What is the preferred initial test for investigating thyroid function, particularly when measuring T4 (thyroxine) and T3 (triiodothyronine) levels in patients with suspected thyroid disease?

TSH is the Preferred Initial Test for Thyroid Function Investigation

When investigating thyroid function, TSH (thyroid-stimulating hormone) should be the first-line test, with free T4 measured subsequently if TSH is abnormal, and T3 reserved only for specific clinical scenarios where hyperthyroidism is suspected despite normal TSH and free T4. 1, 2

Rationale for TSH as Initial Test

• TSH has superior diagnostic performance with 98% sensitivity and 92% specificity when used to confirm suspected thyroid disease in patients referred to specialty clinics 1
• TSH is the most sensitive index of thyroid gland failure, becoming elevated before T4 drops in primary hypothyroidism 3, 2
• TSH measurement constitutes an indispensable test that more reliably establishes thyroid status than direct hormone measurements 4

Algorithmic Approach to Testing

Step 1: Measure TSH First

• Order TSH as the initial screening test for suspected thyroid dysfunction 2
• TSH values below 0.1 mU/L are considered low; values above 6.5 mU/L (or 4.5 mU/L by some definitions) are considered elevated 1

Step 2: Add Free T4 if TSH is Abnormal

• If TSH is elevated: Measure free T4 to distinguish subclinical hypothyroidism (normal T4) from overt hypothyroidism (low T4) 1
• If TSH is suppressed: Measure free T4 to confirm hyperthyroidism (elevated T4) versus subclinical hyperthyroidism (normal T4) 1
• Free T4 measurements have superseded total T4 due to higher diagnostic performance and independence from binding protein variations 5, 4

Step 3: Measure T3 Only in Specific Situations

• Order T3 only if: TSH is undetectable AND free T4 is normal, to detect T3 thyrotoxicosis (occurs in approximately 5% of hyperthyroid patients) 3, 2
• T3 measurement is more appropriate for diagnosing hyperthyroidism specifically, not hypothyroidism 4
• Circulating T3 is less reliable than T4 because 80% is produced extrathyroidally from T4 deiodination, which can be influenced by various non-thyroidal conditions 5

Why T4 is Preferred Over T3 for Routine Testing

• T4 more accurately reflects thyroid hormone production: The thyroid gland secretes 80% T4 and only 20% T3, with most circulating T3 derived from peripheral conversion 5
• T4 is more stable: T4 shows narrower individual variation over time compared to T3, even in patients with thyroid disease 6
• T3 is less diagnostically reliable: T3 conversion from T4 is affected by non-thyroidal illness, medications, and nutritional status, making it less specific for thyroid dysfunction 5
• Free T4 more correctly identifies hypothyroid patients, while free T3 is more appropriate specifically for hyperthyroidism diagnosis 4

Critical Exceptions Where TSH Cannot Be Used

In these situations, measure free T4 (and possibly T3) directly, as TSH is diagnostically misleading: 5

• Central (secondary/tertiary) hypothyroidism due to pituitary or hypothalamic dysfunction 1, 5, 2
• Unstable thyroid status: first months of thyroid treatment, altered levothyroxine dose, or subacute thyroiditis 5
• Suspected hypophysitis (presents with low TSH and low free T4, indicating central hypothyroidism) 1
• Monitoring adequacy of treatment in central hypothyroidism requires free T4 and T3, not TSH 2

Important Caveats

• Low positive predictive value in screening: When TSH is used for screening primary care populations (rather than confirming suspected disease), the positive predictive value is low, and interpretation is complicated by underlying illness or frailty 1
• Non-thyroidal illness interference: Hospitalized patients and those with severe non-thyroidal illness can have false positive TSH results 7
• Always confirm abnormal results: Repeat testing before initiating treatment to rule out laboratory error or transient abnormalities 7
• Binding protein interference: Total T4 and T3 measurements are affected by changes in thyroid hormone transport proteins (especially T4-binding globulin), making them unreliable without free hormone measurement 5, 4