Common nutrition misconceptions
Widely repeated dietary beliefs that are wrong, oversimplified, or actively promoted by food marketing — and what's actually true.
"Some sugars are healthier than others"
Honey, agave nectar, coconut sugar, brown sugar, maple syrup, "raw cane sugar" — all marketed as healthier alternatives to white sugar. Biologically, they are almost identical. Once digested, they all break down into glucose and fructose in similar ratios. Agave is actually worse than table sugar by some measures (70-90% fructose vs sugar's 50%). Your liver and pancreas respond the same way regardless of the marketing.
"Fructose is fine — it doesn't spike blood sugar"
Half-true and entirely misleading. Fructose doesn't raise blood glucose much — because it bypasses glucose metabolism entirely and is processed almost exclusively by the liver. The liver converts excess fructose directly into fat (de novo lipogenesis), driving MASLD (fatty liver), raising triglycerides, and worsening insulin resistance. A "low-GI" fructose-heavy food looks innocent on a glucose monitor while quietly damaging the liver.
"Natural sugars in fruit juice are different"
Once the fibre is removed (juicing) or the cellular structure is destroyed (blending), the body processes the sugar in fruit essentially the same way as Coca-Cola. A 250ml glass of orange juice contains roughly the same sugar load as a similar volume of soda. The NHS specifically caps fruit juice at 150ml per day for a reason. "It's natural" is not a metabolic free pass.
"Carbs alongside a big meal don't count"
Eating carbs alongside protein, fat, and fibre does meaningfully blunt the glucose spike — but it doesn't eliminate it, and it doesn't change the total sugar load your liver has to process. A large pasta dish followed by a sugary dessert still delivers a substantial carbohydrate dose; the body just processes it more gradually. Pairing is a moderating tactic, not a cancellation.
"I exercised, so I can eat what I want"
Exercise improves insulin sensitivity and helps clear glucose — genuinely useful — but it doesn't undo a sugar overload. A 30-minute run burns roughly 300 kcal; a milkshake can contain 600+. More importantly, the fructose half of sugar still goes straight to the liver regardless of how much you've exercised. Movement is a multiplier on a good diet, not a license for a bad one.
"Low-fat means healthy"
When manufacturers remove fat, they almost always replace it with sugar (and salt) to maintain palatability. "Low-fat" yoghurts, "fat-free" salad dressings, and "diet" cereals are frequently higher in added sugar than their full-fat counterparts. Many low-fat products are metabolically worse than the original. Read the sugar content on the label, not the marketing on the front.
"Granola and smoothie bowls are healthy breakfasts"
Most commercial granolas contain 15-25g of sugar per 50g serving — comparable to a chocolate bar. "Healthy" cereal bars are often 30-40% sugar by weight. Acai and smoothie bowls in cafés frequently contain 40-60g of sugar in a single serving. The marketing positions these as health foods; the metabolic impact is closer to dessert. If you eat them, treat them as treats, not breakfast.
"Blending a banana into my berry smoothie makes it healthier"
A banana-and-berry smoothie looks like an obvious win — but bananas are rich in polyphenol oxidase (PPO), the browning enzyme, and it rapidly degrades the heart- and brain-protective flavanols in berries, cocoa and grapes. In a controlled UC Davis / University of Reading crossover trial (Food & Function), adding a single banana cut flavanol absorption by roughly 84% versus a flavanol control. Bananas are still a healthy food — the catch is only blending them with your flavanol-rich ingredients. Easy fix: blend berries or cocoa with a low-PPO partner (pineapple, mango, orange or yoghurt) and eat the banana on its own.
"Diet drinks and sweeteners are healthier"
In May 2023, the WHO issued a guideline recommending against non-sugar sweeteners (NSS) for weight control or chronic disease prevention. Observational studies link long-term NSS use with increased risk of type 2 diabetes, CV disease, and mortality. In July 2023, the IARC classified aspartame as "possibly carcinogenic to humans" (Group 2B). NSS also maintain the brain's expectation of sweetness, which may drive cravings for other sweet foods. The goal is to reduce overall sweetness — not swap the source.
"Children need sugar for energy"
They don't. Children need calories, which they get from whole carbohydrates (oats, rice, potatoes, bread), protein, and fats. Added sugar contributes nothing nutritionally that those foods don't provide more steadily. The "growing kids need sugar" line is largely a legacy of cereal and confectionery advertising — and is one of the main reasons paediatric MASLD is now among the most common chronic liver conditions in children in high-income countries.
"My cholesterol was normal for years and I still had plaque, so cholesterol is not the problem"
The observation is real and the conclusion does not follow. People with unremarkable LDL-C do develop coronary disease, and this site spends an entire chapter on why: LDL-C is a concentration, apoB is a count, and in roughly one person in five they disagree. A “normal” LDL-C carried on an unusually large number of particles is a normal number and a high risk. That is a case for measuring the right thing, not for concluding the lipid is innocent.
And there is a simpler mechanism behind the observation, which almost nobody is told. “Normal” on a UK lipid panel is a population reference range — where most people sit — not a risk-based target. Those are different things. An LDL-C of 3 mmol/L is comfortably inside the normal range and is also close to the middle of the distribution for people having a first heart attack. A number can be entirely typical and entirely compatible with disease, because the typical population is one in which coronary disease is the commonest cause of death.
What the argument has to get past is the genetics. Lowering apoB-carrying lipoproteins reduces events across statins, ezetimibe, PCSK9 antibodies and bempedoic acid — four pharmacologically distinct mechanisms — and Mendelian randomisation points the same way across the corresponding genes. Insulin resistance is a genuine and under-measured driver, and it does not displace that.
Both things are true at once, which is the version worth holding: raised apoB is causal, and metabolic dysfunction changes what those particles do and how long they stay in circulation. This site measures both, and the calculator flags explicitly when a normal LDL sits beside a raised TG/HDL or HbA1c.
"A test can predict a heart attack with 94% accuracy"
No cardiovascular test does this, and “accuracy” on its own is not an interpretable number. Without the sensitivity, the specificity, the population and the time horizon, a single percentage says nothing — and in a population where most people will not have an event inside the study window, a test that says “no” to everybody scores impressively.
Where these figures usually come from is a swapped endpoint. The LP-IR score is the current example: its striking hazard ratios — up to about tenfold — are for incident type 2 diabetes, while its cardiovascular associations are described by the laboratory that sells it as comparable to HOMA-IR. Quoting the diabetes number in a discussion of heart attacks is how a real test becomes an overstated one.
To be clear about what is being criticised: the endpoint swap is a habit of the marketing, not a property of the test. LP-IR measures what it measures perfectly well, and this site grades it Moderate for insulin resistance and diabetes risk. What does not follow is the leap from that to predicting heart attacks.
How to read a trial covers the general version, and it is worth knowing because this is the commonest way a genuine finding gets oversold.
"Berberine is a natural alternative to statins"
Berberine is heavily promoted online as "nature's Ozempic" or a natural statin. It does modestly lower LDL, triglycerides and blood sugar — but the trials are mostly small and low-quality, and crucially there is no evidence it prevents heart attacks or strokes, which is exactly what statins are proven to do. At best it is a metabolic-support supplement, not a statin replacement. It also carries real cautions: it interacts with many medicines — statins (raising the risk of muscle injury), blood thinners, and diabetes drugs (risk of low blood sugar) — supplement purity is unregulated, and it must not be taken in pregnancy or breast-feeding. If you are considering it, and especially if you already take a statin, tell your GP or pharmacist first. Clinicians: see the screening note in the "For doctors" section.
Saturated fat — what the modern science actually says
A focused explainer for the genuine nuance in the saturated-fat literature. This is a topic where the public has been exposed to contradictory messaging — and where some of that contradiction is justified. Engaging with it honestly, rather than ignoring it, gives you the foundation to make sensible dietary decisions.
You have probably encountered competing claims: that saturated fat is the primary driver of heart disease (the conventional public-health position) and that saturated fat has been unfairly demonised (the position taken by some low-carb / keto communities and a small but vocal group of researchers). Neither extreme is fully correct, and the truth is more interesting and more useful than either.
The historical hypothesis (Keys, 1960s onward)
Ancel Keys and colleagues built the "diet-heart hypothesis" from the Seven Countries Study in the 1960s — observing that countries with higher saturated fat intake also had higher rates of coronary heart disease. The chain of reasoning became: saturated fat raises blood cholesterol; high blood cholesterol causes atherosclerosis; atherosclerosis causes heart attacks; therefore reduce saturated fat. This became the basis of UK and US dietary guidelines for over five decades.
What's been confirmed — and what's been nuanced
✓ Confirmed
- Saturated fat raises LDL-cholesterol in most people. This is one of the most reliably reproduced findings in nutrition. Effect size varies widely by individual.
- LDL-cholesterol is causal for atherosclerosis. The strongest evidence comes from Mendelian randomisation — natural genetic variants that lower lifetime LDL (like PCSK9 loss-of-function) produce dramatic reductions in coronary disease, independent of diet or behaviour. This is causal evidence, not correlation.
- APOE ε4 carriers are particularly sensitive. Same saturated fat intake produces a larger LDL rise — and a larger plaque-formation risk over time. ε4/ε4 carriers show the strongest effect of any single common gene variant in this domain.
- Replacing saturated fat with unsaturated fat reliably reduces CV events. The PREDIMED trial (Mediterranean diet vs. low-fat control) reduced cardiovascular events by ~30% over 5 years — strong randomised trial evidence.
⚠ Nuanced / contested
- "Saturated fat" is not one thing. Different saturated fatty acids have measurably different effects on LDL (see below).
- Food matrix matters. The same gram of saturated fat in cheese, in butter, in beef, and in a doughnut behaves differently in the body. The PURE study (2017) and several recent meta-analyses found weaker associations between total saturated fat and CV events than the historical hypothesis predicted.
- Replacement matters as much as removal. Replacing saturated fat with refined carbohydrates (the typical 1980s-90s response) does not reduce CV risk and may worsen it. Replacing with unsaturated fat or whole-grain carbs does help.
- Total dietary pattern dominates individual macronutrients. Modern preventive cardiology has moved from "limit saturated fat to X%" toward "follow a Mediterranean-style pattern overall" — because this captures the food-matrix and replacement-quality effects that single-nutrient targets miss.
Not all saturated fatty acids behave the same
The historical advice grouped all saturated fats together. The chemistry doesn't support this. The carbon-chain length matters substantially:
Palmitic acid (C16) and myristic acid (C14) — found in palm oil, beef fat, lard, and butter — are the most LDL-raising saturated fats. These are the ones the historical hypothesis was implicitly about.
Stearic acid (C18) — found in dark chocolate, beef fat, and cocoa butter — is essentially LDL-neutral. The liver converts much of it to oleic acid (the same monounsaturate found in olive oil). This is why dark chocolate consumption does not behave like "saturated fat" in epidemiological studies.
Short and medium-chain saturated fats (C4–C12) — found in dairy fat and coconut oil — have intermediate or variable effects depending on the food matrix they arrive in. Lauric acid (C12, in coconut oil) does raise LDL but also raises HDL more than longer-chain saturates do; the net CV effect remains debated.
Practical implication: dark chocolate (high stearic acid, plus flavonoids) genuinely is in a different category than processed meat (high palmitic acid, plus nitrates and TMAO precursors), even though both have similar total-saturated-fat figures on the label. Coconut oil sits awkwardly in the middle and is best used sparingly rather than as a staple.
The dairy paradox
This is the strongest example of food-matrix effects in action. In observational cohorts, total dairy fat intake is associated with neutral or even mildly favourable cardiovascular outcomes in most large studies — including PURE (2018), the Karolinska cohort (2017), and several Mediterranean cohorts. Fermented dairy (yoghurt, cheese) shows the strongest neutral-to-favourable signal; full-fat milk shows mixed results.
Several mechanisms have been proposed: the calcium in dairy binds fatty acids in the gut and reduces absorption; the milk fat globule membrane (MFGM) appears to attenuate the LDL-raising effect; fermentation generates bioactive peptides and short-chain fatty acids that may be cardioprotective; dairy-derived branched-chain fatty acids (like phytanic acid) activate PPAR-α with potentially favourable lipid effects.
What this does NOT mean: it does not mean you can drink unlimited whole milk or eat unlimited butter without consequence. The dairy paradox is most consistent in fermented dairy (yoghurt, cheese in moderate amounts) and in otherwise healthy dietary patterns. For someone with elevated LDL or established atherosclerosis, the conservative practical advice (lean toward semi-skimmed milk, moderate cheese, plain yoghurt) is still appropriate — see the milk panel in the diet section. The paradox is a useful corrective to the assumption that all saturated fat sources are interchangeable, not a licence to ignore the conservative position.
The current consensus position
Major preventive-cardiology bodies (AHA 2021, ESC 2021, NICE 2023) all maintain a recommendation to reduce saturated fat intake — but the framing has shifted meaningfully in the past decade:
- From "limit saturated fat to X% of energy" toward "follow a Mediterranean-style dietary pattern overall"
- From "all saturated fat is equally bad" toward "food source matters: limit ultra-processed and processed meat sources particularly"
- From "low-fat is better" toward "replacement quality matters more than total fat: swap toward olive oil, nuts, fish, not toward refined carbohydrate"
- From "LDL is the only target" toward "LDL plus apoB plus the inflammation axis plus the metabolic axis" (see the Integrated Risk card in section 3)
The reduction recommendation has not been withdrawn. What has changed is the recognition that the historical "saturated fat is the enemy" framing was both too narrow (it missed sugar, refined carbohydrate, and ultra-processed foods as bigger drivers in modern diets) and too broad (it treated dairy fat, dark chocolate, processed meat, and palm oil as interchangeable when they aren't).
Practical takeaway given the genuine uncertainty
The conservative position remains the right default. Limit ultra-processed and processed meat sources of saturated fat as a priority (these have the worst supporting biology — palmitic acid plus nitrates plus refined carbohydrates plus salt). Moderate dairy fat is acceptable for most people, with semi-skimmed milk a sensible default if you consume large volumes. Stearic-acid-rich foods (dark chocolate, lean beef) are not the same risk profile as palmitic-acid-rich foods and can be eaten with less restriction. Mediterranean replacement (olive oil, nuts, oily fish, vegetables) is more clearly evidence-based than blanket fat reduction. If you carry an APOE ε4 allele, all of the above tightens substantially — the targets and personalisation in your APOE marker feedback (section 3) reflect this. If your LDL is elevated or you have established atherosclerosis, lean toward the conservative end of all of the above regardless of dietary-pattern preferences.
In short: the science has become more nuanced, but the practical advice for someone managing cholesterol is largely unchanged. The nuance helps you make better choices within a fat-conscious diet — not abandon it.