🦋
Subclinical Hypothyroidism — Raised TSH with Normal Free T4 Confirm on repeat · TSH ≥10 vs <10 · treat / 6-month trial / monitor · NICE NG145 · GEMS / Red Whale
Progress 0 / 9
The full reasoning pathway — subclinical = raised TSH with normal free T4; confirm on a repeat at 3 months, then TSH level + age + symptoms decide treat vs trial vs monitor. Titrate carefully, advise on adherence, and safety-net.StartDecisionInvestigateActionReferStop / Admit
PresentationRaised TSH, normal free T4
Subclinical hypothyroidism. Always repeat TSH + free T4 after 3 months — only act if confirmed. Check anti-TPO antibodies if confirmed. Not for this pathway: pregnant / trying to conceive (treat & refer).
Step 1 · Confirm — is it really subclinical?Confirmed on 2nd test & T4 still normal?
~50% of mild elevations normalise. Exclude transient causes: recent non-thyroidal illness, recovery from treated hyperthyroidism, drugs, assay/biotin interference. Low TSH + low T4 → central (refer).
No — resolved
ReassureTSH normalised
No action; no ongoing monitoring needed unless features of thyroid disease.
Yes — confirmed
StratifyBy TSH level, age & symptoms
Branch on TSH ≥10 vs <10, age <65 vs ≥65, symptomatic vs not, antibody status.
Step 3 · treat · trial · monitor
TSH ≥10 (×2, 3m apart)
Consider levothyroxine
NICE: consider treatment (may help symptoms / long-term CVD; TSH >10 carries higher CHD risk). Start as in overt disease. Be more hesitant if >65y — over-suppression → AF/osteoporosis.
TSH <10, <65y & symptomatic
6-month trial of levothyroxine
Consider a 6-month trial. Stop if symptoms persist once TSH is normal — the elevation was likely not the cause. Treat earlier if antibody-positive or features of thyroid disease.
TSH <10, ≥65y or asymptomatic
Monitor (don't treat)
About half return to normal untreated. Recheck TSH+T4: annually if features of thyroid disease (antibodies +ve, prior surgery/RAI), otherwise every 2–3 years. Avoid attributing vague symptoms to it in older adults.
Step 7 · if treating
Step 7 · Action · levothyroxineStart & titrate to a normal TSH
Start as in clinical hypothyroidism: 1.6 µg/kg/day (≈100 µg at 60kg, 125 at 80kg, 150 at 100kg). If ≥65y or known CVD, start 25–50 µg/day and titrate. Take ≥30 min before breakfast, caffeine and other drugs. Recheck TSH ~3 months; aim for a normal TSH (don't over-suppress).
ReferEndocrinology / different pathway
Endocrinology pregnancy or trying to conceive, age <16, suspected central cause (low TSH + low T4), or diagnostic uncertainty.
Step 8 · adherence & modifiable factors
Step 8 · Adherence & modifiable factorsGet the most from a trial
Take levothyroxine consistently on an empty stomach, ≥30 min before food/caffeine, and separated from iron, calcium and PPIs (which impair absorption). Review drugs that raise TSH (amiodarone, lithium). General wellbeing measures for fatigue — sleep, activity, mood and iron/B12 status — since these often explain symptoms better than a mildly raised TSH.
Step 9 · monitoring & safety-net
Step 9 · Monitoring & safety-netRecheck & review the trial
Recheck TSH ~3 months after starting/adjusting; aim for a normal (not suppressed) TSH. Untreated: recheck annually if higher-risk (antibody-positive, prior surgery/RAI) or every 2–3 years otherwise. Stop the trial if symptoms persist once TSH is normal. In older adults, avoid over-treatment — watch for AF and osteoporosis; seek review if palpitations, weight loss or tremor develop.
⚠️ Confirm before you commit: a single raised TSH is not subclinical hypothyroidism — repeat at 3 months. In adults >65 be cautious: symptoms rarely correlate with mild TSH elevation, and over-treatment risks AF and osteoporosis.
1
Safety

Before You Label It "Subclinical" — Exclude the Things It Isn't

Subclinical hypothyroidism = raised TSH with a normal free T4. First make sure the picture isn't something that needs a different, more urgent pathway.

Overt hypothyroidism Raised TSH with a LOW free T4 — this is overt disease, not subclinical → treat with levothyroxine regardless of symptoms (use the Hypothyroidism pathway).
Pregnant or trying to conceive This pathway does NOT apply → lower treatment threshold, treat promptly and refer. Untreated disease risks miscarriage & fetal neurodevelopment.
Central (secondary) hypothyroidism Low/normal TSH with low free T4, ± headache/visual symptoms/other pituitary hormone loss → refer endocrinology; do not titrate on TSH.
Very high TSH or unwell TSH well above 10 with marked symptoms, or systemically unwell/intercurrent illness → treat as overt and assess clinically; don't just "watch and repeat".
Thyroid mass / compressive features Rapidly enlarging or fixed goitre, hoarseness, dysphagia, neck node → 2WW thyroid (NICE NG12), independent of the TSH.
Amiodarone / lithium / immunotherapy Drug-induced thyroid dysfunction can evolve quickly and needs monitoring with the initiating team — manage alongside, don't simply observe.
Child / young person <16 Different thresholds and causes → paediatric/endocrine referral rather than this adult pathway.
Suspected adrenal insufficiency Hypotension, hyponatraemia, fatigue, pigmentation — check before any thyroxine; levothyroxine can precipitate adrenal crisis if Addison's coexists.
The whole pathway hinges on the free T4 being normal — if it is low, the patient has overt hypothyroidism and the "consider/monitor" logic no longer applies. NICE NG145 explicitly carves out pregnancy and conception (managed on a separate, lower-threshold basis) because maternal hypothyroidism — even subclinical with positive antibodies — is associated with a 2–3× miscarriage risk and adverse neurodevelopment. Central hypothyroidism is the classic trap: the TSH is misleadingly low/normal, so treating "to a normal TSH" chronically under-replaces the patient.
2
Diagnose

Confirm It — Always Repeat the TSH After 3 Months

A single raised TSH is not a diagnosis. Only act once subclinical hypothyroidism is confirmed on a second test ~3 months later with the free T4 still normal.

Repeat at 3 months
Recheck TSH + free T4 after 3 months. About half of mild elevations have returned to normal by then. Only diagnose and consider acting if still raised with a normal T4.
Exclude transient causes
Recovery from any non-thyroidal illness (sick-euthyroid rebound), recent treatment of hyperthyroidism, recent thyroiditis (e.g. post-partum, subacute), and post-viral states can transiently raise TSH.
Drugs that raise TSH
Amiodarone, lithium, and some others. Review the medication list — manage with the initiating prescriber rather than reflexively starting levothyroxine.
Assay / biotin interference
High-dose biotin supplements and assay artefacts can distort results — ask the patient to stop biotin ~2 days before re-testing if there's a mismatch.
Time of day / TSH variation
TSH has diurnal variation (higher at night/early morning). Mild fluctuations around the upper reference limit are common and often not pathological.
Check anti-TPO antibodies
Once confirmed, check anti-TPO once — positivity raises the risk of progression to overt disease and shapes both the treatment decision and the monitoring interval (Step 3).
NICE NG145 (2019) is unambiguous: repeat the TSH (with free T4) after 3 months before acting, because roughly half of mildly raised TSH values normalise spontaneously. Treating on a single result commits a patient to lifelong therapy, repeated blood tests and the small but real harms of over-treatment for what may have been a transient blip. Excluding non-thyroidal illness is essential — TSH commonly overshoots above the reference range during recovery from any significant illness.
3
Diagnose

Risk of Progression — Antibodies & Features of Thyroid Disease

Confirmed subclinical hypothyroidism doesn't always progress — but some patients are far more likely to. This shapes whether to treat and how often to monitor.

Background rate
2–6% per year of people with subclinical hypothyroidism progress to overt hypothyroidism (NEJM 2017;376:2556).
Higher risk if…
Female, older, and anti-TPO antibody positive — these are the established risk factors for progression.
Anti-TPO positive
Marks underlying autoimmune (Hashimoto's) thyroid disease → higher progression risk → lower threshold to treat and monitor annually.
Other features of thyroid disease
Previous radioactive iodine treatment, thyroid surgery, or a goitre — all point to ongoing thyroid pathology and favour treating / closer monitoring.
Higher TSH = higher risk
The higher the confirmed TSH (and especially ≥10), the greater the chance of progression and of cardiovascular harm — this drives the Step 4 stratification.
Symptoms — interpret cautiously
Tiredness, low mood and weight change are common and non-specific. In older adults especially, symptoms correlate poorly with mild TSH elevation (see Step 9).
Antibody status is the most useful single modifier here: anti-TPO positivity roughly doubles the annual progression rate and is why NICE recommends annual (rather than 2–3-yearly) monitoring in these patients. Identifying who is likely to progress lets you treat the right people earlier while sparing low-risk patients from unnecessary lifelong therapy.
4
Diagnose

Stratify — The Three Inputs That Decide Management

Once confirmed, sort every patient by three things — TSH level, age, and symptoms — plus antibody status. These feed directly into the treat / trial / monitor decision.

TSH ≥ 10 mU/L
(Confirmed on 2 tests 3 months apart.) Higher progression and cardiovascular risk — NICE says consider levothyroxine (Step 5). The most clear-cut group to treat.
TSH above normal but < 10
The genuinely grey zone — NICE says "consider", individualised by the factors below. Most will be monitored; some symptomatic younger patients warrant a trial.
Age < 65
More likely to benefit and tolerate treatment → lower threshold for a trial if symptomatic.
Age ≥ 65
Be more hesitant. Over-suppression risks (AF, osteoporosis) rise with age and symptom–biochemistry correlation is poor → favour monitoring.
Symptomatic vs not
Symptoms attributable to hypothyroidism support a time-limited trial (if <65); asymptomatic patients are generally monitored.
Antibody / disease features
Anti-TPO positive or prior RAI/surgery shifts the balance toward treating and toward annual monitoring.
The TSH-10 cut-off is the single most important branch point: a JAMA 2010 meta-analysis (304:1365) found no overall increase in coronary heart disease in subclinical hypothyroidism except in those with TSH >10, who had increased CHD events and mortality. Age and symptoms then refine the decision in the sub-10 group, where the evidence for benefit is weakest.
5
Treat

Treat, Trial, or Monitor — The Core Decision

Match the confirmed picture to one of three actions. Treatment, when chosen, is started exactly as in clinical hypothyroidism.

TSH ≥ 10 (×2, 3m apart)
Consider levothyroxine NICE
NICE: consider treatment — may improve symptoms / long-term CVD risk (note evidence caveats, Step 9). More hesitant if >65y.
TSH < 10, age < 65 & symptomatic
6-month trial time-limited
Consider a 6-month trial of levothyroxine. Stop if symptoms persist once TSH is in range — the elevation likely wasn't the cause.
TSH < 10, ≥ 65y or asymptomatic
Monitor no treatment
About half normalise untreated. Don't treat; set a monitoring interval (Step 7). Avoid over-treatment harms.
Antibody +ve / prior RAI or surgery
Lower threshold to treat
Higher progression risk — favour treating (or a trial) and annual monitoring if not treating.

If starting levothyroxine — start just as in clinical hypothyroidism:

DoseEstimated full dose 1.6 µg/kg/day, rounded to the nearest 25 µg (≈100 µg at 60kg, 125 µg at 80kg, 150 µg at 100kg) for most adults.
≥65y / CVDStart low: 25–50 µg/day and titrate up as required — full doses can precipitate angina or arrhythmia in older / cardiac patients.
How to takeTake at least 30 minutes before breakfast, caffeine and other drugs (BNF). Separate from calcium, iron and PPIs.
TargetTitrate to a normal TSH — there is no benefit aiming for the low-normal end, and over-suppression raises AF / osteoporosis risk.
ℹ️ The hard part of subclinical disease is the <10, sub-65 group: a 6-month trial is a diagnostic test of symptom benefit, not an automatic commitment to lifelong treatment — set that expectation with the patient up front.
NICE NG145 recommends considering levothyroxine for confirmed TSH ≥10, and — for TSH below 10 — a 6-month trial only in those under 65 with symptoms, stopping if symptoms don't improve once the TSH is normal. The 1.6 µg/kg/day estimated-dose approach (with a cautious 25–50 µg start in the elderly/cardiac) is identical to overt disease because, once you've decided to replace, the pharmacology is the same. Framing the sub-10 trial as time-limited prevents the common error of indefinitely continuing levothyroxine that never actually helped.
6
Treat

Follow-Through — Titration, Review & When to Stop

Recheck TSH
Repeat TSH ~3 months after starting or any dose change (the pituitary response lags weeks behind a dose change — don't test sooner). Adjust in 25 µg steps to a normal TSH.
6-month trial review
At ~6 months, with the TSH in range, reassess symptoms honestly. If they haven't improved → stop levothyroxine (the elevation wasn't the cause) and return to monitoring.
If treatment continued
Once stable on a dose with a normal TSH, monitor annually — as for any patient on long-term levothyroxine.
Avoid over-treatment
Don't push the TSH below normal. A suppressed TSH (especially <0.1) raises the risk of atrial fibrillation and osteoporosis, particularly in older and post-menopausal patients.
Persistent symptoms, normal TSH
Look elsewhere — anaemia, low ferritin, B12/folate deficiency, menopause, depression, sleep apnoea, coeliac disease. Don't keep escalating thyroxine for non-thyroidal symptoms.
Adherence / absorption
If TSH stays high despite treatment: check timing relative to food/coffee, and interactions (calcium, iron, PPIs). Separate doses by ≥4 hours.
The 6-month "stop if no benefit" rule is what distinguishes good subclinical management from drift into unnecessary lifelong therapy — randomised trials (e.g. TRUST, NEJM 2017;376:2534) showed no meaningful symptom or quality-of-life benefit from treating subclinical hypothyroidism in older adults, so continuing a trial that hasn't worked is hard to justify. Over-replacement is not benign: suppressed TSH carries a roughly 3-fold AF risk and accelerates bone loss.
7
Lifestyle

If Not Treating — Monitoring Schedule & Counselling

Choosing to monitor is an active management plan, not "doing nothing". Set the interval by the risk of progression and safety-net the patient.

Features of thyroid disease → annually If anti-TPO positive, previous thyroid surgery or radioactive iodine, or a goitre — recheck TSH + free T4 every year (higher progression risk).
No features → every 2–3 years If no antibodies or other features of underlying thyroid disease, periodic checks every 2–3 years are sufficient (NICE NG145).
Re-test sooner if symptomatic Bring testing forward if the patient develops new or worsening symptoms of hypothyroidism — don't wait for the scheduled interval.
Counsel on progression Explain 2–6% per year progress to overt disease; being female, older or antibody-positive increases that risk. Most people remain stable or normalise.
Safety-net the symptoms Tell the patient which symptoms (fatigue, cold intolerance, weight gain, constipation, low mood) should prompt an earlier recheck.
Flag pregnancy plans If she becomes pregnant or plans to conceive, she needs prompt reassessment and treatment — the monitoring approach no longer applies.
Don't over-attribute symptoms Especially in older adults, resist pinning vague non-specific symptoms on a mildly raised TSH — look for other explanations too.
General cardiovascular health Address modifiable CVD risk (BP, lipids once euthyroid, smoking, activity) on its own merits rather than relying on thyroid treatment.
NICE NG145 sets the monitoring intervals explicitly: annually where there are features of underlying thyroid disease, otherwise every 2–3 years. A cross-sectional study of older adults (BJGP, doi:10.3399/bjgp20X708065) found that the presence or absence of symptoms was no different between euthyroid, subclinical hypothyroid and subclinical hyperthyroid groups — a powerful reminder to be cautious about attributing symptoms to mild biochemical abnormalities.
8
Refer

Referral — Who Needs More Than Primary Care

Most subclinical hypothyroidism is managed entirely in primary care. Refer for the specific situations below.

Endocrinology
Suspected central hypothyroidism (low/normal TSH + low free T4), diagnostic uncertainty, or atypical / unstable biochemistry.
Pregnancy / pre-conception
Pregnant or trying to conceive → treat promptly and arrange obstetric/endocrine input — separate, lower-threshold pathway with trimester-specific TSH targets.
Age < 16
Children and young people → paediatric / endocrine referral (different causes and thresholds).
2WW thyroid
Any suspected thyroid malignancy — fixed/rapidly enlarging goitre, hoarseness, dysphagia, neck node (NICE NG12), regardless of the TSH.
Complex drug effects
Amiodarone- or lithium-associated thyroid dysfunction — manage jointly with the initiating specialist.
Primary care manages
Confirming the diagnosis, the treat/trial/monitor decision, levothyroxine initiation & titration, and the monitoring schedule for the great majority of patients.
The referral list is deliberately short because subclinical hypothyroidism is a primary-care decision in almost all cases. The exceptions — central disease, pregnancy/conception, children, and possible malignancy — each need a fundamentally different approach (different targets, different urgency, or a different organ system) that a TSH-led primary-care algorithm isn't designed for.
9
Safety

Evidence Caveats & Safety-Netting

Hold the uncertainties in mind — they keep you from both over-treating and falsely reassuring.

Limited symptom benefit
Treatment trials (JAMA 2018;320:1349; NEJM 2017;376:2534) did not show significant improvements in symptoms or quality of life — temper expectations, especially in older adults.
Cardiovascular picture
Overall no increased CHD or mortality in subclinical hypothyroidism — except TSH >10, where CHD risk and mortality are increased (JAMA 2010;304:1365). Trials haven't shown treatment changes CVD risk.
Symptoms ≠ biochemistry
In ≥65s, symptom prevalence is no different across euthyroid vs subclinical groups — don't over-attribute (BJGP 20X708065).
Harms of over-treatment
Suppressing TSH (over-replacement) risks atrial fibrillation and osteoporosis — a real downside to treating people who won't benefit.
Re-test sooner if
New/worsening hypothyroid symptoms, or a new goitre/neck change. Bring the next TSH forward rather than waiting for the routine interval.
Switch pathway if
Free T4 falls below range (now overt) → treat. Pregnancy/conception → treat & refer. New compressive neck symptoms → 2WW.
Document the plan
Record the decision (treat / 6-month trial / monitor), the agreed interval, and the symptoms that should prompt earlier review — this is what good safety-netting looks like.
The honest evidence position — that treating subclinical hypothyroidism has not been shown to improve symptoms or quality of life, with benefit (if any) concentrated in younger, symptomatic, or TSH >10 patients — is exactly what NICE NG145's cautious "consider" language reflects. Communicating this uncertainty, setting a monitoring interval, and naming the symptoms that should trigger an earlier review is both good shared decision-making and the safety-netting that exams (and patients) reward.
Educational use only. Pathway based on: NICE NG145 Thyroid disease: assessment & management (2019, subclinical hypothyroidism in adults) · NICE CKS Hypothyroidism · NICE NG12 Suspected Cancer Recognition & Referral · GEMS (Guidelines & Evidence Made Simple) / Red Whale, Nov 2024 · BJGP 2016;66:538 & 20X708065 · Clinical Endocrinology 2016;84:799 · JAMA 2010;304:1365 · JAMA 2018;320:1349 · NEJM 2017;376:2534 & 2556. Not applicable to pregnant women or those trying to conceive. Always adapt to individual patient context and current guidance.