Most COVID-related EUAs backed by low-quality data, review finds

FDA sign and building
FDA sign and building

FDA / Wikimedia Commons

Most US Food and Drug Administration (FDA)-issued emergency use authorizations (EUAs) for COVID-19–related medical products were not supported by high-quality data, and the FDA revoked authorizations for 17 products because of effectiveness or safety problems, according to a research letter yesterday in JAMA Internal Medicine.

The authors say that relying on suboptimal supporting data is understandable at the beginning of a pandemic, but standards should be raised later on.

EUAs allow the FDA to expedite medical products to market under public health emergency declarations from the US Department of Health and Human Services. After the first EUA for an anthrax vaccine was issued in February 2005, few medical products were authorized until after the emergence of COVID-19, when the FDA used EUAs to speed product availability.

Researchers from Tel Aviv University studied COVID-19–related EUAs issued from Jan 1, 2020, to Jan 22, 2021, on the FDA website and categorized them by their trial designs. The team ascertained product EUA status on Oct 1, 2021.

Most devices had little or no supporting data

Among the 393 COVID-19–related medical products given EUAs, 329 (84%) were diagnostic tests, 54 (14%) were medical devices, and 10 (3%) were drugs or vaccines.

The EUAs for diagnostic tests were mainly supported by comparisons with different previously authorized assays (254 of 329 [77%]), while some relied on analytical or in vitro studies (47/254 [14%]). Most medical devices (32/54 [59%]) provided no supporting evidence, and those that did cited only lab data (20/54 [37%]).

In contrast, of the 10 drug and vaccine EUAs, 7 (70%) were supported largely by randomized clinical trials, and 8 (80%) had not been previously approved by the FDA for other indications.

The FDA revoked the EUAs of 17 products (4%) because of effectiveness or safety issues (10 [59%]) after a median of 230 days. These products included 2 drugs, 7 medical devices, and 9 diagnostic tests.

A well-known example was the antimalarial drug hydroxychloroquine, whose EUA was supported by only low-quality retrospectively collected data, the authors noted. The FDA approved 1 drug (remdesivir) and 1 vaccine (Pfizer-BioNTech) during the study period.

Call for higher standards of evidence

The authors noted that, because of the lack of diagnostic product testing against standard criteria and the absence of correct identification of participants who had or did not have COVID-19, their true sensitivity and specificity and the real-world frequency of false-positive and -negative results are only estimates.

"Authorizations that were supported by such low-quality data are reasonable at the beginning of a pandemic, when diagnostic products are urgently required," the researchers wrote. "In future public health emergencies, the FDA should consider raising the standard of evidence required for EUAs after several diagnostic products are marketed to ensure their accuracy."

Medical devices authorized without supporting or clinical data could result in consumer harm, the researchers added. "The rapid uptake of potentially beneficial technologies should be balanced against device-related safety problems," they wrote.

The study findings could inform regulators of the current status of EUAs and help guide improvement efforts. "Regulators should resist political pressure to authorize products that are not supported by high-quality clinical data," the authors concluded. "Doing so might prevent confusion and mistrust in the soundness of medical science and bolster the FDA's reputation."

This week's top reads