Diagnosis and treatment of late-onset hypogonadism: Systematic review and meta-analysis of TRT outcomes

Late-onset hypogonadism (LOH) is a relatively common conditions affecting the aging male. The aim of this review is to summarize the available evidence regarding LOH and its interaction with general health. LOH is often comorbid to obesity and several chronic diseases. For this reason lifestyle modifications should be strongly encouraged in LOH subjects with obesity, type 2 diabetes mellitus (T2DM) and metabolic syndrome (MetS) and good treatment balance of chronic diseases. Medical therapy of LOH should be individualized depending on the etiology of the disease and the patient's expectations. Available evidence seems to suggest that testosterone replacement therapy is able to improve central obesity (subjects with MetS) and glycometabolic control (patients with MetS and T2DM), as well as to increase lean body mass (HIV, chronic obstructive pulmonary disease), along with insulin resistance (MetS) and peripheral oxygenation (chronic kidney diseases). However, it should be recognized that the number of studies on benefits of T supplementation is too limited to draw final conclusions. Longer and larger studies are needed to better clarify the role of TRT in such chronic conditions.

Introduction

Male hypogonadism is defined as the failure of the testes to produce testosterone (T), because of a distant (pituitary/hypothalamus) or a local (testicular) deficiency.1, 2 However, a hypogonadism-like syndrome can also be the result of an impaired action of the sex steroids, such as in the resistance to androgens, because of alterations in receptor sensitivity or in free hormone bioavailability. Based on these considerations, male hypogonadism can be divided into three main forms: primary hypogonadism (testis failure), secondary hypogonadism (hypothalamic/pituitary failure) or impaired sex steroid action (failure of biological action).1, 2 The clinical correlates of hypogonadism are indeed quite similar in these three conditions, regardless of the different sites of origin of the disease. However, these categories are strictly dependent on the period of life in which the defect becomes clinically overt. We therefore recently introduced a new classification of male hypogonadism essentially based on the age of onset of the hypogonadal symptoms and signs, thus distinguishing very early, early and late-onset hypogonadism.1, 2

When androgen deficiency manifests very early on – such as during the first to second trimesters of pregnancy – it results in disorders of male sexual differentiation, with a wide range of clinical presentations from ambiguous or female genitalia, to various defects in virilization (micropenis, hypospadias, cryptorchidism), owing to the lack – or to insufficient levels of – T secretion and/or its conversion to DHT during the critical window of male sex differentiation (Very Early Onset Hypogonadism, VEOH).

When hypogonadism manifests during infancy or childhood, it may remain undiagnosed until the age of puberty. At pubertal age, hypogonadism can provoke the complete lack, or the arrest, or just a delay, of the pubertal development of secondary sex characteristics (Early Onset Hypogonadism, EOH). A typical condition is the Klinefelter's syndrome, where, despite the congenital abnormality of additional X chromosome(s), the testicular failure starts to become apparent only during the pubertal process.³ In EOH, the clinical symptoms and signs can be: eunuchoid proportions, lack of deepening of voice, female distribution of sexual hair, anemia, underdeveloped muscles and genitalia, whereas spermatogenesis and sexual functions are not initiated or are impaired in their progression (delayed puberty).⁴

During adulthood and senescence, hypogonadism leads to infertility and/or impaired androgen-regulated physiological functions, with relatively vague symptoms and signs (fatigue, weakness, decreased libido and energy, erectile dysfunction, reduced muscle and bone mass, and increased abdominal fat). The latter form (late-onset hypogonadism, LOH) is a relatively common adverse conditions affecting the aging male, but also one of the most difficult to recognize and treat, since it does not show clear-cut differences in relation to the physiological aging process.*5, 6, *7, 8, 9 Hence, the diagnosis of LOH is problematic, because of the difficulty of knowing to what extent the clinical features are due to aging, to T deficiency or to both.*5, 6, *7 In the case of LOH, the site of origin is often unclear, due to a mixed contribution of testicular and hypothalamic/pituitary failure.

During the last 10 years, the major Andrological/Endocrine Societies have delivered several guidelines providing the clinicians with a puzzling combination of biochemical thresholds of total T (TT) plasma levels (from below 7 up to 12 nmol/L) and various symptoms and signs, however, always including sexual symptoms.*7, 8 There is also a general agreement that a total T level above 12 nmol/L (3.46 ng/mL or 346 ng/dL) does not require substitution. When total T is repetitively >8 and <12 nmol/L, in the presence of typical hypogonadal symptoms, a T treatment trial should be considered.*7, 8 Measurement of free T (FT) by equilibrium dialysis or calculated (cFT) according to Vermuelen formula (at http://www.issam.ch/freetesto.htm) might help in conditions when sex hormone binding globulin (SHBG) could be decreased (obesity, acromegaly, hypothyroidism) or increased (aging, hepatic illness, hyperthyroidism, use of anticonvulsants).

The association between LOH and adverse metabolic conditions, such as obesity, metabolic syndrome (MetS), and type 2 diabetes mellitus (T2DM), is well known,9, 10, 11, 12 with a rather complex, and often multi-directional, pathogenetic background. An obesity-induced estrogen increase seems to play a major role in determining the negative feedback at the pituitary level, therefore inducing hypogonadotropic hypogonadism with low, or inappropriately normal, gonadotropin levels.13, *14, 15 Also insulin resistance may contribute to the low T levels seen in obese men.13, *14, 15 Pathogenetic mechanisms linking obesity, other metabolic derangements and LOH have been described elsewhere.13, *14, 15

An evidence-based study on LOH has recently been published in a general population cohort of middle-aged and elderly men from 8 European countries (European Male Aging Study, EMAS, 16). The aim of that study was to identify the thresholds of T below which symptoms, assumed to be specific for androgen deficiency, become increasingly prevalent, and to define essential criteria for the LOH syndrome on the basis of the presence of symptoms associated with a low T level. They found a consistent syndromic clustering of three sexual symptoms (decreased morning erection and sexual thoughts and erectile dysfunction) with a low T level, suggesting that the presence of at least these three sexual symptoms with a total T level of less than 11 nmol/L and free T < 225 pmol/L be required in order to diagnose LOH. The association between the presence of three sexual symptoms and a low T level was attenuated, but still present, by adjustment for age, body-mass index, and the number of coexisting illnesses (see above).¹⁶

The association between psychological symptoms and hypogonadism is controversial.17, 18, 19, 20 Data from interventional studies are scanty, but a recent meta-analysis of randomized clinical trials (RCTs) showed a significant, positive effect of T replacement therapy (TRT) in depressed patients when compared with placebo.²¹ These observations suggest a direct role of T in modulating psychological health, and consequently, the appearance of psychological disturbances, such as depression, anxiety, irritability, insomnia, memory impairment or reduced cognitive function, must elicit in the physician the clinical suspicion of androgen deficiency. Although data from the same EMAS study demonstrated an increased prevalence of psychological symptoms, such as sadness, when cFT is lower than 0.160 mol/L, no association was found with total T (see ref. ¹⁶).

In conclusion, from the EMAS survey, the presence of at least three sexual symptoms along with a total T level of less than 11 nmol/L and a FT level of less than 0.225 nmol/L is considered as the minimum criteria for the diagnosis of LOH.¹⁶

In male subjects consulting for sexual dysfunction (SD), in contrast to other forms of male hypogonadism, LOH is a relatively common condition, as opposed the general population where its prevalence is only 2% (ref. ¹⁶, see below). However, in this particular population, the presence of sexual symptoms may not help in raising the suspicion of hypogonadism among clinicians, because of the lack of specificity,² as it may be just one of the main reasons for medical consultation. Fig. 1 (Panel A) shows the age-adjusted risk for having LOH, according to the previously mentioned EMAS definition, in a population consulting our Andrology Clinic for SD at the University of Florence (UNIFI). Since erectile dysfunction (ED) is the most common complaint, it is obvious that having any form of ED is not associated with an increased risk of hypogonadism. However, the most severe form of ED, as well as hypoactive sexual desire (HSD) and perceived reduction in spontaneous/nocturnal erections, double the risk of having LOH. Hence, even in a symptomatic population, sexual symptoms might help in recognizing LOH.2, 22

The Endocrine Society recommends the evaluation of total T in subjects with ED, hypothalamic/pituitary diseases and in subjects taking medications interfering with T release or action.²³ However, in the same report, screening for low T is also recommended in several metabolic conditions and clinical illnesses, such as T2DM, HIV and chronic obstructive lung disease.²³ Advantages and disadvantages of T supplementation in these conditions are systematically analyzed later. Fig. 1 (Panel B) reports the age-adjusted risk of having LOH (according to EMAS criteria) according to different metabolic derangements or chronic diseases in LOH symptomatic subjects consulting for SD. Chronic illnesses (here computed using the Chronic Disease Score algorithm, see ref. ²⁴), a consistent (>20%) predicted risk of developing cardiovascular (CV) diseases in the next 10 years (here computed using the Progetto Cuore algorithm, see ref. ²⁵) and metabolic conditions related to overweight and insulin resistance, such as obesity, T2DM and MetS, double or triple the chance of having LOH (Fig. 1, Panel B). MetS refers to a diagnostic category summarizing the co-existence of some of these conditions (impaired glucose tolerance, hypertension, visceral obesity, dyslipidaemia) that might help clinicians identify subjects at a greater risk for forthcoming adverse CV and/or metabolic diseases.*14, 15 In particular, all the components of MetS, apart from hypertension, are associated with an increased, age-adjusted, risk of LOH, an increased waist circumference being the most predictive sign. A detailed description of the link(s) between MetS and LOH is beyond the aim of the present review, because it has been extensively covered elsewhere.13, *14, 15

LOH in the general population is clearly an age-dependent condition, with an estimated prevalence that is quite variable, ranging from 7% to 40%, according to the different criteria employed to define it and to the population studied.⁷ By using the EMAS criteria, the prevalence of LOH in 40–80-year-old European men was much lower than previously reported, i.e. 2.1%.¹⁶ By using the strict EMAS criteria, the prevalence of LOH in Florence (one of the centers of the EMAS study) increased from being quite an unusual condition in the younger age bands (40–49 years old) to more than 3% in the oldest age group (>69 years old). Using the same diagnostic category, the prevalence of LOH in subjects seeking medical care at an Andrology unit for SD at UNIFI – hence from the same geographic area – is 5 times more prevalent than in the general population (15% in the oldest age band as opposed to 3% in the unselected general population). Data from the EMAS study showed an unadjusted annual age decline for total T of 0.04 nmol/L per year.⁹ In line with the EMAS results, a significant age-dependent decline of total T was observed in subjects with SD studied at UNIFI, although with a faster rate (0.09 nmol/L/year; see ref. ¹⁸).

There are generally two types of instruments for assessing LOH symptoms: self-report questionnaires (SRQs) and structured interviews (SIs) (for review, see refs. 2, *5). Although all these instruments cannot replace a detailed history and physical examination, they can be very useful for collecting systematic anamnestic data, representing a guide for a correct collection of symptoms in clinical practice. In addition, they can be used as screening tools for outpatients who need to be more carefully studied. To help in recognizing hypogonadal stigmata, four different SRQs have been developed (see for review refs. 2, *5). All these instruments explore sexual and psychological symptoms, along with physical symptoms and signs of hypogonadism, and have demonstrated a good sensitivity in cross-sectional surveys,²⁶ however, their relative specificity is variable.²⁶ In addition, they are designed for the screening of hypogonadism in the general population, and not specifically targeted toward individuals with sexual dysfunction. ANDROTEST is a structured interview, specifically designed for the screening of hypogonadism in subjects with SD.²⁷ ANDROTEST has a sensitivity and specificity close to 70% in detecting low total or free T.2, 27 In particular, ANDROTEST investigates sexual symptoms (ED, HSD and delayed ejaculation), as well as the presence of hypogonadism-related comorbidities. However, it does not take into account psychological symptoms. We recently described the ability of adding psychological symptoms to ANDROTEST to improve hypogonadism screening.²⁸ We used psychological items derived from a SRQ, Middlesex Hospital Questionnaire (MHQ), a tool designed for the scoring symptoms of mental disorders in a non-psychiatric setting.²⁹ We found that the algebraic addition of scores from two psychological items (exploring unduly fatigue and lack of self-centeredness) to scores derived from the original ANDROTEST increases the accuracy of the latter in predicting hypogonadism.²⁸ As ANDROTEST is a clinician-used and the MHQ is a self-reported tool, the two instruments must be run separately, the scores computed and only combined at the end. Although ANDROTEST, along with its integrated new version PsychoANDROTEST,²⁸ could help in the identification of patients who should undergo further evaluation for LOH, the determination of T remains mandatory to confirm the diagnosis of hypogonadism and/or to start any substitutive therapy according to accepted clinical guidelines.

Since their introduction in the late 1960s, immunoassay (IA) methods have been widely used for the measurement of sex steroids. The majority of IAs show good accuracy in the estimation of high (adult male) TT levels, however they usually overestimate low TT concentration, reducing the specificity and sensitivity for diagnosing hypogonadism.*5, *7, 8, 23 Furthermore, there is a general agreement that IAs produce unreliable results during the evaluation of free T.*5, *7, 8, 23 Mass spectrometry has been considered the gold standard of steroid analysis, however, their complexity and cost still represent a major limitation for widespread use.*7, 8, 23 Recently the EMAS group compared different measurement methods for TT, i.e. a common validated platform IA (Modular E170 electrochemiluminescence Roche Diagnostics) and a gas chromatography mass spectrometry (MS) method.³⁰ Although the two methods show an overall high correlation (r = 0.93, p < 0.001), the IA was less robust in the hypogonadal range (<11 nmol/L r = 0.72; p < 0.001). Nonetheless, the conclusion was that validated IA platforms can be as good as MS in the clinically important discrimination between eugonadal and hypogonadal men, especially when combined with clinical signs of androgen deficiency.³⁰ In line with this view, the Position statement of the Endocrine Society recognizes that most indirect assays can detect low TT concentrations even in the hypogonadal range if such assays are well validated.³¹ Hence, serum TT, preferably obtained with more than one morning sampling, is the recommended screening test for hypogonadism.*7, 8, 23, 31 Since SHBG concentrations can be influenced by many factors (see above), the calculation of free T with the Vermeulen formula should be preferred in clinical practice, because it is the most easily available (an easy-to-use free calculator of calculated FT can be found on http://www.issam.ch) and one of the only two equations that has been validated by comparison with the equilibrium dialysis assay.

We can essentially recognize two distinct levels of intervention options for LOH: lifestyle modification, essentially based on increasing physical activity and weight loss, and medical therapy. The latter offers several approaches, essentially based on the nature of the hypogonadism and on the patient's requirement of potential fertility in the near future, as previously described.1, *5 The nature of the hypogonadism should dictate the most appropriate medical treatment.¹ Removal and/or proper treatment of LOH-associated comorbidities is essential in treating LOH, however, its efficacy seldom has been tested systematically, and the majority of studies are dedicated to reverting or reducing deleterious effects of visceral obesity on male reproductive axis activity.