Body Composition and DEXA: What Weight and BMI Leave Out

Bottom line

Weight and BMI only say how heavy you are. They do not say what the weight is made of. What matters more is whether there is enough muscle, how much visceral fat there is, and whether the bones are solid.

• The same weight can be a body of muscle or a belly of visceral fat, and the health trajectories are nearly opposite ¹¹⁴.
• What people call "losing weight" is really about losing fat while keeping muscle. One line is that the real aim is losing fat ¹.
• A soft tape measure finishes the first screen. The waist-to-height ratio is written into the British NICE guideline ¹⁷.
• DEXA can split the body into muscle, fat, and bone. The value is in watching change over time, not in a single score ⁴¹⁸.
• A shared position holds that resistance training and enough protein are the two pillars for keeping muscle while weight is coming down ³⁷¹⁰.

What it is

Body composition is weight taken apart: how much is muscle, how much is fat, and how much is bone and water.

BMI, the body mass index, is weight divided by height squared. The World Health Organization counts 25 and above as overweight and 30 and above as obesity ¹⁵. It weighs the truck. It does not say whether the load is steel or foam.

DEXA, dual-energy X-ray absorptiometry, also written DXA, sends two very-low-energy X-ray beams through the body and, from the difference in absorption, splits weight into lean tissue (muscle and organs), fat, and bone mineral ⁴.

Visceral fat, VAT, is the fat packed in the abdominal cavity around the liver and the intestines. It is not the same thing as the subcutaneous fat a hand can pinch, and its effect on metabolism is not the same either ².

ALMI and FFMI are two "muscle-mass indexes." ALMI is the weight of lean tissue in the limbs divided by height squared. FFMI is fat-free weight divided by height squared, so tall and short people share a ruler ⁶¹⁸.

Waist-to-height ratio is waist divided by height. NICE treats 0.4 to 0.49 as no increase in central-adiposity risk, 0.5 to 0.59 as increased risk, and 0.6 and above as a further increase ¹⁷. A rough way to remember it: once the waist passes half of height, you have entered the band of increased risk.

Why it matters

BMI misreads two kinds of people: lots of muscle as "fat," and little muscle with lots of fat as "fine."

WHO estimates that in 2022, 2.5 billion adults were overweight, 890 million of them with obesity, and that the overweight share rose from 25% in 1990 to 43%. BMI, it says, is only a stand-in for fat, and waist measures help ¹⁵.

A U.S. DEXA study of 11,687 adults found that at any BMI of 22 or above, less muscle meant higher body fat, more diabetes, and a higher risk of death ¹⁴. When body fat rose, BMI rose 30% to 40% less in people with less muscle. Adding waist, the extra risk attached to a high BMI disappeared ¹⁴. A consensus statement from the International Atherosclerosis Society and partners says decades of evidence show waist circumference adds information for predicting disease and death that is independent of BMI, and that it belongs among the routine vital signs measured in the clinic ¹⁶.

From age 50, one review states, muscle falls about 1% a year and strength about 3% ¹⁰. Weight can stay flat while muscle falls and fat rises.

Different perspectives

Main positions

One view holds that BMI is almost useless in a single patient. In large studies it may be the least bad tool. For the person in front of a clinician, it helps very little ². That view trusts DEXA and grabs four numbers: body-fat percentage, visceral fat, bone density, and ALMI ². Subcutaneous fat is largely genetic and loosely tied to metabolic health. Visceral fat tracks insulin sensitivity and liver fat ².

Weight loss is rewritten there as fat loss, and not bouncing back is the harder part ¹. Leptin, from fat cells, reports that reserves are adequate. When fat falls, the signal weakens and the brain treats the body as if it were starving ¹.

There are reservations about GLP-1 drugs such as semaglutide. Among 140 STEP 1 participants who had a DEXA, lean mass was about 39% of the weight lost. In 178 people in a SUSTAIN 8 substudy, the share was about 40% ³. For someone with obesity, if overall composition is improving, that cost can be accepted. For someone whose weight was not over the line, the benefit is small and the risk is large. People with sarcopenic obesity need extra caution. The countermeasure is more loading, strength training, and enough protein even when appetite has fallen ³.

Repeated scans beat one. In 3,075 adults ages 70 to 79, using DEXA plus clinical data to predict death within ten years, AUROC rose from 0.71 on a single set to 0.79 on repeated data. An AUROC of 0.5 is a coin toss. 1 is every case right ⁴.

In a solo episode, intake minus expenditure is granted as the fundamental formula for losing fat, and low-fat, low-carbohydrate, or intermittent fasting can all create a deficit if a person stays with it ⁵. The angle is the nervous system. NEAT — walking, standing, fidgeting — and shivering from cold, on this account, both promote fat mobilization ⁵. The episode description says small movements can burn thousands of calories of fat a day ⁵.

Andy Galpin calls DEXA the gold standard and says bioelectrical impedance comes reasonably close ⁶. He looks at FFMI: about 20 and above for men and 18 and above for women as enough muscle; below 17 for men and below 15 for women as a serious lack, if body fat is under 30% in men and under 35% in women ⁶. No high-intensity training for 48 hours before a scan. A common rescan interval runs from every six months to once a year ⁶.

Zachary Knight says that without more protein and without resistance training, about 25% to 33% of weight lost is muscle, from diet or from a drug. Serious lifting plus enough protein can nearly eliminate that loss ⁷. He also passes on the opposing view: a lighter body needs less muscle, and how serious the loss is remains unsettled ⁷. Layne Norton adds that once protein is already sufficient, leanness still mostly follows calories in and calories out ⁸.

Another account comes in through exercise. It describes a meta-analysis of 116 studies and more than 6,800 people with overweight or obesity ⁹. Each extra 30 minutes of aerobic work a week lined up with about 0.52 kilograms less weight, 0.56 centimeters less waist, and 0.37 percentage points less body fat. Subcutaneous and visceral fat fell too ⁹¹³. A clinically meaningful change in body fat and waist took at least 150 minutes a week ⁹.

Exercise alone often disappoints, that account explains. Hunger can replace a few hundred calories in minutes. People walk and stand less. The "constrained energy expenditure" hypothesis says the body compresses energy use elsewhere, so the total rises less than expected ⁹. Weight lost was similar across intensities, but waist, body fat, and visceral fat improved most at higher intensity ⁹. High-intensity intervals can lower fat mass and raise fat-free mass ¹¹.

On muscle: people who do not train lose about 4% of strength a year by age 75 ¹⁰. During fat loss, high protein plus resistance training lets more of what is lost be fat ¹⁰. The same source has also passed on a report that older adults with more muscle have a lower risk of early death, and that body composition predicts death better than BMI ¹².

Where views agree and differ

The views point the same way overall. The splits are what to measure with, how serious drug-related muscle loss is, and what fat loss depends on.

Agreement: BMI is not enough to judge one person ²⁶¹². Weight loss has to ask what was lost ¹⁷¹⁰. Resistance training and enough protein are the two pillars for keeping muscle ³⁷¹⁰. Visceral fat deserves more attention than weight ²⁹.

Disagreement one is body-fat scales. One view uses DEXA as the main tool ². Galpin says impedance comes reasonably close ⁶. NICE does not accept bioelectrical impedance in place of BMI for overall fat in adults ¹⁷. We stand with NICE.

Disagreement two is GLP-1 muscle loss. One view treats a lean-mass share near four-tenths as high enough to worry about ³. Knight treats it as still open, because a lighter person needs less muscle ⁷. Both say training and protein can reduce the loss ³⁷.

Disagreement three is what fat loss depends on. One account stresses aerobic amount and intensity ⁹. Another stresses nervous-system tools beyond energy balance, such as small movements and cold ⁵. Constrained energy expenditure is a hint that extra activity is partly offset ⁹.

PearHealth's synthesis

We adopt losing fat while keeping muscle. The waist-to-height ratio is a first step anyone can take. DEXA is a tool for the trend. We hold specific cutoffs and exaggerated fat-burning claims back.

What we adopt: replace "how many pounds came off" with three questions — did the waist change, was muscle kept, and did visceral fat fall ¹¹⁶. The ratio has a guideline behind it and costs almost nothing ¹⁷. During weight loss, training and protein are necessary companions. Those accounts agree ³⁷¹⁰.

What we hold back: FFMI lines such as 20 and 18 come from experience, and reference intervals differ by ancestry, age, and device. We treat them as a rough scale ⁶¹⁸. Thousands of extra calories a day from small movements is not backed, in what we read, by data at that magnitude, and it conflicts with constrained expenditure ⁵⁹. GLP-1 lean-mass shares come from substudies of one or two hundred people. DEXA "lean tissue" includes organs and water, so a layer still sits between that number and true muscle loss ³⁴.

Not everyone needs a DEXA. The International Society for Clinical Densitometry lists narrow uses, for example an expected large weight loss, and states that the effect on clinical outcomes is still uncertain ¹⁸.

How strong is the evidence

BMI's blind spot in a single person is hard evidence. More muscle and less visceral fat tracking a longer life is mostly observational. The causal chain is not closed.

Harder evidence:

• The link between aerobic exercise and weight, waist, and body fat comes from a meta-analysis of 116 randomized trials, the design best able to speak to cause ¹³.
• Waist predicting risk independently of BMI has the backing of an international consensus statement built on decades of research ¹⁶.

Evidence of medium strength:

• Muscle mass and death come from an observational cohort, so the two can only be said to show up together. Illness may come first and muscle second. Setting aside involuntary weight loss and chronic disease did not change the result ¹⁴.

Weaker or disputed evidence:

• The lean-mass share on GLP-1 drugs comes from trial substudies, each with only one or two hundred people ³.
• How much extra fat cold exposure and small movements can burn: we have not seen reliable long-term data ⁵⁹.

Measurement itself has error. The International Society for Clinical Densitometry notes that DEXA machines of different brands are not cross-calibrated, so results cannot be compared directly. Fasting or not, clothing, time of day, prior exercise, and whether the bladder was emptied all affect the result ¹⁸.

What you can do

Watch the trend. A specific plan is worth discussing with a doctor.

1. Measure waist on a regular schedule, calculate the waist-to-height ratio, and see which of NICE's three bands you fall in ¹⁷.
2. People of Chinese, South Asian, other Asian, Middle Eastern, Black African, and African-Caribbean ancestry can use lower BMI lines: NICE counts overweight from 23 and obesity from 27.5 in these groups, because cardiometabolic risk appears at a lower BMI ¹⁷.
3. If you have a DEXA, use the same machine under the same conditions when you can: the same fasting state, the same time of day, and no high-intensity training in the two days before ⁶¹⁸.
4. Reading a DEXA report, focus on body-fat percentage, visceral fat, ALMI or FFMI, and bone density, and compare them with the last scan ².
5. During weight loss, put resistance training and enough protein into the plan, rather than watching only how little you eat ³⁷¹⁰.

Questions worth raising with a doctor:

• If a GLP-1 drug is under consideration, how are muscle and bone tracked while it is used ³?
• If BMI is not high, but the waist-to-height ratio falls in the band of increased risk, are glucose, lipids, and other cardiometabolic checks worth doing ¹⁷?
• For someone over 65, how is BMI read together with the rest of the picture ¹⁷?

Common misconceptions

The most common misconception is treating weight as the whole of health.

• Misconception one: a lower number on the scale means better health. If three or four tenths of what was lost is lean mass, that is not all good news ³⁷.
• Misconception two: a BMI under the cutoff means there is nothing to worry about. A person with little muscle and a lot of fat can look acceptable on BMI while risk has already risen ¹⁴. Waist fills that blind spot ¹⁶.
• Misconception three: exercise does not take many pounds off, so it was wasted. The consensus statement notes that a fall in waist from exercise or diet can appear even when weight does not change, and visceral fat can fall when weight has not ¹⁶.
• Misconception four: a body-fat scale is as accurate as DEXA. Body-fat scales mostly use bioelectrical impedance, and the reading is pushed around by water in the body ⁶. NICE also does not accept it in place of BMI ¹⁷.
• Misconception five: one DEXA can score the body. A single number is moved by the machine and by the state of the body. Repeated measurements, and the trend, mean more ⁴¹⁸.
• Misconception six: the fat you can pinch on the belly is the most dangerous. One view holds the opposite. What matters more is the visceral fat you cannot see ².

How it connects to your blood work

Tape and imaging measure body composition. Visceral fat and muscle mass leave traces on a few blood markers.

• Fasting insulin (fasting-insulin), HOMA-IR (homa-ir), fasting glucose (fasting-glucose), and hemoglobin A1c (hba1c): one view ties visceral fat to insulin sensitivity ². The mortality model described there puts glucose, insulin, and hemoglobin A1c in with DEXA ⁴. Insulin resistance means more insulin is required to push glucose down.
• Triglycerides (triglycerides) and HDL cholesterol (hdl-c): fat piled in the viscera and liver tracks disordered lipids and other cardiometabolic risk ¹⁶.
• Alanine aminotransferase (alt) and gamma-glutamyl transferase (ggt): liver markers. More visceral fat often comes with signs of liver fat ². Read them with the waist.
• Creatinine (creatinine), estimated glomerular filtration rate (egfr), and cystatin C (cystatin_c): creatinine is a muscle waste, so muscle mass moves it ²⁰. eGFR is usually calculated from it. With a great deal of muscle, or at older ages, cystatin C may estimate kidney filtration more accurately ¹⁹.