- 01Recent intake vs long-term state
A blood test needs fasting when it measures what the last meal put into the blood (glucose, insulin, triglycerides). Tests that reflect weeks to months (A1C, LDL, TSH, vitamin D) read the same after breakfast. A few tests (testosterone, cortisol) are timed to the morning for the body clock, not for food.
A blood test needs fasting when it measures what your last meal put into your blood. Glucose, insulin and triglycerides rise within hours of eating, so their tests need an empty stomach. A1C, LDL cholesterol, TSH and vitamin D reflect weeks to months of your body's work, and one breakfast barely moves them. That split explains nearly every fasting instruction on a lab order. A few tests, such as testosterone and cortisol, care more about the time of day than about food, so they are booked for the morning.
Which blood tests require fasting, and which don't
The table below covers the tests people are most often told to prepare for. "Fast" means no food, and no drink other than water, for the hours shown. When your clinician's or lab's instructions differ from this table, follow theirs: they know which method and cut-offs your lab uses.
| Test | Fast? | How long | Why |
|---|---|---|---|
| Fasting glucose | Yes | At least 8 hours | Glucose rises within an hour of eating, and the diabetes cut-offs assume at least 8 hours without calories 3 |
| Fasting insulin and HOMA-IR | Yes | Overnight, as your lab instructs | Insulin rises after food to clear glucose; HOMA-IR, a score for insulin resistance, is calculated from both fasting values |
| Oral glucose tolerance test (OGTT) | Yes | At least 8 hours, then the glucose drink | The test measures how you handle a set sugar load from a fasting start 3 |
| Triglycerides | Not usually | 8 to 12 hours if asked | Rises after a fatty meal; a fasting repeat is advised when a non-fasting result is high 1 |
| Total, LDL, HDL and non-HDL cholesterol, ApoB (a count of LDL-type particles), Lp(a) (an inherited lipoprotein) | No | None | These change by a few percent after a normal meal 1 |
| Hemoglobin A1c | No | None | Reflects roughly three months of glucose; any time of day is fine 3 |
| TSH and thyroid hormones | No | None | A meal has little effect; testing at the same time of morning each time makes results comparable. See reading a thyroid panel |
| Vitamin D (25-OH D) | No | None | Reflects weeks of intake and sun; one meal does not move it |
| Vitamin B12 and folate | No | None (some labs ask for a short fast) | A recent meal has little effect; follow your lab if it asks |
| Serum iron and transferrin saturation (how full your iron-carrying protein is) | Often preferred | Morning draw, as your lab instructs | Serum iron varies through the day and rises after iron-rich food or supplements; ferritin is far less affected. See iron studies explained |
| Testosterone | Preferred | Overnight, morning draw | Guidelines advise a morning fasting sample, and a repeat morning fasting sample to confirm a low result 6. See the male hormone panel |
| Cortisol | No | None | Cortisol peaks in the early morning, so it is drawn at the morning time your lab sets; a meal has little effect |
| CBC, electrolytes, kidney and liver tests | No | None | Food has little effect on these. See CBC explained |
| Comprehensive metabolic panel (CMP) | Often yes | 8 to 12 hours | The panel includes glucose, which needs fasting to be read against fasting cut-offs |
The markers behind the table
Why food moves some numbers and not others
After a meal, glucose rises within the first hour, and insulin rises with it to move that glucose into cells. Dietary fat is absorbed more slowly and travels as triglyceride-rich particles, so triglycerides climb over the following hours; the 2016 European consensus (EAS/EFLM) looked at changes up to about 6 hours after usual meals 1. LDL and total cholesterol are made and cleared over days, so a single meal hardly registers. In the Copenhagen General Population Study of about 33,000 people, lipid levels changed only minimally after normal food intake, and non-fasting values still predicted heart attacks and strokes 5. In about 200,000 people tested in Calgary, the time since the last meal changed calculated LDL by less than 10% and triglycerides by less than 20% 7.
| Marker | Largest average change after a usual meal |
|---|---|
| Triglycerides | Up about 26 mg/dL |
| Total cholesterol | Down about 8 mg/dL |
| LDL cholesterol | Down about 8 mg/dL |
| HDL cholesterol | No clinically relevant change |
Those figures come from the EAS/EFLM consensus 1. A1C works on a different clock. Glucose sticks to hemoglobin inside red blood cells, which live about three months, so A1C records an average across that span. That is why the American Diabetes Association lets A1C be measured at any time of day without fasting 3. For how each blood sugar number fits together, see blood sugar tests, explained.
The cholesterol test: why most labs stopped asking you to fast
For decades, a cholesterol test meant skipping breakfast. In 2016 the European Atherosclerosis Society and the European Federation of Clinical Chemistry and Laboratory Medicine (EAS/EFLM) reviewed the evidence and changed the default.
Fasting is not routinely required for determination of a lipid profile.Title of the EAS/EFLM joint consensus statement, Nordestgaard BG et al., European Heart Journal, 2016 1
The 2018 AHA/ACC cholesterol guideline reached the same place for US practice: in adults not taking a cholesterol-lowering drug, a fasting or non-fasting lipid profile is effective for estimating heart risk and baseline LDL 2. A non-fasting sample also reflects how your blood looks for most of the day, since people spend much of their waking time digesting a meal.
There are exceptions. AHA/ACC suggests a fasting repeat when non-fasting triglycerides are 400 mg/dL or higher 2, and EAS/EFLM sets the line at 440 mg/dL (5 mmol/L) 1. Triglycerides of 500 mg/dL or higher, and above all 1,000 or higher, raise the risk of pancreatitis and need prompt follow-up with your clinician, not just a fasting repeat 2. A fasting lipid panel also still makes sense for someone with a known inherited triglyceride disorder, someone recovering from pancreatitis caused by high triglycerides, or anyone whose clinician is tracking triglycerides specifically. Some US labs still ask everyone to fast because they estimate LDL with an older formula (the Friedewald equation) that works it out from your triglycerides, so it goes wrong when triglycerides are high after a meal. Newer equations handle non-fasting samples better; how to read your lipid panel explains which one your report uses.
CalculatorLDL calculatorRun your own panel through the newer LDL equation, which holds up better on a non-fasting sample.Work it out How long to fast before a blood test, and how long is too long
The American Diabetes Association defines fasting for a glucose test as no caloric intake for at least 8 hours 3. The EFLM standard for a fasting sample is stricter: a 12-hour fast with the blood drawn in the morning 4. Most US labs ask for 8 to 12 hours. The easiest way to hit that window is overnight: finish dinner by about 8 pm and book the earliest draw you can, around 7 to 9 am.
Longer is not better. Fasting well beyond what the lab asks does not make results more accurate, and very long fasts can shift some of them: ketones climb, and bilirubin rises, most visibly in people with Gilbert's syndrome, a common harmless trait (see high bilirubin). An afternoon appointment after skipping breakfast and lunch tends to produce odd numbers, and the light-headedness that comes with them.
What breaks a fast: water, coffee, gum, smoking, medicines
The EFLM standard sets out what is and is not allowed before a fasting sample 4. It is stricter than many lab leaflets and websites, which say black coffee is fine.
| Before a fasting draw | Allowed? | Why |
|---|---|---|
| Plain water | Yes | Allowed before a fasting sample 4; being well hydrated makes veins easier to find |
| Black coffee or plain tea | No | Not allowed before a fasting sample 4. Caffeine can nudge glucose, insulin and cortisol for a few hours; see coffee before a blood test |
| Chewing gum or mints, even sugar-free | No | Not allowed before a fasting sample 4 |
| Smoking | No | Not allowed on the morning of a fasting draw 4 |
| Alcohol | Not for 24 hours | Avoid for 24 hours before the draw 4; alcohol raises triglycerides |
| Hard exercise | Not for 24 hours | Avoid for 24 hours before the draw 4; hard effort can raise muscle markers such as CK |
| Your usual medicines | Usually yes, with water | Unless your clinician says otherwise. If you take thyroid hormone (levothyroxine), ask whether to take it after the draw. Insulin and sulfonylureas: see the safety section below |
| Supplements, including biotin | Tell your clinician or lab | High-dose biotin can distort some hormone and heart tests for a day or more; our biotin and blood tests guide covers how long to pause it, and supplements that affect blood tests covers the rest. Check with your clinician before pausing anything prescribed |
Fasting safely if you have diabetes, are pregnant or get light-headed
Pregnancy screening often uses a glucose tolerance test, and the prenatal team gives specific instructions for which version you are having; follow those over any general guide. If you tend to faint at blood draws, say so when you arrive so you can lie down, drink water beforehand, and bring a snack to eat right after. Children usually only fast when a test requires it, and for the shortest time the test allows; the ordering clinician will say.
I ate by accident: what now?
Tell the person drawing your blood. Most of the order is still valid: a CBC, A1C, thyroid tests, vitamin D, kidney and liver tests and most cholesterol results read the same after a meal. Fasting glucose, fasting insulin and a fasting triglyceride check are the ones to reschedule, because a fed result cannot be read against fasting cut-offs. The American Diabetes Association uses a separate threshold for glucose taken at any time of day: 200 mg/dL or higher, with classic symptoms, points to diabetes; a clinician confirms any diagnosis with repeat or additional testing 3. If any glucose result is 200 mg/dL or higher, contact the clinician who ordered it promptly. Very high glucose with vomiting, heavy thirst, fast breathing or confusion needs urgent care the same day.
Example: a glucose of 112 mg/dL taken two hours after toastmg/dL
Reading results that were taken fasting, or not
Many reports print a fasting status (often "Fasting: Y" or "N") near the collection time; lab report flags explained covers the other codes your report may show. When the status is missing, a missed fast has a recognisable pattern: glucose, insulin and triglycerides all high while A1C sits in the normal range. A normal A1C beside a high glucose often points to a fed sample, though A1C has its own blind spots, so the clinician who ordered the test is the one to say whether to repeat it fasting. A HOMA-IR score needs both a fasting glucose and a fasting insulin, so a single fed value makes the whole score unreadable. A high HOMA-IR on a proper fasting draw is a sign of insulin resistance.
CalculatorHOMA-IR calculatorEstimates insulin resistance from fasting glucose and fasting insulin taken in the same draw.Work it out For the rest of test-day preparation, including timing, hydration and what to wear, see how to prepare for a blood test. For a panel that bundles glucose with kidney and liver tests, see the metabolic panel explained.
Citations
- Nordestgaard BG et al. Fasting is not routinely required for determination of a lipid profile: clinical and laboratory implications including flagging at desirable concentration cut-points. A joint consensus statement from the European Atherosclerosis Society and European Federation of Clinical Chemistry and Laboratory Medicine. European Heart Journal 2016;37:1944-58
- Grundy SM et al. 2018 AHA/ACC/AACVPR/AAPA/ABC/ACPM/ADA/AGS/APhA/ASPC/NLA/PCNA Guideline on the Management of Blood Cholesterol. Circulation 2019;139:e1082-e1143
- American Diabetes Association Professional Practice Committee. 2. Diagnosis and Classification of Diabetes: Standards of Care in Diabetes 2025. Diabetes Care 2025;48(Suppl 1)
- Simundic AM et al. Standardization of collection requirements for fasting samples: for the Working Group on Preanalytical Phase (WG-PA) of the European Federation of Clinical Chemistry and Laboratory Medicine (EFLM). Clinica Chimica Acta 2014;432:33-7
- Langsted A, Freiberg JJ, Nordestgaard BG. Fasting and nonfasting lipid levels: influence of normal food intake on lipids, lipoproteins, apolipoproteins, and cardiovascular risk prediction. Circulation 2008;118:2047-56
- Bhasin S et al. Testosterone Therapy in Men With Hypogonadism: An Endocrine Society Clinical Practice Guideline. Journal of Clinical Endocrinology and Metabolism 2018;103:1715-44
- Sidhu D, Naugler C. Fasting time and lipid levels in a community-based population: a cross-sectional study. Archives of Internal Medicine 2012;172:1707-10
Educational context only, not medical advice or a diagnosis. Always discuss your results with a clinician.
