Divide your waist measurement by your height. Keep the answer under 0.5 — your waist should be less than half your height.
That is the rule, it is a good rule, and almost nobody tells you the part that decides whether your answer means anything: it depends on where you put the tape. The threshold was validated with the tape at the midpoint between your lowest rib and the top of your hip bone. Measure at your navel instead and you get a different number — in women, by as much as two centimetres, which is enough to move you across the line in either direction.
This article is about what the tape measure genuinely tells you about visceral fat, which numbers were established on which measurement, and one finding from 2026 that complicates the reassuring story about not needing a scan.
Quick fact: In Health Survey for England data, 31% of adults with a BMI in the “healthy” range had a waist-to-height ratio of 0.5 or above — 36% of men and 27% of women. After adjusting for age, sex and BMI, that group had higher HbA1c, a worse cholesterol ratio and systolic blood pressure 3.4 mmHg higher.
Two kinds of fat: visceral fat versus the kind you can pinch
Subcutaneous fat sits just under the skin, spread across the body. It is the fat you can pinch, and it is relatively metabolically quiet.
Visceral fat sits deep in the abdomen, wrapped around the liver, pancreas and intestines. You cannot pinch it, and it behaves differently: it is metabolically active tissue that releases hormones and inflammatory signals, and it is the depot tied most directly to type 2 diabetes, fatty liver disease and cardiovascular disease.
That is the core reason the scale misleads. Total body weight cannot distinguish the fat that mostly sits there from the fat that is actively doing something.
The measurement each number was actually built on
Here is the problem that runs through this entire subject. There are four common measures, and the famous findings were each established on a different one. They are not interchangeable, and articles that mix them — including the earlier version of this one — hand you a threshold from one study and a justification from another.
| Measure | What it is | What was actually shown with it | On which protocol |
|---|---|---|---|
| BMI | Weight ÷ height² | J-shaped mortality curve, nadir around 24.9 — risk rises at both ends | None needed |
| Waist circumference | Tape around the abdomen | Best single anthropometric discriminator of hypertension in a 2026 NHANES analysis (AUC 0.791), beating BMI | Depends — NIH and WHO protocols differ |
| Waist-to-height ratio | Waist ÷ height | 31% of healthy-BMI adults above 0.5, with worse markers after adjusting for BMI | WHO midpoint, by trained nurses |
| Waist-to-hip ratio | Waist ÷ hips | The strongest, most consistent mortality association in 387,672 UK Biobank adults | Waist and hips, so both sites matter |
Note the last row against this article’s own title. The most-quoted study in this field tested waist-to-hip ratio, and did not analyse waist circumference at all. It is excellent evidence that body shape beats BMI. It is not evidence about the specific number a tape around your waist gives you.
The two protocols, and why it matters that nobody says which
There are two standard places to put the tape, and they are not the same place.
| NIH protocol | WHO protocol | |
|---|---|---|
| Where | At the superior border of the iliac crest — the top edge of the hip bone | At the midpoint between the lower border of the ribs and the iliac crest |
| Used by | US guidelines, NHANES | The Health Survey for England, most European work |
| Difference between them | Not significantly different in men | 1.8 to 2 cm in women |
Two things follow, and they are the most practically useful things in this article.
The navel is neither of these. Using your navel as the landmark is extremely common advice — the earlier version of this page offered it as interchangeable with the WHO midpoint — and it is not a validated site. Worse, the navel’s position on the torso shifts with abdominal fat, which means the landmark moves as the visceral fat you are trying to track changes. For tracking yourself over months, that is the one property you least want in a reference point.
Pick one protocol and never change it. The consensus statement is explicit: the protocol used for your first measurement should be used for every later one. A 2 cm shift from changing landmarks will look exactly like real progress or real deterioration, and you will have no way to tell the difference. Self-measurement is fine, incidentally — self-measured and technician-measured waists correlate strongly.
If you want a single recommendation: use the WHO midpoint, because that is the protocol the 0.5 waist-to-height threshold was validated on.
Interpreting your waist-to-height ratio
| Waist-to-height ratio | General interpretation |
|---|---|
| Under 0.4 | May indicate underweight; worth discussing rather than targeting |
| 0.4 – 0.49 | Healthy range |
| 0.5 – 0.59 | Increased risk |
| 0.6 and above | Substantially increased risk |
General guidance. Thresholds vary by population and source, and these bands were not derived from a single study — discuss your own numbers with a provider.
The useful thing about this ratio is that it needs no age or sex-specific chart, because dividing by height scales it automatically. That is a real advantage over the alternative, which has aged less well: the familiar waist cut-offs of 102 cm for men and 88 cm for women come from a 1995 study, were developed in White adults, and are applied regardless of BMI category. The consensus group’s own working paper notes the criticism that a single threshold is inadequate across BMI categories, and — pointedly — that sex-specific thresholds can make a person with a normal BMI and a large waist appear lower risk than they are.
Why BMI falls short, with the actual numbers
The 2023 UK Biobank analysis by Irtiza Khan, Guillaume Paré and colleagues in JAMA Network Open is the best single demonstration, and it contains more than the headline.
| Marker | Mortality hazard ratio per SD | Shape of the association |
|---|---|---|
| BMI | 1.14 (1.13–1.16) | J-shaped, nadir at 24.9 |
| Fat mass index | 1.15 (1.13–1.17) | J-shaped, nadir at 6.15 |
| Waist-to-hip ratio | 1.41 (1.38–1.43) | Linear — rising the whole way, no safe lower end |
Three findings sit underneath that table and none of them made it into most coverage.
The shapes differ, and that is the real story. BMI’s J-curve is why “overweight” keeps appearing protective in studies and why BMI’s linear risk estimate understates it. Waist-to-hip ratio rose monotonically: no nadir, no flat zone.
It held at every level of BMI. The waist-to-hip ratio association did not vary across BMI quantiles, whereas BMI’s and fat mass index’s associations did. That is the actual statistical case for measuring your waist as well as weighing yourself — your shape carries risk information your BMI does not contain, whatever your BMI is.
They ran a causal analysis, not just a correlation. Using Mendelian randomisation — which uses genetic variants to approximate a natural experiment — genetically determined waist-to-hip ratio gave an odds ratio of 1.51 (1.32–1.72) against BMI’s 1.29 (1.20–1.38), with heterogeneity P = 0.02. The effect was far stronger in men (OR 1.89) than women (1.20).
One detail in the same table is worth pausing on, because it shows how much the adjustment does. Before adjustment, fat mass index had a hazard ratio of just 1.06; after adjustment it rose to 1.15. BMI barely moved, from 1.17 to 1.14. Waist-to-hip ratio went from 1.52 to 1.41 and stayed far ahead of both. A marker whose apparent risk changes direction depending on what you control for is a marker to be careful with, and two of the three here behave that way.
The study’s own limitations are the familiar ones: participants were a genetically homogeneous White British population, adiposity was measured only once at baseline so changes over time could not be assessed, and the authors note their sample was too small to study those changes even if it had been measured twice. The causal arm rested on 25,297 deaths matched to the same number of controls.
The problem with dividing by height
Before adopting the 0.5 rule wholesale, it is worth knowing what the denominator does, because it is a criticism the rule’s own supporting data illustrates.
In the Health Survey for England analysis, the group above 0.5 differed from the group below it in more than their waists. They were 17 years older on average — 54 against 37 — and they were about 3 cm shorter, at 167 cm against 170. Their weights differed by only 3 kg and their BMIs by 1.7 points.
Some of that is simply what ageing does to body shape, and the headline comparison was adjusted for age, sex and BMI, so the worse blood markers survive it. But the height difference is a property of the arithmetic rather than of anyone’s health: dividing by height means a shorter person crosses 0.5 at a smaller waist than a taller person does. A 160 cm adult reaches the threshold at an 80 cm waist; a 185 cm adult not until 92.5 cm. Whether shorter people genuinely carry more risk at a smaller absolute waist, or whether the ratio simply flags them sooner, is a live criticism of waist-to-height guidance and not a settled question.
That is not a reason to abandon the measure — it remains the best simple option, and it needs no charts. It is a reason to treat 0.49 and 0.51 as the same reading rather than as a verdict, and to care far more about the direction your own number moves over months than about which side of a line it sits on today.
Do you need a scan to measure visceral fat? The answer changed in 2026
The reassuring version of this question — a free tape measure captures essentially all the signal, so imaging is unnecessary — is the one the earlier version of this article gave, and it now needs qualifying.
| Study | Outcome studied | Did direct visceral measurement add anything beyond the tape? |
|---|---|---|
| Dejenie et al., Hypertension Research, 2026 — NHANES, 7,817 adults | Hypertension | Yes, substantially. Waist circumference alone gave AUC 0.791. Adding DXA visceral fat raised it to 0.865 — a gain of 0.074, statistically significant. Adding total fat mass added nothing (gain of 0.002, P = 0.37) |
| Same study, sensitivity analysis | Hypertension by measured blood pressure only | The visceral advantage shrank to a gain of 0.012 |
| AJCN, 2025 — NHANES, >8,000 adults | Metabolic syndrome, type 2 diabetes, insulin resistance | I could not verify this. The paper and a companion comparison piece exist and address exactly this question; neither full text was reachable, so the direction of their finding is not established here |
So the honest position is more interesting than either extreme. Visceral fat specifically — not total body fat — appears to carry information a tape measure misses, at least for hypertension, and at least on one outcome definition. The contrast within that single study is the striking part: measuring visceral fat added something, and measuring total fat added nothing at all. Per standard deviation, visceral fat gave an odds ratio of 2.65 against roughly 1.81–1.85 for BMI, waist circumference and waist-to-height ratio alike.
What has not changed is the practical conclusion, and the 2026 authors state it themselves: simple anthropometrics should still underpin screening. The tape measure is where you start, it costs nothing, and it does the large majority of the work. It is just not true that imaging adds nothing.
Two limits on that 2026 study are worth stating, because they decide who it applies to. It is cross-sectional, so it cannot establish that the fat came before the blood pressure — 39.9% of the sample already had hypertension when they were measured. And it covered adults aged 20 to 59 only, because whole-body DXA scans were not available for people over 59 in those survey cycles. The age group carrying the highest cardiovascular risk is therefore the group this study says nothing about. Its discrimination figures were also calculated within the same dataset rather than validated against an independent cohort.
There is one further caveat that belongs here and is rarely mentioned: adding waist circumference to established cardiovascular risk calculators — Framingham, the Pooled Cohort Equations — does not improve their prediction. Those models are 63–80% driven by non-modifiable factors such as age and sex. Waist measurement is a useful personal signal and a useful tracking tool; it is not an upgrade to formal risk scoring.
How to measure, and why each step matters
| Step | Why |
|---|---|
| Measure against bare skin | Even light fabric adds to the reading and varies between sessions |
| Stand upright, feet shoulder-width apart | Posture changes abdominal shape |
| Use the WHO midpoint — halfway between your lowest rib and the top of your hip bone | It is the protocol the 0.5 threshold was validated on, and the landmark does not move as your abdomen changes |
| Keep the tape horizontal all the way round | A tilted tape reads long, and tilts differently each time |
| Measure at the end of a normal breath out | Sucking in gives a lower number and a meaningless trend |
| Snug, not compressing | Compression reads short; loose reads long |
| Same protocol, same spot, same time of day, every time | The consensus statement’s own instruction — comparability matters more than which protocol you chose |
Re-measure every two to four weeks, not daily. Waist circumference changes slowly, and the day-to-day variation you would be reacting to is mostly technique and meals.
What reduces visceral fat — and how strong each case is
| Approach | Status of the evidence |
|---|---|
| Consistent aerobic activity | The most commonly recommended, with a caveat worth knowing: the consensus working group’s own literature notes that several studies failed to show a dose-dependent reduction in waist circumference with greater exercise quantity or intensity. So activity helps; “more and harder helps proportionally more” is not established |
| Resistance training | Preserves muscle during fat loss and improves insulin sensitivity. See body recomposition for what that changes and what it costs |
| Reducing refined carbohydrate and added sugar | Plausible mechanism via hepatic fat conversion. How much added sugar is too much covers the intake question |
| More daily movement outside workouts | The largest and most variable part of daily expenditure for most people — see NEAT |
| Adequate sleep | Short sleep is associated with visceral accumulation; I did not verify a specific effect size |
| Limiting alcohol | Associated via liver-mediated pathways; same caveat |
| Abdominal exercises | No. Spot reduction is not supported. Crunches build the muscle under the fat |
A note on expectations: even a modest reduction in waist circumference is worth having, in any BMI category, according to the consensus recommendations — which is a more useful target than a specific number, because the direction is what you control.
What people actually notice
The most common experience is the disconnect between scale and tape, especially when resistance training has been added: the waist shrinks while the weight barely moves. That is a favourable pattern, not a stall — losing visceral fat while holding onto muscle produces a better metabolic profile at almost the same body weight — and tracking only the scale reads real progress as failure.
The second is surprise at their own number — the 31% figure at the top of this article is exactly that experience, quantified. Nearly a third of people whose BMI says they are fine are above the 0.5 threshold.
The third is alarm caused by measurement error. Measuring at a different landmark, over clothing, or after a large meal can shift the number by a centimetre or more, which looks like rapid change and is not. This is the whole reason the protocol section above exists.
A note on the numbers in this article
| Figure | Status |
|---|---|
| 31% of healthy-BMI adults with WHtR ≥ 0.5; the adjusted HbA1c, cholesterol and blood pressure differences | Verified — Health Survey for England, WHO midpoint protocol |
| All mortality hazard ratios, the J-shaped vs linear shapes, and the Mendelian randomisation odds ratios | Verified in the 2023 JAMA Network Open paper |
| The two measurement protocols and the 1.8–2 cm difference in women | Verified |
| 102 cm / 88 cm thresholds, their 1995 origin and White-adult derivation | Verified |
| Waist circumference adds nothing to Framingham or the Pooled Cohort Equations | Verified |
| Failure to find dose-dependent waist reduction with more exercise | Verified |
| All AUC and odds ratio figures for visceral fat versus waist circumference | Verified in the 2026 Hypertension Research paper |
| “No clear advantage of direct visceral measurement” (AJCN 2025) | Not verified — paper unreachable; direction not established here |
| Waist-to-height ratio outperforming BMI for heart disease in Lancet Regional Health, 2025 | Not checked — not cited in this version |
| Waist ratios and physical performance in Aging, 2025 | Not checked — not cited in this version |
| Visceral fat responding better to intervention than subcutaneous fat | Not verified — stated as a general direction only |
| Sleep, stress and alcohol effects on visceral fat | Not verified with effect sizes — given as associations |
When should you see a doctor?
Bring your waist-to-height ratio to a doctor if it is 0.5 or above, particularly alongside raised blood pressure, raised blood sugar or an abnormal cholesterol panel, since that combination points toward metabolic syndrome and benefits from clinical assessment rather than self-management. Raise it too if your waist has grown noticeably without a corresponding change in weight or habits.
Because waist circumference is still not routinely recorded in most consultations, it helps to have the sentence ready. Something like: “My BMI is in the normal range, but my waist divided by my height is above 0.5 — can we check whether my blood pressure, blood sugar and cholesterol should be looked at together?” That names your actual measurement, names the threshold, and asks for the specific panel the research points to. It is a more productive opening than “I’m worried about belly fat,” and it gives a clinician something to act on.
Frequently asked questions
What is a healthy waist-to-height ratio? Under 0.5 — your waist less than half your height. It needs no age or sex-specific chart because dividing by height scales it. Measure at the midpoint between your lowest rib and the top of your hip bone, which is the protocol the threshold was validated on.
Where exactly do I put the tape? The two standard protocols are the WHO midpoint (lowest rib to hip bone) and the NIH site (top of the hip bone). In men they give similar numbers; in women they differ by up to 2 cm. The navel is neither, and its position moves as abdominal fat changes. Pick one and use it every time.
Can a normal BMI hide high visceral fat? Yes, and it is common. 31% of adults with a BMI in the healthy range had a waist-to-height ratio of 0.5 or above, and their blood markers were worse even after adjusting for BMI.
Is visceral fat worse than subcutaneous fat? Metabolically, yes. It is active tissue releasing hormones and inflammatory signals, and it is tied most directly to type 2 diabetes, liver disease and cardiovascular disease. A 2026 analysis found it carried risk information for hypertension that waist circumference alone did not.
Do I need a scan? For screening and tracking, no — simple measurements should underpin that, and the 2026 authors say so themselves. But the claim that imaging adds nothing is too strong: adding DXA visceral fat to waist circumference meaningfully improved hypertension discrimination in that study.
Can I lose visceral fat without losing weight? Yes, and the waist-versus-scale disconnect is the most commonly reported experience in this area. Breaking a weight loss plateau covers why the scale can stall while composition improves.
Do ab exercises reduce belly fat? No. Spot reduction is not supported. They strengthen the muscle beneath the fat.
The bottom line
BMI has been the default for fifty years while missing where the fat is, and a tape measure plus one division closes most of that gap for free. The signal it picks up is visceral fat, which is the part that behaves like a disease process rather than like storage. Two things are worth carrying away beyond that.
The first is procedural and it is the thing nobody tells you: the threshold you are judging yourself against was measured at a specific spot on the body, and your number only means what it is supposed to mean if you use that spot — and the same spot every time after that.
The second is the sentence to take to your next appointment. Your waist is probably not in your medical record, your BMI almost certainly is, and the research says the first one carries information the second does not, at every level of the second. So say it out loud: my BMI is normal, my waist-to-height ratio is above 0.5, can we look at blood pressure, blood sugar and cholesterol together? That is the whole article, in one question you can ask.
This article is for general informational purposes and is not a substitute for personalised medical advice. Discuss your measurements and your individual risk with a healthcare provider. If measuring or tracking your body has become a source of distress, that deserves support in its own right.
References
- Khan I, Chong M, Le A, et al., Paré G. Surrogate adiposity markers and mortality. JAMA Network Open, 2023;6(9):e2334836 — 387,672 UK Biobank participants plus a 50,594-person Mendelian randomisation cohort; all hazard ratios, the J-shaped versus linear associations, and the causal analysis. Full text
- Ashwell M, Gibson S. Nearly one third of adults in the “healthy” BMI range are at early cardiometabolic risk according to their waist-to-height ratio. Proceedings of the Nutrition Society, 2019;78(OCE1):E29 — Health Survey for England data, the 31% figure, and the WHO midpoint protocol
- Dejenie TA, Mannan H, Atlantis E. Incremental value of DXA-derived visceral versus total adiposity beyond routine anthropometric assessment for hypertension discrimination. Hypertension Research, 2026 — 7,817 NHANES adults, the AUC comparisons, and the odds ratios per standard deviation
- Ross R, Neeland IJ, Yamashita S, et al. Waist circumference as a vital sign in clinical practice: a Consensus Statement from the IAS and ICCR Working Group on Visceral Obesity. Nature Reviews Endocrinology, 2020;16(3):177–189 — the measurement protocols, the threshold history, and the stated criticisms, summarised by the American College of Cardiology
- Cross-sectional analysis of visceral adipose tissue associations with obesity-related disease, and the accompanying comparison of visceral adipose tissue with waist-based anthropometry. American Journal of Clinical Nutrition, 2025 — both papers address whether direct visceral measurement outperforms waist-based measures; neither full text was reachable, so no finding from either is quoted here
