Subclinical Hyperthyroidism: A Low TSH the Guidelines Treat Differently by Degree
An 82-year-old woman with atrial fibrillation and osteoporosis has a mildly suppressed TSH and nothing else abnormal. The disagreement isn't whether a very low TSH deserves treatment — it's whether hers is low enough to count.
Josephine K., an 82-year-old retired librarian, has managed paroxysmal atrial fibrillation for six years on apixaban and metoprolol and has a T-score of -2.7 at her hip from a bone density scan done last spring. Neither problem is new, and neither has changed. What's new is a TSH of 0.28 μIU/mL on routine labs, found incidentally at a visit for her annual physical, with a normal free T4 and a normal free T3 — no clinical hyperthyroidism, no palpable goiter, no eye findings, nothing on exam to suggest an acute process. She feels entirely well and was surprised the result was even flagged.
Where her TSH sits inside the subclinical hyperthyroidism range matters more than the label itself. Guidelines treat a TSH persistently below 0.1 μIU/mL as a more clearly actionable finding in an older adult, and the pooled cohort data behind that line is specific: Collet and colleagues, working across five prospective cohorts for the Thyroid Studies Collaboration, found incident atrial fibrillation rising with the degree of suppression and concentrated below 0.10, while Blum and colleagues' meta-analysis in JAMA found the same gradient for fracture risk. Her value, 0.28, sits in the grayer 0.1 to 0.4 range, where the same literature is considerably less consistent about whether treating changes her actual outcomes, and where guidance leans toward confirming the finding is persistent, not immediately acting on a single value. That she already carries both risk factors the suppressed-TSH literature warns about — established AFib, established osteoporosis — cuts in two directions at once: it's the exact profile in which subclinical hyperthyroidism's downstream harms would land hardest if real, and it's also a population where those same diagnoses were already present before her TSH ever dropped, making it genuinely unclear how much of her risk this new finding is actually adding.
Deciding whether a first, mildly suppressed value needs treatment
I wouldn't treat her today. Collet's pooled cohort analysis puts the clearer actionable threshold below 0.10, and she's at 0.28 — a real but milder degree of suppression where the evidence for treatment changing outcomes is considerably less consistent. I'd confirm it's persistent with a repeat in two to three months before doing anything else.
I understand the threshold argument in a patient without her specific risk factors, but she already has atrial fibrillation and osteoporosis — the exact two outcomes the subclinical hyperthyroidism literature is worried about. I'd want to treat this more seriously than a threshold built for an average patient without either diagnosis already present.
You're applying a general population threshold to a patient who isn't average — the 0.1 cutoff comes from studies of populations with a mix of baseline risk, and a milder degree of suppression may matter more in someone who already has both downstream conditions than the same value would in someone with neither.
I'd hold to the confirm-first approach, but for a reason distinct from the threshold argument alone: she has no prior TSH on record, so we don't actually know whether 0.28 is a new finding or a longstanding, stable value that's simply never been checked before. Her existing AFib and bone disease genuinely raise the stakes of getting this right, which is itself an argument for confirming the trend before starting a treatment she'd likely stay on indefinitely, not an argument for skipping confirmation because the stakes feel higher.
Agreed: repeat TSH, free T4, and free T3 in two to three months to confirm persistence before starting any treatment.
Not agreed: what a confirmed, still-mild value in the 0.2 to 0.4 range would justify on its own. The cardiologist would treat a confirmed persistent value given her AFib and osteoporosis even without it crossing below 0.1; the endocrinologist would want to see it either drop further or produce some other objective change before treating outside the range the strongest evidence actually supports.