Food swaps that lower ApoB

A high ApoB result usually sends people to the egg carton first. The foods with the stronger trial record sit elsewhere in the kitchen, in the fat that goes in the pan, the fiber in breakfast, and a short list of cholesterol-lowering foods eaten together. In a 2018 meta-analysis of 7 controlled trial comparisons in 439 adults with high cholesterol, adding viscous fiber, plant sterols, soy protein, and nuts to an already low-saturated-fat diet lowered ApoB by 15% and LDL cholesterol by 17%.¹ In free-living adults over 6 months, the same pattern lowered ApoB closer to 11%.³ That is a drop you can test on your own panel, and a useful benchmark for the studied Portfolio pattern.

01ApoB counts atherogenic particles

LDL-C asks how much cholesterol is riding inside LDL particles. ApoB asks how many atherogenic particles are in circulation. Each LDL, VLDL remnant, IDL, and Lp(a) particle carries one ApoB protein, so the blood result works as a particle-count proxy.⁶

That split explains a common false all-clear. Triglycerides can look better, HDL can look fine, and LDL-C can sit in a comfortable range while ApoB stays high because the cholesterol is spread across more particles. The 2024 National Lipid Association consensus holds that when these markers disagree, ApoB and non-HDL-C often track atherosclerotic risk more closely than LDL-C.⁶

The 2026 ACC/AHA dyslipidemia guideline keeps LDL-C and non-HDL-C as treatment anchors and says ApoB measurement is reasonable in adults on lipid-lowering therapy, particularly those with ASCVD, cardiovascular-kidney-metabolic syndrome, type 2 diabetes, or elevated triglycerides, once LDL-C and/or non-HDL-C goals are achieved.⁷ For a food trial, the useful unit is two draws under similar conditions, one on your usual diet and one after the swaps have had time to settle.

02Why saturated fat replacement comes first

Saturated fat can reduce LDL-particle clearance, so the useful move is replacement rather than subtraction. Mensink's 2016 WHO regression of feeding trials quantified the swap. Each 1% of energy moved from saturated fat to cis-polyunsaturated fat lowered LDL-C by 0.055 mmol/L. The same 1% moved to cis-monounsaturated fat lowered it by 0.042 mmol/L. The same 1% moved to carbohydrate lowered it by 0.033 mmol/L, with a worse triglyceride and HDL picture.⁴

Five percent of energy is a realistic kitchen change on a 2,000 kcal day, about 11 grams of fat. Shifting that much from butter, ghee, coconut oil, cream, and fatty processed meat toward polyunsaturated sources such as canola, walnuts, seeds, and fish works out to roughly 0.28 mmol/L, or 11 mg/dL, of LDL-C in the Mensink math. Olive oil and avocado are mostly monounsaturated, so the same swap toward them lands somewhat lower. These are LDL-C estimates. Mensink's figures are not a direct ApoB forecast.

Saturated fat replacement is also the swap in this article with the most direct cardiovascular-event evidence, and the three main sources measure different things. The 2017 American Heart Association presidential advisory focused on the randomized trials it judged adequate that replaced saturated fat with polyunsaturated vegetable oil, and estimated about a 30% reduction in cardiovascular disease.⁸ Cochrane's 2020 review cast a wider net over trials that reduced saturated fat for at least two years and found a smaller 17% reduction in combined cardiovascular events across 12 trials in 53,758 people, with little or no effect on death.⁹ Li and colleagues added cohort evidence. In two large U.S. cohorts, replacing 5% of energy from saturated fat with polyunsaturated fat was associated with 25% lower coronary heart disease risk, and replacing it with refined carbohydrate was not associated with lower risk.¹⁰ The cohort result agrees with the trials on direction. It cannot prove cause on its own.

The 2025-2030 Dietary Guidelines cap saturated fat at 10% of calories, about 22 grams on a 2,000 kcal day.¹¹ The American Heart Association uses a tighter 5% to 6% target when the job is lowering LDL-C, about 13 grams at the same calorie level.¹² Both limits scale with energy intake, so an athlete eating 3,000 kcal has a 10% ceiling near 33 grams. The audit is the recurring sources, the daily butter, cheese, cream, and processed meat. A Saturday steak still counts toward the seven-day average.

Coconut oil belongs in the saturated-fat column. The AHA advisory reported that coconut oil raised LDL-C the way butter, beef fat, and palm oil do, with no offsetting outcome benefit shown.⁸ The seed oil evidence review covers the inflammation claim. The lipid-relevant move is the swap away from tropical oils and bakery fats.

03The portfolio stack adds fiber, sterols, soy, and nuts

Jenkins built the pattern by putting four already-studied foods on the same plate. Plant sterols, soy protein, viscous fiber from oats, barley, and psyllium, and almonds all went in together. In the 2003 JAMA trial, 46 hyperlipidemic adults ate one of three diets for a month. A very-low-saturated-fat control dropped LDL-C 8.0%. The same diet plus 20 mg lovastatin dropped it 30.9%. The portfolio diet dropped it 28.6%. ApoB fell about 0.34 g/L on the portfolio, from 1.49 to 1.15 g/L, and about 0.38 g/L on the statin. C-reactive protein fell 28.2% on the portfolio and 33.3% on the statin.²

That month was close to a feeding study, with the study foods supplied. Jenkins then tested the same idea in free-living adults for 6 months. Intensive dietary advice produced a 13.8% LDL-C drop. Routine advice produced 13.1%. The low-saturated-fat control produced 3.0%. ApoB fell from 122 to 109 mg/dL with intensive advice and from 126 to 112 mg/dL with routine advice, about 11% in both arms. The control arm stayed at 120 mg/dL. Adherence predicted the LDL-C result.³ The 28.6% month is what a research kitchen can do. The 11% ApoB half-year is the number to plan around, because it is what a grocery list did when people kept buying it.

Chiavaroli's meta-analysis is the pooled estimate. Seven trial comparisons, 439 people, portfolio foods on top of an NCEP Step II diet versus that diet alone. LDL-C fell 0.73 mmol/L, about 17%. ApoB fell 0.19 g/L, about 15%. Non-HDL-C, triglycerides, blood pressure, and CRP moved with it. Body weight and HDL-C did not.¹ Because weight held steady, the ApoB drop cannot be credited to weight loss.

The individual doses are smaller than the stack, which is why adding oats while keeping the butter tends to disappoint. Read the table with one caution. Most single-food numbers are LDL-C results. ApoB was measured directly in the Chiavaroli pooling, both Jenkins trials, the Ho oat analysis, and the Sabaté almond trial. LDL-C and ApoB usually move in the same direction when LDL carries most of the particles, though an LDL-C percentage is not an ApoB percentage.

Plant sterols need a second caution. The sterol evidence cited here measures LDL-C. Neither the Ras meta-analysis nor the 2014 EAS consensus drew on a randomized trial with cardiovascular events as the endpoint, so any outcome benefit is inferred from the LDL-C drop rather than shown directly.¹⁴¹⁵

The percentages in the table do not add like a spreadsheet. Jenkins already stacked them under research conditions and still depended on adherence in free living. Soluble fiber from oats, beans, barley, psyllium, and fruit is the cheapest daily piece. The broader gram target for active adults is in How much fiber active adults need. For lipids, the viscous fraction is what counts. A 30 g fiber day from oat bran, beans, lentils, and an apple is a different event from 30 g of inulin in a protein bar.

04Weight, remnants, and alcohol change the particle mix

ApoB also counts remnant particles from VLDL. Insulin resistance, visceral fat, and fatty liver are commonly linked with higher VLDL output. Remnant particles still carry ApoB, so the count can stay high while LDL-C looks average. That is the discordance the 2026 guideline wants ApoB to catch in people with diabetes or high triglycerides.⁷

Weight loss helps, likely in part by lowering VLDL output. Lamantia's 2016 review of nutritional management of high ApoB found that every study inducing a 6% to 12% weight loss lowered plasma ApoB.¹⁸ You do not need a ketogenic protocol to get that. You need a deficit you can repeat, enough protein to protect lean mass, and a fat quality that does not refill the saturated-fat column while calories drop.

Alcohol matters most as a confounder for your retest. It can raise triglycerides in some people, so a heavy week before one draw and a quiet week before the other muddies the comparison. Hold alcohol near zero during the trial, or keep it identical before both draws. Otherwise you may credit oats for a change that was really last Saturday.

Exercise improves insulin sensitivity, triglycerides, blood pressure, and fitness more reliably than it produces a large ApoB drop on its own. Keep training. Do not treat a hard training week as a substitute for the food swaps.

A low-carbohydrate pattern can lower triglycerides and still raise ApoB if saturated fat climbs. That is the panel people screenshot as a win, with HDL up, triglycerides down, LDL-C a little high, and ApoB never measured. Measure the particle count before you decide the pattern is cardioprotective.

05Where eggs and Lp(a) fit when ApoB is high

Eggs are a real lever, and a smaller and more variable one than the breakfast built around them. A 2020 meta-analysis of randomized trials in healthy adults found higher egg intake raised LDL-C by about 8 mg/dL on average compared with control diets, and individual responses varied.¹⁹ Trial data on eggs and ApoB directly are thinner. One large egg carries about 186 mg of cholesterol and about 1.6 g of saturated fat.²⁰ If your ApoB is high, fix the recurring butter, bacon, and cheese that tend to arrive with the eggs first. If the number is still high after that, cutting some yolks is a reasonable next variable to test on your own panel. The full evidence is in Are eggs bad for you.

Lp(a) is the part diet barely touches. Those particles carry ApoB, so they contribute to the count, and the 2026 guideline recommends measuring Lp(a) once in adulthood and states that lifestyle changes minimally affect it.⁷ A clean food trial that leaves ApoB high may point to Lp(a), familial hypercholesterolemia, or both. Either one is a reason to see a clinician.

06How far food moves ApoB compared with medication

In Jenkins 2003, a 20 mg dose of lovastatin, an older statin, matched a research-kitchen portfolio for one month.² In free-living adults, the same portfolio lowered LDL-C about 13% to 14% and ApoB about 11% over 6 months. Across controlled trials, the pooled estimate was 17% for LDL-C and 15% for ApoB.¹³ Use those numbers to set expectations. A 10% to 15% ApoB drop is meaningful. It will not close every gap.

Familial hypercholesterolemia, LDL-C of 190 mg/dL or higher, established ASCVD, diabetes with a high particle burden, chronic kidney disease, and a very high Lp(a) need clinical care. Food still helps on top of medication.

The 2024 National Lipid Association consensus suggested ApoB thresholds that may inform treatment intensification. Those are about 90 mg/dL for borderline to intermediate risk, 70 mg/dL for high risk, and 60 mg/dL for very high risk.⁶ Those are NLA decision points, not universal home targets or a substitute for the 2026 guideline's risk-specific treatment goals.⁷ A clinician applies those numbers against your full risk profile, so treat them as the context for a conversation rather than a home target. The food trial answers a narrower question, whether diet got a fair test.

07How to run a food trial between two blood draws

Start with a baseline panel on your usual diet. Ask for ApoB, LDL-C, non-HDL-C, and triglycerides. Without that first draw, the retest has nothing to compare against.

Then change the food. Lipids respond within weeks, which is why Jenkins could report a result at one month.² For a personal read, hold the new pattern for 6 to 12 weeks before retesting. That window lets the change settle and tests whether the pattern survives ordinary weeks, travel, and training blocks. Keep body weight, alcohol, training load, and supplement use as steady as you can, and draw the second sample on the same kind of morning as the first.

A Mediterranean or DASH pattern already points in this direction. The portfolio doses make the lipid math explicit. Raising your omega-3 index with fish is a triglyceride and remnant project that sits beside ApoB.

If the retest shows ApoB moving toward the range your clinician set for your risk, keep the pattern and recheck on your clinician's schedule. If the number barely moves after a fair trial, or family history, plaque, diabetes, or a high Lp(a) is in the file, take both panels and your food log to a clinician. Two results and a record of what changed turn that visit into a treatment decision instead of a debate about whether diet was tried. Book the baseline draw before you change the pan.