TSH 0.45 to 4.12 mIU/L

The first instruction in the AACE/ATA guideline is not a number. Recommendation 14.1: "The reference range of a given laboratory should determine the upper limit of normal for a third generation TSH assay. The normal TSH reference range changes with age. If an age-based upper limit of normal for a third generation TSH assay is not available in an iodine sufficient area, an upper limit of normal of 4.12 should be considered." So 4.12 is the fallback, and the 0.45 lower bound comes from the same guideline's treatment-target recommendation, which names "the NHANES III reference population range of 0.45–4.12" when a lab prints no limits. That reference population excluded people with thyroid disease, thyroid antibodies, pregnancy, and estrogen, androgen or lithium use.

Why the top of the range is argued about. The guideline says the 4.12 limit "has not been universally accepted," and reports proposals to lower it to 2.5 or 3.0 because the distribution is skewed and the median sits near 1.5. It also gives the case against: TSH "measurements vary up to 50% above (78) and below the mean on a given day," so a 2.5 cutoff "would lead to more than 10 million additional diagnoses of hypothyroidism in the United States per year" without clear benefit, and "there are virtually no clinical outcome data to support treating patients with subclinical hypothyroidism with TSH levels between 2.5 and 4.5 mIU/L." Recommendation 27 closes the question for treated patients: outside pregnancy, "the evidence does not support targeting specific TSH values within the normal reference range." There is no optimal TSH inside the interval, and this page does not print one.

Age moves the line. From the NHANES III analysis the guideline cites, "For every 10-year age increase after 30–39 years, the 97.5th percentile of serum TSH increases by 0.3 mIU/L," and among people over 80 without thyroid antibodies, 12 percent had a TSH above 4.5. The guideline's conclusion: "very mild TSH elevations in older individuals may not reflect subclinical thyroid dysfunction, but rather be a normal manifestation of aging." The 2014 ATA treatment guideline goes further for people already on levothyroxine, calling it "reasonable to raise the target serum TSH to 4–6 mIU/L in persons greater than age 70–80 years."

What a normal result settles, and what it does not. The guideline calls TSH "the single best screening test for primary thyroid dysfunction for the vast majority of outpatient clinical situations," and the 2016 ATA hyperthyroidism guideline says a normal TSH "precludes the diagnosis of thyrotoxicosis" absent a pituitary tumor, hormone resistance or assay interference. The exceptions are pituitary disease, where Recommendation 12 says to use free T4 rather than TSH, and pregnancy, where the range shifts (the 2017 ATA pregnancy guideline's fallback upper limit is about 4.0 mU/L, with a 2026 revision now published that we have not yet read). Variation of 40 to 50 percent within the range between draws "do not necessarily reflect a change in thyroid status." On who gets tested at all, the guideline says screening "should be considered in patients over the age of 60" and records that the bodies disagree, with the USPSTF not recommending routine screening.