Some people seem to defy time. Their skin stays smooth, their energy levels remain high, and their hair isn’t gray when it should. In other words, some people just don’t look their age. What’s their secret?
Scientists at Osaka University in Japan may have an answer—and they claim it’s hidden in our hormones. The researchers have unveiled an AI-driven model that uses a mere five drops of blood to estimate a person’s biological age, a more precise measure of aging than the number of candles on a birthday cake.
Instead of relying on DNA or protein biomarkers, the new method analyzes 22 key steroids, including cortisol, testosterone, and other hormones that regulate metabolism, immunity, and stress responses. Besides separating biological age from chronological age, this system offers a new way to monitor health, detect disease risk earlier, and even guide personalized interventions that could slow down the aging process.
“Our bodies rely on hormones to maintain homeostasis, so we thought, why not use these as key indicators of aging?” Dr. Qiuyi Wang, co-first author of the study, said in a press release.
What Our Hormones Reveal About Aging
Aging is a complex interplay of genetics, lifestyle, and environmental factors. Traditional biological age tests often hinge on DNA methylation—tiny chemical changes to our genes—or protein levels in the blood. But these methods don’t fully capture the real-time chemical signals that influence aging. Hormone levels may be better suited for this task.
Rather than measuring the absolute levels of these hormones, the researchers focused on hormonal ratios, which account for individual differences and provide a more accurate snapshot of aging at a biochemical level.
The study analyzed serum samples from 148 individuals between the ages of 20 and 73. Using liquid chromatography–tandem mass spectrometry, they measured the concentrations of 22 different steroids, including cortisol (a stress hormone) and testosterone (a key sex hormone).
They then trained a deep neural network (or DNN; a type of AI) to predict biological age based on these hormonal patterns. The AI model learned the relationships between different steroids and how they change with age. Crucially, it accounted for sex-specific differences—men and women age differently at the molecular level, and the model was designed to reflect these variations.
The model’s predictions closely aligned with participants’ known health status. Those whose biological age exceeded their chronological age tended to have markers of accelerated aging, while those with a lower biological age often displayed traits of resilience.
The Stress-Aging Connection
One hormone, in particular, stood out.
“When cortisol levels doubled, biological age increased by approximately 1.5 times,” explains Professor Toshifumi Takao, a co-author and expert in analytical chemistry and mass spectrometry.
Cortisol, often called the stress hormone, is released during fight-or-flight responses. This provides concrete evidence that chronic stress doesn’t just feel exhausting—it physically accelerates aging at the cellular level.
Top Right: The AI-predicted biological age (BA) shows a general correlation with chronological age (CA), but individual differences widen over time.
Bottom: Using the metaphor of a “river widening as it flows downstream,” the illustration visualizes how biological age evolves with the passage of time.
Credit: Zi Wang
Another key insight: male smokers showed a significant acceleration in biological aging compared to their nonsmoking counterparts. Female smokers, on the other hand, surprisingly did not exhibit the same trend. The researchers speculate that hormonal differences might buffer women against some of the aging effects of smoking—at least temporarily.
“Stress is often discussed in general terms, but our findings provide concrete evidence that it has a measurable impact on biological aging,” Takao says.
Could This Blood Test Help Slow Aging?
The study offers a fresh lens through which to view the aging process. Traditional biological age models rely on large datasets with thousands of variables, but this DNN-based approach works with a much smaller set of biomarkers—just 22 steroids. If biological age can be accurately measured through hormone interactions, it could open doors to personalized health monitoring with just a simple blood test.
“This is just the beginning,” says Dr. Zi Wang, co-first and corresponding author of the study. “By expanding our dataset and incorporating additional biological markers, we hope to refine the model further and unlock deeper insights into the mechanisms of aging.”
However, challenges remain. The model was trained on a relatively small sample size, and its accuracy needs to be validated in larger, more diverse populations. Additionally, while the findings reinforce the idea that stress hormones accelerate aging, they don’t establish causation. Do high cortisol levels drive aging, or are they simply a byproduct of the body’s struggle against time? More research m
The findings appeared in Science Advances.
