For months, public health experts have been eagerly watching the companies developing spit tests for the coronavirus that could be used at home, producing results in a matter of minutes.
If these rapid saliva tests worked, as many news articles have pointed out, they could greatly expand the number of people getting tested. Some experts have even said they could perform as well as a vaccine in curbing the spread of the coronavirus and paving a path back to normalcy.
But so far, the technology is not panning out as some have hoped.
E25Bio and OraSure, two companies pursuing rapid at-home coronavirus tests, have abandoned efforts to use saliva in their products. Their tests, which detect pieces of coronavirus proteins called antigens, will for now rely on shallow nose swabs instead.
“If I was placing a bet — which I am, because I’m leading an antigen-based testing company — I would say it’s going to be very difficult for antigen-based testing to work on saliva samples,” said Bobby Brooke Herrera, an E25Bio founder and its chief executive. The notion that the virus sets up shop in the mouth and produces enough antigen to be picked up by today’s technology, he said, “is far-fetched.”
The two companies pursued saliva (or “oral fluids,” in the case of OraSure) for months in the hopes of their tests being more comfortable than swabs, some of which go painfully deep into the nose, and less reliant on sputtering supply chains that have caused long delays for laboratory tests.
“There’s nothing more convenient than spitting in a tube,” said Dr. Valerie Fitzhugh, a pathologist at Rutgers University.
But as they continued to tinker with their tests, researchers at both E25Bio and OraSure found saliva’s performance to be more lackluster than anticipated, and were forced to pivot.
“This was a result of optimization studies,” said Stephen Tang, OraSure’s president and chief executive officer. A saliva antigen test is still theoretically possible, Dr. Tang added. But after comparing the amount of coronavirus antigen found in the nostrils and throat, as well as different parts of the mouth, including the cheeks, gums and tongue, “we decided to optimize around the nasal cavity,” he said.
Spit also differs vastly among people, and can even change over the course of a single day. “We’ve all noticed that there is variable performance,” said Sarah Jung, scientific director of clinical microbiology at Children’s Hospital Colorado.
Both E25Bio and OraSure plan to seek authorization from the F.D.A. to sell at-home antigen tests using nose swabs instead of spit, a technique similar to the one used by the much-talked-about Abbott antigen test that takes about 15 minutes. The E25Bio test would require people to swab their nose, stir the sample into a chemical soup, apply the mixture to a paper strip and wait up to half an hour for bands to appear. E25Bio’s test picks up on about 80 percent of the infections that ultrasensitive laboratory tests detect — the F.D.A.’s bare minimum for a regulatory greenlight. OraSure declined to give any details about its test’s methodology or accuracy.
Saliva does seem to be working when used in laboratory tests known as P.C.R., which look for bits of the virus’s genetic material, or RNA, rather than antigens. P.C.R. tests detect minute amounts of coronavirus RNA, making them far more sensitive than antigen tests. Research teams at Rutgers and Yale have been granted emergency authorization for these spit P.C.R. tests.
At the University of Illinois, some 10,000 of the institution’s in-house P.C.R. tests are performed each day on saliva from students, faculty and staff members — roughly 1 percent of the nation’s daily tests.
Standard P.C.R. tests, however, take hours to run and are subject to shortages of laboratory supplies, such as pipettes and chemicals, often leading to delays in getting results.
His team’s prototype takes just 45 minutes. It uses an array of portable equipment, including two tissue-box-size heaters.
That’s too bulky and expensive for at-home testing. But he said the test could be deployed in places where crowds gather, like schools, offices and travel hubs, granting safe passage to those who test negative while sending the infected back home.
His team has submitted an application for emergency approval from the F.D.A. In the meantime, they’ve partnered with Sorrento Therapeutics, a San Diego-based company, to scale up production.
The Columbia saliva test relies on a technique called LAMP that’s generally faster but a bit less accurate than P.C.R. The spit sample is briefly boiled and mixed into a cocktail of chemicals that then gets incubated at 145 degrees Fahrenheit for half an hour. If the tube’s contents turn from red to yellow, the test is positive.
The latest data shows the Columbia test performs as well as a laboratory deep-nose swab test more than 96 percent of the time, even when using saliva from sick patients who gave messy samples.
“Even if there was food or blood, we took it,” Dr. Williams said. That wouldn’t fly with most other saliva tests in use, he said, which ban eating, tooth-brushing and even gum chewing in the hour or so before depositing drool.
Another saliva LAMP test is being tested by David O’Connor at the University of Wisconsin-Madison. Their technique bears many similarities to the Columbia test, including a color-based readout, but takes slightly longer and involves a couple of extra steps. Early trials of the test on volunteers in Wisconsin have gone well, Dr. O’Connor said, and one school district in Illinois is using the test to screen about 1,400 students and teachers on a weekly basis.
But saliva LAMP tests face their own hurdles. Saliva tends to clump and stick, and can be a difficult substance to handle and transfer from tube to tube, said Jennifer Dien Bard, director of the clinical microbiology and virology laboratory at Children’s Hospital Los Angeles, where her team is working to roll out saliva-based testing for coronavirus RNA. And although color-based results are simple, interpreting them can get messy if a rainbow of hues comes out the other end.
“To me,” Dr. Dien Bard said, “something like this might still belong in a lab.”