Drug Repurposing

Definition

Drug repurposing is the search for and validation of a new therapeutic use for an existing medicine beyond the indication for which it was originally developed or approved.

Current Synthesis

An existing drug can be a valuable starting point because its chemistry, manufacturing, and some human safety information are already known. It is not a shortcut around evidence. A new disease, population, dose, formulation, route, combination, or duration can change both benefit and harm, so approval for one indication does not prove another.

The episode frames repurposing as both a scientific and institutional search problem. Drugs can interact with multiple biological targets, relevant observations may sit in another specialty, and rare-disease clinicians or patients may see signals that have not diffused. Meanwhile, generic status and small patient populations can weaken the commercial return from funding trials. Knowledge mapping and machine learning may help prioritize matches, but responsible translation still requires escalating evidence through mechanism, laboratory work, clinical observation, trials where feasible, safety review, and expert judgment.

The examples are deliberately heterogeneous. Sirolimus in Fajgenbaum’s Castleman disease and pembrolizumab in one angiosarcoma patient are hypothesis-generating clinical stories; the described lidocaine breast-surgery trial and colchicine cardiovascular indication sit at different evidentiary stages. Good synthesis preserves those differences instead of turning every surprising use into a cure claim.

Key Claims

• Known manufacturing and prior human exposure can reduce some development uncertainty without establishing a new indication.
• Multi-target pharmacology and fragmented literature create plausible uses outside a drug’s original disease area.
• Generic drugs and rare diseases can face an incentive gap because no actor can easily capture the value of additional evidence.
• Patient networks and cross-specialty experts can surface observations that ordinary diffusion misses.
• Computational ranking is useful for prioritization, not proof of clinical benefit.
• Evidence must distinguish individual cases, mechanisms, observational associations, controlled trials, and regulatory decisions.
• New use can alter dose, route, formulation, interaction, population, and safety requirements.

Evidence

• Search and incentive problem - Using Existing Drugs in New Ways to Treat & Cure Diseases of Brain & Body | Dr. David Fajgenbaum links multi-target drugs, generic status, rare diseases, and fragmented medical knowledge to missed uses.
• Patient-derived hypothesis - Using Existing Drugs in New Ways to Treat & Cure Diseases of Brain & Body | Dr. David Fajgenbaum describes Fajgenbaum’s mTOR finding and physician-supervised use of sirolimus for Castleman disease.
• Scaled prioritization - Using Existing Drugs in New Ways to Treat & Cure Diseases of Brain & Body | Dr. David Fajgenbaum presents Every Cure as combining knowledge mapping and machine learning with laboratory and clinical validation.
• Evidence heterogeneity - Using Existing Drugs in New Ways to Treat & Cure Diseases of Brain & Body | Dr. David Fajgenbaum ranges from individual patients and small papers to observational findings, a large trial, and indication approval.

Counterevidence & Qualifications

Repurposing can sound safer or faster than it is. Prior approval may not cover the new dose, route, population, combination, chronicity, or disease physiology; rare adverse effects and interactions can remain undiscovered. Positive cases are vulnerable to selection, publication, regression-to-the-mean, and concurrent-treatment effects. The source is a public interview summary and does not independently verify its numerical outcomes or establish a prescribing protocol. Self-directed use is especially unsafe for immunosuppressants, cancer drugs, cardiovascular drugs, psychoactive compounds, and medicines requiring monitoring.

What Changed

• Established drug repurposing as a combined evidence-search, incentive, diffusion, and validation problem.
• Added a hierarchy separating computational rank and individual rescue from controlled and regulatory evidence.

Related Concepts

• AI For Science - computational context for connecting dispersed biomedical findings.
• AI Clinical Validation In Drug Discovery - requires model-prioritized candidates to survive human-outcome testing.
• Patient-Led Rare-Disease Infrastructure / 患者主导的罕见病基础设施 - coordination layer that can surface needs, cases, samples, and experts.
• Self-Experimentation - hypothesis-generating practice with severe single-subject limits.
• Medical Risk Management - safety framework for off-label and novel-indication decisions.
• Clinical Trial Continuity - neighboring model for accumulating and reviewing clinical evidence.