FDA has often dismissed allegations of inspectional inconsistency as more perception than reality. Now, IVD companies are coming forward with examples to prove that such charges are more than just a figment of industry's imagination.

In recent years, members of the medical device industry--including IVD companies--have frequently alleged that FDA is very inconsistent in its inspections of manufacturers. Such allegations are commonly supported by accounts of individual inspections, and encompass both the inspectional practices of the agency's investigators and the outcome of those inspections. Whether or not these allegations are true, the fact that they have been made has chilling consequences for all those involved with the industry and its products.

Allegations of inconsistency are especially problematic for the manufacturers of devices and diagnostics by making it difficult to identify the practices that FDA considers to be industry standards. Most companies know their own practices well, and they may also be aware of the practices of competitors and neighboring companies with respect to manufacturing, labeling, advertising, and so on. Where FDA has established regulations or guidelines to govern such practices--and where industry follows them--individual companies can be comfortable that there is a level playing field. But when existing regulations and guidelines are not uniformly understood, applied, or adhered to, companies often become concerned about whether their practices are in compliance with FDA regulations and about whether overzealous enforcement is putting them at a disadvantage relative to their competitors.

With respect to FDA, allegations of inspectional inconsistency raise several questions, the most obvious of which relate to the objectivity and fairness of the agency and its investigators. Such allegations may also lead to questions about the effectiveness of the agency in educating and controlling its investigators, and in educating regulated industry. Most regulated companies honestly desire to meet regulatory requirements, either because they recognize the business advantages of compliance or because they want to avoid sanctions. But in order for a company to satisfy any set of requirements, they must be clearly conveyed and fairly applied. A regulatory agency that is viewed as inconsistent or unfair will not be respected by industry, and this, in turn, can reduce a company's capacity to learn from the agency as well as its willingness to comply with regulations.

For the purchasers and users of medical devices and diagnostics--and for the patients whose well-being may depend upon them--allegations of inconsistency raise concerns about product quality. Observations made during an FDA inspection may lead end-users to incorrectly conclude that one company lacks the quality systems that its competitors possess. Inspectional inconsistency may also cause end-users to question whether any manufacturer's prod- ucts can be relied upon for quality and consistency. In an era when consumers are extremely sensitive to charges that manufacturers have disregarded public safety, any such suggestion can have a devastating effect on the market for otherwise efficacious medical products.

To FDA's credit, complaints about inspectional inconsistency seem to be less common now than in previous years. But because the results of such occurrences can be so serious, it is still important to identify instances of inconsistency, to evaluate their causes, and to develop corrective actions. To accomplish this, it is essential to look not only at the present but also at the recent past, when complaints of unfairness were at their peak.

Shifting Procedures

In September 1990, FDA issued a statement of enforcement policy indicating its intention to protect the public health through the use of both administrative and judicial enforcement mechanisms. In May 1991, the agency announced two procedural changes intended to strengthen those mechanisms: streamlined procedures that delegated more authority to field offices and reduced requirements for headquarters review; and a new warning letter to replace both the notice of adverse finding and the regulatory letter previously used by the agency. These changes brought with them a host of complications.

Field Authority. The authority that was delegated to FDA district offices permitted them to issue warning letters for a variety of good manufacturing practices (GMP) and other violations without first sending them to be reviewed by the Office of Compliance at the Center for Devices and Radiological Health (CDRH). The intent of this change was to provide companies with quicker notice of violations that district offices considered significant.

Although FDA acknowledged that this change would reduce the ability of the device center to exercise the oversight functions that it had previously carried out, at the time the agency accepted this trade-off in order to speed up its enforcement process. Industry, however, did not necessarily consider the trade-off an even deal. In the view of many device manufacturers, this change in FDA procedure resulted in less consistency, both in the agency's inspectional practices and in other enforcement activities.

Warning Letters.Prior to the 1991 changes, FDA had used two types of letters to communicate warnings to inspected companies. The regulatory letter informed a company of serious violations, and indicated that judicial action would be taken unless it responded satisfactorily within 15 days. The notice of adverse finding was used to inform manufacturers about minor infractions that the agency expected them to correct without threat of further enforcement. Although the regulatory letter was a powerful tool (FDA rarely had to go to court in those days), the agency was less satisfied with the notice of adverse finding. In FDA's view, industry did not regard such notices as having potentially serious consequences, and thus frequently failed to take corrective actions as expected.

By replacing the regulatory letter and the notice of adverse finding with the single warning letter, FDA intentionally blurred the distinction between serious violations and minor infractions. Its intent was to put industry on notice that any such findings were considered very significant, and that it was prepared to take whatever enforcement action might be necessary to bring companies into compliance.

From 1991 through 1994 the number of warning letters issued by FDA increased significantly over the number of notices and letters issued in previous years. For a company in the United States, receipt of a warning letter usually also caused the company to be placed on FDA's so-called reference list. In turn, this resulted in disqualification for (or cancellation of) federal contracts, discontinued review of product applications pending at the device center's Office of Device Evaluation, an increase in inspectional frequency, and other results perceived as punitive. Equally important, because warning letters were issued without regard to the seriousness of the violation, they often caused end-users unjustifiable concern about a company's manufacturing practices and the quality of its products.

During this period, device manufacturers outside the United States were also inspected by FDA, many for the first time. In dealing with these firms, the agency adopted the same strategies it had implemented for U.S. companies. But when a warning letter was issued to a foreign company, it often also resulted in an import ban against its products.

Depending upon the source, a variety of explanations have been forwarded for the increased number of warning letters issued from 1991 through 1994. FDA has said that the increase was a result of more strenuous efforts to identify and take action against companies that were not meeting GMP requirements. Some in industry have suggested that the agency considered the issuance of warning letters to be a measure of its effectiveness, and increased its use of them in order to address congressional concern about weak enforcement. Others have said that the increase merely demonstrated FDA's inability to educate and work with companies. Whatever the reason, warning letters and related inspectional issues rapidly became a significant concern for industry. As a result, companies became more sensitive to real or perceived unfairness or imbalance in the inspectional process.

In 1995 these issues also came to the attention of the newly elected Republican Congress, which quickly made it clear that FDA practices would be high on its list of reform priorities. In response to this pressure, the agency modified slightly its approach to enforcement activities and began looking for ways to work more effectively with companies. Although FDA has not changed its enforcement mechanisms, the number of warning letters issued since the beginning of 1995 is substantially lower than in previous years. It is not clear whether this reduction is a result of better compliance efforts on the part of industry, fallout from FDA's earlier enforcement initiatives, or some other cause.

The Baltimore IVD Roundtable

The Baltimore IVD Roundtable is a group of FDA and IVD industry trade association representatives that meets quarterly to identify and discuss issues of interest to the diagnostics industry and to the medical device industry generally.

The group originated in June 1995, when IVD industry trade associations asked Kenneth Shelin, director of FDA's Baltimore district office, to set up a meeting with FDA representatives. The purpose of the meeting was to explore ways FDA and industry could work better together to achieve common goals.

At that initial meeting, discussion focused on the distrust that had grown between industry and FDA, the sources of that distrust, and possible means of reconciliation. Since then, the roundtable has continued meeting to discuss new issues of interest.

Guideline Inconsistencies

During the period when inspection-related complaints were at their peak, FDA also initiated a number of policy changes specific to the IVD industry. These were embodied primarily in the Guideline for the Manufacture of In Vitro Diagnostic Products, which was published in January 1994. Originally drafted in the mid-1980s, this document had undergone two prior revisions and had long been in the hands of FDA investigators.

Although FDA investigators had access to the guidance before its final publication, they had not been trained to interpret it and were advised not to use it when inspecting companies. Nevertheless, some investigators made use of the guidance prematurely, and as a result the use and interpretation of the document varied widely among investigators and districts.

For example, some investigators assumed that the practices described in the guidance were current industry standards, while others did not. This represented a source of inconsistency with potential to have a significant impact on a company's financial stability. Another source of problems was the fact that compliance with the guideline was considered to be voluntary, which may have led to confusion and misunderstanding by both investigators and industry. A manufacturer that was seeking to match the practices described in the guidance--whether by choice or because an investigator insisted upon it--could find itself investing heavily to develop systems that were not required of its competitors. The following paragraphs offer only a few examples of such potentially bank-breaking practices.

In common with earlier drafts, the final IVD manufacturing guideline classified IVD products as sterile, microbiologically controlled, or microbiologically uncontrolled, depending upon the level of control necessary for the product to meet its stated performance characteristics (section 3.1.1). To achieve an appropriate level of environmental "cleanliness," however, could require a manufacturer to undertake significant changes in its facilities, equipment, processes, and personnel practices (3.4). After publication of the IVD manufacturing guideline, a few companies implemented appropriate environmental controls in accordance with the guideline, while many others did not.

Another area of concern is process validation. Although FDA did not explicitly mention process validation in its 1978 GMP regulation, sections of the regulation had long been interpreted as requiring device manufacturers to conduct this activity whenever appropriate. These requirements were further spelled out by the agency in its 1987 Guideline on General Principles of Process Validation. To these general requirements, the IVD manufacturing guideline added the flat statement that "all new IVDs and/or processes are to be prospectively validated" (3.2). This requirement imposed a new burden on IVD manufacturers--but only if they elected to comply with the guideline or were required to do so as a result of FDA inspections. Other IVD manufacturers were not affected.

Another section of the IVD manufacturing guideline focused on stability studies and expiration dating (3.9). It indicated that FDA would accept only real-time data in support of stability studies, and spelled out a variety of environmental conditions that should be considered for such studies. While some companies undertook to meet these requirements, others did not.

Finally, the IVD manufacturing guideline imposed preproduction controls (3.1.1) and called for the application of statistical techniques to the sampling and testing of IVD products (3.8). For other types of medical devices, however, these two requirements will only become effective when the agency's new quality system regulation goes into effect this June 1.

FDA investigators had access to drafts of the IVD manufacturing guideline even before it was published, and they used them to identify what they believed to be significant deficiencies in company practices relative to the rest of the industry. But FDA investigators did not apply the guideline uniformly, nor did they share with industry a common understanding of its requirements. In conjunction with other inconsistencies in FDA's inspectional practices, these factors eventually had a multitude of adverse effects, including the regulatory side-effects of noncompliance, and, for some companies, financial and competitive disadvantages.

The Baltimore Report on Inspectional Consistency

At the winter 1995 meeting of the Baltimore IVD Roundtable, one of the participants noted that inspectional inconsistency had been identified as a significant issue that merited resolution.

Meanwhile, FDA had already gathered a nationwide task force to investigate such allegations. Under the direction of Ron Johnson, former director of the Office of Compliance at FDA's device center and now director of the agency's Pacific Region, the task force included a number of compliance officers and investigators. The mission of the group was to seek out and identify evidence of inspectional inconsistency.

Despite the efforts of the task force in reviewing establishment inspection reports and FDA-483s for examples of inconsistency, none were identified. In the summer of 1996, the group reported that it was prepared to complete its work and to conclude that inspectional inconsistency in the medical device industry is more perceived than real.

Learning at its summer 1996 meeting of the task force's imminent report, the Baltimore IVD Roundtable formed a working group to address the issue of inspectional consistency with a focus on the IVD industry. Members of the working group include Ken Shelin (Baltimore district office, FDA), Leif Olsen and Ray Watkins (Association of Medical Diagnostics Manufacturers), Paul Touhey (Medical Device Manufacturers Association), Bill Gilbert (Independent Reagent Manufacturers Association), and Pat Shrader (Joint Council of Immunohistochemical Manufacturers).

The working group on inspectional consistency adopted the following objectives for its work:

• To identify and present examples of inspectional inconsistency in the medical device and diagnostics industry, including inconsistencies among districts and investigators, in order to demonstrate that inspectional inconsistency is a real rather than a perceived problem.

   

• To identify sources of inspectional inconsistency in order to best identify appropriate corrective actions.

   

• To recommend steps that FDA can take, in conjunction with industry, to help ensure both the reality and the perception of greater inspectional consistency.

Working group members solicited information directly from their association members and from other sources. The information requested included examples of inspectional inconsistency as well as recommendations for FDA and industry efforts to address the issue. The report summarized in the accompanying article is the result of the group's efforts.

Inspectional Inconsistency

The inspectional inconsistency that concerns medical device manufacturers takes many forms, few of which are easy to discover. The following examples were contributed by member companies of IVD industry trade associations and other companies in industry. Some come from the contributors' personal experiences, while others are based on their communications within the industry. At the request of the contributors, the companies are not identified, nor are the specific FDA district offices that were involved in the inspections. All of the following information is a matter of public record. For further information, companies may contact the trade associations that participated in the Baltimore working group (see box below). A citation or explanation of the inconsistency is given in parentheses following each example.

1. A company on the West Coast was cited for failure to have a revision date on a label. The product in question was being manufactured for another company, which had supplied the labeling. All of the company's own products had revision dates on the labels. Because the noncompliant labeling was supplied by the company that contracted for the manufacture of the product, the inspected company felt the citation was inappropriate. (21 CFR 809.10(b)(15).)

2. A company on the West Coast was inspected by FDA and no FDA-483 was issued. The investigator returned to the company after the close-out meeting and asked to inspect the documentation relating to an investigational product. The company informed the investigator that investigational products are exempt from GMP requirements and that therefore the investigator did not have a right to review the documentation for this product. The investigator agreed that the company was correct, but told them his supervisor had made the request. The investigator left without reviewing the documentation. (21 CFR 812.1.)

3. A diagnostics company on the West Coast received a warning letter following an inspection. The FDA-483 issued at the close of the inspection listed only three items, all related to a recall the company had performed more than a year earlier. At the time of the inspection, the company had already implemented short-term corrective actions to address the problem leading to the recall and was able to demonstrate that the corrective action had been effective. The company also had a long-term corrective action in progress. Although the company called these facts to the attention of the investigator, the items were listed on the FDA-483 and a warning letter followed. The result of the warning letter was that review was discontinued for all the company's pending applications. (Compliance Policy Guide, 7382.830, attachment A-1.)

4. An IVD company on the East Coast was inspected by FDA in the spring of 1995. Although the IVD manufacturing guideline was in effect, the company was not asked at that time and had never been asked during previous inspections whether its water system had been validated. Some IVD companies in the same district had been asked about water system validation, while others had not. (Guideline for the Manufacture of In Vitro Diagnostic Products [Guideline], sec. 3.7.)

5. Two companies on the East Coast were inspected by the same FDA district during the same two-year period. One company was asked for its complaint files for the previous two years; the investigator identified complaints for which recalls were not conducted and urged the company to recall certain items, which were indicated on the FDA-483. The other company's complaint files also were reviewed and products were identified for which customer complaints were confirmed as lotwide problems. This investigator did not list the items on the FDA-483, nor did he suggest that the company recall the products. (Compliance Policy Guide, 7382.830, attachment A.)

6. A diagnostics company in the southern United States told the investigator that its products were "microbiologically uncontrolled." The investigator did not ask the company for supporting data to show that control of the products was not necessary to achieve acceptable performance. (Guideline, 3.1.1.)

7. An IVD company with a core facility for the manufacture of "microbiologically controlled" IVD products was audited by another IVD company from a different district within the same region. The auditing company expressed surprise at the sophistication of the auditee's facilities. The auditing company did not have similar facilities and had never been asked by FDA to identify appropriate environmental controls for the manufacture of similar products, while the auditee company had constructed the core facility in direct response to FDA inspectional observations. (Guideline, 3.4.)

8. Following the effective date of the IVD manufacturing guideline, a diagnostics company in the Midwest received an FDA-483 in which the citations included failure to specify a microbiological assurance level for each product, failure to generate data to support weekly environmental monitoring, failure to have statistically significant sampling plans for microbial load analysis, and failure to describe the practice of alternating disinfectants. The company received a warning letter follow-ing the inspection, despite promises of prompt corrective action. The company had a long history of acceptable inspection outcomes and had demonstrated its willingness and ability to implement corrective actions in a timely fashion. A competitor in another district that had recently been inspected by FDA did not have any of the specified controls in place. (Guideline, 3.1.2; 3.4; 3.6; 3.8.)

9. A diagnostics company in the eastern United States labeled IVD products "for laboratory use." An FDA investigator determined, based solely on the company's statements and without verifying the recipients of the company's products, that inspection under the GMP regulation was inappropriate. Other companies in the same area making similar products and selling them to the same facilities were routinely inspected under the GMP regulation. (21 CFR 809.10(d).)

10. Prior to implementation of the IVD manufacturing guideline, a diagnostics company outside the United States had an import ban placed on its products due to failure to adequately validate its water system. There was no indication that this failure had any effect on the performance of the company's products. (Inappropriate use of a draft guideline.)

11. A component supplier for IVD companies was inspected by FDA and received an FDA-483, even though company officials told the agency that they did not believe the company was required to comply with the GMP regulation because it did not manufacture finished products. (21 CFR 820.1.)

12. An IVD company outside the United States received a warning letter for GMP compliance deficiencies, even though the company's products were labeled for research use only. (21 CFR 809.10(d).)

13. A company on the West Coast was inspected in conjunction with a premarket approval (PMA) application. The investigator carried a copy of the Center for Drug Evaluation and Research guidance for WFI water systems. Despite repeated protests by the company that WFI was not required for its products and that the company's water met established specifications, the FDA-483 included observations regarding deficiencies in water system validation. Although there were no performance issues associated with the product, the warning letter that followed the inspection caused a delay in PMA approval. (Inappropriate use of CDER guidance.)

14. An IVD company on the East Coast was issued a warning letter that included an observation that the company lacked a statistical rationale to support the sample size for the serum panel being used in final product testing. The company hired an outside statistician to establish the appropriate sample size and rationale. The statistician was well qualified and had recently retired from FDA after serving as a division director. FDA rejected the consulting statistician's recommendation. After continued negotiations reached no satisfactory conclusion, the company dropped the requirement for use of a serum panel in final product testing, satisfying FDA and resulting in the close out of the warning letter. (21 CFR 820.160.)

15. An East Coast company was cited on an FDA-483 for not manufacturing its IVD products as sterile, even though those products were not labeled as sterile. The investigator later accepted the validation studies demonstrating bioburden control in accordance with the IVD manufacturing guideline. This involved expensive facility renovations to establish a high level of environmental control. The same level of adherence to the guideline had not been required in other districts. (Guideline, 3.1.1.)

Conclusion

The examples cited here demonstrate that inconsistency in FDA inspection and enforcement practices is real, not merely an industry perception. But knowing this is by no means the end of the process. The next steps must be to determine the causes that lead to such inconsistencies, and to develop corrective actions to prevent their recurrence. These will be the subject of another article in the next issue of IVD Technology.

Leif Olsen is vice president for regulatory affairs and quality assurance at Bio Whittaker, Inc. (Walkersville, MD), and a member of the IVD Technology editorial advisory board. Patricia Shrader is director of corporate regulatory affairs at Becton Dickinson & Co. (Sparks, MD).

NOTE: This is the first part of a two-part article. Part 2 is also available for on-line viewing.

A formulation can be defined as a liquid medium in which one or more active components (chemical or biological) remain in a stable environment and maintain specified potency limits for some period of time. For example, a vaccine formulation may still be effective and safe to use after several days of storage at 4°C. After about a week of storage, it must be discarded. However, if the kinetic clock for degradation of the vaccine could be slowed so that 1 second was extended to 1 hour, then the useful life of the formulation would be increased to almost 20 years.

Lyophilization stabilizes the formulation by slowing the kinetic clock of the degradation process. It alters the clock by removing the solvent component or components to levels that no longer support chemical reactions or biological growth. This removal is accomplished, first, by freezing the formulation, that is, separating the solutes from the solvent or solvents and immobilizing any solvent in the interstitial region between the solvent crystals. Then the solvent is removed by sublimation (primary drying) and next by desorption (secondary drying).

The nature of the formulation determines:

• Time and temperature needed to obtain a completely frozen matrix.

   

• Time, shelf temperature, and chamber pressure to conduct primary and secondary drying.

   

• The nature of the container system.

   

• Design and construction of the freeze-drying equipment.

Because time is equated to productivity, the formulation affects not only the process and drying equipment, but also the cost of manufacturing. As process time increases, profitability decreases.

Components of a Formulation

The formulation consists of three basic components--active ingredient, excipient, and solvent system. In general, the active ingredient in the pharmaceutical industry is defined by its potency and, in the diagnostic industry, by its reactivity. Depending on means of production, there may be variations in the composition of the active component from batch to batch.

The formulation of a product affects the time required to process it and, ultimately, the cost of the product. Here, products are readied for lyophilization at Medicus Technologies (West Chester, PA).

 

Excipients serve several functions.^1,2 They primarily provide a stable liquid environment for the active ingredient for some finite time. The excipient may also cryoprotect the active ingredient during the freezing process.^3­5 In the freezing of formulations containing biological organisms, the formation of ice within can lead to the organism's destruction by cell membrane rupture. Sucrose, glucose, and dextran are excipients used to cryoprotect organisms.^2,4

The excipient may also serve only as a bulking agent.² When solid concentrations of a formulation reach <2%, the resulting cake may have poor structural qualities and leave the container during the drying process. The addition of bulking agents such as mannitol and dextran strengthen cake structure. The role of the solvent system is often overlooked. Most formulations are totally aqueous solutions, although others contain solvents such as tertiary butyl alcohol to increase the solubility of some compounds. The solvent system is removed during drying, but its thermal properties have a major impact on the cosmetic properties of the final product (Figure 1).

Figure 1. Typical lyophilization cakes. (A) uniform distribution of constituents; (B) nonuniform distribution of constituents in the cake with a crust or glaze on the upper surface; (C) a cake with poor self-supporting structural properties; (D) a cake showing signs of collapse; (E) example of meltback; (F) disappearing cake, i.e., dissolution of the cake by excess water; and (G) puffing resulting from incomplete freezing of the matrix before evacuation of the dryer.

 

Freezing and Drying the Formulation

Freezing. Formation of ice during freezing results in dramatic changes in concentrations of the active ingredient and the excipient or excipients of the formulation. As an example, consider the freezing of an isotonic (0.9% w/v) NaCl aqueous solution. At a temperature of ­20°C, the frozen matrix consists of ice crystals interlaced with a 23% NaCl solution. At a temperature of <­23°C, the interstitial region in the matrix consists of a eutectic mixture of NaCl * 2H^₂O and ice crystals. (Eutectic temperature is a point on a phase diagram where the temperature of the system or the concentration of the solution at the point cannot be altered without changing the number of phases present.)

In most formulations (>99%; based on thermal analysis of >1000 formulations at Phase Technologies, Inc., Conshohocken, PA) containing an active ingredient and an excipient, freezing greatly increases the concentration of the active ingredient and the excipient or excipients, but does not produce a well defined eutectic mixture. Instead, freezing produces a complex, glassy system that may be homogeneous or heterogeneous. This complex system, at this time, can only be produced in the interstitial region of ice crystals as a result of the freezing process.

Another property of frozen matrix is the degree of crystallization (D), the ratio of the quantity of ice formed during the freezing process to the total freezable water in the formulation (Jennings TA, "Lyophilization Seminar" [Notes], Conshohocken, PA, Phase Technologies, Inc., p 26, 1994). As D approaches 1 (Figure 2), most water is in the form of ice crystals, and only a small quantity forms part of the interstitial region. The ice crystals interconnect to form vapor paths. With decreasing values of D (e.g., D = 0.5), the volume of glassy interstitial region approaches that of the ice crystals.

Figure 2. Illustration of a frozen matrix of a formulation in which the degree of crystallization approaches 1.

 

In the frozen state, the mobility of the water in the glassy interstitial region approaches 0, and the formulation is considered completely frozen. As the temperature of the matrix increases, the glassy interstitial region softens, the electrical resistivity of the interstitial region decreases, or the conductivity of the system increases. Such a change in the electrical nature of the matrix is associated with the onset of mobile water within its interstitial region. As temperature further increases, the interstitial region slowly takes on liquidlike characteristics, while surrounding ice crystals remain frozen.

In most glassy systems, onset temperature for the mobility of the water in the interstitial region is not as sharp and well defined as that for a eutectic mixture. The onset temperature for the mobility of water in the matrix interstitial region is referred to as collapse temperature. This definition differs from that of MacKenzie, who called collapse temperature "disappearance of the freezing pattern, more or less extensive flow of the residual material, and the generation of new patterns."⁶

Drying. For lyophilization to occur, the solvent is first removed by sublimation while the temperature of the frozen matrix is maintained below the eutectic or collapse temperature of the formulation. This is the primary drying process. The chamber pressure and product and shelf temperatures, during primary drying, are based on the formulation's eutectic or collapse temperature. A lyophilized cake is typified by Figures 1A and 1B. The resulting cake volume approaches the original fill-volume.

Primary drying at temperatures greater than that of the collapse or eutectic temperature of the formulation (sometimes referred to as vacuum drying or cryodrying) can lead to a product typified by Figures 1D and 1E, respectively. Figure 1D illustrates some collapse of the cake resulting from the presence of mobile water in the matrix interstitial region during primary drying. Figure 1E, though, illustrates meltback, a result of liquid states in the interstitial region.

After primary drying, the residual moisture on the resulting cake surface is reduced to levels that no longer support biological growth and chemical reaction. This process is secondary drying. The reduction of moisture in the cake during secondary drying is accomplished by increasing the shelf temperature and reducing the partial pressure of water vapor in the container. The required partial pressure of water vapor and shelf temperature are ascertained from stability studies of lyophilized or vacuum-dried products having varied amounts of residual moisture.

References

1. Bashir J, and Avis KE, "Evaluation of Excipients in Freeze-Dried Products for Injection," Bull Parent Drug Assoc, 27:68­83, 1973.

2. Wang YJ, and Kowal RR, "Review of Excipients and PH's for Parenteral Products Used in the United States," Bull Parent Drug Assoc, 34:452­462, 1980.

3. Greaves RIN, "Fundamental Aspects of Freeze-drying Bacterial and Living Cells," in Aspects Théoriques et Industriels de la Lyophilisation, Rey L (ed), Paris, Herman, pp 407­410, 1964.

4. Smith AU, "Some Problems in Supercooling and Freezing Living Cells, Tissues and Organisms," in Aspects Théoriques et Industriels de la Lyophilisation, Rey L (ed), Paris, Herman, pp 257­278, 1964.

5. MacKenzie AP, "Comparative Studies on Freeze-drying Survival of Various Bacteria, Gram Type, Suspending Medium and Freezing Rates," in International Symposium on Freeze-drying of Biological Products, Washington, DC, Develop. Biol. Standard, 36 S, Basel, Switzerland, Karger, pp 263­277, 1977.

6. MacKenzie AP, "Collapse during Freeze-drying--Qualitative and Quantitative Aspects" in Freeze Drying and Advanced Food Technology, Goldblith SA, Rey L, and Rothmayr WW (eds), New York, Academic Press, pp 277­306, 1975.

Thomas A. Jennings, PhD, is CEO, Phase Technologies, Inc. (Conshohocken, PA).

Continue on to Part 2 of this article.

Fluorescence in situ hybridization (FISH) is an emerging IVD technology ideal for the clinical evaluation and characterization of complex biological specimens--such as blood, amniotic fluid, and solid tumors--for genetic anomalies. Of special value to pathologists and geneticists, FISH facilitates the characterization of a broad range of molecular genetic events, such as aneuploidy, gene amplification, gene deletion, and chromosome translocations, that are difficult to detect with karyotype analysis, PCR, or LCR. FISH-based IVDs have the potential to significantly increase the survival of cancer patients by making possible earlier detection of malignancy and more accurate prognostic assessments following tumor surgery. FISH IVDs also can be applied to prenatal and postnatal genetic analysis. Moreover, the technology is especially useful for simultaneous detection of multiple genetic anomalies in an individual cell, potentially saving assay time and limiting specimen requirements.

Evolution

FISH uses nucleic acid probes--segments of labeled DNA designed to bind, or "hybridize," with the target DNA of a specimen--usually fixed to a glass slide. The probes are labeled with fluorescence molecules to make identification of the probe-target hybrid possible by use of a fluorescence microscope. The hybrid is further analyzed with computer imaging equipment. Since hybridization occurs between two complementary strands of DNA, labeled probes can be used to detect genetic abnormalities, providing valuable information about prenatal disorders, cancer, and other genetic diseases. Unlike other molecular DNA-based tests, which require cell lysis to free nucleic acids for analysis, FISH allows analysis of DNA in situ, that is, in its native, chromosomal form within the cell nucleus. This attribute permits the analysis of chromosomes and genes of individual cells.

A key advantage of FISH is its ability to target only those genetic sequences of interest. In 1986 Joe Gray, PhD, and Dan Pinkel, PhD, while at the Lawrence Livermore Laboratories, found that the inclusion of appropriate blocking DNA sequences in FISH using cloned human genes or unique sequence DNA regions suppressed hybridization of the interspersed repetitive DNA sequences (so-called "junk DNA") scattered throughout the genome. This suppression largely eliminated nonspecific signals. The discovery was patented by the University of California and is licensed exclusively to Vysis, Inc. (Downers Grove, IL).

Development of direct-labeled DNA probes by covalent attachment of the fluorophores to the probe has further simplified the technology: older detection methodologies rely on labeling the probe-target DNA hybrid indirectly by binding a signal-generating moiety (e.g., fluorescein isothiocyanate [FITC]­avidin) to a non-signal-generating ligand on the DNA probe (e.g., biotin) (Figure 1). The indirect methods are more complex, require additional steps, and tend to have more nonspecific signals due to the inherent stickiness of the FITC-avidin complex. For example, in a recent study to measure the amplification of HER-2/neu in breast cancer specimens, a comparison of direct-labeled and indirect-labeled probes found a methodological rate of 99.3% with the direct-labeled probe compared with a rate of only 79.7% with the indirect-labeled probe.¹ Thus, with the combination of the Pinkel and Gray invention and direct-labeled probes, FISH has the potential to become a routine technology for clinical identification of chromosome anomalies.

Glossary

alpha satellite DNA: Tandem arrays of different copies of an approximately 171-base-pair sequence found at the centromeric region of each human chromosome.

aneuploidy: Deviation from a normal chromosome complement, either fewer (monosomy) or greater (for example, trisomy [three]) than normal.

spot counting: Counting of the distinctive fluorescent signals, or "spots," produced by FISH probes specific to the alpha satellite region of the chromosome when hybridized.

Procedure

FISH, using direct-labeled probes, is a simple procedure consisting of six basic steps (Figure 2). The first step, which involves sample pretreatment, ends with the denaturation of target DNA sequences. Next, the probe is applied to the target area, and the target DNA and probe are hybridized. Following hybridization, the target is washed to remove nonspecifically bound probe. Finally, a counterstain is applied and the results of the hybridization are visualized.

The sample must be prepared in such a way that the probe has access to the target DNA while specimen morphology is preserved. Tissue type, fixation, and sample embedding processes influence the type of pretreatment necessary before denaturation. For lymphocytes, simple denaturation conditions, such as high temperature and low salt or alkali, are adequate. However, formalin-fixed paraffin-embedded tissues require more steps, including deparaffinization and enzymatic and/or chemical treatment to allow the probe access to the target DNA. Hybridization and washing parameters are fairly standard but in some assays must be varied to accommodate the complexity and specificity of some probes.

(A) Color-composite image of an interphase nucleus hybridized with combinations of repetitive-sequence DNA probes labeled with three different fluorophores. The seven chromosomes identified are X (red), Y (aqua), 18 (green), 17 (blue), 12 (violet), 8 (yellow), and 7 (pink).

 

(B) Metaphase chromosomes stained using three fluorophore labels and combination coding to identify seven pairs of chromosomes.

 

(C) Normal metaphase hybridized by comparative genomic hybridization.

 

(D) Sperm cells hybridized with SpectrumOrange CEP 12 and SpectrumGreen CEP 8.

 

For direct-labeled probes, the results are detected by viewing the samples under a fluorescence microscope with appropriate filters. Indirect detection demands additional labeling steps, which typically require streptavidin or antibody-enzyme conjugates or fluorophore-labeled counterparts, and additional washing steps once the probe is bound to the target. Furthermore, indirect methodologies limit the user's ability to simultaneously score several genetic events in a sample.

A fluorescence microscope that is equipped with a 100-W mercury-arc lamp and oil-immersion fluorescence objectives with numerical apertures > 0.75 is required to view the results of a FISH assay. The probes and counterstain can be visualized separately or simultaneously depending on the filter set used. A single-band-pass filter allows one fluorophore to be viewed; a multi-band-pass filter allows viewing of several different fluorophores.

Many photographic films are available to capture FISH images. Although not impossible, it is usually very difficult to balance the bright probe signals against a dark background to produce acceptable photographic results. A better option is to capture images digitally using a microscope equipped with a digital or video camera. Digital imaging allows not only production of a printed image but also an analysis and enhancement of the image that is not possible with a standard photograph.

Performing FISH assays is not difficult for any technician who is familiar with standard laboratory procedures. However, interpreting FISH assays, especially those used to enumerate chromosomes in interphase nuclei, requires training. The trainee might observe the procedure once, perform it with supervision until competent and comfortable, and finally perform it alone. For interpretation of results, training might include a review of FISH analysis guidelines (developed either in the laboratory or by the assay manufacturer), observation and analysis of results from assays of abnormal and normal specimens (initially under supervision), and finally blinded evaluation of the same slides.

Interpretation becomes more precise with practice. Each laboratory should establish criteria for interpreting the results of a FISH assay based on the types and methods of preparation for the specimens they routinely handle.

By following CLIA and establishing product-specific performance characteristics for FISH probes and reagents, a laboratory can minimize the effects of specimen preparation and assay conditions on the final results. A laboratory can also test its ability to detect a small number of aberrant cells in a specimen that contains mostly normal cells by using appropriate positive and negative controls, creating internal guidelines to standardize the evaluation process, and empirically establishing baselines.

Figure 1. Comparison of direct-labeled and indirect-labeled probe procedures. At the point marked by an asterisk, the fluorophore is covalently linked to a direct-labeled probe; on an indirect-labeled probe, a ligand (e.g., biotin) is linked to the probe and a reagent that carries a detectable label (e.g., FITC-avidin) must be added to the reaction once the probe is bound to detect the probe-target hybrid.

Multievent Capabilities on Individual Cells

Use of multicolored probes allows for simultaneous detection by FISH of multiple genetic events within a single nucleus. A greater number of fluorophores can be used together in direct-labeled probes than in indirect-labeled probes.² Multievent screening can identify the clonal nature of a specimen as well as the chromosomal rearrangements in metaphase and interphase cells. Additionally, simultaneous target detection can increase assay accuracy by permitting the inclusion of probes that hybridize to control loci (versus test loci). With these probes, it is possible to monitor for hybridization success or failure within each nucleus.

There are several approaches to multievent detection in FISH.² One is to use a probe labeled with a different spectrally distinct fluorophore for each target detected in the hybridization. At least seven different chromosomes have been detected simultaneously by use of this method.² Another approach, called color coding, uses different fluorophores singly and in combination to identify more chromosome targets than the number of fluorophores used to label the probes.³ By combining fluorophores, 2^N­1 targets can be detected (N = number of fluorophore labels).

Recently, all 24 chromosomes have been detected simultaneously in metaphase chromosomes by use of only five fluorophore labels.^4,5 An example using three fluorophore labels to determine the aneuploidy of seven different chromosomes in interphase nuclei is shown in Figure A on page 27. When each combination consists of different proportions of each fluorophore label, the number of targets detectable is limited only by the ability to distinguish the different label proportions. For example, eight chromosomes have been reproducibly detected using only two fluorophore labels.²

The scope of FISH probes for detection of genetic defects continues to increase. For example, probes consisting of tandemly repeated human DNA sequences, such as centromeric alpha satellite sequences, are often used to identify or enumerate specific chromosomes. Cloned, unique-sequence, DNA-region probes may be used to detect small regions, or loci, of the genome. Mixtures of such clones may be used to stain, or "paint," large sections of the entire chromosome. Such staining allows analysis of the chromosome number as well as identification of additions and translocations in metaphase cells. The information provided by these probes makes them valuable in assessing the genome for a range of applications, including prenatal/perinatal diagnosis of genetic anomalies, studies of cancer, diagnosis of myeloid disorders, and rare cell analysis.

Prenatal/Perinatal Applications

Chromosomal aneuploidies, such as Down's syndrome, are by far the most common genetic abnormalities associated with birth defects in newborns. Traditional cytogenetics on banded metaphase chromosomes has been the standard prenatal test offered to women at increased risk of having fetuses with chromosomal abnormalities. Cytogenetics can examine all chromosomes for both aneuploidies and structural abnormalities; however, this technique requires a large specimen volume, culture of the fetal cells for several days, isolation of metaphase spreads, and a highly trained technician to analyze the results. Moreover, the method usually takes more than a week to complete. In contrast, FISH performed with a set of chromosome-specific probes covering only the most common aneuploidies (i.e., abnormal number of sex chromosomes or trisomy for chromosomes 13, 18, and 21) can detect these major chromosomal defects with high sensitivity and specificity from uncultured amniocytes in less than 24 hours.⁶ Thus, in certain clinical situations where rapid detection of the most common aneuploidies is desired, FISH on interphase nuclei provides an initial rapid screen preceding the full cytogenetic evaluation. Results from a FISH assay could be used to prioritize cases in a busy cytogenetics lab by moving those specimens that are positive by FISH to the top of the list for analysis by traditional cytogenetic methods.

Figure 2. The six steps for performing FISH using direct-labeled DNA probes.Standard Cytogenetic Analysis

In the perinatal setting, FISH is used to detect both structural and numerical chromosomal abnormalities. It can play a valuable role in perinatal testing to detect small chromosome deletions often missed by conventional cytogenetics. Examples of such microdeletion syndrome tests include those that detect deletions on chromosome 7 that are associated with Williams syndrome and on chromosome 22 that are associated with DiGeorge and velocardiofacial syndromes.

Cancer Applications

The ability of FISH to rapidly test interphase and metaphase chromosome defects makes it especially useful in the study of cancer. In solid tumors, conventional cytogenetics is rarely used because obtaining metaphases is difficult and those cells that do proceed to mitosis may not be representative of the tumor. Other molecular techniques, such as PCR and Southern, Northern, and Western analysis, require extraction of the tissue. Extraction procedures net both normal and abnormal cells, so sensitivity is lower and quantitation less reliable than with FISH probes.

FISH allows cell-by-cell analysis and thus provides for a more sensitive and reliable assessment of chromosomal aneuploidy, gene amplifications and deletions, and chromosome translocations. A reliable determination of whether a gene is amplified in a specimen is often possible with evaluation of only 20 to 50 cells.

HER-2/neu gene amplification is a significant, independent prognostic indicator for breast cancer recurrence and survival. Amplification and overexpression of HER-2/neu have been correlated with a poor prognosis--a shorter disease-free period following treatment and a shorter overall survival. Studies have also shown that HER-2/neu overexpression is useful as a marker of tumor resistance to chemotherapy and to hormone therapy. Thus, for the management of breast cancers, HER-2/neu amplification has the potential to predict response to treatment and to determine the choice of therapy.

While Southern blot analysis for HER-2/neu gene amplification and immunohistochemistry (IHC) for HER-2/neu protein expression are often used to quantitate gene amplification, each presents technical and interpretative limitations. The results of IHC vary with use of different antibodies and tissue treatment, use of different criteria for positivity, and use of different procedures. Press et al. reported the performance characteristics of several HER-2/neu antibodies to have a sensitivity ranging from 6 to 80%.⁷ Because of these inherent technical difficulties, the IHC and Southern blot analysis have not been standardized to yield consistent results.

FISH is an alternative technique without the technical and interpretative limitations of Southern blot and IHC. It reliably and accurately indicates gene amplification. For quantitation of HER-2/neu gene amplification, the distinct advantage of FISH is that it assesses the level of amplification directly in the tumor cells while the specimen retains its characteristic morphology.

A group of researchers recently analyzed a cohort of breast cancer specimens by various methods and found FISH to be more sensitive than Southern blot analysis and more accurate than Northern and Western blot analysis and to have a greater applicability and sensitivity in paraffin-embedded tissue than IHC. Against the best of the methodologies (IHC on frozen tissue sections), the sensitivity of FISH on paraffin sections was 97.2%, while its specificity was 100%.¹

Myeloid Disorders

Myeloid disorders--including chronic myelogenous leu- kemia (CML), acute myeloid leukemia (AML), myelopro-liferative disorder (MPD), and myelodysplastic syndrome (MDS)--have in common the chromosomal abnormality trisomy 8. Thus, the presence of this trisomy is of prognostic value to the physician. FISH on interphase and metaphase cells analyzed with the Vysis CEP 8 probe (a direct-labeled chromosome-enumerator probe specific to the alpha satellite DNA contained within the centromere region of chromosome 8) was compared to standard cytogenetic analysis in a multicenter, blinded, controlled study. Four laboratories provided a total of 364 archived bone marrow specimens for assay. The results of this study, shown in Table I, exemplify the reliability of FISH technology.

In the management of many hematologic malignancies, bone marrow transplantation (BMT) is a critical therapeutic strategy for successful cure. Following the transplantation, the ability to detect the presence of clonal neoplasms and to assess engraftment by cytogenetics is important. Although in many patients a stable chimeric state evolves between donor and recipient bone marrow, in others an unstable one forms and the malignancy eventually recurs, with an increasing number of host cells appearing in the bone marrow or peripheral circulation.

Approximately one-half of all heterologous BMTs are mismatched with respect to sex. Rapid and accurate identification of the genetic sex of the bone marrow cells offers a simple method for evaluating the donor/recipient status of the bone marrow. While several methods are currently used for this evaluation, enumerating chromosomes X and Y via FISH probes can provide the most rapid and easily quantifiable result.

Using the Vysis direct-label CEP X SpectrumOrange/Y SpectrumGreen DNA FISH probe assay, a pilot study was conducted to evaluate bone marrow specimens in 139 patients who had undergone an opposite-sex BMT.⁸ The clinical sensitivity for interphase FISH analysis was estimated to be 100%. The estimate for conventional cytogenetics analysis was 77%, and analysis was impeded by hypocellularity in some cases. In 148 patients who had undergone a same-sex BMT, the clinical specificity for interphase FISH analysis was estimated to be 100%. The assay was shown to be highly reproducible, even across laboratories.

Table I. Results of pilot study comparing Vysis CEP 8 (chromosome enumerator probe) FISH interphase analysis with standard cytogenetic analysis for detection of trisomy 8.

Rare Cell Analysis

FISH has been used extensively in attempts to identify fetal cells in the maternal bloodstream for prenatal diagnosis. Cells from amniocentesis or chorionic villus sampling (CVS), traditional sources of fetal material for genetic analysis, are obtained at the expense of a small but significant risk to the fetus. If fetal cells could be obtained from the maternal bloodstream instead, prenatal genetic testing could be noninvasive to the fetus. Such testing, being safer than amniocentesis or CVS, might be used in a greater number of pregnancies.

By screening maternal blood samples using FISH or PCR with probes to the Y chromosome as a marker of male fetal origin, researchers report values ranging from one fetal cell in 10⁵ to one in 10⁹ nucleated maternal cells.^9,10 Others have combined IHC detection of fetal hemoglobin with FISH in a technique known as FICTION (Fluorescence Immunophenotyping and Interphase Cytogenetics as a Tool for the Investigation of Neoplasms) as a means of identifying female fetal cells.¹¹

Some reports also cite the use of FISH in detection of minimal residual disease in leukemia.^12,13 Unlike reverse-transcription PCR (RT-PCR), which requires complex extraction of intact RNA, FISH provides a cell-by-cell analysis of materials simply fixed to a microscope slide. FISH is thus an easy method of quantifying the level of disease present in the specimen.

Practical Considerations

As the diagnostic industry continues to search for creative ways to contain costs, FISH is likely to play an increasing role. FISH offers the ability to obtain, in most cases, definitive results quickly by employing a simple detection strategy that can be interpreted by basic technical staff with proper training.

Automation of FISH analyses will increase assay throughput, reduce technician time, remove operator bias, and simplify the interpretation of complex results. Automated preparation and staining of slides is possible. Image processing methods and instrumentation for the automatic enumeration of chromosomes in interphase nuclei have already been demonstrated. When commercially available, automated "spot counting" will relieve technical personnel of hours of tedious microscope viewing and eliminate errors due to fatigue and misinterpretation or inconsistent application of counting criteria. Karyotyping systems similar to those designed for semiautomated analysis of conventionally stained chromosomes allow the identification of fluorescent-banded chromosomes. These capabilities, once integrated into a complete system, will allow for a turnkey FISH analysis system.

Conclusion

The utility of FISH in clinical research laboratories is growing. In a recent CAP Today article, Arthur Brothman, PhD, director of the Cytogenetics Laboratory at the University of Utah Health Sciences Center, states, "If FISH is not offered in some cytogenetic evaluations, patients are not being given what I believe to be the standard of care."¹⁴ Several driving factors are enabling FISH testing to move into mainstream practice in pathology laboratories. These include improvements in methods of cloning unique sequence probes and labeling nucleic acids, simplification of assay procedures to allow for automatic slide staining, high information density through the use of multicolor event detection, sophisticated image analysis tools, and increased knowledge of the clinical importance of genotyping of solid tumors for diagnosis, prognosis, and monitoring.

Acknowledgment

The authors gratefully acknowledge the assistance of Ping H. Hsu, PhD; Walter King, PhD; Dave Lane, PhD; Larry Morrison, PhD; Uwe Muller, PhD; Peter Osella; John Proffitt, PhD; Chris Shasserre; and Doug Taron, PhD, of Vysis, Inc.

References

1. Pauletti G, Godolphin W, Press MF, et al., "Detection and Quantitation of HER-2/neu Gene Amplification in Human Breast Cancer Archival Material Using Fluorescence In Situ Hybridization," Oncogene, 13:63­72, 1996.

2. Morrison LE, and Legator MS, "Multi-color Fluorescence In Situ Hybridization Techniques," in An Introduction to Fluorescence In Situ Hybridization, Pinkel D, and Andreev M (eds), New York, John Wiley, in press.

3. Fox JL, Hsu P-H, Legator MS, et al., "Fluorescence In Situ Hybridization: Powerful Molecular Tool for Cancer Prognosis," Clin Chem, 41:1554­1559, 1995.

4. Speicher MR, Ballard SG, and Ward DC, "Karyotyping Human Chromosomes by Combinatorial Multi-fluor FISH," Nature Genet, 12:368­375, 1996.

5. Schröck E, du Manior S, Veldman T, et al., "Multicolor Spectral Karyotyping of Human Chromosomes," Science, 273:494­497, 1996.

6. Ward B, Gersen SL, Carelli MP, et al., "Rapid Prenatal Diagnosis of Chromosomal Aneuploidies by Fluorescence In Situ Hybridization: Clinical Experience with 4,500 Specimens," Am J Hum Genet, 52:854­865, 1993.

7. Press MF, Hung G, Godolphin W, et al., "Sensitivity of HER-2/neu Antibodies in Archival Tissue Samples: Potential Source of Error in Immunohistochemical Studies of Oncogene Expression," Cancer Res, 54:2771­2777, 1994.

8. Dewald GW, Schad CR, Christensen ER, et al., "Fluorescence In Situ Hybridization with X and Y Chromosome Probes for Cytogenetic Studies on Bone Marrow Cells after Opposite Sex Transplantation," Bone Marrow Transplantation, 12:149­154, 1993.

9. Hamada H, Arinami T, Kubo T, et al., "Fetal Nucleated Cells in Maternal Peripheral Blood: Frequency and Relationship to Gestational Age," Hum Genet, 91:427­432, 1993.

10. Reading JP, Huffman JL, Wu JC, et al., "Nucleated Erythrocytes in Maternal Blood: Quantity and Quality of Fetal Cells in Enriched Populations," Hum Reprod, 10:2510­2515, 1995.

11. Soenen V, Fenaux P, Flactif M, et al., "Combined Immunophenotyping and In Situ Hybridization (FICTION): A Rapid Method to Study Cell Lineage Involvement in Myelodysplastic Syndromes," Brit J Haematol, 90:701­706, 1995.

12. White DM, Crolla JA, and Ross FM, "Detection of Minimal Residual Disease in Childhood Acute Lymphoblastic Leukaemia Using Fluorescence In-Situ Hybridization," Brit J Haematol, 91: 1019­1024, 1995.

13. Zhao L, Chang K-S, Estey EH, et al., "Detection of Residual Leukemia Cells in Patients with Acute Promyelocytic Leukemia by the Fluorescence In Situ Hybridization Method: Potential for Predicting Relapse," Blood, 85:495­499, 1995.

14. Check WA, "Cytogenetic Labs Embrace FISH," CAP Today, 10(3):1­18, 1996.

Steven Seelig, MD, PhD, is chief medical officer and vice president for research and development, and Susan E. Tibedo is a technical writer at Vysis, Inc. (Downers Grove, IL).

Ever since the Clinical Laboratory Improvement Amendments of 1988 (CLIA) were implemented, the Centers for Disease Control and Prevention (CDC) have been wrestling with the issue of quality control regulation for in vitro diagnostic products (IVDs).

CDC has been obtaining input from many groups as part of its ongoing program to address this subject. The effort was expanded last September, when CDC sponsored a workshop to gather information from industry and user representatives regarding current quality control systems and practices. Specifically, CDC requested comments on the following concerns:

• The ability of current test systems to assess the potential for error in the total testing process.

   

• The types of processes necessary to monitor the quality of clinical laboratory results over time.

   

• The types of processes necessary to assess appropriate test performance by operators.

In response to this call for comments, the Association of Medical Diagnostics Manufacturers prepared a presentation on the first and third of the topics under discussion at the workshop. Founded in 1973, AMDM is a trade association made up of in vitro diagnostic manufacturers representing a wide range of clinical laboratory products, technologies, methodologies, and analytes. Its presentation at the workshop was based on a recognition of the complexity of the IVDs and QC procedures represented among its membership.

The current requirements for QC testing are spelled out in several sets of regulations and standards. Unfortunately, these documents sometimes appear vague or contradictory. For instance, the CLIA regulation requires that laboratories "perform and document control procedures using at least two levels of control materials each day of testing" (42 CFR 493.1202( c)(4)). The Health Care Financing Administration's section on "Survey Procedures and Interpretive Guidelines for Laboratories and Laboratory Services" states that "instrument or procedural control checks (at least two levels) may be used to meet this requirement" (State Operations Manual, app. C, p. 108). The College of American Pathologists Commission on Laboratory Accreditation's inspection checklist for diagnostic immunology and syphilis serology says that "for certain reagent systems with built-in controls, external controls may not be required. Such systems must have been classified as moderate complexity or waived tests under CLIA '88 and the internal control must measure reactivity, not process." These somewhat contradictory statements have created confusion and uncertainty for diagnostic manufacturers and field inspectors alike.

One of the reasons such standards and regulations can be confusing is that the world of IVDs encompasses a tremendous range of complexity. This fact is well known among those in industry, and has been highlighted by the struggle to categorize IVD tests according to their level of complexity, as required by CLIA.

Current IVD products range from manual tests containing several operator-dependent steps, to manual tests unitized with minimal operator-dependent steps, to automated tests containing several operator-dependent steps, and finally to fully automated systems requiring little operator intervention. To make matters more complicated, IVDs tend to evolve over the course of their effective lifetimes, beginning as manual tests and becoming increasingly automated as they undergo the constant process of product improvement. Thus, IVDs are not only complex in their initial format but are also dynamic in their individual complexity over time.

Overlying these wide differences in the complexity of their technologies, IVDs also display a wide range in the quality control procedures developed by their manufacturers. IVD quality control systems can range from none at all, to monitors of all internal systems, to systems that monitor everything short of adding the analyte to a sample cup. Manufacturers of manual assays with several operator-dependent steps sometimes provide or recommend external quality controls to monitor all reagents and steps. Manufacturers of manual assays with minimal operator-dependent steps may incorporate into an assay controls that monitor its vital reagents but are not needed for operator monitoring. Manufacturers of automated assays with several operator-dependent steps may recommend some type of quality control procedure to monitor the operator steps, or may include separate built-in controls to monitor instrument function. And finally, manufacturers of automated assays that require little operator intervention may build in controls that monitor instrument and reagent function and that, once again, don't require operator monitoring.

The results of all these varieties of quality control procedures and technologies may also provide a wide range of data. Some results may be strictly qualitative, while others are either semi- or fully quantitative. The implications of these variations on the need for operator interpretation are obvious.

Although CDC's effort to develop a single, clear-cut requirement for IVD quality control is a worthy one, the wide variations in current technologies, monitoring systems, and results are likely to make such an effort extremely frustrating. Given this complexity, the overriding question posed by CDC's September workshop--what type of quality control system should be recommended for all IVDs?--is difficult to answer in a single recommendation. The answer, it seems, is "it depends." A number of factors are involved, including the following:

• What needs to be monitored based on reagent composition.

   

• How sensitive the test's reactions are to environmental changes.

   

• How many and what type of operator-dependent steps are used.

   

• What shelf-life studies have been performed by the manufacturer in support of product performance over time.

For this reason, AMDM believes that no single system can be established for all IVDs. At the CDC workshop, the association presented an alternative proposal that would take into account some of the many variables that characterize the world of IVDs. AMDM recommended that a task team representing FDA, industry, and user groups be appointed to establish technology groupings of IVDs and to recommend general QC procedures for each. Once the groupings had been established, the team would meet periodically to review changes in technology and QC systems, and to adapt the groupings and recommended QC procedures as appropriate.

This system would provide manufacturers with specified levels of quality control toward which they could direct their product development efforts. By meeting the QC requirements specified for a particular grouping, manufacturers would be able to reduce their burden of proof at the end of the product development process. Moreover, the task team would be able to identify quality control issues relevant to particular groupings so that manufacturers could develop QC systems specifically to address them.

If the program were implemented appropriately, new technology groupings could be added within the same regulation, thus providing flexibility without creating the need to revise the regulation. Until the QC groupings were established, users would follow the manufacturers' QC recommendations as described in each product's package insert.

Reaction to the AMDM recommendation was mixed. Many participants supported the association's approach, agreeing that the tremendous diversity of IVD products precludes the use of a one-size-fits-all quality control method. They also commented that the slow progress of change in regulations should be taken into account, and that whatever regulation is developed now must be flexible enough to keep up with rapidly evolving technologies in the future.

Others at the CDC workshop disagreed, arguing that products already undergoing FDA 510(k) review and CDC complexity categorization should not have to undergo a third product-specific level of review. This criticism permitted the association to make it clear that its proposal was not intended to create another product-specific system, but to address broad technology categories of IVDs.

In the opening remarks of its September workshop, CDC presented the key elements of the CLIA regulation as a three-legged stool--quality control, personnel, and proficiency testing. If all three of these legs are in balance, then quality assurance is achieved. CDC's difficult task is to revise QC requirements while maintaining that balance. To do so, it plans to review all the comments it has received from various forums and to develop a proposal by early this year. IVD manufacturers and users look forward to reviewing the proposal and providing additional input.

Judith J. Smith is director of regulatory affairs and quality management at Sienna Biotech, Inc. (Columbia, MD), and a member of the Association of Medical Diagnostics Manufacturers (Washington, DC).

Studies on new treatment approaches for preterm labor are paving the way for use of diagnostic tests that predict prematurity.

That's bright news for two Northern California biomedical companies: Sunnyvale-based Adeza Biomedical Corp. and Dublin-based Biex, Inc. Thanks to therapeutic advances, Adeza's upcoming point-of-care kit--based on its fetal fibronectin enzyme assay--is generating considerable buzz in the obstetric community, as is Biex's estriol assay.

Predicting preterm labor has been a medical challenge throughout the ages. Causes are multifactorial and, despite intense research, they remain somewhat mysterious. Both IVD tests demonstrate better sensitivity and specificity in prediction than does reliance on a risk factor, although physical signs--contractions, cervical dilatation, vaginal bleeding--have until now been the gold standard.

By measuring saliva estriol levels, the Biex Salest test promises improved prediction for preterm labor.

 

The only FDA-approved treatment, ritodrine, often proves intolerable to patients. Racing heartbeats in both mother and baby are a common side effect of the drug. These side effects frequently lead to maternal noncompliance, says T. Murphy Goodwin, MD, an obstetrician at the University of Southern California School of Medicine. He and his colleagues demonstrated that an experimental therapy--an antagonist to oxytocin known as antocin--could slow uterine contractions with the same effectiveness as ritodrine, but without producing the high degree of side effects (Obstet Gynecol, September, 88:33­ 36, 1996). Magnesium sulfate is also coming into more use, according to the medical literature.

But it isn't just new approaches that have Goodwin and his coauthors feeling upbeat. In his quest for a better way to halt preterm labor's progress, Goodwin has also found that the two new diagnostic tests shed some light on cause and effect.

For instance, a positive result from the Adeza assay seems to be associated with subclinical infection, he says. Biex's product, on the other hand, seems to identify what Goodwin calls the fallout from "a precocious fetus"--a cascade of events that seem to start with early biochemical changes originating with the developing baby.

"One of the past problems is that preterm labor was treated as if it was just one thing," he says. Ritodrine, which slows smooth-muscle contractions, doesn't target the cause but treats the result, a contracting uterus.

But the two new diagnostics, while both aimed at detecting early labor onset, seem to point to separate causes, at least to some extent, Goodwin observes. Biex's test, which measures estriol levels in saliva, seems linked to changes of fetal origin, which cause elevation of levels of this weak estrogen. The Adeza test, on the other hand, may be positive more often in cases of microbial infection, Goodwin says.

There are big advantages to having such information. Antocin seems to work most beneficially for preterm labor associated with biochemical alterations; antibiotics are the appropriate treatment for infection-related uterine activity. The two tests may soon be used together to get more clinically useful information.

Adeza currently has a 10-minute, one-step, point-of-care product in a premarket approval (PMA) clinical trial. Meanwhile, PMA-related clinical trials are under way for Biex's estriol-collection test, which simply requires women to "drool into a tube," explains Fred Voss, PhD, Biex's vice president for R&D. The saliva sample is sent to a lab for analysis, usually with an 18- to 24-hour turnaround. The company expects to make its PMA application early this year. Voss says the Biex assay will give physicians a way of detecting the changes directly linked to "the uterus's preparation for contraction and labor." Still unanswered, he acknowledges, is why early-onset labor can't be fully explained by fetal changes or microbial invasion. "For all of our understanding of some of the causes, we still don't know the full story," Voss concedes. "In some women, stress seems to play a role," he says, adding that this is another piece of the puzzle and "something harder to pin down."--A.S.

Manufacturers of HIV-1 tests and federal scientists are working earnestly to expand the sensitivity of test kits to include group O HIV-1. Currently marketed assays to detect infection with HIV-1 are optimized to detect group M viruses, and their track record has been less than stellar when it comes to detecting viruses of group O, which is associated with a relatively small population in central Africa (Lancet, 344:1333­ 1334, 1994).

The antibody response elicited by group O strains can slip through the screening process of current commercial enzyme immunoassays because in them this form of the virus is "weakly reactive," notes Patrick Sullivan, DVM, PhD, an epidemiologist who has been working on surveillance of "O" infections at the Centers for Disease Control and Prevention in Atlanta. This situation was thought to be "exceedingly rare," however, because fewer than 100 cases of group O HIV infection had been reported worldwide, and none had been recognized in the United States.

But then group O was found in a patient whose test results had been negative for HIV. With the help of CDC researchers, the Los Angeles County Department of Health identified a group O strain in a woman with symptoms of AIDS but a history of negative HIV testing (MMWR, 45:561­565, 1996). Since then, another case has been found in another area of the country. "Now we are in a mode of systematic surveillance; however, the conclusion--that this is very rare--is still the same," stresses Sullivan.

But not so rare that in some regions of the country--Southern California, for example--the health departments aren't sending out some warnings. The Los Angeles County Department of Health has mailed letters to primary-care physicians across the region urging them to consider double-checking patients who have symptoms suggestive of HIV infection but negative test results, and naming group O as a possible culprit.

This increased concern has fueled stepped-up efforts to refine HIV test kits, and manufacturers have had a willing partner in the process: the federal government. "CDC and FDA are collecting sera from persons infected with group O HIV for use in evaluating reconfigured test kits," affirms Sullivan. The substance is so scarce that CDC and the agency have collaborated closely to collect the material, and are working together with business to expedite evaluation of the reconfigured kits. FDA has given the effort top priority.

The news of group O has cast a shadow over the AIDS research community at a time when cautious optimism seemed justified. Last summer the international AIDS conference in Vancouver unveiled findings showing that a powerful cocktail of antiviral drugs could prolong life, even largely eradicate the virus. It seemed that the elusive virus had met its match in emerging biotechnology, and that demand for HIV assays would increase.

The widely read Annals of Internal Medicine followed up with a study that demonstrated the efficacy of rapid HIV testing. Accompanying the article was a glowing editorial, which stated that the assay may have the potential for improving clinical outcomes, thanks in part to early detection (Ann Intern Med, 125: 509, 1996).

Now some in the field are privately wondering if health-care providers or companies could be held liable for undiagnosed cases. Could infected individuals claim that an IVD test's failure to detect the group O virus meant delayed access to drugs that could have staunched the infection if caught earlier? Ordinarily, says Michael D. Roth, a Los Angeles attorney who formerly chaired the Medicine and Law Committee of the American Bar Association, a case in which a lab test result is a "negative when it should be a positive should generally be an easy one for a plaintiff to win in court."

But, in the case of group O, there is a question of whether there is any liability. The courts haven't clearly defined where standard of care begins or ends in areas of evolving technology. A case can rest largely on whether changing a test truly constitutes an innovation or is merely a correction of an original oversight, Roth says. If a judge or jury concludes the latter, damages can be sought. And some in public health are openly concerned about the impact of the current efforts to improve HIV test kits.

"CDC has a way to identify group O that most commercial labs don't have," notes Paul Simon, MD, the Los Angeles County epidemiologist who helped uncover the index case of group O. Combining the methods presents a formidable technical challenge, Simon points out. Current testing picks up the overwhelming majority of HIV. Whatever the new test is, "you just don't want it to be something that compromises the ability to detect these more common strains," he warns.

On a misty Southern California morning recently, I was called to attend a press conference cosponsored by American Red Cross Biomedical Services and the San Diego­based biotechnology firm, Gen-Probe, Inc. Freshly named to take over the helm of IVD Technology (more about that later), I was naturally eager to make a good impression. But as I entered the conference room, my rain-slick shoes encountered an equally slick patch of marble, and I found myself suddenly sprawled on the floor.

Recovering my footing and composure, I turned to the business of the press conference, which was to announce that the National Heart, Lung, and Blood Institute (NHLBI) had awarded Gen-Probe a three-year, $7.7-million contract to develop a rapid test for HIV and hepatitis C (HCV), to be used in screening the nation's blood supply. The company's winning proposal makes use of its patented transcription mediated amplification (TMA) technology, which was first approved for use in the United States in 1995 as part of the company's tuberculosis diagnostic.

The unique characteristic of TMA is that it amplifies ribosomal RNA instead of the DNA amplified in polymerase chain reaction. Gen-Probe officials expect that the new test will reduce the time required to detect HIV from 16 to 11 days. The fully automated test will be priced "somewhere between $6 and $10," adding slightly to the current cost of testing a unit of blood, but efficiencies will be gained by combining the tests for HIV and HCV. Company officials expect to obtain FDA approval for the test before the end of 1999.

There's no question about the importance of this award to Gen-Probe. According to Gen-Probe president and CEO Hank Nordhoff, the NHLBI contract will increase the company's annual R&D budget by about 14%. Equally striking, however, is the speed with which the federal government has moved to support development of a diagnostic based on the relatively new TMA technology. With that kind of support, Nordhoff likes the chances of TMA's success. "We think we have a very strong technology," he says. The company hopes that TMA will become the amplification method of choice for many applications.

No doubt, the competition among companies and technologies that characterizes the field of IVDs is also a marker of the industry's vitality. As the new editor of IVD Technology, I hope that I will be able to capture some of that liveliness and bring it to our pages.

As my debut, this issue of IVD Technology is certainly more graceful than my entry to the Gen-Probe press conference. It has benefited from the influence of my predecessor, Jenevieve Blair Polin, who ably guided the publication through its launch and first year as an independent journal, and will remain a contributor to the publication in her new role as editor at large. But the pratfalls--editorial and otherwise--are clearly my own.