Combining Frailty and Trabecular Bone Score Did Not Improve Predictive Accuracy in Risk of Major Osteoporotic Fractures
Jerilynn C Prior, Tanveer Towheed, Kelly Davison
Lehana Thabane, Mitchell Levine, Jie Zeng, Junzhan Tian , Jonathan D Adachi, for the Canadian Multicentre Osteoporosis Study (CaMos) Research Group , Canadian Multicentre Osteoporosis Study (CaMos) Research Group , Alexandra Papaioannou, Christopher S Kovacs, David Goltzman, Guowei Li, K Shawn Davison, Yanfei Qi , Robert G Josse, William D Leslie
It is recognized that the trabecular bone score (TBS) provides skeletal information, and frailty measurement is significantly associated with increased risks of adverse health outcomes. Given the suboptimal predictive power in fracture risk assessment tools, we aimed to evaluate the combination of frailty and TBS regarding predictive accuracy for risk of major osteoporotic fracture (MOF). Data from the prospective longitudinal study of CaMos (Canadian Multicentre Osteoporosis Study) were used for this study. TBS values were estimated using lumbar spine (L1 to L4 ) dual-energy X-ray absorptiometry (DXA) images; frailty was evaluated by a frailty index (FI) of deficit accumulation. Outcome was time to first incident MOF during the follow-up. We used the Harrell's C-index to compare the model predictive accuracy. The Akaike information criterion, likelihood ratio test, and net reclassification improvement (NRI) were used to compare model performances between the model combining frailty and TBS (subsequently called "FI + TBS"), FI-alone, and TBS-alone models. We included 2730 participants (mean age 69 years; 70% women) for analyses (mean follow-up 7.5 years). There were 243 (8.90%) MOFs observed during follow-up. Participants with MOF had significantly higher FI (0.24 versus 0.20) and lower TBS (1.231 versus 1.285) than those without MOF. FI and TBS were significantly related with MOF risk in the model adjusted for FRAX with bone mineral density (BMD) and other covariates: hazard ratio (HR) = 1.26 (95% confidence interval [CI] 1.11-1.43) for per-SD increase in FI; HR = 1.38 (95% CI 1.21-1.59) for per-SD decrease in TBS; and these associations showed negligible attenuation (HR = 1.24 for per-SD increase in FI, and 1.35 for per-SD decrease in TBS) when combined in the same model. Although the model FI + TBS was a better fit to the data than FI-alone and TBS-alone, only minimal and nonsignificant enhancement of discrimination and NRI were observed in FI + TBS. To conclude, frailty and TBS are significantly and independently related to MOF risk. Larger studies are warranted to determine whether combining frailty and TBS can yield improved predictive accuracy for MOF risk.
trabecular bone score frailty fracture prediction, frailty index osteoporotic fracture risk CaMos, TBS FRAX fracture risk assessment elderly women, Prior JC bone mineral density fracture prediction, Canadian Multicentre Osteoporosis Study fracture risk model, frailty deficit accumulation major osteoporotic fracture, trabecular bone score lumbar spine DXA fracture, net reclassification improvement fracture prediction model, BMD T-score frailty combined predictive accuracy, Cox proportional hazards fracture risk elderly women, Leslie WD trabecular bone score fracture prediction
PMID 31995642 31995642 DOI 10.1002/jbmr.3971 10.1002/jbmr.3971
Cite this article
Li, G., Leslie, W. D., Kovacs, C. S., Prior, J., Josse, R. G., Towheed, T., Davison, K. S., Thabane, L., Papaioannou, A., Levine, M. A., Goltzman, D., Zeng, J., Qi, Y., Tian, J., Adachi, J. D., & Canadian Multicentre Osteoporosis Study (CaMos) Research Group (2020). Combining Frailty and Trabecular Bone Score Did Not Improve Predictive Accuracy in Risk of Major Osteoporotic Fractures. Journal of bone and mineral research : the official journal of the American Society for Bone and Mineral Research, 35(6), 1058-1064. https://doi.org/10.1002/jbmr.3971
Li G, Leslie WD, Kovacs CS, Prior J, Josse RG, Towheed T, et al. Combining Frailty and Trabecular Bone Score Did Not Improve Predictive Accuracy in Risk of Major Osteoporotic Fractures. J Bone Miner Res. 2020;35(6):1058-1064. doi:10.1002/jbmr.3971
Li, G., et al. "Combining Frailty and Trabecular Bone Score Did Not Improve Predictive Accuracy in Risk of Major Osteoporotic Fractures." Journal of bone and mineral research : the official journal of the American Society for Bone and Mineral Research, vol. 35, no. 6, 2020, pp. 1058-1064.
Related articles
Bone HealthBody Mass Index AssociationFracture Risk FactorsFracture Epidemiology
Several recent studies suggest that obesity may be a risk factor for fracture. The aim of this study was to investigate the association between body mass index (BMI) and future fracture risk at different skeletal sites. In prospective cohorts from more than 25 countries, baseline data on BMI were available in 398,610 women with an average age of 63 (range, 20-105) years and follow up of 2.2 million person-years during which 30,280 osteoporotic fractures (6457 hip fractures) occurred. Femoral neck BMD was measured in 108,267 of these women. Obesity (BMI ≥ 30 kg/m(2) ) was present in 22%. A majority of osteoporotic fractures (81%) and hip fractures (87%) arose in non-obese women. Compared to a BMI of 25 kg/m(2) , the hazard ratio (HR) for osteoporotic fracture at a BMI of 35 kg/m(2) was 0.87 (95% confidence interval [CI], 0.85-0.90). When adjusted for bone mineral density (BMD), however, the same comparison showed that the HR for osteoporotic fracture was increased (HR, 1.16; 95% CI, 1.09-1.23). Low BMI is a risk factor for hip and all osteoporotic fracture, but is a protective factor for lower leg fracture, whereas high BMI is a risk factor for upper arm (humerus and elbow) fracture. When adjusted for BMD, low BMI remained a risk factor for hip fracture but was protective for osteoporotic fracture, tibia and fibula fracture, distal forearm fracture, and upper arm fracture. When adjusted for BMD, high BMI remained a risk factor for upper arm fracture but was also a risk factor for all osteoporotic fractures. The association between BMI and fracture risk is complex, differs across skeletal sites, and is modified by the interaction between BMI and BMD. At a population level, high BMI remains a protective factor for most sites of fragility fracture. The contribution of increasing population rates of obesity to apparent decreases in fracture rates should be explored.
Bone HealthImminent Risk PredictionRisk Factor ModelingStructural Equation Modeling
We examined the underlying relationship between fracture risk factors and their imminent risk. Results suggested that having past year fracture, worse past year general health, worse past year physical functioning, and lower past year BMD T-score directly predicted higher imminent fracture risk. Past year falls indirectly predicted imminent risk through physical functioning and general health.
This study aimed to examine direct and indirect effects of several factors on imminent (1 year) fracture risk. Data from women age 65 and older from population-based Canadian Multicentre Osteoporosis Study were used. Predictors were identified from study years 5 and 10, and imminent fracture data (1-year fracture) came from years 6 and 11 (year 5 predicts year 6, year 10 predicts year 11). A structural equation model (SEM) was used to test the theoretical construct. General health and physical functioning were measured as latent variables using items from the 36-Item Short Form Health Survey (SF-36) and bone mineral density (BMD) T-score was a latent variable based on observed site-specific BMD data (spine L1-L4, femoral neck, total hip). Observed variables were fractures and falls. Model fit was evaluated using root mean square error of approximation (RMSEA), Tucker Lewis index (TLI), and comparative fit index (CFI). The analysis included 3298 women. Model fit tests showed that the SEM fit the data well; χ2(172) = 1122.10 < .001, RMSEA = .03, TLI = .99, CFI = .99. Results suggested that having past year fracture, worse past year general health, worse past year physical functioning, and lower past year BMD T-score directly predicted higher risk of fracture in the subsequent year (p < .001). Past year falls had a statistically significant but indirect effect on imminent fracture risk through physical functioning and general health (p < .001). We found several direct and indirect pathways that predicted imminent fracture risk in elderly women. Future studies should extend this work by developing risk scoring methods and defining imminent risk thresholds.
Bone HealthCognitive Decline AssociationMedication EffectsBias by Indication
We are grateful to Dr. Naharci for the interest in our study reporting the association between cognitive decline and bone loss and fracture risk. (1) We agree that bisphosphonates (BPs) have a proven effect on reducing bone loss and fracture risk. (2) Medication with anticholinergic (ACH) side effects may also affect cogni-tive function as well as propensity to fall and fracture, although these effects have not been demonstrated in all studies. (3) We did not include these medication classes in our models because in observational studies the relationship between medication and outcomes is likely driven by factors associated with medication use. This bias by indication can only be avoided by specific study design (ie, propensity score matching), which was beyond the scope of our study. (4) However, we have conducted additional analyses to determine the prevalence of BP and ACH medication in our cohort, the association between these medication classes and our study outcomes and the impact of the addition of these medication classes to our findings. BP and ACH use were self-reported and obtained by questionnaire at baseline, and years 5 and 10. Bone mineral density (BMD) was assessed by dual-energy X-ray absorptiometry (DXA) and cognitive function using the Mini Mental State Examination (MMSE) test during all clinical visits. Follow-up time for BMD
Bone HealthRisk Prediction ModelsFracture Risk AssessmentMultistate Prediction
Existing fracture risk assessment tools are not designed to predict fracture-associated consequences, possibly contributing to the current undermanagement of fragility fractures worldwide. We aimed to develop a risk assessment tool for predicting the conceptual risk of fragility fractures and its consequences. The study involved 8965 people aged ≥60 years from the Dubbo Osteoporosis Epidemiology Study and the Canadian Multicentre Osteoporosis Study. Incident fracture was identified from X-ray reports and questionnaires, and death was ascertained though contact with a family member or obituary review. We used a multistate model to quantify the effects of the predictors on the transition risks to an initial and subsequent incident fracture and mortality, accounting for their complex interrelationships, confounding effects, and death as a competing risk. There were 2364 initial fractures, 755 subsequent fractures, and 3300 deaths during a median follow-up of 13 years (interquartile range [IQR] 7-15). The prediction model included sex, age, bone mineral density, history of falls within 12 previous months, prior fracture after the age of 50 years, cardiovascular diseases, diabetes mellitus, chronic pulmonary diseases, hypertension, and cancer. The model accurately predicted fragility fractures up to 11 years of follow-up and post-fracture mortality up to 9 years, ranging from 7 years after hip fractures to 15 years after non-hip fractures. For example, a 70-year-old woman with a T-score of -1.5 and without other risk factors would have 10% chance of sustaining a fracture and an 8% risk of dying in 5 years. However, after an initial fracture, her risk of sustaining another fracture or dying doubles to 33%, ranging from 26% after a distal to 42% post hip fracture. A robust statistical technique was used to develop a prediction model for individualization of progression to fracture and its consequences, facilitating informed decision making about risk and thus treatment for individuals with different risk profiles.