How Long to Fast Before a Blood Test? Past 12 Hours, It Made No Difference

Few things feel as pointless as skipping breakfast for a blood test, only to find out you didn’t need to.

A huge new study suggests that happens a lot.

Researchers analysed nearly 10 million test results from over 100,000 patients. Most routine tests barely changed whether people had eaten recently or not.

But a handful did, and for those, fasting really matters.

How long should you fast before a blood test?

If your doctor asks you to fast, 8 to 12 hours is the useful window. This study found nothing gained by going longer.

The tests most affected were blood sugar (glucose) and triglycerides, a type of blood fat. Two others, the liver enzyme GGT and the pancreatic enzyme lipase, were also higher after recent eating.

Most other routine tests were barely affected. That includes full blood counts, salts, kidney and liver function tests, and total and HDL cholesterol.

Always follow your own doctor’s instructions, though. As the NHS explains, you may be asked not to eat or drink anything other than water for a short time before some tests.

What the research shows

The study, published in JAMA Internal Medicine, used routine data from a 670-bed university hospital near Seoul, South Korea.

Why do fasting rules exist at all? Fasting is meant to reduce the variation recent meals cause in results.

But whether those differences matter clinically, across a broad range of everyday tests, had never been checked systematically using real clinical data.

A huge real-world dataset

Researchers analysed outpatient blood samples taken between February 2021 and December 2023. Each patient’s time since their last meal was recorded.

They compared results against established limits for how much a test can vary before it matters clinically.

The trick that removed false alarms

At first, 27 tests looked different across fasting groups.

But people who fast for 12 hours may differ from people who come in after lunch. So the team checked the same patients across different visits.

Only 10 of the 27 differences held up. The other 17 mostly reflected differences between people, not fasting.

The tests that really changed

Comparing the same people after under 8 hours versus over 12 hours of fasting:

Another pancreatic enzyme, amylase, rose only 4.6% after recent eating, which stayed within its clinical limit.

Glucose fell by about 4.3 mg/dL for every extra hour of fasting in the first few hours after eating. That’s why fasting still matters for diabetes screening.

What happens in the hours after a meal

The team also looked closely at people tested within 8 hours of eating, in 30-minute steps.

Glucose was highest soon after a meal and fell steadily. Adjusted values dropped from 134.0 mg/dL at half an hour to 113.3 mg/dL at four hours.

Lactate, a by-product of energy use, fell even more: about 27.9% lower by four hours. The authors link this to a short rise after eating that clears within a few hours.

Blood fats behaved differently. Within that first 8 hours, lipid levels showed no clear trend with time since eating.

Small shifts in white blood cells

Some white blood cell measures also moved a little. Neutrophils were 8.2% higher after recent eating, while two other cell percentages were slightly lower.

All stayed close to or within clinical limits. The authors note these cells naturally vary through the day, and people who eat first tend to be tested later.

Surprises that weren’t real

The paired check also flagged two long-term blood sugar markers, including HbA1c, as slightly higher after recent eating.

That can’t be a fasting effect: these markers build up over weeks to months. The authors suggest unscheduled visits, during periods of poorer control, explain it.

It’s a useful reminder. Even careful comparisons can’t remove every source of bias.

Longer isn’t better

Between 8 to 12 hours and over 12 hours of fasting, no test changed beyond clinical limits.

Glucose differed by just 0.7% between those two groups. So skipping breakfast and lunch doesn’t make your results more accurate.

What earlier research found

Cholesterol barely moves with food

A 2012 study in Archives of Internal Medicine looked at lipid tests from 209,180 people in Calgary, Canada.

Total cholesterol and HDL cholesterol differed little by fasting time. Calculated LDL varied by up to 10%, and triglycerides by up to 20%.

Europe dropped routine fasting for cholesterol

In 2016, the European Atherosclerosis Society and European laboratory medicine experts issued a joint statement. They recommended routine use of non-fasting cholesterol tests.

They found that typical meals changed triglycerides by about 26 mg/dL and cholesterol by about 8 mg/dL, which isn’t clinically significant.

They added that HDL cholesterol, apolipoprotein B and lipoprotein(a) aren’t affected by fasting at all. A fasting retest may be considered only when non-fasting triglycerides come back above 440 mg/dL.

How it fits together

The cholesterol finding was already established. This study extends it to dozens of other tests and shows that fasting longer than 12 hours doesn’t help.

How much should you trust this?

Promising. It’s an enormous dataset with a smart within-person check, but it has real limits.

What makes it convincing

• Nearly 10 million results from over 100,000 patients.
• Paired comparisons of the same people removed many false differences.
• Results were judged against established clinical limits, not just statistics.
• The findings match earlier large lipid studies.

What makes me cautious

• It’s one hospital in Korea, so diets and lab methods may differ elsewhere.
• Meal content and reported fasting times couldn’t be checked.
• It didn’t measure how often results crossed diagnostic cut-offs.
• Some tests were only included when ordered alongside a fasting test.

What this means for you

• Follow your doctor’s instructions. They know which tests you’re having.
• If fasting is required, aim for 8 to 12 hours. Longer doesn’t improve accuracy.
• Water is usually fine. Ask if you’re unsure.
• Diabetes screening needs proper fasting. Glucose is the most affected test.
• Ask whether fasting is really needed. For many tests, including standard cholesterol in some countries, it may not be.
• Don’t stop medicines without advice. Some tests need special preparation, but ask first.

The NHS page on blood tests explains how to prepare and what to expect.

In this short video from Texas A&M Health, a primary care physician explains the do’s and don’ts of fasting before routine bloodwork:

What we still don’t know

1. How often do results change a diagnosis? The study didn’t count threshold crossings.
2. Does what you eat matter more than when? Meal content wasn’t recorded.
3. Do results apply elsewhere? It was one hospital with one population.
4. What about rarer tests? Many weren’t ordered often enough to analyse.
5. Could labs adjust for meal timing? The authors’ time-based data hint at it.

My take: fast smarter, not longer

What I like about this study is the within-person check. It quietly removed 17 false alarms that a simpler analysis would have reported.

I’m less convinced by anyone telling you fasting is pointless. For glucose and triglycerides, it clearly matters.

So follow the instructions, aim for 8 to 12 hours when fasting is needed, and don’t go hungry longer than that.