WHAT IS SCREENING?
Screening was first defined as “the presumptive identification of unrecognised disease or defect by the application of tests, examinations, or other procedures which can be applied rapidly to sort out apparent well persons who probably have a disease from those who probably do not” (1). Screening is justified when 1) a disease is common and associated with serious morbidity and mortality; 2) screening tests are sufficiently accurate in detecting early stage disease, are acceptable to patients, and are feasible in general clinical practice; 3) treatment after detection by screening has been shown to improve prognosis relative to treatment after usual diagnosis; and 4) evidence exists that the potential benefits outweigh the potential harms and costs of screening (2). These criteria were recently updated by a National Screening Committee from the UK Health Department (3).

Three tiers of screening exist and must be considered in a screening programme (4):
1. The nanolevel of care delivered to an individual patient by a medical practitioner in the office or out-patient clinic
2. The microlevel of care delivered usually by an individual practitioner or practitioners within a specialty to a community population served by a district hospital
3. The macrolevel of care delivered to a whole population by a regional or national organisation.

EPIDEMIOLOGY OF THYROID DYSFUNCTION
The earliest biochemical abnormality in hypothyroidism is an increase in serum thyrotrophin (TSH) concentration associated with normal serum thyroxine (T₄) and triiodothyronine (T₃) concentrations (subclinical hypothyroidism), followed by a decrease in serum T₄ concentration, at which stage most patients have symptoms and benefit from treatment (overt hypothyroidism). The prevalence of spontaneous hypothyroidism is between 1% and 2%, and it is more common in older women and ten times more common in women than in men (5). In cross-sectional community studies, the prevalence of newly diagnosed overt hypothyroidism is 3-4 per 1000 (6,7,8). The cause is either chronic autoimmune disease (atrophic autoimmune thyroiditis or goitrous autoimmune thyroiditis (Hashimoto's thyroiditis)) or destructive treatment for thyrotoxicosis, which may account for up to one-third of cases of hypothyroidism in the community. Subclinical hypothyroidism is found in 8% of women (10% of women over 55 years of age) and 3% of men (6,7,8) (see Figure ) and can progress to overt hypothyroidism, particularly if antithyroid antibody positive (9).

LEGEND TO FIGURE The percentage of 25,682 subjects with a high serum thyrotropin (TSH) concentration, by sex and decade of age, in the Colorado Thyroid Study. (From Canaris GJ, Manowitz NR, Mayor G, et al. The Colorado Thyroid Disease Prevalence Study. Arch Intern Med 160:526-534,2000)

The prevalence of a history of thyrotoxicosis in women is between 0.5 and 2%, and it is also ten times more common in women than in men (10). In cross-sectional community studies, the prevalence of undiagnosed thyrotoxicosis is 1-2 per 1000 (7,8) and the most common causes are Graves' disease, followed by toxic multinodular goiter, whilst rarer causes include an autonomously functioning thyroid adenoma, or subacute thyroiditis. Subclinical hyperthyroidism is defined as a low serum TSH concentration and normal serum T₄ and T₃ concentrations, in the absence of hypothalamic or pituitary disease, non-thyroidal illness, or ingestion of drugs that inhibit TSH secretion. The available studies differ in the definition of a low serum TSH concentration and whether the subjects included were receiving T₄ therapy for hypothyroidism. The reported overall prevalence ranges from 0.5% to 6.3%, with men and women over 65 years having the highest prevalence; approximately half of them are taking thyroxine (7,8,9).

RECOGNISED INDICATIONS FOR SCREENING
Screening programmes for congenital hypothyroidism were developed in the 1970s in which TSH or T₄ were measured in heel-prick blood specimens to detect this condition as early as possible. Certain groups within the population who should have an assessment of thyroid function at least once include those with atrial fibrillation, dyslipidaemia, subfertility and osteoporosis and those patients who present with a suspected goitre (5). Routine testing of thyroid function in patients admitted acutely to hospital is not warranted unless specific clinical indications exist (11). Surveillance of thyroid function is recommended following destructive treatment for thyotoxicosis by either radioiodine or surgery, in women with a past history of post-partum thyroiditis, patients with diabetes particularly women with type 1 diabetes, Down Syndrome and Turner’s Syndrome, post neck irradiation and patients receiving lithium and amiodarone therapy (5).

POPULATION SCREENING?
Controversy exists as to whether healthy adults in the community would benefit from screening for thyroid disease (10). The prevalence of unsuspected overt hypothyroidism or thyrotoxicosis is low but a significant proportion of subjects tested will have evidence of mild thyroid failure or excess. Different recommendations and position papers have been reported by various physician organisations as to whether subclinical thyroid disease is of sufficient clinical importance to warrant screening (12-18). A recent cost-utility analysis using a computer decision model has however suggested that the cost-effectiveness of screening for mild thyroid failure compares favourably with other preventive medical practices (19). These data led the ATA to recommend population-based screening for thyroid dysfunction by measurement of serum TSH, beginning at age 35 years and every five years thereafter (20). However, a recent scientific review from representatives of the ATA, the Endocrine Society and AACE concluded that there was insufficient evidence to support population screening (21).

CRITERIA FOR POPULATION SCREENING

1. The Condition
Hypothyroidism is an insidious condition with a significant morbidity and the subtle and non-specific symptoms and signs may be mistakenly attributed to other illnesses, particularly in post-partum women and the elderly. Thyrotoxicosis has a significant short-term morbidity and long-term morbidity and mortality. Apart from progression to overt hypothyroidism, subclinical hypothyroidism may result in non-specific symptoms, dyslipidaemia and increased risk of cardiovascular disease. The possible consequences of subclinical hyperthyroidism include progression to overt thyrotoxicosis, systemic symptoms, atrial fibrillation and adverse cardiac end points and osteoporosis. The quality of evidence on the strength of association with these outcomes in subclinical hypothyroidism and hyperthyroidism has recently been reviewed (21).

2. The Screening Test
The first test in any screening programme is that which identifies those within the population who would benefit from screening. In the absence of the confounding effects of non-thyroidal illness or drugs, a normal serum TSH concentration has a high negative predictive value in ruling out thyroid disease in the healthy ambulant subject (22). Normal serum TSH concentrations may be recorded in hypothyroidism secondary to pituitary or hypothalamic disease but this is rare. In nearly all populations screened a raised serum TSH above 5mU/l is accepted as being unequivocally raised. A simple blood test is usually acceptable in most populations and is often included in a health screening process.

The standard follow-up investigation for subjects with a raised serum TSH greater than 5mU/l is a repeat measurement plus a serum FT₄. Measurement of thyroid peroxidase antibodies in subjects with a borderline raised serum TSH that is found by screening the general population may be justified (10). If there is no intervention, then an annual test of thyroid function is warranted.

Few subjects screened will have overt hyperthyroidism but the consequences of finding a suppressed serum TSH have to be addressed when using a TSH assay. Subjects with subclinical hyperthyroidism can be categorised into those with low but detectable serum TSH (0.1-0.4mU/L) and those with a clearly low serum TSH (less than 0.1mU/L). If a subject has a low serum TSH value between 0.1 and 0.4mU/L and is not on thyroxine therapy, then the first step is to repeat the measurement together with free T₄ and free T₃ to exclude overt hyperthyroidism (and also central hypothyroidism) within one or two months. In most circumstances serum TSH will have returned to within the reference range. If the repeat serum TSH measurement remains between 0.1-0.4mU/L with normal free T₄ and/or free T₃ concentrations, then repeat testing every 12 months is all that is required. In those subjects with a serum TSH less than 0.1mU/L, free T₄ and free T₃ should be measured to exclude overt hyperthyroidism. Usually, subclinical hyperthyroidism will be confirmed, and although no consensus exists, it has been strongly argued that further diagnostic testing is warranted (21,23).

3. The Treatment
Treatment of overt hypothyroidism with T₄ is cheap and usually effective. The risks of T₄ therapy include the long-term consequences of inadequate or over-treatment with T₄. There is considerable evidence for poor compliance with T₄ therapy and community studies consistently show that 50% of treated hypothyroid patients have serum TSH levels either above or below the reference range (7,9,24). There has been some concern that T₄ given in doses that suppress serum TSH to undetectable concentrations might promote osteoporosis (25-27) or atrial fibrillation (28). In a recent observational 10-year cohort study from Birmingham, a low serum TSH (<0.05mU/l) but not a raised serum TSH was associated with an increased risk of all cause mortality and cardiovascular mortality (29).

Evidence that T₄ therapy to normalise serum TSH in subjects identified with subclinical hypothyroidism at screening improves symptoms or cardiovascular outcomes is lacking (21). There is observational data suggesting that subclinical hypothyroidism during pregnancy may be associated with sub-optimal intellectual performance in children but these are based on relatively small numbers of cases (30). Some studies have even suggested that the maternal serum free T₄ level is more sensitive than the serum TSH in predicting the likelihood of adverse intellectual outcomes in the offspring (31). No intervention data on the effect of T₄ therapy in pregnancy exist. Most clinicians (at the nanolevel) treat those patients who have both raised serum TSH concentrations and positive thyroid-antibody tests, even if symptoms are absent, provided that no contraindication is present, in view of the annual risk of developing hypothyroidism of approximately 5% (10). If serum TSH alone is raised the annual risk of developing hypothyroidism is approximately 3% per year. The higher the serum TSH level, the greater is the prognostic significance for the development of overt hypothyroidism. No consensus exists regarding the treatment of subclinical hyperthyroidism but any potential benefits of therapy must be weighed against the substantial morbidity associated with the treatment of thyrotoxicosis.
4. The Screening Programme
No high quality randomised controlled trials of a complete screening programme for thyroid disorders exist. A case can be argued from a cost-utility analysis using a computer decision model that has assessed the consequences and costs of including serum TSH screening with cholesterol screening (19,32). It concluded that testing women aged 35 years and older with repeat serum TSH every five years for 50 years would be beneficial. The cost-effectiveness of screening for mild thyroid failure or subclinical hypothyroidism was comparable with that of other commonly performed preventive and therapeutic health practices such as hypertension, exercise, breast cancer screening and oestrogen replacement therapy in women. The costs of identifying and treating subclinical hyperthyroidism were not considered in this study.

Healthy subjects are exposed to adverse effects of screening without a guarantee of benefit and may be harmed by a screening programme. Subjects identified as having subclinical thyroid disease may suffer from the labelling effect described among hypertensive patients (33). Other costs include further investigations in those with borderline results and those who are falsely reassured and therefore fail to realise the significance of symptoms occurring later. Subjects with positive screening tests do not always comply with treatment. It is also essential that any screening programme be evaluated within a randomised controlled trial that has been designed with mortality as the outcome (5).

CONCLUSION
Screening for thyroid disease is clearly warranted in certain patient groups. Case-finding for thyroid disease in women older than 50 years or if visiting a doctor in primary care with non-specific symptoms might be justified in view of the high prevalence of mild thyroid failure. There is an urgent need for long–term studies of the effects of treatment of both subclinical hypothyroidism and hyperthyroidism, to determine if there is indeed benefit from screening for thyroid dysfunction in adults.