Bottleneck: FDA COVID-19 Emergency Use Authorization reviews
From toilet paper and testing swabs to elected and appointed leadership, the pandemic has exposed the many ways our systems have been unable to respond quickly and effectively at scale to the plethora of challenges. This post will examine a testing regulatory bottleneck that does not have the social media appeal of empty supermarket shelves, yet still has a real impact on dealing with the pandemic and attempts to resuscitate the crippled economy.
Most readers who have gotten this far will be familiar with the FDA’s Emergency Use Authorization (EUA) process that is called into play when circumstances warrant to allow diagnostics to rapidly be approved for use after evaluation of testing on limited sets of real or contrived disease and control samples. Recent examples include: 2016 (Zika virus, 14 tests approved), 2015 (Enterovirus D68, 1 test approved), 2014 (Ebola virus, 10 tests approved), and 2013 (MERS coronavirus, 2 tests approved). Note the small number of tests approved under these past EUAs; we will discuss the SARS-CoV-2/COVID-19 numbers shortly. Note also that the normal non-emergency route of FDA approval for a disease test can take months or years.
To make this post more accessible to a non-specialist audience, here are some brief and simplified explanations of key concepts (expert readers can skip this paragraph): There are 2 general categories of tests for viruses and bacteria that cause human disease. Molecular tests (PCR and all its variants) look for the RNA or DNA of the causative agent in a human sample. Serological (antibody) tests look in human blood for the various types of antibodies that are generated by the human immune system in response to assault by a disease-causing agent. Each category of test has its strengths and weaknesses. PCR tests may be negative during the earliest days of infection (the pre-symptomatic phase) but tend to be robust during much of the symptomatic phase. For some diseases, the causative agent may be cleared by the body (and PCR tests thus are negative) before the patient is fully recovered. One type of antibody (IgM) generally becomes detectable about a week after COVID-19 infection but generally fades within another ~2 weeks. A longer-lasting antibody (IgG) generally becomes detectable about 2-3 weeks after infection; how long it lasts is still being determined at this time. Thus, a combination of PCR and antibody tests are useful for detecting COVID-19 infection given these rough time windows. No tests are perfect; for a wide variety of reasons, false positive and false negative results occur with both molecular and serological tests. In general, PCR tests have greater sensitivity than serological tests because PCR involve molecular signal amplification. There are high-throughput and low-throughput and point-of-care formats of both molecular and serological tests. Serological tests (particularly IgG) are useful for population surveys because they can identify people who had COVID-19 but whose symptoms were not bad enough for them to have been diagnosed with a PCR test at the time (or they were sick at a time when PCR testing for whatever reason was not available to them.) Much remains to be learned about how this disease affects humans and how our immune systems respond to it.
Recall the small numbers of EUAs issued in prior emergencies, as listed above. COVID-19 in contrast has completely overwhelmed the FDA. All numbers listed here are derived from FDA web pages and were accurate as of June 26, 2020:
93 Commercial Molecular tests for COVID-19 have received EUAs
37 Laboratory Developed Molecular tests for COVID-19 have received EUAs
23 Serology tests for COVID-19 have received EUAs; (only 6 are for lateral flow rapid tests)
183 Serology tests are awaiting EUA (~140 appear to be rapid lateral flow tests)
53 Serology tests are listed as “Should no longer be distributed for COVID-19”
Note that the FDA does not list how many molecular tests (Commercial or Lab-developed) are still awaiting EUA. At least 130 COVID-19 molecular tests have been approved to date; 29 of which were approved from 6/1/20-6/26/20. In that same period, 9 serological tests were granted EUA. Only 23 serological tests have been granted EUA to date, with a backlog of 183 as of 6/26/20. May saw a total of 39 EUAs granted (34 molecular and 5 serological.)
The FDA has clearly been overwhelmed by the unprecedented volume of EUA applications for this pandemic. They have been remarkably transparent, holding weekly Town Hall calls for the past 14 weeks where they have noted that they are seeing a 60X increase in applications over normal. Their staff of examiners has already been doubled by bringing in other trained FDA staff, but reviewing tests that can have life or death implications is not a skill that can be gained by watching a few YouTube videos. Some tests have data from extensive testing and use in parts of the world where COVID-19 struck early; others have data only from limited sets of contrived or real samples. Inconsistent or missing metadata, particularly about the number of days since the onset of symptoms to the test date, further complicate the analysis and review.
The FDA has also been completely transparent (if somewhat confusing) about their priorities, stating that high-throughput tests (both molecular and serological) and point-of-care tests (molecular) are highest priority, “along with home sample-collection and home testing” (no home testing for COVID-19 has yet been approved as of this writing.) This large jumble of “highest priority” applications leaves rapid serological tests as the clear losers (only 6 granted EUA to date out of the 153 total EUAs granted.)
Assuming the FDA’s approval rate of ~40 EUAs/month for May and June does not increase, and considering the current backlog of 183 serological tests alone (remember, the FDA does not list the backlog of molecular tests), it would take another 4.5 months to clear this backlog even if no additional EUA applications are filed and no additional molecular tests are still in the queue. We know for sure there is at least one: in early June, the CDC announced that they had submitted for EUA a new PCR assay that combines COVID-19 with Flu A/B. It will be interesting to see how long that approval will take, given that it will presumably be given high-priority treatment by the FDA to prepare for the Fall beginning of flu season. (UPDATE: EUA was granted in early July 2020 for this assay, which will be in use nationwide by the Fall.)
Not surprisingly, the developers of the 183 serological tests stuck in the backlog are unhappy (especially the ~140 developers of low-priority rapid tests), even though current FDA rules allow them to market their devices while awaiting their EUA ruling. Given that the FDA has already listed 53 serological tests that should not be distributed (either they failed to submit EUA data or their data indicated sub-par performance) the market is now telling the manufacturers whose EUAs are still pending to “come back when you have been granted EUA.”
Much like the toilet paper shortage, this regulatory clog will eventually pass. The FDA may train more reviewers and find ways to accelerate the review process, and at some point, the rate of EUA submissions will lessen. The market will eventually sort out the winners and losers as it always does.
We can hope that the FDA will learn some lessons that can be applied to future emergencies where a similar scale of EUAs can be anticipated. The EUA templates that were eventually produced for COVID-19 should be readied for use for the next virus that emerges out of some jungle to threaten us all. Electronic templates, particularly for data submission, should be used instead of Word templates. These would permit rapid automated triaging of submissions (likely passes, certain failures, and ones that need close human inspection) to augment the current method of assigning a human lead reviewer to each Word document submission. The FDA needs to sort out how serological tests should be handled: the change in attitude from March to April from “EUAs are not required to market serological tests” to “EUAs are now required but you can still market until we say you cannot” led to the market confusion that is still rampant today. FDA priorities for review need to be determined and publicly posted, along with data on the sizes of the various queues for review. But all that can happen later when (we hope) effective vaccines have gotten this pandemic under control. For now, let us applaud the FDA for doing the best they can, given the current unprecedented circumstances, and hope for the best.
(Tom Slezak had a long career leading bioinformatics teams at Lawrence Livermore National Lab. He now does independent consulting and advising of biotech companies, including one whose COVID-19 rapid serological test is stuck in the FDA EUA backlog discussed in this post.)