FDA shifts toward one pivotal trial plus confirmatory evidence for most new drugs

The FDA announced via a Sounding Board in the New England Journal of Medicine that one pivotal trial, accompanied by confirmatory evidence, will be the default standard for approval of most new drugs. The agency argues that modern science is more precise and that biomarkers and Bayesian approaches can complement traditional replication, while critics argue the shift risks betting public health on a single performance.

For decades, the expectation that new drugs demonstrate efficacy in two independent, adequate, and well-controlled clinical trials served as a simple safeguard: replicate the finding before exposing patients. The two-trial norm dates to reforms of the early 1960s, when Congress required adequate and well-controlled investigations before drug approval. One positive result can occur by chance, while two independent trials that converge reduce the risk of false positives.

The FDA said two trials may provide false assurance if design flaws persist and argued that lowering the default requirement will reduce development costs, shorten time to market, and spur innovation. The rationale is that modern science is more precise, mechanisms are better understood, and supportive data can substitute for replication. Critics counter that mechanism is not outcome, and that improvements in laboratory values or surrogate markers do not necessarily translate into longer survival, fewer hospitalizations, or better quality of life.

More than half of recent approvals have relied on a single pivotal trial plus supportive evidence, particularly in oncology and rare diseases, and statutory authority for single-trial approval has existed since 1997. But single-trial approvals have often addressed dire unmet needs, including advanced cancers or rare disorders, where large replication trials may be impractical, and many were accelerated approvals with confirmatory trials required post-marketing.

The new policy extends the one-trial default beyond those settings and may now apply to drugs for common conditions such as hypertension, diabetes, and psychiatric illness, where millions will be exposed. The article said the risk calculus changes when the denominator is not 5,000 patients but 50 million.

Proponents argue that eliminating a second pivotal study, often costing $30 million to $150 million, will reduce development time and expense. But the article said there is little historical evidence that research and development savings reliably reach patients, while the potential cost of error remains substantial.

The FDA paired the shift with promises of enhanced post-market surveillance and real-world data collection, but the article said post-market monitoring has been uneven and adverse event systems are incomplete and subject to bias. It also pointed to a broader reproducibility problem, noting that single studies, even statistically significant ones, often fail to replicate and that replication is how science separates signal from chance, particularly when millions of patients may ultimately be exposed.