Medical Device & Diagnostic Industry
Magazine [3] |
Originally published April 1996
A number of FDA actions and proposals over the last several months appear to indicate a more flexible and, hopefully, more practical approach to the regulation of in vitro diagnostic (IVD) devices. Communication has improved noticeably between industry and the Division of Clinical Laboratory Devices (DCLD), the division within FDA's Office of Device Evaluation that clears or approves IVDs prior to marketing. In this improved atmosphere, companies can more easily discuss proposed clinical studies and premarket submissions with the agency. The agency encourages both conference calls and face-to-face meetings. DCLD has even suggested meetings with companies to examine questions and concerns about proposed products or studies and to exchange scientific information with company scientists and clinicians. Although DCLD has always met with regulated companies to discuss submissions, recently the agency appears to have a more open attitude and is making a greater effort to cooperate with industry.
At the same time, FDA has raised the bar for attaining regulatory clearance of a number of IVDs, particularly regarding the kinds and amount of clinical data required. In fact, DCLD has indicated that some Class II devices will be sent to a device advisory panel for review and possibly to a formal panel meeting. These are similar to those for premarket approval applications, and requirements for clinical trial data and panel review seem to reflect the concerns DCLD has raised repeatedly regarding the effects of IVDs on health and safety.
Notwithstanding these somewhat mixed signals, a number of developments in the IVD regulatory arena warrant close attention. A new compliance policy guide (CPG) addressing the commercialization of unapproved IVDs labeled for research or investigational use will likely be made public for comment in 1996. This will coincide with FDA's proposal for regulating analyte-specific reagents (ASRs). The ASR initiative, designed to address the difficult issue of home brew products, was recently endorsed by an FDA advisory panel. The agency is now working on a Federal Register notice that will formally present its proposal for regulating home brew ASRs.
Industry may also benefit from DCLD's willingness to take a second look at applying FDA's triage initiative to IVD submission reviews.1 DCLD's initial assessment of the initiative was that most IVDs requiring only a tier 1 review had already been classified as Class I premarket notificationexempt devices. It is now, however, working with industry to develop new models for assessing the risk posed by a diagnostic device.
Another promising development is FDA's proposed policy for simplifying submission requirements for primary and secondary reagents on automated analyzers. FDA has sent the proposal to manufacturers for comment and expects to finalize it in 1996.
These initiatives and their implications for the device industry are the focus of this article.
RUO AND IUO COMMERCIALIZATION
The commercialization of IVDs labeled for research use only (RUO) or investigational use only (IUO) has been a particularly thorny issue for FDA and IVD companies. An appropriately labeled diagnostic device can be shipped to investigators without an approved investigational device exemption (IDE) application if the testing "(i) is noninvasive, (ii) does not require an invasive sampling procedure that presents significant risk, (iii) does not by design or intention introduce energy into a subject, and (iv) is not used as a diagnostic procedure without confirmation of the diagnosis by another, medically established diagnostic product or procedure."2,3
By the early 1990s, when FDA issued a draft CPG on the commercialization of IVDs, a few IVDs distributed under this exemption had become widely used standard tests for diagnosing or monitoring some conditions or diseases. Recognizing that the abrupt removal of these devices from the market would be disruptive to the practice of medicine, FDA proposed that these IVDs be placed on an accommodation list. Listed devices would be allowed to remain on the market for a specified period while safety and effectiveness studies were conducted to prepare premarket submissions. FDA abandoned this approach, however, after its general counsel advised that an accommodation list might be illegal.
FDA's draft CPG, released in August 1992 under the title "Commercialization of Unapproved In Vitro Diagnostic Devices Labeled for Research and Investigation," described a certification program under which these diagnostic devices could be distributed.4 The certification program was to provide an audit trail from the manufacturer or importer through the distributor and laboratory to the end-user, confirming that these devices were not sold or diverted for unapproved uses.
In the draft CPG, FDA indicated that home brew products would be subject to the same regulatory requirements as unapproved medical devices. Such products are brewed in the laboratory by combining components or reagents. These components and reagents are often sold in bulk for further manufacturing, and in some cases are modified for in-house clinical assays. If they are modified, they must be validated according to the laboratory quality regulations of the Clinical Laboratory Improvement Amendments of 1988 (CLIA).5
Because laboratories were not shipping finished devices across state lines, FDA had not previously regulated home brew IVDs. Although its authority to do so has been questioned, FDA has asserted its jurisdiction over IVD components, including reagents intended for use as finished medical devices, whether they are labeled for stand-alone use or not.
In the years since FDA issued the draft CPG, industry has wrestled with whether and how to comply with it. Some companies distributing RUO and IUO devices have chosen not to establish a certification program, and many of these companies have played a waiting game to see whether they would escape FDA's notice. Some struggled with the fairly significant logistical issues of instituting such a program. These issues include:
* Obtaining written certification of compliance from laboratories and individuals.
* Procuring copies of clinical protocols from independent investigators.
* Determining whether distribution to a site should cease because of noncompliance, which can create customer relations issues.
* Maintaining what can be a massive paper trail, depending upon the number of products distributed under certification programs.
Rumors that a revised CPG would be released created a dilemma. Companies had to decide whether to proceed with a certification program or wait to see what a new guidance might entail. In the meantime, FDA did not issue any new guidances on home brew products and therefore seemed to be backing off its original stance that these products would be regulated as unapproved medical devices.
In October 1995, the IVD industry got an indication of FDA's current position regarding commercialization of unapproved IVDs. At a meeting of the Association of Medical Diagnostics Manufacturers, Betty Collins, deputy chief of the IVD branch of the Office of Compliance (OC) of FDA's Center for Devices and Radiological Health, presented an overview of a revised CPG proposal. Collins stressed that FDA considers commercialized IUOs in violation of the Federal Food, Drug, and Cosmetic Act if they are labeled or promoted as safe and effective prior to agency clearance or approval, or if they are distributed to persons other than those conducting a clinical investigation to collect safety and effectiveness data.
Under the proposed CPG, as described by Collins, FDA will provide a period of "enforcement discretion" for most companies distributing IUO IVDs. A company would have 30 months from the effective date of the CPG to gather safety and effectiveness data to support a 510(k) notice, a premarket approval (PMA) application, or an IDE application, and to obtain clearance or approval for the device or IDE application. If, at the end of the 30-month period, FDA has not cleared a product or approved an IDE application, it would remove the product from the market.
The proposal includes conditions for receiving the 30-month grace period. If a company has not already begun a clinical study for a device once the CPG becomes effective, it must do so within 6 months. This means that a company must make a good faith effort to obtain institutional review board (IRB) approval for the study, initiate the clinical study with a reasonable protocol, and demonstrate that it intends to monitor the study to ensure compliance with the protocol and IRB requirements. Furthermore, a company cannot label or promote the product in any way that indicates it is safe and effective. If the company makes any claim through labeling, advertising, conference booth displays, or even a salesperson's oral comments regarding performance characteristics such as sensitivity or specificity, the device could be considered misbranded and adulterated and therefore subject to regulatory action.
The proposed CPG's 30-month grace period, however, would not automatically be available to all IVDs. Devices not receiving it would include the following:
* IVDs that may lead to a significant medical decision without confirmation by another medically established procedure or diagnostic product, and for which adequate safety and effectiveness data are not available.
* IVDs for which data to support a human diagnostic use either do not exist or are only anecdotal.
* IVDs sold by companies that fail to abide by FDA's conditions for receiving the 30-month grace period.
It is likely that OC would consult with DCLD as to what constitutes a significant medical decision, and entirely possible that DCLD's assessment of a product's effect on medical decision making will differ from that of the manufacturer. DCLD's position has generally been conservative, and top division personnel have alluded to FDA's experience with a wide range of devices that can affect health and safety, compared to a single company's more limited perspective.
The proposed CPG has not been finalized and could undergo further modifications before it is released, particularly because it relates to FDA's initiative on ASRs. A final version that incorporates most of what has been described would offer a reasonable approach to companies seeking clearance for their IUO devices. Assuming that companies comply with FDA's conditions, they would have time to collect data for 510(k) submissions and to obtain clearance. Gathering data and obtaining approval for a PMA application within 30 months is clearly more challenging, but not impossible. Nevertheless,while studies are under way, FDA is unlikely to remove products from the market that have become important adjuncts to diagnostic armamentaria unless they pose a health and safety risk. Conducting studies and preparing premarket submissions for IUO IVDs requires resources, but addressing compliance issues also exhausts valuable resources. For its part, FDA will be under pressure from industry to review and act on these submissions to avoid unnecessarily removing products from the market.
HOME BREWS
After the 1992 draft CPG was released, little was heard from FDA regarding its intentions to become more actively involved in regulating home brews. In January 1995, however, the agency conducted a video teleconference on IVDs. During the teleconference, Steven Gutman, DCLD's director, indicated that although the agency might have some role to play in regulating home brews, laboratories are already regulated by the Health Care Financing Administration under CLIA.
For their part, laboratories have faced a different dilemma. Products that have already been validated, manufactured, and labeled in compliance with government regulations may be unavailable because the potential income may not justify the manufacturer's cost to obtain FDA approval. On the other hand, creating home brew tests is risky and resource intensive. Under the CLIA regulations, laboratories are required to validate the performance of a home brew or modified manufacturers' assay, a task that can entail significant work for more-complex diagnostic tests. Most laboratories prefer to use FDA-approved assays without modifying them.
Fortunately, several solutions have been proposed to streamline the device approval process, including reclassifying certain cancer diagnostic tests from Class III to Class II, down-classifying many transitional and preamendment Class III devices to Class II, and automatically placing new IVDs into Class II rather than into Class III. In December 1995, the Immunology Devices Advisory Panel met on a manufacturer-sponsored petition and recommended to FDA that it reclassify tumor markers for patient monitoring from Class III to Class II. DCLD appears to be actively looking for ways to use the 510(k) clearance path for devices that heretofore were likely to require a PMA application. Clearance of these products would decrease the need for laboratories to brew many tests on their own.
CLASSIFYING ASRs
FDA has also pursued alternative regulatory approaches for handling home brew products. In 1995, industry worked closely with FDA on a proposal to create a subcategory for special-purpose reagents (SPRs) within the general-purpose reagent category. SPRs, which could have included polyclonal and monoclonal antibodies, antigens, and genetic probes, would not have been stand-alone products and would not have been labeled for specific medical uses. Under the proposal, SPRs would have been regulated as Class I devices that are exempt from 510(k) notification requirements but subject to general controls, including registration, listing, and compliance with good manufacturing practice (GMP) requirements as well as labeling restrictions.
FDA subsequently chose not to create a subcategory within the general-purpose reagent category but to proceed with a proposal to classify these products as IVDs. At an advisory panel meeting on January 22, 1996, the agency presented its proposal to classify analyte-specific reagents as Class I devices that are exempt from premarket notification requirements but subject to general controls. In addition, FDA proposed placing certain labeling restrictions on ASRs.
After reviewing definitions of analyte-specific reagents proposed by both FDA and the American Association of Clinical Chemistry, the panel agreed on the following definition: "Antibodies (both monoclonal and polyclonal), specific receptor proteins, nonhuman nucleic acids and fragments of nonhuman nucleic acids, and similar biological reagents which, through specific chemical binding or reaction, are intended for identification or quantitation of specific analytes in a biological specimen."6 The panel voted in favor of the classification proposal, concurring with FDA that ASRs used in blood banking and in certain infectious disease tests should be excluded from the Class I, premarket notificationexempt category. The panel also recommended expanding the exclusions to cover ASRs used in genetic disease tests and tests to screen for predisposition to disease. These products would thus be subject to all FDA device and diagnostics regulations.
Other panel recommendations included restricting the sale of ASRs to laboratories with high-complexity-testing certificates under CLIA. It also recommended restrictions on labeling. Proposed language for ASR labeling is expected to be included in the Federal Register notice regarding the proposed classification of ASRs. FDA plans to publish the notice after reviewing the panel meeting proceedings.
Although the ASR proposal appears to be a positive step toward alleviating the regulatory uncertainties associated with home brew products, it raises new concerns for some product manufacturers. Under the proposal, companies that are not manufacturing finished IVDs, but are making components that qualify as ASRs, would be subject to GMP requirements. Their counterparts in the device industry that manufacture other components are not subject to the GMP requirements. This issue, as well as the exclusions from the ASR classification and the labeling restrictions, will likely generate considerable discussion during the upcoming notice-and-comment rule making process.
TRIAGE MODELS
Along with proposals relating to reclassification and ASRs, FDA and industry have jointly discussed the use of the agency's triage initiative introduced in 1993. Under this initiative, FDA reviewers categorize devices into one of three tiers, based on the agency's assessment of the risk posed by the device, and subject the premarket submission to a level of review commensurate with that risk. Tier 1 devices represent the lowest level of risk, and submissions generally undergo a labeling review. Tier 2 devices undergo both a labeling and a scientific review, and tier 3 devices undergo a more intensive labeling and scientific review. Currently, submissions for most IVDs are subjected to tier 2 reviews; a few IVDs receive tier 1 reviews.
FDA and the Health Industry Manufacturers Association (HIMA) have proposed decision trees for determining review levels. These decision trees were discussed at an October 1995 workshop at which FDA convened a panel of advisory committee members, government agency personnel, and representatives from professional associations. Because the workshop was announced in the Federal Register only the day before it was held, industry attendance was minimal. FDA's purpose in holding the workshop was to solicit opinions from workshop participants rather than to achieve consensus or gather recommendations.
During the workshop, HIMA representatives took the position that a test using a new methodology for detecting a previously cleared analyte in the same matrix in which it was cleared could be sufficiently reviewed as a tier 2 device. They also voiced concerns that FDA's model fails to adequately address whether an IVD is used as a primary determinant of diagnosis or treatment or as an adjunct to other diagnostic information. In addition, FDA's model does not allow for products not used in professional settings to be reviewed as tier 1 devices. Some panel members presented a more conservative position, indicating that FDA should not place point-of-care and physician-office laboratory IVDs in tier 1. They expressed concern that a tier 1 review could mislead users into thinking the device had received a more thorough review than a tier 1 review entails. FDA is expected to form an agency-industry task force to formulate proposals for simplifying IVD review processes or to develop a pilot initiative.
The IVD industry has generally believed that FDA has subjected its devices to more-stringent premarket reviews than other devices that pose potentially greater risk. FDA's willingness to engage in dialogue on this issue is encouraging. In addition, FDA has maintained that it is looking for ways to focus its resources on newer and more complex submissions. Allowing a tier 3, Class II 510(k) submission for an IVD that might otherwise require a PMA application would ease the burden on companies with newer technologies and "first of a kind" devices. The required clinical data would still undoubtedly be significant. Additionally, having 510(k) notices reviewed by an advisory panel would be more complicated than the submission process for most 510(k)-cleared devices. The alternative of submitting a PMA application, nevertheless, is much more daunting. If DCLD assigns the tier 3 designation injudiciously, however, industry would still face ever-increasing delays in product clearances.
REAGENTS USED IN AUTOMATED ANALYZERS
One of the more exciting developments in the IVD regulatory arena has been FDA's proposal to simplify submission requirements for reagents used in automated analyzers. The proposal's introduction, in the version dated October 3, 1995, described the draft guidance as an effort to "clarify, standardize, and streamline" data requirements for reagent and analyzer system applications.7 Such an effort would be welcome to the IVD industry, which has faced requirements that seemed to differ from branch to branch within DCLD. Existing submission requirements, outlined in the proposal, likely came as a surprise to some secondary reagent manufacturers who may have been unaware that they should have been submitting 510(k) notices for use of their reagents on each analyzer for which they distributed reagent application sheets.
Under the proposal, entitled "Data Required for Commercialization of Primary, Secondary and Generic Reagents for Automated Analyzers," submission requirements would become simpler for both primary and secondary reagent manufacturers and for manufacturers of "open" analyzers who wish to distribute reagent application sheets. Primary reagents are defined in the proposal as reagents produced or obtained by an analyzer manufacturer specifically for use on that analyzer. Secondary reagents are those produced for use on specified analyzers that are open systems.
Requirements remain the same for both generic reagent manufacturers and manufacturers of analyzers for which no specific claims are made and for which no reagent application sheets are distributed. These manufacturers must still submit a 510(k) notification showing compliance with labeling regulations and demonstrating the performance characteristics of reagents for a given analyte on one analyzer.
The proposal requires that primary reagent manufacturers continue to submit 510(k) notifications when introducing a new reagent and instrument system employing a new operating principle, technology, or intended use. The manufacturer would include a protocol by which subsequent models of the same system would be tested. FDA must approve the protocol and, if changes are made, the manufacturer must submit a new one. Once a new model has been tested according to the approved protocol and prior to its marketing, the manufacturer would send a letter to FDA that references the initial 510(k) submission, describes the new model, and provides a 510(k) summary or statement. FDA would file this letter with the original 510(k) and send an acknowledgment letter to the manufacturer. The manufacturer must maintain a complete record of the testing performed to qualify the new model.
Under the proposal, secondary reagent manufacturers (or manufacturers of analyzers on which secondary reagents can be used) would submit an initial 510(k) notification for an analyte on an analyzer, and would include a protocol for qualifying the reagents on additional analyzers (or for qualifying additional reagents for the same analyte on the same analyzer). Once testing is completed using the FDA-approved protocol and prior to making specific claims or distributing reagent application sheets, the manufacturer would send a letter to FDA that references the initial 510(k) submission, lists the new qualified analyzers (or new reagents for the same analyte), summarizes the performance testing, and includes the new reagent application sheets. FDA would add this letter to the original 510(k) file and send an acknowledgment letter to the manufacturer. Again, the manufacturer must maintain a complete record of the testing done to qualify the new analyzers or reagents.
Given the permutations of reagents and analyzers that exist, a number of questions arise regarding how FDA would implement the policy for existing reagents and analyzers, how long FDA would take to approve protocols, what protocol changes would require resubmission, and how long FDA would take to provide an acknowledgment letter. Moreover, the policy does not address assay reagents cleared for manual use but not for use on automated analyzers. Some manufacturers have maintained that if no assay parameters are changed when the assay is automated, then no submission should be required. FDA's position has appeared to vary, depending upon the DCLD branch.
Notwithstanding the issues still to be addressed, the proposal appears to demonstrate a willingness by DCLD to simplify the submission process. The proposal has undergone three revisions in an effort to address questions raised by industry, indicating FDA's commitment to putting this policy into effect.
CONCLUSION
Recent proposals in IVD regulation are encouraging in light of the chasms that existed between FDA and industry not long ago. Nevertheless, lengthy approval times still exist and clinical data requirements are increasing significantly for many IVDs. For its part, the IVD industry must take an active role in examining proposals that affect its products and in drafting proposals to address emerging issues on which FDA has not yet focused. Device companies should not assume that other manufacturers will provide feedback to FDA that reflects their own positions. They should also not assume that when FDA or professional associations solicit comments, the comments will be ignored. The time for offering constructive input to proposals on the table for discussion, or for offering alternative solutions, is now. FDA seems to be listening.
REFERENCES
1. "PMA/510(k) Triage Review Procedures," General Program Memorandum G94-1, Rockville, MD, FDA, Center for Devices and Radiological Health (CDRH), May 20, 1994.
2. Code of Federal Regulations, 21 CFR 809.10(c).
3. 21 CFR 812.2(c)(3).
4. "Commercialization of Unapproved In Vitro Diagnostic Devices Labeled for Research and Investigation," Draft Compliance Policy Guide, Rockville, MD, FDA, CDRH, August 1992.
5. Clinical Laboratory Improvement Amendments of 1988, 42 USC 201, October 1988.
6. "Analyte Specific Reagents' Class I, 510(k)-Exempt Status Endorsed by Panel; Few GMP Inspections of ASR Manufacturers Foreseen by FDA," MDDI Reports--The Gray Sheet, pp 35, January 29, 1996.
7. "Data Required for Commercialization of Primary, Secondary and Generic Reagents for Automated Analyzers," draft, Rockville, MD, FDA, CDRH, October 3, 1995.
Constance Finch is a regulatory specialist for the law firm of Hogan & Hartson (Washington, DC). She was formerly with Baxter Diagnostics and with Abbott Diagnostics Div. of Abbott Laboratories.