College of American Pathologists
Printable Version

  Feature Story


cap today

FDA puts ASR rule back on the table

CAP genetic testing oversight efforts

October 2003
Karen Lusky

A visitor to the FDA recalls once seeing a placard in someone’s office there that read: “In God we trust . . . Everyone else must bring data.”

“Everyone else” has, to date, not included clinical laboratories that develop and offer their own high-complexity tests, known colloquially as homebrews, which are regulated by the Clinical Laboratory Improvement Amendments program. Yet the Food and Drug Administration announced in August that it has a proposal on an expedited track that could pave the way for the agency to review laboratory-developed tests, including genetic tests.

The FDA rule is likely to take into account recommendations by the former Secretary’s Advisory Committee on Genetic Testing, says Steve Gutman, MD, director of the FDA’s Office of In Vitro Diagnostic Device Evaluation and Safety. “SACGT suggested a risk-based regulatory approach with maximum transparency that will not chill new technology. The committee also strongly emphasized that FDA work be collaborative with involved stakeholders,” he says.

The College will join other clinical laboratory groups and the FDA to debate the oversight issue at an FDA roundtable next month. Representatives from both sides of the fence appear to agree on at least one major point: Someone should be looking at the clinical validity and utility of laboratory-developed genetic tests, the majority of which screen for or predict disease.

“Currently CLIA only addresses the analytical validity of clinical laboratory testing, not the clinical validity,” says CLIA program director Judy Yost. “The latter has to do with whether the test works on the patient.” In other words, does the patient who tests positive for a condition or disease actually have it and, if so, will that person become symptomatic?

The College advocates strengthening the existing regulatory and accreditation infrastructure as the best tack to ensure the quality of lab-developed testing without impeding new technology, which can take years to wend its way through the FDA.

The CAP is, in fact, putting the finishing touches on a conceptual framework for genetic testing oversight that it originally presented at an FDA roundtable in May 2002. Debra Leonard, MD, PhD, chair of the College’s Molecular Test Validation Project Work Group, gave the Clinical Laboratory Improvement Advisory Committee an update on the proposal at its September 2003 meeting in Atlanta (for more information, see “CAPgenetic testing oversight efforts,” page 20). The College’s proposal includes adding new questions to the Laboratory Accreditation Program checklist, as well as enhanced proficiency testing programs for genetic testing.

“The thrust of the concern [and the College’s efforts] has been in the [clinical] validation area in that a lot of these molecular diagnostic tests are in-house-developed and have not been FDA-approved,” reports Ronald Lepoff, MD, chair of the CAP Commission on Laboratory Accreditation.

The College focused on oversight of genetic testing because there are few homebrew tests in other parts of the laboratory, Dr. Lepoff adds. “So it’s not clear that the CAP needs to generate a whole series of questions for chemistry or hematology, as examples.”

The College is moving ahead with its proposal to strengthen oversight of genetic testing regardless of what happens on the regulatory front, affirms Phil Bongiorno, CAP assistant director of public health and scientific affairs.

FDA eyeing ASR loophole
The FDA says it is considering plans to strengthen oversight of laboratory-developed tests by reopening its 1997 analyte specific reagent, or ASR, rule. In its existing form, the ASR regulation allows CLIA-certified clinical laboratories to use ASRs as individual building blocks for developing genetic and other assays. Laboratories are required to develop and maintain the test’s analytical (but not clinical) performance. They also must report test results with the boilerplate disclaimer: “This test was developed and its performance characteristics determined by [laboratory name]. It has not been cleared or approved by the FDA.”

Under the ASR rule, test manufacturers are not required to seek FDA premarket approval for class 1 (low-risk) ASRs, which consist primarily of the active ingredients for genetic tests. To qualify for the regulatory exemption, the manufacturer cannot make analytical performance or clinical claims for the ASR. Nor can it provide clinical labs with instructions on how to use the ASR.

Class 2 and 3 ASRs represent a small subset used in blood banking tests for infectious disease (for example, certain cytomegalovirus tests) or in the diagnosis of potentially deadly contagious diseases (HIV or TB). The FDA designated HIV and TB as the highest risk, or class 3, ASRs, which require test makers to get the reagent approved as part of a test kit through a premarket approval or 510(k) before they can sell the individual reagents separately.

Dr. Gutman says the FDA may rewrite the ASR rule using a risk-stratification approach but not based on any notion of “genetic exceptionalism.

“Instead,” he says, “the risk stratification would be based more on the risk of an individual test to the patient than the fact that the test is a genetic test per se.”

Device makers push the envelope
Some observers say the FDA was forced to revisit the ASR regulation when test makers began marketing avant-garde ASRs that walked and talked more like full-blown test kits, which require 510(k) or a premarket approval application.

The ASR rule did fail to predict all of the permutations for marketing of ASRs that the FDA has since seen evolve, Dr. Gutman says. “These include the use of custom devices, the use of closed systems, the use of open systems, and the use of both commercially purchased ASRs as well as active ingredients made by the same laboratory developing the lab tests,” he says. In his view, open systems provide the flexibility envisioned for ASRs. “Closed systems provide less flexibility and seem more consistent with test kits than with ASRs,” he says. “The closer one gets to a building block for a lab-developed test, the more consistent that is to at least the spirit of the ASR regulation.”

By definition, an ASR is one reagent (or perhaps a single pair of matched reagents) or a building block for an in-house-developed or homebrew test, agrees Daniel H. Farkas, PhD, HCLD, associate professor of pathology at Baylor College of Medicine and director of molecular pathology at The Methodist Hospital, Houston. Yet Dr. Farkas cites at least two examples, both of which happen to be genetic tests, of “what look, smell, and feel like” full-blown kits being marketed as ASRs. One of the tests is used to manage cystic fibrosis; the other is the AmpliChip CYP450, a pharmacogenomic microarray announced by Roche Diagnostics in June of this year.

Dr. Gutman, in fact, wrote a letter to Roche in July inviting company officials to “explain the basis for their apparent conclusion that the DNA chip is properly regulated under the ASR provisions of FDA regulations.” The letter maintains that, based in part on the company’s press release announcing the product, the DNA chip appears to be a medical device requiring premarket review.

“The FDA’s response to the Roche product was the first evidence that the agency is putting its foot down,” says Wayne Grody, MD, PhD, director of the UCLA Diagnostic Molecular Pathology Laboratory and a member of the College’s Molecular Test Validation Project Work Group. Dr. Grody suspects that the Roche press release, which included some claims about the chip, seemed to trigger the FDA’s reaction, rather than the product itself.

Roche’s June 25 press release said AmpliChip microarrays can include tens of thousands of individual DNA pieces, called probes, assembled on a thumbnail-sized glass plate, functioning like “gene antennas.” The statement said that Roche planned to initially sell the AmpliChip CYP450 microarray as an ASR in the United States for use by CLIA-certified high-complexity laboratories performing CYP2D6 and CYP2C19 genotyping tests. The press release also said Roche expects the test to be available as an in vitro diagnostic in the United States and Europe in late 2004.

Shadow cast by shady labs
The FDA is concerned with “truth in labeling,” which Dr. Gutman says is “very consistent with where the commissioner wants to go . . . as the FDA is here to help consumers be informed about what they are getting.”

And the current regulatory framework does leave the door open for so-called fringe laboratories to develop and market tests with no known scientific validity. All they have to do is establish the tests’ analytical performance and report test results using the boilerplate disclaimer the ASR rule requires.

“These ‘fringe’ labs do tests beyond the boundaries of traditional laboratory medicine,” says Dr. Lepoff. “Most of the tests are not genetic tests, but they are unusual applications of traditional tests to non-traditional areas.” Examples include in-house-developed assays for functional disorders, such as Alzheimer’s disease, autism, and sleep disorders, which actually require extensive clinical examination and testing to diagnose.

“The vast majority of testing done under ASRs is aboveboard and well done, but it is abused by the fringe elements,” says Richard Friedberg, MD, PhD, chair of the College’s Patient Safety and Performance Measures Committee.

To address that issue, the CAP implemented a policy this year that will prevent the organization from inspecting and accrediting labs whose primary test menu includes tests outside the College’s areas of medical expertise. “That means testing beyond what we would usually consider as part of allopathic medicine,” Dr. Lepoff says.

CLIA may play a role
In addition to the CAP’s genetic testing oversight, enhanced regulations under CLIA ’88 could help assuage the FDA’s concerns about the clinical validity of lab-developed genetic tests. “CLIA has an important role in terms of ASRs and genetic testing, and that role is also under scrutiny in terms of how it can be strengthened,” Dr. Gutman says. “Since CLIA’s regulatory tools are different than those available to the FDA, some reasonable combination of both might be used to address the concerns regarding genetic testing.”

Yost reports that the Centers for Medicare and Medicaid Services and the Centers for Disease Control and Prevention will issue a proposed rule on genetic testing probably next year based on CLIAC recommendations.

CLIAC has recommended changes to CLIA regulations to introduce a subspecialty in genetic testing. “If CLIAC’s recommendations are adopted in the upcoming proposed rule, there would be new requirements to ensure genetic tests’ analytical and clinical performance,” says former CLIAC member Andrea Ferreira-Gonzalez, PhD, director of the Molecular Diagnostics Division, Virginia Commonwealth University, Richmond. “The genetics subspecialty would also have specific personnel requirements for individuals overseeing the testing process, and it would provide different avenues for proficiency testing programs.”

Yost says someone with the right expertise should be overseeing the clinical validity and utility of genetic testing so it’s not just “caveat emptor.” Yet she doesn’t see the need to invent a new bureaucracy for that purpose. “We have to be careful to avoid duplicating oversight efforts,” Yost says. “For example, if CLIA were to do clinical reviews of lab-developed tests, and the FDA did some form of premarket reviews, that’s double work. Also, CLIA oversees labs, not manufacturers. And when a lab takes an ASR and creates a new test system, the lab becomes a manufacturer. CLIA is not really intended to oversee that [function].”

Industry to the rescue?
The diagnostics industry is floating a proposal for discussion at the November FDA roundtable that it believes could provide a way out of the regulatory quagmire. The model involves what some are calling a 510(k) “lite” approval process for analytical kits too complex to qualify as class 1 ASRs, but lacking the clinical validity data to satisfy existing 510(k) or premarket approval application requirements.

“Using the IVAT [in vitro analytical tests] process, a device maker would only have to establish the analytical validity for a test to obtain permission to market a product,” explains Rick Naples, vice president of regulatory submissions, reimbursement, and government affairs for Roche Diagnostics.

Here’s how it would work: A manufacturer would develop a kit and assess its analytical performance against a particular target—for example, SARS or another emerging pathogen. “The test maker would develop analytical characteristics for the test to define how it would perform against samples that actually contained the pathogen, and make a claim that the test could, in the case of SARS, as an example, detect the corona-virus,” Naples explains. “But that’s as far as the manufacturer could go with the product claims.”

At that point, the clinical laboratory would pick up the ball. “The test maker would provide labs with the standardized analytical kit, which it had been granted permission to market by the FDA, to use in developing clinical validity parameters. The parameters establish how the test might best be used with the lab’s patient population,” Naples says.

The IVAT approach is meant for any new test, but Naples believes it would be used mostly for genetic or low-prevalence infectious disease tests.

While the College was still reviewing the proposal at CAP TODAY press time, the IVAT idea was generating positive feedback from key College members involved in the issue, says a CAP staffer.

Dr. Friedberg says the IVAT proposal would “at least ascertain that the kits must be used as intended, which is not the case with ASRs. The FDA’s role would be to determine characteristics of analytical utility, which with ASRs only happens at the individual lab.”

Dr. Ferreira-Gonzalez likes the IVAT proposal because it would allow manufacturers to give labs not only the components necessary to perform a test, but also the instructions for their use. “Since the IVAT would be a complete kit, it would come with specific instructions that will need to be followed in order to achieve performance. This would provide an avenue to compare results from different labs and allow us to gain more insight not only about the analytical performance of the test, but also the clinical validity of the test,” she says. “When every lab is performing a test in the same manner, you can compare results and gain more information in terms of diagnostic utility of the test.”

Roche predicts the IVAT model, if adopted by the FDA, could actually head off premarket review of laboratory-developed tests. “If the analytical kits were to become widely used,” Naples says, “the FDA might reason, OK, at least we have seen their analytical performance, so there is truth in labeling that labs can rely on in developing in-house tests. And ultimately, when clinical consensus is reached on how the test should be used, the manufacturer will probably be coming back for full-blown 510(k) clearance or PMA approval, and labs would have the complete package.”

Naples concedes that the IVAT proposal does not address FDA concerns over how labs establish clinical validity of testing. He maintains, however, that the FDA could work closely with CMS and the College to develop guidelines for clinical laboratory validation of analytical kits. “Then CLIA and/or programs with deeming status, such as the College, could take over and look at that test in-house to see how the lab validated it analytically and clinically,” Naples says. “The test maker would have the analytical portion covered, but the lab would look at why the doctor ordered the test and whether it really predicts disease.”

All bets are on
The FDA is reviewing the IVAT proposal. Yet how the agency will address its concerns with ASRs and lab-developed tests is anyone’s best educated guess. “FDA is considering all options and working with other agencies within Health and Human Services to make sure it develops a reasonable path forward,” Dr. Gutman says.

In Dr. Ferreira-Gonzalez’s view, the worst-case scenario would be for the FDA to require labs doing homebrew assays to register with the agency and submit their in-house-developed tests to some form of premarket review. “Even if it is a lighter version of premarket review or an expedited one, the lab would still have to put an application together for FDA review. And that could possibly affect how fast labs can bring tests on line, which would ultimately affect patient access to new tests,” she says.

Dr. Farkas agrees: “Technology has presented the molecular pathology laboratory with the opportunity to do high-quality tests. It would be disappointing if the FDA passed the buck by applying inappropriate bureaucratic measures on labs because it cannot keep pace with technological advancement.”

Karen Lusky is a writer in Brentwood, Tenn.