This story has been updated to include additional manufacturing capabilities of authorized reagent suppliers.
NEW YORK – Labs developing their own assays for SARS-CoV-2 virus detection in patients are coming across unexpected bottlenecks and hurdles as they rush toward US Food and Drug Administration emergency use authorization.
Nationally, some labs that are in the assay validation process are facing reagent supply shortages and lack of instrument access, as well as increased staffing needs. Individual labs are doing their best to overcome these obstacles, guided by frequent regulatory guidance updates from FDA, in order to make testing for coronavirus COVID-19, the disease caused by SARS-CoV-2 virus, available at local levels quickly.
As of Wednesday mid day, there were 1,101 confirmed cases of COVID-19 in the US, including 273 in Washington state, 212 in New York, 157 in California, and 92 in Massachusetts, but there are also reports that a lack of tests has prevented patients with symptoms from being diagnosed. The World Health Organization has declared the outbreak a pandemic, meaning it will likely spread across the world.
In order to meet an increased demand for clinical testing, the US Food and Drug Administration opened the emergency use authorization process to high-complexity CLIA labs on Feb. 28, creating a new pathway for regulated assay development in the SARS-CoV-2 emergency. Commercial labs are also beginning testing and pursing EUA under these guidelines.
This expansion is in addition to testing that relies on an EUA assay developed by the Centers for Disease Control and Prevention that is used at state and local public health labs. The CDC reported Tuesday that 78 public health labs are now running its test, with a current capacity in the labs to test 75,000 people.
The FDA has also extended the CDC's emergency authorization to cover certain kit lots from reagent manufacturers that have been validated at the CDC labs. The first two lots of kits are from Danaher's Integrated DNA Technologies (IDT). Another manufacturer, LGC Biosearch Technologies, had a kit lot validated and authorized earlier this week. LGC has said that it has manufactured 625 kits and has capacity to make 1,000 kits per week going forward, with each kit supporting 1,000 tests. IDT has also said that sufficient primer and probe kits have been manufactured and validated to enable over one million tests, and that the company expects to manufacture enough kits to enable 2.5 million tests this week, and five million tests each week starting March 16th.
FDA has said that these kits can be used both by public health labs and other high-complexity labs, as well as by diagnostics developers.
Validating with viral material
High-complexity labs can go a few different routes to bring on testing. They can run an already-approved protocol exactly as is, in which case labs do not need to seek EUA, FDA suggested in webinars with labs on March 2 and March 6.
However, the agency also seemed to suggest that any variations on the CDC protocol — even mixing and matching reagents from CDC and IDT kits, for example — would require labs to perform analytical and clinical validation and file for EUA. However, FDA has also referred labs to the agency's frequently asked questions page, which was updated recently to include certain acceptable protocol modifications.
If labs choose to make their own tests, they can begin testing patients as soon as the assay is validated to the standards of FDA's recent guidance and must then file for EUA within 15 days.
At this point in time, it appears that nearly every step in the overall workflow of the PCR test — from sample collection and extraction to thermal cycling — has the potential to cause labs trouble.
Specifically, the clinical diagnostic workflow that needs to be validated entails reagents like inactivated virus or viral constructs, extraction kits and instruments, real-time PCR systems, as well as positive and negative control materials.
On a good day, reagents can be ordered, or are already on hand in the lab, but now, there are signs that demand may exceed supply during the current emergency.
Also, with a novel pathogen like SARS-CoV-2, just having the viral nucleic acids available to test if the assay is working has been the first challenge.
The FDA guides that labs should determine the limit of detection using 20 samples, and subsequent clinical evaluation should include 30 contrived clinical specimens. Getting viral material to do this first step has stymied some labs, while others have found workarounds.
For example, in Seattle and California — the US areas initially hit hardest by the COVID-19 outbreak – validation in two high-complexity labs is complete and testing has already begun.
Alex Greninger is the assistant director of the University of Washington Medicine Clinical Virology Laboratory, a service lab which launched last year. He and his team have managed to overcome the limiting step of obtaining viral RNA.
"We used positive specimens, which we are awash in now," Greninger said in an email.
Meanwhile, Ben Pinsky and his colleagues at Stanford University are also well on their way to EUA and are testing patient samples in California. Pinsky, who is the medical director of the clinical virology laboratory at Stanford Health Care and Stanford Children's Health said that to validate its test, the lab used dilutions of a highly positive clinical specimen that was quantitated using single-stranded DNA as a calibrator.
"Once we had access to the positive clinical specimen, validation proceeded very rapidly," Pinsky noted.
Other labs developing tests may not have the same access to positive samples, however. There were many queries about this issue on the recent FDA webinars, specifically about whether labs might use naked viral RNA, or plasmids, and they were guided to obtain inactivated virus.
Greninger explained that spiking viral RNA into patient sample matrix does not make a good validation material, especially for limit of detection calculations, because it is quickly broken down by RNAses. "There is a reason viruses are encapsulated. … Snot is full of RNAses," he explained.
The CDC and FDA have been directing developers needing to obtain viral materials to BEI Resources, a reagent provider established by the National Institute of Allergy and Infectious Diseases, as previously reported.
BEI has said that it has live and inactivated virus in stock. Some end users have reported that there is a one-time step involving registration in order to receive shipments of virus, which may be a small bottleneck.
There are also intermediate suppliers, such as ZeptoMetrix, that can obtain live virus from BEI and inactivate it. ZeptoMetrix said on Wednesday that it has viral material ready to ship to customers.
Dwight Oliver, medical director of the molecular diagnostics laboratory at the University of Texas Southwestern's pathology department, said that his team started the process of creating a lab-developed test protocol this week.
"We plan to validate using whole viral genome, and [we are] waiting for that resource to arrive," Oliver said in an email.
He said he believes there are many labs currently waiting for the same viral genome or attenuated virus that his lab is also requesting. "I fear getting the viral RNA to spike into samples and determine our limit of detection will not be fast, but it is too early to tell," he said.
Sample types, collection methods, primer and probe choices
Assay design is a series of choices, based on science and, in the current situation, constrained by ever-evolving federal regulatory guidance.
Validating each sample type that a lab expects to receive used to be a potential sticking point, given the short supply of viral materials and time pressure.
However, the FDA guided that labs do not need to test every type of specimen but instead can validate testing only on "the most challenging" specimen type that they expect to run, with sputum considered to be the extreme.
Sample collection is also an issue. There are labs that want to use nasal swab specimens but have patients self-collect them, either in the clinic, at home, or possibly as a "drive-through" service.
On the webinar, FDA noted that labs that wish to validate self-collected samples for at-home collection would need to do a shipping study, involving having patients collect a sample and then essentially having the lab mail the sample to itself, to demonstrate there is no difference in results.
An example of this might be a project at UW, funded by the Gates Foundation, that is piggybacking on a self-collection study for influenza testing. However, for that study, Greninger said his lab is not testing self-collected samples yet, as they are focused on scaling up SARS-CoV-2 testing to meet the required volume first.
Labs are also exercising a range of options in developing the primers and probes for a test.
Oliver at UTSW said his lab will compare three assays — one combining a unique N gene and published Orf1a target, another using the IDT research-use-only CDC kit, and a third using multiplexed IDT and CDC primers and probes — and will then use the most robust test.
The UW and Stanford teams, meanwhile, have settled on different methods.
The UW lab validated both the WHO and CDC assays and found the WHO assay to be very sensitive, according to Greninger. But, having the IDT kit lots available made the CDC kit "a great choice" because "it has relieved all supply chain issues," he said.
Stanford, meanwhile, validated a modified version of the test developed by the Drosten group at Charité Virology in Germany, which was the first to be publicly available. The test, which is the basis for the WHO test, targets the SARS-CoV-2 envelope gene, or E gene, and the RNA-dependent RNA polymerase gene, or RdRp gene, Pinsky said.
For this protocol, Pinsky said the Stanford group has not yet encountered reagent supply issues.
"However, as testing increases globally, we anticipate reagents may be more difficult to obtain," he added.
Extraction and instrument bottlenecks
The Stanford lab is now working on expanding testing capacity, which will require validation of high-throughput nucleic acid extraction instruments.
This could also present a bottleneck. In the FDA webinars, labs asked about workarounds to address shortages and backorders in human extraction control samples from CDC, and FDA has since updated its FAQ to include an alternative that involves labs using human RNA that they extract themselves from specimens or cultured cells.
UTSW's Oliver noted that his lab has not yet encountered any shortage of commercial reagents but that he was aware that some of the assay instrumentation for the CDC protocol, including automated RNA extraction and real-time PCR instruments, is on back order from the instrument manufacturers.
Extraction instrumentation and reagents have also been a bottleneck for a few labs participating in the recent FDA webinars who wish to use the CDC protocol but didn't happen to have a Qiagen setup. In its FAQ, the FDA has recently added other instruments and kits from Roche that might substitute for the Qiagen kits.
In the meantime, at least one researcher has registered dismay on social media that the Qiagen extraction kits might become a bottleneck. Michael Mina, an epidemiologist at Harvard School of Public Health, tweeted this week that "with the test kits finally arriving in labs, we enter a new phase. Now, we are finding a shortage of a critical component required for the test to run (RNA extraction kits). These commercial kits are quickly becoming out of stock."
Qiagen replied to Mina via Twitter, saying that supplies of materials and components so far are sufficient, but that "extraordinary demand for coronavirus testing workflows, which include [Qiagen] products, is challenging our capacity for certain items and may lead to backorders with delayed delivery or reduced allocation of affected products."
Thomas Theuringer, senior director of corporate communications at Qiagen, expanded on this in an email yesterday. The current demand is "challenging our capacity to supply certain products used for SARS-CoV-2-related LDTs, and we are doing our best to manage our supply chains and meet the needs of customers in the most timely way possible," he said.
Specifically, Qiagen has ramped up production at two of its manufacturing sites in Europe, "moving to three shifts working seven days a week," he added.
Qiagen is encouraging the use of larger kit sizes to allow for more sample preps to be manufactured, and it is also increasing the utilization of its manufacturing capacity in Maryland to address the situation.
"We have also put a task force in place that is evaluating the incoming orders," Theuringer said, adding that the firm is working directly with customers to understand their flexibility and specific needs in order to be able to ensure broad availability of products.
Similarly, in terms of real-time PCR instruments, labs had been directed by FDA that Thermo Fisher Scientific's Applied Biosystems 7500 Dx is the preferred instrument to run the testing and that even performing a software upgrade to make an existing instrument compliant with its guidance would likely require labs to perform a validation and EUA application.
However, the most recent FAQ now states that certain instruments that were previously cleared to run CDC's RNA-based influenza panel could be used, as well.
The acceptable instruments now include the Applied Biosystems 7500 Fast Dx Real-Time PCR Instrument with SDS software version 1.4, the Applied Biosystems QuantStudio Dx with version 1.0.3 software, and the Qiagen Rotor-Gene Q MDx with AssayManager version 1.0.4.1 and Epsilon version 1.0.1 software.
Although this expansion is an improvement, it still doesn’t cover the instruments his lab has, said Linoj Samuel, division head of clinical microbiology at the Henry Ford Health System in Detroit, Michigan.
"We do not have the CDC-listed platforms, so unfortunately we will have to file an EUA," he said in an email. Samuel noted that his lab is developing essentially its own version of the CDC assay and expects to have the commercial kits it needs this week.
The increased demand for testing also seems poised to highlight a well-known short supply of skilled technicians.
To meet demand, Qiagen is bringing on new staff, as are the labs at Stanford and UW. Training new staff takes time, however, as does training existing staff to run a newly acquired instrument or kit.
At Roswell Park Comprehensive Cancer Center in Buffalo, New York, Jan Nowak, clinical chief of molecular pathology, said his lab is validating a test under the New York State Department of Health EUA. But he foresees a bottleneck in finding and training a sufficient number of medical technologists "to perform the assay, [and] to accommodate the volume of tests that I suspect we will see soon."
The path forward, in historical context
The regulation of lab-developed tests has been a huge issue in the industry for some time. A new regulatory framework, called the Verifying Accurate, Leading-edge in vitro clinical tests (IVCT) Development (VALID) Act, was proposed last year and was introduced by Congress last week.
How the regulatory aspects will all play out remains to be seen, and the big picture is hard to see in the weeds of a global public health emergency. The authors of an article on diagnostic testing for novel coronavirus published in JAMA this week suggested that FDA should "reassess the new approach as testing becomes more available, and transparently review its experience to develop future policy for public health emergencies."
In addition, the experiences of the H1N1 pandemic a little over a decade ago could inform the current SARS-CoV-2 situation, according to Roswell Park's Nowak.
H1N1 was also a global emergency that instigated implementation of emergency-use diagnostic testing, with labs and commercial entities stepping up to develop tests. The CDC received EUA on May 2, 2009 after the emergency was declared in the US on April 26, and commercial tests started receiving EUA about five months later.
Labs were also able to develop their own tests in that emergency. "During the 2009 H1N1 flu epidemic, FDA did not prohibit the use of LDTs, and community hospital laboratories played a major role in testing for that virus," Nowak said in an email.
He and others wrote a review article after the H1N1 emergency on the role that community molecular diagnostics labs played. H1N1 was "the first pandemic in the age of molecular diagnostics," they wrote, and it begged questions about preparedness and oversight in the lab community.
The conclusion at the time was that labs did a great job. They were able to pioneer tests specific for H1N1 and deploy them rapidly to begin screening — an aspect that, along with surveillance testing, was initially delayed in the current SARS-CoV-2 outbreak due to issues with the first batch of CDC kits.
Indeed, the H1N1 review suggests that the "rich resource" of skilled molecular diagnostics experts was "generally undervalued, unappreciated, and … underutilized." It also noted that a survey by the Association for Molecular Pathology at the time revealed that labs relied on more than five different commercial assays, both analyte-specific reagents and in vitro diagnostic tests, and some also used noncommercial laboratory-developed tests.
A decade ago, the labs' H1N1 assay validation process was not regulated by FDA, and the new oversight is unprecedented.
"FDA enforcement of the EUA requirement for [the current] testing was unfortunate and short sighted," Nowak said. He added that the strategy has "seriously delayed the availability of testing in the US."
Still, the current FDA oversight and evolving guidance is being managed and adapted to by labs as they move forward to face the unexpected.
Pinsky said that since the new guidelines for LDTs were issued, "the process has gone very quickly, and the FDA has been very responsive."
At the UW lab — which is now trying to accelerate its throughput, and is testing hundreds of samples each day — Greninger noted that the team faced an unexpected hiccup this past weekend, during the daylight savings time transition.
"One of our extractors went down briefly when the clocks shifted," he said. The lab had previously never needed to run the instrument at 2AM and didn't anticipate that the instrument's clock could be an issue. Fortunately, "no harm, no foul, and it's back up and running," he said.
Now, his lab needs to submit its EUA this week. "I question whether the best use of my time is to file an EUA right now, given the critical needs of more testing. But we follow the rules," he said.