Stress Hormone Levels in Urine Show No Clear Link to Osteoporosis in Older Adults

For decades, scientists have suspected that chronic stress might quietly erode the human skeleton. The idea is biologically plausible: prolonged activation of the body’s central stress machinery floods tissues with cortisol, a hormone known to accelerate bone loss when present in excess. Now, a large population-based study from Taiwan has put that hypothesis to one of its most rigorous community-level tests, and the results suggest the story is more complicated than many researchers hoped.

A team of investigators led by researchers at Hsinchu MacKay Memorial Hospital, MacKay Medical University, and National Taiwan University analyzed data from 847 community-dwelling adults aged 53 and older who took part in the 2006 Social Environment and Biomarkers of Aging Study, a landmark Taiwanese survey known as SEBAS that integrates self-reported health information with objective biological measurements. Their question was deceptively simple: does naturally occurring variation in cortisol, measured noninvasively in overnight urine, predict osteoporosis in ordinary aging populations? Their answer, published in BMC Endocrine Disorders, was a careful and statistically robust no, at least not in any straightforward dose-response sense.

The study’s methodological strength lies in how it captured cortisol. Rather than relying on blood samples, which reflect momentary hormone spikes influenced by time of day, meals, and acute anxiety, the researchers measured cortisol in twelve-hour overnight urine specimens. They then expressed the result as a urinary cortisol-to-creatinine ratio, a standard technique that corrects for dilution differences in urine volume and provides a more stable index of integrated hormone secretion during sleep, when the confounding noise of daily activity is minimized. Creatinine, a byproduct of muscle metabolism excreted at a relatively constant rate, serves as the internal normalizing yardstick.

Participants were sorted into four groups, or quartiles, according to their cortisol-to-creatinine ratios, from lowest to highest. The average participant was 65.6 years old, and women made up 46.5 percent of the sample, a distribution well suited to studying osteoporosis, which disproportionately affects postmenopausal women. The outcome measure was self-reported osteoporosis, an acknowledged limitation, but one that captures clinically recognized diagnoses within a large, real-world population.

The statistical analysis employed multivariable logistic regression, the workhorse method of observational epidemiology. By adjusting for a panel of potential confounders, the model isolates the association between cortisol exposure and osteoporosis while holding other risk factors statistically constant. Compared with participants in the lowest cortisol quartile, the adjusted odds ratios for self-reported osteoporosis were 1.04 in the second quartile, 1.07 in the third, and 1.21 in the fourth. Although the point estimates crept upward with higher cortisol, none of the confidence intervals excluded the null value of one, meaning none of the associations reached statistical significance. A formal test for linear trend across quartiles returned a p-value of 0.463, far from the conventional threshold of 0.05, effectively ruling out a consistent gradient of risk.

The one genuinely intriguing signal emerged in exploratory subgroup analyses stratified by age. The researchers detected a statistically significant interaction by age group, with a p-value for interaction of 0.004. In plain terms, the relationship between cortisol and osteoporosis appeared to differ between younger and older participants in the sample. Yet, crucially, neither age group displayed a consistent monotonic pattern across cortisol quartiles, meaning the interaction did not resolve into a clean story of harm in one group and protection in the other. The authors themselves urge caution, noting that such findings can arise by chance in secondary analyses and require confirmation before any clinical interpretation is attempted.

Why does the broader biological hypothesis remain compelling even in the face of a null result? The hypothalamic-pituitary-adrenal axis, or HPA axis, is the body’s master stress circuit. The hypothalamus signals the pituitary gland, which in turn prompts the adrenal cortex to release cortisol. In controlled medical contexts, such as Cushing’s syndrome, where the body produces pathological excess cortisol, or during long-term glucocorticoid therapy, the bone-destroying power of the hormone is undisputed: cortisol suppresses osteoblast activity, promotes osteocyte apoptosis, and shifts the delicate balance of bone remodeling toward resorption. The open question has always been whether the subtler, physiologic range of cortisol variation seen in everyday stress reaches a threshold capable of measurable skeletal damage at the population level.

The new findings suggest that, at least as captured by a single overnight urinary measurement, ordinary endogenous cortisol variation does not register as a meaningful osteoporosis risk factor. Several explanations deserve consideration. First, a single twelve-hour urine specimen is a snapshot of HPA axis activity on one night, whereas cumulative cortisol exposure over years or decades may be what matters for bone. Repeated sampling across multiple days and time points would better approximate the long-term hormonal milieu that bone cells actually experience. Second, the outcome was self-reported osteoporosis rather than objective bone mineral density measured by dual-energy X-ray absorptiometry, the diagnostic gold standard. Misclassification of the outcome can dilute true associations toward the null. Third, the cross-sectional design captures a single moment in time, making it impossible to establish whether any hormonal difference preceded the development of bone disease.

These caveats are not reasons to dismiss the study; they are precisely the roadmap the authors propose for future work. Larger prospective cohorts with repeated cortisol measurements and direct imaging-based assessment of bone health would allow researchers to track whether HPA axis activity predicts subsequent bone loss, rather than merely coexisting with it. The SEBAS infrastructure, which links the Taiwan Longitudinal Study of Aging with biomarker collection, offers an unusually rich foundation for such inquiries, and the current analysis demonstrates both the feasibility and the limits of single-biomarker cross-sectional approaches.

The study, supported by the National Science and Technology Council, Taiwan, and approved by the Research Ethics Committee of National Taiwan University Hospital, carries practical implications for clinicians and the public alike. People worried about osteoporosis should not interpret the findings as a signal that stress is irrelevant to bone health; rather, the results indicate that a single overnight cortisol measurement is not a useful screening tool for identifying osteoporosis risk in middle-aged and older adults. Established risk factors, including age, sex, menopausal status, body weight, physical inactivity, smoking, and inadequate calcium and vitamin D intake, remain the cornerstones of risk assessment. For researchers, the study is a sobering reminder that plausible mechanisms do not guarantee detectable population effects, and that null findings, honestly reported and statistically transparent, are essential building blocks of reliable science. As the authors conclude, the exploratory age interaction and the overall null association both demand confirmation in larger, longitudinal studies before the stress-bone hypothesis can be either advanced or retired.

Subject of Research: Association of overnight urinary cortisol-to-creatinine ratio with self-reported osteoporosis in community-dwelling Taiwanese adults

Article Title: Association between the overnight urinary cortisol-to-creatinine ratio and self-reported osteoporosis in community-dwelling Taiwanese adults: a population-based cross-sectional study

Article References: Kuo, Y.-C., Chen, H.-H., Hsu, H.-Y., Wang, J., Chien, K.-L., Yeh, T.-L., Tsai, S.-Y., & Lee, Y.-S. (2026). Association between the overnight urinary cortisol-to-creatinine ratio and self-reported osteoporosis in community-dwelling Taiwanese adults: a population-based cross-sectional study. BMC Endocrine Disorders. https://doi.org/10.1186/s12902-026-02568-5

Image Credits: AI Generated

DOI: 10.1186/s12902-026-02568-5

Keywords: osteoporosis, urinary cortisol-to-creatinine ratio, hypothalamic-pituitary-adrenal axis, cortisol, bone health, healthy aging, SEBAS, Taiwan, cross-sectional study, endocrinology, epidemiology, stress