Long-term consequences to individual health and societal healthcare costs of untreated osteoporotic vertebral fractures: Towards opportunistic identification

Background and aim
Vertebral fractures (VFs) are a common consequence of osteoporosis, and they are widely prevalentamong the aging population. Most VFs are not clinically diagnosed, yet they are often present onradiological imaging performed in the context of daily clinical practice, in some studies exceeding 1in 3 exams. Identifying and reporting such opportunistically occurring VFs may be the first steps toprevent future adverse health events and thus improve patient outcomes. However, most of theseVFs are left unreported.While VFs have been extensively studied in general, very few publications address theconsequences of opportunistically identifiable VFs. This may contribute to the aforementionedunderreporting and hinders assessment of whether proper reporting and management of these VFs –observed amidst competing morbidities – is clinically warranted. Therefore, to help bridge thisknowledge gap, the aim of this project is to evaluate the key health outcomes and societal costs ofmen and women with VFs identified on CT scans performed as part of daily clinical practice, andnot treated with osteoporosis medications.

Methods
This PhD-project was designed as an observational cohort study in men and women aged 50 yearsor older. From the Holbæk Hospital radiology database we identified the first 2,000 such subjects –as of 1st January 2010 and onwards – with a CT scan including the chest and/or abdomen. PrevalentVFs were identified from these scans according to the Genant Semiquantitative method, and thefindings linked with the national Danish registers for further analyses.Subjects with VF on the CT scan were matched on age and sex against subjects without VF (1:2-ratio) to form the analysis population, and against a general population sample (1:3-ratio) to formthe scaling analysis population. Subjects were excluded if treated with osteoporosis medication(s)in the year prior to baseline.We evaluated the risk of subsequent fractures and mortality, respectively, during up to seven yearsof follow-up. The main outcomes measures were examined by fitting Cox proportional hazardsregression models.We also evaluated healthcare costs (including cost of admissions, outpatient visits, and prescriptionmedications) for up to 6 years after the CT scan. Primary healthcare costs were evaluatedseparately. Differences between the cohorts were assessed by linear mixed-effects models.

Results
In the 2,000 subjects with a CT scan, re-evaluation of the scans yielded 1,111 VFs in 423 subjects.After exclusion and matching, the VF and no VF cohorts were constituted of 321 and 606 subjects(analysis population), while the VF and general population cohorts were constituted of 332 and 996subjects (scaling analysis population). Median age was 73 years in all cohorts, and 53-55% of thecohorts were men.Publication I examines the risk of fractures, and show an increased risk of a first major osteoporoticfracture in subjects with VF as compared to those with no VF (adjusted HR [HRadj] 1.72 [95%confidence interval 1.03-2.86]) and when compared to the general population cohort (HRadj 2.04[1.29-3.23]). This was driven by hip and vertebral fractures. In terms of any subsequent fracture(except face, skull, and fingers), there were no significant difference between the VF and no VFcohort, while the risk was increased in subjects with VF as compared to the general populationcohort (HRadj 1.60 [1.07-2.40]).Publication II reports on mortality, and shows mortality rates of 276 and 155 deaths per 1,000subject-years in the VF and no VF cohorts (analysis population), respectively. There was anincreased risk of death in the VF cohort versus the no VF cohort (HRadj 1.51 [1.27-1.79]), and alsowhen compared to the general population cohort (HRadj 4.25 [3.53-5.12]). Subgroup analysesshowed a higher relative risk in subjects with moderate or severe VFs, and in those with a lowerburden of comorbidities.Publication III shows that healthcare costs per day at risk were numerically higher in the VF cohortas compared to the no VF cohort in the first years after the CT scan, yet this did not achievestatistical significance. Costs per year were similar between these cohorts. When compared to thegeneral population cohort, the costs in the VF cohort were consistently significantly higher.

Conclusions
In this project, it is demonstrated that subjects with VFs available on routine CT scans and nottreated with osteoporosis medications are at an increased risk of major osteoporotic fractures anddeath, when compared to those with no VFs on the CT scan. When compared to a generalpopulation cohort, these risk differences are even more pronounced. Healthcare costs weresignificantly higher in the VF cohort only when compared to the general population cohort.Altogether, this project highlights the rationale for improved reporting of VFs from routine CTscans, to allow proper management of these patients in order to improve patient outcomes.Questions on implementation of systematic opportunistic screening for VFs remains to beelucidated, and care pathways for these patients are yet to be agreed upon.