Where Failed Drug Data Goes to Sleep — and How Researchers Wake It Up

When Pfizer and Eli Lilly discontinued their nerve growth factor inhibitor tanezumab in 2021, the decision ended one of the most closely watched non-opioid pain programs in a generation. But it did not end the life of the data. Thousands of patients across multiple Phase III trials had contributed joint assessments, imaging results, and adverse event reports. What becomes of that archive — and whether anyone outside the sponsoring companies can examine it — is a question that extends far beyond one molecule.

The Default Is Silence

A discontinued drug program has no regulatory mandate to publish further. Unlike an approval pathway, which compels labeling, risk evaluation and mitigation strategies, and post-marketing commitments, a voluntary withdrawal carries few disclosure obligations. Sponsors typically issue a press release, notify investigators, and file the minimum documentation with the FDA. The underlying datasets — case report forms, adjudication committee minutes, imaging reads — remain the property of the companies that paid for them.

This is not secrecy in the legal sense. It is simply the absence of a requirement. And for patient communities who enrolled in trials on the promise that their participation would advance medical knowledge, the gap between contribution and publication can feel like a broken covenant.

What the Law Actually Requires

The regulatory scaffolding for transparency in the United States rests on several pillars, each with real limits.

ClinicalTrials.gov, mandated under FDAAA 2007, requires registration and summary results posting for most trials of approved products — but enforcement historically applied only to studies involving FDA-regulated products, and civil penalties for noncompliance have been levied sporadically. Trials of drugs that never reach approval occupy a gray zone; summary results must still be posted for registered Phase II and III trials, but the granularity is coarse. A reader sees event counts, not adjudicated case narratives.

The Freedom of Information Act offers another route, but a blunt one. FOIA applies to records held by federal agencies, not to company files. A request to the FDA can surface review memos, complete response letters, and advisory committee transcripts — documents that often reveal the agency's reasoning in detail. What FOIA cannot do is unlock the sponsor's raw data. Exemption 4 protects confidential commercial information, and pharmaceutical companies invoke it routinely and successfully.

The Secondary Market for Data

A quieter mechanism has emerged over the past decade: controlled data sharing. Platforms such as Vivli and the Clinical Study Data Request Consortium allow qualified researchers to request access to de-identified patient-level datasets from participating sponsors. Pfizer was an early adopter of this model, and datasets from its pain portfolio — including NGF inhibitor studies — have been listed for request.

The process is not frictionless. Applicants submit a research plan, pass review by an independent committee, and accept terms restricting re-identification and secondary transfer. Critics note that the gatekeeping can shape what gets studied; a proposal examining a sponsor's least-favorable subgroup faces the same reviewers as any other, but the reviewer pool and criteria are set within a framework the sponsor helped design. Still, peer-reviewed reanalyses of abandoned programs have emerged through this channel, including independent examinations of joint-safety events in osteoarthritis populations that refined the risk estimates companies had reported only in summary form.

How Buried Findings Resurface

History offers instructive examples. The reanalysis of antidepressant trial data by researchers at Oregon Health & Science University, published in 2015, used data obtained through such portals and revealed publication bias that had inflated the apparent efficacy of an entire drug class. In the pain field, patient advocacy organizations — notably groups representing arthritis patients during the opioid crisis — have pressured sponsors and the FDA to release adjudicated safety findings from halted programs, arguing that a withdrawal decision itself constitutes safety information the public deserves.

The FDA's advisory committee records have proven especially valuable. When the agency convened a public meeting on NGF inhibitors in 2012 following reports of rapidly progressive osteoarthritis, the transcripts and briefing documents became a public archive of the controversy — accessible to anyone, indefinitely, regardless of what the sponsors later chose to publish.

Why This Matters for Pain Research

Failed analgesics are not merely commercial losses; they are negative knowledge. A dataset showing precisely how a drug damaged cartilage in a subset of patients can inform the design of the next NGF-targeting agent, refine patient screening, or rule out an entire mechanism. Conversely, when that data sleeps in a corporate archive, the field risks repeating the same trial — with the same enrolled patients bearing the same risks.

For clinicians counseling patients about experimental pain treatments, the practical advice is concrete: before enrollment, ask whether the trial is registered with a results-posting commitment, whether the sponsor participates in a data-sharing consortium, and what the consent form says about future use of the participant's data. These questions are no longer academic. In a research ecosystem where the most informative datasets increasingly belong to the programs that failed, transparency after termination is not a courtesy. It is the residue of the public's investment — in money, in trust, and in the bodies of the volunteers who made the science possible.