How should we manage patients with mildly increased serum thyrotrophin concentrations?
Mark P.J. Vanderpump, Consultant Physician and Honorary Senior Lecturer in Endocrinology and Diabetes, Department of Endocrinology, Royal Free Hampstead NHS Trust, Pond Street, London, NW3 2QG, UK. Tel: (+44) 0207 4726280; Fax: (+44) 0207 4726487; E-mail: email@example.com
A mildly increased serum thyrotrophin (TSH) is usually because of mild thyroid failure, and the most common aetiology in iodine-replete communities is chronic autoimmune thyroiditis. It is more common in women, and the prevalence increases with age in both men and women and is associated with the presence of antithyroid antibodies. The majority will have serum TSH levels between 5–10 mIU/l, normal free thyroxine (T4) levels and relatively few symptoms. In 2004, US evidence-based consensus guidelines concluded that there were no adverse outcomes of a mildly increased serum TSH other than a risk of progression to overt hypothyroidism and few data to justify levothyroxine therapy. There is still debate as to what constitutes an increased serum TSH, particularly in older subjects. Although some subjects will progress to overt hypothyroidism, recent data suggest a significant proportion revert to a serum TSH within the reference range without treatment. Two recent meta-analyses have suggested that the possible cardiovascular risks may be more significant in younger adults. Other data suggest that mild thyroid failure may be the only reversible cause of left ventricular diastolic dysfunction. No appropriately powered prospective, randomized, controlled, double-blinded interventional trial of levothyroxine therapy for a mildly increased serum TSH exists. However, treatment in subjects who are symptomatic, pregnant or preconception, aged less than 65 years and older subjects with evidence of heart failure appear justified.
Serum thyrotrophin (TSH) is the first-line diagnostic test for hypothyroidism. Normal serum TSH levels in thyroid disease-free individuals are typically cited as about 0·4–4·0mIU/l.1–3 Mild thyroid failure or subclinical hypothyroidism is normally defined in the absence of hypothalamic or pituitary disease, nonthyroidal illness or drugs.
A mildly increased serum TSH may be found either postradioiodine therapy or postsurgery in up to 50% of apparently euthyroid patients and usually represents a stage in the progression towards overt thyroid failure. If present 1 year or more after radioiodine or surgical treatment, then the annual rate of progression to overt hypothyroidism after either treatment is 2–6%. Less frequent causes include external beam irradiation of malignant tumours of the head and neck, and drugs including lithium, amiodarone and interferon, undiagnosed Addison’s disease and renal failure. Circulating antibodies against TSH or mouse immunoglobulins, when used as assay reagents, can also yield falsely raised TSH immunoassay readings. Thyroid hormone resistance and TSH-secreting pituitary tumours can present as an inappropriately elevated serum TSH but raised thyroid hormone levels.1
Congenital hypothyroidism may present with subclinical hypothyroidism and is a treatable cause of mental retardation. In iodine-replete areas, 85% of the cases are due to sporadic developmental defects of the thyroid gland (thyroid dysgenesis) such as the arrested migration of the embryonic thyroid (ectopic thyroid) or a complete absence of thyroid tissue (athyreosis). The remaining 15% have thyroid dyshormonogenesis defects transmitted by an autosomal recessive mode of inheritance. Iodine deficiency (below 25 μg a day), particularly in preterm infants, accounts for many cases in Eastern Europe, Asia and Africa.
In epidemiological surveys of iodine-replete communities, a mildly increased serum TSH (between 4 and 10 mIU/l) is associated with positive thyroid peroxidase (TPO) antibodies in 60–80% of the subjects.4 The Whickham survey in North East England found that 7·5% of women (10% of women over 55 years of age) and 2·8% of men had an elevated serum TSH and a normal T4 level.5 A cross-sectional screening survey in Colorado, USA of 25,682 adult subjects found that 9·4% of the population, excluding 1525 on levothyroxine therapy, had an increased serum TSH level, of whom 9% had a normal free T4.6 Among those with a high serum TSH concentration, 74% had a value between 5·1 and 10 mIU/l and 26% had a value greater than 10 mIU/l. The percentage of subjects with an increased serum TSH concentration was higher for women than men in each decade of age, and ranged from 4 to 21% in women and 3 to 16% in men. In the third National Health and Nutrition Examination Survey (NHANES III) of 17,353 US subjects, 4·3% had a mildly increased serum TSH which rose to 13·7% of individuals aged 65 years and older.7 Approximately, 2% of adolescents aged 12 to 19 years had a serum TSH greater than 4·5 mIU/l. Studies of elderly persons have confirmed the high prevalence of a raised serum TSH in this age group, with approximately 10% of subjects over 60 years in a single general practice in Birmingham, UK having serum TSH values above the reference range.8 More recent data from Birmingham suggest that there is now an increased awareness of thyroid disease and testing of thyroid function and perhaps earlier use of levothyroxine in mild thyroid failure. A community-based sample of 5960 participants aged 65 years and older demonstrated a lower prevalence of subclinical hypothyroidism of 2·9% (95% confidence interval 2·6–3·1%).9 In surveys of hospital inpatients, the point prevalence rates were similar being between 3 and 6% with most subjects reverting to normal thyroid function 3 months following the acute illness.3
The management of patients with overt primary hypothyroidism is usually straightforward, and levothyroxine is offered to anyone with characteristic clinical features, an increased serum TSH and a low T4 concentration. However, there has been considerable debate as to whether patients with a mildly increased serum TSH above the reference range with normal circulating T4 levels warrant levothyroxine. Some authorities have suggested that the upper limit of the reference range be lowered to 2·5 mIU/l in view of an increased prevalence of TPO antibodies.1,2 A recent further analysis of the NHANES III data demonstrated that 11% of 20–29 year-olds had a serum TSH greater than 2·5 mIU/l, increasing to 40% in the subjects aged 80 and over. The 97·5 centile for the subjects aged 80 and over was 7·49 mIU/l and 70% had a serum TSH greater than the population defined upper limit of the reference range of 4·5 mIU/l of which only 40% were antithyroid antibody positive.10
In 2004, US evidence-based consensus guidelines concluded that there were no adverse outcomes in those subjects with a serum TSH between 5–10 mIU/l other than a risk of progression to overt hypothyroidism and few data to justify levothyroxine therapy.11 More recent guidance has supported this view.3,12,13 However, a mildly increased serum TSH remains a common reason for referral to an endocrinologist. For example, I was recently asked whether a 40-year-old woman who had presented to her general practitioner following a miscarriage with a serum TSH 5·5 mIU/l (reference range 0·4–4 mIU/l) and free T4 12·8 pmol/l (reference range 11–20 pmol/l) warranted levothyroxine therapy? Expert guidelines provide advice but what factors do I consider with an individual patient in my clinic before deciding on the appropriate management for this biochemical abnormality?
Is the serum TSH persistently elevated?
In practice, the serum TSH level should be repeated, along with a test for TPO antibodies, within 3 months of the initial test to ascertain the consistency of the abnormal TSH. A transient rise in serum TSH suggests recovery from a nonthyroidal illness or a thyroiditis, either viral or autoimmune in aetiology. The likely aetiology, if the mildly increased serum TSH persists, is chronic autoimmune thyroiditis and TPO antibodies will be present in the majority. Clinical features associated with autoimmune thyroiditis may include a personal or family history of thyroid autoimmunity or other organ-specific autoimmune diseases (such as Addison’s disease, type 1 diabetes mellitus, primary ovarian failure, coeliac disease, vitiligo, alopecia and pernicious anaemia), Down’s syndrome and Turner’s syndrome and rarely the presence of a diffuse goitre.1
If a normal serum T4 and a mildly raised TSH (defined as more than 15 mIU/l) are found in neonates at screening for congenital hypothyroidism, it is recommended that treatment with levothyroxine is commenced promptly in view of the critical dependence of the developing central nervous system on adequate amounts of thyroid hormone. There remains controversy about the need to treat older children if a mildly raised serum TSH, usually due to autoimmune thyroiditis, is confirmed on repeat testing. The abnormalities most frequently associated in the paediatric population include goitre, weight gain, growth delay, lethargy and impaired psychomotor and cognitive development. Although a mildly raised serum TSH may show a very slow evolution to overt hypothyroidism, levothyroxine therapy is usually warranted.14–16
Is there a possibility of pregnancy?
There is some evidence that women with autoimmune thyroiditis are more likely to have an increased risk of recurrent miscarriage although it is not known whether the risk is related to thyroid autoimmunity or to subtle thyroid failure.17 Acceleration of thyroid hormone metabolism during pregnancy can lead to hypothyroidism in women with a limited thyroid hormone reserve. Untreated mild hypothyroidism can lead to subtle impairment of subsequent childhood neuropsychological development.1,2 Thyroid dysfunction in relation to pregnancy was considered in the recent Endocrine Society Guidelines.18 Early correction of maternal hypothyroxinaemia was recommended aiming to maintain the serum TSH in the lower half of the reference range prior to conception if possible.
Does the patient have symptoms?
The presence of nonspecific symptoms in patients with a mildly elevated serum TSH remains controversial. Symptoms are probably related to disease severity, disease duration and individual sensitivity to thyroid hormone deficiency.2 Most patients will be diagnosed in primary care after investigation of symptoms or following screening in asymptomatic subjects. The Colorado study found certain symptoms such as dry skin, cold intolerance and easy fatigability appear more often in those with a mildly elevated serum TSH compared with euthyroid subjects.6 The available double-blind placebo-controlled trials of the value of levothyroxine therapy on symptoms associated with an elevated serum TSH are inconclusive until the serum TSH is more significantly raised more than 10 mIU/l.19 However, if patients have presented with symptoms of hypothyroidism, then a 3–6 month trial of levothyroxine therapy appears warranted.
If the patient is asymptomatic and been identified although a screening programme then it is reasonable to be more cautious before recommending levothyroxine. If identified as having a disease, i.e. mild thyroid failure, some subjects might suffer from the labelling effect similar to that seen in hypertension. Subjects with positive screening tests do not always comply with treatment, particularly if their symptoms do not change or they suffer from side effects. No randomized controlled trial of a complete screening programme for subclinical hypothyroidism exists and there is insufficient evidence to support population screening.11
Does the patient have evidence of dyslipidaemia?
The evidence of adverse effects of a mildly elevated serum TSH on cholesterol concentrations remains inconsistent.2 Most but not all studies have found an association with a raised low-density lipoprotein cholesterol, low high-density lipoprotein cholesterol and raised lipoprotein (a). Levothyroxine treatment of patients with mild thyroid failure has been estimated to be lower serum total cholesterol by 0·2–0·4 mmol/l and low-density lipoprotein cholesterol by 0·2 mmol/l. Among the studies that have stratified patients by serum TSH level, the magnitude of the change in lipids following levothyroxine therapy appeared to depend upon the severity of thyroid dysfunction. In the only one of five studies that was a randomized controlled trial, reductions in serum lipids were not significant in those individuals with a serum TSH less than 10 mIU/l. Mild thyroid failure is however associated with a higher risk of statin induced myopathy.20 I therefore recommend levothyroxine in high-risk cardiovascular patients starting statin therapy.
What is the likelihood of progression to overt hypothyroidism?
Spontaneous recovery has also been described in subjects with a mildly raised serum TSH, although the frequency of this phenomenon is unclear.2 In one study, 37% of patients normalized their serum TSH levels over a mean follow-up time of 32 months. Normalization of serum TSH concentrations was more likely to occur in patients with negative antithyroid antibodies and serum TSH levels less than 10 mIU/l, and within the first 2 years after diagnosis.21
All the available longitudinal studies indicate that the higher the serum TSH value, the greater the likelihood of development of overt hypothyroidism in subjects with chronic autoimmune thyroiditis. At the Whickham 20-year follow-up study, either a raised serum TSH or positive antithyroid antibodies alone or in combination were associated with a significantly increased risk of hypothyroidism.22 In women, the annual risk of spontaneous overt hypothyroidism is 4% in those who had both high serum TSH and antithyroid antibody concentrations, 3% if only their serum TSH concentrations were high and 2% if only their serum thyroid antibody concentration were high; at the time of follow-up the respective rates of hypothyroidism were 55%, 33% and 27%. A serum TSH above 2 mU/l was associated with an increased risk of developing hypothyroidism. Patients may find it helpful to assess their individual risk of hypothyroidism over a 20-year-period calculated from the Whickham data (see Table 1). In a recent 5-year follow-up study of 790 women with a family history of autoimmune thyroid disease in the Netherlands, the incidence of hypothyroidism was 1·9–2·7 times higher than seen in the general female population of the UK and was also higher in those with serum TSH greater than 2·0 mIU/l and high TPO antibody concentrations at baseline.23
Table 1. The probability (%) of a woman developing hypothyroidism in 20 years calculated from age, serum TSH and anti-thyroid antibody status (defined as Thy Ab + if thyroid microsomal antibodies titre >1:100)
Is treatment likely to improve cardiovascular morbidity or mortality?
Two recent meta-analyses of ten selected population-based cohort studies have examined whether mild thyroid hormone failure increase coronary heart disease (CHD) risk.24,25 The relative risks for CHD events and cardiovascular and overall mortality were 1·2, 1·2 and 1·1 respectively. Limiting analyses to studies with the most rigorous methodologies slightly decreased the risk estimates. However, both analyses suggested that the cardiovascular risks may be more significant in younger adults, approximately 1·5 vs. 1·0 for populations with mean ages younger or older than 65 years respectively. There is also evidence in the very elderly people that mild thyroid failure may even be associated with longevity.26,27 Other recent epidemiological data suggest that mild thyroid failure may be the only reversible cause of left ventricular diastolic dysfunction, particularly in those subjects with a serum TSH greater than 10 mIU/l.28 The most consistent cardiac abnormality reported in patients with a mildly elevated serum TSH is impaired left ventricular diastolic function, which is characterized by slowed myocardial relaxation and impaired ventricular filling.2 Treatment with levothyroxine has been demonstrated to reverse echocardiogram abnormalities associated with a mildly raised serum TSH 5–10 mIU/l.2
Normalization of serum TSH with levothyroxine is often not achieved in clinical practice6,22,29 and detrimental effects on the skeleton, the cardiovascular system and even mortality have been suggested by subclinical hyperthyroidism, which can be a consequence of over-treatment with levothyroxine therapy. No appropriately powered prospective, randomized, controlled, double-blinded interventional trial of thyroxine therapy for a mildly increased serum TSH exists. Adopting a ‘wait and see’ policy rather than intervention may avoid unnecessary treatment or the potential for harm.
If the serum TSH is greater than 10 mIU/l, then levothyroxine is indicated. If the serum TSH is mildly increased between 4 and 10 mIU/l, and the patient is TPO antibody positive, an annual check of serum TSH is recommended with the commencement of levothyroxine, once the serum TSH rises above 10 mIU/l. If the patient is thyroid-antibody negative, then ensuring a check of serum TSH every 3–5 years is all that is required. If the patient with a serum TSH 4–10 mIU/l has symptoms consistent with hypothyroidism and the TSH elevation persists, then a 3–6 month therapeutic trial of levothyroxine appears justified. If the patient feels improved by therapy, it is reasonable to continue treatment. If the patient does not have symptoms and the serum TSH level appears stable with positive TPO antibodies, the annual risk of progression to overt hypothyroidism is 4% per year and so an annual serum TSH surveillance strategy is warranted. If the TPO antibodies are negative, then 3-yearly serum TSH surveillance is recommended, with the risk of progression to overt hypothyroidism 3% per year.
The exception to the above is in neonates and children, pregnancy, or in someone trying to conceive, when a mildly increased serum TSH should always be treated as it is associated with adverse outcomes for both mother and foetus. Recent data suggest that levothyroxine treatment for those with serum TSH 5–10 mIU/l may also now be justified in subjects less than 65 years and in those older subjects with documented evidence of heart failure on echocardiography. In view of the history of miscarriage and the possibility of a future pregnancy, I recommended treatment to my 40-year-old patient who had presented following a miscarriage with a serum TSH 5·5 mIU/l and achieved a target serum TSH 1·2 mIU/l on levothyroxine 50 μg daily.
Competing interests/financial disclosure
Nothing to declare.