On December 17, 2012, the U.S. Food and Drug Administration (“FDA” or “Agency”) released a draft guidance for industry entitled, Enrichment Strategies for Clinical Trials to Support Approval of Human Drugs and Biological Products. The draft guidance document is intended to provide guidance to industry regarding enrichment strategies that can be used in clinical trials to support safety and effectiveness claims in new drug application (NDAs) and biologic license applications (BLAs). In brief summary, the draft guidance document defines specific types of enrichment strategies, outlines examples of potential clinical trial designs, and describes potential regulatory considerations when utilizing such strategies in clinical trials.

Enrichment strategies are used in product development to enhance efficiency and available data. Their use can lower development costs, reduce development times, and decrease the number of trial subjects required for evaluation. The term “enrichment,” as used in the draft guidance document, means the “prospective use of any patient characteristic to select a study population in which detection of a drug effect (if one is in fact present) is more likely than it would be in an unselected population.” Many different characteristics — demographic, historical, genetic, psychological — have been used for enrichment. The term may also refer generally to the population to be evaluated within a broader population.

The draft guidance document categorizes enrichment strategies as either prognostic, predictive, or designed to decrease heterogeneity. Prognostic enrichment strategies select patients with a greater likelihood of having a disease-related endpoint event or a substantial worsening in condition, which increases the absolute effect differences between groups but does not impact relative effect. Predictive enrichment strategies select patients more likely to respond to treatment than others, which may lead to a larger effect size in both absolute and relative terms thereby allowing the use of a smaller patient population. Strategies that decrease heterogeneity select patients with baseline measurements in a narrow range and exclude patients whose symptom or disease spontaneously improve or are highly variable; decreasing inter-patient variability increases study power, or the ability of a clinical trial to demonstrate a treatment effect if one exists. These categories are discussed in the draft guidance document, which also provides study design options for each, discusses the advantages and disadvantages of the several designs, and highlights the issues associated with interpreting the results of enrichment strategies. As an overarching theme, FDA reminds readers in the draft guidance document that the primary concern with implementing enrichment strategies relates to the generalizability and applicability of the study results.

Prognostic Enrichment Strategies
Prognostic indicators, such as clinical and laboratory measures, medical history, and genomic or proteomic measures, have been used to identify potential subjects with a greater likelihood of having the event of interest to be studied in a trial. Selecting optimal patients allows a treatment effect to be detected more readily. Trials of preventative treatment that reduce the rate of death or other serious event are generally more successful if the patients enrolled have a high event rate; a high-risk event rate will increase the power of the study to detect risk reduction. Prognostic strategies may also be employed for treatments intended to delay progression of conditions, where patients with more rapid progression could be selected, acknowledging however that rapidly progressing patients may be less responsive to treatment. Prognostic enrichment does not increase the relative risk reduction but will increase the absolute effect size, which typically provides for a smaller sample size. From a design perspective, patients at high risk for events are commonly used in an initial outcome study, and if the study is successful, the sponsor may then proceed to larger studies in lower-risk patients.

The draft guidance document outlines experience with prognostic strategies in cardiovascular and oncology studies of enalapril and statin, and tamoxifen, respectively. The draft guidance document also outlines additional approaches to prognostic enrichment for cardiovascular, prostate cancer, and breast cancer studies, for example.

Predictive Enrichment Strategies
Predictive strategies, which identify subjects more likely to respond to a particular intervention, have long been employed when subject selection has been based on a specific aspect of pathophysiology, past history of response, or a disease characteristic related to the study treatment’s mechanism of action. Identifying and studying a responding population can increase study efficiency or feasibility and enhance the benefit-to-risk balance for the subset relative to the overall population. Identifying a high-responding population can significantly increase the chance a study will be able to detect a treatment effect. The strategy can be useful in early effectiveness studies by providing proof of concept and helping to define appropriate dose selection.

The draft guidance document further divides predictive enrichment strategies into five (5) categories, namely: empiric strategies, pathophysiological strategies, genomic strategies, randomized withdrawal studies, and studies in non-responders or patients intolerant to other therapy. Under an empiric strategy, subjects are selected based on observed response during screening periods or based on prior experience with the treatment or a related treatment (as compared to selection based on an underlying understanding of why patients respond differently). Relevant designs include open trial followed by randomization, enrollment based on an individual’s historic response to a particular product or therapeutic class, and utilization of factors identified from previous study results. Pathophysiological strategies rely on the selection of likely responders based on the subject’s own physiology or the disease pathophysiology. This may include evaluating how the subject treatment is metabolized, its effect on the target site of action (e.g., tumor metabolism), or proteomic markers or genetic markers linked to a proteomic marker. The third category, genomic strategies, may or may not be linked to a pathophysiologic property, making it better classified as an empiric strategy itself rather than a subset of a pathophysiological strategy. With respect to a randomized withdrawal study, patients responding to treatment during an open-label period or in a treatment arm of a randomized clinical trial are subsequently randomized to continue drug treatment or placebo. This design may be employed to establish long-term effectiveness where use of a placebo would not be acceptable or as an initial trial to show effectiveness in an open-label treatment situation. Finally, a study can be enriched by including non-responders or those intolerant to other therapy, which provides useful information regarding between-drug comparisons of effectiveness particularly when evaluating therapeutic agents from the same or a similar therapeutic class.

Decreasing Heterogeneity
Employing strategies to increase study power by decreasing heterogeneity is well accepted and routinely employed. Useful strategies include, but are not limited to

• defining entry criteria so that enrolled patients actually have the condition of interest and training investigators adhere to those criteria and definitions included in study protocols; identifying and selecting patients likely to comply with treatment to reduce drug exposure variability;
• using placebo lead-in periods before randomization to exclude patients whose condition improves spontaneously or who have large placebo responses;
• excluding patients unlikely to tolerate the treatment; or
• excluding patients who are likely to drop out for non-medical reasons.

The draft guidance document places additional focus on encouraging compliance and reducing spontaneous improvers/placebo responders.

Regulatory Considerations
Using an enrichment strategy in clinical trials has direct regulatory implications. In the draft guidance document, FDA points out two key considerations for sponsors, namely: (1) can the enrichment strategy be used to identify the patients to whom the drug should be given, and (2) would the drug be useful in a broader population than was studied? The resolution of these questions clearly affects the product labeling, which the draft guidance document acknowledges. Enrichment designs will likely impact product labeling as the design elements are integral components relayed in the Indications and Usage, Dosage and Administration, and Clinical Studies sections. The use of enrichment studies in clinical trials should be contemplated in light of the potential regulatory outcomes affecting the overall use of the drug and the potential available market. Recognizing the limitations of enriched populations and the generalizability of study results, the Agency states that post-market commitments or requirements may be requested to better understand the true extent of a treatment’s effect (i.e. efficacy and/or safety studies in a broader population). As always, FDA recommends engaging the Agency review staff to discuss potential study design and how much information should be developed prior to drug approval.

The comment period for the draft guidance document was originally set to close in February 2013, but to allow additional time for interested parties to comment, the Agency has extended the comment period through March 18, 2013. Although those interested are welcome to provide comments on FDA guidance at any time, the Agency encourages individuals to submit electronic or written comments on the draft guidance document (Docket No. FDA-2012-D-1145) before it begins work on the final version. The draft guidance document is available at: http://1.usa.gov/YRCQKf (PDF); the original Federal Register notice is available at: http://1.usa.gov/ Y4OqPq (PDF); and the notice of extension is available at: http://1.usa.gov/X027Db (PDF).

---

Colleen Heisey is a partner in the Washington, D.C. office of Hunton & Williams LLP (hunton.com) and a member of the firm’s Food and Drug Practice. She can be reached by email at cheisey@hunton.com.