What each test measures, and where patterns get complicated

TSH, free T4, total T4, total T3, and the thyroid antibody tests (thyroid peroxidase antibody, TPOAb; TSH-receptor antibody or thyroid-stimulating immunoglobulin, TRAb/TSI; thyroglobulin antibody) each answer a different question. TSH screens the axis. Free T4 confirms and quantifies a peripheral abnormality. Total T3 helps when a hyperthyroid pattern is T3-predominant, meaning TSH is suppressed but FT4 is still normal. Antibody tests support an autoimmune cause once a functional pattern is already established; they indicate what is detected and when they are added, while interpreting antibody titers is a separate topic.

Use the pattern matrix as the enumeration. The decision is to first ask whether TSH and free hormone move in a physiologically expected direction; when they do not, investigate central disease, illness, medication, timing, binding-protein effects, or assay interference rather than forcing a pattern label.

Total T4 and total T3 measure hormone bound to carrier protein plus the free fraction. Anything that changes TBG concentration or binding affinity, pregnancy, estrogen therapy, nephrotic syndrome, liver disease, familial TBG variants, some drugs, changes total hormone without necessarily changing the free, biologically available fraction. That is a source of apparent disagreement between total and free results that has nothing to do with thyroid gland function.

Most routine free T4 immunoassays are indirect, analog methods performed in the intact serum matrix rather than by physical separation. They are more susceptible to binding-protein and matrix disturbance than the reference procedure, equilibrium dialysis or ultrafiltration paired with isotope-dilution mass spectrometry. Different FT4 platforms can disagree, particularly in pregnancy or severe binding-protein derangement, which is one reason a laboratory sometimes compares platforms when a result does not fit.

In pregnancy, the 2017 American Thyroid Association (ATA) guideline recommends a laboratory- and trimester-specific TSH reference interval derived from a local iodine-sufficient, TPOAb-negative pregnant population. Where a validated local range is unavailable, the guideline discusses an approximate upper TSH limit near 4.0 mIU/L, mainly relevant to the late first trimester, gradually returning toward the nonpregnant range later in gestation. This replaced the fixed 2.5 and 3.0 mIU/L trimester cutoffs from the 2011 guideline; postpartum, TSH is generally interpreted with the nonpregnant interval, and postpartum thyroiditis can produce a transient hyperthyroid phase followed by a hypothyroid phase.

Nonthyroidal illness syndrome (NTIS, sometimes called sick euthyroid syndrome) produces thyroid-test abnormalities in acutely or chronically ill patients without primary thyroid-gland disease. The classic pattern is a low T3 with a normal TSH and normal FT4; with greater illness severity, FT4 and then TSH may also fall, and a TSH rebound is sometimes seen during recovery. A markedly elevated TSH is not typical of NTIS and favors true primary hypothyroidism instead. A pattern that does not match one of the named boxes is a reason to look for pregnancy, illness, or an assay event before it is a reason to force a diagnosis.