Men with lower testosterone at 31 reported good health 15 years later

By Vijay Kumar MalesuReviewed by Susha Cheriyedath, M.Sc.Oct 27 2026

A 15-year Finnish birth cohort study examines how early-adulthood testosterone levels relate to metabolic disease, mood, sexual activity, and quality of life in middle age.

Study: Associations Between Testosterone Levels in Men at Age 31 and Quality of Life, Self-Reported Health, and Morbidity at Age 46: The Northern Finland Birth Cohort 1966. Image Credit: Panadesign / Shutterstock

In a recent study published in the journal Endocrine Connections, researchers found that testosterone concentrations at age 31 were not associated with substantial differences in self-reported health or quality of life (QoL) at age 46.

Background

Testosterone is an important hormone involved in reproductive function, muscle and bone development, metabolic regulation, and psychological well-being. Low testosterone concentrations have been shown to correlate with obesity, type 2 diabetes, cardiovascular disease, sexual dysfunction, and depression.

Testosterone levels have also declined over recent decades, including among healthy men. But evidence linking testosterone concentrations to mental health and health-related QoL remains inconsistent. Understanding these relationships is important for interpreting hormone measurements and evaluating men's long-term health. 

About the study

The researchers used data from the Northern Finland Birth Cohort 1966 (NFBC1966), a population-based cohort of men born in Northern Finland in 1966. The final study population included 2,691 men with testosterone measurements at age 31.

The men were divided into four quartiles (groups of roughly equal size) based on testosterone concentrations, ranging from below 17.2 nmol/L in the lowest group (Q1) to at least 25.5 nmol/L in the highest (Q4). The fourth quartile served as the reference group.

Fasting morning blood samples were collected, and serum testosterone concentrations were measured using liquid chromatography–mass spectrometry (LC/MS). Anthropometric measurements and questionnaire data provided information on body mass index (BMI), waist circumference (WC), smoking, alcohol consumption, education, physical activity, self-reported health, and medical conditions.

At age 46, responding participants completed health surveys and the 15D questionnaire, which assessed 15 dimensions of health-related QoL. Higher 15D scores indicated better QoL, with scores of 0.9–1.0 classified as excellent.

The researchers compared health outcomes across testosterone quartiles using appropriate statistical tests and regression models. Adjusted analyses took WC, smoking, and alcohol consumption at age 46 into account. Missing observations were excluded from the relevant analyses, and participants provided informed consent.

About 30% of the men lacked self-reported health assessments, while approximately 39% lacked usable 15D quality-of-life data. The researchers reported comparable missing-data rates across the testosterone groups, but nonresponse could still influence the results.

Study results

The median testosterone concentrations were 14.5, 19.2, 22.8, and 29.0 nmol/L in Q1, Q2, Q3, and Q4, respectively. The median in Q1 remained within the usual clinical reference range, so membership in the lowest quartile did not necessarily indicate testosterone deficiency.

At age 31, obesity decreased progressively across the quartiles, while smoking increased. Active smoking was reported by 27.1% of men in Q1 and 44.5% in Q4. Men in Q1 also reported lower alcohol consumption than those in Q4. Similar patterns were observed at age 46, but educational differences were no longer significant.

At age 46, 63.9% of men who answered the health questionnaire reported excellent or good health, 32.1% reported average health, and 4.0% reported poor or very poor health. The distribution did not differ significantly across testosterone quartiles.

Medical conditions showed some variation across the quartiles. Compared with Q4, men in Q1 and Q2 had higher unadjusted odds of metabolic and gastrointestinal diseases. These differences were no longer statistically significant after adjustment for WC. Men in Q3 had higher odds of pulmonary diseases than those in Q4, even after adjustment.

Separate analyses found no significant adjusted associations for asthma, emphysema, or obstructive sleep apnea. The authors cautioned that the pulmonary disease findings were uncertain because 76% of the relevant questionnaire data were missing.

Men in Q1 had a lower prevalence of substance use disorders, a finding that persisted in adjusted analyses. No significant differences were identified for several other conditions, including thyroid, neurological, rheumatic, psychiatric, and musculoskeletal disorders, or cancer.

Overall QoL was excellent across all groups, with mean 15D scores above 0.9. Unadjusted analyses found no significant differences. After adjustment for WC, smoking, and alcohol consumption, Q1 had slightly better overall QoL than Q4 when the groups were compared directly. The difference between the reported unadjusted mean scores was approximately 0.01 points, and the fully adjusted comparison across all four quartiles was not statistically significant.

Men in Q1 also had higher 15D depression scores, and this difference remained significant in the fully adjusted model. Their lower sexual activity score was no longer significant after adjustment for WC, smoking, and alcohol consumption. Men in Q3 initially had better breathing scores than Q4, but the difference weakened when smoking was included in the analysis.

Conclusions

Men with lower testosterone concentrations at age 31 generally reported good health and excellent QoL at age 46. The higher odds of metabolic and gastrointestinal diseases and the lower sexual activity score were no longer statistically significant once WC was taken into account.

The findings highlight the importance of considering obesity when interpreting testosterone-related health associations, but the study cannot establish a causal relationship. The cohort comprised Finnish men assessed at age 46, so the results may not apply to older men or populations in other settings. Testosterone was measured only once, medical conditions were self-reported, and the study involved numerous statistical comparisons.

The authors cautioned against unnecessary testosterone replacement therapy, but this study did not evaluate treatment or separately examine men with clinically diagnosed hypogonadism. Further longitudinal research is needed to clarify these associations.