Thyroid disease can be broadly categorised as thyroid dysfunction (hypothyroidism, hyperthyroidism) and structural disease (goitre, nodules and cancer). Management is often straightforward, but there are pitfalls that may lead to misdiagnosis, overdiagnosis and inappropriate treatment. This article reviews the approach to common thyroid problems in general practice.

Epidemiology

Worldwide, iodine deficiency is the most common cause of thyroid disease. Iodine deficiency has long been known in Tasmania, and a survey published in 2006 found evidence of iodine deficiency in mainland Australia, particularly New South Wales and Victoria.1 In 2009, use of iodised salt in bread became mandatory in Australia and New Zealand (following a Tasmanian program started in 2001), and Australia is now considered iodine sufficient.2

Autoimmune thyroid disease is the commonest cause of thyroid dysfunction in Australia, with prevalence as shown in Box 1. Some 10–15% of the population have positive thyroid antibodies, most commonly to thyroid peroxidase (TPOAb), with a higher prevalence in women than men.3 Patients who are euthyroid with positive thyroid antibodies do not require treatment, but are at increased risk of thyroid dysfunction (particularly hypothyroidism),4,5 and should be followed with annual measurement of serum thyroid-stimulating hormone (TSH).

TSH testing is the most sensitive means of detecting thyroid dysfunction. Debate surrounding the laboratory reference interval for TSH has largely resolved, and a range of about 0.4–4.0 mU/L is generally accepted.6,7 TSH concentrations increase with age,7,8 and some laboratories have adopted age-related reference intervals, with upper limits of up to 7 mU/L in older patients.

Hypothyroidism

Diagnosis

Classic symptoms of hypothyroidism include fatigue, weight gain, cold intolerance, arthralgia, constipation, menorrhagia, and dry skin and hair. Physical signs include pallor, coarse skin and hair, bradycardia and goitre, but may be absent in mild hypothyroidism. These symptoms and signs are non-specific and common in people without thyroid disease,9 so laboratory diagnosis is required. Serum TSH should be measured; if this is in the reference range, then additional tests such as free thyroxine (T4), free triiodothyronine (T3) or thyroid antibodies are rarely helpful. Tests such as basal metabolic rate and reverse free T3 have no diagnostic value.

Overt hypothyroidism (high TSH, low free T4) is usually symptomatic, readily diagnosed and can be treated without further investigation. A more common presentation in general practice is an elevated level of serum TSH with normal free T4. This may indicate subclinical hypothyroidism caused by autoimmune thyroid disease, but can arise from non-thyroidal, systemic illness, particularly in the recovery phase. In patients with a mildly elevated TSH (up to 10 mU/L), TSH normalises without treatment in over 50% of cases,10 so treatment need not be offered immediately. Instead, serum TSH testing should be repeated 6–8 weeks later, together with free T4 and TPOAb, and treatment offered if the abnormality persists. Thyroid imaging (including ultrasound) is not indicated in the investigation of hypothyroidism.11

The importance of detecting and treating subclinical hypothyroidism remains uncertain.12,13 Randomised controlled trials of thyroxine have shown inconsistent results on symptoms and quality of life, particularly in older patients and when TSH levels are mildly elevated (up to 10 mU/L). Untreated subclinical hypothyroidism with TSH levels above 10 mU/L is associated with increased risks of cardiovascular disease.14,15 The benefits of treating subclinical hypothyroidism are probably greater in younger and middle-aged patients than in older people.13,16

Treatment

Thyroid hormone replacement is indicated for overt hypothyroidism and for subclinical hypothyroidism with TSH levels above 10 mU/L. Patients with persistent mild subclinical hypothyroidism (TSH, 4–10 mU/L) and minimal or no symptoms can be offered a choice between thyroxine treatment or observation with annual follow-up testing to detect progressive hypothyroidism. Progression is more likely in TPOAb-positive patients.4 Women with subclinical hypothyroidism who are planning pregnancy should be treated (see below). When it is uncertain whether non-specific symptoms are caused by, or merely coexist with, mild subclinical hypothyroidism, a 3-month trial of thyroxine is reasonable to assess symptomatic benefit. Thyroid replacement therapy is not indicated for individuals with symptoms suggestive of hypothyroidism if TSH levels are within the reference interval.17

Thyroxine is the standard treatment for hypothyroidism.18 The usual approach is an initial dose of 50–100 μg/day with subsequent titration based on thyroid function tests checked 6–8 weeks later. Smaller initial doses (25 μg/day) should be used in very frail or elderly patients and in those with symptomatic ischaemic heart disease. Ideally, thyroxine should be taken in a fasting state, 1 hour before breakfast, but this may be inconvenient and reduce adherence, and it is probably more important that daily dosing is consistent with regard to time of day and relationship to meals.18 The long half-life of thyroxine means that if a dose has been missed, a catch-up dose can be taken later in the day or the following day.

When treating hypothyroidism, the targets are relief of symptoms and return of TSH to within the reference interval. In thyroxine-treated patients, serum free T4 may be within the reference range or elevated; the latter is not an indication to reduce dosage if TSH is within the reference interval. Measurement of free T3 is unhelpful in monitoring thyroxine replacement.11,18 In some patients, TSH levels remain elevated despite apparently adequate thyroxine dosage. The most common cause is non-adherence; other causes and ways to address these are shown in Box 2.

Three thyroxine preparations are available in Australia. Two of these (Eutroxsig and Oroxine, Aspen Pharma) are identical and interchangeable. A third preparation, Eltroxin (Aspen Pharma), has recently been marketed. It has a different formulation, a wider range of tablet strengths and (unlike Eutroxsig/Oroxine) does not require refrigeration, so may be more convenient. Eltroxin product information states that it is not bioequivalent to Oroxine/Eutroxsig, but this is based on a study using single, large doses of thyroxine in healthy volunteers, which may not predict clinically relevant differences during clinical use.18 Because of this uncertainty, patients should not be switched between Eltroxin and Eutroxsig/Oroxine (or vice versa), except where explicitly intended by the prescribing doctor. Patients who do switch brands should have their serum TSH checked 6 weeks later, and dosage adjusted if necessary.

Dissatisfaction with thyroxine replacement

In some patients, symptoms of ill health persist despite adherence to treatment and normalisation of TSH.18,19 There are three likely explanations for this. First, the persistent symptoms may be unrelated to thyroid dysfunction. For example, a patient may present with fatigue and be found to have mild subclinical hypothyroidism, but without a causal relationship between the two. In such cases, comorbidities including coeliac disease (which is associated with autoimmune thyroid disease) and depression should be sought to account for the symptoms, but often, none can be identified.

Second, standard thyroid replacement therapy may in some way be suboptimal for some patients. Anecdotally, some patients feel better if thyroxine dosage is increased until serum TSH levels are in the lower part of the reference interval (0.4–2.0 mU/L) or below the interval (0.1–0.4 mU/L), although this has not been confirmed in clinical trials.20 This approach appears to be safe, particularly in younger, otherwise healthy patients, so long as TSH is not suppressed below 0.1 mU/L.21 The healthy thyroid secretes small amounts of T3 as well as T4, and randomised controlled trials have explored whether the addition of T3 to thyroxine treatment results in symptomatic benefit. Overall, no convincing superiority of combined thyroxine/T3 treatment has been shown,18,22,23 but it remains possible that there is a subgroup of patients who respond symptomatically. Dessicated thyroid extract (from porcine thyroid) is sometimes prescribed for hypothyroidism. In a clinical trial, its effects on hypothyroid symptoms and neurocognitive function were equivalent to those of thyroxine, but resulted in modest weight loss compared with thyroxine (1.3 kg over 4 months) and was preferred by some participants.24 It is not approved by the Therapeutic Goods Administration or listed on the Pharmaceutical Benefits Scheme, and is not considered standard therapy.

Third, autoimmune thyroiditis is an inflammatory disorder associated with increased cytokine production, and it is possible that this causes symptoms of ill health, independent of thyroid dysfunction. In some studies, selenium supplementation (100–200 μg/day, equivalent to ingestion of two to four Brazil nuts per day) reduces inflammatory markers and improves quality of life in patients with Hashimoto’s disease.25 Although not an established treatment, this can be considered for patients with persistent symptoms.

Central (secondary) hypothyroidism

In patients with pituitary disease, serum TSH is unreliable in diagnosing central hypothyroidism and in monitoring thyroxine replacement. Monitoring is based on free T4 measurements and clinical assessment,18 and should include specialist input.

Hyperthyroidism

Overt hyperthyroidism

Hyperthyroidism is less common than hypothyroidism. The clinical picture is often characteristic, with symptoms including weight loss, heat intolerance, palpitations, breathlessness, anxiety, diarrhoea, tremor and proximal muscle weakness. Physical signs include tremor, tachycardia, ophthalmopathy, goitre and difficulty rising from a squatting position. The diagnosis is confirmed by thyroid function tests showing suppressed TSH (usually undetectable) with elevated free T4 and/or free T3.26

Hyperthyroidism is most commonly caused by Graves’ disease, thyroiditis or toxic nodular goitre. It is important to establish the cause of hyperthyroidism before starting treatment: Box 3 shows the key clinical features and diagnostic tests. Positive TSH-receptor antibodies (TRAb) establish a diagnosis of Graves’ disease. Radionuclide thyroid scanning is often helpful (Box 4). Thyroid ultrasound is not routinely indicated; it does not distinguish reliably between Graves’ disease and thyroiditis, and frequently identifies nodules that are unrelated to the hyperthyroidism, resulting in diagnostic confusion and further unnecessary investigation.11,26

Thyroiditis characteristically has a triphasic course of hyperthyroidism followed by hypothyroidism, resolving to euthyroidism. In subacute (viral) thyroiditis, full recovery is the rule, whereas in autoimmune thyroiditis (with positive TPOAb), hypothyroidism may persist. Often no treatment is required during the thyrotoxic phase, although β-blockers may be helpful symptomatically. TSH, free T4 and free T3 should be checked every 6–8 weeks until resolution. If hypothyroidism persists, then thyroxine treatment is indicated.

Graves’ disease should be treated initially with carbimazole (15–20 mg daily for mild to moderate hyperthyroidism, 30–40 mg for severe hyperthyroidism). In patients who respond well, it can be continued for an 18-month course, aiming for long term remission. Other treatment options are radioactive iodine treatment and thyroidectomy.

Toxic nodular goitre can be treated with surgery or radioactive iodine. Antithyroid drugs can be used, but need to be continued lifelong as remission of hyperthyroidism is unlikely, and are not the preferred option.

Hyperthyroid patients with Graves’ disease, toxic nodular goitre and those in whom the diagnosis is unclear should generally be referred.

Subclinical hyperthyroidism

Mild subclinical hyperthyroidism, with TSH levels between 0.1 and 0.4 mU/L, may be caused by autonomous thyroid nodules but may also be found in healthy individuals (healthy outliers). It often resolves without treatment,10 so follow-up with repeat testing may be all that is required. Subclinical hyperthyroidism with TSH levels persistently below 0.1 mU/L is classified as mild hyperthyroidism and should be managed as above.

Thyroid nodules and cancer

Palpable thyroid nodules

Palpable thyroid nodules are present in about 5% of the population.27,28 Most are benign, commonly colloid nodules, cysts, nodular thyroiditis or benign neoplasm, whereas about 5% are malignant. Large nodular goitres can be symptomatic and require surgery for relief of pressure symptoms, but most thyroid nodules are asymptomatic, and the diagnostic work-up is aimed at assessing the risk of thyroid cancer. The diagnostic approach to palpable thyroid nodules is shown in Box 5. TSH levels should be measured but are usually normal, and the key investigation is ultrasound-guided fine needle aspiration (FNA) biopsy. Where clinical assessment, sonographic features and cytology are all consistent with a benign pathology, no further assessment is required. When cytology is suspicious of cancer or indeterminate, or if clinical suspicion persists, referral to an endocrine surgeon or head and neck surgeon is indicated. If there is uncertainty regarding the need for surgery, an opinion from an endocrinologist may be helpful.

When TSH is suppressed, radionuclide scanning may detect one or more autonomous (“hot”) nodules. These are rarely malignant, and do not routinely require biopsy. When TSH is normal or raised, radionuclide scanning is not indicated.

Thyroid nodules and cancer: the risk of overdiagnosis

Although 5% of people have a palpable thyroid nodule, the prevalence of nodules detectable by ultrasound is much higher, and increases with age up to 70% in the elderly.27 Clinically diagnosed thyroid cancer is uncommon, with a lifetime risk of less than 1%. In autopsy studies, however, small thyroid cancers are present in up to 36% of individuals.29 Most of these are small papillary cancers < 1 cm in size.

In Australia and many other countries, the incidence of thyroid cancer is increasing dramatically.30–32 The increase is largely accounted for by small papillary cancers found using ultrasound, most of which would never have become clinically apparent, an example of overdiagnosis.33 The most extreme example is in Korea, where screening with thyroid ultrasound has resulted in a 15-fold increase in thyroid cancer incidence, with no reduction in mortality (which is very low).34

Australia can avoid an epidemic of thyroid cancer by judicious use of thyroid ultrasound as follows:

• Thyroid ultrasound should be performed to assess clinically detected, visible or palpable thyroid nodules or goitre. It is not indicated for hypothyroidism or hyperthyroidism in the absence of goitre; nor is it indicated for globus, non-specific symptoms, or for screening.

• FNA biopsy should be considered for nodules > 1 cm in size, based on sonographic appearances. Nodules < 1 cm in size should not be routinely biopsied.27,33

Thyroid disease in pregnancy and postpartum

Pregnancy requires a 30–50% increase in thyroid hormone secretion because of stimulatory effects of chorionic gonadotropin (hCG) on the thyroid, increased circulating levels of thyroxine-binding globulin and degradation of thyroid hormone by the placenta.35 Maternal T4 crosses the placenta and is important for fetal brain development until 18–20 weeks’ gestation, when the fetal thyroid is fully functional. Mild iodine deficiency during pregnancy may impair fetal brain development,36 and dietary sources of iodine may not be sufficient for increased requirements during pregnancy.37 Iodine supplementation (150 μg/day) is therefore recommended for women who are pregnant or trying to conceive.

It remains controversial whether universal screening of pregnant women for thyroid dysfunction is indicated.38,39 The main value of screening is probably the detection of rare cases of overt hypothyroidism and hyperthyroidism for which treatment is clearly indicated, rather than minor abnormalities of uncertain significance (which are more common).35 Where screening is performed, TSH should be measured during the first trimester.

General laboratory reference intervals for TSH and free T4 do not apply to pregnancy. American Thyroid Association guidelines recommend that laboratories should develop trimester- and method-specific reference ranges from local populations. Instead, many laboratories have simply adopted suggested TSH reference intervals from the guidelines as follows: first trimester, 0.1–2.5 mU/L; second trimester, 0.2–3.0 mU/L; and third trimester, 0.3–3.0 mU/L.38 These intervals were originally based on expert opinion and limited data. Subsequent reference interval studies from different countries have produced widely varying results; in most, the upper limit for TSH is higher than 2.5 or 3.0 mU/L,40–43 suggesting that these cut-offs may not be appropriate.

Hypothyroidism

Overt hypothyroidism during pregnancy is associated with adverse outcomes, including miscarriage, pre-eclampsia, placental abruption, preterm birth, low birth weight and reduced IQ in offspring.38,39 The impact of subclinical hypothyroidism during pregnancy is uncertain. Although some observational studies have shown adverse outcomes,44 the data are inconsistent and some studies have found no association.45,46

There are few clinical trials of thyroxine in pregnancy. In one study, thyroxine treatment of TPOAb-positive, euthyroid pregnant women resulted in fewer miscarriages and preterm births.47 More recently, thyroxine treatment of pregnant women with elevated TSH or reduced free T4 concentrations at a mean gestation of 12 weeks had no effect on obstetric outcomes or on cognitive function in the offspring.48 Other trials are in progress.

Until better data are available, the following approach is recommended. Women with pre-existing hypothyroidism who are planning a pregnancy should have their thyroxine dose optimised, aiming for serum TSH in the lower reference range (0.4–2.5 mU/L). When pregnancy is confirmed, thyroxine dosage should be increased by about 30%, conveniently achieved by doubling the thyroxine dose on 2 days per week (with no change on other days). Serum TSH should be checked every 4–6 weeks until 20 weeks’ gestation, then once in the third trimester. After delivery, thyroxine dosage can be reduced to pre-pregnancy levels and TSH levels checked at 6–8 weeks postpartum.38,39

In women found to have elevated serum TSH during pregnancy, free T4 and TPOAb should be measured. If TSH is > 4 mU/L or free T4 is reduced, thyroxine treatment should be started, at an initial dose of 75–100 μg/day.40

It is uncertain whether pregnant women with TSH values in the 2.5–4.0 mU/L range benefit from thyroxine treatment. American Endocrine Society guidelines recommend thyroxine treatment for all such women,39 but this is increasingly challenged as overly simplistic, and is likely to result in overdiagnosis of subclinical hypothyroidism and unnecessary treatment.41–43,49 American Thyroid Association guidelines recommend treating pregnant women with TSH levels > 2.5 mU/L if they are TPOAb-positive.38

Hyperthyroidism in pregnancy

Overt hyperthyroidism during pregnancy is uncommon. Gestational hyperthyroidism can occur in the first trimester because of the stimulatory effect of hCG on the thyroid, mediated by the TSH receptor in women with very high hCG levels, particularly in hyperemesis gravidarum or multiple pregnancy. It usually resolves rapidly without treatment. Persistent hyperthyroidism in pregnancy is usually caused by Graves’ disease. Radionuclide scanning is contraindicated in pregnancy, and the key diagnostic test is measurement of TRAb. Hyperthyroidism increases the risk of pregnancy loss and other adverse outcomes, and patients should be referred urgently.

Subclinical hyperthyroidism can be a normal variant in pregnancy, or can be caused by thyroid disease. It is not associated with adverse outcomes and may not require treatment,50 but should be monitored closely.

Postpartum thyroid dysfunction

Thyroid dysfunction occurs in up to 10% of women in the first year postpartum. Hyperthyroidism may be caused by postpartum (autoimmune) thyroiditis or Graves’ disease, which can be distinguished by measurement of serum TRAb or radionuclide scanning (although scanning is contraindicated in breastfeeding women). Postpartum thyroiditis can be managed as above; it results in long term hypothyroidism in 10–20% of affected women.38,39

Box 4 –
Radionuclide scanning appearances of the thyroid in hyperthyroidism

---

A: Graves’ disease, with diffusely increased tracer uptake. B: Thyroiditis, with absent thyroidal uptake of tracer. C: Solitary autonomous nodule with focal tracer uptake in left lobe and reduced uptake in right lobe. D: Toxic multinodular goitre, with multiple areas of increased and reduced uptake.

Box 5 –
Algorithm for investigation of a palpable thyroid nodule

---

FNA = fine needle aspiration. T3 = triiodothyronine. T4 = thyroxine. TSH = thyroid-stimulating hormone.

The Endocrine Society of Australia formulated guidelines for testosterone prescribing in 2000, aiming to restrict inappropriate usage.1 Since then, prescriptions of testosterone have risen dramatically in Australia and elsewhere without any new proven indications, consistent with its use extending beyond the treatment of men with pathological hypogonadism due to pituitary or testicular disease.2–4 Controversy has arisen over the role of testosterone treatment in older men with medical comorbidities who have low levels of circulating testosterone, in the absence of hypothalamic, pituitary or testicular disease.5 There are gaps in the evidence base in relation to the potential benefits of testosterone treatment in men with obesity or type 2 diabetes and those receiving long term glucocorticoid or opioid therapy, who may exhibit low levels of circulating testosterone. There is also ongoing debate about the risk of cardiovascular adverse events related to testosterone treatment.6 In view of the rising rates of testosterone prescription, in 2015 the Australian Government tightened the criteria for which testosterone therapy would be subsidised in the absence of pathological hypogonadism.7 In this context, the Endocrine Society of Australia commissioned a position statement to update its 2000 guidelines and to inform the recommended management of men with androgen deficiency.

Methods

The Council of the Endocrine Society of Australia invited the authors of the 2000 guidelines and new authors with recognised expertise in this field to participate in a working group in 2014. A distinguished endocrinologist (H G B) was appointed to chair the working group. Extensive communication within the working group took place by email before and after a face-to-face meeting in Adelaide on 26 August 2015. All competing interests of participating authors were declared. Members of the group were asked to identify, consider and cite relevant evidence, sourced from their personal knowledge and searches of the literature, and to consider previous published guidelines. Controversies were resolved by discussion within the group. The draft statement was submitted to the Council of the Endocrine Society of Australia, who provided feedback. The working group responded to the feedback, and the final version of the position statement was approved and submitted for publication in April 2016. This article forms Part 1 of the position statement, focusing on assessment of male hypogonadism, including the indications for testosterone therapy. Part 2 will deal with treatment and therapeutic considerations.8

Recommendations

Pathological hypogonadism: clinical diagnosis

Hypogonadism refers to a pathological disorder of the hypothalamic–pituitary–testicular (HPT) axis that results in the testes being unable to produce both physiological levels of testosterone (androgen deficiency) and adequate numbers of functional sperm for paternity (male infertility).9,10 Thus, in its complete form, hypogonadism affects both virility and fertility. In many infertile men, virility may be relatively preserved, as the germinal epithelium is more vulnerable to both congenital and acquired defects.

Most often, hypogonadism is due to congenital or acquired defects of the testes (primary testicular failure), in which case levels of gonadotropins (luteinising hormone [LH] and follicle-stimulating hormone [FSH]) are elevated (hypergonadotropic hypogonadism). A failure of hypothalamic–pituitary stimulation with onset before or after puberty (hypogonadotropic hypogonadism) is less common.11 Rarely, there can be combined defects at both levels, presenting special diagnostic and therapeutic challenges (Box 1).

Although inaccurate, the term “hypogonadism” is often used interchangeably with “androgen deficiency”, as most often the problems concern whether testosterone treatment is required or justified for androgen deficiency. Accordingly, this position statement focuses primarily on the management of androgen deficiency, while fertility is mentioned only when relevant.

The diagnosis of androgen deficiency requires an initial careful clinical assessment to determine whether a compatible presenting symptom complex exists (Box 2). A comprehensive medical history, including information on smoking and the use of medication, alcohol and other recreational drugs, in particular androgens, is required. In addition, a focused review of the reproductive system should include any developmental history of undescended testes or other genital abnormalities, pubertal development, prior fertility, erectile function, sexual desire and any history of pelvic surgery, genital trauma or infection. The physical examination must include height and weight (and, if obese, waist circumference), a check for gynaecomastia, the adequacy of age-appropriate virilisation and especially scrotal palpation, using an orchidometer to assess testicular volume (usually 15–35 mL in men aged 21–35 years with normal reproductive function12).

Initial hormonal assessment

Having identified the possibility of pathologically based androgen deficiency on clinical grounds, laboratory testing is undertaken. Measurement of serum testosterone levels is not otherwise warranted (eg, for population screening). A serum total testosterone level, together with serum LH and FSH levels, should be measured early in the morning, between 08:00 and 10:00, as testosterone levels may be lower during the remainder of the day.13 Fasting results in higher serum testosterone levels.14 This time frame should be adjusted for shiftworkers or men who are sleep-deprived for any reason, in whom measurement shortly after waking may be preferred.15

Accurate testosterone assays are required. As most laboratories offer immunoassays that exhibit non-specificity and method-dependent bias, mass spectrometry is preferred where available.16 Serum testosterone level peaks at about the age of 20 years17 and gradually declines thereafter as age-related comorbidities accumulate.18 In men aged 21–35 years with normal reproductive function (ie, proven normal testes and semen analysis), the reference interval for total testosterone measured using mass spectrometry is 10.4–30.1 nmol/L.12 The reference interval using mass spectrometry in unselected young men is 7.4–28.0 nmol/L.19 In very healthy men aged 70–89 years, the reference interval using mass spectrometry is 6.4–25.7 nmol/L.20 The differences between the reference intervals for younger and older men may represent unresolved confounding by underlying reproductive and other health disorders that accumulate with age. At high and low ends of the reference interval, serum testosterone level must be interpreted in relation to serum sex hormone-binding globulin (SHBG) level. Furthermore, reference intervals of testosterone immunoassays vary between methods, as these direct immunoassays do not correspond accurately to mass spectrometry-based measurements. Even lower thresholds defined by mass spectrometry are not 100% sensitive or specific for the detection of androgen deficiency due to pathological hypogonadism. Clearly normal serum testosterone and LH levels (using clinical judgement to assess assay results for individual men) exclude androgen deficiency.

Confirmatory blood testing

An initial low serum testosterone level, especially when accompanied by a serum LH level that is within the reference interval, should be confirmed, as up to 30% of men (of any age) who have an isolated initial low serum testosterone level will have a normal serum testosterone level on repeat measurement.21,22 Unless one measurement of serum testosterone, LH and FSH levels is clearly diagnostic of pathological hypogonadism, at least two measurements of serum testosterone, LH and FSH are recommended to diagnose androgen deficiency in the appropriate clinical and pathological context. SHBG measurement can be helpful, as certain conditions are associated with marked changes in SHBG level that in turn affect total testosterone levels and their interpretation. Mildly elevated SHBG levels are seen with ageing, but SHBG can be markedly increased by hyperthyroidism, liver disease and anti-epileptic therapies; and accompanied by elevated total testosterone levels without evidence of androgen excess. Conversely, SHBG levels are suppressed with obesity, insulin resistance and exposure to exogenous androgens; in which case, serum testosterone concentrations below the reference interval, especially with normal serum LH and FSH levels, do not confirm a diagnosis of androgen deficiency. There is no evidence that free or bioavailable testosterone levels, which are usually not directly measured but calculated with various formulae (lacking validation in some cases),23 are a better measure of androgen status than total testosterone level. Moreover, reference intervals for these are even less well defined than those for measured testosterone concentrations. Therefore, they are not recommended for clinical decision making.

Additional investigation of hypogonadal men

A further round of investigations is recommended for men with suspected hypogonadotropic hypogonadism, to identify uncommon but treatable underlying disorders. These include:

• serum prolactin test (for prolactinoma and macroadenoma with pituitary stalk compression)

• iron studies and full blood count (for haemochromatosis and thalassaemia)

• hypothalamic–pituitary magnetic resonance imaging, if there is a clinical or biochemical suspicion of pituitary or hypothalamic disease

• anterior pituitary function test (for hypopituitarism and/or hyperfunctioning adenoma).

For men with suspected hypergonadotropic hypogonadism (primary testicular failure), additional evaluation is warranted in situations where knowing the aetiology will inform clinical management for fertility through in vitro fertilisation procedures:

• karyotyping (for suspected Klinefelter syndrome)

• Y chromosome microdeletion analysis.

Management decisions should only be made after a systematic approach to the diagnosis of hypogonadism, seeking to distinguish between pathological and functional causes of a low serum testosterone level. The features of pathological androgen deficiency depend on the age of onset and its severity and duration. Pre-pubertal androgen deficiency may manifest as micropenis and testicular maldescent (although not all cases of micropenis or testicular maldescent are due to hypogonadism) and, later, with delayed puberty and excessive long bone growth (eunuchoidal proportions). Post-pubertal onset produces typical yet often non-specific features that vary depending on the rate and extent of the fall in testosterone levels (Box 2). Conversely, the positive predictive value of non-specific symptoms for androgen deficiency can be low even in conjunction with a reduced serum testosterone level; hence the necessity for a careful clinical evaluation to look for an underlying pathological cause. In some cases, the diagnosis is easily appreciated, but, in others, limited or subtle features or an insidious onset leave even profound androgen deficiency overlooked. A striking example is the failure to diagnose Klinefelter syndrome in more than half of all men with the condition, which is the commonest chromosomal disorder in males (about one in 580) and a cause of adult androgen deficiency. Yet, invariably, a clinical examination reveals the characteristically small testes, and most affected men benefit from lifelong testosterone replacement therapy (TRT).24 The cause of this diagnostic failure is that many men go through life without a medical genital examination, where the clinical findings would promptly suggest the diagnosis.

Testosterone replacement in men with pathological hypogonadism

Indications and non-indications for testosterone therapy are shown in Box 3. The justification for testosterone treatment is to restore physiological androgen status to that comparable with eugonadal men. This will relieve the symptoms and signs of androgen deficiency. In this context, TRT refers to replacing the deficient hormone to restore physiological levels. The case for TRT is well established for androgen-deficient men with proven hypothalamic, pituitary or testicular disease, comparable with hormone replacement in other established deficiency states, such as adrenal failure (Addison disease) or hypothyroidism. TRT aims to establish or maintain secondary sexual characteristics, sexual function, body composition (including bone density) and wellbeing. In certain cases, such as late-diagnosed Klinefelter syndrome, symptoms may be subclinical, as men may be accustomed to their hypogonadism, yet will nevertheless note marked improvement in their health with TRT.

TRT should never be commenced before diagnostic work-up is complete and a clear pathological diagnosis has been made, for several reasons. First, failure to do so may overlook underlying abnormalities, such as a pituitary tumour. Second, exogenous testosterone suppresses the HPT axis so that once testosterone treatment has commenced, accurate evaluation for an underlying aetiology is problematic. An extended period of monitoring after testosterone is withdrawn may be needed to determine whether adequate endogenous testosterone production is present. Third, TRT for pathological hypogonadism is expected to be lifelong. Fourth, testosterone treatment in men without pathological hypogonadism compromises fertility, whereas in men with secondary hypogonadism, spermatogenesis can be restored with gonadotropin treatment.25 Contraindications to TRT are shown in Box 4. Age per se is neither an indication for testosterone treatment nor a contraindication to treating men who are androgen deficient due to pathological hypogonadism. In very old men with other comorbidities, individual consideration of the benefits versus risks of treatment is warranted. Ongoing testosterone replacement in older men with pathological hypogonadism is generally no more risky than in younger men, except for the occasional need to initiate treatment slowly and cautiously.

Hypogonadotropic hypogonadism in the presence of a proven aetiology (eg, pituitary tumour or its treatment by surgery or radiotherapy) usually requires TRT and treatment of coexisting anterior pituitary hormone deficiencies. Specific dopamine agonist therapy is required for hyperprolactinaemia, to not only suppress serum prolactin and restore testosterone but also to reduce tumour size and progression, as well as compressive symptoms. Iron chelation therapy may be required for congenital or acquired haemochromatosis.

Testosterone not justified in older men with chronic disease in the absence of pathological hypogonadism

Declines in serum testosterone level with age, in the order of 0.5–2% per year in reportedly healthy men, have been taken by some to imply a gradual change in the HPT axis with age.18,26 Modest changes in gonadotropin-releasing hormone pulsatility and Leydig cell responsiveness have provided mechanistic support but no aetiological basis for age-related dysfunction. Terms such as “late-onset hypogonadism” or “andropause” have been created to foster the idea that this decline in serum testosterone level, in isolation and without regard to pathological hypogonadism, can be considered an androgen-deficient state.27

In men (of any age) with intrinsically normal HPT function, systemic illness and accumulation of comorbidities, including renal, cardiac, inflammatory and mental health disorders, are associated with reduced levels of circulating testosterone. Not surprisingly, such men have a higher risk of serious intercurrent illness and earlier mortality. These men may have symptoms and signs resulting from the underlying (non-reproductive) disorder that overlap with those of androgen deficiency. In this setting, a low testosterone level is a biomarker for the underlying poor health.

While healthy older men exhibit lower testosterone concentrations on average than do healthy, reproductively normal younger men,12,20 there are no convincing data that healthy ageing necessarily results in a lowering of serum testosterone level to an extent that constitutes a clinical deficiency. Observational studies have reported independent associations of lower circulating androgen levels with poorer health outcomes in older men, including cognitive decline,28 frailty,29,30 incidence of stroke31 and mortality.32 However, other longitudinal studies indicate that a substantial proportion of the age-related fall in testosterone may be accounted for by obesity, burden of disease and medication use.33,34 In one cross-sectional study, a sample of men recruited for self-reported excellent health showed stable serum testosterone and LH concentrations into their eighth decade.14 The prevalence of middle-aged and older men who have both low circulating testosterone levels and sexual dysfunction is relatively low — about 2% in the European Male Aging Study.35 However, while that study interpreted declining serum testosterone levels as causing sexual dysfunction, there is evidence that sexual activity modestly increases serum testosterone levels36 and sexual inactivity is associated with decreased serum testosterone levels.37

In the United States, the Testosterone Trials showed a moderate improvement in sexual function in older men with a baseline testosterone level < 9.5 nmol/L who were treated with testosterone.38 At present, there is limited evidence showing clear benefit of treating older men with low–normal testosterone concentrations in the absence of pathological hypogonadism. While further research in this area is important, the immediate priority for these men is to identify and optimally manage any underlying medical comorbidity or mood disorder, such as depression. Where the diagnosis of hypogonadism is unclear, review by an endocrinologist may be warranted.

Obesity

Obesity is consistently associated with lower levels of circulating testosterone in cross-sectional studies and predicts declining testosterone levels in longitudinal studies. Serum LH and FSH levels are almost invariably normal in this setting. This association likely reflects predominantly reversible suppression of the HPT axis mediated by central adiposity, although testosterone treatment exhibits moderate effects in reducing fat mass and increasing lean mass.39–42 Successful weight loss, whether by diet or surgery, can lead to substantial increases in testosterone levels in obese men, whereas an 18-week randomised controlled trial (RCT) of testosterone treatment in obese men with sleep apnoea showed no benefit compared with placebo in terms of bodyweight or metabolic syndrome.43 The increase in testosterone level is proportional to the amount of weight lost: 10% weight loss increases testosterone by 2–3 nmol/L, whereas in morbidly obese men, profound weight loss after bariatric surgery can raise testosterone levels by > 10 nmol/L.44

Metabolic syndrome and diabetes

Men with metabolic syndrome have lower circulating testosterone levels compared with men without, and lower testosterone and SHBG levels are associated with increased risk of developing metabolic syndrome.45 Similarly, men with diabetes have lower testosterone concentrations than men without, and men with low testosterone levels are more likely to develop type 2 diabetes.46 Serum LH and FSH levels are almost invariably normal in these men. A recent meta-analysis of RCTs of testosterone treatment in men with metabolic syndrome or diabetes showed a marginal improvement in indices of insulin sensitivity with testosterone, but no evidence of better glycaemic control.47 Priority should be given to implementing lifestyle measures (especially exercise and dietary measures to lose excess weight) and optimising glycaemic control in the setting of diabetes. Such weight loss has many other demonstrable symptomatic and metabolic benefits. Persistently low testosterone levels despite successful weight loss should prompt reassessment to ensure a diagnosis of pathological hypogonadism has not been missed.

Glucocorticoid and opioid use

Use of systemic glucocorticoids has been associated with lower levels of circulating testosterone in men.48 In an RCT of 51 men receiving long term glucocorticoid therapy (5 mg or more of prednisone daily for at least 6 months or 1000 μg or more of inhaled steroid plus at least one course of oral prednisone within the past 6 months), 12 months of treatment with either testosterone or nandrolone (a minimally aromatisable androgen) increased muscle mass and reduced fat mass.49 In that study, testosterone increased lumbar bone mineral density and quality of life, while nandrolone did not.49 Opioid analgesia suppresses gonadotropin release, resulting in low levels of circulating testosterone, and there are data indicating that opioid treatment is associated with lower circulating testosterone levels.50–53 In a recent RCT, in which 84 men receiving opioid analgesia for chronic non-cancer pain and with baseline testosterone levels < 12.2 nmol/L (average, 8.7 nmol/L) were randomly assigned to receive testosterone or placebo, there was no significant effect of testosterone treatment on the primary outcome of self-reported pain or on opioid dosage.54 Testosterone improved one out of eight measures of experimental pain perception and reduced body fat, compared with placebo.54 Further studies are needed to determine whether testosterone therapy would effectively and safely improve health and functional outcomes in men receiving glucocorticoid or opioid therapy.

Where possible, removal of relevant glucocorticoid or opioid medications should be attempted, as this should reverse functional gonadal axis suppression and safely raise testosterone levels, with effective improvement in health and functional outcomes. The underlying strategy directs the focus back to primary health disorders, as these may be treatable and reverse the decline in serum testosterone levels. Where continuation of glucocorticoid or opioid therapy is necessary, review by an endocrinologist may be warranted for assessment and discussion of the risks versus benefits of different interventions. It is possible that therapeutic roles for androgens exist; however, additional evidence is needed from properly designed placebo-controlled RCTs with outcomes related to patient health and functional improvement.

Conclusions

Pathological hypogonadism can present in many ways, such as psychosexual symptoms or unexplained osteoporosis, anaemia or reduced muscle mass, and the diagnosis can be missed. It is a clinical diagnosis with a pathological basis, confirmed by low testosterone levels. TRT is warranted in men with pathological hypogonadism to alleviate symptoms and signs of androgen deficiency, and it may have marked benefits in these men. By contrast, in men with preserved spermatogenesis and without pathological hypogonadism, testosterone treatment is expected to impair fertility. In the absence of hypothalamic, pituitary or testicular disorders, testosterone therapy is not justified in older men with medical comorbidities. Obese men who may have reduced testosterone levels should be encouraged to lose weight using a combination of diet and exercise. Currently, there is limited evidence for the efficacy and safety of testosterone treatment in men with diabetes and those receiving glucocorticoid or opioid therapy. Further evidence from well-conducted RCTs is required before such treatments can be fully evaluated.

Importantly, men who have classic causes of pathological hypogonadism that have been missed earlier in life, or men who have developed disorders of the HPT axis, may present at any age; there is universal agreement that such men should be identified and considered for treatment. It is important not to miss a diagnosis of pathological hypogonadism in a man who may coincidentally have diabetes or another comorbidity (including among older men with chronic disease), or to confuse it with non-specific effects of systemic disease on the reproductive system.

The recommendations given in this position statement are based on data from a limited number of RCTs, as well as non-randomised clinical studies and observational studies. As such, further research is warranted, which may have an impact on the recommendations.