Abstract
The regulatory standards of the United States Food and Drug Administration (FDA) require substantial evidence of effectiveness from adequate and well-controlled trials that typically use a valid comparison to an internal concurrent control. However, when it is not feasible or ethical to use an internal control, particularly in rare disease populations, relying on external controls may be acceptable. To better understand the use of external controls to support product development and approval, we reviewed FDA regulatory approval decisions between 2000 and 2019 for drug and biologic products to identify pivotal studies that leveraged external controls, with a focus on select therapeutic areas. Forty-five approvals were identified where FDA accepted external control data in their benefit/risk assessment; they did so for many reasons including the rare nature of the disease, ethical concerns regarding use of a placebo or no-treatment arm, the seriousness of the condition, and the high unmet medical need. Retrospective natural history data, including retrospective reviews of patient records, was the most common source of external control (44%). Other types of external control were baseline control (33%); published data (11%); and data from a previous clinical study (11%). To gain further insights, a comprehensive evaluation of selected approvals utilizing different types of external control is provided to highlight the variety of approaches used by sponsors and the challenges encountered in supporting product development and FDA decision making; particularly, the value and use of retrospective natural history in the development of products for rare diseases. Education on the use of external controls based on FDA regulatory precedent will allow for continued use and broader application of innovative approaches to clinical trial design, while avoiding delays in product development for rare diseases. Learnings from this review also highlight the need to update regulatory guidance to acknowledge the utility of external controls, particularly retrospective natural history data.