Elevated Luteinising Hormone despite Normal Testosterone Levels in older Men – Natural History , Risk Factors , and Clinical Features Short Running Title : Elevated LH in Ageing Men

Objective Elevated LH with normal testosterone (T) suggests compensated dysregulation of the gonadal axis. We describe the natural history, risk factors and clinical parameters associated with the development of high LH (HLH, LH>9.4 U/L) in ageing men with normal T (T≥10.5 nmol/L). Design, Patients and Measurements We conducted a 4.3 year prospective observational study of 3,369 community-dwelling European men aged 40-79 years. Participants were classified as: incident (i) HLH (n=101, 5.2%); persistent (p) HLH (n=128, 6.6%); reverted (r) HLH (n=46, 2.4%); or persistent normal LH (pNLH, n=1667, 85.8%). Potential predictors and changes in clinical features associated with iHLH and rHLH were analysed using regression models. Results Age >70 years (OR=4.12[2.07–8.20]), diabetes (OR=2.86[1.42–5.77]), chronic pain (OR=2.53[1.34–4.77]), pre-degree education (OR=1.79[1.01–3.20]) and low physical activity (PASE≤78, OR=2.37[1.24–4.50]) predicted development of HLH. Younger age (40-49 years, OR=8.14[1.35-49.13]) and non-smoking (OR=5.39[1.48–19.65]) predicted recovery from HLH. Men with iHLH developed erectile dysfunction, poor health, cardiovascular disease (CVD) and cancer more frequently than pNLH men. In pHLH men, co-morbidities, including A cc ep te d A rt ic le This article is protected by copyright. All rights reserved. CVD, developed more frequently, and cognitive and physical function deteriorated more, than in pNLH men. Men with HLH developed primary hypogonadism more frequently (OR=15.97[5.85–43.60]) than NLH men. Men with rHLH experienced a small rise in BMI. Conclusions Elevation of LH with normal T is predicted by multiple factors, reverts frequently and is not associated with unequivocal evidence of androgen deficiency. High LH is a biomarker for deteriorating health in aged men who tend to develop primary hypogonadism. Introduction Ageing is associated with myriad alterations in endocrine functions, some of which may contribute to a decline in health and quality of later life. Elevated LH with normal testosterone (T) is encountered commonly in ageing men and may present a diagnostic and therapeutic dilemma to the practicing clinician. High LH may signify the onset of dysregulation in the hypothalamic-pituitary-testicular (HPT) axis, whereby compensatory increase in luteinising hormone (LH) is required to mobilise the Leydig cell steroidogenic reserves to maintain normal T production. We have suggested that this may represent a state of ‘compensated’ hypogonadism, where high LH concentration in the face of normal T (HLH) marks an intermediate stage before transitioning towards overt primary hypogonadism (PHG) when testicular reserve eventually becomes exhausted. The rate of transition (progression and reversal) from normal LH (NLH) to HLH and the factors that influence its development in ageing men are important issues that are yet to be explored in prospective studies. Whether men with HLH experience symptoms, functional deficits or morbidity, without the development of a low T, is unclear. In particular, the proportion of men with HLH who exhibit evidence of androgen deficiency is currently unknown. These uncertainties pose important questions regarding the clinical significance of abnormally increased LH in ageing men. The European Male Ageing Study (EMAS) provides an opportunity to help rectify these knowledge gaps through longitudinal observations of unselected men from the general population. A cc ep te d A rt ic le This article is protected by copyright. All rights reserved. In this study, we aimed to: 1) identify factors predisposing normal LH men to develop high LH and to identify factors, which predict recovery from high LH; 2) determine the clinical features associated with high LH; and 3) assess whether men with high LH are at increased risk of developing PHG. Materials and Methods Participants and study design The study design and recruitment for EMAS have been described previously. Briefly, an age-stratified sample of 3,369 men aged 40–79 (mean±SD: 60±11) years was recruited from population registers in eight European centres. Participants completed a postal questionnaire and attended a clinic for further assessments. These men were invited to participate in the follow-up assessment a median of 4.3 years later (range: 3.0–5.7 years). During this period, 168 men died and 407 men were lost to follow-up. Ethical approval for the study was obtained in accordance with local requirements in each centre. All participants provided written, informed consent. They completed questionnaires both at baseline and at follow-up about smoking, alcohol consumption, and currently treated medical conditions. Anthropometric measurements, Reuben’s physical performance test (PPT), and psychomotor processing speed estimation (digit symbol substitution test [DSST]) were performed according to standardized methods. Physical, sexual and psychological function was assessed using responses to the Medical Outcome Study (MOS) 36-item Short-Form health survey (SF36), the EMAS Sexual Function Questionnaire and the Beck’s Depression Inventory. Hormone measurements A single fasting morning (before 1000) blood sample was obtained at baseline and at followup. T was measured by liquid chromatography–tandem mass spectrometry (LC–MS/MS), with paired baseline and follow-up samples analysed simultaneously. LH, FSH and SHBG were measured by the E170 platform electrochemiluminescence immunoassay (Roche A cc ep te d A rt ic le This article is protected by copyright. All rights reserved. Diagnostics). The Vermeulen formula was used to calculate free (f) T using participantspecific testosterone, SBHG and albumin values. Intraand inter-assay coefficients of variation (CV) were: T 4.0 and 5.6%; LH 1.9 and 2.7%; FSH 0.9 and 1.9%; and SHBG 1.9 and 3.2%, respectively. The detection limits for the sex hormones were: T (0.17 nmol/L; 0.05 ng/mL), LH (0.10 U/L), FSH (0.10 U/L) and SHBG (0.35 nmol/L). Insulin was assayed using chemiluminescence (CVs: 3.9% and 5%). Biochemistry/haematology Standardized measurements were undertaken in laboratories in each centre. Insulin resistance was calculated using the homeostasis model assessment of insulin resistance (HOMA-IR = fasting insulin [U/mL] x fasting glucose [mmol/L]/22.5). Categorisation of participants by LH criteria Following the criteria described previously, participants were categorised according to LH and T levels: normal LH (NLH, LH≤9.4 U/L and T≥10.5 nmol/L); high LH (HLH, LH>9.4 U/L and T≥10.5 nmol/L). A LH cut-off of 9.4 U/L was chosen because this was the 97.5 centile value for men aged 40 to 44 years in the EMAS population. Participants were further categorised by their change in LH status into one of four groups: incident HLH (iHLH) in which men had a normal LH at baseline and an high LH at follow-up; persistent HLH (pHLH) in which men had HLH at both baseline and follow-up; reverted HLH (rHLH) in which men had HLH at baseline and NLH at follow-up; and pNLH in which men had NLH at both baseline and follow-up. Statistical analysis Baseline and follow-up differences between pNLH, iHLH, pHLH and rHLH men in hormone levels, anthropometrics, biochemistry, sexual, physical and psychological function, and health and lifestyle measures were compared between groups using initially the independent-samples t-test or analyses of variance (ANOVA) for continuous variables and χ A cc ep te d A rt ic le This article is protected by copyright. All rights reserved. tests for categorical variables. Post-hoc analyses were performed using Tukey-Kramer tests to allow for multiple pairwise comparisons. Longitudinal within-group differences in parameters were compared using the paired t-test for continuous variables and the McNemar test for categorical variables. Multiple regression models, adjusted for centre as a random effect, were used to account for the hierarchical study design (individuals nested within centres). The relationships between LH status and putative predictors were assessed using multilevel binary logistic regression models, where gonadal status was the outcome, with the pNLH or pHLH group being the referent for the analyses of the potential predictors for iHLH or rHLH respectively. Nine factors were included as fixed effects: age (40-49 years, 50-59 years, 60-69 years or ≥70 years), diabetes, BMI (<25 kg/m, 25-29.9 kg/m or ≥30 kg/m), smoking status (current, ex or never), alcohol intake (≥5 days per week or <5 days per week), chronic pain, pre-degree education (education obtained to below a university degree level), partner status (living with a partner or not) and physical activity (PASE≤78 or >78). The factors chosen were not collinear – the variance inflation factors were less than 3.33. The relationships between gonadal status and clinical features of high LH were investigated using binary logistic regression models when associations between iHLH, pHLH or rHLH status with sexual, physical and psychological symptoms or medical conditions were assessed, with adjustments for age, comorbidities, BMI, centre, smoking, chronic pain, education and physical activity. Linear regression analysis was used when associations between iHLH, pHLH or rHLH status with functional ratings were assessed, with adjustments for age, comorbidities, BMI, centre, smoking, chronic pain, education and physical activity. Results from linear regression models are presented as β-coefficients (for standardized variables) with 95% confidence intervals (CI) and results from logistic regression models are presented as odds ratios (OR) with 95% CI. All statistical analyses were conducted using STATA SE version 13 (StataCorp, College Station, TX). A cc ep te d A rt ic le This article is protected by copyright. All rights reserved. Results Natural history of High LH Of the 3,369 men recruited into EMAS, 1,942 men made up the analytical sample after exclusion of those with known pituitary, testicular or adrenal disease (n=93)