TSH under 0.45 mIU/L

A TSH below your lab's lower limit (about 0.4 to 0.45 mIU/L, depending on the assay) is the screening signal for an overactive thyroid, and the 2016 ATA hyperthyroidism guideline is the document that governs it. Its first move is to add two tests: "when thyrotoxicosis is strongly suspected, diagnostic accuracy improves when a serum TSH, free T4, and total T3 are assessed at the initial evaluation." If free T4 or T3 is high, that is overt hyperthyroidism, in which TSH is usually under 0.01. If both are normal, it is subclinical hyperthyroidism, and the guideline splits that by depth: "suppressed" under 0.1 mU/L, "low" between 0.1 and 0.4. In a representative US sample, 0.7 percent of adults had a TSH under 0.1 and 1.8 percent had one under 0.4.

Rule out the things that are not thyroid disease first. The guideline's differential for an isolated low TSH is "exogenous thyroid hormone use, nonthyroidal illness, drug effects, and pituitary/hypothalamic disease, all of which need to be ruled out." It adds that mean TSH runs lower in Black non-Hispanic Americans, "some of whom may have slightly low TSH levels without thyroid disease," and that some healthy older people carry a low TSH with low-normal hormones and no disease. The AACE/ATA hypothyroidism guideline lists the same pitfalls from its side: TSH is suppressed in acute illness, by dopamine and high-dose glucocorticoids, by octreotide, and "typically falls, but infrequently to below 0.1 mU/L, during the first trimester of pregnancy." The 2017 ATA pregnancy guideline puts numbers on that: a TSH below 0.4 "is observed in as many as 15% of healthy women during the first trimester." And one supplement matters: the 2016 guideline says high-dose biotin "causes spuriously low results" for TSH and falsely high free T4, and tells patients to stop biotin and repeat "at least 2 days later."

Then repeat, and wait. The 2016 guideline: "it is important to document that SH is a persistent problem by repeating the serum TSH at 3–6 months, prior to initiating therapy," except in people at high risk of complications, who are retested "within 2–6 weeks." The reason is that the natural history "is variable," with 5 to 12 percent a year reverting to normal on their own; in one study half had a normal TSH when next checked. A low TSH on repeated measurement over 3 to 6 months "is considered persistent, effectively ruling out transient thyroiditis."

Who is treated. Recommendation 73: "When TSH is persistently <0.1 mU/L, treatment of SH is recommended in all individuals ≥65 years of age; in patients with cardiac risk factors, heart disease or osteoporosis; in postmenopausal women who are not on estrogens or bisphosphonates; and in individuals with hyperthyroid symptoms." Under 0.1 without those factors, treatment "should be considered" (Recommendation 74). For a TSH between 0.1 and the lower limit, treatment is considered at 65 and older or with cardiac disease, osteoporosis or symptoms (Recommendation 75), while asymptomatic people under 65 without those "can be observed without further investigation" (Recommendation 76), monitored every 6 to 12 months. The guideline is candid that "Treatment of SH is controversial, since few intervention studies showing benefit have been performed." The risk it is weighing: pooled data on 52,674 people found subclinical hyperthyroidism carried a 24 percent higher overall mortality, with cardiovascular risk higher under 0.1 than between 0.1 and 0.4. When treatment happens, the goal "is to render the patient euthyroid with a normal TSH," by the same methods as overt disease, and for hyperthyroid symptoms the guideline says "a trial of b-adrenergic blockers may be useful"; it recommends beta blockade outright only in symptomatic overt thyrotoxicosis.