DEXA Body Composition: What "Normal" Really Means and Where Optimal Begins

What a DEXA Body Composition Scan Actually Measures

A whole-body DEXA scan (technically abbreviated DXA, for dual-energy X-ray absorptiometry) passes two low-dose X-ray beams through the body to separate tissue into three compartments: fat mass, lean soft tissue mass, and bone mineral content [1]. The scan takes roughly seven minutes. It produces a region-by-region map of where fat and muscle sit, divided into arms, legs, trunk, and android (waist) versus gynoid (hip) zones.

This three-compartment model is what sets DEXA apart from a bathroom scale, bioelectrical impedance, or even skinfold calipers. A scale tells you total weight. Impedance devices estimate fat percentage but shift with hydration status by as much as 5 percentage points in a single day [2]. DEXA quantifies lean mass in each limb separately, making it the reference standard for diagnosing sarcopenia and tracking whether a weight-loss intervention is preserving muscle [3]. The International Society for Clinical Densitometry (ISCD) issued formal positions in 2019 endorsing whole-body DXA for body composition assessment in clinical practice, stating that "DXA total body composition measures are sufficiently precise for clinical use when obtained on the same instrument using standardized protocols" [4].

What clinicians pull from a DEXA report: total body fat percentage, appendicular lean mass index (ASMI, the combined lean mass of arms and legs divided by height squared), estimated visceral adipose tissue area, and the android-to-gynoid fat ratio. Each of these values carries a "normal" reference range. The problem is what "normal" actually means.

The Problem with Standard Reference Ranges

Most commercial DEXA systems (Hologic and GE Lunar dominate the U.S. market) report body fat percentage against age- and sex-matched population percentiles drawn from NHANES data [5]. A 45-year-old man at the 50th percentile of the U.S. population sits near 28% body fat. The lab report marks this "normal." It is normal only in the sense that it is average.

The distinction matters. Population-based norms reflect prevalence, not health. When 74% of U.S. adults meet criteria for overweight or obesity [6], the median body fat percentage drifts upward with each NHANES cycle. A lab range anchored to the 5th through 95th percentiles will always flag relatively few patients. By the time a result falls outside "normal," the patient may already carry significant visceral adiposity.

The AACE 2016 Comprehensive Clinical Practice Guidelines for Medical Care of Patients with Obesity explicitly warned against relying on single-metric cutoffs: "Diagnosis and assessment of the patient with obesity should include evaluation of complications and not be based solely on BMI" [7]. Body fat percentage measured by DEXA provides exactly that layer of complication-focused detail, but only if clinicians apply thresholds tied to metabolic risk rather than population averages.

Functional Optimal Ranges for Body Fat Percentage

A functional optimal range is defined by the body fat percentage below which cardiometabolic risk markers (fasting insulin, hsCRP, triglyceride-to-HDL ratio, liver fat fraction) remain consistently low in prospective cohort data. These are not arbitrary wellness targets. They come from outcomes research.

For men, the inflection point appears near 20%. A 2012 analysis of 13,236 adults in the Dallas Heart Study using DEXA-derived body fat found that men above 25% total body fat had significantly higher odds of metabolic syndrome (OR 3.6 to 95% CI 2.9 to 4.5), while those between 10 and 20% showed the lowest prevalence of insulin resistance and hepatic steatosis [8]. For women, the corresponding low-risk window fell between 18 and 28%.

These thresholds produce a practical clinical framework:

Men (functional optimal)

• Body fat 10 to 20%: low cardiometabolic risk bracket
• Body fat 20 to 25%: borderline, warrants metabolic screening
• Body fat above 25%: elevated risk regardless of BMI

Women (functional optimal)

• Body fat 18 to 28%: low cardiometabolic risk bracket
• Body fat 28 to 33%: borderline
• Body fat above 33%: elevated risk

Compare this to a typical Hologic printout, which may label a 50-year-old man at 28% body fat as "within expected range." The functional framework flags the same result as borderline. That difference changes clinical decisions: it might prompt a fasting insulin draw, a liver ultrasound, or an earlier conversation about GLP-1 pharmacotherapy.

Appendicular Lean Mass: The Metric GLP-1 Prescribers Should Track

Total weight loss is a crude endpoint. What patients and prescribers actually care about is the composition of weight lost. In STEP-1 (N=1,961), semaglutide 2.4 mg produced 14.9% mean total body weight loss at 68 weeks versus 2.4% with placebo [9]. A substudy using DEXA found that approximately 39% of total weight lost was lean mass [10]. That proportion is consistent with other pharmacologic weight-loss interventions but draws concern when patients start from a low lean mass baseline.

The appendicular skeletal muscle mass index (ASMI) quantifies this risk. ASMI equals the sum of lean mass in both arms and both legs (measured by DEXA) divided by height in meters squared. The European Working Group on Sarcopenia in Older People (EWGSOP2) set diagnostic cutoffs at <7.0 kg/m² for men and <5.5 kg/m² for women [3]. Falling below these values during GLP-1 therapy raises the question of sarcopenic obesity, a state where excess fat coexists with inadequate muscle mass.

Dr. Caroline Apovian, co-director of the Center for Weight Management and Wellness at Brigham and Women's Hospital, noted in a 2023 review: "Clinicians prescribing GLP-1 receptor agonists should monitor lean mass, not just scale weight, to ensure that the metabolic benefits of fat loss are not offset by muscle depletion" [11].

Functional optimal ASMI targets, based on EWGSOP2 and the Foundation for the National Institutes of Health (FNIH) Sarcopenia Project [12], look like this:

• Men: 7.0 kg/m² (minimum), 7.5 kg/m² or higher (optimal)
• Women: 5.5 kg/m² (minimum), 6.0 kg/m² or higher (optimal)

A DEXA scan before starting a GLP-1 agonist and again at 6 and 12 months provides three time points to calculate lean mass velocity. If ASMI drops more than 5% from baseline, many clinicians add or intensify resistance exercise prescriptions, increase protein targets to 1.2 to 1.6 g/kg/day, and consider whether concurrent testosterone therapy (in hypogonadal men) could mitigate losses [13].

Visceral Fat and the Android-to-Gynoid Ratio

Not all fat is equal. Visceral adipose tissue (VAT), the fat surrounding abdominal organs, drives insulin resistance, systemic inflammation, and MASLD progression more aggressively than subcutaneous stores [14]. Newer DEXA software (Hologic Horizon with APEX, GE Lunar enCORE) estimates VAT area in square centimeters at the L4 vertebral level, validated against CT as the reference standard.

A VAT area below 100 cm² on DEXA corresponds to low metabolic risk. Values between 100 and 160 cm² are intermediate. Above 160 cm², the association with type 2 diabetes incidence rises sharply [15]. These cutoffs are more specific to cardiometabolic outcomes than total body fat percentage alone.

The android-to-gynoid (A/G) fat ratio adds another dimension. Android fat is measured in a band between the iliac crest and 20% of the distance to the chin. Gynoid fat is measured across the hips and upper thighs. An A/G ratio above 1.0 in either sex signals disproportionate central adiposity. In the Framingham Heart Study Third Generation cohort, each 0.1-unit increase in A/G ratio was associated with a 30% higher odds of metabolic syndrome independent of BMI [16].

For patients on testosterone replacement, DEXA-tracked A/G ratio changes can document the fat redistribution that TRT produces. The Testosterone Trials (TTrials) showed that one year of testosterone gel in men aged 65 and older with low testosterone reduced total fat mass by 1.65 kg, with preferential loss from the android region [17].

How Hormone Therapy Shifts the DEXA Picture

Testosterone replacement in hypogonadal men consistently increases lean mass and decreases fat mass on serial DEXA. A meta-analysis of 59 randomized controlled trials (N=5,331) found that testosterone therapy increased lean body mass by a weighted mean of 1.6 kg (95% CI 1.3 to 2.0) and decreased fat mass by 2.0 kg (95% CI 1.4 to 2.6) over a median treatment duration of 40 weeks [18]. The Endocrine Society's 2018 Clinical Practice Guideline recommends monitoring body composition as one metric of testosterone therapy response, noting that "body composition changes may be detected by DXA and are typically evident within 3 to 6 months of treatment initiation" [19].

Estrogen-based HRT in postmenopausal women shows a more modest but reproducible effect. The Women's Health Initiative (WHI) observational data indicated that women using hormone therapy had 1.0 to 1.5 kg less trunk fat and 0.5 to 1.0 kg more appendicular lean mass than nonusers after 3 years [20]. This shift matters for fracture risk and functional independence, not just aesthetics.

GLP-1 receptor agonists complicate the picture by producing large total weight reductions that include both fat and lean compartments. Tirzepatide (the dual GIP/GLP-1 agonist) at 15 mg in SURMOUNT-1 produced 22.5% mean weight loss at 72 weeks [21]. DEXA substudies from the tirzepatide program showed lean mass losses proportional to total weight loss, though the percentage of weight lost as fat was slightly higher with tirzepatide than with semaglutide in cross-trial comparisons [22].

The clinical takeaway: any patient on GLP-1 pharmacotherapy, TRT, or HRT should have baseline and follow-up DEXA body composition scans. A scale cannot distinguish a favorable 15 kg fat loss from a concerning 6 kg lean mass deficit.

Setting Retest Intervals and Interpreting Change

DEXA body composition precision error (the expected variability between two scans on the same machine within a short interval) is approximately 1 to 2% for total body fat percentage and 1 to 3% for regional lean mass [4]. This means a change must exceed roughly 2 to 4 percentage points for body fat, or 0.5 to 1.0 kg for regional lean mass, to be considered real rather than measurement noise.

Practical retest timing depends on clinical context:

• Starting GLP-1 therapy: baseline scan, repeat at 6 months and 12 months
• Starting TRT or HRT: baseline scan, repeat at 6 months, then annually
• Stable maintenance: annually, or when a clinical change (new medication, significant weight shift, fracture) warrants reassessment
• Post-bariatric surgery: baseline presurgical scan, then at 6, 12, and 24 months to track lean mass preservation

Patients should be scanned on the same machine each time. Switching between Hologic and GE Lunar devices introduces systematic differences of 2 to 4% in body fat estimates due to different calibration algorithms [23]. Cross-platform comparisons are not valid for tracking individual change.

Fasting state, hydration, and clothing affect results. The ISCD recommends scanning in a hospital gown, after a 2-hour fast, at a consistent time of day [4]. These details sound minor, but they can account for 1 to 2 percentage points of apparent fat change, enough to obscure a real therapeutic effect.

What to Do When Results Fall Outside Optimal

A DEXA scan is diagnostic, not therapeutic. The value comes from what the clinician does with the numbers.

When body fat exceeds functional optimal thresholds and VAT is above 100 cm², the metabolic workup should include fasting glucose, HbA1c, fasting insulin, a lipid panel with triglyceride-to-HDL ratio, and hepatic steatosis screening (either ultrasound or FibroScan). If metabolic syndrome criteria are met, GLP-1 receptor agonist therapy carries strong evidence from STEP and SURMOUNT programs [9][21].

When ASMI falls below 7.0 kg/m² in men or 5.5 kg/m² in women, structured resistance training 2 to 3 days per week and protein intake of 1.2 to 1.6 g/kg/day are first-line interventions supported by the EWGSOP2 consensus [3]. In hypogonadal men with low ASMI, testosterone replacement may produce additive lean mass gains of 1 to 2 kg over 6 months when combined with exercise [18].

When the A/G ratio exceeds 1.0 despite an acceptable total body fat percentage, preferential visceral fat accumulation is the concern. This pattern responds particularly well to GLP-1 agonists, which reduce VAT disproportionately to total fat loss [10], and to testosterone therapy in men, which preferentially depletes android fat stores [17].

The scan's most actionable data point may be the one patients never see on a standard lab report: the lean mass trajectory over 6 to 12 months of active pharmacotherapy. A body fat drop from 32% to 24% is a success only if ASMI stayed above the sarcopenia cutoff. Track both, or the DEXA report tells half the story.