Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
Subclinical thyrotoxicosis may be defined as a low serum thyrotrophin (TSH) concentration in an asymptomatic patient with normal serum free thyroxine (T4) and triiodothyronine (T3) concentrations. The secretion of TSH may be suppressed even in the presence of normal serum thyroid hormone levels. This reflects the highly sensitive response that the pituitary gland mounts to minor changes in serum free T4 and T3concentrations inside the normal range of the population—exemplified by the log-linear relation between serum TSH and thyroid hormone concentrations.1
In this review I shall address the various disorders characterised by a low serum TSH and how to differentiate them from subclinical thyrotoxicosis. I also discuss the causes of subclinical thyrotoxicosis, its potential pathophysiological and clinical consequences, and its management.
The introduction in the mid-1980s of sensitive assays for TSH allowed the measurement of serum TSH concentrations well below the normal range of 0.5 to 4.0 mIU/l. Unlike the first generation radioimmunoassays, which have a sensitivity of about 0.1 mIU/l, the new immunometric second and third generation assays have greater sensitivities of approximately 0.05 and 0.005 mIU/l, respectively, with about a 10-fold increase in sensitivity per generation.1 2 The thyrotrophin releasing hormone (TRH) stimulation test has since played a diminishing diagnostic role.3
Causes of low serum TSH concentrations
The main causes of low serum TSH concentrations are overt thyrotoxicosis, non-thyroidal illness, secondary (central) hypothyroidism, physiological causes, and subclinical thyrotoxicosis.
Overt thyrotoxicosis is characterised by high serum free T4 and T3 and very low serum TSH concentrations, which are undetectable in more than 95% of patients, even with third generation assays,4 and it is associated in the majority of patients with overt symptoms and signs of thyrotoxicosis. Two small subgroups of thyrotoxic patients are characterised by a normal serum T4 in association with either a raised total T3 and a normal thyroxine binding globulin (T3toxicosis)5 or, more rarely, a normal total T3 but a raised free T3(free T3toxicosis).6 The diagnosis in these overtly thyrotoxic patients can therefore be readily made by the demonstration of a rise in either serum free T4 or total or free T3, or both.
Some patients with non-thyroidal illnesses may have low serum TSH concentrations. This is invariably associated with low serum concentrations of T3 and often T4. Although the pathogenesis of the reduction in serum TSH in these patients remains uncertain, decreased secretion of TRH, increased secretion of somatostatin, cortisol, or cytokines, and inhibition of TSH secretion by drugs such as dopamine or glucocorticoids may contribute. In addition to a low serum TSH concentration,secondary (central) hypothyroidism is often associated with other pituitary hormone deficiencies, and may be accompanied by neurological abnormalities related to the local effect of a hypothalamic-pituitary tumour.
Physiological causes of a reduction in serum TSH include pregnancy and old age. The thyroid stimulating activity of chorionic gonadotrophin, particularly in women with hyperemesis gravidarum, may increase the thyroid secretory activity enough to lower serum TSH concentration to below normal. This is a common finding near the end of the first trimester or at the beginning of the second trimester, and is significantly more common in Asian women than in those of European origin (15.7% v 4.8%).7 The pulsatile secretion of TSH may result in an interpulse nadir value that is below normal in some elderly subjects.
The causes of subclinical thyrotoxicosis, in which TSH secretion is suppressed but the concentrations of circulating thyroid hormones remain normal—albeit in the upper normal range—can be divided intoexogenous andendogenous. Drug related subclinical thyrotoxicosis may occur, as in the administration of supraphysiological doses of thyroid hormone, and in drug induced thyroiditis (amiodarone, α interferon, iodine in patients with multinodular goitre). Overtreatment with thyroid hormones may be intentional, as in patients with thyroid carcinoma in order to reduce TSH secretion to below normal, or unintentional, as in approximately one quarter of hypothyroid patients receiving treatment.8L-thyroxine overtreatment is by far the commonest cause of subclinical thyrotoxicosis.9
When subclinical thyrotoxicosis results from an autonomous oversecretion of thyroid hormones by a diseased thyroid gland, it is termed subclinical hyperthyroidism. Approximately one fifth of patients with Graves' disease who have received medical (radioactive iodine)10 or surgical (subtotal thyroidectomy)11 treatment and are euthyroid may be insufficiently treated and have low serum TSH concentrations. A similar proportion of patients with Graves' disease who remain euthyroid after discontinuation of antithyroid drug treatment may have low serum TSH concentrations which indicate a state of subclinical hyperthyroidism.11 It is possible that subclinical hyperthyroidism may be caused by Graves' disease in some patients who have never been overtly thyrotoxic, as diffuse isotope uptake by the thyroid gland and TSH receptor antibodies can be demonstrated in these patients.12 Subclinical Graves' hyperthyroidism accounts for some patients with “euthyroid Graves' ophthalmopathy”.13
Some patients with an autonomously functioning thyroid adenoma14 and about a quarter of patients with multinodular disease15 have persistently low serum TSH concentrations. Multinodular goitre and Graves' disease are the commonest of endogenous causes of subclinical hyperthyroidism.15 Transient subclinical hyperthyroidism lasting for a few weeks or months may occur in patients with thyroiditis, and after initiation of treatment in overtly thyrotoxic patients who later become euthyroid. Thyroiditis can be “painless,” such as in silent, postpartum, and drug induced thyroiditis, or “painful,” as in subacute thyroiditis.12
The clinical relevance of subclinical thyrotoxicosis
Although subclinical thyrotoxicosis is not an uncommon disorder and may be identified in 2–16% of the population in large community and clinic surveys,16-20 low serum TSH concentrations are often a transient finding of unknown clinical importance.16 The significance of subclinical thyrotoxicosis, however, is related to the small but important risk of progression to overt thyrotoxicosis in patients with thyroid disease. Furthermore, subclinical thyrotoxicosis may have important pathophysiological and subtle or previously unrecognised clinical effects. Treatment of these patients may therefore be beneficial.
RISK OF PROGRESSION TO OVERT THYROTOXICOSIS
Although spontaneous regression to normal serum TSH concentrations in subjects with low levels on initial testing may occur,18 progression to overt thyrotoxicosis in some patients with subclinical hyperthyroidism associated with multinodular goitre21 or Graves' disease22 has been observed. Overt thyrotoxicosis may also be induced in patients with subclinical hyperthyroidism by the administration of iodine containing drugs such as radiocontrast media or amiodarone.23 24
PATHOPHYSIOLOGICAL EFFECTS OF SUBCLINICAL HYPERTHYROIDISM
Although the effects of overt thyrotoxicosis on cardiovascular function are well recognised and include cardiomegaly, atrial fibrillation, and congestive heart failure,25 the effects of subclinical thyrotoxicosis are less certain. Nonetheless, patients with subclinical thyrotoxicosis appear to be at increased risk of atrial fibrillation. In the subjects of the Framingham cohort, who were followed up for up to 10 years, atrial fibrillation developed in 21% of those with low serum TSH concentrations (⩽ 0.1 mIU/l), 12% of subjects with slightly low serum TSH concentrations (> 0.10 to 0.40 mIU/l), and 8% of subjects with normal TSH concentrations.26 When adjusted for other known risk factors for atrial fibrillation, a low serum TSH concentration conferred a threefold relative risk for the development of atrial fibrillation.26 This is an important observation, as atrial fibrillation is a recognised independent risk factor for the development of arterial thromboembolism, stroke, and congestive heart failure, and is associated with an increased risk of death.27 The prevalence of subclinical thyrotoxicosis in patients with idiopathic atrial fibrillation, however, is low (5.5% in a Canadian survey28). The effects of subclinical thyrotoxicosis on cardiac function are unpredictable. While some studies in patients receiving long term TSH suppressive treatment with thyroxine have shown an increase in left ventricular mass resulting in diastolic dysfunction, with impairment of cardiac contractility and exercise capacity,29 others have reported normal30 or improved31 ventricular function.
Patients with overt thyrotoxicosis are at risk of osteoporosis caused by accelerated bone turnover32—thyroid hormones lead to a greater degree of bone resorption than bone formation.33 Bone mineral density at the femoral neck, lumbar spine, and midshaft or distal radius is also decreased in patients with longstanding subclinical thyrotoxicosis caused by prolonged TSH suppressive doses of thyroxine34 35 or a solitary autonomously functioning thyroid adenoma.14 This osteopenia is more pronounced in, or limited to, postmenopausal women compared with premenopausal women and men, and affects cortical bone (forearm and hip) more than trabecular bone (vertebrae).14 35 Although other investigators failed to show an association between subclinical thyrotoxicosis and a reduction in bone density,36 37 this may in part be related to the cross sectional design of most studies, with the inherent discrepancy between patients and controls in some of the relevant variables, the size of the study population,38 and the inclusion of patients who remained biochemically thyrotoxic for some time or who had non-suppressed TSH levels during treatment with L-thyroxine.37 Nonetheless, it seems possible that long lasting endogenous or exogenous subclinical thyrotoxicosis may be a contributing factor in the development of osteoporosis in some postmenopausal women, mostly at sites where cortical bone preponderates.35
Despite evidence suggesting a slight decrease in bone mineral density in subjects with subclinical thyrotoxicosis, especially in postmenopausal women, there are no studies indicating an increase in fracture rate in these subjects. In patients treated with L-thyroxine, the fracture rate was not increased in those with low compared with normal TSH concentrations.39
Metabolic and biochemical changes
Subclinical thyrotoxicosis may be associated with metabolic and biochemical changes that are similar, albeit mild, to those observed in overtly thyrotoxic patients. These changes, however, are uncommon and appear to be of no clinical consequence. They may include a slight increase in resting energy expenditure,40 slight reductions in serum total and low density lipoprotein cholesterol concentrations,41 a small increase in serum sex hormone binding globulin concentrations,42 and a slight increase in markers of bone turnover (serum osteocalcin, urinary pyridinoline, and deoxypyridinoline).43
DIFFERENTIAL DIAGNOSIS OF LOW SERUM TSH CONCENTRATION
The unexpected finding of a low serum TSH but normal free T4 concentration in a patient who does not have a significant non-thyroidal illness and ingests no pharmacological agents known to suppress TSH production (such as thyroxine) may suggest hyperthyroidism. However, if goitre is not clinically evident and symptoms and signs of thyrotoxicosis are absent, the decrease in serum TSH may be a transient and insignificant phenomenon. Additional investigations can, therefore, be reserved for patients who, on further measurement few weeks or months later, show persistently low serum TSH concentration. If goitre or symptoms and signs of thyrotoxicosis are present, or if serum TSH concentrations remain low then serum total T3 and, if normal, free T3 concentrations should be measured. The detection of a raised total or, if normal, free T3 strongly suggests T3 and free T3toxicosis, respectively. This diagnosis may be substantiated by the demonstration of a high isotope uptake on radionuclide imaging.
If the serum concentrations of total T3 and free T3 are normal, then subclinical hyperthyroidism should be distinguished from subclinical central hypothyroidism. The finding of free T4 or free T3 concentrations, or both, within the upper half of the normal range suggests the former disorder, while levels within the lower half of the normal range suggest the latter. This interpretation may be further corroborated by radionuclide imaging of the thyroid gland. A normal or high isotope uptake implies autonomous thyroid function and supports a diagnosis of subclinical hyperthyroidism, while a low isotope uptake suggests central hypothyroidism. A low isotope uptake, however, in the presence of serum free T4 or free T3 concentrations within the upper half of the normal range, may suggest thyroiditis, surreptitious ingestion of thyroid hormones, or a high iodine intake.
In patients with subclinical hyperthyroidism, the detection of serum TSH receptor antibodies or thyroid stimulating antibodies22 46 is useful for detecting subclinical Graves' disease. The detection of high titres of autoantibodies to thyroid antigens (thyroglobulin, thyroid peroxidase) is characteristic of autoimmune thyroiditis.47
TREATMENT OF SUBCLINICAL THYROTOXICOSIS
The indications for treatment in patients with subclinical thyrotoxicosis are not clearly defined. Factors that favour the administration of antithyroid treatment, whether pharmacological or surgical, include older age, the presence of even mild symptoms of thyrotoxicosis, risk of osteoporosis, the presence of atrial arrhythmias, and a large goitre. Periodic observation, however, is all that may be required in younger patients lacking any of the above factors, as the risk of progression to overt thyrotoxicosis is small.
On the other hand, the recommendation for treatment should be tempered by its effectiveness and its potential risks and disadvantages. Radioiodine is an effective treatment in the management of nodular thyroid disease (autonomously functioning thyroid adenoma and multinodular goitre), and carries a low risk of hypothyroidism. However, it may be preferable to manage patients with subclinical Graves' disease conservatively with observation alone, as treatment with antithyroid drugs carries some risk and may need to be prolonged, while radioiodine treatment is associated with a high risk of hypothyroidism.
Treatments may also be directed towards the correction or prevention of some of the deleterious pathophysiological effects of subclinical thyrotoxicosis on the cardiovascular and skeletal systems. Concurrent administration of β adrenergic receptor antagonists has been shown to offset the changes in ventricular mass and diastolic dysfunction in patients receiving long term TSH suppressive treatment with L-thyroxine,29 and improves the quality of life in these patients.48 Cardiomyopathy associated with subclinical hyperthyroidism may be reversed by treatment with antithyroid drugs.49 On the other hand, treatment with L-thyroxine may improve cardiac and exercise performance in non-hyperthyroid patients with idiopathic dilated cardiomyopathy.50
Oestrogen replacement treatment has a known beneficial effect on bone density, and has been shown to attenuate the reduction in bone mineral density in postmenopausal women taking TSH suppressive doses of L-thyroxine.51 Antithyroid treatment may decelerate bone loss in patients with endogenous subclinical hyperthyroidism.52 Bisphosphonate treatment may also preserve bone density in these patients.53 In postmenopausal women on L-thyroxine, a reduction in T4 dose in those with suppressed serum TSH levels can result in a decrease in bone turnover and an increase in bone mineral density.43