Col NF, Surks MI, Daniels GH. Subclinical Thyroid DiseaseClinical Applications. JAMA. 2004;291(2):239-243. doi:10.1001/jama.291.2.239
Author Affiliations: Division of Women's Health and Department of Medicine, Harvard Medical School, Brigham and Women's Hospital, Boston, Mass (Dr Col); Department of Medicine and Pathology, Montefiore Medical Center and the Albert Einstein College of Medicine, Bronx, NY (Dr Surks); and Thyroid Unit and Department of Medicine, Massachusetts General Hospital, Harvard Medical School, Boston (Dr Daniels).
Subclinical hypothyroidism and hyperthyroidism are diagnoses based on
laboratory evaluation with few if any clinical signs or symptoms. Subclinical
hypothyroidism is defined as an elevation in serum thyroid-stimulating hormone
(TSH) above the upper limit of the reference range (0.45-4.5 mIU/L) with normal
serum FT4 concentration; subclinical hyperthyroidism is defined
as a decrease in serum TSH below the reference range with normal serum FT4 and T3 concentrations. Though these conditions represent
the earliest stages of thyroid dysfunction, the benefits of detecting and
treating subclinical thyroid disease are not well established. Most persons
found to have subclinical thyroid disease will have TSH values between 0.1
and 0.45 mIU/L or between 4.5 and 10 mIU/L, for which the benefits of treatment
are not clearly established; treatment may be beneficial in individuals with
serum TSH lower than 0.1 mIU/L or higher than 10 mIU/L. This article illustrates
approaches to managing patients with subclinical hypothyroidism and hyperthyroidism
through 5 case scenarios that apply the principles of evidence-based medicine.
Because of the substantial uncertainty concerning the consequences of untreated
subclinical hypothyroidism and hyperthyroidism, as well as the benefit of
initiating treatment, patient preferences are important in deciding on management
of subclinical disease.
Subclinical hypothyroidism and hyperthyroidism represent the earliest
stages of thyroid dysfunction. Subclinical hypothyroidism is defined as an
elevation in serum thyroid-stimulating hormone (TSH) above the upper limit
of the reference range (0.45-4.5 mIU/L) with a normal serum free T4 (FT4) concentration; subclinical hyperthyroidism is defined as a decrease
in serum TSH concentration below the reference range, with normal serum FT4 and T3 concentrations. Most patients have few if any signs
or symptoms of thyroid dysfunction; therefore, it is a diagnosis based on
laboratory evaluation. Because the risk for subclinical thyroid dysfunction,
particularly subclinical hypothyroidism, increases with age,1 the
number of cases should increase as the US population ages.
Management of patients with thyroid dysfunction remains controversial
because the body of scientific evidence available to guide clinical decisions
is limited. Fundamental questions such as whom to screen and when to initiate
treatment remain largely unanswered. Based on the available data, withholding
treatment for individuals with serum TSH values that are slightly above or
below the reference range (4.5-10 or 0.1-0.45 mIU/L) likely poses no harm,
and initiating treatment likely poses no clear gains. Individuals with serum
TSH concentrations lower than 0.1 or higher than 10 mIU/L are more likely
to benefit from treatment, though some uncertainty remains. Subclinical thyroid
dysfunction predicts future progression to overt disease; however, TSH levels
in some individuals with subclinical hypothyroidism or hyperthyroidism return
to the reference range. Initiating treatment for subclinical hypothyroidism
does not alter the natural history of the disease but may prevent symptoms
and signs of overt disease.
In disorders for which the definition of the condition is imprecise
and the benefits and risks of treatment are not clearly established, patient
preferences play a critical role in treatment decisions. To make informed
decisions, patients should understand the benefits and risks of initiating
vs withholding treatment, as well as the expense and inconvenience of treatment
vs monitoring. Figure 1 and Figure 2 depict simplified clinical algorithms
for approaching subclinical hypothyroidism and hyperthyroidism.
Patient 1. A 28-year-old aerobics instructor
in excellent health had laboratory tests performed by her primary care physician
as part of a routine office visit. All results were normal, including complete
blood cell count and lipid profile, except for the serum TSH concentration
of 7.9 mIU/L.
Before making a treatment decision, guidelines recommend repeating the
serum TSH and measuring FT4 within 2 to 12 weeks, depending on
the clinical setting, to exclude transient forms of hypothyroidism. Transient
hypothyroidism is most commonly caused by destructive thyroiditis (including
painful subacute thyroiditis, silent subacute thyroiditis, or postpartum thyroiditis)
or recovery from severe nonthyroidal illness. The clinician should assess
the patient for symptoms and signs of hypothyroidism, including fatigue, lethargy,
slow cerebration, diminished sweating, dry skin, cold intolerance, dry hair,
weight gain, constipation, hoarseness, paresthesias, menstrual alterations,
and muscle pain. The thyroid gland should be examined carefully. The patient
should be asked about previous radioactive iodine treatment, thyroid surgery,
levothyroxine treatment, a family history of thyroid disease, and her lipid
profile should be determined.
Three months later, the patient's serum TSH was 6.8 mIU/L and FT4, 1.4 ng/dL (18 pmol/L); antithyroid peroxidase (TPO) antibodies were
absent. She was referred to an endocrinologist for evaluation and treatment
of subclinical hypothyroidism. The patient reported no symptoms of hypothyroidism.
She was taking no medications other than oral contraceptives, her menses occurred
regularly, and she was not intending to become pregnant. Her family history
included one maternal aunt with hypothyroidism and a maternal grandmother
with diabetes mellitus. Her examination findings were unremarkable and her
thyroid gland was palpable but not enlarged.
Routine measurement of anti-TPO levels is controversial. Although antibody
positivity predicts an increased risk of progressing to overt hypothyroidism
(2.6% per year if negative, 4.3% if positive),2 it
does not affect the effectiveness of treatment. Nonetheless, the increased
rate of progression may tip the balance toward treatment in some cases. The
presence of anti-TPO antibodies does predict an increased risk of miscarriage
as well as postpartum thyroiditis.3
Having confirmed the diagnosis of subclinical hypothyroidism, deciding
whether to initiate levothyroxine would depend on her serum TSH concentration,
the presence of any signs or symptoms suggestive of hypothyroidism, her risk
of progression to overt disease, and her preferences. There is no evidence
that levothyroxine treatment in healthy, asymptomatic patients with TSH between
4.5 and 10 mIU/L results in significant improvements in either quality of
life or clinical outcomes. Initiating treatment would prevent symptoms and
signs of hypothyroidism should this patient eventually progress to overt hypothyroidism.
The risk of progression to overt hypothyroidism is 2% to 5% per year. However,
serum TSH decreases to the reference range in a similar percentage during
the 2 to 4 years after elevated serum TSH is discovered.2 No
risks of delaying detection of overt hypothyroidism have been demonstrated
as long as individuals are carefully monitored.
In the absence of symptoms, the only other benefit of levothyroxine
treatment in this patient would be possible improvement in cardiac function.
Several small, unblinded studies suggest that subclinical hypothyroidism might
be associated with subtle declines in cardiac contractility.4 However,
evidence concerning the impact of subclinical hypothyroidism, treated or untreated,
on clinical cardiac end points is limited. Because of the lack of clear clinical
benefits, routine treatment with levothyroxine is not recommended in such
patients. There is no compelling evidence that low-density lipoprotein (LDL)
cholesterol is higher in individuals with serum TSH in the range of 4.5 to
After discussing the risks and benefits of treatment, this patient elected
not to be treated but to be followed with annual determination of serum TSH
and to be alert for signs and symptoms of overt hypothyroidism.
If she were contemplating pregnancy, treatment with levothyroxine should
be started, and serum TSH should be restored to the reference range. Although
some data suggest that withholding treatment may result in an increased risk
of fetal loss and neuropsychological complications in the offspring,5,6 there is no compelling evidence that
levothyroxine decreases the risk of miscarriage or improves the neuropsychiatric
complications. The requirement for levothyroxine in treated hypothyroid patients
often increases during pregnancy and serum TSH should be monitored each trimester.
Patient 2. During an annual evaluation, a healthy
70-year-old woman complained of mild fatigue, dry skin, and constipation.
Physical examination results were normal, including a nonpalpable thyroid
gland and normal relaxation phase of deep tendon reflexes. The serum TSH was
8.1 mIU/L; serum total cholesterol, 215 mg/dL (5.57 mmol/L); high-density
lipoprotein (HDL) cholesterol, 47 mg/dL (1.22 mmol/L); LDL cholesterol, 148
mg/dL (3.83 mmol/L); and triglycerides, 78 mg/dL (0.88 mmol/L). Repeat testing
2 months later revealed a serum TSH of 8.3 mIU/L and an FT4 of
1.4 ng/dL (18 pmol/L).
Up to 20% of women older than 60 years have subclinical hypothyroidism;
75% of those have serum TSH between 4.5 and 10 mIU/L.1,7,8 Routine
treatment with levothyroxine is not recommended because data are insufficient
to link this degree of hypothyroidism with any adverse health outcomes, and
no clear benefits of treatment have been demonstrated. However, because this
patient has symptoms that may be associated with hypothyroidism, levothyroxine
treatment may be considered. Although one study found patients with subclinical
hypothyroidism to have more symptoms than euthyroid patients,1 other
studies found no differences between patients and euthyroid controls.9- 11 There are no data
clearly demonstrating that treatment will improve symptoms in patients with
elevated serum TSH concentrations higher than 4.5 but lower than 10 mIU/L.
While there appear to be no adverse effects of initiating levothyroxine treatment
in this setting, inadvertent overtreatment occurs in about 20% (range, 14%-21%)12,13 of levothyroxine-treated patients,
carrying the potential risks of osteoporosis and atrial fibrillation when
serum TSH falls below 0.1 mIU/L. Treatment also involves the costs and inconvenience
of taking a daily medicine for the rest of one's life. Although follow-up
diagnostic testing to adjust dosage is also necessary, periodic thyroid function
tests are also necessary in individuals who are not treated.
Patient 3. A 58-year-old sedentary, obese man
with well-controlled hypertension, type 2 diabetes mellitus, and hypercholesterolemia
who presents for medical clearance before embarking on an exercise program
is found to have a serum TSH of 12 mIU/L. Repeat testing 2 months later reveals
a serum TSH of 14 mIU/L, FT4 of 1.3 ng/dL (17 pmol/L), and total
serum cholesterol of 235 mg/dL (6.09 mmol/L). Current medications included
metformin, 850 mg twice each day, hydrochlorothiazide, 25 mg/d, and pravastatin,
20 mg/d. He reports occasional constipation and fatigue. His physical examination
findings are unremarkable other than obesity (body mass index of 35).
Among patients with a serum TSH higher than 10 mIU/L and normal serum
FT4 who have signs and symptoms possibly consistent with hypothyroidism,
there is suggestive evidence supporting treatment with levothyroxine. Potential
benefits include a lowering of serum total and LDL cholesterol concentrations,
an improvement in cardiac function, and a possible improvement in symptoms.
However, the study suggesting that levothyroxine treatment reduced cholesterol
levels among individuals with subclinical hypothyroidism did not compare treated
with untreated groups,14 and the small, unblinded
trials suggesting that treatment improved cardiac function examined intermediate
end points of uncertain clinical significance.4 No
blinded, randomized controlled studies have assessed the impact of levothyroxine
on important clinical cardiac end points. Because the risk of progression
to overt hypothyroidism may be higher in these patients than in patients with
serum TSH between 4.5 and 10 mIU/L, treatment may prevent the manifestations
and consequences of hypothyroidism in those who would have progressed.
After discussing the potential benefits of treatment on his serum cholesterol
concentration, exercise tolerance, and constipation, the patient decided to
begin treatment with levothyroxine.
Patient 4. A 36-year-old healthy woman was
found to have a serum TSH concentration of 0.26 mIU/L, which is in the low
(<0.45 mIU/L) but detectable (>0.1 mIU/L) range. She has no personal or
family history of thyroid disease. She has one child and has monthly menstrual
periods. Her examination results were unremarkable, and her thyroid gland
was not palpable.
When the serum TSH concentration is found to be low but detectable,
the assay should be repeated along with a serum FT4 and total or
free T3 concentration within several months, sooner if any cardiac
signs or symptoms are present such as atrial fibrillation or palpitations.
If free thyroid hormone concentrations are within their reference ranges and
serum TSH remains low, the following causes of low serum TSH should be excluded:
treatment with levothyroxine, high-dose glucocorticoid or dopamine therapy,
severe nonthyroidal illness, or pregnancy. Although a low serum TSH may also
be due to hypothalamic or pituitary disease or anorexia nervosa, FT4 is usually low in these situations.
Patients with serum TSH levels between 0.1 and 0.45 mIU/L infrequently
progress to overt hyperthyroidism, defined as serum TSH lower than 0.1 mIU/L
and elevated concentrations of FT4 and/or FT3. However,
the rate of progression varies according to the underlying etiology. Patients
with large autonomously functioning adenomas (>3.0 cm diameter) or toxic multinodular
thyroids are at greater risk for progression to overt hyperthyroidism, especially
when exposed to high concentrations of iodine, most commonly after treatment
with amiodarone or radiocontrast agents.
Routine treatment is likely not beneficial in young asymptomatic persons
because there are no data that this condition is associated with adverse health
outcomes. Additional studies such as a 24-hour radioiodine uptake and radionuclide
thyroid scan to determine the etiology of the subclinical hyperthyroidism
are usually not necessary in asymptomatic individuals with serum TSH concentrations
in this range, if therapy is not being considered. In older patients, however,
treatment should be based on clinical judgment because of a possible association
of subclinical hyperthyroidism with increased cardiovascular mortality and
osteoporosis and the higher risk of progression to hyperthyroidism.
Patient 5. An active 77-year-old woman with
a history of myocardial infarction and osteoporosis feels well other than
experiencing rare exertional angina. She takes atenolol and atorvastatin daily,
alendronate weekly. Her pulse is 75/min and she is normotensive. Her thyroid
gland is difficult to examine due to kyphosis but feels somewhat prominent.
Thyroid function tests reveal serum TSH concentration lower than 0.01 mIU/L
and FT4 and FT3 within the reference ranges. A thyroid
ultrasound reveals a multinodular goiter. She has not received iodinated contrast
material in the past year.
Although other causes of low serum TSH should be sought, subclinical
hyperthyroidism due to a toxic multinodular goiter is the likely etiology.
A 24-hour radioiodine uptake and radionuclide thyroid scan is the most efficient
method to distinguish the various etiologies of hyperthyroidism. Hyperthyroidism
due to the various forms of destructive thyroiditis is transient and self-limited;
the 24-hour radioioidine uptake is close to zero when these conditions are
present. In contrast, individuals with Graves hyperthyroidism have normal
or elevated 24-hour radioiodine uptakes and a homogeneous pattern on radionuclide
scan. Autonomous adenomas appear as "hot" nodules on scans. Individuals with
toxic nodular goiters may have single or multiple hot areas but commonly have
a heterogeneous pattern of uptake.
Thyroid autonomy in a single-nodule or multinodular goiter increases
the likelihood of progression to subclinical or overt hyperthyroidism, especially
if patients receive excess iodine such as radiocontrast dyes or amiodarone.
However, nodules that do not concentrate radioactive iodine ("cold" nodules)
also occur in this setting and in some individuals with Graves disease. These
nodules may require a fine-needle aspiration biopsy.
Untreated subclinical hyperthyroidism with suppressed serum TSH (<0.1
mIU/L) carries the potential risks of atrial fibrillation,15 cardiovascular
mortality,16 and osteoporosis.17,18 Because
these risks are higher among patients older than 60 years, the balance is
shifted toward treatment among older patients. There are no studies that compare
the risk of atrial fibrillation in treated vs untreated patients, but bone
density is higher in treated compared with untreated postmenopausal women
with subclinical hyperthyroidism and decreased serum TSH.19,20 The
risks of treatment include allergic reactions (rash, fever, arthralgias, agranulocytosis,
hepatotoxicity, vasculitis) after antithyroid drug administration, and transient
worsening hyperthyroidism, permanent hypothyroidism, or worsening Graves ophthalmopathy
after therapy with iodohippurate sodium I 131.21
The definition of subclinical thyroid dysfunction is based on serum
TSH determination and is of necessity somewhat arbitrary. There is substantial
uncertainty concerning the consequences of untreated subclinical hypothyroidism
and hyperthyroidism, as well as the benefit of initiating treatment. Although
treatment may be beneficial in individuals with serum TSH lower than 0.1 mIU/L
or higher than 10 mIU/L, most persons found to have subclinical thyroid dysfunction
will have values between 0.1 and 0.45 mIU/L or between 4.5 and 10 mIU/L, for
which the benefits of treatment are not clearly established.
It is impossible to assess the merits of determination of serum TSH
to screen for occult subclinical disease without addressing the merits of
screening for occult overt thyroid disease, since both use the same test.
Any attempt to screen for overt disease will likely yield a far greater number
of cases of subclinical than overt disease. Unlike subclinical disease, the
benefits of detecting and treating overt thyroid disease are established.
Until clear therapeutic benefits are established for treating subclinical
thyroid dysfunction, general population screening for these conditions is
not recommended. However, the benefits of TSH determination to detect occult
thyroid dysfunction will be greater among those populations at higher risk
for developing overt disease, including women, older persons, and individuals
with previous or family history of thyroid disease, type 1 diabetes mellitus,
radioactive iodine treatment for hyperthyroidism, recurrent miscarriages,
or administration of medications that may affect thyroid function, such as
lithium carbonate or interferon. Vigorous case finding is recommended in these
populations. Among those found to have subclinical disease, patient preferences
are important in deciding on management.