Skipping breakfast before a blood test? Ten million results say it matters for only a few tests

Many people arrive for a blood test hungry, because the instructions said to fast. A study published in JAMA Internal Medicine on 28 September 2026 asked how much that fast actually changes the numbers. Laboratory doctors from Seoul used the records of adult outpatients at a 670-bed university hospital in the Seoul area, collected between February 2021 and December 2023, where the time since the patient's last meal was written down at the blood draw.

The dataset held 9,755,547 results from 101,148 patients across 372 different tests. Comparing groups who had fasted under 8 hours, 8 to 12 hours, and over 12 hours, 27 tests showed differences big enough to matter by European laboratory standards. But when the researchers compared the same people tested on different days, only 10 held up, which suggests the rest came from differences between patients rather than from food. The clearest effects with under 8 hours of fasting were higher glucose (10.8%), triglycerides (20.1%), the liver enzyme GGT (14.8%) and the pancreatic enzyme lipase (36.4%). Fasting beyond 12 hours added nothing: no test differed meaningfully between the 8 to 12 hour and the over 12 hour groups.

The authors conclude that fasting rules should be set test by test rather than applied to every blood draw. This is one hospital in one country, and the data are observational, so it does not change what your own doctor asks of you. If you are told to fast for a glucose or lipid test, do so; and if a long fast makes you dizzy, or you take medicine for diabetes, ask whether you really need to.