Lawrence Livermore Scientists Found 27 Drugs Linked to Longer ALS Survival in Veterans

Researchers at Lawrence Livermore National Laboratory identified 27 existing medications associated with longer survival in people with amyotrophic lateral sclerosis, using a machine learning analysis of veterans' health records that represents one of the largest ALS electronic health datasets ever assembled.

The study, published in The Lancet Digital Health, analyzed records from more than 11,000 U.S. military veterans diagnosed with ALS between 2009 and 2019 and treated within the Veterans Health Administration. Rather than developing new compounds from scratch, the LLNL-led team screened 162 drugs already prescribed for other conditions, searching for any associated with meaningful differences in survival among ALS patients.

ALS, the fatal neurodegenerative disease that destroys the nerve cells controlling movement, currently has no cure and only limited approved treatments that modestly slow its progression. The average survival after diagnosis is two to five years. For the roughly 30,000 Americans living with ALS at any given time, identifying existing drugs that might extend survival represents a faster and less costly path than a traditional drug-development pipeline.

The team combined causal-inference methods with machine learning to work through the records, a pairing designed to move beyond simple correlations and account for the complex, messy reality of real-world patient data. LLNL researchers Andre Goncalves, Priyadip Ray, Jose Cadena Pico, and Braden Soper were among the scientists and computational engineers who conducted the work.

The study was performed in collaboration with Stanford University School of Medicine, the VA Palo Alto Health Care System, and the University of California, Los Angeles.

The scale of the dataset was itself a significant feature of the research. Most ALS clinical trials enroll hundreds of patients; this analysis drew on more than a decade of treatment records for more than 11,000 veterans, a cohort large enough to detect signals that smaller studies would miss. The VA's integrated electronic health record system, which tracks prescriptions, diagnoses, and outcomes across a unified national network, made that scale possible.

The approach reflects a broader shift in how computational researchers are attacking diseases with limited treatment options. Drug repurposing, the strategy of finding new uses for approved medications, bypasses years of safety testing because these compounds already have established human safety profiles. If any of the 27 identified medications hold up in prospective clinical trials, patients could potentially benefit far sooner than through conventional drug discovery.

The study's language is deliberately cautious. The findings show associations, not proof that any drug extends life. Translating these signals into confirmed treatments will require randomized controlled trials. The specific list of 27 medications, their therapeutic classes, and the magnitude of the survival differences observed were not detailed in LLNL's public release materials; the full data are available in The Lancet Digital Health paper.

For the veteran community, which faces ALS at roughly twice the rate of the general population, the findings carry particular weight. Military service has long been recognized as an independent risk factor for the disease, making the VA's patient population both a practical and scientifically relevant cohort for this kind of large-scale analysis.