On October 18, the Food and Drug Administration (FDA) issued a letter reminding developers of in vitro diagnostics (IVDs) that institutional review board (IRB) approval is required before initiating a clinical investigation involving human subjects if data from that clinical investigation will be submitted to the FDA. While that proposition will strike most as unremarkable for an interventional study, many may find it surprising that the FDA believes IRB approval is required even for retrospective research involving only leftover, de-identified specimens.

Many laboratories still perform IVDs under what is commonly referred to as the “laboratory developed test” (LDT) exception. The LDT exception primarily allows laboratories to conduct IVD testing within their laboratory without FDA oversight. However, more and more laboratories want to sell their tests directly to consumers or for use as point-of-care testing outside of the laboratory, making them a manufacturer of an IVD. The FDA regulates the IVD as a medical device in these situations, subject to the same pre-and post-market regulatory requirements as other medical devices. Often, this requires providing the FDA with clinical data showing that the IVD is safe and effective for its intended use.

Clinical investigations of medical devices are ordinarily subject to the investigational device exemption (IDE) requirements of 21 CFR Part 812 and the informed consent and IRB approval requirements at 21 CFR Parts 50 and 56, respectively. However, many IVD studies are exempt from most provisions of 21 CFR Part 812, and in 2006 the FDA issued a Guidance expressing its intent to exercise its enforcement discretion not to require informed consent for certain IDE-exempt IVD studies using only leftover, de-identified (or in some cases coded) specimens.

In addition, the federal “Common Rule” – the regulations at 45 CFR Part 46 governing federally-supported human subjects research and, historically, all human subjects research conducted at many research institutions – has never required IRB review of research using only leftover, de-identified specimens.

Because the Common Rule and the FDA’s regulations on IRB review are identical in many respects, many research institutions and IRBs assume that these studies are not subject to IRB review, unaware of subtle differences between the Common Rule’s definition of “human subjects research” and the FDA’s definition of “clinical investigation.” Comments submitted in response to the FDA’s Human Subject Protection: Acceptance of Data from Clinical Studies for Medical Devices proposed rule show that some IVD manufacturers and trade groups shared this assumption, partly due to internal inconsistencies in various FDA regulations.

The Secretary of Health and Human Services (HHS) has missed a deadline imposed by Section 3023 of the 21st Century Cures Act to harmonize the FDA human subject protection regulations with the Common Rule. The FDA previously indicated that it intended to work with others at HHS in carrying out this statutory directive, including with respect to de-identified human specimens. Nevertheless, it has been evident since the FDA finalized the proposed rule mentioned above that the FDA considers research involving human specimens, including leftover specimens that are de-identified, to be “clinical investigations” within the meaning of its IRB review regulations if data from the research will be used to support an IDE, device marketing application, or another submission to the FDA.

Unless and until the FDA’s IRB review regulations change due to the HHS-wide harmonization efforts, organizations planning to use stored specimens in support of FDA submissions related to an IVD should obtain IRB review of their study protocols and any relevant policies and procedures. The IRBs should be looking to ensure that all of the following are true:

  • The investigation meets the IDE criteria at 21 CFR 812.2(c)(3).
  • The study uses leftover specimens, that is, remnants of specimens collected for routine clinical care or analysis that would have been discarded. The study may also use specimens obtained from specimen repositories or leftover specimens previously collected for other research purposes.
  • The specimens are not individually identifiable, i.e., the subject’s identity is not known to and may not readily be ascertained by the investigator or any other individuals associated with the investigation, including the sponsor. If the specimen is coded, it will be considered not individually identifiable if neither the investigator(s) nor any other individuals associated with the investigation or the sponsor can link the specimen to the subject from whom the specimen was collected, either directly or indirectly through coding systems.
  • The specimens may be accompanied by clinical information as long as this information does not make the specimen source identifiable to the investigator or any other individual associated with the investigation, including the sponsor.
  • The individuals caring for the patients are different from and do not share information about the patient with those conducting the investigation.
  • The specimens are provided to the investigator(s) without identifiers and the supplier of the specimens has established policies and procedures to prevent the release of personal information.

IVD manufacturers sponsoring studies should ensure these elements are met to facilitate IRB approval. In addition, if any of the above elements are not met, the study will not qualify for enforcement discretion under the FDA’s informed consent regulations, application of which would make studies involving leftover specimens infeasible.

If you have any questions about research using de-identified specimens and compliance with the FDA guidance, please contact the authors.