Total Body Fat

You can have a normal weight, a normal BMI, and a clean blood panel, and still carry enough fat to raise your odds of diabetes and heart disease. Total body fat can catch this because it measures fat itself, not weight as a stand-in.

In US adults aged 20 to 49, body fat percentage measured by bioimpedance predicted 15-year death risk better than BMI did. In that same study a simple waist measurement predicted death nearly as well as the bioimpedance reading (adjusted hazard ratios of about 1.59 for waist versus 1.78 for body fat), so a tape measure captured much of the same signal. That doesn't prove a DXA scan will beat BMI in every group. It shows the larger point: direct body composition can reveal risk that height-and-weight math misses.

What This Number Measures

Total body fat is the amount of fat tissue you carry, reported either as kilograms of fat or as a percentage of body weight. It is not a molecule in your blood. DXA is a low-dose X-ray scan that separates fat, lean tissue, and bone across your whole body.

Two people can weigh the same and have very different amounts of fat. The scale and BMI treat a pound of muscle and a pound of fat as identical. This measurement does not, which is the whole point of ordering it.

Fat tissue isn't padding. It releases hormones and immune signals that can shift blood sugar, cholesterol-carrying particles, and blood pressure. The more fat you carry, and the more of it lies deep in your abdomen, the louder those signals tend to get.

Why the Scale and BMI Miss This

BMI is a rough stand-in for fatness, and it fails in a specific way. It is better at spotting some people with high body fat than at reassuring you that body fat is low. In older adults, using DXA as the comparison, a BMI in the obese range identified only about a third of those who actually had high body fat. The threshold matters as much as the metric: a BMI of 25 or higher caught most people with high fat in the same cohort.

The caveat matters. Total fat is not a better risk marker for every purpose. In a study of more than 60,000 adults, BMI predicted death from heart disease as well as or better than body-fat measures based on skinfolds or underwater weighing. In UK Biobank, body fat measured by bioimpedance lost much of its heart-disease signal once waist-to-height ratio was included. Where fat sits often matters more than total fat.

This number helps most when BMI is likely to lie: hidden fat in someone with normal weight, a weight-loss plan that may be stripping muscle, and younger adults, where BMI performed poorly in one 15-year mortality study.

Type 2 Diabetes

Excess fat helps cause type 2 diabetes. Genetic studies using BMI as a proxy for general body fat found that higher inherited tendency toward adiposity raises diabetes risk, rather than merely traveling alongside it. These studies don't measure DXA total fat, but they strengthen the causal case.

Where risk starts climbing differs by sex and population. In about 5,600 Chinese adults whose body fat was estimated with bioimpedance, higher groups of total body fat carried higher diabetes risk in men, and the highest group did in women. Other groups land on different numbers. The useful part is the direction: more fat, more diabetes risk, with women carrying more fat than men at comparable levels of risk.

Normal weight can be misleading. In NHANES adults 40 and older, DXA-measured high body fat within a normal BMI was linked to more abnormal blood sugar than overweight with lower body fat. Adding fat measurement to BMI improved risk classification, though the study was cross-sectional.

Heart Disease and Stroke

Genetic evidence using BMI and waist-to-hip ratio as proxies supports a causal link between higher body fat and coronary heart disease. Central fat also appears more relevant for ischemic stroke than total body fat alone.

One useful refinement is the balance between fat and muscle. In nearly 469,000 UK Biobank participants, using bioimpedance-estimated fat and muscle, a higher fat-to-muscle ratio predicted more cardiovascular disease and earlier death. Part of the risk ran through the same cholesterol and inflammation markers a blood panel can show. Your fat number means more when you read it next to your muscle mass.

The Mortality Curve Bends at Both Ends

Total body fat is not simply lower is better. Pooled cohort data using several body-fat methods found a J-shaped pattern: very high fat carried risk, and very low body fat also looked worse. The low end is harder to read because illness, smoking, and low muscle can make people lighter before they die. The signal also weakens with age: in those same pooled cohorts, higher body fat tracked with higher death risk in general adult populations but not in adults over 60, where the link disappeared.

That low end looks less paradoxical when you separate fat from lean mass. A study of about 38,000 US men estimated fat and lean mass from body measurements validated against DXA. Much of the extra risk at low body weight came from low lean mass, not low fat. Excess fat raises risk; lean mass often protects. Genetic studies point the same way: when fat mass is set by inherited variants rather than by illness, higher fat mass tracks with higher mortality in a nearly straight line, and the observed J-shape mostly reflected smoking and inactivity. Among never-smokers and physically active people the relationship was close to linear. Read your result as a pattern, not a scorecard, and never in isolation from lean mass.

Source: Lee et al. 2018 (BMJ); Mainous et al. 2025 (Annals of Family Medicine); Jayedi et al. 2022 (International Journal of Obesity).

Takeaway: chasing the lowest possible fat number is the wrong goal. The useful target is enough fat loss to move you out of the high-risk pattern while you hold onto or build lean mass. Tracking fat and lean mass together beats tracking either alone.

Not All Fat Sits in the Same Place

Total body fat counts everything: the fat under your skin and the fat deep around your organs. Those depots do not carry equal risk. Visceral fat around your organs is more tightly tied to insulin resistance and heart disease than total fat alone. Fat that builds up inside the liver is another clue about metabolic risk, but a body-composition scan cannot single it out; that takes a different test.

This test gives you the total, which is the right place to start. If your total is high, a companion visceral fat measurement tells you how much of it is in the riskier location. Two people with the same total fat can carry different risk depending on how it is distributed.

Why One Reading Is Not Enough

A single fat measurement is a snapshot, and body composition wobbles from day to day. An older study using daily bioimpedance found within-person body fat percentage varied by around 10% in relative terms, not 10 percentage points. Bioimpedance estimates fat from body water. Modern standardized methods are far tighter, with day-to-day differences under about 1 to 2 percentage points and DXA the most precise. Either way, small changes between two readings can be noise.

The value is in the trajectory. A single number tells you roughly where you are. A series of numbers, taken the same way under the same conditions, tells you whether your diet, training, medication, or surgery is shifting fat, and whether you are protecting muscle while you do it.

When a Reading Can Mislead

A few things can distort the number and send you to the wrong conclusion. Lead with the biggest one: never compare across methods.

• Method mismatch: a DXA scan and a home bioimpedance scale can disagree by several percentage points on the same body. They track better across a group than for one individual, so pick one method and stick with it.
• Hydration and recent activity: bioimpedance estimates fat from body water, so heavy dehydration can nudge the reading, more so in women. In controlled tests, a normal meal, a drink, or a workout moved the number only modestly, usually within about 2 percentage points. DXA is even less affected. Keeping conditions similar still helps.
• DXA scanner differences: DXA results can vary by scanner maker, model, and software version. If you are tracking change, use the same facility when you can.
• Day-to-day biological variation: even under good conditions, body fat percentage drifts within a person, so a small change between two readings may mean nothing.
• High body fat and simpler tools: in one youth study with overweight or obesity, a consumer bioimpedance device underestimated DXA body fat by about 10%. Simpler tools can flatter the reading in some groups.

What to Do With an Out-of-Pattern Result

A high or rising fat number is a prompt, not a diagnosis. The next move is to find out whether the fat is already doing metabolic damage. Order the markers that show it: a measure of insulin resistance, HbA1c for blood sugar, triglycerides, and ApoB for heart risk. Add a lean mass reading so you can see the fat-to-muscle balance.

The pattern matters more than any single value. High fat with rising insulin resistance, climbing triglycerides, and a shrinking muscle share is the combination that calls for action. High fat with strong metabolic markers and strong muscle is worth watching closely. If several markers are drifting the wrong way together, bring in a clinician who manages metabolic risk and treat the trend seriously rather than wait for a diagnosis.