# Your genes and saturated fat: what a 2026 study of 21,000 women found about cholesterol

> A 2026 JACC Advances WHI study: saturated fat, LDL and heart risk linked most in the top genetic quartile. What it means for food and training.

Published 2026-11-20 · Updated 2026-11-20 · revns editorial

**Your genes can change how much saturated fat appears to move LDL cholesterol and later heart risk**, according to a study of **20,940** genotyped postmenopausal women in the U.S. **Women's Health Initiative**, published online **18 November 2026** in *JACC: Advances*. Women in the **top quarter** of a [polygenic risk score](/dictionary/polygenic-risk-score) for LDL showed a **stronger** link between saturated fat and cholesterol and, over **10 years**, a **higher** rate of heart attacks, strokes and related events when saturated fat rose in statistical models. Women in the **lower three quarters** showed **no clear** added event risk from the same saturated-fat shift in those models. The work helps explain why diets feel unfair — but the authors stress that **polygenic scores are not ready to dictate your plate**.

**TL;DR**

| Fact | Detail (Barad et al., *JACC: Advances*, 18 Nov 2026) |
| --- | --- |
| Who | **20,940** WHI participants with genotyping and lipid data; median age **65** |
| Diversity | **47.9%** non-Hispanic White, **33.0%** Black, **14.4%** Hispanic (self-report) |
| Baseline LDL | Median **148 mg/dL**, adjusted when on lipid-lowering drugs |
| Main finding | Polygenic LDL score **modified** saturated fat ↔ LDL link (**P = 0.01**) |
| Heart events | **1,582** incident ASCVD cases in **16,835** without prior ASCVD; **~10 y** follow-up |
| Top genetics quartile | **5% isocaloric** swap toward saturated fat ↔ **HR 1.26** (95% CI **1.08–1.47**) |
| Lower 3 quartiles | Same swap ↔ **HR 1.00** (95% CI **0.89–1.12**); interaction **P = 0.006** |
| Clinical takeaway | Authors: **no evidence yet** to use polygenic scores for diet advice |



## What did researchers study?

Cardiometabolic researchers led by **Alexa Barad** and **Shoa Clarke** asked whether **inherited tendency toward high LDL** changes how **saturated fat intake** relates to **LDL cholesterol** and to **atherosclerotic cardiovascular disease (ASCVD)** — heart attack, stroke, revascularisation and related deaths.

LDL is a **causal** risk factor for ASCVD, and guidelines still urge limiting saturated fat to protect lipids. Yet people argue about how much saturated fat matters, and **two-thirds of U.S. adults** exceed recommended saturated fat limits, according to the paper's introduction. Part of the confusion is **variable response**: average diet-trial effects hide wide person-to-person swings.

Earlier work tied **specific gene variants** to diet responses. This study used a **genome-wide polygenic score** that captures thousands of small effects — a broader "genetic background" dial than single-gene tests.

Medical Xpress and Stanford Medicine coverage highlighted the same peer-reviewed paper we cite below. We rely on the **journal article and ACC summary**, not press release claims alone.

## Who was in the Women's Health Initiative sample?

The WHI enrolled **161,808** postmenopausal women aged **50 to 79** across **40** U.S. clinical centres between **1993 and 1998**. This analysis used **20,940** women who had:

- Genotyping from WHI ancillary studies (**42,400** started; exclusions for close relatives, low-quality arrays and ancestry groups outside the four largest genetic clusters),
- Baseline **fasting LDL**,
- A baseline **food frequency questionnaire** (122 foods; prior three months),
- Complete covariate data.

Median age was **65.0 years**. Self-reported race and ethnicity in the analytic cohort was **47.9%** non-Hispanic White, **33.0%** non-Hispanic Black, **14.4%** Hispanic and **4.8%** other pooled groups. Genetically inferred ancestry was about **half** European, **one-third** African, **14%** admixed American and **2%** East Asian — reflecting the genotyping design.

**Who this does not represent:** men, premenopausal women, teenagers, athletes in heavy blocks, or anyone whose diet today differs from a **1990s U.S. questionnaire**. Associations in an observational cohort **cannot prove** that changing saturated fat will change **your** future events.

## How did they measure diet and genetics?

**Saturated fat** came from the WHI food frequency questionnaire, expressed as **percent of total daily energy**. Nutrients were computed with the Nutrition Data System for Research. In prior WHI validation, saturated fat correlated **0.63** with multiple recalls — reasonable for population research, imperfect for one person.

**Polygenic score (PGS LDL)** used score **PGS000889**, validated across ancestries in prior work. It was standardised within ancestry group (mean 0, SD 1). Results are reported **per one SD** increase in genetic propensity.

**LDL** was calculated from fasting lipids (Friedewald equation in **96%**) or measured directly. If a woman reported lipid-lowering therapy, the team **divided LDL by 0.7** to estimate untreated levels — a common correction, not a perfect individual measure.

For **incident ASCVD**, they excluded **1,457** women with prevalent disease and **2,648** randomised to the low-fat diet trial arm (to avoid post-randomisation diet shifts), leaving **16,835** for prospective models.

## What were the main results on LDL?

In multivariable linear models with **isocaloric substitution** (replacing **5%** of energy from carbohydrates with saturated fat, holding total calories stable), the association between saturated fat and LDL depended on **PGS LDL** (**P interaction = 0.01**).

Spline plots showed a **nonlinear** pattern: the saturated-fat effect **grew** toward the **upper end** of the genetic-risk distribution — not a simple on/off switch at one gene.

Median LDL in the full sample was **148 mg/dL** (interquartile range **123–176**). Women **above the median** polygenic score had higher median LDL (**162** vs **135 mg/dL** among below-median scores) — the score captured real lipid differences.

Sensitivity analyses in subgroups without diabetes, without ASCVD history, without lipid drugs and across ancestry groups broadly supported the interaction pattern, per the paper.

## What about heart attacks and strokes over 10 years?

Among women free of ASCVD at baseline, **1,582** developed a first composite ASCVD event over **median 10.0 years** of follow-up (quartiles **9.4–10.0**).

When the same **5% isocaloric** increase in saturated fat (from carbohydrates) was modeled:

- **Top quartile** of PGS LDL: **hazard ratio 1.26** (95% CI **1.08–1.47**),
- **Bottom three quartiles combined:** **HR 1.00** (95% CI **0.89–1.12**),
- **P interaction = 0.006**.

That means **in this cohort and model**, extra saturated fat aligned with **higher event rates** only among the **most genetically LDL-prone quarter**. It does **not** mean saturated fat is harmless for everyone else — other outcomes, shorter horizons and different foods were not your personal experiment.

The American College of Cardiology journal scan notes authors' caution: the comparison is **one categorical swap** in an observational design, not a randomised feeding trial.

## What do the authors say you should do with this?

They are explicit: **"The clinical utility of using genetic information to guide dietary recommendations remains to be established."** Senior author **Shoa Clarke** told Stanford Medicine and Medical Xpress that **polygenic risk scores should not yet guide diet recommendations**, though future trials may test targeted advice.

That matters if you see ads for **"DNA diets"** or cholesterol scores. The science here is **explain variability**, not sell a supplement or a strict low-carb plan. **Familial hypercholesterolaemia** and other single-gene conditions still need specialist care — a polygenic score is a different tool.

For everyday trainers, the practical line is humbler:

1. **Know your numbers** with periodic blood tests if you are over 40, have family risk or abnormal prior results (NHS guidance).
2. **Keep saturated fat within usual guideline ranges** unless a dietitian personalises more — whole-food patterns beat chasing one macro.
3. **Train and sleep** as separate levers. [Does exercise lower cholesterol?](/blog/does-exercise-lower-cholesterol-what-the-evidence-shows) — yes, modestly in pooled trials — but not enough alone for many high-LDL people.
4. **Do not n=1 compare** with a friend who dropped LDL on keto or vegan swaps; genetics may be part of the story, but so are weight change, adherence and starting point.



## How does this fit women's wellness after menopause?

Every participant was **postmenopausal**, when LDL often rises and **heart disease risk** climbs. The WHI itself shaped how clinicians think about **hormone therapy, diet trials and women's heart health**.

If you are in **perimenopause or menopause** and training hard, this paper is a reminder that **lipids can shift** even when weight looks stable. It is **not** a reason to abandon strength work — bone and muscle still matter. Pair training with **blood pressure and lipid checks** your clinician recommends, especially if migraines, smoking or family early heart disease add risk. Our [menopause exercise guide](/blog/menopause-exercise-strength-sleep-bone-health) covers strength and bone; use it alongside medical follow-up, not instead of it.

## What foods count as saturated fat (for any diet)?

Guidelines target **saturated fat as a share of calories**, not a ban on one culture's foods. Common sources include:

- **Vegetarian and vegan:** coconut oil and full-fat coconut milk (high saturated fat), palm oil in some processed foods, hard margarines, many bakery items and fried snacks.
- **Eggetarian:** butter, ghee, cheese, whole milk, cream-based sauces.
- **Non-vegetarian:** fatty cuts of red meat, sausages, bacon, poultry skin, butter and cheese.

Swaps that often lower saturated fat **without** dropping protein: lentils and beans, tofu, fish, skimmed or soy milk, olive or rapeseed oil for frying, nuts and seeds. **Fibre** from oats, barley, beans and vegetables can lower LDL in many people — another lever this genetics paper did not replace.

Use [meal tracking](/features/meal-tracking) or a simple food log if you are debugging patterns with a dietitian, not guessing from one week's recall.

## What should you do this week?

| If you are… | Starting action |
| --- | --- |
| Healthy, under 40, no family lipid issues | Keep training, eat mostly whole foods; ask for a baseline lipid panel at your next routine check if offered |
| Over 40 or family history of early heart disease | Book a cholesterol test if you have not had one in the last year (NHS: ask your GP) |
| Already on statins or lipid drugs | Take medicines as prescribed; use exercise and food as **add-ons**, not swaps, unless your clinician agrees |
| Curious about genetic tests | Discuss results with a GP or cardiologist; do not overhaul diet from an app score alone |
| Training for performance | Keep saturated fat moderate on heavy weeks; prioritise sleep and easy days — see [rest days](/blog/rest-days-how-many-you-need-and-what-to-do) |

Log **steps, sleep and workouts** in one place if you use [Apple Health through the iPhone app](/features/multi-sport-tracker) (read-only sync for steps, sleep, heart rate, HRV and workouts). Patterns help clinicians see lifestyle context beside lab dates.

## When should you see a clinician?

- **LDL or total cholesterol above your clinician's target**, or a first abnormal test.
- **Family history** of heart attack or stroke before age 60 in a close relative, or known familial hypercholesterolaemia.
- **Chest pain, pressure or tightness** with exertion; **fainting**; **sudden severe breathlessness**; **weakness or numbness** on one side — emergency care, not a diet tweak.
- **Thyroid disease, diabetes, kidney disease or pregnancy** — lipid targets and diets need professional tailoring.

High cholesterol rarely causes symptoms. A blood test is the only way to know your numbers.

## Related

- [Does exercise lower cholesterol? What the evidence shows](/blog/does-exercise-lower-cholesterol-what-the-evidence-shows)
- [Heart health for men: the numbers to track](/blog/heart-health-for-men-the-numbers-to-track)
- [Blood pressure by age: what is normal and why it rises](/blog/blood-pressure-by-age-what-is-normal-and-why-it-rises)
- [Menopause exercise: strength, sleep and bone health](/blog/menopause-exercise-strength-sleep-bone-health)

This is general training guidance, not medical advice. It does not diagnose high cholesterol, inherited lipid disorders or cardiovascular disease. Food and exercise plans for abnormal lipids belong with a GP, nurse or registered dietitian.

## Sources

- [Barad A, Ritter V, Nudy M, et al. Association Between Saturated Fat Intake and Low-Density Lipoprotein Cholesterol Across the Genetic Spectrum. JACC: Advances. Published online 18 November 2026. doi:10.1016/j.jacadv.2026.103423](https://www.jacc.org/doi/10.1016/j.jacadv.2026.103423)
- [American College of Cardiology. Do Genetics Alter Effect of Saturated Fats on LDL-C and ASCVD Risk? Journal scan, 19 November 2026](https://www.acc.org/latest-in-cardiology/journal-scans/2026/11/19/11/24/do-genetics-alter-effect)
- [NHS. High cholesterol: how to lower your cholesterol](https://www.nhs.uk/conditions/high-cholesterol/how-to-lower-your-cholesterol/)
- [NHS. High cholesterol: getting tested](https://www.nhs.uk/conditions/high-cholesterol/diagnosis/)

## Questions

### Do your genes affect how saturated fat changes cholesterol?

In a study of 20,940 genotyped postmenopausal women in the U.S. Women's Health Initiative, published November 18, 2026 in JACC: Advances, authors found that a polygenic score for LDL cholesterol modified the link between saturated fat intake and LDL levels (P interaction = 0.01). The effect grew larger toward the top of the genetic-risk distribution. That fits older ideas that some people respond more to diet than others, but the study measured associations, not a diet prescription for any one reader.

### What is a polygenic risk score for cholesterol?

A polygenic risk score (PGS) sums many small genetic variants that together estimate tendency toward higher or lower LDL cholesterol. The WHI study used a validated multi-ancestry score (PGS000889). Senior author Shoa Clarke told outlets that polygenic scores are not yet evidence-based for telling someone what to eat. Treat home tests as conversation starters with a clinician, not a solo diet switch.

### Did saturated fat raise heart attack risk for everyone in the study?

Not equally. Among 16,835 women without prevalent cardiovascular disease at baseline, 1,582 had a first atherosclerotic cardiovascular disease event over a median 10 years. A model replacing 5% of calories from carbohydrates with saturated fat was linked to higher risk only in the top quartile of the LDL polygenic score (hazard ratio 1.26, 95% CI 1.08-1.47). In the bottom three quartiles the hazard ratio was 1.00 (95% CI 0.89-1.12). Interaction P = 0.006.

### Why does my friend lose cholesterol on a diet and I do not?

The new paper is one explanation among many: genetics, starting LDL, weight change, fibre, alcohol, smoking, sleep, training and how long you stick with a plan all matter. The study only included postmenopausal women enrolled in the 1990s, with one food questionnaire snapshot. Men, younger adults and people who cycle diets may not match. Track your own blood tests with your clinician instead of comparing anecdotes.

### Should I cut saturated fat if I have a high genetic cholesterol score?

General guidelines still recommend limiting saturated fat for heart health, and this study does not replace them. Authors explicitly wrote that clinical use of polygenic scores to guide diet is unproven and needs trials. If your LDL is high or heart disease runs in the family, work with a GP or dietitian on food, activity and medicines. Exercise can nudge lipids modestly; see our exercise and cholesterol guide for pooled trial data.

### Does exercise still matter if cholesterol runs in the family?

Yes for most people. A 2025 meta-analysis of 148 exercise trials found modest improvements across lipid markers, separate from genetics. Inherited high cholesterol often needs medicines as well. The WHI paper does not test training. Standard advice remains aerobic plus strength work on most days, aligned with WHO and NHS activity guidance, unless a clinician limits you.

---

Canonical: https://revns.com/blog/genes-saturated-fat-cholesterol-whi-study-2026
Index of everything: https://revns.com/llms.txt
revns gives training guidance, not medical advice.
