Making peace with saliva, pooled testing

Karen Titus

November 2020—Adam Barker, PhD, D(ABMM), was ready to call it quits.

For weeks, he had been working to bring saliva-based SARS-CoV-2 testing to ARUP Laboratories and the University of Utah. Dr. Barker, director of ARUP’s COVID-19 rapid response lab, and his colleagues had done studies comparing saliva with nasopharyngeal swabs, which seemed to be following the flight of the passenger pigeon out of existence. They had wrestled with the FDA over emergency use authorization. They’d developed their own transport media, since that supply was also becoming extinct.

He had begun building kits for saliva collection and figured out what sample size worked best. Kits had been delivered to collection sites on campus, and staff were being trained in their use.

He was, in other words, creating a laboratory success story, one of the many that have been written since March.

He was not basking in this fact.

“I have to tell you: I lost so much sleep because of saliva,” says Dr. Barker, who is also director, ARUP Institute for Clinical and Experimental Pathology, medical director of special microbiology and R&D special operations, and associate professor, Department of Pathology. Despite his lab’s incredible investment and the fact that he was an early proponent, he’d had enough—a thought he did not keep to himself.

“I went to the powers that be at my hospital and said, ‘Listen, I was the first person who loved saliva. I am working so hard to get this to work. How about we just go back to the old way?’” Dr. Barker recalls. The lab was seeing too many invalids on saliva samples and watching turnaround times increase—perturbing to any administrator. “I thought they would jump on it and say, Yeah, let’s stop.

“They did the exact opposite.”

Which is why Dr. Barker now finds himself running a large-scale, saliva-sample–based SARS-CoV-2 testing operation. The operation launched in early September. By mid-October, when he spoke to CAP TODAY, it was being used at drive-through collection centers throughout the university to test symptomatic patients or those who’ve been in contact with someone who has tested positive. They also do saliva-based screening for patients admitted to the hospital, as well as inpatients. As a large reference lab, ARUP is also providing kits to its numerous clients.

Dr. Barker’s superiors had their reasons for wanting to stick with saliva, including ongoing PPE shortages and an easier collection process. They told him, Patients like it better, we like it better. It’s win-win. Keep going.

And so laboratories go, continuing to pursue testing strategies they would have found unfathomable a year ago. Some, like ARUP, are turning to saliva. Others are handling outsized demands by pooling specimens. Choices are being shaped in part by maddening supply shortages outside the lab as well as in, by hospital executives, by media coverage and patient expectations.

The pressures within the lab are equally exhausting. At the University of Chicago, Kathleen Beavis, MD, has implemented pooling. It, too, appears to be a success story, but it’s impossible to ignore the tensions of deciding who needs to be tested and how, says Dr. Beavis, professor, Department of Pathology. When does one risk outweigh another? In a pandemic, with no real acme, can good enough actually be great?

In finding answers, labs will continue to devote hours to bringing up new tests—and, at least in Dr. Barker’s case, listening to leaders explain why sleep will remain elusive for the foreseeable future.

Dr. Barker and his colleagues started taking a more serious look at saliva samples in July, sparked by conversations with colleagues at other institutions as part of an American Board of Medical Microbiology consortium: If they ran out of swabs, everyone wondered, how would they keep testing going? Saliva seemed like a reasonable choice, especially with the promising work being done at Yale and Rutgers universities.

He and colleagues began their own study at Utah. “Supplies were still strained,” he says, “so we just went with what we had,” using empty 50-mL Falcon tubes at their drive-up sample collection sites. “That was basically our clinical trial. We wanted to get our head wrapped around what the numbers really looked like when we compared them to NP swabs.” The trial took about six weeks from start to finish. “Once we looked at the numbers we were quite happy with them,” he says. (Details of their work appeared in the Journal of Clinical Microbiology. Hanson KE, et al. 2020;58[11]:e01824-20.)

First of two parts. Next month: Journeys to alternative SARS-CoV-2 strategies

And yet they stopped the project—if only temporarily. Their contact at the FDA told them that if saliva concordance with NP swabs wasn’t 95 percent or higher, the lab’s EUA would be rejected. The lab’s numbers fell just short: 94.6.

Dr. Barker was frustrated and recalls plenty of back and forth with the agency. The other approved method currently in use—nasal swabs—was performing far worse (around 84 percent) compared with the NP swabs.

While he was trying to convince the FDA that saliva was viable, he was, as director of the reagent laboratory, looking into how to build kits. NP swabs weren’t the only supply casualty; everyone was running out of media as well. So the lab developed ARUP Transport Media, or ATM.

He was also evaluating different collection kits for saliva, calling its use “inevitable” as he watched supplies vanishing. The goal was to find a kit that would not only be easy for clients to use but also simple to make and source, the lab equivalent of a wrap dress. He likewise was in contact with his vendors to determine what tubes might fit on automated systems.

At one point the lab tested 2-mL versus 1-mL tubes. The smaller size wasn’t perfect, so they tried the larger tubes at one of the drive-through sites. The switch backfired, cutting the number of daily samples from 200 to 100. “People were taking forever to spit 2 mLs. So we went back and forth—what’s the ideal situation,” he says, given that nothing is ideal about having patients wait in their cars in long lines to spit into a tube.

Other numbers were much bigger. To maintain a steady stream of kit-making supplies requires about 7,000 funnels a day and 120,000 tubes a week. “And we have to make enough ATM to fill 50,000 tubes a week to keep up with demand.”

Once he felt secure about his supply lines, the lab again contacted the FDA, whose own recent updates suggested to ARUP that they would not have to submit an EUA. When the FDA agreed, “That was the day we decided we’re going to move forward with saliva.”

At the start of September, the lab converted its numerous drive-through collection sites at the University of Utah and its Salt Lake Valley health clinics. The sites collect about 1,500 samples daily. On a Friday night, he says, the lab distributed some 20,000 kits it had made and taught staff how to use them.

Such a testbed “is important for labs to think about,” he says, given that saliva is a hard matrix to work with. Indeed, the following day or two, after the lab started getting kits back from the university, “we saw a very high repeat rate that we didn’t see in our study.”

He blames a few things, mostly linked to the obdurate nature of the sample—saliva is as peevish as a Dickens shopkeeper. Patients were spitting more than the required allotment. “They were going way over the fill line,” Dr. Barker says. “And they were not just drooling saliva in, but they were going deeper, almost like sputum.” The protocol also asks patients not to eat or drink 30 minutes beforehand, but that wasn’t being enforced. It wasn’t, says Dr. Barker, a pretty picture. “All sorts of things that you can imagine with saliva.”

That required them to dilute the samples according to their validation. The lab had planned for repeats, given the nature of saliva, but the high numbers took them by surprise. Their study suggested they’d encounter a three to four percent repeat rate; in practice, the number was closer to 15 percent. At 9,000 to 10,000 samples, that high rate was problematic. More training ensued.

But that wasn’t the end of the challenges. ARUP runs three platforms: Roche, Hologic, and Thermo Fisher. Each uses a different extraction method, which appears to have affected the repeat rates. Roche performed well, at three percent, so the lab began routing all its saliva samples to that platform. ARUP had that ability to maneuver, managing to add the Roche instrument fairly early; likewise, the newly formed rapid response laboratory is completely Thermo Fisher-based, “because it’s really the only platform I could get en masse to get through our 15,000 samples a day.”

If he were considering saliva-based testing and lacked these multiple options, “I would not do it,” Dr. Barker says bluntly.

He’s equally brief when he talks about the keys to managing saliva: “Teaching, dilution, and shuttling it to the right instrument,” he says.

Once he had his come-to-Jesus moment with top administrators and realized there was no going back, he estimated that 20 to 30 percent of the kits they give to the university would be saliva. Instead, they’ve converted 95 percent from NP to saliva. TATs run about 30 hours. “I don’t see saliva going away,” he says. “I’m just trying to get our repeat rate down.”

Dr. Barker has made his peace with saliva and now talks knowledgeably and with some degree of comfort about the matrix that he was once ready to jilt at the altar.

Despite its challenges, saliva has some cheering aspects. Though he’s had no time to delve into the ever-changing numbers, he’s drawn some early conclusions, based on ARUP’s experience. “We think that saliva is very good,” he says. “We also think it will capture some cases that NP will miss, and vice versa. But we think they’re equivalent as far as sensitivities go.”

If the repeat rates can be kept low, he continues, saliva “would be superior for many reasons. It makes automation much easier. I don’t have to worry about swabs.” At current count, ARUP is receiving 35 different kits, he says. “We’re talking every swab known to mankind, in every media.” With that overwhelming variety, “It’s weird to say, but saliva’s the most consistent” element he now works with. “I see spit in the tube, and that’s what it is. We know they’ve collected enough, because they can see the volume. From a lab point of view, that makes it more trustworthy.”

The initial collection problems have largely disappeared, Dr. Barker reports. Three weeks into the launch, and with two weeks of clients sending in the saliva samples, things were falling into place. He credits the drop in repeats to the considerable training the lab did, including videos and photos.

His lab’s own study confirmed what other studies have shown, he says—that both saliva and the virus are stable. “We saw no degradation at room temperature for five days” with neat saliva, he says.

Nor did they see any degradation when ATM was added. “We did that study because we knew we’d be diluting the patient sample right off the bat,” he says. The tubes contain 1 mL of solution, diluting the saliva sample 1:1. Dr. Barker calls that ratio the sweet spot for sensitivity.

Moreover, “We wanted to make sure we didn’t see any degradation in the RNA virus,” he says. “That sounds like a simple thing, but it isn’t.”

That was, in fact, the stumbling block to his initial plan—just buy saliva collection kits, which might have gained him back some of those lost hours of sleep. “We were going to,” he says. “To be honest with you, we wanted the easy way out—just buy a kit and send it to our clients.”

They found, however, that many of the supposed DNA virus-protecting kits on the market contained a detergent, usually SDS, which degraded the virus. “If I had time to do my science work,” he says, in a familiar pandemic lament, “it’s more than likely that the detergent is ripping apart the lipid wall of the virus and exposing the RNA.” The lab would need to build its own kits, Dr. Barker reluctantly concluded.

The DIY model continues to evolve. The kit uses a funnel adapter to make it easier for patients to deposit their specimen, and early on Dr. Barker was able to secure that supply line. “But at the same time I commissioned a group in California to make a mold for me,” so ARUP could manufacture its own funnels. He anticipated it being ready in October.

Dr. Barker has in a sense taken on the understudy role of manufacturer. His motivation is self-evident: “I don’t trust the supply lines anymore,” he says. With ARUP owning its supply line, “Those molds give us about 150,000 funnels a week. And we won’t have to deal with companies that might run short.”

Contemplating the scene, he continues: “That’s what COVID does. When we get a nice supply line in from a company for one week, the following week it will be wiped out. That’s why we moved off on our own—making our own kits is the only reliable way I can make sure I get about 50,000 to 60,000 of them a week.”

Dr. Barker took his first steps toward saliva in March, when he saw on the news that northern Italy was shutting down. Knowing that Copan’s manufacturing plant was located in the region, “I remember turning to my wife and saying, ‘That’s where we get—where the whole world gets—swabs and media for viruses.’

“And the next day Copan called all of us and said, ‘What you have is all you’re going to get,’” he continues. ARUP’s clients—from all 50 states—started calling soon afterward, asking the lab to send everything it had in turn. The lab keeps about 1 million various kits on hand in its warehouse, which seemed sufficient at first. “But within a week we realized we weren’t going to make it through the month,” Dr. Barker says.