said the lag time between identifying the faulty reagent and issuing another EUA was "the frustrating part of the story."
"We would be telling each other a different story if it were a manufacturing issue," he said. "The honest truth is, whatever it was, they just got rid of the reagent."
Greninger's lab at the University of Washington began using its own COVID-19 test on March 2 — just two days after the FDA started allowing certified academic hospital labs to perform their own tests.
His lab is now administering 1,800 to 2,000 tests a day and diagnosing more than 100 to 150 patients each day in the Seattle area, he said. He said he's still getting kits with the faulty reagent included.
"We take that tube and we throw it in the trash so we don't accidentally use it," Greninger said. "It's still there."
Update: The CDC provided the following statement to Business Insider three weeks after our initial inquiry.
"There were 19 state labs that ran the test successfully, with controls working. As a general matter of quality control, CDC supports any State Public Health Laboratory that has trouble validating tests, and collaboratively explores alternative options. Both avenues were occurring — it was expeditious for the labs that could validate to continue using the test, and for us to work with the labs that couldn't in order to narrow down the issue and resolve it.
CDC quickly identified the component that was not functioning correctly, which may have been the result of a design or manufacturing problem; validated the test's performance without the component; re-manufactured the test kits; obtained FDA permission to drop the problematic component from the Emergency Use Authorization; and rolled the test kit back out to the states to get testing up and running again. By March 8 public health labs in all 50 states had validated the assay and were testing for COVID-19."
