How often to retest common markers

By Alina Radu, founder of Sera · November 10, 2026

Most guideline retest intervals apply after a change: a new dose, a new treatment or a first abnormal result. For thyroid, B12, HbA1c and lipids, that interval is usually 6 weeks to 3 months. Once things are stable, guidelines move to yearly or longer, and for ferritin and vitamin D in healthy adults they set no routine interval at all. The yearly check that many practices run is a convention, not a guideline.

I build Sera, a lab-review workspace for practitioners, and it sets retest dates. So I've kept this to what the guidelines say, with a note on where they stop and convention starts.

Guideline or convention

Every interval below falls into one of two groups, and it helps to know which one you're using.

  • Guideline interval: a published body such as NICE, the American Diabetes Association (ADA) or the American Thyroid Association (ATA) names a time for a stated situation. I opened each source and checked the wording.
  • Practice convention: a common habit, such as rechecking ferritin three months after an out-of-range result or running a yearly panel. It may be sensible. It isn't in a guideline.

Most guidelines are written for people with a diagnosis or on medication. Few say anything about retesting a healthy adult with an optimal result, and I've said so where that's the case.

Summary

MarkerSituationIntervalSource
TSHAfter a levothyroxine dose change4 to 6 weeksATA 2014
TSHOn levothyroxine, until stableEvery 3 months, then yearlyNICE NG145
TSH, free T4Untreated subclinical hypothyroidismYearly with raised antibodies, every 2 to 3 years withoutNICE NG145
TSHUntreated subclinical hyperthyroidismEvery 6 monthsNICE NG145
TPO antibodiesAfter the first positive testDo not repeatNICE NG145
Blood countTreating iron deficiency anemiaWithin 4 weeks, then 2 to 4 monthsNICE CKS
FerritinRoutineNo guideline interval foundBSG 2021
Vitamin DAfter starting treatment, some groups3 to 6 monthsNICE CKS
Vitamin DHealthy adultsRoutine testing not advisedEndocrine Society 2024
B12After starting replacementFollow-up at 3 monthsNICE NG239
HbA1cNormal screening resultEvery 3 yearsADA 2026, USPSTF
HbA1cPrediabetesAt least yearlyADA 2026, NICE PH38
LipidsAfter starting or changing treatment2 to 3 monthsNICE NG238
LipidsOn lipid-lowering treatmentYearlyNICE NG238

Thyroid: TSH, free T4 and TPO antibodies

Thyroid has the clearest intervals of any marker here.

After a levothyroxine dose change, the ATA guideline says TSH should be checked 4 to 6 weeks later. NICE NG145 says that for adults on levothyroxine, TSH can be measured every 3 months until it's stable, meaning two similar results within the reference range 3 months apart, and then once a year. It suggests adding free T4 when symptoms continue after starting treatment.

For adults with untreated subclinical hypothyroidism, NICE suggests TSH and free T4 once a year when there are signs of underlying thyroid disease, such as raised thyroid antibodies, and every 2 to 3 years when there aren't. Untreated subclinical hyperthyroidism gets TSH every 6 months. NICE also says not to repeat thyroid tests sooner than 6 weeks after the last one, even when symptoms change.

TPO antibodies are where guidelines and practice part ways most clearly. NICE says to consider measuring them when TSH is above the reference range, and then not to repeat the test. Some practices track antibodies yearly in Hashimoto's anyway. That's a practice convention, and it's good to know it goes against NICE if a client's GP pushes back.

For a healthy adult with a normal TSH, I found no guideline interval. Yearly is convention. The TSH, free T4 and TPO antibody range pages list the sources behind each band.

Ferritin and iron

The guidelines here are about iron deficiency anemia, and they monitor the blood count more than ferritin.

NICE's clinical summary on iron deficiency anemia says to recheck hemoglobin within the first 4 weeks of iron treatment, then the full blood count at 2 to 4 months. Once hemoglobin is normal, iron continues for 3 more months, and the blood count is then checked periodically, for example at 3, 6, 12 and 24 months.

The British Society of Gastroenterology guideline suggests a blood count every 3 months for a year and then every 6 months for 2 to 3 years. It also says there isn't enough data to recommend routine ferritin monitoring.

So a ferritin recheck 3 months after a low result, or after starting iron, is convention. It's a reasonable one, since that's roughly when the guideline blood count happens, but no guideline names it. The ferritin range page shows the bands and their sources.

Vitamin D

The guidelines lean against retesting. NICE's clinical summary on vitamin D deficiency says routine monitoring isn't necessary. It suggests considering a recheck 3 to 6 months after starting treatment in specific groups: people with symptoms, a malabsorption disorder, medicines that affect vitamin D, or where adherence is in doubt.

The Endocrine Society's 2024 guideline suggests against routine testing in generally healthy adults under 50, including follow-up tests to guide the dose.

A yearly vitamin D test is convention. If you do retest, compare like with like: a late-winter result and a late-summer result can differ for reasons of season alone. See the vitamin D range page.

Vitamin B12

NICE NG239 sets a first follow-up appointment 3 months after starting B12 replacement, or 1 month in pregnancy or breastfeeding. That's a review of symptoms, not necessarily a blood test. For people on intramuscular B12, NICE says not to repeat the initial diagnostic test. If a first result made deficiency unlikely but symptoms are still there 3 to 6 months later, NICE suggests considering a repeat.

For clients on long-term metformin, ADA's 2026 Standards of Care suggest checking B12 periodically, without naming an interval. Yearly is convention. The B12 range page has the bands.

HbA1c

HbA1c has intervals for almost every situation.

  • Normal screening result: ADA says repeat screening at least every 3 years is reasonable, sooner with symptoms or a change in risk such as weight gain. The US Preventive Services Task Force also says every 3 years may be reasonable, for adults aged 35 to 70 with overweight or obesity.
  • Prediabetes: ADA says to check at least once a year. NICE PH38 says the same for an HbA1c of 42 to 47 mmol/mol (6.0 to 6.4%).
  • Diabetes: ADA says at least twice a year, and about every 3 months after a treatment change or when goals aren't met.

So a yearly HbA1c for a client with a normal result is more often than ADA requires. Some practices do it anyway, alongside the rest of the panel. See the HbA1c range page.

Lipids

NICE NG238 says to measure a full lipid profile 2 to 3 months after starting or changing lipid-lowering treatment. After that, it offers a yearly lipid profile for secondary prevention and suggests considering one for primary prevention.

For adults with no condition and no treatment, the NHS Health Check, which includes a cholesterol test, is offered every 5 years between 40 and 74.

After a diet change, I found no guideline interval. Rechecking at about 3 months, to match the drug interval, is convention. The LDL range page has the bands and sources.

Keeping track across a caseload

Knowing the interval is the easy part. Remembering that a client is due is harder, and it falls apart quietly: nothing reminds you, and the client assumes you'll get in touch.

A simple setup works for a small practice:

  • A due-date list. One row per client and marker: last draw date, interval, due date, and why (guideline or convention). Sort by due date and look at it every week. A spreadsheet is fine for this.
  • A calendar reminder. When you review a result, put a reminder on the due date with the client's name and the markers. It takes a minute and doesn't depend on you remembering to open the list.
  • Count from each marker's own date. If a client retests vitamin D alone, that shouldn't move the date for their lipids.
  • Write the reason down. "TSH 3 months, NICE NG145, dose change in March" is something you can explain to a client or their doctor. "3 months" on its own isn't.

The weak point of both methods is that they rely on you updating them after every result. Miss one and the list is wrong without telling you.

Where Sera fits

Sera suggests a retest date for any result out of range, and a yearly one for the markers on a typical annual check-up list. You can set your own date for any marker, and it drafts the retest request. It reads lab PDFs and shows each value next to its page in the report, so you can check every result before it's saved. Trends run across years and labs, with diet, supplement and medication changes marked on the chart. It doesn't diagnose or recommend treatment, and the decision on when to retest stays with you.

It's free for three clients, so you can try it on the clients you most often lose track of. There's a 3-minute tour if you'd rather watch first.

Alina Radu is the founder of Sera and a software engineer. The idea started with her own Hashimoto's tests, scattered across years without the big picture.

More guides