Advertisement

Increased risks of vertebral fracture and reoperation in primary spinal fusion patients who test positive for osteoporosis by Biomechanical Computed Tomography analysis

Open AccessPublished:November 09, 2022DOI:https://doi.org/10.1016/j.spinee.2022.10.018

      Highlights

      • Osteoporosis testing for spinal fusion patients using a pre-operative CT scan
      • Identify osteoporosis on basis of low bone strength and/or bone mineral density
      • With osteoporosis, 4.7X higher risk of vertebral fracture (than without osteoporosis)
      • If low bone strength and low bone mineral density, 3.7X higher risk of reoperation
      • Similar trends for patients with short (≤ 3 fused levels) or long fusion construct

      Abstract

      Background Context

      While osteoporosis is a risk factor for adverse outcomes in spinal fusion patients, diagnosing osteoporosis reliably in this population has been challenging due to degenerative changes and spinal deformities. Addressing that challenge, biomechanical computed tomography analysis (BCT) is a CT-based diagnostic test for osteoporosis that measures both bone mineral density and bone strength (using finite element analysis) at the spine; CT scans taken for spinal evaluation or previous care can be repurposed for the analysis.

      Purpose

      Assess the effectiveness of BCT for preoperatively identifying spinal fusion patients with osteoporosis who are at high risk of reoperation or vertebral fracture.

      Study Design

      Observational cohort study in a multi-center integrated managed care system using existing data from patient medical records and imaging archives.

      Patient Sample

      We studied a randomly sampled subset of all adult patients who had any type of primary thoracic (T4 or below) or lumbar fusion between 2005 and 2018. For inclusion, patients with accessible study data needed a preop CT scan without intravenous contrast that contained images (before any instrumentation) of the upper instrumented vertebral level.

      Outcome Measures

      Reoperation for any reason (primary outcome) or a newly documented vertebral fracture (secondary outcome) occurring up to 5 years after the primary surgery.

      Methods

      All study data were extracted using available coded information and CT scans from the medical records. BCT was performed at a centralized lab blinded to the clinical outcomes; patients could test positive for osteoporosis based on either low values of bone strength (vertebral strength ≤ 4,500 N women or 6,500 N men) and/or bone mineral density (vertebral trabecular bone mineral density ≤ 80 mg/cm3 both sexes). Cox proportional hazard ratios were adjusted by age, presence of obesity, and whether the fusion was long (four or more levels fused) or short (3 or fewer levels fused); Kaplan-Meier survival was compared by the log rank test. This project was funded by NIH (R44AR064613) and all physician co-authors and author 1 received salary support from their respective departments. Author 6 is employed by, and author 1 has equity in and consults for, the company that provides the BCT test; the other authors declare no conflicts of interest.

      Results

      For the 469 patients analyzed (298 women, 171 men), median follow-up time was 44.4 months, 11.1% had a reoperation (median time 14.5 months), and 7.7% had a vertebral fracture (median time 2.0 months). Overall, 25.8% of patients tested positive for osteoporosis and no patients under age 50 tested positive. Compared to patients without osteoporosis, those testing positive were at almost five-fold higher risk for vertebral fracture (adjusted hazard ratio 4.7, 95% confidence interval = 2.2–9.7; p<.0001 Kaplan-Meier survival). Of those positive-testing patients, those who tested positive concurrently for low values of both bone strength and bone mineral density (12.6% of patients overall) were at almost four-fold higher risk for reoperation (3.7, 1.9–7.2; Kaplan-Meier survival p<.0001); the remaining positive-testing patients (those who tested positive for low values of either bone strength or bone mineral density but not both) were not at significantly higher risk for reoperation (1.6, 0.7–3.7) but were for vertebral fracture (4.3, 1.9–10.2). For both clinical outcomes, risk remained high for patients who underwent short or long fusion.

      Conclusion

      In a real-world clinical setting, BCT was effective in identifying primary spinal fusion patients aged 50 or older with osteoporosis who were at elevated risks of reoperation and vertebral fracture.

      Keywords

      Introduction

      Several studies in spinal fusion patients have shown higher rates of complications to be associated with osteoporosis, including reoperation, nonunion, screw loosening, proximal junction kyphosis, cage subsidence, and vertebral fracture [
      • Shirley M
      • Wanderman N
      • Keaveny T
      • Anderson P
      • Freedman BA
      Opportunistic computed tomography and spine surgery: a narrative review.
      ,
      • Bjerke BT
      • Zarrabian M
      • Aleem IS
      • Fogelson JL
      • Currier BL
      • Freedman BA
      • et al.
      Incidence of osteoporosis-related complications following posterior lumbar fusion.
      ,
      • Khalid SI
      • Nunna RS
      • Maasarani S
      • Belmont E
      • Deme P
      • Chilakapati S
      • et al.
      Association of osteopenia and osteoporosis with higher rates of pseudarthrosis and revision surgery in adult patients undergoing single-level lumbar fusion.
      ,
      • Jain N
      • Labaran L
      • Phillips FM
      • Khan SN
      • Jain A
      • Kebaish KM
      • et al.
      Prevalence of osteoporosis treatment and its effect on post-operative complications, revision surgery and costs after multi-level spinal fusion.
      ,
      • Yagi M
      • Fujita N
      • Tsuji O
      • Nagoshi N
      • Asazuma T
      • Ishii K
      • et al.
      Low bone-mineral density is a significant risk for proximal junctional failure after surgical correction of adult spinal deformity: a propensity score-matched analysis.
      ,
      • Duan PG
      • Mummaneni PV
      • Rivera J
      • Guinn JMV
      • Wang M
      • Xi Z
      • et al.
      The association between lower Hounsfield units of the upper instrumented vertebra and proximal junctional kyphosis in adult spinal deformity surgery with a minimum 2-year follow-up.
      ,
      • Kim JS
      • Phan K
      • Cheung ZB
      • Lee N
      • Vargas L
      • Arvind V
      • et al.
      Surgical, radiographic, and patient-related risk factors for proximal junctional kyphosis: a meta-analysis.
      ,
      • Bredow J
      • Boese CK
      • Werner CM
      • Siewe J
      • Löhrer L
      • Zarghooni K
      • et al.
      Predictive validity of preoperative CT scans and the risk of pedicle screw loosening in spinal surgery.
      ,
      • Zou D
      • Muheremu A
      • Sun Z
      • Zhong W
      • Jiang S
      • Li W.
      Computed tomography Hounsfield unit-based prediction of pedicle screw loosening after surgery for degenerative lumbar spine disease.
      ,
      • Gupta A
      • Cha T
      • Schwab J
      • Fogel H
      • Tobert D
      • Razi AE
      • et al.
      Osteoporosis increases the likelihood of revision surgery following a long spinal fusion for adult spinal deformity.
      ,
      • Okuyama K
      • Abe E
      • Suzuki T
      • Tamura Y
      • Chiba M
      • Sato K
      Influence of bone mineral density on pedicle screw fixation: a study of pedicle screw fixation augmenting posterior lumbar interbody fusion in elderly patients.
      ,
      • Löffler MT
      • Sollmann N
      • Burian E
      • Bayat A
      • Aftahy K
      • Baum T
      • et al.
      Opportunistic osteoporosis screening reveals low bone density in patients with screw loosening after lumbar semi-rigid instrumentation: a case-control study.
      ,
      • Schwaiger BJ
      • Gersing AS
      • Baum T
      • Noël PB
      • Zimmer C
      • Bauer JS.
      Bone mineral density values derived from routine lumbar spine multidetector row CT predict osteoporotic vertebral fractures and screw loosening.
      ,
      • Liu Y
      • Dash A
      • Krez A
      • Kim HJ
      • Cunningham M
      • Schwab F
      • et al.
      Low volumetric bone density is a risk factor for early complications after spine fusion surgery.
      ,
      • Xi Z
      • Mummaneni PV
      • Wang M
      • Ruan H
      • Burch S
      • Deviren V
      • et al.
      The association between lower Hounsfield units on computed tomography and cage subsidence after lateral lumbar interbody fusion.
      ]. As a result, best practices for spinal fusion care have recently recommended the need for a formal preoperative evaluation of bone health [
      • Sardar ZM
      • Coury JR
      • Cerpa M
      • DeWald CJ
      • Ames CP
      • Shuhart C
      • et al.
      Best practice guidelines for assessment & management of osteoporosis in adult patients undergoing elective spinal reconstruction.
      ]. Appropriate evaluation can help surgeons identify which patients might benefit the most from treating with an osteoporosis medication or altering the surgical plan, or both, and help establish the medical necessity for any risk-mitigating steps.
      Until recently, obtaining a reliable diagnosis of spinal osteoporosis in spinal fusion patients has been challenging. The current clinical standard, measuring bone mineral density (BMD) by dual-energy x-ray absorptiometry (DXA) imaging, can often show false negatives in the presence of spinal deformities and degenerative changes [
      • Liu Y
      • Dash A
      • Krez A
      • Kim HJ
      • Cunningham M
      • Schwab F
      • et al.
      Low volumetric bone density is a risk factor for early complications after spine fusion surgery.
      ,
      • Tenne M
      • McGuigan F
      • Besjakov J
      • Gerdhem P
      • Akesson K.
      Degenerative changes at the lumbar spine–implications for bone mineral density measurement in elderly women.
      ,
      • Schneider P
      • Borner W.
      The impact of degenerative spinal changes on the correlation of peripheral and axial bone density.
      ,
      • Guglielmi G
      • Floriani I
      • Torri V
      • Li J
      • van Kuijk C
      • Genant HK
      • et al.
      Effect of spinal degenerative changes on volumetric bone mineral density of the central skeleton as measured by quantitative computed tomography.
      ,
      • Li N
      • Li XM
      • Xu L
      • Sun WJ
      • Cheng XG
      • Tian W.
      Comparison of QCT and DXA: osteoporosis detection rates in postmenopausal women.
      ,
      • St Jeor JD
      • Jackson TJ
      • Xiong AE
      • Kadri A
      • Freedman BA
      • Sebastian AS
      • et al.
      Osteoporosis in spine surgery patients: what is the best way to diagnose osteoporosis in this population?.
      ,
      • Zou D
      • Li W
      • Deng C
      • Du G
      • Xu N.
      The use of CT Hounsfield unit values to identify the undiagnosed spinal osteoporosis in patients with lumbar degenerative diseases.
      ] because DXA only provides planar images of the bone. Three-dimensional imaging, such as computed tomography (CT), can overcome this limitation. Some have suggested that quantitative CT should be used instead of DXA to measure BMD diagnostically for spinal fusion patients [
      • Kulkarni AG
      • Thonangi Y
      • Pathan S
      • Gunjotikar S
      • Goparaju P
      • Talwar I
      • et al.
      Should Q-CT be the gold standard for detecting spinal osteoporosis?.
      ]. However, quantitative CT typically requires the use of an external calibration phantom during imaging and specialized software and is not widely used. Others are investigating obtaining CT-based BMD measurements without a calibration phantom [
      • Löffler MT
      • Sollmann N
      • Burian E
      • Bayat A
      • Aftahy K
      • Baum T
      • et al.
      Opportunistic osteoporosis screening reveals low bone density in patients with screw loosening after lumbar semi-rigid instrumentation: a case-control study.
      ,
      • Schwaiger BJ
      • Gersing AS
      • Baum T
      • Noël PB
      • Zimmer C
      • Bauer JS.
      Bone mineral density values derived from routine lumbar spine multidetector row CT predict osteoporotic vertebral fractures and screw loosening.
      ], although generalization of some of those investigative techniques remains to be shown. Measurements of Hounsfield Units of attenuation from a CT scan have also been investigated [
      • Zaidi Q
      • Danisa OA
      • Cheng W.
      Measurement techniques and utility of Hounsfield unit values for assessment of bone quality prior to spinal Instrumentation: a review of current literature.
      ]. However, that approach does not constitute a diagnostic test due to various technical limitations [
      • Shirley M
      • Wanderman N
      • Keaveny T
      • Anderson P
      • Freedman BA
      Opportunistic computed tomography and spine surgery: a narrative review.
      ,
      • Xi Z
      • Mummaneni PV
      • Wang M
      • Ruan H
      • Burch S
      • Deviren V
      • et al.
      The association between lower Hounsfield units on computed tomography and cage subsidence after lateral lumbar interbody fusion.
      ,
      • Sardar ZM
      • Coury JR
      • Cerpa M
      • DeWald CJ
      • Ames CP
      • Shuhart C
      • et al.
      Best practice guidelines for assessment & management of osteoporosis in adult patients undergoing elective spinal reconstruction.
      ,
      • Zou D
      • Li W
      • Deng C
      • Du G
      • Xu N.
      The use of CT Hounsfield unit values to identify the undiagnosed spinal osteoporosis in patients with lumbar degenerative diseases.
      ,
      • Engelke K
      • Keaveny TM.
      Letter to the Editor.
      ,
      • Pinto EM
      • Neves JR
      • Teixeira A
      • Frada R
      • Atilano P
      • Oliveira F
      • et al.
      Efficacy of Hounsfield units measured by lumbar computer tomography on bone density assessment: A systematic review.
      ] and the lack of a consistent diagnostic cut point [
      • Gausden EB
      • Nwachukwu BU
      • Schreiber JJ
      • Lorich DG
      • Lane JM.
      Opportunistic use of CT imaging for osteoporosis screening and bone density assessment: a qualitative systematic review.
      ], therefore requiring a confirmation DXA for diagnostic purposes [
      • Sardar ZM
      • Coury JR
      • Cerpa M
      • DeWald CJ
      • Ames CP
      • Shuhart C
      • et al.
      Best practice guidelines for assessment & management of osteoporosis in adult patients undergoing elective spinal reconstruction.
      ].
      Another CT-based test for osteoporosis, named biomechanical computed tomography analysis (BCT), identifies spinal osteoporosis on the basis of both BMD and bone strength, the latter obtained from finite element analysis and providing additional diagnostic information over BMD alone [
      • Keaveny TM
      • Clarke BL
      • Cosman F
      • Orwoll ES
      • Siris ES
      • Khosla S
      • et al.
      Biomechanical Computed Tomography analysis (BCT) for clinical assessment of osteoporosis.
      ]. BCT has been validated in the general population for identifying women and men with spinal osteoporosis who are at high risk of vertebral fracture [
      • Melton LJ
      • Riggs BL
      • Keaveny TM
      • Achenbach SJ
      • Hoffmann PF
      • Camp JJ
      • et al.
      Structural determinants of vertebral fracture risk.
      ,
      • Melton LJ
      • Riggs BL
      • Keaveny TM
      • Achenbach SJ
      • Kopperdahl D
      • Camp JJ
      • et al.
      Relation of vertebral deformities to bone density, structure, and strength.
      ,
      • Wang X
      • Sanyal A
      • Cawthon PM
      • Palermo L
      • Jekir M
      • Christensen J
      • et al.
      Prediction of new clinical vertebral fractures in elderly men using finite element analysis of CT scans.
      ,
      • Anderson DE
      • Demissie S
      • Allaire BT
      • Bruno AG
      • Kopperdahl DL
      • Keaveny TM
      • et al.
      The associations between QCT-based vertebral bone measurements and prevalent vertebral fractures depend on the spinal locations of both bone measurement and fracture.
      ,
      • Kopperdahl DL
      • Aspelund T
      • Hoffmann PF
      • Sigurdsson S
      • Siggeirsdottir K
      • Harris TB
      • et al.
      Assessment of incident spine and hip fractures in women and men using finite element analysis of CT scans.
      ,
      • Allaire BT
      • Lu D
      • Johannesdottir F
      • Kopperdahl D
      • Keaveny TM
      • Jarraya M
      • et al.
      Prediction of incident vertebral fracture using CT-based finite element analysis.
      ,
      • Johannesdottir F
      • Allaire B
      • Kopperdahl DL
      • Keaveny TM
      • Sigurdsson S
      • Bredella MA
      • et al.
      Bone density and strength from thoracic and lumbar CT scans both predict incident vertebral fractures independently of fracture location.
      ]. In addition, in part via patient-specific phantomless calibration that utilizes the patient's internal tissues as calibrating references [
      • Lee DC
      • Hoffmann PF
      • Kopperdahl DL
      • Keaveny TM.
      Phantomless calibration of CT scans for measurement of BMD and bone strength — Inter-operator reanalysis precision.
      ], BCT can repurpose most types of spine-containing CT scans, without intravenous contrast, that are taken for any indication. BCT is FDA-cleared in the US as a diagnostic test for osteoporosis, and does not require confirmation by DXA [
      • Keaveny TM
      • Clarke BL
      • Cosman F
      • Orwoll ES
      • Siris ES
      • Khosla S
      • et al.
      Biomechanical Computed Tomography analysis (BCT) for clinical assessment of osteoporosis.
      ,
      • Keaveny TM.
      Letter to the editor concerning "Best practice guidelines for assessment and management of osteoporosis in adult patients undergoing elective spinal reconstruction".
      ].
      Previously, we described the use of BCT in a cohort of female spinal fusion patients, age 50–70, in a large academic medical center [
      • Burch S
      • Feldstein M
      • Hoffmann PF
      • Keaveny TM.
      Prevalence of poor bone quality in women undergoing spinal fusion using biomechanical-CT analysis.
      ]. In that study, 29% of patients tested positive for spinal osteoporosis by either bone strength or BMD criteria. That study did not include any clinical outcomes. Here, we assessed the effectiveness of BCT for identifying primary spinal fusion patients with osteoporosis who are at high risk of subsequent reoperation or vertebral fracture. To do so, we conducted an observational study in a real-world clinical setting to assess risks of reoperation and vertebral fracture for patients who tested positive for osteoporosis by BCT.

      Materials and methods

      Study design

      This was an observational cohort study in a multi-center integrated managed care system, using existing data available from patient medical records and imaging archives. The study was IRB approved with a waiver of informed consent.

      Patient population

      We sampled from all adult patients (age ≥ 18 years, both sexes) in a health care system in southern California who underwent any type of spinal fusion surgery at one of eight different hospitals between Jan 1, 2005, and Dec 31, 2018. In generating our analysis cohort from that population of 20,920 patients (Fig. 1), patients needed to have: 1) no previous fusion surgery; 2) if no reoperation or vertebral fracture, at least 3 months of follow-up; 3) a CT scan of the chest, abdomen, abdomen-pelvis or spine without intravenous contrast, taken within 1 year before surgery or immediately post-operatively. Due to resource constraints on both collecting data and analyzing the CT scans, from these eligible patients (N=4,027) we selected a random subset (N=1,348) for analysis, stratified by sex. From that subset, successful BCT analyses were performed on all those patients who had an archived CT covering vertebrae up to T4 that was suitable for BCT as well as an archived post-operative X-ray displaying the upper instrumented vertebral (UIV) level from T4–L5. The resulting analysis cohort comprised 469 patients with complete study data. The CT scans for these patients were taken on 21 different types of CT scanners, across five different manufacturers; 85% of the CT scans were for the thoracic or lumbar spine, the others comprising either pelvic/abdominal, chest, or whole-body CT scans and all were acquired without intravenous contrast.
      Fig 1
      Fig. 1Flow chart for development of the analysis cohort. Starting with a patient population of 20,920 primary fusion patients, a total of 469 patients with complete study data were included in this analysis. A) Any fusion was found by EMR search using CPT codes (22533, 22534, 22556, 22558, 22610, 22612, 22630, 22632, 22633, 22634, 22532, 22840, 22841, 22842, 22843, 22844, 22845, 22846, 22847, 22853, 22854); ICD-9 codes (81.0, 81.62, 81.63, 81.64, 84.51); or ICD-10 codes (0RG6, 0RG7, 0RG8, 0RGA, 0SG0, 0SG1). B) Revision fusion was found by EMR search using all codes above plus CPT codes (22849, 22850, 22852, 22855); ICD-9 (81.3) or ICD-10 codes (0RW604Z, 0RW60AZ, 0RW634Z, 0RW63AZ, 0RW644Z, 0RW64AZ, 0RW6 × 4Z, 0RW6XAZ, 0RWA04Z, 0RWA0AZ, 0RWA34Z, 0RWA3AZ, 0RWA44Z, 0RWA4AZ, 0RWAX4Z, 0RWAXAZ, 0SW004Z, 0SW00AZ, 0SW034Z, 0SW03AZ, 0SW044Z, 0SW04AZ, 0SW0 × 4Z, 0SW0XAZ). C) An eligible CT scan was a spine-containing CT scan acquired without intravenous contrast, taken within one year before the fusion surgery, day of surgery, or within 2 weeks postoperatively. Scans were found by an EMR search using CPT codes (71250, 71270, 72128, 72130, 72131, 72133, 72292, 74150, 74170, 77078, 74176, 74178) or ICD-10 codes (BR27ZZZ, BR29ZZZ, BW24ZZZ, BW20ZZZ, BW21ZZZ, BW25ZZZ, BW2000Z, BW2010Z, BW20Y0Z). D) Due to resource constraints, a sex-stratified random sample was selected for analysis (56% women), that proportion based on the eligible population with reoperation who were female. CPT is a registered trademark of the American Medical Association.

      Clinical outcomes

      The primary clinical outcome was a second spinal fusion surgery for any reason (“reoperation”), during the observation period of up to 5 years. A secondary clinical outcome was a vertebral fracture documented during the same observation period. Each patient's observation period was defined as the time span from their first surgery to death, disenrollment from the managed care program, 5 years, or the end of the study observation period of March 31, 2019, whichever came first. Any reoperation or vertebral fracture event must have occurred within the patient's observation period. All procedure, covariate, and outcome data were identified by electronic screening algorithms utilizing billing, procedure, and prescription codes (Fig. 1). Covariates that were available in the coded database were included in the analysis if deemed relevant to the clinical outcomes or known to affect bone strength. Included covariates were age, sex, race/ethnicity, weight, height, body mass index, diabetes, smoking, other bone conditions (see Table 1 for details), use of steroids or osteoporosis medications during the observation period, and number of fused levels (obtained also from postop x-rays). Full study data were available for all patients in the analysis cohort and were validated by cross-checking a subset against those in a more detailed spine registry [
      • Guppy KH
      • Harris J
      • Paxton LW
      • Alvarez JL
      • Bernbeck JA.
      Reoperation rates for symptomatic nonunions in anterior cervical fusions from a national spine registry.
      ,
      • Paxton EW
      • Inacio MCS
      • Kiley M-L.
      The Kaiser permanente implant registries: effect on patient safety, quality improvement, cost effectiveness, and research opportunities.
      ].
      Table 1Characteristics of the cohort (N=469 patients)
      Characteristic (categorical if %)Mean (SD) or % of cohort
      Demographics
       Age (y)63.6 (11.6)
       Age range (years)22–93
       Age ≥ 50 years (%)88.7
       Sex
      Female (%)63.5
      Male (%)36.5
       Race/ethnicity
      Non-Hispanic White (%)56.7
      Hispanic (%)24.7
      Black (%)11.9
      Asian or Pacific Islander (%)5.8
      Other (%)0.4
      Characteristics
       Height (cm)167 (10)
       Mass (kg)82.1 (19.1)
       Body mass index (kg/m²)29.2 (5.8)
       Vertebral trabecular BMD (mg/cm3)
      Measured at the nominal level (N=461 patients), which is the standard measurement site for assessment of osteoporosis (one vertebral level from T12–L3, preferably L1). Cut point for BMD-defined osteoporosis is 80 mg/cm3 for both sexes; and for fragile bone strength is 4,500 N for women and 6,000 N for men; patients have osteoporosis by BCT if either measurement is less than or equal to the respective cut point.
      118 (38)
       Vertebral strength (N)
      Measured at the nominal level (N=461 patients), which is the standard measurement site for assessment of osteoporosis (one vertebral level from T12–L3, preferably L1). Cut point for BMD-defined osteoporosis is 80 mg/cm3 for both sexes; and for fragile bone strength is 4,500 N for women and 6,000 N for men; patients have osteoporosis by BCT if either measurement is less than or equal to the respective cut point.
      Women6,080 (1,890)
      Men8,610 (2,580)
      Comorbidities
       Obese (BMI ≥ 30, %)38.4
       Diabetes (%)36.7
       Other bone conditions
      Other bone conditions that could impact bone health were defined as: any history of rheumatoid arthritis, osteopetrosis, Paget's disease, hypophosphatasia, osteogenesis imperfecta; a diagnosis from 2 years before surgery through the end of follow-up of osteomalacia, multiple myeloma, malignant neoplasms of the spine, or use of tumor necrosis factor (TNF) inhibitor; or a diagnosis during the follow-up period of hypocalcemia or hypercalcemia.
      (%)
      11.9
       Smoker
      During the observation period.
      (%)
      52.2
       Osteoporosis drug use
      During the observation period.
      (%)
      16.2
       Corticosteroid use
      During the observation period.
      (%)
      26.2
      Surgical (any type of primary spinal fusion)
       Number of levels fused3.4 (1.8)
       Short or Long fusion
      Short (≤ 3 levels fused; %)80.6
      Long (≥ 4 levels fused; %)19.4
      Post-Surgical
       Follow-up time (months; median, IQ range)
      All follow-up times started on the date of the patient's surgery. The observation period could not exceed 5 years but could be shorter if the patient entered the study within 5 years of the end date of the study or if the patient died or left the health system within the 5 years. Any reoperation or vertebral fracture had to occur within the patient's observation period.
      For observation (N=469 patients)44.4 (16.5–60.0)
      To reoperation (N=52 patients)14.5 (7.5–30.9)
      To vertebral fracture (N=36 patients)2.0 (0.3–29.4)
       Reoperation (%)11.1
       Vertebral fracture (%)7.7
      SD = standard deviation; IQ = interquartile (25–75%)
      low asterisk Measured at the nominal level (N=461 patients), which is the standard measurement site for assessment of osteoporosis (one vertebral level from T12–L3, preferably L1). Cut point for BMD-defined osteoporosis is 80 mg/cm3 for both sexes; and for fragile bone strength is 4,500 N for women and 6,000 N for men; patients have osteoporosis by BCT if either measurement is less than or equal to the respective cut point.
      Other bone conditions that could impact bone health were defined as: any history of rheumatoid arthritis, osteopetrosis, Paget's disease, hypophosphatasia, osteogenesis imperfecta; a diagnosis from 2 years before surgery through the end of follow-up of osteomalacia, multiple myeloma, malignant neoplasms of the spine, or use of tumor necrosis factor (TNF) inhibitor; or a diagnosis during the follow-up period of hypocalcemia or hypercalcemia.
      During the observation period.
      § All follow-up times started on the date of the patient's surgery. The observation period could not exceed 5 years but could be shorter if the patient entered the study within 5 years of the end date of the study or if the patient died or left the health system within the 5 years. Any reoperation or vertebral fracture had to occur within the patient's observation period.

      Biomechanical computed tomography analysis

      All BCT measurements were performed blinded to the clinical outcomes. As described elsewhere for spinal fusion patients [
      • Burch S
      • Feldstein M
      • Hoffmann PF
      • Keaveny TM.
      Prevalence of poor bone quality in women undergoing spinal fusion using biomechanical-CT analysis.
      ], BCT was performed at a centralized facility using the VirtuOst software system (version 2.3; O.N. Diagnostics, Berkeley, CA, USA) to measure vertebral trabecular BMD (in units of mg/cm3) and vertebral compressive strength (in units of newtons, N; Fig. 2). Because the bone quality at the UIV level is most relevant to surgical planning, measurements were made at the UIV level for all patients (N=469). We also report measurements for one “nominal” vertebral level (T12–L3, L1 being preferable), when available (for N=461 patients); these measurements are cleared by the FDA for identifying osteoporosis and assessing fracture risk [
      • Keaveny TM
      • Clarke BL
      • Cosman F
      • Orwoll ES
      • Siris ES
      • Khosla S
      • et al.
      Biomechanical Computed Tomography analysis (BCT) for clinical assessment of osteoporosis.
      ]. Using cut points previously established for the nominal level, we classified patients as having BMD-defined osteoporosis, BDO (BMD ≤ 80 mg/cm3) [
      American College of Radiology
      ACR–SPR–SSR practice parameter for the performance of musculoskeletal quantitative computed tomography (QCT).
      ] or fragile bone strength, FBS (vertebral strength ≤ 4,500 N for women or 6,500 N for men) [
      • Kopperdahl DL
      • Aspelund T
      • Hoffmann PF
      • Sigurdsson S
      • Siggeirsdottir K
      • Harris TB
      • et al.
      Assessment of incident spine and hip fractures in women and men using finite element analysis of CT scans.
      ,
      • Zysset P
      • Qin L
      • Lang T
      • Khosla S
      • Leslie WD
      • Shepherd JA
      • et al.
      Clinical use of quantitative computed tomography-based finite element analysis of the hip and spine in the management of osteoporosis in adults: The 2015 ISCD official positions-part II.
      ]. Using BCT, a patient is considered to have osteoporosis if they test positive for FBS and/or BDO (one or both) since both the FBS and BDO classifications independently predict fracture risk and both have established diagnostic cut points for identifying osteoporosis. [
      • Keaveny TM
      • Clarke BL
      • Cosman F
      • Orwoll ES
      • Siris ES
      • Khosla S
      • et al.
      Biomechanical Computed Tomography analysis (BCT) for clinical assessment of osteoporosis.
      ] The absence of osteoporosis is therefore signified by testing negative for FBS and testing negative for BDO. Because of variations in vertebral size along the spinal column, when classifying FBS at the UIV level we adjusted the measured strength at the UIV level by a level-specific scaling ratio (ranging from 1.7 at T4 to 0.8 at L5). Each ratio represented a mean strength for the L1 level to the non-L1 level, as measured separately in an independent cohort of 455 non-fusion subjects. This type of scaling approach has been previously validated for predicting incident vertebral fracture [
      • Allaire BT
      • Lu D
      • Johannesdottir F
      • Kopperdahl D
      • Keaveny TM
      • Jarraya M
      • et al.
      Prediction of incident vertebral fracture using CT-based finite element analysis.
      ,
      • Johannesdottir F
      • Allaire B
      • Kopperdahl DL
      • Keaveny TM
      • Sigurdsson S
      • Bredella MA
      • et al.
      Bone density and strength from thoracic and lumbar CT scans both predict incident vertebral fractures independently of fracture location.
      ]. We did not scale the BMD values, which are directly correlated with the apparent density of the trabecular bone [
      • Kaneko TS
      • Bell JS
      • Pejcic MR
      • Tehranzadeh J
      • Keyak JH.
      Mechanical properties, density and quantitative CT scan data of trabecular bone with and without metastases.
      ] and its volume fraction (1 - porosity) [
      • Zioupos P
      • Cook RB
      • Hutchinson JR.
      Some basic relationships between density values in cancellous and cortical bone.
      ]. Thus, values of BMD below the 80 mg/cm3 cut point denote clinical osteoporosis when measured at the nominal vertebral level [
      • Keaveny TM
      • Clarke BL
      • Cosman F
      • Orwoll ES
      • Siris ES
      • Khosla S
      • et al.
      Biomechanical Computed Tomography analysis (BCT) for clinical assessment of osteoporosis.
      ,
      American College of Radiology
      ACR–SPR–SSR practice parameter for the performance of musculoskeletal quantitative computed tomography (QCT).
      ] and highly porous trabecular bone when measured at any level.
      Fig 2
      Fig. 2BCT images for three subjects (age/sex), showing the nominal (left) and upper instrumented vertebral (UIV) levels (right). Images depict the finite element model loaded to failure (colored regions showed failed tissue), a mid-transverse section showing the region of interest (yellow ellipse) of the BMD measurement, and a mid-sagittal section showing the analyzed vertebral level highlighted in red. For each subject, the measured values of vertebral strength and vertebral trabecular BMD are shown (strength value is scaled for the UIV level); * denotes a positive test result (FBS or BDO) for that measurement.

      Statistical analysis

      For each clinical outcome (reoperation and vertebral fracture), we report the crude (unadjusted) and adjusted Cox proportional hazard ratios for the BCT classifications. Adjustments were made for covariates that had significant associations (p<.05) with the outcomes based on univariate hazard ratio analyses for the demographics, patient factors, comorbidities and surgical factors (Table 1). Follow-up time for each patient was the shorter of their observation period and any event time. In these analyses, our main BCT classification for osteoporosis was testing positive for FBS and/or BDO (one or both) at the UIV level. We also performed separate hazard ratio analyses to dissect out the positive-testing patients into two complementary subsets: a) those who tested positive concurrently for both FBS and BDO; and b) those who tested positive for only one of FBS or BDO but not both. For each analysis, the hazard ratio for testing positive was calculated relative to the reference of not having osteoporosis (negative for each of FBS and BDO). Thus, the hazard ratio is interpreted as the relative risk of reoperation (or vertebral fracture) when testing positive versus not having osteoporosis; to indicate a statistically significant level of increased risk compared to the reference, the lower 95% confidence interval needed to have a value ≥ 1.0. In addition, we performed Kaplan-Meier survival analyses corresponding to the hazard ratio analyses. The crude 5-year cumulative incidences of reoperation and vertebral fracture (separately) were calculated as one minus the Kaplan-Meier estimators (with right-censoring) and the log-rank test was used to compare survival curves.
      For basic-science insight and to facilitate comparisons with the literature, hazard ratios are also reported for the individual FBS and BDO classifications if used separately (FBS positive relative to FBS negative; BDO positive relative to BDO negative). For example, a BMD test has no information on bone strength and therefore would only use the BDO classification. In a series of secondary analyses, we also compared the adjusted hazard ratios for the UIV versus nominal levels (for the 461 patients with measurements at both levels) because the nominal level is currently the standard for assessing osteoporosis using BCT. In addition, for the UIV level, to complement the adjusted hazard ratio analyses, stratified analyses were used to measure the crude hazard ratio in sub-groups of patients aged 50 and older, and for “short” (≤ 3 levels fused) versus “long” (≥ 4 levels fused) fusion. All analyses were performed for the two sexes pooled. Data analysis was performed using the JMP Pro software (version 16.0.0, SAS Institute, Cary, NC).
      After our statistical analysis, one of the physician co-authors performed a post-hoc chart read on the subset of patients who correctly tested positive by BCT for reoperation. The objective was to determine if the documented clinical indication for reoperation for these true-positive patients was biomechanical in nature (eg, hardware loosening, pedicle perforation, adjacent level fracture, non-union, or degenerative adjacent level disease) or due to some other non-biomechanical cause. In addition, although we did not design this study to compare BCT against DXA, we realize that a limited comparison with any available DXA data might provide additional clinical context. Thus, for the 103 patients in our analysis cohort (28%) who also had DXA within 1 year before the surgery, we report prevalence rates for osteoporosis for DXA (BMD T-score at the hip or spine ≤ -2.5) and BCT (FBS and/or BDO). For each clinical outcome, we also directly compared the hazard ratios by including osteoporosis by BCT and by DXA as separate predictors in the same model.

      Results

      The cohort was 64% female, racially diverse with 25% Hispanic and 12% Black, 89% were aged 50 or older, 38% were obese (BMI ≥ 30 kg/m2), 37% had diabetes, and 52% were smokers (Table 1). Eighty-nine patients (19%) had long fusions (≥ 4 levels fused) and 180 patients (38%) had just one level fused; 391 patients (83%) had a lumbar fusion (UIV at L1 or below) and L4 was the most common UIV level (Fig. 3).
      Fig 3
      Fig. 3Distribution of the N=469 patients in the analysis cohort by the number of fused levels and the location of the UIV level.
      Median times for observation and to reoperation and vertebral fracture were 44.4 months, 14.5 months, and 2.0 months, respectively, and the overall rates for reoperation and vertebral fracture were 11.1% and 7.7%, respectively (Table 1). Univariate Cox proportional hazard ratio analysis indicated that none of the covariate characteristics were significantly associated with reoperation: sex (p=.86), age (p=.58), race (p=.86), bodyweight (p=.24), height (p=.64), BMI (p=.28), steroid use (p=.32), smoking (p=.22), diabetes (p=.22), obesity (p=.59), rheumatoid arthritis (p=.99), other bone conditions (p=.59), osteoporosis medication (p=.60); nor was the surgical occurrence of dural tear (p=.75), infection (p=.70), epidural hematoma (p=.87), or any reported surgical complication (p=.87). For vertebral fracture, obesity (p=.0036) and higher body mass index (p=.02) decreased risk, whereas having a long fusion (p<.0001) and a greater number of fused levels (p<.0001) increased risk. Other bone-related risk factors had no effect, including steroid use (p=.25), smoking (p=.37), diabetes (p=.41), rheumatoid arthritis (p=.96), nor did sex (p=.15), age (p=.48), bodyweight (p=.21), or height (p=.24). Osteoporosis medication (during the observation period) was strongly associated with vertebral fracture (p=.0001), presumably because patients with newly identified fractures were placed on osteoporosis treatment. Based on these results, adjustments for hazard ratios were made for both outcomes using age (a standard adjustment factor in bone studies), obesity (a stronger predictor than body mass index), and long fusion (a stronger predictor than number of levels fused).
      At the UIV level, patients testing positive for osteoporosis by BCT (FBS and/or BDO) were at statistically significant increased risks of reoperation and vertebral fracture, before and after adjustment for covariates, compared to patients without osteoporosis (Table 2). Hazard ratios for the different classifications had substantially overlapping confidence intervals, with or without adjustment. Overall, 25.8% of patients tested positive for osteoporosis (FBS and/or BDO). Of those, the numerically highest hazard ratios occurred for the 12.6% of patients who tested positive for both FBS and BDO concurrently. For those patients, compared to patients without osteoporosis, the adjusted hazard ratio was 3.7 (95% CI: 1.9–7.2) for reoperation and 5.0 (2.2–11.7) for vertebral fracture. For reoperation, the additional 13.2% of patients who tested positive for only one of either FBS or BDO (but not both) were not at significantly higher risk (1.6, 0.7–3.7). By contrast, for vertebral fracture, risk remained significantly higher for these patients (4.3 1.9–10.2) and was similar to the risk for all patients testing positive for osteoporosis (4.7, 2.2–9.7). While the hazard ratios (crude and adjusted) for reoperation did not differ between the FBS and BDO classifications when each was used alone, models containing both classifications together showed a significant hazard ratio for FBS (crude: 2.4, 1.2–4.9; adjusted: 2.4, 1.2–5.0) but not for BDO (crude: 1.3, 0.6–2.7; adjusted: 1.5, 0.7–3.3), demonstrating an advantage of FBS over BDO for identifying high-risk patients; similar trends occurred for vertebral fracture. Taken together, these results indicate that, compared to patients without osteoporosis, the 25.8% of patients who tested positive for osteoporosis were at 4.7-fold higher risk for vertebral fracture and, of those, the 12.6% of patients who tested positive concurrently for both FBS and BDO were at 3.7-fold higher risk for reoperation.
      Table. 2Cox proportional hazard ratio (HR, crude and adjusted, with 95% confidence intervals) for reoperation and vertebral fracture, and the proportion of patients testing positive for the BCT, for different BCT classifications. The hazard ratio denotes the relative risk of reoperation (or vertebral fracture) when testing positive by each classification compared to not having osteoporosis. All measurements were made at the UIV level. Data for N=469 patients having complete BCT data at the UIV level
      Classification
      FBS = fragile bone strength; BDO = BMD-defined osteoporosis. The standard clinical test for osteoporosis by BCT is to test positive for either FBS and/or BDO (one or both can be positive); the last two rows break down these positive testing patients into those testing positive concurrently for both FBS and BDO versus those testing positive for only one (but not both) condition. Testing negative concurrently for both FBS and BDO denotes not having osteoporosis, which was used as the reference for all classifications that considered both bone strength and BMD; for the classifications involving FBS and BDO when considered alone (first two rows), not having osteoporosis was denoted by testing negative for the respective test.
      Proportion test positive (%)Reoperation (52/469 = 11.1%)Vertebral fracture (36/469 = 7.7%)
      Test positive for:(%)Crude HRAdjusted HR
      Adjustments were made for age, obesity (BMI ≥ 30 kg/m²), and long fusion (4 or more levels fused); crude = unadjusted.
      Crude HRAdjusted HR
      Adjustments were made for age, obesity (BMI ≥ 30 kg/m²), and long fusion (4 or more levels fused); crude = unadjusted.
      FBS21.82.8 (1.6–4.9)3.0 (1.7–5.4)4.4 (2.3–8.4)3.7 (1.9–7.5)
      BDO16.62.3 (1.3–4.2)2.7 (1.4–5.0)3.3 (1.7–6.5)3.3 (1.6–6.9)
      FBS and/or BDO (one or both)25.82.3 (1.3–4.0)2.6 (1.4–4.7)4.9 (2.5–9.5)4.7 (2.2–9.7)
      FBS and BDO (both)12.63.3 (1.8–6.1)3.7 (1.9–7.2)4.9 (2.2–10.8)5.0 (2.2–11.7)
      FBS or BDO (one, not both)13.21.4 (0.6–3.3)1.6 (0.7–3.7)4.8 (2.2–10.6)4.3 (1.9–10.2)
      low asterisk FBS = fragile bone strength; BDO = BMD-defined osteoporosis. The standard clinical test for osteoporosis by BCT is to test positive for either FBS and/or BDO (one or both can be positive); the last two rows break down these positive testing patients into those testing positive concurrently for both FBS and BDO versus those testing positive for only one (but not both) condition. Testing negative concurrently for both FBS and BDO denotes not having osteoporosis, which was used as the reference for all classifications that considered both bone strength and BMD; for the classifications involving FBS and BDO when considered alone (first two rows), not having osteoporosis was denoted by testing negative for the respective test.
      Adjustments were made for age, obesity (BMI ≥ 30 kg/m²), and long fusion (4 or more levels fused); crude = unadjusted.
      Reflecting those findings, the Kaplan-Meier survival analysis (Fig. 4) indicated distinct event rates for patients testing positive by the different classifications. For reoperation, those who tested positive concurrently for both FBS and BDO separated out compared to those testing positive for either FBS or BDO but not both and for those without osteoporosis (p<.0004), the latter two classifications having similar failure profiles. For vertebral fracture, the response for those without osteoporosis was distinct from the two other classifications (p<.0001), which in turn had similar failure rates. Overall, compared to patients with no osteoporosis, the failure rates were most statistically distinct for patients who tested positive concurrently for both FBS and BDO when reoperation was the outcome (p<.0001), and for those tested positive for either FBS and/or BDO when vertebral fracture was the outcome (p<.0001).
      Fig 4
      Fig. 4Kaplan-Meier failure curves (1 – survival) for reoperation (top row) and vertebral fracture (bottom row), comparing the failure rates for different BCT classifications (N=469 patients, BCT at the UIV level). In each column, the reference for comparison is not having osteoporosis (red line, concurrent negative test for each of FBS and BDO, N=348). The left column compares those non-osteoporotic patients to two groups: those testing positive concurrently for both FBS and BDO (blue line, N=59), and those testing positive for only one of FBS or BDO (but not both; green line, N=62); the middle column compares against those testing positive for osteoporosis (test positive for FBS and/or BDO, gray line, N=59+62=121); and the right column compares only against those testing positive concurrently for both FBS and BDO (blue line, N=59). For each panel, the p-values are from the log-rank test, which detects differences in clinical failure rates between any of the classifications in the panel. Shaded regions represent ± 95% confidence intervals (not shown on the left-most column for clarity).
      For the N=461 patients having BCT measurements at both the UIV and nominal levels, risks of reoperation and vertebral fracture remained significantly higher when the BCT measurements were made instead at the nominal level and trends were all the same as for the UIV level (Table 3). For both clinical outcomes, the hazard ratios consistently trended higher when the BCT measurements were made at the UIV than nominal level.
      Table 3Cox proportional hazard ratio (HR, adjusted, with 95% confidence intervals) for reoperation and vertebral fracture, and the proportion of patients testing positive, for the BCT measurements made at the UIV versus nominal levels. Data for N=461 patients having complete BCT data at the UIV and nominal levels. See Table 2 for additional legends
      ClassificationProportion test positive (%)Reoperation HR
      Adjustments were made for age, obesity (BMI ≥ 30 kg/m²), and long fusion (4 or more levels fused)
      Vertebral fracture HR
      Adjustments were made for age, obesity (BMI ≥ 30 kg/m²), and long fusion (4 or more levels fused)
      UIVNominalUIVNominalUIVNominal
      Test positive for:levellevellevellevellevellevel
      FBS21.721.02.9 (1.6–5.2)2.1 (1.1–3.9)3.9 (1.9–8.1)2.6 (1.2–5.4)
      BDO16.715.42.5 (1.3–4.8)2.3 (1.2–4.5)3.6 (1.7–7.6)3.1 (1.4–6.7)
      FBS and/or BDO (one or both)25.824.72.5 (1.4–4.6)2.0 (1.1–3.7)5.0 (2.3–10.8)3.1 (1.5–6.7)
      FBS and BDO (both)12.611.73.5 (1.8–6.9)2.8 (1.4–5.7)5.4 (2.2–13.2)3.6 (1.5–8.5)
      FBS or BDO (one, not both)13.213.01.6 (0.7–3.8)1.4 (0.6–3.3)4.6 (1.9–11.3)2.7 (1.1–6.9)
      low asterisk Adjustments were made for age, obesity (BMI ≥ 30 kg/m²), and long fusion (4 or more levels fused)
      Risk also remained significantly higher and trends similar for various sub-groups of patients aged 50 and older. Stratified analyses indicated that no patients under age 50 tested positive with BCT. For patients aged 50 and over, the reoperation rate did not differ (p=.34) between patients who underwent a long (14.6%) vs. short (10.8%) fusion. For patients testing positive concurrently for both FBS and BDO, the crude hazard ratio for reoperation remained statistically significant for long (4.6, 1.3–15.8) and short (2.9, 1.3–6.1) fusion (Table 4). For vertebral fracture, the observed rate in patients aged 50 and older was almost six-fold higher (p<.0001) for those who underwent long (23.2%) versus short (3.9%) fusion, indicating a strong association between long fusion and vertebral fracture; even so, the crude hazard ratio for testing positive for osteoporosis (FBS and/or BDO) remained high for both long (4.6, 1.7–12.1) and short (6.4, 2.0–20.7) fusion (Table 4), consistent with the results for the adjusted hazard ratios for all patients (Table 2).
      Table 4Cox proportional hazard ratio (crude, with 95% confidence intervals) for reoperation and vertebral fracture, for stratified analyses of different patient sub-groups from the full analysis cohort (N=469 patients). All measurements were made at the UIV level. See Table 2 for additional legends
      ClassificationPatient sub-group
      All PatientsAge ≥ 50Age ≥ 50 &Age ≥ 50 &
      Short fusionLong fusion
      (N=469)(N=416)(N=334)(N=82)
      Reoperation
      FBS2.8 (1.6–4.9)2.8 (1.6–5.0)2.6 (1.3–5.0)3.4 (1.1–10.7)
      BDO2.3 (1.3–4.2)2.3 (1.3–4.2)2.0 (1.0–4.1)3.3 (1.1–10.5)
      FBS and/or BDO
      test positive for one or both.
      2.3 (1.3–4.0)2.3 (1.3–4.1)2.1 (1.1–4.1)3.0 (0.9–9.4)
      FBS and BDO
      test positive for both.
      3.3 (1.8–6.1)3.3 (1.7–6.2)2.9 (1.3–6.1)4.6 (1.3–15.8)
      FBS or BDO
      test positive for one but not both.
      1.4 (0.6–3.3)1.4 (0.6–3.3)1.4 (0.5–3.7)1.6 (0.3–8.3)
      Vertebral fracture
      FBS4.4 (2.3–8.4)5.0 (2.5–10.1)4.3 (1.5–12.9)4.1 (1.6–10.4)
      BDO3.3 (1.7–6.5)3.5 (1.8–7.1)3.9 (1.3–11.6)3.6 (1.4–8.9)
      FBS and/or BDO
      test positive for one or both.
      4.9 (2.5–9.5)5.9 (2.8–12.5)6.4 (2.0–20.7)4.6 (1.7–12.1)
      FBS and BDO
      test positive for both.
      4.9 (2.2–10.8)5.9 (2.5–14.0)5.6 (1.4–22.4)5.5 (1.8–16.5)
      FBS or BDO
      test positive for one but not both.
      4.8 (2.2–10.6)5.9 (2.5–13.8)7.1 (1.9–26.6)3.8 (1.2–12.0)
      low asterisk test positive for one or both.
      test positive for both.
      test positive for one but not both.
      In a post-hoc statistical analysis, for the sub-group of 103 patients with both BCT (UIV level) and DXA (hip/spine) data, prevalence of osteoporosis by DXA was over four-fold lower than by BCT. One hundred of those patients were aged 50 or older and 82 were women. Of all 103 patients, only 3/99 (3.0%) patients tested positive for osteoporosis by spinal DXA; 8/100 (8.0%) tested positive by hip DX, and 9/103 (8.7%) tested positive by hip/spine DXA; prevalence of osteoporosis by BCT (FBS and/or BDO) was over four-fold higher at 36.9%. All three of the spine-DXA positives tested positive by BCT, as did six of the eight hip-DXA positives. When osteoporosis by BCT and hip/spine DXA were both entered into the same hazard ratio analysis, despite the small sample size and wide confidence intervals the (crude) hazard ratio for vertebral fracture was statistically significant for BCT (7.3, 1.5–35.5) but not for DXA (1.7, 0.3–8.2), and likewise for reoperation: BCT (4.2, 1.0–17.4), DXA (0.9, 0.1–7.6); adjusted models were not run due to the small sample size.
      Post-hoc chart reads were performed on the 15 high-risk patients with concurrent FBS and BDO who correctly tested positive (true positives) for being at significantly higher risk of reoperation (11 women, ages 60–79; 4 men, ages 58–79). Those reads indicated that these patients had reoperation due to hardware pullout (N=1), pedicle perforation (N=1), adjacent level fracture (N=0), non-union (N=5), degenerative adjacent level disease (N=4, one of whom also had non-union), or “other” (N=5, 3 of whom had a prescheduled or unrelated reoperation at the cervical level). Excluding the three patients with prescheduled or unrelated reoperation, these data indicate that 10/12 (83%) of these true-positive patients underwent reoperation because of some biomechanically related issue.

      Discussion

      These results demonstrate that spinal fusion patients aged 50 or older who had osteoporosis by BCT were at elevated risk for vertebral fracture and reoperation, compared to patients without osteoporosis, regardless of whether the patient underwent long or short fusion or whether measurements were made at the UIV (preferable) or nominal levels. While BMD testing represents the clinical standard for identifying osteoporosis, BCT uses measurements of both BMD and bone strength. This expanded approach is consistent with the Surgeon General of the United States’ definition of osteoporosis as “a skeletal disorder characterized by compromised bone strength, predisposing to an increased risk of fracture” [
      U.S. Department of Health and Human Services
      Bone Health and Osteoporosis: A Report of the Surgeon General.
      ,
      NIH Consensus Development Panel
      Osteoporosis prevention, diagnosis, and therapy.
      ]. The BCT test has been validated previously for vertebral fracture risk assessment in the general population [
      • Melton LJ
      • Riggs BL
      • Keaveny TM
      • Achenbach SJ
      • Hoffmann PF
      • Camp JJ
      • et al.
      Structural determinants of vertebral fracture risk.
      ,
      • Melton LJ
      • Riggs BL
      • Keaveny TM
      • Achenbach SJ
      • Kopperdahl D
      • Camp JJ
      • et al.
      Relation of vertebral deformities to bone density, structure, and strength.
      ,
      • Wang X
      • Sanyal A
      • Cawthon PM
      • Palermo L
      • Jekir M
      • Christensen J
      • et al.
      Prediction of new clinical vertebral fractures in elderly men using finite element analysis of CT scans.
      ,
      • Anderson DE
      • Demissie S
      • Allaire BT
      • Bruno AG
      • Kopperdahl DL
      • Keaveny TM
      • et al.
      The associations between QCT-based vertebral bone measurements and prevalent vertebral fractures depend on the spinal locations of both bone measurement and fracture.
      ,
      • Kopperdahl DL
      • Aspelund T
      • Hoffmann PF
      • Sigurdsson S
      • Siggeirsdottir K
      • Harris TB
      • et al.
      Assessment of incident spine and hip fractures in women and men using finite element analysis of CT scans.
      ,
      • Allaire BT
      • Lu D
      • Johannesdottir F
      • Kopperdahl D
      • Keaveny TM
      • Jarraya M
      • et al.
      Prediction of incident vertebral fracture using CT-based finite element analysis.
      ,
      • Johannesdottir F
      • Allaire B
      • Kopperdahl DL
      • Keaveny TM
      • Sigurdsson S
      • Bredella MA
      • et al.
      Bone density and strength from thoracic and lumbar CT scans both predict incident vertebral fractures independently of fracture location.
      ]. The current study extends those results to primary spinal fusion patients. As in those previous studies, high-risk patients for vertebral fracture tested positive for FBS and/or BDO. However, for reoperation, only those patients who concurrently tested positive for both FBS and BDO were at significantly higher risk. Those testing positive for only one of FBS or BDO (but not both) were not at significantly increased risk of reoperation but were at increased risk of vertebral fracture. Assessing risk of reoperation pre-operatively is complex and involves many factors [
      • Veeravagu A
      • Li A
      • Swinney C
      • Tian L
      • Moraff A
      • Azad TD
      • et al.
      Predicting complication risk in spine surgery: a prospective analysis of a novel risk assessment tool.
      ,
      • Guppy KH
      • Royse KE
      • Norheim EP
      • Harris JE
      • Brara HS.
      PLF versus PLIF and the fate of L5-S1: Analysis of operative nonunion rates among 3065 patients with lumbar fusions from a regional spine registry.
      ,
      • Lee MJ
      • Cizik AM
      • Hamilton D
      • Chapman JR.
      Predicting medical complications after spine surgery: a validated model using a prospective surgical registry.
      ,
      • Ratliff JK
      • Balise R
      • Veeravagu A
      • Cole TS
      • Cheng I
      • Olshen RA
      • et al.
      Predicting occurrence of spine surgery complications using "big data" modeling of an administrative claims database.
      ,
      • Sato S
      • Yagi M
      • Machida M
      • Yasuda A
      • Konomi T
      • Miyake A
      • et al.
      Reoperation rate and risk factors of elective spinal surgery for degenerative spondylolisthesis: minimum 5-year follow-up.
      ,
      • Yagi M
      • Hosogane N
      • Fujita N
      • Okada E
      • Suzuki S
      • Tsuji O
      • et al.
      The patient demographics, radiographic index and surgical invasiveness for mechanical failure (PRISM) model established for adult spinal deformity surgery.
      ,
      • Akins PT
      • Harris J
      • Alvarez JL
      • Chen Y
      • Paxton EW
      • Bernbeck J
      • et al.
      Risk factors associated with 30-day readmissions after instrumented spine surgery in 14,939 patients: 30-day readmissions after instrumented spine surgery.
      ]. To date, such clinical risk assessments have mostly omitted osteoporosis as a risk factor, due in part to the technical limitations [
      • Tenne M
      • McGuigan F
      • Besjakov J
      • Gerdhem P
      • Akesson K.
      Degenerative changes at the lumbar spine–implications for bone mineral density measurement in elderly women.
      ,
      • Schneider P
      • Borner W.
      The impact of degenerative spinal changes on the correlation of peripheral and axial bone density.
      ,
      • Guglielmi G
      • Floriani I
      • Torri V
      • Li J
      • van Kuijk C
      • Genant HK
      • et al.
      Effect of spinal degenerative changes on volumetric bone mineral density of the central skeleton as measured by quantitative computed tomography.
      ,
      • Li N
      • Li XM
      • Xu L
      • Sun WJ
      • Cheng XG
      • Tian W.
      Comparison of QCT and DXA: osteoporosis detection rates in postmenopausal women.
      ,
      • St Jeor JD
      • Jackson TJ
      • Xiong AE
      • Kadri A
      • Freedman BA
      • Sebastian AS
      • et al.
      Osteoporosis in spine surgery patients: what is the best way to diagnose osteoporosis in this population?.
      ] of using DXA to reliably assess spinal osteoporosis in fusion patients. Our results indicate that preoperative BCT, utilizing an existing CT scan, can provide a reliable means of incorporating osteoporosis as a clinical risk factor into surgical planning and management of spinal fusion patients.
      Together with the literature, our results demonstrate that for spinal fusion patients, more cases of clinically-relevant osteoporosis at the spine can be detected by BCT than by BMD alone, regardless of whether BMD is measured by quantitative CT or DXA. In our limited group of 103 patients with DXA, the prevalence of osteoporosis by DXA (at the hip or spine) was 8.7%, which was over four-fold lower than the 36.9% prevalence of osteoporosis by BCT for that group. Recognizing the small size of that group, this DXA rate for patients encountered from 2005 to 2018 in our study is consistent with rates for the general US population of adults aged 50 years and older of 9% from 2005 to 2008 [
      • Looker AC
      • Borrud LG
      • Dawson-Hughes B
      • Shepherd JA
      • Wright NC.
      Osteoporosis or low bone mass at the femur neck or lumbar spine in older adults: United States, 2005-2008.
      ] and 10.3% in 2010 [
      • Wright NC
      • Looker AC
      • Saag KG
      • Curtis JR
      • Delzell ES
      • Randall S
      • et al.
      The recent prevalence of osteoporosis and low bone mass in the United States based on bone mineral density at the femoral neck or lumbar spine.
      ] and a rate 10.0% for 140 consecutive spinal fusion patients at a single medical center from 2007–2014 [
      • Bjerke BT
      • Zarrabian M
      • Aleem IS
      • Fogelson JL
      • Currier BL
      • Freedman BA
      • et al.
      Incidence of osteoporosis-related complications following posterior lumbar fusion.
      ]. Clinically with CT, the nominal level is usually used for diagnosing osteoporosis whether by BMD [
      American College of Radiology
      ACR–SPR–SSR practice parameter for the performance of musculoskeletal quantitative computed tomography (QCT).
      ] or BCT [
      • Keaveny TM
      • Clarke BL
      • Cosman F
      • Orwoll ES
      • Siris ES
      • Khosla S
      • et al.
      Biomechanical Computed Tomography analysis (BCT) for clinical assessment of osteoporosis.
      ] testing. In the present study at the nominal level, the prevalence of osteoporosis by CT-based BMD testing alone (BDO) versus BCT (FBS and/or BDO) was 15.4 % vs. 24.7% (Table 3), consistent with respective rates of 14.3% and 28.6% in our earlier BCT study of 98 female fusion patients aged 50–70 at a major academic center from 2003 to 2012 [
      • Burch S
      • Feldstein M
      • Hoffmann PF
      • Keaveny TM.
      Prevalence of poor bone quality in women undergoing spinal fusion using biomechanical-CT analysis.
      ]. This collective evidence suggests that at least 60% more spinal fusion patients with spinal osteoporosis can be identified by BCT testing (FBS and/or BDO) compared to CT-based BMD testing (BDO), and 2–4 times more compared to DXA testing. As evidenced by the equivalence of the hazard ratios, this greater number of patients identified as having osteoporosis by BCT also appear to be at an equivalent level of elevated risk as the smaller number of patients with osteoporosis by CT-based BMD. Further, in our ad hoc hazard ratio models that contained both BCT and DXA for the 103 patients with both DXA and BCT data, higher hazard ratios for fracture and reoperation were higher by BCT than DXA. Despite the small sample size and wide confidence intervals for that ad hoc analysis, it nevertheless indicates that BCT better identified spinal fusion patients at higher risk of vertebral fracture and reoperation than did DXA.
      One noteworthy advance of BCT over CT-based BMD testing alone is that risk of reoperation was significantly elevated only when patients tested positive concurrently by both BMD and bone strength. Related, models that contained both the FBS and BDO as independent classifiers showed a statistical advantage of FBS over BDO. Both these findings demonstrate an added prognostic value of considering both bone strength and BMD rather than just BMD alone. Mechanistically, in fusion patients some vertebrae can contain highly porous trabecular bone (test positive by the BDO classification) but can also have thickened cortices or endplates or other features that can strengthen the overall vertebra despite that porosity (test negative by the FBS classification). Our findings suggest that such strengthening can uniquely protect against risk of reoperation, presumably by mitigating certain failure mechanisms not directly depending on high trabecular porosity, such as endplate failure and cage subsidence. By contrast, for vertebral fracture, patients with either weak or porous vertebrae were at high risk of vertebral fracture. That result might reflect the more systemic nature of vertebral fractures, which can occur anywhere along the spinal column, versus the more local nature of reoperation, which is likely more related to problems at the UIV level and thus the specifics of both BMD and bone strength at that or an adjacent level.
      As with previous studies on fusion patients that showed a higher risk of reoperation in the presence of osteoporosis [
      • Khalid SI
      • Nunna RS
      • Maasarani S
      • Belmont E
      • Deme P
      • Chilakapati S
      • et al.
      Association of osteopenia and osteoporosis with higher rates of pseudarthrosis and revision surgery in adult patients undergoing single-level lumbar fusion.
      ,
      • Gupta A
      • Cha T
      • Schwab J
      • Fogel H
      • Tobert D
      • Razi AE
      • et al.
      Osteoporosis increases the likelihood of revision surgery following a long spinal fusion for adult spinal deformity.
      ], our study was not designed to address if osteoporosis was the cause of reoperation. Reoperation has many causes [
      • Veeravagu A
      • Li A
      • Swinney C
      • Tian L
      • Moraff A
      • Azad TD
      • et al.
      Predicting complication risk in spine surgery: a prospective analysis of a novel risk assessment tool.
      ,
      • Guppy KH
      • Royse KE
      • Norheim EP
      • Harris JE
      • Brara HS.
      PLF versus PLIF and the fate of L5-S1: Analysis of operative nonunion rates among 3065 patients with lumbar fusions from a regional spine registry.
      ,
      • Lee MJ
      • Cizik AM
      • Hamilton D
      • Chapman JR.
      Predicting medical complications after spine surgery: a validated model using a prospective surgical registry.
      ,
      • Ratliff JK
      • Balise R
      • Veeravagu A
      • Cole TS
      • Cheng I
      • Olshen RA
      • et al.
      Predicting occurrence of spine surgery complications using "big data" modeling of an administrative claims database.
      ,
      • Sato S
      • Yagi M
      • Machida M
      • Yasuda A
      • Konomi T
      • Miyake A
      • et al.
      Reoperation rate and risk factors of elective spinal surgery for degenerative spondylolisthesis: minimum 5-year follow-up.
      ,
      • Yagi M
      • Hosogane N
      • Fujita N
      • Okada E
      • Suzuki S
      • Tsuji O
      • et al.
      The patient demographics, radiographic index and surgical invasiveness for mechanical failure (PRISM) model established for adult spinal deformity surgery.
      ,
      • Akins PT
      • Harris J
      • Alvarez JL
      • Chen Y
      • Paxton EW
      • Bernbeck J
      • et al.
      Risk factors associated with 30-day readmissions after instrumented spine surgery in 14,939 patients: 30-day readmissions after instrumented spine surgery.
      ], and in our study, for example, 58% of the 52 reoperation patients did not have osteoporosis by BCT. However, that osteoporosis is a causative factor is mechanistically feasible since low bone density weakens the bone-screw interface [
      • Okuyama K
      • Abe E
      • Suzuki T
      • Tamura Y
      • Chiba M
      • Sato K
      Influence of bone mineral density on pedicle screw fixation: a study of pedicle screw fixation augmenting posterior lumbar interbody fusion in elderly patients.
      ] and has been linked clinically with proximal junctional kyphosis, cage subsidence, screw loosening, and pseudoarthrosis [
      • Khalid SI
      • Nunna RS
      • Maasarani S
      • Belmont E
      • Deme P
      • Chilakapati S
      • et al.
      Association of osteopenia and osteoporosis with higher rates of pseudarthrosis and revision surgery in adult patients undergoing single-level lumbar fusion.
      ,
      • Duan PG
      • Mummaneni PV
      • Rivera J
      • Guinn JMV
      • Wang M
      • Xi Z
      • et al.
      The association between lower Hounsfield units of the upper instrumented vertebra and proximal junctional kyphosis in adult spinal deformity surgery with a minimum 2-year follow-up.
      ,
      • Kim JS
      • Phan K
      • Cheung ZB
      • Lee N
      • Vargas L
      • Arvind V
      • et al.
      Surgical, radiographic, and patient-related risk factors for proximal junctional kyphosis: a meta-analysis.
      ,
      • Bredow J
      • Boese CK
      • Werner CM
      • Siewe J
      • Löhrer L
      • Zarghooni K
      • et al.
      Predictive validity of preoperative CT scans and the risk of pedicle screw loosening in spinal surgery.
      ,
      • Zou D
      • Muheremu A
      • Sun Z
      • Zhong W
      • Jiang S
      • Li W.
      Computed tomography Hounsfield unit-based prediction of pedicle screw loosening after surgery for degenerative lumbar spine disease.
      ,
      • Löffler MT
      • Sollmann N
      • Burian E
      • Bayat A
      • Aftahy K
      • Baum T
      • et al.
      Opportunistic osteoporosis screening reveals low bone density in patients with screw loosening after lumbar semi-rigid instrumentation: a case-control study.
      ,
      • Schwaiger BJ
      • Gersing AS
      • Baum T
      • Noël PB
      • Zimmer C
      • Bauer JS.
      Bone mineral density values derived from routine lumbar spine multidetector row CT predict osteoporotic vertebral fractures and screw loosening.
      ,
      • Liu Y
      • Dash A
      • Krez A
      • Kim HJ
      • Cunningham M
      • Schwab F
      • et al.
      Low volumetric bone density is a risk factor for early complications after spine fusion surgery.
      ,
      • Xi Z
      • Mummaneni PV
      • Wang M
      • Ruan H
      • Burch S
      • Deviren V
      • et al.
      The association between lower Hounsfield units on computed tomography and cage subsidence after lateral lumbar interbody fusion.
      ]. Beyond the direct effects of osteoporosis in weakening bone, a spinal fusion operation can also accelerate subsequent decreases in BMD over time in adjacent vertebral levels, further increasing risk of vertebral failure [
      • Okano I
      • Jones C
      • Salzmann SN
      • Miller CO
      • Shirahata T
      • Rentenberger C
      • et al.
      Postoperative decrease of regional volumetric bone mineral density measured by quantitative computed tomography after lumbar fusion surgery in adjacent vertebrae.
      ,
      • Okano I
      • Jones C
      • Rentenberger C
      • Sax OC
      • Salzmann SN
      • Reisener MJ
      • et al.
      The association between endplate changes and risk for early severe cage subsidence among standalone lateral lumbar interbody fusion patients.
      ,
      • Jones C
      • Okano I
      • Salzmann SN
      • Reisener MJ
      • Chiapparelli E
      • Shue J
      • et al.
      Endplate volumetric bone mineral density is a predictor for cage subsidence following lateral lumbar interbody fusion: a risk factor analysis.
      ]. And there is emerging evidence that elements of bone fragility and osteoclastic stimulation [
      • Fields AJ
      • Ballatori A
      • Liebenberg EC
      • Lotz JC.
      Contribution of the endplates to disc degeneration.
      ,
      • Wang Y
      • Videman T
      • Battié MC.
      ISSLS prize winner: Lumbar vertebral endplate lesions: associations with disc degeneration and back pain history.
      ,
      • Feng Z
      • Liu Y
      • Yang G
      • Battié MC
      • Wang Y.
      Lumbar vertebral endplate defects on magnetic resonance images: classification, distribution patterns, and associations with modic changes and disc degeneration.
      ,
      • Sun Z
      • Zheng X
      • Li S
      • Zeng B
      • Yang J
      • Ling Z
      • et al.
      Single impact injury of vertebral endplates without structural disruption, initiates disc degeneration through Piezo1 mediated inflammation and metabolism dysfunction.
      ,
      • Okano I
      • Salzmann SN
      • Jones C
      • Ortiz Miller C
      • Shirahata T
      • Rentenberger C
      • et al.
      The impact of degenerative disc disease on regional volumetric bone mineral density (vBMD) measured by quantitative computed tomography.
      ] are linked to the etiology of disc degeneration, suggesting that osteoporosis may also contribute to or be coincident with disc degeneration. Consistent with this overall experience, our post-hoc chart reads showed that, excluding three patients who had pre-scheduled or unrelated second surgeries, 10/12 (83%) of the true-positive patients underwent reoperation because of biomechanically-related issues. That osteoporosis is causative of reoperation is also consistent with results from a large claims-based analysis, which found an association between therapeutic treatment of osteoporosis and reduced rates of reoperation [
      • Jain N
      • Labaran L
      • Phillips FM
      • Khan SN
      • Jain A
      • Kebaish KM
      • et al.
      Prevalence of osteoporosis treatment and its effect on post-operative complications, revision surgery and costs after multi-level spinal fusion.
      ].
      Our study has a number of strengths. We used only data from a non-academic clinical setting, including BCT measurements and classifications at the nominal level that are FDA-cleared for clinical diagnostic use and fracture risk assessment [
      • Keaveny TM
      • Clarke BL
      • Cosman F
      • Orwoll ES
      • Siris ES
      • Khosla S
      • et al.
      Biomechanical Computed Tomography analysis (BCT) for clinical assessment of osteoporosis.
      ]. The study included eight large medical centers, involved over 20 different CT scanners, and sampled from a racially diverse patient population, including all adults and both sexes. We included all types of primary fusion procedures regardless of the approach (eg, ALIF, TLIF, PLIF) or number of fused levels. Our use of a random sample of existing data from all available patients, pending availability of their data, minimized any selection bias associated with surgeon or patient participation and resulted in similar overall reoperation rates for the analysis cohort (11.1%) and full eligible population (11.9%). Our primary outcome was reoperation, which is the most important outcome in terms of clinical relevance but is not often addressed due to sample size constraints. Our median observation time of 44 months was much longer than the median 14 months to reoperation and 2 months to vertebral fracture, indicating that our observation time was sufficiently large to capture most reoperation (and vertebral fracture) cases. Our Cox proportional hazard ratio analyses accounted for the variable follow-up and event times across patients. Those analyses also demonstrated increased risk for patients who tested positive by BCT independently of age, obesity, and long vs. short fusion, which was further supported by our stratified analyses.
      Despite these strengths, some elements of our study design warrant discussion. The first issue is whether the nature of our study design might have compromised the internal validity of the analysis by introducing some type of bias that would have altered the ability of BCT to identify high-risk patients compared to what would be expected clinically. Our main inclusion criterion required patients to have had a spine-containing CT as part of their medical care before surgery, which in this health care system comprised of approximately 20% of all primary fusion patients. For approximately 85% of those included patients, that CT comprised a lumbar or thoracic spine CT. In our clinical practice, there is no link between any indication for ordering such a spine CT pre-operatively and the likelihood of having osteoporosis. Thus, it is unlikely that this inclusion criterion introduced significant bias into the analysis. In addition, to minimize effects of any short-term post-operative mortality (or other loss to follow-up) on our results, we only included patients without reoperation or vertebral fracture if they had at least 3 months of follow up. However, only 4.4% of all fusion patients did not meet this criterion and again, any links to the presence of osteoporosis are tenuous.
      We were also forced to exclude 65% of the 1,348 potentially eligible patients in our random sample due to missing or inappropriate data. Compared to that random sample, the final analysis cohort of 469 patients had significantly more women (64% vs 52%), older age (63.6 vs 56.6 years), shorter height (167 vs 170 cm), higher BMI (29.2 vs 28.3 kg/m2), and more diabetes (37% vs 29%); there were also fewer long fusions (19% vs 34%) and ultimately fewer vertebral fractures (8% vs 13%). A lack of preoperative CT scans that extended up the spine to the UIV level might explain the observed shorter fusion length in the analysis cohort than the overall random cohort, and thus the lower observed fracture rate. However, our adjusted and stratified hazard ratio analyses demonstrated that osteoporosis by BCT was associated with vertebral fracture for both short and long fusions, indicating that any such differences in the cohort regarding fusion length should not compromise the ability of BCT to assess fracture risk. Reoperation rate did not differ between the random and analysis groups (p=.10), and none of the above factors affected reoperation rate. Thus, while our analysis cohort differed slightly from the collective group of all primary fusion patients with CT in our system, any different characteristics were either accounted for in our statistical analysis or did not affect the clinical outcomes.
      From a statistical perspective, while a fully prospective study design would have resulted in more control over data collection and more fine-grained detail on surgical approaches and clinical outcomes, our analysis cohort did have a complete set of study data and the sample size was sufficiently large to support the statistical validity of our study conclusions. That said, we made multiple statistical comparisons but did not adjust p-values; some of the reported hazard ratios had 95% confidence intervals that were wide or close to unity; and the overall statistics depended on a relatively small number of events. Further, in retrospect we found that some pre-planned second surgeries were included, which could have diluted statistical power. Thus, replication of our findings in a larger study would further support generality.
      One limitation relates to the nature of our clinical outcomes. For our reoperation outcome, we did not distinguish between different causes for reoperation and thus it is not possible to use our current results to predict risk of any particular type of bone-related failure, for example screw loosening, cage subsidence, or proximal junctional kyphosis. That issue remains a topic of ongoing research. Some have shown, for example, that a localized measurement of BMD around the endplate is associated with cage subsidence [
      • Jones C
      • Okano I
      • Salzmann SN
      • Reisener MJ
      • Chiapparelli E
      • Shue J
      • et al.
      Endplate volumetric bone mineral density is a predictor for cage subsidence following lateral lumbar interbody fusion: a risk factor analysis.
      ,
      • Okano I
      • Jones C
      • Salzmann SN
      • Reisener MJ
      • Sax OC
      • Rentenberger C
      • et al.
      Endplate volumetric bone mineral density measured by quantitative computed tomography as a novel predictive measure of severe cage subsidence after standalone lateral lumbar fusion.
      ]. Going forward, challenges in conducting clinical studies with granular clinical outcomes include precisely defining such outcomes [
      • Bjerke BT
      • Zarrabian M
      • Aleem IS
      • Fogelson JL
      • Currier BL
      • Freedman BA
      • et al.
      Incidence of osteoporosis-related complications following posterior lumbar fusion.
      ] while ensuring a sufficiently large sample size in order to provide the required statistical power to differentiate between prediction of different modes of clinical failure. For our vertebral fracture outcome, we had no record of the vertebral level of any new fracture, nor could we distinguish between a new fracture after surgery and one that might have existed before surgery but was only noticed and entered into the patient's record after surgical follow-up. The short median time to vertebral fracture of 2.0 months might therefore reflect that many new vertebral fractures occurred shortly after surgery, or, that many fractures existed before surgery but were only noted in the medical record during the first post-surgery visit. Given the consistency across findings from our previous studies in which BCT predicted both existing and new vertebral fractures in the general population [
      • Melton LJ
      • Riggs BL
      • Keaveny TM
      • Achenbach SJ
      • Hoffmann PF
      • Camp JJ
      • et al.
      Structural determinants of vertebral fracture risk.
      ,
      • Melton LJ
      • Riggs BL
      • Keaveny TM
      • Achenbach SJ
      • Kopperdahl D
      • Camp JJ
      • et al.
      Relation of vertebral deformities to bone density, structure, and strength.
      ,
      • Wang X
      • Sanyal A
      • Cawthon PM
      • Palermo L
      • Jekir M
      • Christensen J
      • et al.
      Prediction of new clinical vertebral fractures in elderly men using finite element analysis of CT scans.
      ,
      • Anderson DE
      • Demissie S
      • Allaire BT
      • Bruno AG
      • Kopperdahl DL
      • Keaveny TM
      • et al.
      The associations between QCT-based vertebral bone measurements and prevalent vertebral fractures depend on the spinal locations of both bone measurement and fracture.
      ,
      • Kopperdahl DL
      • Aspelund T
      • Hoffmann PF
      • Sigurdsson S
      • Siggeirsdottir K
      • Harris TB
      • et al.
      Assessment of incident spine and hip fractures in women and men using finite element analysis of CT scans.
      ,
      • Allaire BT
      • Lu D
      • Johannesdottir F
      • Kopperdahl D
      • Keaveny TM
      • Jarraya M
      • et al.
      Prediction of incident vertebral fracture using CT-based finite element analysis.
      ,
      • Johannesdottir F
      • Allaire B
      • Kopperdahl DL
      • Keaveny TM
      • Sigurdsson S
      • Bredella MA
      • et al.
      Bone density and strength from thoracic and lumbar CT scans both predict incident vertebral fractures independently of fracture location.
      ], it is unlikely that our conclusions regarding risk of vertebral fracture would have changed appreciably had our outcome comprised of only radiographically confirmed new fractures.
      Regarding generality, one caveat is that our results might apply only to the types of patients included in our cohort, namely only to primary fusion patients who had a UIV level at or below T4. In addition, our patients, being specific to one region of the U.S. and one health care system, may have specific characteristics that differ from patients elsewhere. Likewise, surgical practices may vary across different health care systems. While the diverse nature of our cohort, the multiple medical centers, the multi-year time span over which patients were sampled, and the strong statistical significance of the Kaplan-Meier analyses together support the generality of our main conclusions, specific metrics such as revision rates and prevalence of osteoporosis can vary geographically and thus may not apply broadly. And because we did not account for factors related to the type of surgical approach, beyond our stratification of long and short fusions, our results best apply to primary fusion patients in general and do not address any specific sub-categories. Future studies are therefore needed to integrate the BCT metrics reported here into other preoperative risk assessments [
      • Veeravagu A
      • Li A
      • Swinney C
      • Tian L
      • Moraff A
      • Azad TD
      • et al.
      Predicting complication risk in spine surgery: a prospective analysis of a novel risk assessment tool.
      ,
      • Guppy KH
      • Royse KE
      • Norheim EP
      • Harris JE
      • Brara HS.
      PLF versus PLIF and the fate of L5-S1: Analysis of operative nonunion rates among 3065 patients with lumbar fusions from a regional spine registry.
      ,
      • Lee MJ
      • Cizik AM
      • Hamilton D
      • Chapman JR.
      Predicting medical complications after spine surgery: a validated model using a prospective surgical registry.
      ,
      • Ratliff JK
      • Balise R
      • Veeravagu A
      • Cole TS
      • Cheng I
      • Olshen RA
      • et al.
      Predicting occurrence of spine surgery complications using "big data" modeling of an administrative claims database.
      ,
      • Sato S
      • Yagi M
      • Machida M
      • Yasuda A
      • Konomi T
      • Miyake A
      • et al.
      Reoperation rate and risk factors of elective spinal surgery for degenerative spondylolisthesis: minimum 5-year follow-up.
      ,
      • Yagi M
      • Hosogane N
      • Fujita N
      • Okada E
      • Suzuki S
      • Tsuji O
      • et al.
      The patient demographics, radiographic index and surgical invasiveness for mechanical failure (PRISM) model established for adult spinal deformity surgery.
      ,
      • Akins PT
      • Harris J
      • Alvarez JL
      • Chen Y
      • Paxton EW
      • Bernbeck J
      • et al.
      Risk factors associated with 30-day readmissions after instrumented spine surgery in 14,939 patients: 30-day readmissions after instrumented spine surgery.
      ] for specific patient and surgical categories.
      Notwithstanding these limitations and caveats, our results have implications for clinical care. Given the mechanistic nature of BCT, and its prior validation for assessing vertebral fracture risk in multiple cohorts [
      • Melton LJ
      • Riggs BL
      • Keaveny TM
      • Achenbach SJ
      • Hoffmann PF
      • Camp JJ
      • et al.
      Structural determinants of vertebral fracture risk.
      ,
      • Melton LJ
      • Riggs BL
      • Keaveny TM
      • Achenbach SJ
      • Kopperdahl D
      • Camp JJ
      • et al.
      Relation of vertebral deformities to bone density, structure, and strength.
      ,
      • Wang X
      • Sanyal A
      • Cawthon PM
      • Palermo L
      • Jekir M
      • Christensen J
      • et al.
      Prediction of new clinical vertebral fractures in elderly men using finite element analysis of CT scans.
      ,
      • Anderson DE
      • Demissie S
      • Allaire BT
      • Bruno AG
      • Kopperdahl DL
      • Keaveny TM
      • et al.
      The associations between QCT-based vertebral bone measurements and prevalent vertebral fractures depend on the spinal locations of both bone measurement and fracture.
      ,
      • Kopperdahl DL
      • Aspelund T
      • Hoffmann PF
      • Sigurdsson S
      • Siggeirsdottir K
      • Harris TB
      • et al.
      Assessment of incident spine and hip fractures in women and men using finite element analysis of CT scans.
      ,
      • Allaire BT
      • Lu D
      • Johannesdottir F
      • Kopperdahl D
      • Keaveny TM
      • Jarraya M
      • et al.
      Prediction of incident vertebral fracture using CT-based finite element analysis.
      ,
      • Johannesdottir F
      • Allaire B
      • Kopperdahl DL
      • Keaveny TM
      • Sigurdsson S
      • Bredella MA
      • et al.
      Bone density and strength from thoracic and lumbar CT scans both predict incident vertebral fractures independently of fracture location.
      ], our findings suggest that therapeutic treatment [
      • Sardar ZM
      • Coury JR
      • Cerpa M
      • DeWald CJ
      • Ames CP
      • Shuhart C
      • et al.
      Best practice guidelines for assessment & management of osteoporosis in adult patients undergoing elective spinal reconstruction.
      ] should be considered for all spinal fusion patients age 50 and older who test positive by BCT for either fragile bone strength or BMD-defined osteoporosis, whether at the nominal or UIV levels. Such treatment should at least reduce the risk of any future vertebral fracture. Additional mitigating steps might be justified for some of these osteoporotic patients, such as patients undergoing a long fusion, who are at particularly high risk for vertebral fracture, or those testing positive concurrently for both FBS and BDO, who are at almost four-fold higher risk for reoperation compared to patients without osteoporosis. While BCT provides an objective basis to identify such high-risk patients, a determination of what particular mitigating steps might be most appropriate for any particular patient remains an issue of clinical judgement for the surgeon and an important topic of ongoing research.

      Conflicts of interest

      1. TMK: Consulting; Amgen; Bone Health Technologies; O.N. Diagnostics; UCB Pharma. Equity: O.N. Diagnostics.O.N. Diagnostics LLC is the developer and provider of the FDA-cleared BCT test that was utilized in this study. Technicians at O.N. Diagnostics performed all BCT analyses for this study, blinded to the clinical outcomes. 2. ALA: None. 3. HF: None. 4. HSB: None. 5. SB: None. 6. KHG: None. 7. DLK: Employee of O.N. Diagnostics, LLC

      Acknowledgments

      This research was supported by the National Institute of Arthritis and Musculoskeletal and Skin Diseases of the National Institutes of Health under award number R44AR064613 . Authors #1, 4, 5, 6 received salary support from their respective departments.

      References

        • Shirley M
        • Wanderman N
        • Keaveny T
        • Anderson P
        • Freedman BA
        Opportunistic computed tomography and spine surgery: a narrative review.
        Global Spine J. 2020; 10: 919-928
        • Bjerke BT
        • Zarrabian M
        • Aleem IS
        • Fogelson JL
        • Currier BL
        • Freedman BA
        • et al.
        Incidence of osteoporosis-related complications following posterior lumbar fusion.
        Global Spine J. 2018; 8: 563-569
        • Khalid SI
        • Nunna RS
        • Maasarani S
        • Belmont E
        • Deme P
        • Chilakapati S
        • et al.
        Association of osteopenia and osteoporosis with higher rates of pseudarthrosis and revision surgery in adult patients undergoing single-level lumbar fusion.
        Neurosurg Focus. 2020; 49: E6
        • Jain N
        • Labaran L
        • Phillips FM
        • Khan SN
        • Jain A
        • Kebaish KM
        • et al.
        Prevalence of osteoporosis treatment and its effect on post-operative complications, revision surgery and costs after multi-level spinal fusion.
        Global Spine J. 2022; 12: 1119-1124
        • Yagi M
        • Fujita N
        • Tsuji O
        • Nagoshi N
        • Asazuma T
        • Ishii K
        • et al.
        Low bone-mineral density is a significant risk for proximal junctional failure after surgical correction of adult spinal deformity: a propensity score-matched analysis.
        Spine (Phila Pa 1976). 2018; 43: 485-491
        • Duan PG
        • Mummaneni PV
        • Rivera J
        • Guinn JMV
        • Wang M
        • Xi Z
        • et al.
        The association between lower Hounsfield units of the upper instrumented vertebra and proximal junctional kyphosis in adult spinal deformity surgery with a minimum 2-year follow-up.
        Neurosurg Focus. 2020; 49: E7
        • Kim JS
        • Phan K
        • Cheung ZB
        • Lee N
        • Vargas L
        • Arvind V
        • et al.
        Surgical, radiographic, and patient-related risk factors for proximal junctional kyphosis: a meta-analysis.
        Global Spine J. 2019; 9: 32-40
        • Bredow J
        • Boese CK
        • Werner CM
        • Siewe J
        • Löhrer L
        • Zarghooni K
        • et al.
        Predictive validity of preoperative CT scans and the risk of pedicle screw loosening in spinal surgery.
        Arch Orthop Trauma Surg. 2016; 136: 1063-1067
        • Zou D
        • Muheremu A
        • Sun Z
        • Zhong W
        • Jiang S
        • Li W.
        Computed tomography Hounsfield unit-based prediction of pedicle screw loosening after surgery for degenerative lumbar spine disease.
        J Neurosurg Spine. 2020; 32: 716-721
        • Gupta A
        • Cha T
        • Schwab J
        • Fogel H
        • Tobert D
        • Razi AE
        • et al.
        Osteoporosis increases the likelihood of revision surgery following a long spinal fusion for adult spinal deformity.
        Spine J. 2021; 21: 134-140
        • Okuyama K
        • Abe E
        • Suzuki T
        • Tamura Y
        • Chiba M
        • Sato K
        Influence of bone mineral density on pedicle screw fixation: a study of pedicle screw fixation augmenting posterior lumbar interbody fusion in elderly patients.
        Spine J. 2001; 1: 402-407
        • Löffler MT
        • Sollmann N
        • Burian E
        • Bayat A
        • Aftahy K
        • Baum T
        • et al.
        Opportunistic osteoporosis screening reveals low bone density in patients with screw loosening after lumbar semi-rigid instrumentation: a case-control study.
        Front Endocrinol (Lausanne). 2020; 11552719
        • Schwaiger BJ
        • Gersing AS
        • Baum T
        • Noël PB
        • Zimmer C
        • Bauer JS.
        Bone mineral density values derived from routine lumbar spine multidetector row CT predict osteoporotic vertebral fractures and screw loosening.
        AJNR Am J Neuroradiol. 2014; 35: 1628-1633
        • Liu Y
        • Dash A
        • Krez A
        • Kim HJ
        • Cunningham M
        • Schwab F
        • et al.
        Low volumetric bone density is a risk factor for early complications after spine fusion surgery.
        Osteoporos Int. 2020; 31: 647-654
        • Xi Z
        • Mummaneni PV
        • Wang M
        • Ruan H
        • Burch S
        • Deviren V
        • et al.
        The association between lower Hounsfield units on computed tomography and cage subsidence after lateral lumbar interbody fusion.
        Neurosurg Focus. 2020; 49: E8
        • Sardar ZM
        • Coury JR
        • Cerpa M
        • DeWald CJ
        • Ames CP
        • Shuhart C
        • et al.
        Best practice guidelines for assessment & management of osteoporosis in adult patients undergoing elective spinal reconstruction.
        Spine (Phila Pa 1976). 2022; 47: 128-135
        • Tenne M
        • McGuigan F
        • Besjakov J
        • Gerdhem P
        • Akesson K.
        Degenerative changes at the lumbar spine–implications for bone mineral density measurement in elderly women.
        Osteoporos Int. 2013; 24: 1419-1428
        • Schneider P
        • Borner W.
        The impact of degenerative spinal changes on the correlation of peripheral and axial bone density.
        Nuklearmedizin. 1994; 33: 138-143
        • Guglielmi G
        • Floriani I
        • Torri V
        • Li J
        • van Kuijk C
        • Genant HK
        • et al.
        Effect of spinal degenerative changes on volumetric bone mineral density of the central skeleton as measured by quantitative computed tomography.
        Acta Radiol. 2005; 46: 269-275
        • Li N
        • Li XM
        • Xu L
        • Sun WJ
        • Cheng XG
        • Tian W.
        Comparison of QCT and DXA: osteoporosis detection rates in postmenopausal women.
        Int J Endocrinol. 2013; 2013895474
        • St Jeor JD
        • Jackson TJ
        • Xiong AE
        • Kadri A
        • Freedman BA
        • Sebastian AS
        • et al.
        Osteoporosis in spine surgery patients: what is the best way to diagnose osteoporosis in this population?.
        Neurosurg Focus. 2020; 49: E4
        • Zou D
        • Li W
        • Deng C
        • Du G
        • Xu N.
        The use of CT Hounsfield unit values to identify the undiagnosed spinal osteoporosis in patients with lumbar degenerative diseases.
        Eur Spine J. 2019; 28: 1758-1766
        • Kulkarni AG
        • Thonangi Y
        • Pathan S
        • Gunjotikar S
        • Goparaju P
        • Talwar I
        • et al.
        Should Q-CT be the gold standard for detecting spinal osteoporosis?.
        Spine (Phila Pa 1976). 2022; 47: E258-EE64
        • Zaidi Q
        • Danisa OA
        • Cheng W.
        Measurement techniques and utility of Hounsfield unit values for assessment of bone quality prior to spinal Instrumentation: a review of current literature.
        Spine (Phila Pa 1976). 2019; 44: E239-EE44
        • Engelke K
        • Keaveny TM.
        Letter to the Editor.
        Br J Radiol. 2019; 9220190115
        • Pinto EM
        • Neves JR
        • Teixeira A
        • Frada R
        • Atilano P
        • Oliveira F
        • et al.
        Efficacy of Hounsfield units measured by lumbar computer tomography on bone density assessment: A systematic review.
        Spine (Phila Pa 1976). 2022; 47: 702-710
        • Gausden EB
        • Nwachukwu BU
        • Schreiber JJ
        • Lorich DG
        • Lane JM.
        Opportunistic use of CT imaging for osteoporosis screening and bone density assessment: a qualitative systematic review.
        J Bone Joint Surg Am. 2017; 99: 1580-1590
        • Keaveny TM
        • Clarke BL
        • Cosman F
        • Orwoll ES
        • Siris ES
        • Khosla S
        • et al.
        Biomechanical Computed Tomography analysis (BCT) for clinical assessment of osteoporosis.
        Osteoporos Int. 2020; 31: 1025-1048
        • Melton LJ
        • Riggs BL
        • Keaveny TM
        • Achenbach SJ
        • Hoffmann PF
        • Camp JJ
        • et al.
        Structural determinants of vertebral fracture risk.
        J Bone Miner Res. 2007; 22: 1885-1892
        • Melton LJ
        • Riggs BL
        • Keaveny TM
        • Achenbach SJ
        • Kopperdahl D
        • Camp JJ
        • et al.
        Relation of vertebral deformities to bone density, structure, and strength.
        J Bone Miner Res. 2010; 25: 1922-1930
        • Wang X
        • Sanyal A
        • Cawthon PM
        • Palermo L
        • Jekir M
        • Christensen J
        • et al.
        Prediction of new clinical vertebral fractures in elderly men using finite element analysis of CT scans.
        J Bone Miner Res. 2012; 27: 808-816
        • Anderson DE
        • Demissie S
        • Allaire BT
        • Bruno AG
        • Kopperdahl DL
        • Keaveny TM
        • et al.
        The associations between QCT-based vertebral bone measurements and prevalent vertebral fractures depend on the spinal locations of both bone measurement and fracture.
        Osteoporos Int. 2014; 25: 559-566
        • Kopperdahl DL
        • Aspelund T
        • Hoffmann PF
        • Sigurdsson S
        • Siggeirsdottir K
        • Harris TB
        • et al.
        Assessment of incident spine and hip fractures in women and men using finite element analysis of CT scans.
        J Bone Miner Res. 2014; 29: 570-580
        • Allaire BT
        • Lu D
        • Johannesdottir F
        • Kopperdahl D
        • Keaveny TM
        • Jarraya M
        • et al.
        Prediction of incident vertebral fracture using CT-based finite element analysis.
        Osteoporos Int. 2019; 30: 323-331
        • Johannesdottir F
        • Allaire B
        • Kopperdahl DL
        • Keaveny TM
        • Sigurdsson S
        • Bredella MA
        • et al.
        Bone density and strength from thoracic and lumbar CT scans both predict incident vertebral fractures independently of fracture location.
        Osteoporos Int. 2021; 32: 261-269
        • Lee DC
        • Hoffmann PF
        • Kopperdahl DL
        • Keaveny TM.
        Phantomless calibration of CT scans for measurement of BMD and bone strength — Inter-operator reanalysis precision.
        Bone. 2017; 103: 325-333
        • Keaveny TM.
        Letter to the editor concerning "Best practice guidelines for assessment and management of osteoporosis in adult patients undergoing elective spinal reconstruction".
        Spine (Phila Pa 1976). 2022; 47: E466-E4E7
        • Burch S
        • Feldstein M
        • Hoffmann PF
        • Keaveny TM.
        Prevalence of poor bone quality in women undergoing spinal fusion using biomechanical-CT analysis.
        Spine (Phila Pa 1976). 2016; 41: 246-252
        • Guppy KH
        • Harris J
        • Paxton LW
        • Alvarez JL
        • Bernbeck JA.
        Reoperation rates for symptomatic nonunions in anterior cervical fusions from a national spine registry.
        Spine (Phila Pa 1976). 2015; 40: 1632-1637
        • Paxton EW
        • Inacio MCS
        • Kiley M-L.
        The Kaiser permanente implant registries: effect on patient safety, quality improvement, cost effectiveness, and research opportunities.
        Perm J. 2012; 16: 36-44
        • American College of Radiology
        ACR–SPR–SSR practice parameter for the performance of musculoskeletal quantitative computed tomography (QCT).
        Am Coll Radiol. 2018; (Report No. 9.): 1-14
        • Zysset P
        • Qin L
        • Lang T
        • Khosla S
        • Leslie WD
        • Shepherd JA
        • et al.
        Clinical use of quantitative computed tomography-based finite element analysis of the hip and spine in the management of osteoporosis in adults: The 2015 ISCD official positions-part II.
        J Clin Densitom. 2015; 18: 359-392
        • Kaneko TS
        • Bell JS
        • Pejcic MR
        • Tehranzadeh J
        • Keyak JH.
        Mechanical properties, density and quantitative CT scan data of trabecular bone with and without metastases.
        J Biomech. 2004; 37: 523-530
        • Zioupos P
        • Cook RB
        • Hutchinson JR.
        Some basic relationships between density values in cancellous and cortical bone.
        J Biomech. 2008; 41: 1961-1968
        • U.S. Department of Health and Human Services
        Bone Health and Osteoporosis: A Report of the Surgeon General.
        U.S. Department of Health and Human Services, Office of the Surgeon General, Rockville, MD2004
        • NIH Consensus Development Panel
        Osteoporosis prevention, diagnosis, and therapy.
        JAMA. 2001; 285: 785-795
        • Veeravagu A
        • Li A
        • Swinney C
        • Tian L
        • Moraff A
        • Azad TD
        • et al.
        Predicting complication risk in spine surgery: a prospective analysis of a novel risk assessment tool.
        J Neurosurg Spine. 2017; 27: 81-91
        • Guppy KH
        • Royse KE
        • Norheim EP
        • Harris JE
        • Brara HS.
        PLF versus PLIF and the fate of L5-S1: Analysis of operative nonunion rates among 3065 patients with lumbar fusions from a regional spine registry.
        Spine (Phila Pa 1976). 2021; 46: E584-EE93
        • Lee MJ
        • Cizik AM
        • Hamilton D
        • Chapman JR.
        Predicting medical complications after spine surgery: a validated model using a prospective surgical registry.
        Spine J. 2014; 14: 291-299
        • Ratliff JK
        • Balise R
        • Veeravagu A
        • Cole TS
        • Cheng I
        • Olshen RA
        • et al.
        Predicting occurrence of spine surgery complications using "big data" modeling of an administrative claims database.
        J Bone Joint Surg Am. 2016; 98: 824-834
        • Sato S
        • Yagi M
        • Machida M
        • Yasuda A
        • Konomi T
        • Miyake A
        • et al.
        Reoperation rate and risk factors of elective spinal surgery for degenerative spondylolisthesis: minimum 5-year follow-up.
        Spine J. 2015; 15: 1536-1544
        • Yagi M
        • Hosogane N
        • Fujita N
        • Okada E
        • Suzuki S
        • Tsuji O
        • et al.
        The patient demographics, radiographic index and surgical invasiveness for mechanical failure (PRISM) model established for adult spinal deformity surgery.
        Sci Rep. 2020; 10: 9341
        • Akins PT
        • Harris J
        • Alvarez JL
        • Chen Y
        • Paxton EW
        • Bernbeck J
        • et al.
        Risk factors associated with 30-day readmissions after instrumented spine surgery in 14,939 patients: 30-day readmissions after instrumented spine surgery.
        Spine (Phila Pa 1976). 2015; 40: 1022-1032
        • Looker AC
        • Borrud LG
        • Dawson-Hughes B
        • Shepherd JA
        • Wright NC.
        Osteoporosis or low bone mass at the femur neck or lumbar spine in older adults: United States, 2005-2008.
        NCHS Data Brief. 2012; : 1-8
        • Wright NC
        • Looker AC
        • Saag KG
        • Curtis JR
        • Delzell ES
        • Randall S
        • et al.
        The recent prevalence of osteoporosis and low bone mass in the United States based on bone mineral density at the femoral neck or lumbar spine.
        J Bone Miner Res. 2014; 29: 2520-2526
        • Okano I
        • Jones C
        • Salzmann SN
        • Miller CO
        • Shirahata T
        • Rentenberger C
        • et al.
        Postoperative decrease of regional volumetric bone mineral density measured by quantitative computed tomography after lumbar fusion surgery in adjacent vertebrae.
        Osteoporos Int. 2020; 31: 1163-1171
        • Okano I
        • Jones C
        • Rentenberger C
        • Sax OC
        • Salzmann SN
        • Reisener MJ
        • et al.
        The association between endplate changes and risk for early severe cage subsidence among standalone lateral lumbar interbody fusion patients.
        Spine (Phila Pa 1976). 2020; 45: E1580-E15E7
        • Jones C
        • Okano I
        • Salzmann SN
        • Reisener MJ
        • Chiapparelli E
        • Shue J
        • et al.
        Endplate volumetric bone mineral density is a predictor for cage subsidence following lateral lumbar interbody fusion: a risk factor analysis.
        Spine J. 2021; 21: 1729-1737
        • Fields AJ
        • Ballatori A
        • Liebenberg EC
        • Lotz JC.
        Contribution of the endplates to disc degeneration.
        Curr Mol Biol Rep. 2018; 4: 151-160
        • Wang Y
        • Videman T
        • Battié MC.
        ISSLS prize winner: Lumbar vertebral endplate lesions: associations with disc degeneration and back pain history.
        Spine (Phila Pa 1976). 2012; 37: 1490-1496
        • Feng Z
        • Liu Y
        • Yang G
        • Battié MC
        • Wang Y.
        Lumbar vertebral endplate defects on magnetic resonance images: classification, distribution patterns, and associations with modic changes and disc degeneration.
        Spine (Phila Pa 1976). 2018; 43: 919-927
        • Sun Z
        • Zheng X
        • Li S
        • Zeng B
        • Yang J
        • Ling Z
        • et al.
        Single impact injury of vertebral endplates without structural disruption, initiates disc degeneration through Piezo1 mediated inflammation and metabolism dysfunction.
        Spine (Phila Pa 1976). 2022; 47: E203-EE13
        • Okano I
        • Salzmann SN
        • Jones C
        • Ortiz Miller C
        • Shirahata T
        • Rentenberger C
        • et al.
        The impact of degenerative disc disease on regional volumetric bone mineral density (vBMD) measured by quantitative computed tomography.
        Spine J. 2020; 20: 181-190
        • Okano I
        • Jones C
        • Salzmann SN
        • Reisener MJ
        • Sax OC
        • Rentenberger C
        • et al.
        Endplate volumetric bone mineral density measured by quantitative computed tomography as a novel predictive measure of severe cage subsidence after standalone lateral lumbar fusion.
        Eur Spine J. 2020; 29: 1131-1140