A useful report conversation starts with three details: the exact measurement, the units and the reason the scan was performed. Keep those details beside the number as you read. The reference material below explains how published comparisons work; it cannot decide whether your individual result needs treatment.
If your report uses a different scanner, region or reference population, ask the clinic which comparison applies. Avoid substituting a convenient online table just because the units look similar. Discuss bone density or visceral fat findings with your GP.
Read the measurement
Fat mass index adjusts measured fat mass for height. It is calculated as fat mass in kilograms divided by height in metres squared, and is reported in kg/m².
| Source category | Men | Women |
|---|---|---|
| Severe fat deficit | <2 | <3.5 |
| Moderate fat deficit | 2 to <2.3 | 3.5 to <4 |
| Mild fat deficit | 2.3 to <3 | 4 to <5 |
| Normal | 3 to 6 | 5 to 9 |
| Excess fat | >6 to 9 | >9 to 13 |
| Obese class I | >9 to 12 | >13 to 17 |
| Obese class II | >12 to 15 | >17 to 21 |
| Obese class III | >15 | >21 |
Fat mass index references
The source uses a reversed fitness-style ranking: 90th means the lower end of the distribution, 10th the higher end. These labels do not measure fitness or health.
| Age | n | Mean ± SD | 90th | 80th | 70th | 60th | 50th | 40th | 30th | 20th | 10th |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 20-29 | 562 | 7.6± 3.6 | 4.30 | 5.00 | 5.60 | 6.00 | 6.60 | 7.30 | 8.20 | 9.50 | 11.90 |
| 30-39 | 196 | 10.0±5.2 | 4.40 | 5.40 | 6.40 | 7.50 | 8.90 | 10.20 | 11.90 | 14.30 | 18.00 |
| 40-49 | 258 | 11.0±5.1 | 4.90 | 7.00 | 7.80 | 8.60 | 9.70 | 11.30 | 12.80 | 15.40 | 19.20 |
| 50-59 | 437 | 12.0±4.9 | 6.40 | 7.60 | 9.10 | 10.10 | 11.30 | 12.80 | 14.40 | 16.30 | 18.70 |
| 60-69 | 440 | 12.1±4.5 | 6.80 | 8.00 | 9.40 | 10.30 | 11.20 | 12.90 | 14.30 | 15.60 | 18.10 |
| 70-79 | 183 | 10.8±3.8 | 6.40 | 8.00 | 8.80 | 9.60 | 10.50 | 11.60 | 12.80 | 14.40 | 16.30 |
Choose an age to highlight a row. No values are sent to the server.
Population reference, not a target. 3,327 US adults, 95% Caucasian; GE Lunar Prodigy or iDXA. These are not Australian norms or athlete targets. Source and table ↗ · Checked 23 September 2026.
Calculate it from the report
Take the total fat mass, not total body weight. Convert height to metres, multiply that height by itself, then divide fat mass by the result. For the invented example, 24.5 ÷ (1.68 × 1.68) gives approximately 8.68 kg/m². Dividing 68 kg body weight by height squared would calculate BMI instead.
Height errors matter because the denominator is squared. If you cannot find height on the report, ask the clinic which value it entered. A report without FMI can still contain enough information to calculate it, provided the fat mass covers the whole body rather than just the trunk or a limb.
Read the classification in context
Kelly and colleagues proposed FMI categories by matching their prevalence to established BMI categories at age 25 in the US reference population. The category names in the table are the authors’ terminology. They are not a personal diagnosis, a treatment plan or an Australian screening recommendation.
FMI separates fat mass from lean tissue in its numerator, unlike BMI. That makes it useful for describing body composition, but it does not measure where fat is stored, physical fitness, nutrition adequacy or metabolic health. Two people with the same FMI can have different reasons for testing and different needs.
The classification and a percentile table are also different tools. A category uses a specified boundary; a percentile locates a value in a selected distribution. Do not treat a shift across a rounded category boundary as proof of an important biological change.
What to compare over time
If adult height is unchanged, FMI mainly follows measured fat mass. Check that repeat scans use the same whole-body definition and comparable preparation. A change in the reported height can alter FMI without any change in fat mass. Keep the original report values so the calculation can be checked.
When looking at a GE report, the scanner-specific Imboden FMI percentiles provide a separate population comparison. The Kelly classification originated from Hologic NHANES data and should not be treated as an interchangeable manufacturer-specific diagnostic scale. Ask which interpretation your clinic uses and why.
Avoid turning a category into a goal
The purpose of the table is to explain a term you may see on a report. It cannot determine a suitable body composition for an athlete, someone recovering from illness or a person with a history of restrictive eating. A practitioner considers symptoms, training demands, medical history and the reason for testing alongside the scan.
There is no recommended timetable for moving between FMI categories. If you are changing nutrition or training, agree on what outcome matters and how it will be assessed. Energy, recovery and performance may be more useful immediate observations than a small movement in one index.
Questions for your practitioner
- Was FMI calculated from total fat mass and the correct height?
- Is this a percentile or a proposed classification?
- What would a change in FMI add to the decisions we are making?
Sources checked 23 September 2026: Source 1 · Source 2. New reference material prepared by DEXA Scans Australia; not included in Kristof’s earlier review.
Common questions about results
What is a good body fat percentage on a DEXA scan?
There is no single percentage that suits everyone. Published tables describe a reference population, not a target. Check the scanner, age group and sex group, then discuss the result alongside your health, strength, nutrition and reason for testing.
Why is my DEXA percentage different from my smart scale?
DEXA and bioimpedance scales estimate body compartments using different measurements and assumptions. Hydration and preparation can also affect comparisons. Neither an automatic correction factor nor combining both methods into one trend is reliable.
What does ALMI mean?
Appendicular lean mass index is arm and leg lean tissue divided by height in metres squared. It is one measure of muscle quantity. It does not measure strength and cannot diagnose sarcopenia by itself.
Is a Z-score of -1 bad?
ISCD describes adult Z-scores above -2.0 as within the expected range for age. That wording is a reference comparison, not a complete assessment of bone health. Ask your GP about the scan site, report conclusion and your medical history.
How often should I have another scan?
Agree on an interval that answers a useful clinical or training question. The reason for testing, expected change, scanner precision and whether the scan is for bone density or body composition all matter. There is no universal schedule for everyone.
Does lean mass mean muscle mass?
No. Lean soft tissue includes water and organs as well as muscle. Bone mineral content is reported separately. A change in lean tissue can reflect more than a change in muscle.
Can I compare results from GE and Hologic?
Do not treat values from different scanner families as interchangeable. Software, calibration, region definitions and reference databases differ. Ask the clinic whether a meaningful comparison is possible before interpreting a change.
Does a whole-body scan diagnose osteoporosis?
A body composition scan is not a substitute for a diagnostic bone density assessment at appropriate sites. Your GP or specialist interprets bone health findings alongside age, menopause status, fracture history and other relevant information.
How do I know if a change is larger than measurement variation?
Ask the clinic for its least significant change for the measurement. At 95% confidence, the usual calculation is 2.77 times measured precision error. Our calculator is an illustration and cannot replace the clinic’s precision assessment or confirm why a number changed.
Keep the result useful
Save the original report, including its scanner details and reference notes. At a later appointment, bring the earlier copy rather than relying on a rounded value entered in an app. Ask whether preparation, positioning and software were comparable, and whether the change is large enough to influence the decision you are making.
Return to all report measurements ↗ · Plan your next conversation ↗