Low-fat diets cut non-HDL cholesterol by 8 mg/dL but leave ApoB untouched

The low-fat versus high-fat argument has mostly been fought over two familiar numbers, LDL and HDL cholesterol. But lipid metabolism is richer than that: non-HDL cholesterol captures all the cholesterol carried outside HDL, apolipoproteins measure the proteins on the particles themselves, and ratios try to summarise overall balance. Whether cutting fat to 30% of energy or less actually improves these less familiar markers has never been properly settled. This review set out to pool every randomised trial comparing lower-fat with higher-fat eating on exactly those measures, to give a fuller picture of what a low-fat diet does to the blood.

The authors searched PubMed, Scopus and Web of Science from each database's start to June 2024, then updated the search through August 2025, with no language restrictions. They included randomised controlled trials in adults that compared diets providing 30% or less of energy from fat against diets providing more than 30%. Independent reviewers extracted the data, judged each trial's risk of bias and rated how certain the evidence was. Ninety-two trials qualified, and their results were combined statistically to estimate the average difference between the diets on non-HDL cholesterol, apolipoproteins and lipoprotein ratios.

Low-fat diets lowered non-HDL cholesterol by 8.00 mg/dL on average compared with higher-fat diets, and the authors rated that evidence high quality. They also nudged down ApoA1 by 2.51 mg/dL and the HDL subfractions HDL-C2 and HDL-C3 by 0.72 and 1.32 mg/dL. Crucially, ApoB and its related measures, including the ApoB to ApoA1 ratio, showed no significant change, though that evidence was rated low to very low quality. Two ratios, total cholesterol to HDL and triglycerides to HDL, rose slightly on the low-fat diets, and lipoprotein(a) did not move at all.

The result confirms that trimming dietary fat reliably lowers the cholesterol carried outside HDL, which fits what conventional lipid testing has long suggested. It complicates the picture too: the particle marker ApoB stayed flat, and small drops in HDL fractions pushed two ratios slightly the wrong way. The trials also disagreed with each other considerably on the size of the non-HDL effect. The open question is whether a diet that lowers non-HDL cholesterol without clearly moving ApoB delivers the heart benefit people expect, something marker-based trials cannot answer on their own.

If you eat lower fat partly for your cholesterol, this is a reason to expect your non-HDL number to improve, and the evidence behind that specific change is strong. It is equally a reason not to assume every marker improves: ApoB did not clearly budge here, HDL fractions dipped slightly, and two ratios crept upward. None of the outcomes in these trials were heart attacks or deaths, only blood measures, so this informs expectations about a lipid panel rather than proving anything about events. The authors themselves describe the evidence on ApoB and the ratios as inconclusive.