In laboratories, mice live longer thanks to interventions that have not yet produced comparable effects in humans. This gap between animal results and real human outcomes is becoming one of the main problems in biomedicine: it forces us to rethink how exactly to use preclinical research to understand aging.
An article by Steven Austad, scientific director of the American Federation for Aging Research, published in Nature Aging in June 2026, raises a pressing question: is pragmatism needed when working with animal models? Without it, resources are wasted, and promises of extending healthy life remain hypotheses. The crux of the problem: how to choose models that have proven valuable for understanding basic mechanisms but often fail to predict success in humans?
Austad is a leader in comparative biology of longevity. His research draws on observations of wild opossums, which became a turning point in his career. In the 1980s, he discovered that an isolated population of opossums on Sapelo Island off the coast of Georgia lived 25–50% longer than their mainland relatives—a gap that reflected evolutionary differences in adaptation to predator pressure. Work at the Nathan Shock Center and Austad's publications show that short-lived rodents are useful for studying cellular processes, but their very evolutionary history limits the universality of conclusions.
Animal models have an obvious advantage: they allow manipulation of genes and living conditions while observing the full life cycle in a reasonable time. But the translatability of results to humans remains a weak point. In oncology, the average translation success rate is less than 8%; in Alzheimer's disease research, success is even lower—only 58% of animal experiments correlate with clinical trial results, while 99,6% of compounds that pass preclinical tests ultimately fail to receive regulatory approval. Austad calls for change: select models based on similarity in key biological pathways, focus on health and functionality rather than just lifespan, and necessarily supplement animal data with studies on human cells and biomarkers.
Imagine renovating a historic building: blueprints help understand the structure, but they do not account for how real loads and materials behave over time. The same happens with mice—they reveal fundamental processes, but without considering species differences in metabolism, immune function, and lifespan, any conclusions remain approximate.
Strategies to improve translation include using a broader range of animal models, computational modeling to integrate data, and prioritizing combination interventions—that is, strategies tested at multiple levels of complexity simultaneously. At the same time, research funding requires greater transparency: conflicts of interest can subtly distort priorities and divert attention from fundamental questions.
Pragmatism in preclinical work does not guarantee rapid breakthroughs in understanding or treating aging in humans. But it can prevent wasted years and resources—and allow for a more honest dialogue between the laboratory and the clinic.
