Fasting changed only a few routine blood tests, and fasting beyond 12 hours changed none

Design

Retrospective single-centre cross-sectional study with within-person analyses

Population

101,148 adult outpatients; 9.8 million results across 372 tests

Primary outcome

Differences by fasting duration exceeding analytical performance specifications

Effect

Under 8 h vs longer: glucose +10.8%, triglycerides +20.1%, GGT +14.8%, lipase +36.4%; no test differed between 8–12 h and over 12 h

A Korean study used nearly 9.8 million results from 101,148 adult outpatients across 372 laboratory tests at one 670-bed hospital, comparing results after less than 8, 8 to 12, and more than 12 hours of fasting against biological-variation-based analytical tolerances.

In between-group comparisons 27 tests differed more than the tolerance, but within-person comparisons confirmed only 10, suggesting most apparent differences were confounding. The clearest were glucose (10.8% higher), triglycerides (20.1%), gamma-glutamyl transferase (14.8%) and lipase (36.4%) after less than 8 hours of fasting. No test differed meaningfully between 8–12 hours and more than 12 hours.

For patients this matters: older adults, people on insulin or sulfonylureas and outpatients travelling long distances are often told to fast overnight for tests that do not need it. The study supports asking for a fast only when glucose, triglycerides, GGT or lipase are the point of the test, and never asking for more than 8–12 hours. It is single-centre and cross-sectional, so local laboratory policy still applies.

• Ask for fasting only when glucose, triglycerides, GGT or lipase are central to the question
• Do not ask patients to fast for more than 12 hours — longer fasts changed nothing
• Allow non-fasting samples for most other routine tests
• Avoid unnecessary overnight fasts in patients on insulin or sulfonylureas, who risk hypoglycaemia
• Note the time of the last meal on the request when it matters for interpretation

Why it matters

Blanket overnight fasting burdens patients and risks hypoglycaemia for little analytical gain.

Don't overread it

This was a single-centre cross-sectional study; it supports analyte-specific fasting but does not override local laboratory requirements.

The statistics, in plain English

Comparing different people fasting different lengths of time can mislead, because those groups differ in other ways. Comparing the same person on different visits removes much of that, which is why only 10 of 27 apparent differences survived.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.