Osteoporotic vertebral compression fracture (OVCF) is a common and clinically important manifestation of osteoporosis in older adults and is associated with pain, disability, progressive spinal deformity, impaired mobility, and reduced quality of life [1,2]. In spine practice, OVCF represents a fracture event and a marker of vertebral skeletal fragility that influences sagittal alignment, functional status, treatment decisions, and future fracture risks, even when vertebral fractures are clinically silent or asymptomatic [3,4].

Bone mineral density (BMD) measured by dual-energy X-ray absorptiometry (DXA) is the current reference standard for osteoporosis diagnosis based on the World Health Organization T-score threshold of ≤−2.5 [5]. However, BMD explains only 60%–70% of bone strength, as vertebral strength also depends on bone quality, including trabecular microarchitecture, geometry, and tissue material properties [6]. This limitation is particularly relevant in vertebral compression fractures because many fragility fractures occur despite nonosteoporotic BMD, and prevalent vertebral fractures are frequently observed among patients with clinical risk factors and nonosteoporotic lumbar BMD [7].

Lumbar spine DXA interpretation in older adults is further complicated by degenerative changes, osteophytes, vertebral deformity, and overlying abdominal aortic calcification (AAC), which may artifactually elevate lumbar BMD and mask spinal fragility [8,9]. This diagnostic gap is clinically important because patients with radiographic OVCF may be classified as nonosteoporotic using conventional BMD thresholds, potentially delaying the recognition of bone fragility and appropriate osteoporosis care.

Trabecular bone score (TBS), a texture-based index derived from lumbar spine DXA images, provides complementary information on trabecular microarchitectural integrity not captured by BMD [6]. Low TBS indicates degraded bone microarchitecture and predicts osteoporotic fracture risk independently of BMD [10]. TBS has also been integrated into fracture-risk assessment models, including the Fracture Risk Assessment Tool (FRAX), to enhance future risk stratification beyond BMD alone [11]. In patients with nonosteoporotic BMD, TBS may help detect prevalent vertebral fractures and occult vertebral fragility that may be missed by BMD-based assessment [7,9].

AAC is a radiographic marker of vascular calcification that may reflect the systemic bone–vascular interplay underlying skeletal fragility [12,13]. AAC can be readily identified on lateral lumbar radiographs using validated semiquantitative scoring methods [12]. A greater AAC burden has been linked to lower BMD and elevated vertebral and nonvertebral fracture risk and lower TBS independent of BMD [14,15]. However, whether AAC burden provides additional clinical insight into BMD–TBS discordance and radiographically confirmed OVCF remains uncertain.

Although BMD, TBS, and AAC have each been investigated in relation to fracture risk, few studies have evaluated these measures together within a matched framework focused on radiographically confirmed OVCF. The present study addresses this gap by concurrently assessing the lumbar spine, femoral neck, and total hip BMD, TBS, and high-burden AAC in an Asian spine-practice cohort. This approach enabled us to evaluate whether TBS and high-burden AAC provide additional clinical context beyond conventional DXA-derived BMD assessment.

This study aimed to examine the relationship between TBS and OVCF after accounting for BMD at conventional DXA sites, including the lumbar spine, femoral neck, and total hip. Secondary objectives were to investigate BMD–TBS discordance, defined as nonosteoporotic BMD at all measured DXA sites with low TBS, and to evaluate whether high-burden AAC was associated with OVCF after adjusting for BMD and TBS.

This matched case-control study provides clinically oriented relevant evidence on BMD–TBS discordance in radiographically confirmed OVCF within an Asian spine-practice cohort. Building on prior evidence linking TBS and AAC to fracture risk, we evaluated lumbar spine, femoral neck, and total hip BMD, TBS, and high-burden AAC within a single matched framework focused on Genant-confirmed OVCF. Three findings are most relevant. First, low TBS remained independently associated with OVCF after accounting for the lowest all-site BMD T-score. Second, more than half of OVCF cases did not meet the osteoporotic threshold at any measured DXA site, and low TBS was approximately twice as common among these patients as in matched controls. Third, although high-burden AAC was more prevalent among OVCF cases, its association was attenuated after accounting for all-site BMD and low TBS, supporting its interpretation as a contextual vascular marker rather than a standalone skeletal fragility predictor. These findings are relevant to spine practice because radiographic OVCF may reflect vertebral fragility despite nonosteoporotic BMD at conventional DXA sites.

A major challenge in vertebral fracture evaluation is that BMD alone does not fully reflect skeletal fragility. Although a T-score ≤−2.5 remains the diagnostic threshold for osteoporosis, many fragility fractures occur in individuals whose BMD is in the osteopenic or even normal BMD range [5,22–24]. This limitation is particularly relevant in the spine, where degenerative changes, vertebral deformity, and vascular calcification may artifactually elevate areal BMD and obscure underlying vertebral fragility [6,9]. The present study extends these observations by demonstrating that this diagnostic gap was not confined to lumbar spine BMD alone: 53.2% of OVCF cases did not meet the osteoporotic threshold even after incorporating femoral neck and total hip BMD. This finding supports the clinical concern that conventional DXA thresholds may fail to identify vertebral fragility in a substantial proportion of patients with radiographically confirmed OVCF.

Hip BMD plays an important role in osteoporosis diagnosis and fracture-risk assessment; however, because femoral neck and total hip BMD are measured at nonvertebral skeletal sites, they may not fully reflect local spinal fragility. TBS, derived from the lumbar DXA texture, offers complementary information by indirectly assessing the trabecular microarchitecture [6,25,26]. Prior cohort studies and meta-analyses have demonstrated that lower TBS predicts vertebral and major osteoporotic fractures independently of BMD and clinical risk factors [11,20,27]. Among patients with nonosteoporotic BMD, TBS has also been shown to improve the detection of prevalent vertebral fractures beyond BMD alone [7]. The present findings extend these observations to an all-site BMD framework: low TBS remained significantly associated with OVCF after adjusting for the lowest all-site BMD T-score, and this association persisted even when BMD was represented categorically as any-site osteoporosis. These results support TBS as a clinically useful adjunct for identifying vertebral fragility that may not be detected using conventional BMD thresholds [20,21]. Methodologically, TBS and lumbar BMD were derived from the same evaluable vertebral levels, minimizing regional mismatch. The association of low TBS in the model using the conventional any-site osteoporosis threshold further supports its clinical relevance when BMD is interpreted categorically. However, the model incorporating the continuous lowest all-site BMD T-score was emphasized because it preserves more densitometric information across conventional DXA sites.

The AAC findings should be interpreted cautiously because the observed associations varied across statistical models. In the primary cohort, high-burden AAC was more common among OVCF cases, but its association with OVCF was attenuated after adjustment for the lowest all-site BMD T-score and low TBS. In contrast, high-burden AAC remained associated with OVCF in the lumbar spine BMD-based sensitivity models. This discrepancy may be explained by several non-mutually exclusive factors. First, the AAC–BMD relationship may be site-dependent. Because AAC is anatomically adjacent to the lumbar spine, it may influence lumbar DXA interpretation [9,14], whereas femoral neck and total hip BMD provide non-spinal skeletal information that is less affected by local aortic calcification. Second, lumbar spine-only models may not fully account for skeletal fragility because lumbar BMD can be artifactually increased by degenerative change, vertebral deformity, or vascular calcification [6,9]. Incorporating hip BMD may therefore minimize residual confounding and attenuate the apparent AAC association. Third, the relatively smaller number of OVCF cases may have contributed to sampling variability and limited precision across model specifications. Taken together, these findings suggest that high-burden AAC should not be interpreted as a direct measure of bone quality or as a standalone skeletal predictor in this cohort.

This interpretation is consistent with the bone–vascular axis concept, which proposes that vascular calcification and skeletal deterioration may share overlapping biological pathways [13]. Previous observational evidence has linked greater AAC burden to lower BMD and increased vertebral and nonvertebral fracture risk [14], while aortic vascular calcification has been inversely associated with TBS in patients undergoing dialysis [15]. In the present hospital-based cohort of non-dialysis patients, AAC was best interpreted as complementary vascular information that may coexist with skeletal fragility and degraded trabecular microarchitecture, rather than as a universal threshold for severe AAC or clinical intervention.

Methodologically, OVCF was conservatively defined using Genant grade 2 or 3 fractures rather than by diagnostic codes or mild grade 1 deformities, which have lower reproducibility [16,28,29]. In exploratory case-only analyses, the Genant-based fracture-burden score was not significantly associated with BMD, TBS, AAC-24 score, low TBS, high-burden AAC, or BMD–TBS discordance. Because these analyses were restricted to patients with confirmed OVCF and were not powered to evaluate fracture-burden gradients, they should be interpreted descriptively. Prior evidence also indicates that TBS and VFA may capture complementary skeletal risk dimensions [30].

This study has several strengths, including sex- and age-matched design, radiographic confirmation of OVCF, inclusion of a hospital-based Asian cohort, good-to-excellent inter-rater reliability, and alignment of TBS analysis with the same evaluable vertebral levels used for lumbar BMD. Nevertheless, several limitations should be acknowledged. First, the sample size was determined by the number of eligible OVCF cases with complete imaging and DXA-derived data during the study period, and no formal a priori sample size calculation was performed. A post hoc Wald-based precision assessment indicated that the study was primarily powered to detect relatively large associations, corresponding to an OR of approximately 2.9 or greater for a binary exposure with precision comparable to that of low TBS. Consequently, smaller independent associations may have been missed, particularly for high-burden AAC, interaction testing, and the combined exploratory phenotype. Second, the retrospective case-control design and the timing of DXA, TBS, and AAC assessments preclude causal or temporal inferences. Accordingly, low TBS observed in OVCF patients should not be interpreted as evidence that degraded trabecular microarchitecture necessarily preceded fracture occurrence. Third, osteoporosis treatment before DXA assessment was not excluded and may have influenced BMD and TBS differently. Prior antiresorptive or anabolic therapy may preserve or increase BMD despite persistent microarchitectural degradation, potentially increasing apparent BMD–TBS discordance. Conversely, effective treatment could mitigate fracture risk and attenuate associations with OVCF. Fourth, residual confounding by body mass index, comorbidities, medication exposure, fall risk, and other clinical fracture-risk factors cannot be excluded. Finally, the TBS cutoff of ≤1.230 was based on widely used international classification thresholds, but population-specific normative data and fracture-based validation for Thai patients and broader Asian spine-practice cohorts remain limited. In addition, the single-center, hospital-based design may limit the generalizability of the findings.

Clinically, these findings support the integrated assessment of BMD, TBS, and vertebral imaging when assessing vertebral fragility in older spine patients, particularly in those with radiographic OVCF despite nonosteoporotic BMD at conventional DXA sites.

Low TBS remained independently associated with radiographically confirmed OVCF after accounting for BMD across the lumbar spine, femoral neck, and total hip in this Asian spine-practice cohort. More than half of OVCF cases did not meet the osteoporotic threshold at any measured DXA site, highlighting a clinically relevant diagnostic gap with conventional BMD-based assessment. These findings support the use of TBS as an adjunct to conventional DXA-derived BMD for detecting vertebral skeletal fragility that may otherwise be overlooked by BMD thresholds alone. High-burden AAC should be interpreted as contextual vascular information rather than as an independent marker of skeletal fragility or a treatment threshold.