Fasting before bloodwork mattered for just 10 of 121 tests in a Seoul hospital's records

The team, led by laboratory medicine physician Sunghwan Shin, did not assign anyone a fasting time ^[2][3]. The records covered adult outpatients whose fasting before the draw was documented, sorted afterwards into under eight hours, eight to 12, and more than 12 ^[4]. The obvious worry with that design is that people who arrive fed may differ from people who arrive fasted in ways that also move their results. So the researchers ran within-person analyses alongside the between-group comparisons, setting a patient's results against that same patient's other draws so that whatever is stable about the person stays fixed ^[15].

Every difference had to clear a bar: the established analytical performance specification for that test ^[5]. The 111 analytes that stayed under it make up about 92 percent of the 121 examined ^[1][2]. The bar is a laboratory standard. The thing it does not tell you is whether a difference just past it would change a diagnosis or a dose, and the study as reported does not measure clinical outcomes.

Two of the four named exceptions, glucose and triglycerides, are the tests the rule was built around ^[7]. The researchers wrote that fasting is "a practice that originated with lipid and glucose testing" and that it "is now applied by default to much broader test panels" ^[10]. The spread ran ahead of the evidence. Prior research on pre-test fasting comes mostly from lipid tests. Comparable data are limited for liver and pancreatic enzymes, blood-cell indices, hormones and vitamins, and fasting recommendations differ from one lab to another ^[11].

"These findings support analyte-specific rather than uniform fasting requirements," the authors wrote in JAMA Internal Medicine ^[9][13]. In an accompanying commentary, clinical biochemist Liv Tybjaerg Nordestgaard of Copenhagen University Hospital and colleagues wrote that the study "provides strong support for accumulating evidence that fasting blood sampling is not necessary" ^[12]. ScienceAlert, reporting the study, was more careful. It said the results alone may not yet warrant changing the pre-test standard, and that they do not mean patients should disregard an instruction to fast ^[14].

I think the case for test-specific instructions is strong, with conditions attached. The records come from one 670-bed university hospital between 2021 and 2023 ^[3]. A lab with a different patient population or different assays would want its own check before copying the result.

Patients carry the cost of the current default. The usual instruction is nothing but water for 8 to 12 hours ^[1], and the burden that places on patients was one reason the authors took up the question, according to ScienceAlert ^[16]. For the 111 analytes outside the exceptions, that fast bought no difference large enough to clear the bar ^[1].