Subclinical thyroid disease: when to investigate, when to treat, and when to wait

Once subclinical hypothyroidism is confirmed, the next step is to stratify by the TSH value, because the clinical significance is not uniform across the range. The individual-participant-data meta-analysis of the Thyroid Studies Collaboration, with more than 55,000 participants, associated a TSH equal to or above 10 mIU/L with higher risk of coronary heart disease and cardiovascular mortality; in the 4.5 to 9.9 mIU/L range, this increase in risk was not statistically significant (Rodondi et al., 2010). It is essential to read this finding as an association, not as proof of cause: the studies do not demonstrate that elevated TSH causes the events, nor that correcting it prevents them.

This gradation supports the practical threshold of 10 mIU/L as the treatment divide adopted by the guidelines. Above this value, and always with the finding confirmed on a new measurement, levothyroxine tends to be recommended; in the intermediate range of 4.5 to 9.9 mIU/L, the benefit is uncertain and the decision becomes individualized (Garber et al., 2012).

Two modifiers help individualize this gray zone. The first is anti-TPO positivity, which signals autoimmune thyroiditis and is associated with a higher rate of progression to overt hypothyroidism — an argument in favor of monitoring more closely or treating. The second is age: in older people, a slightly elevated TSH may be a physiological phenomenon of aging, and the upper reference limit tends to rise with age (Biondi and Cooper, 2008). Applying a single threshold to all ages leads to treating older adults unnecessarily.

In summary, the confirmed TSH defines the range; anti-TPO and age calibrate the decision within it. A young person with positive anti-TPO and an asymptomatic older adult with the same TSH of 6 mIU/L do not belong to the same management category.

The cautious stance in mild subclinical disease has well-defined exceptions. The most important is pregnancy and the intention to become pregnant. During gestation, period-specific TSH reference ranges are used, and the American Thyroid Association guideline considers the use of levothyroxine in subclinical hypothyroidism — especially when anti-TPO is positive — a situation in which the treatment threshold is lower than outside pregnancy (Alexander et al., 2017). Here, the anti-overtreatment reasoning yields to maternal-fetal protection, and what would apply to an older adult does not apply to a pregnant woman.

The second exception is a persistently very high TSH, equal to or above 10 mIU/L, confirmed on a new measurement — the range in which the associated risk is greater and in which the guidelines tend to recommend treatment (Rodondi et al., 2010; Garber et al., 2012).

The third involves unequivocal symptoms of hypothyroidism and documented progression, especially with positive anti-TPO. In these cases, a careful, reassessed therapeutic trial may be reasonable — provided that no outcomes are promised and there is a willingness to stop treatment if there is no objective benefit. What is not justified is automatically attributing every nonspecific complaint of tiredness to TSH.

Outside these situations, and without an evidence base, supplements should not be recommended as treatment for subclinical thyroid disease — iodine, selenium, glandular extracts, or "thyroid support" formulas. Excess iodine, in particular, can worsen thyroid function. Suggesting an intervention without robust evidence is the opposite of the rigor this topic demands.

Gathering what the evidence supports, management of subclinical thyroid disease fits into three verbs. INVESTIGATE always means confirming the finding with a new TSH measurement — with free T4 and anti-TPO when relevant — after 2 to 3 months, because TSH variability, nonthyroidal illness, and recovery from acute illness produce transient changes that should not be labeled or treated.

TREAT applies to defined scenarios: in hypothyroidism, a TSH persistently equal to or above 10 mIU/L; in qualified pregnancy, especially with positive anti-TPO and using gestational ranges; and in unequivocal symptoms with documented progression, with a reassessed therapeutic trial and no promise of result. In subclinical hyperthyroidism, treat when the TSH is persistently below 0.1 mIU/L in those over 65 years of age or in the presence of heart disease, osteoporosis, or symptoms.

WAIT and monitor is the appropriate management in the intermediate, mild, and low-risk ranges: subclinical hypothyroidism with a TSH of 4.5 to 9.9 mIU/L without modifiers, the mildly elevated TSH in the asymptomatic older adult, and mild subclinical hyperthyroidism outside the risk groups. In these cases, periodic reassessment replaces prescription, and "not treating now" is an active decision, not an omission.

The thread of this article is the discipline of distinguishing association from cause and of not promising what the evidence does not deliver. The large cohort studies and meta-analyses show that subclinical dysfunction is associated with cardiovascular, arrhythmic, and bone outcomes — but they do not prove that the abnormal test causes them, nor that treating it prevents them in mild subclinical disease. Less of the isolated test turned into a diagnosis, less levothyroxine prescribed for nonspecific symptoms, and more attention to the few scenarios in which treatment really changes management: that is how subclinical thyroid disease is practiced with rigor.