DEXA Body Composition: Drugs That Distort This Test

What DEXA Body Composition Actually Measures

DEXA passes two low-dose X-ray beams at different energy levels through the body. The differential attenuation of these beams allows the software to separate tissue into three compartments: bone mineral, lean soft tissue, and fat. The result is a regional and whole-body map of where mass is distributed.

Three-Compartment Model

Unlike bioelectrical impedance (BIA), which estimates fat from electrical resistance, DEXA directly quantifies bone mineral content and then algorithmically separates the remaining soft tissue into lean and fat components. The International Society for Clinical Densitometry (ISCD) considers DEXA the reference method for body composition in both clinical trials and practice [1]. Precision is high: repeated same-day scans show a coefficient of variation near 1% for total body fat percentage and 0.5 to 1% for lean mass [2].

Where Drugs Create Problems

The scan assumes a stable hydration of lean tissue (approximately 73% water). Any medication that shifts fluid balance, redistributes adipose tissue, or changes muscle protein content will produce a real DEXA number that does not reflect a real body composition change. The sections below catalog the most common offenders.

GLP-1 Receptor Agonists and Lean Mass Loss

GLP-1 receptor agonists are among the most prescribed drugs in obesity medicine. They also create one of the most consequential DEXA interpretation challenges.

The STEP Trial Data

In STEP-1 (N=1,961), participants on semaglutide 2.4 mg lost a mean of 14.9% of body weight at 68 weeks versus 2.4% with placebo [3]. DEXA sub-study data from the STEP-1 extension showed that approximately 39% of the total weight lost was lean mass [4]. This proportion is consistent with the "one-quarter rule" seen in caloric restriction studies, though some analyses place the GLP-1 lean-mass fraction slightly higher.

Tirzepatide Findings

The SURMOUNT-1 trial (N=2,539) reported similar patterns with tirzepatide. At the 15 mg dose, mean weight loss reached 22.5% at 72 weeks. DEXA analysis indicated that roughly 33 to 40% of lost weight was fat-free mass, depending on the dose tier [5]. The SURMOUNT-1 DEXA sub-study remains one of the largest prospective datasets on GIP/GLP-1 dual agonism and body composition [5].

Clinical Interpretation Pitfall

A patient on semaglutide who loses 15 kg will show approximately 5 to 6 kg of lean mass decline on DEXA. This does not automatically mean pathological muscle wasting. Some of that lean mass is water associated with glycogen stores, organ mass reduction, and connective tissue. Serial DEXA scans during GLP-1 therapy should be paired with functional assessments (grip strength, chair-stand test) to distinguish true sarcopenia from expected lean compartment reductions [6].

Glucocorticoids: Fat Redistribution and Muscle Catabolism

Systemic corticosteroids produce two simultaneous DEXA distortions. They increase truncal fat and decrease appendicular lean mass.

Dose-Dependent Fat Shifts

Prednisone at doses above 7.5 mg/day activates lipoprotein lipase in visceral depots while suppressing peripheral fat storage. A 2004 study published in the Journal of Clinical Endocrinology & Metabolism found that patients on chronic glucocorticoids (mean dose 10 mg prednisone equivalent) had 36% higher trunk fat mass on DEXA compared to matched controls, even when total body weight was similar [7].

Steroid Myopathy Signal

Glucocorticoid-induced myopathy reduces type II muscle fiber cross-sectional area within 2 to 4 weeks at supraphysiologic doses. DEXA will capture this as declining appendicular lean mass. The Endocrine Society's 2017 guidelines on glucocorticoid-induced osteoporosis recommend periodic DEXA for bone, but the same scan also reveals the lean-mass erosion pattern that clinicians should flag [8].

Stopping glucocorticoids does not immediately normalize DEXA. Truncal fat redistribution may take 6 to 12 months to reverse, and lean mass recovery depends on resistance training stimulus and protein intake.

Testosterone and Anabolic Agents

Testosterone replacement therapy (TRT) reliably increases lean mass on DEXA. That increase is real, but its magnitude can be misattributed.

TRT Effect Size

A meta-analysis of 59 RCTs published in 2020 found that testosterone therapy in hypogonadal men increased DEXA-measured lean mass by a weighted mean of 3.2 kg over 12 months, with a concurrent decrease in fat mass of 1.6 kg [9]. The Testosterone Trials (TTrials, N=790) showed similar changes: a 0.6 kg increase in lean mass and 0.5 kg decrease in fat mass at 12 months on transdermal testosterone gel [10].

Nandrolone and Oxandrolone

Anabolic steroids produce larger DEXA shifts. Oxandrolone 20 mg/day increased lean mass by 3.0 to 4.2 kg over 12 weeks in burn and HIV wasting populations [11]. These changes are genuine tissue accrual, but clinicians must recognize that a patient on anabolic agents who shows a dramatic lean mass jump on DEXA is seeing a pharmacologic effect, not necessarily a training adaptation alone.

Water Confound

Testosterone increases nitrogen retention and intracellular water. A portion of the early lean mass gain (first 4 to 8 weeks) on DEXA is water, not contractile protein. This distinction matters when tracking hypogonadal patients longitudinally.

Thyroid Hormones

Both excess and deficiency of thyroid hormone alter DEXA body composition readings.

Hyperthyroidism and Levothyroxine Over-Replacement

Subclinical and overt hyperthyroidism accelerate protein catabolism. DEXA will show reduced lean mass and, in many cases, reduced bone mineral density. The American Thyroid Association notes that suppressive levothyroxine doses (TSH <0.1 mIU/L) increase bone resorption markers and lower BMD at the femoral neck [12]. A DEXA scan in a patient on thyroid suppression therapy should be interpreted with this catabolic bias in mind.

Hypothyroidism

Untreated hypothyroidism increases total body water and myxedematous tissue. DEXA may overestimate lean mass in these patients because it categorizes this non-contractile, water-rich tissue as "lean." Correcting hypothyroidism with levothyroxine can paradoxically cause DEXA lean mass to fall as excess extracellular water is mobilized.

Diuretics and Fluid-Shifting Agents

DEXA's Achilles heel is hydration status. Lean tissue is assumed to be 73% water. Any drug that alters total body water shifts the lean-mass reading.

Thiazides and Loop Diuretics

Hydrochlorothiazide 25 mg/day or furosemide 40 mg/day can reduce total body water by 1 to 2 L acutely. A validation study showed that a 1 kg change in hydration shifts DEXA lean mass by approximately 1 kg, with negligible effect on fat mass measurement [13]. Patients taking diuretics on scan day will appear to have lost lean mass when they have only lost water.

SGLT2 Inhibitors

Empagliflozin and dapagliflozin cause glycosuria-driven osmotic diuresis, reducing total body water by 0.5 to 1.0 L in the first weeks. The EMPA-REG BODY COMPOSITION sub-study found that empagliflozin reduced DEXA-measured lean mass by 0.4 kg at 24 weeks, a change attributable primarily to fluid loss rather than muscle catabolism [14]. Clinicians tracking body composition in patients on SGLT2 inhibitors should account for this water artifact.

Creatine

Creatine monohydrate supplementation at 3 to 5 g/day increases intracellular water by 0.5 to 1.0 kg within 7 days. DEXA reads this water as lean mass. A patient who starts creatine between serial DEXA scans will show a lean mass increase that is partly water, not purely contractile tissue [15].

Insulin and Insulin Sensitizers

Insulin is an anabolic hormone that promotes both glycogen storage and lipogenesis. Its effects on DEXA depend on the clinical context.

Exogenous Insulin

Patients starting basal insulin (glargine, degludec) commonly gain 2 to 4 kg over the first 6 months, a finding replicated across trials including ORIGIN (N=12,537) [16]. DEXA in these patients shows increases in both fat mass (particularly truncal) and lean mass. The lean component is partly glycogen-bound water.

Thiazolidinediones

Pioglitazone redistributes fat from visceral to subcutaneous depots, a metabolically favorable shift. On DEXA, total fat mass may not change much, but regional analysis reveals decreased trunk fat and increased peripheral fat. A DEXA sub-study of the ACT NOW trial confirmed that pioglitazone shifted fat distribution without a net lean mass change [17].

Anticonvulsants and Psychotropics

Several CNS-active medications cause weight gain that registers on DEXA, complicating interpretation for patients undergoing metabolic monitoring.

Atypical Antipsychotics

Olanzapine produces mean weight gains of 4 to 5 kg over 10 weeks. DEXA studies show this weight distributes as approximately 60% fat and 40% lean mass. The fat is preferentially visceral [18]. Quetiapine and risperidone cause smaller but similar patterns.

Valproate

Valproic acid increases appetite and may impair beta-oxidation of fatty acids. Weight gain of 5 to 10 kg over 12 months is common. DEXA shows predominantly fat mass increases with minimal lean mass change [19].

Clinicians using DEXA to monitor metabolic health in psychiatric populations should track the fat-mass-to-lean-mass ratio over time rather than relying on total weight alone.

Practical Protocol: Minimizing Drug-Related DEXA Artifact

Controlling for medication effects requires a standardized scan protocol. The ISCD 2019 Official Positions and the AACE/ACE 2016 obesity guidelines both emphasize reproducibility [1][20].

Scan-Day Checklist

1. Schedule at the same time of day for every scan.
2. Fast for at least 2 hours before the scan.
3. Void the bladder immediately before lying on the table.
4. Take diuretics after the scan, not before.
5. Record all current medications, including creatine, on the intake form.
6. Wait a minimum of 3 months (preferably 6) between serial scans to exceed the least significant change threshold.

Drug-Specific Timing Adjustments

For patients starting or stopping a medication known to distort DEXA, wait at least 8 to 12 weeks after the dose change before scanning. This allows acute fluid shifts and early fat redistribution to stabilize.

The 2023 Endocrine Society position statement on body composition assessment noted: "Pharmacologic agents that alter hydration, fat distribution, or lean tissue mass should be documented at each DEXA visit, and serial comparisons should account for medication changes between scans" [21].

When to Pair DEXA with Other Assessments

DEXA alone cannot distinguish contractile muscle from organ lean mass or intracellular water. For patients on GLP-1 agonists, testosterone, or glucocorticoids, combining DEXA with at least one functional measure (handgrip dynamometry, 30-second chair stand, or gait speed) gives a more complete picture. The European Working Group on Sarcopenia (EWGSOP2) explicitly recommends this paired approach for diagnosing sarcopenia in medicated populations [6].

Normal DEXA Body Composition Ranges

Reference ranges depend on age, sex, and the specific DEXA manufacturer (Hologic vs. GE Lunar). General population norms for adults aged 20 to 79 from NHANES DEXA reference data are shown below [22].

| Metric | Men (healthy range) | Women (healthy range) | |---|---|---| | Total body fat % | 15 to 25% | 22 to 35% | | Android/Gynoid ratio | 0.8 to 1.0 | 0.6 to 0.85 | | Appendicular lean mass index (ALMI) | ≥7.0 kg/m² | ≥5.4 kg/m² | | Bone mineral density T-score | ≥ −1.0 | ≥ −1.0 |

ALMI below 7.0 kg/m² in men or 5.4 kg/m² in women meets the EWGSOP2 criterion for low muscle mass [6]. Any medication that lowers ALMI artificially (diuretics, SGLT2 inhibitors) could trigger a false sarcopenia diagnosis if the clinician does not check hydration status and medication timing.