When Does Subclinical Hypothyroidism Warrant Levothyroxine? What the TRUST Trial Showed

A mildly elevated thyrotropin (TSH) with a normal free T4 is one of the most common lab abnormalities a person will ever see on a report, and in most cases it does not warrant levothyroxine. The TRUST trial, a randomized, double-blind, placebo-controlled study of 737 adults over 65, tested exactly this question and found that lowering an elevated TSH with medication produced no measurable improvement in hypothyroid symptoms or fatigue. Pooled data from two randomized trials later showed no cardiovascular benefit either. The number on the page turned out to matter far less than what treating it actually did for the people in the trial.

What "subclinical" actually means

Subclinical hypothyroidism is a biochemical picture, not a felt illness. It describes an elevated serum TSH, conventionally above roughly 4.5 mIU/L, paired with a free T4 that still sits inside the normal range. The Cleveland Clinic Journal of Medicine review divides it into two grades: grade 1, with a TSH of 4.0 to 10.0 mIU/L, and grade 2, with a TSH above 10.0 mIU/L. Roughly 90 percent of cases are grade 1, the mild end, and that is precisely the group where the decision to treat has been most contested.

Two facts complicate the picture before any drug is prescribed. First, TSH drifts upward with age, so a value that looks abnormal against a young-adult reference range may be unremarkable for someone in their late seventies. Second, a mildly high TSH often normalizes on its own when repeated weeks or months later, which means a single reading can send someone toward lifelong medication for a lab blip.

What the TRUST trial actually measured

TRUST enrolled 737 adults aged 65 and older with a TSH between 4.6 and roughly 20 mIU/L and randomly assigned them to levothyroxine or an identical placebo. The two co-primary outcomes were deliberately patient-centered: a hypothyroid symptoms score and a tiredness score, both drawn from a validated thyroid quality-of-life questionnaire.

The drug did what it is designed to do biochemically. The mean TSH fell from about 6.40 mIU/L at baseline to 3.63 mIU/L in the levothyroxine group at one year, while the placebo group drifted down to 5.48 mIU/L on its own, a reminder of how often these values regress without intervention. Yet the symptom and tiredness scores moved essentially the same amount in both groups. Normalizing the number did not translate into feeling better. Secondary measures, including handgrip strength, cognitive function, and blood pressure, showed no meaningful separation between the groups either.

This is the central lesson: TRUST was not a study that failed to find a treatment. It was a well-powered study that found the treatment did not help the outcomes patients care about, in the population studied.

What the pooled cardiovascular data added

Symptoms are one question; hard cardiovascular events are another. Observational studies had long hinted that a higher TSH tracks with cardiovascular risk, which fueled an argument that treatment might protect the heart even if it did not lift fatigue. A 2021 pooled analysis in Frontiers in Endocrinology combined individual patient data from TRUST and the IEMO 80-plus trial, 842 older adults with a median age of 75, to test that idea directly.

The result was neutral. Levothyroxine did not significantly reduce cardiovascular events, atrial fibrillation, or heart failure, and the neutral pattern held whether or not participants had prior cardiovascular disease and across age groups. An observed association in a population, in other words, did not convert into a benefit when the exposure was assigned at random. That gap between correlation and randomized effect is one of the most reliable reasons a plausible treatment fails to pan out, and it is why guideline writers weight trial evidence over epidemiology.

Why the guideline threshold is an evidence threshold

Reading TRUST and the pooled data together, professional bodies moved away from treating a number and toward treating a situation. The European Thyroid Association favors a wait-and-see approach for patients over 80 with a TSH at or below 10 mIU/L, and recommends age-specific reference ranges rather than a single young-adult cutoff. The American Thyroid Association similarly frames the decision around individual factors, symptoms, thyroid antibodies, and cardiovascular history, rather than universal treatment, and suggests relaxing the TSH target upward in older patients.

The grade 2 cutoff of 10 mIU/L functions as the practical dividing line in most guidance, not because something magical happens at that value, but because the randomized trials that found no benefit were populated mostly by grade 1 patients. The threshold marks the edge of where the reassuring evidence applies, not a proven point of harm below it. Younger adults, people with a persistently rising TSH, those with clear symptoms or positive thyroid antibodies, and women who are pregnant or seeking pregnancy sit outside the TRUST population and are reasonably weighed differently.

This article is educational and is not medical advice; whether to start or stop any medication is an individual decision to make with your own clinician. The value of the randomized evidence is that it replaces a reflexive impulse to correct an abnormal lab with a fair question: does treatment change how this specific person does, and at what cost of pills, monitoring, and the small but real risk of overtreatment. For most older adults with a mildly high TSH, the honest answer from TRUST is that it does not.