As described in our November, 1999 article on the regulation of HIV drug resistance assays, the FDA's Center for Biologics Evaluation and Research (CBER) is responsible for reviewing applications to market resistance assays that are produced and sold as kits (see related HIVresistanceWeb article). Assays that are developed and run within individual labs, known as "home brews," are not regulated by the FDA. The FDA regulates tests (considered "devices") according to three categories (class I, II and III). Each category requires progressively more detailed information to be submitted to the agency in order for the FDA to determine whether the device is safe and effective prior to marketing. On September 17, 1999, CBER convened a Medical Device Panel that recommended that HIV genotypic resistance assays be reclassified from class III to class II devices.

Class III devices require the submission of a Pre-Market Approval (PMA) application which includes data from prospective clinical trials, while Class I and II device applications [called a 510(K)] may or may not require such clinical trial-derived data. Class II and III devices are also subject to specified "special controls" that must be submitted and approved prior to clearance (in the case of a class II device) or approval (in the case of a class III device). When special controls are required, the FDA publishes a Guidance Document in the Federal Register that outlines the specific special controls and attendant requirements.

With the recent approval of Visible Genetics' TruGene genotyping assay as a class II device, FDA has now reclassified these assays as class II devices. In addition, the FDA has developed a draft Guidance Document that is published for comment in the Federal Register. The final revised version of the Guidance Document will also be published in the Federal Register for use by manufacturers in preparing their 510(K) submissions. Written comments regarding the draft must be received by October 29, 2001. (Note that FDA does not propose reclassification when a genotypic test includes un-approved analyte specific reagents (ASRs). ASRs remain class III devices when they are used as a component of a genotype test.)

There was significant discussion at the September 1999 CBER meeting regarding the inclusion of clinical trials data as a special control for the clearance of gentoype assays as class II devices. Clinical trials data is not required for submission in the current draft. The Guidance Document proposes two approaches to gaining marketing approval. The "Analytical Data Only" track requires rigorous demonstration of analytical sensitivity for specified mutations. The "Clinical Trial Track" requires less rigorous demonstration of analytic sensitivity of the same mutations provided data from clinical studies provides evidence that use of the test will provide medical benefit. There is a warning that "reliance on clinical data and less extensive analytical data may limit the claims of intended use set out in the labeling." In other words, if a clinical trial did not include all the mutations listed in the guidance document, FDA may determine that only those evaluated in the trial may be included in the labeling as approved mutations. Details of analytic sensitivity testing and the range of detectability (both plasma HIV RNA levels and proportions of a specific mutation in the viral swarm at which the test is effective) are outlined in the document.

Two categories of mutations are outlined in the draft Guidance Document, those proven to be associated with viral resistance to certain drugs ("established"), and those of interest but not yet proven to be associated with drug resistance ("implicated"). Analytical data must be presented for both sets of mutations, although FDA may clear assays with "limited claims for subsets of the mutations listed…if…doing so would benefit the public health," e.g., point mutation assays, which detect individual point mutations as opposed to any deviation from consensus sequences of wild-type RT and protease genes (see related HIVresistanceWeb article). The document states that FDA will work with the manufacturer to "determine the correlation between the use of the assay and benefit to the patient for mutations that are currently not generally recognized as being associated with HIV resistance to anti-retroviral drugs." Analytic sensitivity needs to be demonstrated for all mutations for which the manufacturer proposes to make a claim in its labeling.

The document recognizes that a critical component of overall assay performance is an interpretation algorithm and all submissions must include an interpretation algorithm. A "minimal interpretation algorithm" is provided in the guidance document, as well as guidelines for developing different interpretation rules. Unfortunately, the algorithm provided simply lists RT and protease mutations associated with antiretroviral resistance, but does not take into account mutational interactions. Thus, if it is used as is, it may provide inaccurate information. For example, there is no note of the impact of multiple "AZT mutations" on viral susceptibility to other nucleoside analogs, nor of the "resensitizing" effect of the mutation at codon 184 on some AZT and nucleotide analog mutations.

Finally, specific product labeling instructions are provided. Clinicians relying on tests cleared by the FDA will need to learn how to critically read and interpret these product labels as a number of clinically important limitations of any given assay will be outlined therein.

FDA's CBER Defining Regulatory Requirements for Genotypic Resistance Tests
(Michelle Roland, December, 1999)

Genotypic Assays for Determining HIV Drug Resistance
(Mike Kozal, February, 1999)