Sex-differential testosterone response to long-term weight loss

Obesity is associated with chronic low-grade inflammation, altered adipocytokines, proinflammatory cytokines, and hypothalamic hormones that can impair gonadal function. Prior work has largely focused on premenopausal women, especially those with polycystic ovary syndrome (PCOS), showing that bariatric surgery reduces hyperandrogenism and improves menstrual regularity and fertility. In men, obesity is associated with reduced testosterone via increased aromatization to estrogen in adipose tissue, reduced SHBG, insulin resistance, and potential nutrient deficiencies. Weight loss and negative energy balance modulate the hypothalamic-pituitary-adrenal, thyroid, and somatotropic axes, influencing cortisol, stress responses, and growth hormone release. The authors previously showed bariatric surgery yields more effective long-term weight loss and improvements in inflammation and insulin resistance than diet-based therapy. Hypothesis: Effective weight loss interventions would lead to alterations in hypothalamic-pituitary hormones affected by obesity, with more pronounced changes after surgical procedures due to greater weight loss. Primary aim: assess the incremental effect of weight loss on endogenous sex hormones (testosterone, estradiol, SHBG) in men and women over 3 years. Secondary aim: compare long-term effects of RYGB, SG, and LAGB vs dietary intervention on gonadal axes, and explore effects on corticotropic (ACTH, cortisol), somatotropic (IGF-1), and thyroid axes (TSH, FT4, FT3) over 3 years.

The literature consistently reports that bariatric surgery alleviates clinical and biochemical hyperandrogenism in women with PCOS, reducing total and free testosterone and improving reproductive function. In men with obesity, cross-sectional and longitudinal studies show lower testosterone, linked to increased aromatization in adipose tissue, reduced SHBG, insulin resistance, and possible micronutrient deficiencies. Obesity and weight loss interventions influence multiple neuroendocrine axes, including HPA, thyroid, and somatotropic axes. Prior studies are often short-term and focus on younger populations or PCOS; there is limited evidence on androgen changes in postmenopausal women with obesity undergoing weight loss, especially when preoperative androgen levels are normal. Existing data suggest inverse relationships between adiposity and GH/IGF-1, with mixed findings on IGF-1 levels in obesity and potential reversibility after weight loss. Meta-analyses indicate that bariatric surgery can reduce TSH and FT3. The present study addresses gaps by assessing sex hormones and broader pituitary axes over 3 years in both sexes, including postmenopausal women.

Design: Prospective, observational cohort study of adults with obesity undergoing bariatric surgery (RYGB, SG, LAGB) or dietary intervention, followed for 36 months. Participants: Adults aged 18–70 years with BMI ≥30 kg/m² and at least 5 years of obesity despite prior weight loss attempts. Exclusions: Pregnancy or planning pregnancy within 2 years, within 5 years post-onset of menopause, or active psychiatric problems limiting adherence. Recruitment: From Obesity clinics at Royal Prince Alfred and Royal North Shore Hospitals and private bariatric services at St George Private and St Vincent's Hospitals, Sydney (2009–2012); follow-up until Oct 2015. Allocation: To bariatric or dietary interventions based on probability of diabetes remission (criteria congruent with DiaRem score). Sample: 64 recruited; 3 withdrew; 61 analyzed (RYGB n=7, SG n=21, LAGB n=12, Diet n=21). Timeline and measurements: Assessments at baseline and 1, 3, 6, 12, 24 months for all groups; 36 months for SG and RYGB (LAGB had BMD only at 36 months). Fasting blood samples (12 h overnight) collected at each visit. Assays: Total testosterone, SHBG (CV <4%), estradiol (E2) by electrochemiluminescent immunoassay (Roche Elecsys). Female free androgen index (FAI) calculated as (total testosterone/SHBG)×100. In males, calculated free testosterone (CFT) estimated by the Vermeulen equation. Cortisol and ACTH measured by Roche Modular E170 immunoassay (CV <6%). IGF-1 measured using Roche Elecsys with sex- and age-based reference intervals. Thyroid axis: TSH by Roche Modular analyzer; FT4 and FT3 measured per standard methods (details in Supplementary). Additional measures included LH and FSH. Ethics: Approved by St Vincent's Hospital HREC (HREC/09/SVH/64); informed consent obtained; registered at ANZCTR (12613000188730). Statistical analysis: Baseline comparisons between surgical and Diet groups used ANOVA with Tukey HSD or Kruskal–Wallis with Dunn’s post hoc as appropriate; power >90% to detect observed differences between RYGB and Diet at α=0.05. Primary analyses assessed within-person changes at each timepoint vs baseline (Wilcoxon signed-rank for nonparametric variables; paired t-test for thyroid parameters where applicable). Random intercept linear mixed-effects models (log-transformed hormone outcomes) evaluated the incremental contribution of weight loss (per kg) to hormone changes over 36 months. Covariates: age for all participants and menopausal status for women. Secondary analyses used mixed models to estimate differences in mean percentage changes over time between each surgery type and Diet (95% CI, p-values). Exploratory analyses assessed longitudinal changes in ACTH, cortisol, IGF-1 (Wilcoxon) and TSH, FT4, FT3 (paired t-test), plus mixed models for weight loss and procedure effects.

• Cohort: 61 participants (72% female), mean age 51 years (SD 11). Baseline groups were similar in sex hormones and pituitary axes; SG participants were heavier than Diet. 31% of women were premenopausal. - Weight change over 3 years (women; mean % vs baseline): RYGB −38% (95% CI: −48%, −29%), SG −22% (−30%, −14%), LAGB −17% (−30%, −4%), Diet −2% (−7%, 4%). Men: RYGB −21% (−29%, −14%), SG −22% (−27%, −17%), LAGB −16% (−28%, −4%), Diet +4% (−25%, 33%). - Incremental effect per kg lost over 36 months: • Females: Total testosterone decreased by 0.8% per kg (95% CI: −1.4%, −0.3%, p=0.003), remaining significant after adjustment for age and menopausal status. SHBG tended to increase (0.4% per kg; 95% CI: −0.1%, 0.9%); FSH and LH associations were not significant after adjustment. • Males: Total testosterone increased by 0.6% per kg (95% CI: 0.2%, 1.0%, p=0.002), adjusted 0.7% (0.3%, 1.1%); CFT increased by 0.5% per kg (95% CI: 0.0%, 0.9%, p≈0.04), adjusted 0.7% (0.3%, 1.1%). SHBG increased by 0.4% per kg (95% CI: 0.0%, 0.8%, p=0.033), not significant after age adjustment. - Absolute longitudinal changes: • Females: Median testosterone fell from 0.9 to 0.6 nmol/L by 6 months and remained lower; SHBG rose from 29.1 to 47.8 nmol/L at 6 months and remained elevated at 36 months. LH and FSH increased over time; estradiol unchanged. • Males: Median testosterone rose from 11.8 to 17.8 nmol/L at 36 months; SHBG rose from 20.1 to 41.0 nmol/L at 24 months; LH increased at 24 months; FSH and estradiol unchanged. - Procedure vs Diet (mean % differences over time to 3 years): • Women: RYGB associated with lower total testosterone by 54% (95% CI: −90%, −17%, p=0.004) and lower FAI by 65% (95% CI: −114%, −17%, p=0.009); SG associated with lower FAI by 39% (95% CI: −77%, 0%, p=0.05) and higher SHBG by 51% (95% CI: 27%, 76%, p=0.0001). Some LAGB-related differences in LH, FSH, estradiol vs Diet were attenuated after adjustment for age and weight loss. • Men: No significant differences in testosterone or SHBG between surgical groups and Diet. LH declined after LAGB by 51% vs Diet (95% CI: −90%, −13%, p=0.010) and remained significant after adjustment; FSH declined in SG by 39% (95% CI: −73%, −5%, p=0.024) but not after adjustment. - Other axes: Per kg weight loss, ACTH decreased by 0.3% (95% CI: 0.0%, −0.5%, p=0.05); IGF-1 increased by 0.4% (95% CI: 0.2%, 0.7%); FT4 and FT3 showed small declines (FT3 −0.1% per kg). Across time, ACTH decreased by 12 months, IGF-1 increased by 24 months, and TSH decreased by 36 months. Procedure effects vs Diet: RYGB reduced ACTH by 22% (95% CI: −42%, −1%, p=0.032); SG increased cortisol by 13% (95% CI: 0.1%, 25%, p=0.048) and reduced TSH by 54% (95% CI: −81%, −27%, p<0.001); LAGB increased IGF-1 by 26% (95% CI: 9%, 44%, p=0.003). Overall, androgen changes were proportional to weight loss, with sex-differential directions.

The study demonstrates that long-term weight loss is associated with sex-specific changes in androgen levels: decreasing total testosterone in women and increasing total and free testosterone in men. These findings support the hypothesis that weight loss modulates gonadal axes and that the magnitude of hormonal change is proportional to weight reduction. In women, the reduction in androgens was independent of age and menopausal status and was more pronounced following bariatric surgery (particularly RYGB for total testosterone and SG for FAI/SHBG), indicating that surgical weight loss may exert additional influences beyond weight reduction alone. Given that low androgen levels in older women have been linked to cognitive and functional outcomes, these results raise important clinical questions regarding long-term effects of androgen reduction after bariatric procedures. In men, weight loss restored testosterone and SHBG levels over time, consistent with decreased aromatization and improved metabolic milieu; however, no clear differences between surgical and dietary interventions were observed, likely reflecting limited male sample sizes. Beyond gonadal axes, weight loss and bariatric procedures differentially affected pituitary-adrenal, somatotropic, and thyroid axes: decreased ACTH (especially after RYGB), increased IGF-1 with weight loss (notably after LAGB), and reduced TSH (particularly after SG). These neuroendocrine adaptations may contribute to improved metabolic outcomes, stress responses, and body composition following weight loss interventions. Overall, the results underscore the relevance of neuroendocrine regulation in mediating the metabolic benefits of bariatric surgery and sustained weight loss.

This 3-year prospective study shows that weight loss elicits sex-differential androgen responses: decreasing total testosterone in women and increasing total and free testosterone in men, with effects proportional to the amount of weight lost. In women, androgen reductions were independent of age and menopausal status and were more pronounced after bariatric surgery (RYGB and SG), alongside increases in SHBG. Weight loss also modulated other pituitary axes, including reductions in ACTH and TSH and increases in IGF-1, with some effects differing by surgical procedure. These hormonal adaptations likely contribute to the metabolic benefits of bariatric surgery. Future research should investigate the clinical implications of reduced androgens in postmenopausal women (e.g., cognitive, functional, and sexual health), delineate mechanisms underlying procedure-specific neuroendocrine changes, and evaluate long-term outcomes across diverse patient subgroups with adequately powered randomized designs.

• Modest sample size, particularly within surgical subgroups and among male participants, limiting power to detect between-procedure differences. - Non-randomized allocation to interventions, risking residual confounding despite adjustments. - Women were not evaluated for PCOS or menstrual irregularities; clinical significance of androgen reductions within normal ranges is uncertain. - Lack of direct assessments of clinical outcomes (sexual function, mood, cognition, frailty) to relate hormonal changes to patient-centered endpoints. - Potential assay and modeling limitations inherent to observational longitudinal designs; some 36-month measurements were restricted to SG and RYGB groups. - Baseline TSH differences between groups may influence thyroid axis interpretations despite adjustments.