Your Bone Density Is Normal. Your Bone Quality May Not Be.

Every year, millions of people receive a reassuring bone density report. Their T-score is acceptable. Their DEXA looks fine. Their doctor tells them their fracture risk is low.

Some of those people will break a hip anyway.

The problem is not the test. DEXA measures bone density accurately. The problem is that bone density is only half of what determines whether a bone will fracture. The other half is bone quality—and your DEXA says nothing about it.

In cardiovascular medicine, the goal is preventing heart attack and stroke—not achieving a particular cholesterol number. Cholesterol reduction became a proxy for that goal, but the two are not the same thing: patients with normal cholesterol have heart attacks, and patients with elevated cholesterol sometimes do not. Bone medicine has the same problem. The goal is fracture prevention, not achieving a particular T-score. A bone with normal density but heavily glycated, brittle collagen is still more likely to fracture than a bone with lower density but healthier matrix. The number can be reassuring while the real risk goes unmeasured.

Think of bone like a woven basket. A DEXA scan tells you how thick the basket walls are—the amount of mineral packed into your bone. That is density. But it does not tell you whether the fibers weaving that basket together are strong and flexible or dry and brittle. That second quality—the resilience of the collagen that gives bone its shock-absorbing ability—is bone quality. It is a completely separate question from density, and it is often the deciding factor in whether a bone fractures under everyday stress. Two people can have identical DEXA scores and completely different fracture risks depending on the quality of their bone collagen.

That collagen quality is vulnerable to a specific form of molecular damage that accumulates silently over decades, is driven by diet and metabolism, and can now be measured non-invasively—in under a minute, without a blood draw, in a clinical office.

The measurement is called skin autofluorescence. And the evidence linking it to fracture risk is more compelling than most clinicians realize.

WHAT SKIN AUTOFLUORESCENCE MEASURES

Skin autofluorescence, or SAF, is measured by a device called the AGE Reader. It shines a low-intensity ultraviolet light on the forearm and measures the fluorescence emitted from the skin. The fluorescent signal comes from advanced glycation end products—AGEs—that have accumulated in the skin’s collagen over time.

Skin collagen and bone collagen are both type I collagen—the same structural protein, the same long lifespan, and the same vulnerability to AGE accumulation. Studies comparing SAF measurements with direct skin biopsy samples have confirmed a close relationship between the two.

Whether the skin reading mirrors the AGE content of bone specifically is contested: in 100 patients undergoing spinal surgery, SAF correlated with skin pentosidine but not with pentosidine measured directly in lumbar lamina bone. What the outcome data establish is narrower and more useful—SAF predicts fracture, independently of bone mineral density, whatever the tissue-level mechanism turns out to be.

The test takes under a minute. It requires no blood draw, no urine collection, no fasting. The patient places their forearm on the device. A reading appears. That reading reflects years of cumulative AGE accumulation in a way that no single blood test or imaging study captures.

WHAT THE EVIDENCE SHOWS

• The Rotterdam Study: One of the largest population cohort studies in Europe examined SAF alone as a predictor of prevalent fractures in nearly 3,000 older adults. Those in the upper three SAF quartiles had 43% higher odds of major osteoporotic fracture (OR 1.43) and 72% higher odds of vertebral fracture (OR 1.72) compared to those in the lowest quartile, after adjustment for age, sex, BMI, and other confounders. Critically, additional adjustment for bone mineral density did not change the associations—confirming that SAF captures fracture-relevant bone damage independently of what DEXA measures.
• 2022 REACTION Cohort Study: Examined 1,214 patients with type 2 diabetes using a composite index called SAF-AGEage (defined as SAF multiplied by age divided by 100). Compared to patients in the lowest quartile, those in the highest quartile had 6.51 times the odds of low bone density or osteoporosis (OR 6.51), and 170% higher odds of major osteoporotic fracture (OR 2.70), after full statistical adjustment (Liu et al., Journal of Diabetes, 2022). The relationship was consistent and progressive across all four quartiles.
• 2025 Type 1 Diabetes Study: Confirmed that SAF contributes to prediction of impaired bone quality independently of bone mineral density. 

These findings matter because they point to a mechanism, not just a correlation. AGEs accumulate in bone collagen over years, forming crosslinks that change the mechanical behavior of the matrix. A bone with a high AGE burden fractures at a lower load than a bone with the same mineral density but less glycation damage. SAF reflects that accumulated burden.

Those with elevated skin AGE burden had 72% higher odds of vertebral fracture than those with the lowest AGE burden (OR 1.72) – even after adjusting for bone density. The number your DEXA gives you does not change what your AGE burden predicts.

THE LAB TEST THAT CONFIRMS IT: CML

The SAF reading reflects cumulative AGE accumulation across all long-lived tissues. For patients whose SAF is elevated, carboxymethyl-lysine—CML—provides a confirmatory biomarker that links that AGE burden specifically to fracture outcomes.

The Cardiovascular Health Study: Followed over 3,000 older adults prospectively and found that serum CML predicted incident hip fracture with a hazard ratio of 1.17 per standard deviation after adjustment for age, sex, BMI, smoking, alcohol, physical activity, and kidney function. The critical detail: total hip bone mineral density was not significantly associated with CML. The fracture prediction ran independent of density.

The Health, Aging and Body Composition Study: Replicated this in type 2 diabetes: higher serum CML was associated with incident clinical fractures independent of BMD, glycemic control, and diabetes duration.

Pentosidine, a different advanced glycation end product, has been studied the same way, and its results sharpen the picture in a useful way. In the Health, Aging and Body Composition study, higher urine pentosidine was associated with incident clinical fractures and prevalent vertebral fractures in older adults with diabetes, independent of bone mineral density and other fracture risk factors – the first study to link an AGE level to incident clinical fracture risk. In type 1 diabetes, serum pentosidine was independently associated with prevalent fractures while bone mineral density was not. In the Japanese Osteoporosis Intervention Trial-05, pentosidine was the strongest risk factor for nonvertebral fracture, with an approximately linear relationship to fracture incidence and no threshold effect.

Two honest qualifications. No association was found in Health ABC subjects without diabetes. And in the French OFELY cohort of healthy untreated postmenopausal women, fracture risk was higher in the top pentosidine quartile but did not remain significant after adjustment for age, bone density, and prevalent fracture. The AGE-fracture association appears strongest where glycation burden is highest.

The reason peripheral measurements matter at all is worth stating plainly: direct measurement of AGE levels in bone collagen requires invasive sampling and is not feasible in practice. Every marker discussed here, skin autofluorescence included, is a proxy for what is happening in bone.

A practical note on availability, since the research literature and the clinical menu are not the same thing. Serum pentosidine and serum CML are research measurements – quantified by ELISA, HPLC, or validated mass spectrometry methods in study settings – and neither is a routinely orderable clinical test. What is available is urinary CML, measured directly as one of the 16 markers on the Vibrant Wellness Oxidative Stress Profile, alongside skin autofluorescence in the office. Those two are the accessible proxies for AGE burden. The cohort studies above used serum, and their findings are what justify measuring AGE burden at all – but the route to measuring it in a patient is the scan and the urinary panel.

Together, SAF-AGEage and serum CML provide converging evidence from two different measurement approaches—tissue accumulation and circulating and excreted biomarkers—that AGE burden is associated with fracture risk above and beyond what DEXA captures.

THE OXIDATIVE STRESS CONNECTION

SAF and CML measure the end result—AGE accumulation. The Vibrant Wellness Oxidative Stress Profile measures the processes that drive it.

AGEs activate RAGE, the receptor for advanced glycation end products, triggering NADPH oxidase and a surge of reactive oxygen species (ROS). In bone, ROS suppress osteoblast function—the cells responsible for building new bone—impairing their differentiation, activity, and survival. The net result is impaired bone formation and reduced capacity to repair microdamage within the collagen matrix. A 2023 systematic review confirmed that oxidative stress-related genetic polymorphisms associate with BMD and fracture risk across diverse populations.

The Vibrant Wellness Oxidative Stress Profile measures 16 urinary markers of oxidative damage to proteins, lipids, DNA, and RNA—including CML directly—alongside 32 genetic variants in antioxidant enzymes. Its Oxidative Damage Score quantifies the rate of molecular aging from oxidative stress. For bone health, this captures both arms of the AGE-oxidative stress pathway: the glycation damage marker and the reactive oxygen species that drive it.

WHICH INFLAMMATORY MARKERS ACTUALLY PREDICT FRACTURE

The cytokines named above – TNF-alpha, IL-6, IL-1beta – are usually invoked as mechanism. They have also been measured against fracture outcomes prospectively, and the results are more specific than “inflammation is bad for bone.”

In the Osteoporotic Fractures in Men study, a case-cohort analysis nested in a cohort of 5,994 men aged 65 and older, men in the highest quartile of TNF-alpha and its two soluble receptors had a 2.0 to 4.2-fold higher risk of hip and clinical vertebral fracture than men in the lowest quartile. Additional adjustment for total hip bone mineral density yielded similar findings. TNF receptors had already been shown to predict hip fracture in the Women’s Health Initiative.

Composite burden outperformed any single marker. Men with three or more of six pro-inflammatory markers in the highest quartile had roughly twice the hip fracture risk and three times the clinical vertebral fracture risk of men with none. The Health, Aging and Body Composition study found the same pattern in 2,985 older men and women: individual markers were each borderline, but subjects with three or more of seven elevated markers carried 2.65 times the fracture risk of those with none. As the MrOS authors put it, the interrelatedness of human cytokines makes it unlikely that any one biomarker captures an individual’s entire inflammatory burden.

Two findings complicate the simple story. IL-10, an anti-inflammatory cytokine, ran the other way – men in the highest IL-10 quartile had a 49 percent lower risk of clinical vertebral fracture. And IL-6, despite its prominence in the mechanistic literature, was inconsistent: it predicted hip fracture in the Study of Osteoporotic Fractures but lost significance after adjustment in both Health ABC and MrOS.

High-sensitivity CRP sits in the same unsettled position. A 2021 meta-analysis of ten cohort studies covering 29,741 subjects found a 54 percent higher fracture risk in the top hs-CRP tertile compared with the bottom, and the Tromso Study found hsCRP an independent risk factor for non-vertebral fracture in both sexes, with the upper tertile carrying 80 percent higher risk independent of bone density. MrOS found no association at all. The reasonable reading is that hsCRP indicates inflammatory burden rather than functioning as a stand-alone fracture predictor.

 The most useful finding for this discussion is a mediation result. In MrOS, adjusting for appendicular lean mass attenuated the association between inflammatory burden and hip fracture by 27 percent, and adjusting for disability attenuated it by another 27 percent – while adjusting for bone density attenuated it by only 15 percent. There was no association between inflammation and the rate of hip bone loss at all.

Inflammation, in other words, appears to reach fracture largely through muscle and physical function rather than through bone density. That is the mechanistic case for measuring grip strength, skeletal muscle mass and phase angle rather than assuming a DEXA will register the damage.

These markers are clinically accessible. The Vibrant Wellness Cytokines Panel covers TNF-alpha, IL-6 and IL-10 – three of the four cytokines with prospective fracture data – and requires no fasting, though samples ship on dry ice because several markers degrade rapidly. The soluble TNF receptors, which carried the strongest signal in MrOS, remain research assays and are not available clinically.

THE GUT DIMENSION

The inflammatory pathway described above has an upstream source, and it now has human fracture data behind it. In the FINRISK 2002 cohort, 7,043 adults had gut microbiome composition characterized by shotgun metagenomic sequencing and were followed for a median of 18 years, during which 1,092 incident fractures occurred. Greater microbial diversity was associated with lower fracture risk, with a hazard ratio of 0.92 for each standard deviation increase in the Shannon diversity index. Abundance of Proteobacteria was associated with increased fracture risk, hazard ratio 1.14, an effect driven by the Gammaproteobacteria class. Abundance of Tenericutes was associated with decreased risk, hazard ratio 0.90.

The functional analysis is the part that closes the loop. Microbial pathways for lipopolysaccharide biosynthesis were among those associated with fracture risk, while Tenericutes abundance correlated with butyrate synthesis pathways. The authors’ interpretation is that a high Gammaproteobacteria burden increases circulating LPS and produces inflammation-induced bone loss – which is the mechanism described above, measured against fractures over eighteen years.

What you eat shapes which bacteria you have. Populations eating low-fiber diets high in fat, processed food and meat carry more of the bacterial groups linked to higher fracture risk, and far fewer of the fiber-fermenting bacteria. Populations eating high-fiber traditional diets show the reverse, with markedly higher levels of the beneficial short-chain fatty acids their bacteria produce.

Butyrate – the short-chain fatty acid most linked to bone – is made by only a handful of bacterial species, and they make it by fermenting dietary fiber. Nothing else feeds them. Depletion of these butyrate-producing bacteria tracks with Western dietary patterns, food additives, and antibiotic exposure. A high-meat, low-fiber pattern pushes fermentation toward protein breakdown rather than fiber fermentation, and toward the bacterial groups associated with more fracture rather than fewer.

Butyrate also connects to the muscle side of the equation, which matters given the mediation finding in the preceding section. High circulating butyrate has been causally associated with greater lean mass, gut microbial butyrate synthesis was found to affect skeletal muscle mass in healthy menopausal women, and butyrate produced by gut bacteria appears to be involved in exercise-induced gains in bone mineral density.

Two limits before the measurement question. The effect sizes above are modest, 8 to 14 percent per standard deviation, and they are associations rather than demonstrated causes. And the interventional evidence lags well behind: the microbiome is a well-established regulator of bone in rodents, but first-generation probiotic trials in humans have produced only modest effects on bone density, with no fracture endpoints. Diet-first remains the defensible position.

Can the gut be measured?

 Partly, and it is worth being precise about what the numbers do and do not tell you.

The FINRISK analysis rested on three findings: greater alpha diversity was protective, Proteobacteria abundance was associated with higher fracture risk, and the protective taxa correlated with butyrate synthesis pathways. Those three variables are not exotic. A strain-level metagenomic stool test reports all three directly.

The GutID Complete Microbiome Assessment is worth naming because the overlap is unusually close, and its panel continues to expand. It identifies all bacteria present down to the strain level and reports alpha diversity calculated using the Shannon Index, which is the same metric FINRISK used. It reports Proteobacteria as a percentage, describing them as gram-negative bacteria characterized by lipopolysaccharide on the outer membrane, that may activate the immune system response and cause systemic inflammation – which is the mechanism described earlier in this post, measured directly. And it reports the abundance of short-chain fatty acid producing bacteria, noting that dietary fiber is metabolized by these bacteria to generate acetate, propionate, and butyrate. It also reports lipopolysaccharide production, the Firmicutes to Bacteroidetes ratio, and an antibiotic resistome score.

A methodological point worth understanding, because it explains why two gut tests can both be relevant without measuring the same thing. FINRISK inferred butyrate synthesis from sequencing data rather than measuring fecal butyrate concentration. Sequencing tells you which organisms are present and what they are capable of producing; a metabolite assay tells you how much of the compound is actually there. The two answer different questions, and the sequencing question is the one the fracture cohort asked. Markers of barrier integrity and intestinal inflammation – zonulin, calprotectin, lipopolysaccharide-binding protein – answer a third question again, about whether the barrier is leaking. A panel reporting several of these is measuring a pathway from more than one angle, which is a reasonable thing to do, but no single one of those numbers is the fracture-relevant one.

Now the limits, stated plainly.

No gut test has a validated fracture-risk threshold. There is no Proteobacteria percentage above which fracture risk is established to rise, no Shannon index below which a patient is classified at risk. The FINRISK effect sizes were population-level hazard ratios of 0.92 and 1.14 per standard deviation – real, replicable, and far too small to act on in a single patient. A composite microbiome score below 60 on the GutID report indicates known issues with bacterial composition, but that is a general gut-health judgment, not a skeletal one.

Note also what is absent. GutID organizes its findings around gut axes – brain, gastrointestinal, metabolism, heart, and immune among them – and none of them is skeletal. That is not a criticism of the test. The gut-bone literature is recent enough that no commercial panel has yet built a bone axis around it, and any panel reporting diversity, Proteobacteria, and butyrate-producer abundance is already reporting the fracture-relevant variables whether or not it labels them that way. The report also states explicitly that it is not intended for diagnostic purposes.

So the honest use is mechanism and monitoring, not prediction. If a patient has elevated skin autofluorescence, elevated inflammatory markers, and a stool test showing high Proteobacteria with low butyrate-producer abundance, those findings cohere: a pro-inflammatory bacterial population, reduced production of the metabolite that appears protective, and the downstream signaling that suppresses bone formation. That picture justifies dietary and gut-directed intervention on its own terms.

It also gives something to re-measure, which matters more than any single reading. Proteobacteria abundance that falls and butyrate-producer abundance that rises on a higher-fiber, lower-processed-food diet is evidence the intervention is working. That is a reasonable thing to track. A fracture prediction it is not.

GOING DEEPER: THE CELLULAR ZOOMER

For patients whose SAF is elevated and whose serum CML or oxidative stress markers are abnormal, the Cellular Zoomer adds a comprehensive functional assessment of why bone quality may be deteriorating:

• 24 Mitochondrial and Energy Markers: Reveal whether osteoblasts—which are highly ATP-dependent—have the energy production they need for bone matrix synthesis.
• Methylmalonic Acid (B12 Status): Reflects vitamin B12 status, and B12 deficiency is linked to hip fracture risk through homocysteine accumulation. The evidence for that link is stronger than it is usually given credit for. Two prospective population-based studies following 2,406 subjects aged 55 and older, across 11,253 person-years, found that fracture risk rose roughly 30 percent for each one standard deviation increase in total homocysteine after adjustment for age and sex, with a relative risk of 1.4 per standard deviation in the fully adjusted model. Subjects in the highest age-specific quartile carried 1.9 times the fracture risk of those below it. The associations were independent of bone mineral density, and the authors characterized elevated homocysteine as a strong and independent risk factor for osteoporotic fracture, similar in magnitude to established fracture and cardiovascular risk factors.
  The accompanying editorial was titled “Homocysteine and osteoporotic fractures – culprit or bystander?”, and that question is not settled. But as a marker, homocysteine tells you something a DEXA does not.
• Pyridoxic Acid (B6 Status): Reflects vitamin B6 status. Pyridoxamine—a form of B6—is the strongest dietary AGE inhibitor with specific bone evidence in research literature.
• Pyroglutamic Acid (Glutathione Turnover): Reflects glutathione turnover, which the glyoxalase system needs to detoxify the methylglyoxal driving bone collagen glycation.
• Antioxidant Enzyme Variants: Identifies inherited genetic variants that explain why some patients accumulate more oxidative damage from a given AGE burden than others.

The case for assessing B12 and B6 status is not a case for taking more of them. In a cohort of 75,864 postmenopausal women in which 2,304 hip fractures occurred, a combined high intake of vitamins B6 and B12 from food and supplements was associated with an increased risk of hip fracture. The authors’ conclusion was that caution is warranted in vitamin supplementation when there is no apparent deficiency.

 This is worth sitting with, because it runs against the intuition that more of a bone-relevant nutrient is better. It may also help explain a puzzle in the homocysteine literature: elevated homocysteine is a well-replicated fracture risk factor, but trials of homocysteine-lowering B-vitamin treatment have not consistently reduced fracture incidence.

The practical position: measure methylmalonic acid and pyridoxic acid, correct a deficiency that is actually present, and stop there. This is precisely why the Cellular Zoomer markers matter more than a supplement protocol.

A FIFTH DIMENSION: BIOELECTRICAL IMPEDANCE PHASE ANGLE

Bioelectrical impedance phase angle (PhA) is a measurement derived from BIA body composition analysis that reflects cell membrane integrity and cellular health. Where SAF measures AGE accumulation in collagen and CML measures circulating AGE burden, phase angle measures the downstream cellular consequence of that burden: how well individual cells are maintaining their structural integrity.

The bone health evidence for phase angle is substantial:

• The Yakumo Study: A population-based cross-sectional study identified low phase angle as an independent predictor of osteoporosis regardless of age or sex.
• BMD & BMC Correlations: Multiple studies confirm positive associations between phase angle and femoral neck bone mineral density and whole-body bone mineral content in older adults, CKD patients, and athletes.
• Fracture Outcomes & Survival: A study published in Nutrients identified phase angle as an independent predictor of 12-month survival after hip fracture surgery, with critical risk cutoff values of 4.05 degrees for women and 4.65 degrees for men—patients below these thresholds had significantly higher mortality at 1, 3, 6, and 12 months post-surgery.
• Functional Recovery: A separate prospective study found that patients with phase angle below 3.0 degrees had substantially worse walking and balance scores at discharge after hip fracture surgery. Research in rehabilitation settings confirms that patients who increase their phase angle during recovery show direct improvements in daily living activities.

Phase angle measurement is available in this practice as part of BIA body composition assessment. For patients whose SAF is elevated or whose CML is abnormal, phase angle adds the cellular health dimension that explains why some patients fracture more easily and recover more poorly than their bone density alone would predict.

A SIXTH DIMENSION: HAND GRIP STRENGTH

A simple hand dynamometer measurement adds a dimension that no laboratory or imaging study provides: grip strength.

• Kuopio Osteoporosis Study: In a 15-year population-based study of 971 perimenopausal women, grip strength predicted fracture only in the subgroup whose baseline bone density was normal – and there it predicted strongly, with the lowest quartile carrying a hazard ratio of 2.0. Among the 284 women already classified as osteopenic or osteoporotic, grip strength added nothing. In the normal-density group, grip strength and T-score were the only significant predictors of 15-year fracture-free survival. The measurement earns its place exactly where the density scan is least informative.
• Meta-Analysis: A review of 19 prospective cohort studies confirmed handgrip strength as a risk indicator for future fractures across diverse populations.
• Genetic Causality: A 2024 Mendelian randomization study established a genetic causal relationship between low hand grip strength and both osteoporosis and fall risk.

                                                                                                                                                                                                                                                            Grip strength captures three fracture-relevant dimensions simultaneously:

1. Mechanical Loading: Drives bone remodeling through muscle contraction.
2. Fall Risk: The proximate cause of most fractures regardless of bone quality.
3. Sarcopenia Status: Declining muscle mass removes the IGF-1, irisin, and beta-aminoisobutyric acid stimulus that sustains osteoblast activity.

AGE-driven sarcopenia connects dietary AGE burden directly to grip strength decline and, through it, to fracture risk.

A hand dynamometer measurement takes thirty seconds and costs nothing. It belongs in every bone health assessment—and for patients with diabetes, metabolic syndrome, or elevated SAF, a declining grip strength trend is one of the most clinically accessible signals that musculoskeletal AGE burden is translating into functional vulnerability.

A NOTE ON DEXA

None of this makes DEXA obsolete. Bone density remains a valid predictor of fracture risk. A large 2024 Danish cohort of over 42,000 patients found that T-score predicted fracture risk equally in people with and without type 2 diabetes.

 There is also a bone quality measure already built into the DEXA you have. Trabecular bone score, or TBS, is a textural analysis of the lumbar spine DEXA image that estimates bone microarchitecture, and it predicts fracture independently of bone mineral density. A meta-analysis pooling individual-level data from 17,809 men and women across 14 prospective cohorts, with a mean follow-up of 6.7 years, found that TBS predicted major osteoporotic fracture independently of FRAX probability – and that adjusting FRAX for TBS produced more accurate risk estimates than FRAX alone. On average, each one standard deviation decline in TBS corresponds to a 30 to 40 percent increase in fragility fracture risk in postmenopausal women.

Two practical points. TBS is derived from the DEXA images already acquired, so it requires no additional scanning time, no additional radiation, and no additional visit. And bone mineral density and TBS are together described as the two pillars of the World Health Organization’s clinical definition of osteoporosis – so this is not a fringe measurement.

One limitation worth stating: TBS should not be used to monitor antiresorptive therapy, though it may prove useful for monitoring anabolic treatment.

The practical catch is availability. TBS requires licensed software, and while it installs on the DEXA scanners already in use – it is compatible with the GE, Hologic, and OsteoSys systems that cover most of the market – only a minority of DEXA facilities have licensed it. Adoption is concentrated in academic medical centers; the manufacturer reports use at 16 of the top 20 US hospitals, which tells you as much about where TBS is not as where it is. The manufacturer’s online site locator is worth a look but has proven unreliable, listing facilities that turn out to know nothing about it when called.

If you want TBS, the question that gets a reliable answer is a billing question rather than a clinical one. Ask the imaging center whether they bill CPT code 77089, which covers TBS with interpretation and report. If they do not, ask about 77090 – a code that exists specifically for a facility to acquire the DEXA and transmit the data elsewhere for TBS calculation. That second option means a center does not need to own the software at all, only be willing to send the images out. A billing department can answer this in a minute where a front desk cannot, and it can often be applied to DEXA images already on file.

The reason a bone quality measure was developed at all is the same reason this post exists. Most fragility fractures occur in people whose T-score sits above the -2.5 diagnostic threshold. The Rotterdam Study found that only 44 percent of non-vertebral fractures occurred in women at or below it, and the International Working Group on DXA Best Practices notes that these fractures confirm skeletal fragility while evaluation and treatment reach only a small percentage of the people who sustain them. Density is a real risk gradient. It is not, by itself, a decisive test.

The accurate position: DEXA should not be the sole basis for fracture risk assessment in patients with diabetes, metabolic syndrome, or elevated AGE burden. SAF, CML, and oxidative stress assessment add dimensions that DEXA cannot provide—and together they give a complete picture that neither test gives alone.

Thiazolidinediones (pioglitazone, rosiglitazone): Associated with increased fracture risk in women and postmenopausal patients based on randomized-trial safety data. Discuss fracture risk assessment with your clinician.

GLP-1 receptor agonists: A 2024 randomized trial in JAMA Network Open found that liraglutide alone reduced hip and lumbar spine BMD, while liraglutide plus exercise preserved bone. For patients losing weight on a GLP-1 agonist, concurrent resistance training is the specific evidence-backed recommendation for bone preservation.

Additional Medications: Acid blockers (PPIs and H2 blockers), select antidepressants and antipsychotics, anti-seizure medications, aromatase inhibitors (breast cancer treatment), glucocorticoids (long-term steroids), heparin, immunosuppressants, opioids, and excess thyroid hormone. Bisphosphonates—commonly prescribed for osteoporosis—have their own side-effect profile worth discussing.

If you take any of these medications long-term, this practice can help evaluate your overall bone health picture and identify modifiable factors that may help protect bone quality alongside your current treatment.

WHAT ACTUALLY HELPS

No clinical trial has yet shown that reducing AGE intake prevents fractures. The mechanism is coherent, the biomarker associations are real, but the intervention arm is unproven. What follows is ranked by evidential strength:

• Resistance and Impact Loading: The best-supported bone intervention in every risk group. Muscle and bone are coupled—contraction loads bone and drives remodeling, while muscle secretes IGF-1, irisin, and beta-aminoisobutyric acid that promote osteoblast activity. This leads ahead of every dietary or supplement lever.
• Dietary AGE Reduction & Cooking Changes: Less grilled and processed meat, moist heat cooking methods, and high polyphenol density reduce the glycation burden that SAF detects—and simultaneously reduce the inflammatory cytokine production that suppresses osteoblasts. 
• Gut Barrier Support: Metabolic endotoxemia – chronically elevated circulating LPS from a leaky gut – activates TLR4 and NF-kB, driving the same IL-6, TNF-alpha, and IL-1beta that suppress osteoblasts. A plant-diverse, high-fiber diet that supports butyrate-producing bacteria and maintains intestinal barrier integrity reduces this inflammatory bone-damaging pathway as effectively as reducing dietary AGE intake – and the two operate synergistically.
  This is no longer mechanism alone. In the FINRISK 2002 cohort, 7,043 adults had gut microbiome composition characterized by shotgun metagenomic sequencing and were followed for a median of 18 years, during which 1,092 incident fractures occurred. Greater microbial diversity was associated with lower fracture risk, with a hazard ratio of 0.92 for each standard deviation increase in the Shannon diversity index. Abundance of Proteobacteria was associated with increased fracture risk, hazard ratio 1.14, an effect driven by the Gammaproteobacteria class. Abundance of Tenericutes was associated with decreased risk, hazard ratio 0.90.
  The functional analysis is the part that closes the loop. Microbial pathways for lipopolysaccharide biosynthesis were among those associated with fracture risk, while Tenericutes abundance correlated with butyrate synthesis pathways. The authors’ interpretation is that a high Gammaproteobacteria burden increases circulating LPS and produces inflammation-induced bone loss – which is the mechanism described above, measured against fractures over eighteen years.
  Supporting evidence points the same direction from two other angles. In postmenopausal women stratified by T-score, plasma LPS was significantly elevated in those with osteoporosis compared with controls, indicating disrupted intestinal integrity. And butyrate specifically has been linked to the muscle side of the equation: high circulating butyrate has been causally associated with greater lean mass, gut microbial butyrate synthesis was found to affect skeletal muscle mass in healthy menopausal women, and butyrate produced by gut bacteria appears to be involved in exercise-induced gains in bone mineral density. Given that muscle mass mediates more of inflammation’s effect on hip fracture than bone density does, this is not a peripheral finding.
  The limit to state plainly, and it mirrors the AGE position exactly: diet to microbiome and microbiome to fracture are each established, but nobody has tested diet to microbiome to fracture end to end.
  Two honest limits. The effect sizes above are modest – 8 to 14 percent per standard deviation – and they are associations, not demonstrated causes. And the interventional evidence lags well behind: the microbiome is a well-established regulator of bone in rodents, but first-generation probiotic trials in humans have produced only modest effects on bone density, with no fracture endpoints. Diet-first remains the defensible position.                                                                                                                                                                                                                     In a cohort of 7,043 adults followed for a median of eighteen years, gut bacterial composition predicted incident fracture – greater microbial diversity was protective, and the taxa that correlated with butyrate synthesis were the protective ones. See “The Gut Dimension” above for the detail.
• Conventional Bone Nutrition: Adequate protein, calcium, vitamin D, vitamin K2, and magnesium remain valid alongside the glycation framing, and there is prospective data behind several of them. A meta-analysis of four cohort studies and one nested case-control study, covering 80,982 participants and 1,114 fracture cases, found a statistically significant inverse association between dietary vitamin K intake and fracture risk, with a relative risk of 0.78 comparing highest to lowest intake. One caveat matters here: every included study measured phylloquinone, vitamin K1, which is the predominant form in food – so this evidence does not speak directly to the K2 and MK-7 forms usually supplemented. In the Kuopio Ischemic Heart Disease cohort, 2,245 men followed for a median of 25.6 years, low serum magnesium was associated with increased fracture risk in a non-linear relationship. In men followed for a median of 9.5 years, dietary calcium above 1,000 mg per day carried a hazard ratio of 0.27 for osteoporotic fracture compared with 400 mg per day or less. Vitamin D status predicts hip fracture and mortality.
  One correction to a common assumption, worth making because it points back to loading rather than supplements: a systematic review of randomized trials found that vitamin D monotherapy produced no significant improvement in grip strength, gait speed, appendicular lean mass, overall muscle strength, or physical performance in adults aged 50 and above. Benefits were more evident in people with baseline deficiency, and when vitamin D was combined with adequate protein intake and resistance training. Correct a measured deficiency. Do not expect a supplement to do the work that loading does.
• Glycemic Control (Sustained Mean): Rotterdam Study data found that HbA1c at or above 7.5% in type 2 diabetes carried a 47 to 62 percent higher fracture risk – while simultaneously showing 1.1 to 5.6 percent higher bone mineral density and 4.6 to 5.6 percent thicker femoral cortices. Adequately controlled diabetes carried a risk similar to no diabetes at all, which makes glycemic control rather than diabetes status the variable that matters. The ARIC Study found a 1.63-fold increased fracture risk at HbA1c above 8%. Long-term average glycemic burden matters—not hour-to-hour variability. Caveat: Very low HbA1c in older patients on glucose-lowering therapy raises hypoglycemia and fall risk.
• Reducing Visceral Adiposity: Visceral adipose tissue secretes TNF-alpha, IL-6, and IL-1beta continuously. Waist circumference above 80cm in women and 100cm in men identifies the central adiposity pattern most associated with inflammatory bone damage.
• Creatine Monohydrate as an Adjunct: Evidence for direct BMD benefit is mixed. Maintenance dosing is 3-5g/day. Note: Creatine raises serum creatinine without affecting kidney function—use cystatin C to assess renal function in patients taking creatine.

The chronic inflammation driving bone quality loss has many upstream sources beyond dietary AGEs:

• Gut Dysbiosis & Endotoxemia: Delivers LPS through a compromised intestinal barrier.
• Visceral Adipose Tissue: Generates inflammatory cytokines around the clock.
• Sleep Disruption: Elevates IL-6 and TNF-alpha within days.
• Psychological Stress: Activates HMGB1 release and cortisol elevation, both directly suppressing osteoblast function.
• Physical Inactivity: Removes mechanical loading stimulus and anti-inflammatory myokine secretion.
• Periodontal Disease: Maintains persistently elevated systemic CRP and IL-6.
• Acid-Forming Dietary Pattern: Excess processed food, salt, and animal protein without sufficient alkaline-forming vegetables/fruit creates low-grade metabolic acidosis that mobilizes calcium from bone as a buffer.
• Environmental Toxins: Cadmium (tobacco/contaminated food) competes with calcium in bone mineral; lead accumulates in bone and releases during resorption; PFAS compounds correlate to lower bone density.

The dietary and lifestyle framework I apply in my practice addresses these multiple inflammatory drivers simultaneously through gut barrier support, polyphenol-mediated NF-kB suppression, glycemic control, and resistance training.

URINE BONE RESORPTION MARKERS: THE RATE DIMENSION

Skin autofluorescence and CML measure AGE accumulation that has already occurred—cumulative damage over years and decades. A complementary dimension is how fast bone is currently being broken down. Urine bone resorption markers—particularly deoxypyridinoline (DPD) and pyridinoline (PYD)—measure the collagen crosslink fragments released when osteoclasts actively degrade bone matrix.

• DPD: Particularly specific to bone—it comes almost exclusively from bone collagen rather than cartilage or other connective tissue.
• PYD: Present in both bone and cartilage.

The combination is where this becomes clinically decisive. Among women with a femoral bone mineral density of -2.5 standard deviations or below, those who also had high C-terminal telopeptides of type I collagen or high free deoxypyridinoline carried hip fracture odds ratios of 4.8 and 4.1 respectively, compared with women who had only low bone density or only high bone resorption. The predictive value of bone turnover markers persists after adjustment for bone mass, which is what makes them complementary to densitometry rather than redundant with it.

Low density and rapid resorption are not the same finding, and a patient with both is in a different risk category than a patient with either alone.

Elevated urine DPD in a patient with normal or osteopenic DEXA means bone is being resorbed faster than the density number has yet registered. The inflammatory signaling downstream of AGE-RAGE activation suppresses bone formation and, at the systemic level, favors resorption—though AGE-crosslinked matrix locally resists osteoclastic digestion, so the direction depends on the level of analysis. Either way, a patient with high SAF and elevated urine DPD has both the accumulated glycation damage and active ongoing resorption occurring simultaneously.

Together, the three measurements tell a complete story: SAF reveals cumulative glycation damage, urinary CML reflects current AGE burden, and urine resorption markers reveal the rate of active bone destruction happening right now. In this practice, PYD and DPD are measured through the Vibrant Wellness Hormone Zoomer panel.

I am currently developing the Anti-AGEing+ Nutrition Analyzer, an AI-powered tool that will estimate the AGE content of meals and provide personalized dietary guidance. The Anti-AGEing Protocol—my comprehensive dietary framework for reducing AGE burden and its consequences across aging and chronic disease—is also in development. Make sure you are signed up for my newsletter to be notified when both become available.

This assessment is relevant whether your DEXA is normal, osteopenic, or osteoporotic:

• Normal DEXA: Find the fracture risk your scan cannot detect.
• Osteopenia: Stop waiting. The assessment reveals whether your actual risk is higher than your T-score suggests and identifies specific, targeted interventions rather than watchful waiting.
• Osteoporosis: Your medication addresses bone density. This assessment addresses the inflammatory, metabolic, cellular, and musculoskeletal dimensions your medication does not.

For the genuinely low-risk osteopenic patient with no other risk factors, the assessment establishes a baseline across dimensions that DEXA cannot track—so that if those values begin to change, you have an early warning before a fracture tells you something went wrong.

Fee: $150. The assessment takes approximately 15 minutes.

Contact: Call 516-829-1515 or visit drdebe.com to schedule.

THE CLINICAL PICTURE

A patient walks in with a normal DEXA report and asks if their bones are healthy. The complete answer now has more dimensions than it did a decade ago:

• Skin Autofluorescence (SAF): Quantifies AGE accumulation in long-lived collagen.
• Urinary CML: Links tissue accumulation to hip fracture outcomes.
• hs-CRP: Indicates active inflammatory cascades driven by AGE burden.
• Cellular Zoomer: Evaluates mitochondrial energy, B-vitamin status, glutathione capacity, and antioxidant genetics.
• Phase Angle: Predicts cellular health and post-fracture survival.
• Grip Strength: Measures the muscle-bone axis and fall risk in thirty seconds.
• Skeletal Muscle Mass (BIA): Quantifies sarcopenic progression over time.

                                                                                                                                                                                                                                                         Almost all hip fractures happen because of a fall. A small number – under one in ten by most estimates, and one large study of video-captured falls found none at all – appear to happen the other way round, with the bone giving way under ordinary weight-bearing and the fall following. Either way the conclusion is the same: fall prevention is necessary but not sufficient. A fracture requires both a fall and a bone that cannot absorb it, and fall-prevention programs address only one of those two.

None of these assessments replace DEXA. Together, they complete it.

Bone fragility is multifactorial—hormonal, nutritional, inflammatory, metabolic, and mechanical. Glycation is not the entire story. But it is a significant and modifiable chapter of that story that a DEXA report will never show—and that a one-minute SAF measurement can reveal in the time it takes to read this paragraph.

REFERENCES

Skin autofluorescence, AGEs and oxidative stress

Rotterdam Study. Skin autofluorescence and osteoporotic fractures. PMC7687120.
Liu H et al. Skin autofluorescence-AGEage and bone in type 2 diabetes. J Diabetes. 2022;14(9):571-585.
Leanza G et al. Skin autofluorescence and bone quality in type 1 diabetes. eLife. 2024;12:RP90437.
Barzilay JI et al. Circulating carboxymethyl-lysine and incident hip fracture: the Cardiovascular Health Study. J Bone Miner Res. 2014;29:1061-1066.
Dhaliwal R et al. Carboxymethyl-lysine and fracture risk in older adults with type 2 diabetes: Health ABC. J Bone Miner Res. 2022;37(2):265-272. PMID: 34820902.
Schwartz AV et al. Pentosidine and increased fracture risk in older adults with type 2 diabetes. PMC2708944.
Tanaka S et al. Urinary pentosidine and fracture risk in severe osteoporosis: JOINT-05. JBMR Plus. 2022;6(10):e10673.
Rousseau JC et al. Urinary pentosidine and fracture risk in postmenopausal women: the OFELY study. Osteoporos Int. 2009.
Leon-Reyes G et al. Oxidative stress-related polymorphisms, bone mineral density and fracture. Antioxidants. 2023;12(4):915.

Bone density, trabecular bone score and bone turnover

Van Hulten V et al. Fracture risk revisited: bone mineral density T-score and fracture risk in type 2 diabetes. Diabetes Obes Metab. 2024;26(11):5325-5335. doi: 10.1111/dom.15890. PMID: 39228286.
McCloskey EV et al. A meta-analysis of trabecular bone score in fracture risk prediction and its relationship to FRAX. J Bone Miner Res. 2016. PMID: 26498132.
Update on trabecular bone score. PMC10118821.
Medimaps Group. TBS Osteo: compatible DXA platforms and CPT coding (77089, 77090, 77091, 77092).
Clinical utility of biochemical markers of bone turnover: fracture risk prediction and bone healing. PMC5995756.

Inflammatory markers

Cauley JA et al. Inflammatory markers and incident fracture risk in older men and women: the Health, Aging and Body Composition Study. J Bone Miner Res. 2007;22(7):1088-1095.
Cauley JA et al. Inflammatory markers and the risk of hip and vertebral fractures in men: the Osteoporotic Fractures in Men study. J Bone Miner Res. 2016;31(12):2129-2138.
Barbour KE et al. Inflammatory markers and the risk of hip fracture: the Women’s Health Initiative. J Bone Miner Res. 2012;27(5):1167-1176.
Dahl K et al. High-sensitivity C-reactive protein is an independent risk factor for non-vertebral fractures in women and men: the Tromsø Study. Bone. 2015;72:65-70.
Mun H, Liu B, Pham THA, Wu Q. C-reactive protein and fracture risk: an updated systematic review and meta-analysis of cohort studies through the use of both frequentist and Bayesian approaches. Osteoporos Int. 2021;32(3):425-435. doi: 10.1007/s00198-020-05623-6. PMID: 32935169.

The gut axis

Gut microbiome composition and incident fractures: the FINRISK 2002 cohort. npj Biofilms and Microbiomes. 2024. PMID: 39143108. (Shotgun metagenomics, n=7,043, 1,092 incident fractures, median 18-year follow-up.)Links between bone fractures and the gut microbiota. Nat Rev Endocrinol. 2026. doi: 10.1038/s41574-026-01286-1.
The gut-bone axis: impact of diet on gut microbiome and osteoporosis. Bone Res. 2026. doi: 10.1038/s41413-026-00550-4.
Osteoporosis in postmenopausal women is associated with disturbances in gut microbiota and migration of peripheral immune cells. BMC Musculoskelet Disord. 2024. s12891-024-07904-1.
Lv W-Q et al. Gut microbiome impacts skeletal muscle mass via microbial butyrate synthesis. J Cachexia Sarcopenia Muscle. 2021.
Lipopolysaccharide-induced bone loss in rodent models: a systematic review and meta-analysis. J Bone Miner Res. 2023;38(1):198.
GutID Complete Microbiome Assessment. Sample report, 2025. (Strain-level shotgun metagenomic sequencing; reports Shannon alpha diversity, Proteobacteria percentage, and short-chain fatty acid producing bacteria.)

Phase angle, grip strength and muscle

Tanaka S, Ando K, Kobayashi K, Hida T, Ito K, Tsushima M, et al. A low phase angle measured with bioelectrical impedance analysis is associated with osteoporosis and is a risk factor for osteoporosis in community-dwelling people: the Yakumo study. Arch Osteoporos. 2018;13:39. PMID: 29623499.
Sánchez-Torralvo FJ, Pérez-del-Río V, Navas Vela LI, García-Olivares M, Porras N, Abuín Fernández J, Bravo Bardají MF, García de Quevedo D, Olveira G. Phase angle as a predictor of mortality in older patients with hip fracture. Nutrients. 2024;16(14):2221. doi: 10.3390/nu16142221. PMC11279825.
Lim SK, Lim JY. Phase angle and functional recovery after hip fracture surgery. Arch Gerontol Geriatr. 2020. PMID: 32304889.
Sirola J, Rikkonen T, Tuppurainen M, Jurvelin JS, Alhava E, Kröger H. Grip strength may facilitate fracture prediction in perimenopausal women with normal BMD: a 15-year population-based study. Calcif Tissue Int. 2008;83(2):93-100. doi: 10.1007/s00223-008-9155-0. PMID: 18641912.
Ma Y et al. Grip strength, osteoporosis and fall risk: a Mendelian randomization study. Front Endocrinol. 2024.
Research progress on sarcopenia in the musculoskeletal system. Bone Res. 2025. doi: 10.1038/s41413-025-00455-8.

Nutrients and supplementation

van Meurs JBJ et al. Homocysteine levels and the risk of osteoporotic fracture. N Engl J Med. 2004;350:2033-2041.
Hao G et al. Vitamin K intake and the risk of fractures: a meta-analysis. Medicine. 2017;96(17).
Kunutsor SK et al. Low serum magnesium levels are associated with increased risk of fractures. PMC5570773.
Meyer HE et al. Association of high intakes of vitamins B6 and B12 from food and supplements with risk of hip fracture among postmenopausal women: the Nurses’ Health Study. PMC6512300.

Glycemic control and medications

Oei L, Zillikens MC, Dehghan A, et al. High bone mineral density and fracture risk in type 2 diabetes as skeletal complications of inadequate glucose control: the Rotterdam Study. Diabetes Care. 2013;36(6):1619-1628. PMID: 23315602.
Schneider AL, Williams EK, Brancati FL, Blecker S, Coresh J, Selvin E. Diabetes and risk of fracture-related hospitalization: the Atherosclerosis Risk in Communities Study. Diabetes Care. 2013;36(5):1153-1158.
Oei L, Rivadeneira F, Zillikens MC, Oei EHG. Diabetes, diabetic complications, and fracture risk. Curr Osteoporos Rep. 2015;13(2):106-115. doi: 10.1007/s11914-015-0260-5. PMC4352609. (Review; the source of the pooled HbA1c framing.)
Jensen SBK et al. Liraglutide, exercise and bone mineral density after weight loss. JAMA Netw Open. 2024;7:e2416775.

Dr. Joseph Debe, DC, DACBN, FMCP, CDN, CCSP is the Director of the Debe Center for Optimal Nutrition and Functional Medicine in Rockville Centre, NY. He is the author of the upcoming The Anti-AGEing Protocol. Visit drdebe.com.