BMI vs body fat vs waist ratio: which number matters?

Three questions, three numbers

It helps to be blunt about what each measurement physically is. BMI is arithmetic on a scale reading and a height. It has never seen your body. A body fat percentage is the output of a device or an equation that measured something else entirely (X-ray attenuation, body volume, electrical resistance, tissue thickness, circumference) and converted it using population assumptions. A waist measurement is the only one of the three that is a direct measurement of you, with no equation between the tape and the number.

That difference in kind explains a lot of the argument. BMI is criticised for being crude, but it is at least transparent: two inputs, one formula, no hidden assumptions. Body fat percentage feels like the truth because it sounds like a direct statement about your tissue, but every route to it carries an error band of one to eight percentage points depending on the method. The waist reading is the humblest of the three and, per second spent, the most informative.

So the useful question is not 'which number is best' but 'which question am I asking'. If the question is whether a population is getting heavier, BMI. If the question is whether the weight you just lost was fat or muscle, body fat percentage, tracked consistently. If the question is whether your body composition carries cardiometabolic risk, the waist measures, first and by a clear margin.

What each number is good for

BMI is the fastest and least informative: two inputs, no tape, decades of population data behind its thresholds. As a first-pass screen and a language shared with every doctor on earth, it earns its place; as a verdict on an individual body, it never did. Body fat percentage is the composition truth (it distinguishes the lean 90 kg from the fat 90 kg), but every way of measuring it is an estimate with an error band, from ±1–2 points (DEXA) to ±3–4 (tape method).

The waist ratios carry the most underrated signal. Visceral fat, around the organs, drives cardiometabolic risk out of proportion to fat elsewhere, and a waist measurement captures it directly. This is why studies keep finding waist-to-height ratio predicts heart disease and diabetes as well as or better than BMI, and why the 2025 obesity framework made a distribution measure mandatory rather than optional.

Laid out side by side, the trade-offs are easier to see than to argue about. Note in particular the inputs column: the measures that require the most equipment are not the ones that carry the most risk information, which is the single most counter-intuitive fact in this field.

When the numbers disagree

The disagreements are where the insight lives. High BMI + low waist ratio is the athlete pattern: mass without central fat. The 2025 framework explicitly reads this as not meeting adiposity criteria. Normal BMI + high waist ratio is the reverse and more dangerous case (sometimes called normal-weight central obesity): total mass unremarkable, storage pattern adverse; research associates it with mortality risk rivalling or exceeding frank obesity, and BMI alone waves it through. Normal BMI + high body fat + soft waist is the sarcopenic pattern common with age; the fix there is building muscle, not losing weight.

Set out as a matrix, the four combinations of BMI and waist each point to a different next measurement. This is the same logic the 2025 clinical obesity criteria formalised: a raised BMI confirms excess adiposity only when a fat-distribution measure is raised alongside it, and a normal BMI does not rule adiposity out when the distribution measure is elevated.

The waist measures in detail

There are three waist-based numbers and they are not interchangeable. Raw waist circumference has sex- and ethnicity-specific thresholds: the WHO marks increased risk at 94 cm for men and 80 cm for women, with substantially increased risk at 102 and 88, while the International Diabetes Federation's central obesity criteria use 94 cm for Europid men but 90 cm for South Asian, Chinese and Japanese men, with 80 cm for women across all those groups. Those differences are not arbitrary; they reflect genuine population differences in how much visceral fat accompanies a given circumference.

Waist-to-height ratio divides that circumference by your height, which removes most of the need for separate thresholds. One line at 0.5 applies to both sexes and across heights, with a second band at 0.6 marking further increased risk. That simplicity is why it has spread through clinical guidance so quickly, and why it is the most efficient single number in this entire guide: two measurements you can take in a minute, one threshold to remember.

Waist-to-hip ratio adds a hip measurement and asks a slightly different question: not how large the middle is, but how the fat is distributed between the trunk and the lower body. The WHO marks substantially increased risk at 0.90 for men and 0.85 for women. It is useful for characterising shape, but it carries the technique error of two measurements rather than one, and it can look reassuring in someone whose waist and hips are both large.

One caution applies to all three. A tape reads total abdominal circumference and cannot separate subcutaneous fat from the visceral fat that carries most of the risk. Across a population the proxy works extremely well, which is why these measures predict outcomes so reliably. For one individual, it is still a proxy.

What the error bands mean for one reading

Every number in this guide is a range wearing the costume of a point. A tape-based body fat estimate of 22% is honestly saying 'most likely around 22, plausibly 18 to 26, and possibly offset by a few points in a direction that depends on my build'. A smart scale's 22% is saying something considerably vaguer. A DEXA's 22% is the tightest of them, and still carries a point or two, and will not match a DEXA on a different machine.

BMI is the odd one out here: as arithmetic it has no measurement error worth discussing beyond the accuracy of your scale and your height. Its error is of a different kind entirely: it is precisely computed and may still be a poor description of your body, because the formula cannot see composition. A BMI of 27 is exactly 27 and tells you almost nothing on its own.

The consequence for practice is the same in every case: do not act on one reading, and do not act on a difference smaller than the error band. If your body fat estimate moves from 22% to 21% in a month, nothing has been demonstrated. If it moves from 22% to 19% across four consistent monthly readings, something has. Trends survive error bands that snapshots do not, because a consistent bias subtracts itself out of the comparison.

The practical answer

For most people, two numbers cover it: waist-to-height ratio (risk, distribution) plus a consistent body fat estimate (composition, trend). BMI then costs nothing to compute alongside and remains the term your doctor speaks. Our BMI calculator effectively runs this synthesis for you: add a waist measurement and it applies the multi-measure framework, flagging concordance or discordance instead of leaving three numbers staring at each other.

If you are training for muscle, add a third: fat-free mass index, which expresses lean mass relative to height and is the number BMI garbles most badly. It is the honest way to watch muscle accumulate across years, and it settles the recurring argument about whether a rising scale weight is progress or drift.

What none of these numbers do is describe your health, and it is worth saying plainly. They are screens: they identify patterns worth a closer look. A waist-to-height ratio over 0.5 is a reason to check blood pressure, lipids and fasting glucose, which measure the thing the ratio is only predicting. A body fat percentage in the average band says nothing about fitness, strength or cardiovascular health, all of which are separately measurable and all of which matter.

How to run all three in ten minutes

The whole set costs one morning a month and a five-dollar tape. Take the measurements in a fixed order, under fixed conditions, and record the raw numbers rather than only the calculated results. If you later improve your technique or a threshold changes, raw measurements can be recomputed and derived ones cannot.

• Weigh first thing, after the bathroom, before food or drink, on the same scale.
• Measure height once a year without shoes; it does not need re-measuring monthly, but it does change over decades.
• Waist at the rib–hip midpoint or the navel, three passes at the end of a normal exhale, averaged.
• Hips at the widest point of the buttocks, and neck just below the larynx. Both are needed for the ratios and the body fat estimate.
• Compute BMI (weight ÷ height²), waist-to-height (waist ÷ height), waist-to-hip (waist ÷ hip) and a body fat estimate from the same readings.
• Read the BMI and waist pair against the discordance matrix above before looking at anything else.
• File the raw measurements. Compare against three months ago, not against last month.

Frequently asked questions

Put it into practice

Run your own numbers through the BMI calculator, the body fat calculator and the waist-to-height calculator. Related reading: BMI explained, BMI for athletes and Body fat methods.

Sources

1. Rubino F, et al. Definition and diagnostic criteria of clinical obesity. Lancet Diabetes Endocrinol 2025;13:221–262. doi.org/10.1016/S2213-8587(24)00316-4
2. Ashwell M, Gibson S. Waist-to-height ratio as an indicator of early health risk. BMJ Open 2016;6:e010159. doi.org/10.1136/bmjopen-2015-010159
3. Sahakyan KR, et al. Normal-weight central obesity: implications for total and cardiovascular mortality. Ann Intern Med 2015;163:827–35. doi.org/10.7326/M14-2525
4. Yusuf S, et al. Obesity and the risk of myocardial infarction (INTERHEART). Lancet 2005;366:1640–9. doi.org/10.1016/S0140-6736(05)67663-5
5. Ross R, et al. Waist circumference as a vital sign in clinical practice: a consensus statement. Nat Rev Endocrinol 2020;16:177–89. doi.org/10.1038/s41574-019-0310-7
6. Alberti KGMM, Zimmet P, Shaw J. Metabolic syndrome: a new world-wide definition. A consensus statement from the International Diabetes Federation. Diabet Med 2006;23:469–80. doi.org/10.1111/j.1464-5491.2006.01858.x
7. WHO. Waist circumference and waist–hip ratio: report of a WHO expert consultation. Geneva, 2008. www.who.int/publications/i/item/9789241501491