Fasting before a blood test

Fasting exists for a narrow reason: a few markers change with the last meal, and comparing them with a reference range drawn from fasting people only works if you were fasting too. For most other tests a meal makes no meaningful difference. The trouble is that a single draw usually includes several tests, and the strictest one sets the rule for the whole tube.

Which tests need it

MedlinePlus lists glucose, cholesterol, triglycerides and calcitonin among the tests commonly needing fasting, and notes that a creatinine test may ask you to avoid cooked meat beforehand, since meat adds creatine.

How long

Eight hours is the common minimum. Quest's consumer site says fasting usually means eight hours or more, and its lipid panel page recommends 8 to 12 hours of nothing but water. The CDC's cholesterol page uses the same 8 to 12 hour range. Longer is not better: fasting well beyond 12 hours can itself change some results, and an overnight fast with a morning draw is the practical standard.

Water

Yes. Every source here agrees water is allowed, and MedlinePlus notes some tests ask you to drink extra water to make veins easier to find. Dehydration concentrates the blood and nudges up haemoglobin, albumin, BUN and creatinine, so a glass or two before the draw is helpful rather than harmful.

Coffee

Quest defines fasting as "not eating or drinking anything except water", which rules out coffee, black or otherwise. Most lab instructions follow the same line. If a draw is non-fasting, coffee is not a problem for most markers. If it is fasting, save it for afterwards. The same applies to tea, gum, mints and anything else with flavour, and MedlinePlus lists not smoking among common preparations.

Medicines

Quest's guidance is to keep taking prescription medicines during the fast. MedlinePlus says some medicines and supplements may need to be avoided, but not to stop any medicine unless your provider tells you to. Two exceptions are common enough to name:

• Glucose-lowering drugs. People on insulin or sulfonylureas can go low if they fast and still take their usual dose. A 2024 Clinical Chemistry review noted that avoiding hypoglycaemia is one advantage of non-fasting sampling. The prescriber sets how to handle this.
• Biotin. The FDA has warned since 2017, and again in 2019, that biotin in supplements can cause falsely high or low results on certain laboratory tests that use biotin in the assay, including some troponin tests used to diagnose heart attack, where it can give a falsely low result. Hair, skin and nail supplements can contain far more biotin than the daily requirement. Tell whoever orders or draws your test if you take it; the ordering clinician can advise whether to pause it.

Weekly injectables such as GLP-1 receptor agonists do not need to be timed around a draw for fasting purposes, but writing down the days since the last dose helps when results are read later.

Other preparation that matters as much as fasting

• No hard exercise for a few days. In a 2008 study, one hour of weightlifting raised AST and ALT in healthy men for at least seven days.
• No alcohol the day before. MedlinePlus lists alcohol among things to avoid; it raises triglycerides and can affect liver enzymes.
• Same time of day each time. Several markers follow a daily rhythm, and a morning fasting draw is the easiest condition to repeat.
• Not while ill. A cold raises hs-CRP and the white count and lowers cholesterol for a while.

Is fasting on its way out?

For lipids, largely. The 2016 European consensus found that after ordinary meals, triglycerides rose by about 26 mg/dL and LDL fell by about 8 mg/dL at most, on average, and judged neither clinically significant. By 2024, guidelines in Europe, the US, Canada, Brazil, Japan, India and Australia endorsed non-fasting lipid profiles in some or all patients. A 2026 JAMA Internal Medicine piece from the same Copenhagen group was titled "Time to Stop Fasting Before Blood Sampling." For glucose and insulin, fasting still defines the test.

The real rule: same conditions every time

Whether you fast or not matters less than doing it the same way each time. A fasting result compared with a non-fasting one is two different measurements. The biological variation page explains how much noise is left even when conditions match.

Logging it

In Peptide AI, add a note to the draw: fasted or not, hours since eating, time of day, days since last dose, and anything unusual such as a hard session or a cold. It sits with the results in bloodwork tracking and on the same timeline as the dose log, which is what makes two draws comparable months later.