Cassandra's Cortex

In Greek mythology, Cassandra was cursed with the gift of prophecy and the torment of disbelief; she saw Troy's destruction with perfect clarity and was ignored at every turn. The CKD skeleton is our Cassandra. It is already signaling catastrophic cortical deterioration while the T score returns a number that reads as normal, or at worst osteopenic, and sends clinician and patient alike away reassured. Bone mineral density measured by dual-energy x-ray absorptiometry (DXA) and the Fracture Risk Assessment Tool were built, validated, and refined in large cohorts that were largely, or entirely, free of CKD. The result is a diagnostic framework that renders a number where CKD demands a story. The number says the bone is osteopenic. The story says the patient is going to fall, fracture a hip, be hospitalized, and die at rates that eclipse those of people with identical T scores and intact kidneys. Into this diagnostic gap comes the study by Kim et al.,¹ published in this issue of JASN, which uses a nationwide Korean cohort of more than 550,000 postmenopausal women to interrogate the interplay between eGFR, DXA-based bone mineral density (BMD), and site-specific fracture risk. Its findings are clarifying and, in their clarity, alarming.

The central finding of Kim et al. will provide comfort to those who want it: DXA-based BMD predicts fracture risk in patients with CKD in essentially the same directional fashion as in the general population, confirming that DXA retains its value in CKD and supporting current Kidney Disease: Improving Global Outcomes guidelines recommending BMD assessment in patients with reduced eGFR.² Yet, to read the paper only for this finding is to miss its more provocative message. At any given level of BMD—normal, osteopenic, or osteoporotic—patients with lower eGFR face substantially greater hip fracture risk. Among participants with normal BMD and an eGFR below 45 ml/min per 1.73 m², the adjusted hazard ratio for hip fracture was 2.44-higher compared with the normal BMD, normal eGFR reference group. The skeleton looks the same on DXA. The outcome is not the same at all.

This is not a new biologic observation. Ghasem-Zadeh et al. demonstrated with high-resolution peripheral quantitative computed tomography that patients with CKD stages 4–5D have substantially impaired cortical and trabecular microarchitecture and reduced estimated failure loads at any BMD category—and that microarchitecture, not femoral neck BMD, independently predicted bone strength, accounting for over 85% of the variance.³ The prospective Structure of the Aging Men's Bones study by Bobiller et al. extends this picture into milder disease with sobering precision.⁴ Following 826 older men over 8 years with serial high-resolution peripheral quantitative computed tomography, they found that cortical area and density loss was already two- to three-fold faster in men with eGFR below 60 ml/min. More provocatively, even men with CKD stage 2 (eGFR 60–89) showed significantly faster cortical deterioration than the reference group—consistent across three eGFR equations, two skeletal sites, and absolute and percentage-based analyses alike. DXA cannot detect this accelerating decay. It is already underway when the creatinine barely registers concern.

Kim et al. illuminate the mechanistic basis of this diagnostic mismatch through the divergent behavior of hip versus vertebral fracture with declining eGFR. Hip fracture risk escalated steeply with lower eGFR at any given BMD; vertebral fracture risk did not. This maps directly onto skeletal anatomy: the hip is predominantly cortical bone, the vertebrae predominantly trabecular. Secondary hyperparathyroidism—virtually universal as eGFR declines—drives cortical bone loss through intracortical remodeling, while its effects on trabecular bone are inconsistent.⁵ CKD thus creates a skeleton that is disproportionately cortically impoverished, generating a hip fracture phenotype that escapes DXA-based detection. This is the fracture that kills: Postfracture mortality in patients with CKD is more than double that of kidney-healthy individuals, and 30-day mortality after hip fracture exceeds that seen after incident kidney failure.^6,7 The cortex is where CKD attacks the skeleton and precisely where our standard tools are most limited.

These observations demand a rethinking of how, and when, fracture risk is assessed in CKD. If cortical bone is already declining at an accelerated rate in men with eGFR 60–89 ml/min, waiting for a T score to cross the osteoporotic threshold is not watchful waiting. It is watching the clock run out. The 2023 Kidney Disease: Improving Global Outcomes Controversies Conference took meaningful steps in this direction, reframing CKD–mineral and bone disorder around CKD-associated osteoporosis and explicitly acknowledging that renal osteodystrophy is part of the osteoporosis spectrum.² The conference also recognized that the Fracture Risk Assessment Tool underestimates fracture risk in CKD—unsurprisingly, because it was not developed or validated in this population. What is needed is a fracture risk screening algorithm built from CKD data, for patients with CKD, incorporating the kidney-specific drivers of skeletal deterioration: eGFR trajectory, parathyroid hormone, fibroblast growth factor-23, bone turnover markers, and physical performance—and that identifies patients early enough, while there is still cortical bone to protect.

Kim et al. also deliver a practical clinical message through their physical performance subgroup analysis. Participants with CKD and an abnormal single-leg stance test—<5 seconds of unassisted standing—had significantly greater fracture risk, with a statistically significant interaction not replicated for the timed up and go test. This identifies a simple, no-cost assessment performable in any nephrology clinic today. Sarcopenia is endemic in CKD and a primary contributor to fall risk.⁸ Falls in a patient with covertly compromised cortical bone produce hip fractures. The single-leg stance test, completed in under a minute, identifies those most urgently in need of intervention.

And yet what is that intervention? Here, the field has too often answered with silence. Therapeutic nihilism—withholding treatment because of fears about safety, off-label status, or diagnostic uncertainty—is not a neutral act. It is a lethal one. More than one in ten American individuals has CKD, and among older adults with osteoporosis, more than half have some degree of CKD.² The Structure of the Aging Men's Bones data confirm that skeletal deterioration is already accelerating at eGFR levels most clinicians would not flag for fracture concern, and the excess mortality after fracture in this population is unambiguous and large.⁶ Leaving these patients unscreened and untreated because their T score has not crossed a threshold designed for someone else is not caution, it is abandonment.

The most immediately deployable, noncontroversial intervention available is weight-bearing exercise. Resistance training and weight-bearing physical activity improve muscle strength, balance, and bone mass in patients with CKD across all stages, including those on hemodialysis.⁹ Modifying fall risk does not require navigating complex drug–CKD interactions, off-label prescribing, or guideline gray zones. It requires a prescription for exercise, reinforced at every nephrology visit. Combined with broader DXA screening—which Kim et al. confirm remains appropriate and informative in CKD—and systematic identification of patients with poor static balance, this approach could meaningfully reduce hip fracture risk in a population currently failed by the tools we rely on.

Kim et al. have produced an important, large-scale dataset that clarifies where our fracture risk framework succeeds and where it breaks down in CKD. DXA remains a valid and useful tool. But a normal T score in a patient with an eGFR of 55 is not a clearance. It is a number that does not tell the full story of what CKD has done to the cortex, what a fall will do to that cortex, or what that hip fracture will do to the patient. Until we build a framework native to CKD—one that incorporates the cortical predilection of uremic bone disease, the sarcopenic vulnerability of this population, and the early stage at which skeletal damage begins—we will continue to do what the Trojans did to Cassandra: dismiss the warning, walk past the patient, and act surprised when the fracture comes. Cassandra was not wrong. She was simply not believed. The CKD skeleton has been telling us the same story for years. It is time we listened.

Disclosures

Disclosure forms, as provided by each author, are available with the online version of the article at http://links.lww.com/JSN/F742.

Author Contributions

Conceptualization: Thomas L. Nickolas.

Supervision: Thomas L. Nickolas.

Writing – original draft: Hazem El-Bilbeisi, Mahshid Mohseni, Thomas L. Nickolas.

Writing – review & editing: Thomas L. Nickolas.

Funding

None.