CLINICAL TRIALS

FDA Clinical Trials and the Regulatory Process

A look at how initiatives being implemented by the FDA affect clinical investigation of contact lenses.

Dr. Lippman is vice president of Ophthalmic Product Regulatory Affairs for R.P. Chiacchierini & Associates, LLC, Statistical & Regulatory Consulting, Rockville, Md. He is former director, Division of Ophthalmic Devices, Office of Device Evaluation, Center for Devices and Radiological Health, Food and Drug Administration from 1984 to 1993. He has been consulting in regulatory affairs and clinical trials to the ophthalmic device industry for the past 19 years. He is also a consultant or advisor to Bausch + Lomb, Vistakon, CooperVision, and SynergEyes. You can reach him at rel@rpcaconsulting.com.

By Richard E. Lippman, OD, FAAO

The regulation of medical devices in the United States is codified by law and interpreted and implemented by the Code of Federal Regulations. The Medical Device Amendments were promulgated in 1976 when Congress split off devices from drugs and created a separate Center for Medical Devices. Implementation of the device regulations was further defined by specific product guidelines written to address the scientific needs and requirements for each product group. These guidelines were developed in the United States Food and Drug Administration (FDA) ophthalmic division and were further enhanced by input from industry groups and the Ophthalmic Devices Panel experts who met periodically to review applications submitted to the FDA for approval. The guidelines are intended to outline the scope of preclinical and clinical tests needed to determine device safety and efficacy.

Recently, the agency has published new Guidance relating to its initiative to improve the quality of clinical trials used to support both Premarket Approvals (PMAs) and 510(k) clearances.

While this Guidance restates much of what has transpired in regulatory clinical study design in recent memory, it provides a comprehensive review of the nature of clinical investigations and the impact that good clinical design and statistical planning have on medical device approvals. These initiatives come on the heels of changes brought about by Congress regarding the medical device program as well as the introduction of drug study design models brought over from the Center for Drug Evaluation at the FDA.

This article describes many of the initiatives being implemented by the FDA regarding clinical investigation of medical devices and, more specifically, how these initiatives relate to clinical studies of contact lenses.

Medical Device Classification

Medical devices are classified into three specific classes: Class I—general controls, Class II—special controls and substantial equivalence to a predicate device on the market since May 1976, and Class III—premarket approval. All devices are initially assigned a Class II designation until further classified by classification panels. All contact lenses were initially placed into Class III under that paradigm. In 1994, daily wear contact lenses were down-classified to Class II, but extended wear continued to be regulated as premarket approval in Class III. The classification of a given device usually determines what type of clinical investigations are required to support an application. Clinical trial requirements for extended wear contact lens indications far exceed those for a daily wear approval.

Regulatory Clinical Trials

For each regulatory classification, clinical studies take different approaches. Product-specific Guidance provides the means to determine what kind of clinical study is necessary to support a given submission. For daily wear contact lenses, most studies are required to support a 60- patient completion profile of three-months duration for a new daily wear product to demonstrate substantial equivalence to a predicate contact lens already cleared for marketing. This sample size is approximate, and is to be divided into test and control arms, usually in a 2:1 ratio, test lens to a known control lens. To complete a full cohort of 60 subjects, it is often necessary to enroll an additional 15 percent of the proposed sample to cover eventual dropouts and discontinuations over the course of the clinical trial.

Extended wear clinical trials are subject to a greater regulatory demand that includes a required minimum of one year (or more) of follow up with a sample population of at least 400 completed subjects, also divided into test and control arms. Safety concerns for prolonged overnight use defines the rationale for a more closely watched investigational period. Whether the material is polyHEMA or silicone hydrogel (SiHy), the indication for use governs the nature and length of the trial necessary to establish safe and effective determination.

Because new lens materials (SiHy) have been developed over the past 10 years, general clinical guidance rules, while still applicable, have demanded larger cohorts to establish more comprehensive reporting and understanding of the unique characteristics of these materials. Material characteristics of SiHy lenses that are unfamiliar to the agency—and not yet identified into lens groupings—have created an environment in which clinical study requirements are greater in terms of size and duration. Additional specialized testing may include endothelial cell studies, contrast sensitivity analysis, near-point visual analysis, topographic studies, etc., and have been employed in certain circumstances. While daily wear studies remain limited in scope at a 90-day follow-up period, extended wear cohorts of a one-year follow-up period are often saddled with longer-term postapproval follow up of three years on new and novel materials. The FDA has imposed post-market study requirements for this broader group of SiHy materials in the extended wear indication.

Studies are being designed with increased attention to statistical design to better define the study objectives and their outcomes. The early SiHy lens studies included much higher populations compared to that stated in the Guidance Document for Extended Wear Contact Lenses. These early studies included some 800 to 1,100 subjects in their respective protocols. Outcome measures were modeled to demonstrate target outcomes. In general, for targets that are not achieved, additional steps must be taken to demonstrate a safe and effective product. This is still true for new material studies today. Target outcomes have more of a drug model design compared to what has been used traditionally for medical device studies.

One reason for the increase in study size as indicated above has been the challenge of defining the lens groups under which SiHy lenses fall due to unique material properties. There is currently no consensus for identifying SiHy group characteristics, leaving the industry to identify individual material characteristics and to test and validate their specific material profile. This has given rise to a level of regulatory discomfort regarding potential safety and effectiveness with these products. Until now, each new material has been defined as its own group, unlike traditional polyHEMA lenses, which are defined into one of four group categories based on water content and ionic/non-ionic surface characteristics (Tables 1 and 2). The effort to define lens groups for SiHy lenses continues to evolve and will be better defined once the FDA and consultants reach consensus regarding the more generalized material characteristics and predictability as to material behavior.

Clinical Study Categories/Valid Scientific Evidence

The FDA divides clinical investigations into two categories: significant risk and non-significant risk. Significant risk includes investigations of medical devices for which serious safety concerns, such as the potential for life threatening and/or sight threatening conditions, may be present. The corollary is nonsignificant risk device investigation, in which major concerns for catastrophic events are diminished. For example, daily wear contact lenses are defined under the non-significant risk category whereas extended wear is considered as significant risk.

The regulatory determinations associated with significant versus non-significant risk are defined in the Code of Federal Regulations (21CFR 812) regarding investigational device exemptions for conducted clinical trials. For daily wear studies, the non-significant risk designation allows the conduct of studies under the abbreviated rules, while extended wear requires the submission of an Investigational Device Exemption (IDE) application directly to the agency. While Institutional Review Board approval is required for either daily wear or extended wear studies, the abbreviated IDE does not require submission to the agency for prior approval before starting a clinical investigation, whereas an IDE approval is necessary to commence an extended wear study. These distinctions often govern the approach that an industry sponsor will take to introduce a new product in the effort to gain market clearance sooner than would otherwise occur (i.e., daily versus extended wear). Nevertheless, sponsors are encouraged to approach the agency prior to the start of any study for early collaboration and input with the intention of avoiding potential problems in study design.

Premarket Approval and Valid Scientific Evidence

For PMAs, the statutory definition of evidence provided in the Food, Drug, and Cosmetic Act is: reasonable assurance of safety and effectiveness. Interpreted, that means that safety and effectiveness of a device are to be determined:
• With respect to the persons for whose use the device is represented or intended,
• With respect to the conditions of use prescribed, recommended, or suggested in the labeling of the device, and
• Weighing any probable benefit to health from the use of the device against any probable risk of injury or illness from such use.

Further, by regulation, “…the valid scientific evidence used to determine the safety of a device shall adequately demonstrate the absence of unreasonable risk of illness or injury associated with the use of the device for its intended uses and conditions of use.”

For 510(k) type studies, clinical performance and determination of substantial equivalence to a known predicate device determines a successful outcome for regulatory purposes.

Regardless of the class to which a study is assigned, the regulation applied to valid scientific evidence (21CFR860.7(c)(2)) states that it is gained from well-controlled clinical investigations, partially controlled studies, studies without matched controls, or well-documented case histories by qualified experts among other ancillary activities. It is not very often that the FDA has accepted historical controls as valid scientific evidence because of the propensity for anecdotal evidence and because the potential bias of data as reported in published literature does not match well with regulatory study designs. The FDA has thwarted the use of meta-analyses from peer-reviewed literature in support of 510(k)/PMA applications for those very reasons.

Claims

A claim is an assertion of truth about a product. Claims associated with a given device often dictate the clinical endpoints or measurements that indicate device success or failure. In contact lens studies, claims are more often ignored in favor of defining a safe and effective product. After initial studies and approvals are obtained, sponsors may come back and conduct further studies to support specific claims such as increased comfort, better water retention, superior oxygenation, and more.

Endpoints

Studies should define the principal study variables that demonstrate the performance of the device. These are defined as the primary and secondary clinical endpoints. They are further defined as the definition of the study objective. These endpoints should be pre-specified in the final clinical protocol and are further divided into safety and effectiveness endpoints.

Study designs should do the following:
• Characterize the clinical effects of the device (both safety and effectiveness) for the intended use.
• Describe what endpoints or outcomes are being measured, when they are being measured, and how they will be analyzed statistically.
• Be clinically relevant to the stated study objectives.
• Specify who will evaluate and analyze the study data. An independent team of analysts who have no proprietary interest in the study at hand is recommended.
• Be evaluated for specific claims such as lens deposit resistance (which should be evaluated by independent reading centers and be pre-defined in the study protocol).
• Provide patient surveys for use when the outcome of interest is measured from the subject's perspective (e.g., lens comfort).
• Provide scientific rationale when more than one primary or safety endpoint is desired by explaining the role and importance of each endpoint.

Safety Endpoints Primary safety endpoints in contact lens studies are identified by permanent loss of two or more lines of Snellen visual acuity. A percentage of subjects such as 1 percent or more who have lost two or more lines of Snellen acuity is often cause for concern and may impact the study outcome. An additional primary safety outcome would be associated with serious adverse reactions related to visual acuity loss as well as accompanying slit lamp findings and other associated observations.

Secondary safety endpoints are characterized through slit lamp findings by which in a standard clinical grading scale of 0 to 4, grades 3 to 4 are considered serious endpoint concerns for safety. Symptoms, problems, and complaints as well as additional safety-related observations including keratometric changes and adverse reactions that are deemed as serious or complicated all play a part in the safety analysis of a contact lens investigation.

Effectiveness Endpoints A primary effectiveness endpoint is defined as the percentage of subjects with best-corrected contact lens visual acuity of 20/25 and 20/20 or better. Wearing time can also be defined in this category. Additional secondary effectiveness endpoints can include lens comfort over time, symptoms of dryness with the lens for the recommended wearing time, etc.

A new area of interest in contact lens usage is myopia control in children. Recent journal articles and public discourse have expressed the desire to use contact lenses in specific lens designs, including aspheric and reverse geometry as well as other unique peripheral defocus curves, as a means to demonstrate a reduction in the progression of myopia. The effectiveness measure in these clinical outcome studies is the rate of change as compared to a control group in a parallel study format that does not wear the same lens design profile. These studies are currently being defined, and their time durations for follow-up periods are hotly debated.

Bias in Clinical Studies

Studies are designed to minimize the effect of bias— or the introduction of systematic errors. Bias may be reduced in the following ways:
• Careful inclusion and exclusion criteria.
• Use of diverse populations including age, gender, race, and ethnicity.
• Masking investigators, patients, and monitors as to the intervention.
• Coding of test and control lenses.
• Employing randomization.

Sample Size Determination

For regulatory clinical contact lens studies, sample size has been pre-determined by prior Guidance and is addressed above. All studies are comparative to a known control lens. Larger studies lend themselves to better statistical grounding while shorter studies provide for a clinical impression through summary statistics.

Sample size is more often determined by study purpose and material uniqueness—to demonstrate substantial equivalence to a known predicate device, or to demonstrate safety and effectiveness as in extended wear. Newer-generation silicone hydrogel lenses with no defined predicates often need larger sample cohorts to better illustrate their unique clinical characteristics.

Clinical Outcome Studies—Types of Study Designs

Most contact lens studies are designed as clinical outcome studies. Subjects are assigned to either the test lens or to a control lens and at scheduled follow-up visits are examined for various objective and subjective observations including best-corrected visual acuity, slit lamp analysis, subjective symptoms, problems and complaints, keratometry or topographic analysis, corneal thickness (pachymetry) evaluation, and wearing time. Safety analysis includes monitoring for adverse events, both serious (corneal ulcers/bacterial keratitis, iritis, central infiltrates, permanent loss of Snellen visual acuity of two or more lines) and significant (peripheral infiltrates, superficial punctate keratitis, abrasions, hyperemia, conjunctivitis, etc.). Special clinical protocols such as for orthokeratology, for example, include analysis of refractive change over time, stability of change, and regression of visual acuity during the course of the study in addition to the customary clinical study parameters discussed above. Recent open discussion of myopia control is a relatively new area of interest that will be closely evaluated in the coming years.

All study observations are cross-referenced to associated clinical findings and how they relate to a specific finding so as not to miss anything that may be associated with a particular clinical observation.

For example, how adverse events may be related to slit lamp findings, visual acuity loss, symptoms, etc., are evaluated in the analysis process. These kinds of analyses often provide information relating to a contact lens fitting characteristic, lens design issue, lens material compatibility, and other clinical or manufacturing concerns.

Each of these analyses is compared to an active control that is reported in a parallel fashion, upon which the data is compared for outcomes in all subsections of the study.

Types of Comparative Study Designs

There are three types of comparative study designs:

Parallel Group Design Each subject is assigned to only one intervention, and comparisons are made between subjects in each study group. Randomization assures an equal footing between subgroups. These designs are most common in standard contact lens regulatory study plans.

Paired Design Each subject receives both the test and control interventions. For example, one eye receives the test device, the opposing fellow eye receives the control device. Randomization and investigator masking help to reduce inadvertent bias being introduced into the study.

This design assures that every subject is exposed to each device (lens) whereas in parallel designs, variability may exist between subjects that confound observations among the enrollees.

Cross-Over Design A design in which all subjects receive two interventions at different times. One group of patients receives the test lens and the second group receives the control lens. At some point in time, these are reversed. A wash-out period between the change-over is provided to minimize residual effects from the initial lens usage. There is a stronger statistical basis for designs of this nature, which is useful when subtle differences are suspected between the test and control lens.

Controls Used in Contact Lens Studies

Controls used in clinical studies and how they relate to studies of contact lenses include:

Active Concurrent Control The most common type of control used in contact lens studies is the active control—a lens of relatively comparative features and indication for use. Ideally the lenses are similar in material design and indication, although that is not always possible.

Historical Control This type of control involves comparisons of test devices to known predicate devices for which the data has been derived from published peer-reviewed literature or another form of validated study. The FDA has difficulty in accepting this form of control lens because much if not all of published literature bears little resemblance to that of a standardized clinical format for regulatory purposes. Meta analyses from these kinds of studies provide only loose associations to real-time clinical study designs and are therefore prone to poolability challenges for purposes of regulatory trials. Historical controls most often do not meet the agency's definition of valid scientific evidence as defined under 21CFR 860.7.

Placebo Controls This type of control design has no place in contact lens studies. It is more appropriately used in accessory lens product studies such as in comfort drop solution studies.

No Control This design has been useful in situations such as corneal overnight swelling studies in which one of the test arms is no intervention. This provides the opportunity for a baseline comparison when compared to a test lens of unknown characteristics and a control lens with known swelling profiles.

The design has little use in prospective studies of new contact lens materials in which the only viable test would be against a known entity such as comparative contact lens.

Summary

Regulatory clinical studies are specifically defined, limited to clearly identified objectives and well described endpoints, and are highly accountable. These studies are made to support safety and efficacy for market clearances or approvals licensed by the FDA. Most studies follow well established Guidance published by the agency and follow good clinical practice regulations and international clinical harmonization guidelines.

Daily and extended wear contact lens regulatory studies have evolved over many years, but are constantly being scrutinized for best clinical practices in an ever-changing clinical environment. This has been evident with the development of more sophisticated contact lens materials and will be subject to further modifications as even newer materials come into the marketplace.

For silicone hydrogel lenses for which no group class has been defined, it is anticipated that each new material being advanced will be considered as unique, making the demands for clinical trials of these materials more burdensome to sponsors. CLS

To obtain references for this article, please visit http://www.clspectrum.com/references.asp and click on document #202.