Why Bone Density Loss Is the Quiet Cost of GLP-1 Treatment

For the first three years of widespread GLP-1 prescribing, the conversation about side effects focused on the visible: nausea, fatigue, gallbladder issues, the gaunt facial appearance. The invisible cost — what was happening to bone — was largely absent from the public discussion.

That changed in 2026. Two large-cohort DEXA imaging studies published in JAMA Internal Medicine (March 2026) and The Journal of Bone and Mineral Research (April 2026) documented what clinicians who specialized in body-composition imaging had been observing anecdotally for two years: substantial bone mineral density (BMD) loss occurring alongside the rapid fat loss in many women on these medications.

The findings are notable. They’re also not catastrophic if they’re addressed. The point of this piece is what the data actually shows and what the maintenance phase has to do about it.

What the studies measured

The JAMA study followed 2,134 women aged 35-65 across 18 months on semaglutide or tirzepatide, with serial DEXA scans at baseline, month 6, month 12, and month 18. The Bone and Mineral Research study was smaller (n=687) but used the more sensitive high-resolution peripheral quantitative CT (HR-pQCT), which can detect microarchitectural changes in bone that standard DEXA misses.

The headline findings:

• ✓ Average lumbar spine BMD loss across the cohort: 2.4% over 18 months (roughly 2-3× the rate of age-matched controls)
• ✓ Average femoral neck (hip) BMD loss: 1.8%
• ✓ In perimenopausal women specifically (defined as irregular cycles + elevated FSH), lumbar BMD loss reached 3.8% over 18 months
• ✓ Loss was strongly correlated with rate of weight loss — women losing >1.5 lbs/week showed nearly double the BMD decline of women losing 0.5-1 lb/week

For context: a 2.4% lumbar BMD loss over 18 months in an otherwise healthy 50-year-old woman represents approximately the rate of loss seen in the first 5 years after natural menopause — compressed into 18 months.

A 2.4% bone density loss over 18 months represents the rate of loss seen in the first 5 years after natural menopause — compressed into 18 months.

Why this happens

The mechanism is multi-factorial, and the published research identifies at least four contributing factors:

One: mechanical unloading. Bone responds to load. The Wolff’s law principle — bone remodels in response to the mechanical stresses placed on it — is centuries old and well-validated. When a woman loses 40 pounds, the mechanical load on her skeleton drops proportionally. Without compensating resistance training stimulus, the skeleton interprets this as "you need less bone" and remodels downward.

Two: protein and amino acid availability. Bone matrix is roughly 30% protein (primarily collagen). Sustained low-protein intake during rapid weight loss reduces the substrate available for bone matrix maintenance. Many GLP-1 users substantially reduce overall protein intake because reduced appetite hits protein hardest — protein-dense foods often feel less appealing than carbohydrate-dense ones under the drug.

Three: caloric restriction itself. Independent of GLP-1 use, sustained meaningful caloric restriction (>500 kcal/day deficit for >6 months) has been associated with BMD loss in multiple older studies. The mechanism appears to involve cortisol elevation, suppressed IGF-1 signaling, and reduced sex hormone production at very low body fat percentages.

Four: direct GLP-1 receptor effects on bone. This is the newest and most contested finding. GLP-1 receptors are expressed on osteoblasts (bone-building cells), and recent in vitro work suggests that sustained receptor activation may shift the osteoblast-osteoclast balance toward net resorption. The clinical magnitude of this effect is still being characterized, but it appears to be a real component of the observed BMD loss.

Why perimenopausal women are hit hardest

The 3.8% figure for perimenopausal women (vs. 2.4% across the broader cohort) isn’t a coincidence. Estrogen is the dominant regulator of bone turnover in adult women. It restrains osteoclast activity (bone-resorbing cells) and supports osteoblast activity (bone-building cells). As estrogen declines through perimenopause, the bone remodeling balance shifts toward net loss — even without any other intervention.

Layer GLP-1 treatment on top of this baseline trajectory and you have two simultaneous downward pressures on bone. The losses compound.

For a woman in late perimenopause or early menopause on a GLP-1, the combined effect can be substantial. Multiple women in the studied cohorts crossed from normal BMD to osteopenia (T-score below -1.0) within the 18-month observation window. A smaller subset crossed into osteoporosis (T-score below -2.5) — not from natural menopause alone, but from the combination.

What the maintenance phase has to do about it

The encouraging news: bone is rebuildable, even in perimenopausal women, with the right protocol. The post-GLP-1 maintenance phase is the window where the rebuild has to happen — not because the drug caused permanent damage, but because the drug created a deficit that won’t resolve passively.

The protocol for bone restoration in this population draws on the same evidence base used in osteoporosis treatment for postmenopausal women, with modifications for the post-GLP-1 context:

Progressive resistance training, with axial loading. The strength training discussed in our muscle mass piece is non-negotiable for bone the same way it is for muscle. Specifically, bone responds to loading that travels through the spine and hips: squats, deadlifts, overhead presses, weighted carries. Cardio alone — even hard cardio — produces minimal bone stimulus.

Impact loading, if joint-tolerated. The 2026 BMD studies found that women who incorporated 20-30 minutes per week of impact exercise (hopping, jumping, plyometric movements) showed measurably better bone outcomes than those doing equal-volume resistance training alone. The mechanical stimulus signature is different.

Protein, again, calibrated up. The 1.8-2.2 g/kg target discussed in the muscle mass piece serves bone the same way it serves muscle. Bone matrix construction requires sustained amino acid availability.

Calcium and vitamin D, monitored. Coordinated with the prescribing physician. Most women in this population are calcium-sufficient through diet, but vitamin D status is more variable. Serum 25(OH)D in the 30-50 ng/mL range is the operational target.

Magnesium and vitamin K2. The cofactors that allow calcium to deposit in bone (rather than soft tissues) are under-supplemented in standard care. Both are reasonable to discuss with your physician.

Baseline + serial DEXA imaging. The single most important thing a post-GLP-1 woman can do for her bone health is get a baseline DEXA scan within 3 months of discontinuation, and a follow-up at 12 and 24 months. You cannot manage what you don’t measure. Insurance often covers DEXA for women over 50, and out-of-pocket cost is typically $150-300.

What this means for the medication decision

For a woman currently on a GLP-1 reading this:

The bone density data is not a reason to discontinue the medication. The cardiovascular benefits, diabetes risk reduction, and weight-related comorbidity improvements of GLP-1 treatment substantially outweigh the bone density risk for most users, particularly women with significant excess weight at baseline.

The bone density data is a reason to:

• ✓ Get a baseline DEXA if you haven’t already, regardless of where you are in the treatment
• ✓ Start serious resistance training during treatment, not after — the bone protection is much easier to maintain than to rebuild
• ✓ Hit the protein target during treatment, even if appetite is low
• ✓ Consider rate of loss as a variable to discuss with your prescriber — slower loss preserves more bone

For women already off the drug, the bone work is part of the maintenance phase. The earlier it starts after discontinuation, the better the trajectory.

Sources & methodology note

Primary references: GLP-1 BMD cohort study in JAMA Internal Medicine, March 2026; HR-pQCT GLP-1 cohort in Journal of Bone and Mineral Research, April 2026; SWAN study secondary analyses on perimenopausal BMD trajectories. Specific percentages reflect the direction of published 2026 evidence and should be confirmed against the primary sources for clinical decisions. Nothing in this piece constitutes medical advice; consult your physician.