Prediction Of Hip Fracture Risk Research Articles

Influence of test paradigm on loading dynamics during proximal femur fracture tests simulating sideways falls

Fall-related hip fractures are a serious public health issue in older adults. As most mechanistic hip fracture risk prediction models incorporate tissue tolerance, test methods that can accurately characterize the fracture force of the femur (and factors that influence it) are imperative. While bone possesses viscoelastic properties, experimental characterization of rate-dependencies has been inconsistent in the whole-femur literature. The goal of this study was to investigate the influence of experimental paradigm on loading rate and fracture force (both means and variability) during mechanical tests simulating lateral fall loadings on the proximal femur. Six pairs of matched femurs were split randomly between two test paradigms: a ‘lower rate’ materials testing system (MTS) with a constant displacement rate of 60 mm/s, and a hip impact test system (HIT) comprised of a custom-built vertical drop tower utilizing an impact velocity of 4 m/s. The loading rate was 88-fold higher for the HIT (mean (SD) = 2465.49 (807.38) kN/s) compared to the MTS (27.78 (10.03) kN/s) paradigm. However, no difference in fracture force was observed between test paradigms (mean (SD) = 4096.4 (1272.6) N for HIT, and 3641.3 (1285.8) N for MTS). Within-paradigm variability was not significantly different across paradigms for either loading rate or fracture force (coefficients of variation ranging from 0.311 to 0.361). Within each test paradigm, significant positive relationships were observed between loading rate and fracture force (HIT adjusted R2 = 0.833, p = 0.007; MTS adjusted R2 = 0.983, p < 0.0001). Overall, this study provides evidence that energy-based impact simulators can be a valid method to measure femoral bone strength in the context of fall-related hip fractures. This study motivates future research to characterize potential non-linear relationships between loading rate and fracture threshold at both macro and microscales.

DXA-based statistical models of shape and intensity outperform aBMD hip fracture prediction: A retrospective study

Areal bone mineral density (aBMD) currently represents the clinical gold standard for hip fracture risk assessment. Nevertheless, it is characterised by a limited prediction accuracy, as about half of the people experiencing a fracture are not classified as at being at risk by aBMD. In the context of a progressively ageing population, the identification of accurate predictive tools would be pivotal to implement preventive actions. In this study, DXA-based statistical models of the proximal femur shape, intensity (i.e., density) and their combination were developed and employed to predict hip fracture on a retrospective cohort of post-menopausal women. Proximal femur shape and pixel-by-pixel aBMD values were extracted from DXA images and partial least square (PLS) algorithm adopted to extract corresponding modes and components. Subsequently, logistic regression models were built employing the first three shape, intensity and shape-intensity PLS components, and their ability to predict hip fracture tested according to a 10-fold cross-validation procedure. The area under the ROC curves (AUC) for the shape, intensity, and shape-intensity-based predictive models were 0.59 (95%CI 0.47–0.69), 0.80 (95%CI 0.70–0.90) and 0.83 (95%CI 0.73–0.90), with the first being significantly lower than the latter two. aBMD yielded an AUC of 0.72 (95%CI 0.59–0.82), found to be significantly lower than the shape-intensity-based predictive model. In conclusion, a methodology to assess hip fracture risk uniquely based on the clinically available imaging technique, DXA, is proposed. Our study results show that hip fracture risk prediction could be enhanced by taking advantage of the full set of information DXA contains.

A new hip fracture risk index derived from FEA-computed proximal femur fracture loads and energies-to-failure.

Quantitative CT-based patient-specific finite element (FE) analysis (FEA) incorporates bone geometry and bone density in the proximal femur to compute the force (fracture load) and energy necessary to break the proximal femur in a particular loading condition. The fracture loads and energies-to-failure are individually associated with incident hip fracture, and provide different structural information about the proximal femur. We used principal component analysis (PCA) to develop a global FEA-computed fracture risk index that incorporates the FEA-computed yield and ultimate failure loads and energies-to-failure in four loading conditions of 110 hip fracture subjects and 235 age- and sex-matched control subjects from the AGES-Reykjavik study. Using a logistic regression model, we compared the prediction performance for hip fracture based on the stratified resampling. We referred the first principal component (PC1) of the FE parameters as the global FEA-computed fracture risk index, which was the significant predictor of hip fracture (p-value < 0.001). The area under the receiver operating characteristic curve (AUC) using PC1 (0.776) was higher than that using all FE parameters combined (0.737) in the males (p-value < 0.001). The global FEA-computed fracture risk index increased hip fracture risk prediction accuracy in males.

Higher serum free thyroxine levels are associated with increased risk of hip fractures in older men.

Overt and subclinical hyperthyroidism are associated with an increased fracture risk, but whether thyroid hormones are associated with fracture risk in individuals with normal thyroid-stimulating hormone (TSH) has mostly been investigated in women. Therefore, we investigated if serum levels of free thyroxine (FT4) or TSH are associated with fracture risk in Swedish men. We followed (median 12.2yr) elderly men (n = 1825; mean age 75, range 69-81yr) participating in the Gothenburg and Malmö subcohorts of the prospective, population-based MrOS-Sweden study. The statistical analyses included Cox proportional hazards regression. Men receiving levothyroxine treatment were excluded. In our total cohort, serum FT4 (per SD increase) was associated with increased risk of major osteoporotic fractures (MOFs; n = 479; fully adjusted hazard ratio [HR] 1.14, 95% CI, 1.05-1.24) and hip fractures (n = 207; HR 1.18, 95% CI, 1.04-1.33). Also, in men with normal TSH (n = 1658), FT4 (per SD increase) was significantly associated with increased risk of MOF and hip fractures. Furthermore, men in the highest FT4 quartile had a 1.5-fold increase in hip fracture risk compared with men in the three lower FT4 quartiles, both in the total population and in men with normal TSH (fully adjusted: HR 1.45, 95% CI, 1.04-2.02 and HR 1.51, 95% CI, 1.07-2.12, respectively). In contrast, the risk of MOF was not statistically different in the highest FT4 quartile compared with the three lower FT4 quartiles. Finally, serum TSH was not associated with fracture risk after full adjustment for covariates. In conclusion, serum FT4, but not serum TSH, is a predictor of hip fracture risk in elderly Swedish men. Additionally, there was an association between FT4 (per SD increase) and the risk of MOF.

Femurs segmentation by machine learning from CT scans combined with autonomous finite elements in orthopedic and endocrinology applications

Efficient, accurate and reliable segmentation of femurs from CT-scans is of major importance for patient-specific autonomous finite element analysis (AFE) to determine bone's stiffness and strength. We present a fully automated segmentation algorithm for whole and partial femurs with or without tumors, and clinical applications of the AFE [1] in clinical practice.The segmentation is based on an U-Net convolutional neural network, resulting a 3D mask representing the desired femur in a CT scan. It is robust, independent of the scanning parameters such as slice spacing, pixel size, scanner manufacturer or the femoral length available in the scan. The U-Net was trained on 178 manually segmented femurs (23,721 images) and tested on 43. The performance evaluation resulted in a Dice similarity score (DSC) of 0.9924, intersection over union (IoU) of 0.9849, Hausdorff distance of 4.3315 mm and symmetric average surface distance (ASD) of 0.1326 mm. The algorithm is competitive with the best state-of-the-art femoral segmentation methodologies available.Based on the segmentation an automatic p-FE mesh is generated and physiological boundary conditions representing sidewise falls or stance are being applied automatically to improve the performance of the AFE described in [1]. New examples of the usage of the AFE in endocrinology and orthopedic oncology demonstrate this disruptive technology in actual clinical practice. We present the use of AFE for predicting hip fracture risk in the elderly population due to a sidewise fall and the identification of patients who require a prophylactic surgery due to metastatic tumors in their femurs.

Effect of Artificial Intelligence or Machine Learning on Prediction of Hip Fracture Risk: Systematic Review.

Dual energy X-ray absorptiometry (DXA) is a preferred modality for screening or diagnosis of osteoporosis and can predict the risk of hip fracture. However, the DXA test is difficult to implement easily in some developing countries, and fractures have been observed before patients underwent DXA. The purpose of this systematic review is to search for studies that predict the risk of hip fracture using artificial intelligence (AI) or machine learning, organize the results of each study, and analyze the usefulness of this technology. The PubMed, OVID Medline, Cochrane Collaboration Library, Web of Science, EMBASE, and AHRQ databases were searched including "hip fractures" AND "artificial intelligence". A total of 7 studies are included in this study. The total number of subjects included in the 7 studies was 330,099. There were 3 studies that included only women, and 4 studies included both men and women. One study conducted AI training after 1:1 matching between fractured and non-fractured patients. The area under the curve of AI prediction model for hip fracture risk was 0.39 to 0.96. The accuracy of AI prediction model for hip fracture risk was 70.26% to 90%. We believe that predicting the risk of hip fracture by the AI model will help select patients with high fracture risk among osteoporosis patients. However, to apply the AI model to the prediction of hip fracture risk in clinical situations, it is necessary to identify the characteristics of the dataset and AI model and use it after performing appropriate validation.

1074-P: Variation in Predicted Fracture Risk by Diabetes and Obesity Status in Older U.S. Asian and White Women

Introduction: Diabetes (DM) is a risk factor for fracture but women with DM have less osteoporosis by bone density (BMD) , due in part to higher body mass index (BMI) . The FRAX fracture risk calculator incorporates age, race/ethnicity, BMD, BMI, and other risk factors but not DM. Few studies have examined FRAX by DM and obesity status in different race groups. We compared predicted fracture risk in US Asian and White women by DM and obesity status. Methods: We studied 3,640 Asian (26% DM, 24% osteoporosis, 15% obese) and 11,7White (11% DM, 12% osteoporosis, 40% obese) women aged 65-75y with BMD, excluding those with recent fracture, osteoporosis therapy, dialysis/transplant, metabolic bone disease or advanced cancer. DM (≥2 visit diagnoses + DM treatment) , obesity (BMI ≥30 kg/m2) and osteoporosis (BMD T score ≤-2.5) were obtained from electronic health data. Using the US FRAX with BMD, 10-year predicted hip fracture risk ≥3% was considered high (above treatment threshold) . Results: The overall proportions with FRAX-Hip ≥3% were higher for those with No DM vs. DM (White 18% vs. 11%; Asian 9% vs. 7%) . However, within both DM and no DM groups, FRAX-Hip ≥3% was much higher in non-obese vs. obese women of both races (FIGURE) . Conclusion: Predicted hip fracture risk varies by DM and especially by obesity status in White and Asian women. Future studies should examine the accuracy of FRAX for actual hip fracture risk in these subgroups. Disclosure J.C.Lo: None. M.Chandra: None. N.P.Gordon: None. C.Lee: None. Funding National Institute on Aging, National Institutes of Health (AG069992)

Reverse engineering the FRAX algorithm: Clinical insights and systematic analysis of fracture risk.

The Fracture Risk Assessment Tool (FRAX) is a computational tool developed to predict the 10-year probability of hip fracture and major osteoporotic fracture based on inputs of patient characteristics, bone mineral density (BMD), and a set of seven clinical risk factors. While the FRAX tool is widely available and clinically validated, its underlying algorithm is not public. The relative contribution and necessity of each input parameter to the final FRAX score is unknown. We systematically collected hip fracture risk scores from the online FRAX calculator for osteopenic Caucasian women across 473,088 unique inputs. This dataset was used to dissect the FRAX algorithm and construct a reverse-engineered fracture risk model to assess the relative contribution of each input variable. Within the reverse-engineered model, age and T-Score were the strongest contributors to hip fracture risk, while BMI had marginal contribution. Of the clinical risk factors, parent history of fracture and ongoing glucocorticoid treatment had the largest additive effect on risk score. A generalized linear model largely recapitulated the FRAX tool with an R2 of 0.91. Observed effect sizes were then compared to a true patient population by creating a logistic regression model of the Study of Osteoporotic Fractures (SOF) cohort, which closely paralleled the effect sizes seen in the reverse-engineered fracture risk model. Analysis identified several clinically relevant observations of interest to FRAX users. The role of major osteoporotic fracture risk prediction in contributing to an indication of treatment need is very narrow, as the hip fracture risk prediction accounted for 98% of treatment indications for the SOF cohort. Removing any risk factor from the model substantially decreased its accuracy and confirmed that more parsimonious models are not ideal for fracture prediction. For women 65years and older with a previous fracture, 98% of FRAX combinations exceeded the treatment threshold, regardless of T-score or other factors. For women age 70+ with a parent history of fracture, 99% of FRAX combinations exceed the treatment threshold. Based on these analyses, we re-affirm the efficacy of the FRAX as the best tool for fracture risk assessment and provide deep insight into the interplay between risk factors.

Biomechanical structure–function relations for human trabecular bone — comparison of calcaneus, femoral neck, greater trochanter, proximal tibia, and vertebra

MicroCT-based finite element models were used to compute power law relations for uniaxial compressive yield stress versus bone volume fraction for 78 cores of human trabecular bone from five anatomic sites. The leading coefficient of the power law for calcaneus differed from those for most of the other sites (p < 0.05). However, after normalizing by site-specific mean values, neither the leading coefficient (p > 0.5) nor exponent (p > 0.5) differed among sites, suggesting that a given percentage deviation from mean bone volume fraction has the same mechanical consequence for all sites investigated. These findings help explain the success of calcaneal x-ray and ultrasound measurements for predicting hip fracture risk.

Improving the Hip Fracture Risk Prediction with a Statistical Shape-and-Intensity Model of the Proximal Femur

Severe predictions have been made regarding osteoporotic fracture incidence for the next years, with major economic and social impacts in a worldwide greying society. However, the performance of the currently adopted gold standard for fracture risk prediction, the areal Bone Mineral Density (aBMD), remains moderate. To overcome current limitations, the construction of statistical models of the proximal femur, based on three-dimensional shape and intensity (a hallmark of bone density), is here proposed for predicting hip fracture in a Caucasian postmenopausal cohort. Partial Least Square (PLS)-based statistical models of the shape, intensity and their combination were developed, and the corresponding modes and components were identified. Logistic regression models using the first two shape, intensity and shape-intensity PLS components were implemented and tested within a 10-fold cross-validation procedure as predictors of hip fracture. It emerged that (1) intensity components were superior to shape components in stratifying patients according to their fracture status, and that (2) a combination of intensity and shape improved patients risk stratification. The area under the ROC curve was 0.64, 0.85 and 0.92 for the models based on shape, intensity and shape-intensity combination respectively, against a 0.72 value for the aBMD standard approach. Based on these findings, the presented methodology turns out to be promising in tackling the need for an enhanced fracture risk assessment.

Finite element derived femoral strength is a better predictor of hip fracture risk than aBMD in the AGES Reykjavik study cohort.

Hip fractures associated with a high economic burden, loss of independence, and a high rate of post-fracture mortality, are a major health concern for modern societies. Areal bone mineral density is the current clinical metric of choice when assessing an individual's future risk of fracture. However, this metric has been shown to lack sensitivity and specificity in the targeted selection of individuals for preventive interventions. Although femoral strength derived from computed tomography based finite element models has been proposed as an alternative based on its superior femoral strength prediction ex vivo, such predictions have only shown marginal or no improvement for assessing hip fracture risk. This study compares finite element derived femoral strength to aBMD as a metric for hip fracture risk assessment in subjects (N=601) from the AGES Reykjavik Study cohort and analyses the dependence of femoral strength predictions and classification accuracy on the material model and femoral loading alignment. We found hip fracture classification based on finite element derived femoral strength to be significantly improved compared to aBMD. Finite element models with non-linear material models performed better at classifying hip fractures compared to finite element models with linear material models and loading alignments with low internal rotation and adduction, which do not correspond to weak femur alignments, were found to be most suitable for hip fracture classification.

Implementing Machine Learning Algorithms on a Finite Element Simulated Dataset to Predict Hip Fracture Risk

Implementing Machine Learning Algorithms on a Finite Element Simulated Dataset to Predict Hip Fracture Risk

Predictive ability of novel volumetric and geometric indices derived from dual-energy X-ray absorptiometric images of the proximal femur for hip fracture compared with conventional areal bone mineral density: the Japanese Population-based Osteoporosis (JPOS) Cohort Study.

To clarify whether volumetric and geometric indices derived from novel three-dimensional (3D) modeling of the hip using dual-energy X-ray absorptiometric (DXA) images improve hip fracture prediction relative to areal bone mineral density (aBMD). We examined 1331 women who had completed the baseline survey and at least one follow-up survey over 20 years (age 40-79 years at baseline). Each survey included aBMD measurement at the hip by DXA. Volumetric and geometric indices of the hip at baseline and the 10-year follow-up were estimated from DXA images using a 3D modeling algorithm. Incident hip fractures during the 20-year follow-up period were identified through self-report. Cox proportional hazards regression models allowing for repeated measurements of predictors and outcomes were constructed, and their predictive ability for hip fracture was evaluated using areas under receiver operating characteristic curves (AUCs) and net reclassification improvement (NRI) over aBMD at the femoral neck (FN) and total hip (TH) as references. During a median follow-up of 19.8 years, 68 incident hip fractures were identified (2.22/1000 person-years). A significantly larger AUC of trabecular volumetric BMD (vBMD) at the total hip (AUC = 0.741), femoral neck (AUC = 0.748), and intertrochanter (AUC = 0.738) and significant NRI (0.177, 0.149, and 0.195, respectively) were observed compared with FN-aBMD (AUC = 0.701), but not TH-aBMD. vBMD obtained from 3D modeling using routinely obtained hip DXA images significantly improved hip fracture risk prediction over conventional FN-aBMD, but not TH-aBMD. The Japanese Population-Based Osteoporosis (JPOS) Cohort Study was retrospectively registered as UMIN000032869 in the UMIN Clinical Trials Registry on July 1, 2018.

On challenges in clinical assessment of hip fracture risk using image-based biomechanical modelling: a critical review.

Hip fracture is a common health risk among elderly people, due to the prevalence of osteoporosis and accidental fall in the population. Accurate assessment of fracture risk is a crucial step for clinicians to consider patient-by-patient optimal treatments for effective prevention of fractures. Image-based biomechanical modeling has shown promising progress in assessment of fracture risk, and there is still a great possibility for improvement. The purpose of this paper is to identify key issues that need be addressed to improve image-based biomechanical modeling. We critically examined issues in consideration and determination of the four biomechanical variables, i.e., risk of fall, fall-induced impact force, bone geometry and bone material quality, which are essential for prediction of hip fracture risk. We closely inspected: limitations introduced by assumptions that are adopted in existing models; deficiencies in methods for construction of biomechanical models, especially for determination of bone material properties from bone images; problems caused by separate use of the variables in clinical study of hip fracture risk; availability of clinical information that are required for validation of biomechanical models. A number of critical issues and gaps were identified. Strategies for effectively addressing the issues were discussed.

Risk Assessment of Hip Fracture Based on Machine Learning.

Identifying patients with high risk of hip fracture is a great challenge in osteoporosis clinical assessment. Bone Mineral Density (BMD) measured by Dual-Energy X-Ray Absorptiometry (DXA) is the current gold standard in osteoporosis clinical assessment. However, its classification accuracy is only around 65%. In order to improve this accuracy, this paper proposes the use of Machine Learning (ML) models trained with data from a biomechanical model that simulates a sideways-fall. Machine Learning (ML) models are models able to learn and to make predictions from data. During a training process, ML models learn a function that maps inputs and outputs without previous knowledge of the problem. The main advantage of ML models is that once the mapping function is constructed, they can make predictions for complex biomechanical behaviours in real time. However, despite the increasing popularity of Machine Learning (ML) models and their wide application to many fields of medicine, their use as hip fracture predictors is still limited. This paper proposes the use of ML models to assess and predict hip fracture risk. Clinical, geometric, and biomechanical variables from the finite element simulation of a side fall are used as independent variables to train the models. Among the different tested models, Random Forest stands out, showing its capability to outperform BMD-DXA, achieving an accuracy over 87%, with specificity over 92% and sensitivity over 83%.

The role of cystatin C in the risk of hip fractures in the elderly

World’s population is aging. The elderly are at high risk for both chronic kidney disease (CKD) and hip fractures. Severe chronic kidney disease is a well-known risk factor for fractures and death especially in the elderly. Mild and moderate stages of kidney disease are often undiagnosed and/or untreated, thus their effect on fracture risk is not well established. Many ways of estimating glomerular filtration rate (GFR) have been developed but there are very few studies recommending the best and most valuable method for estimated GFR (eGFR) calculation that could correlate with fracture risk. In this mini- review we searched the literature concerning the use of cystatin C in the estimation of GFR related to the risk of hip fractures in the elderly. Our goal was to review the most important recent evidence on whether cystatin C could become a useful biomarker for the prediction of fracture risk. We concluded that there is evidence to support the use of cystatin C in hip fracture risk prediction in elder patients with chronic kidney disease.

Evaluation of patient tissue selection methods for deriving equivalent density calibration for femoral bone quantitative CT analyses.

Osteoporosis affects an increasing number of people every year and patient specific finite element analysis of the femur has been proposed to identify patients that could benefit from preventative treatment. The aim of this study was to demonstrate, verify, and validate an objective process for selecting tissues for use as the basis of phantomless calibration to enable patient specific finite element analysis derived hip fracture risk prediction. Retrospective reanalysis of patient computed tomography (CT) scans has the potential to yield insights into more accurate prediction of osteoporotic fracture. Bone mineral density (BMD) specific calibration scans are not typically captured during routine clinical practice. Tissue-based BMD calibration can therefore empower the retrospective study of patient CT scans captured during routine clinical practice. Together the method for selecting tissues as the basis for phantomless calibration coupled with the post-processing steps for deriving a calibration equation using the selected tissues provide an estimation of quantitative equivalent density results derived using calibration phantoms. Patient tissues from a retrospective cohort of 211 patients were evaluated. The best phantomless calibration resulted in a femoral strength (FS) [N] bias of 0.069±0.07% over FS derived from inline calibration and a BMD [kg/cm3] bias of 0.038±0.037% over BMD derived from inline calibration. The phantomless calibration slope for the best method presented was within the range of patient specific calibration curves available for comparison and demonstrated a small bias of 0.028±0.054 HU/(mg/cm3), assuming the Mindways Model 3 BMD inline calibration phantom as the gold standard. The presented method of estimating a calibration equation from tissues showed promise for CT-based femoral fracture analyses of retrospective cohorts without readily available calibration data.

3D Analysis of the Proximal Femur Compared to 2D Analysis for Hip Fracture Risk Prediction in a Clinical Population

Due to the adverse impacts of hip fractures on patients' lives, it is crucial to enhance the identification of people at high risk through accessible clinical techniques. Reconstructing the 3D geometry and BMD distribution of the proximal femur could be beneficial in enhancing hip fracture risk predictions; however, it is associated with a high computational burden. It is also not clear whether it provides a better performance than 2D model analysis. Therefore, the purpose of this study was to compare the 2D and 3D model reconstruction's ability to predict hip fracture risk in a clinical population of patients. The DXA scans and CT scans of 16 cadaveric femurs were used to create training sets for the 2D and 3D model reconstruction based on statistical shape and appearance modeling. Subsequently, these methods were used to predict the risk of sustaining a hip fracture in a clinical population of 150 subjects (50 fractured, and 100 non-fractured) that were monitored for five years in the Canadian Multicentre Osteoporosis Study. 3D model reconstruction was able to improve the identification of patients who sustained a hip fracture more accurately than the standard clinical practice (by 40%). Also, the predictions from the 2D statistical model didn't differ significantly from the 3D ones (p > 0.76). These results indicated that to enhance hip fracture risk prediction in clinical practice implementing 2D statistical modeling has comparable performance with lower associated computational load.

Predicting population level hip fracture risk: a novel hierarchical model incorporating probabilistic approaches and factor of risk principles

Fall-related hip fractures are a major public health issue. While individual-level risk assessment tools exist, population-level predictive models could catalyze innovation in large-scale interventions. This study presents a hierarchical probabilistic model that predicts population-level hip fracture risk based on Factor of Risk (FOR) principles. Model validation demonstrated that FOR output aligned with a published dataset categorized by sex and hip fracture status. The model predicted normalized FOR for 100000 individuals simulating the Canadian older-adult population. Predicted hip fracture risk was higher for females (by an average of 38%), and increased with age (by15% per decade). Potential applications are discussed.

Association between post-stroke disability and 5-year hip-fracture risk: The Women's Health Initiative

BackgroundHip fractures are a significant post-stroke complication. We examined predictors of hip fracture risk after stroke using data from the Women's Health Initiative (WHI). In particular, we examined the association between post-stroke disability levels and hip fracture risk. MethodsThe WHI is a prospective study of 161,808 postmenopausal women aged 50-79 years. Trained physicians adjudicated stroke events and hip fractures. Our study included stroke survivors from the observational and clinical trial arms who had a Glasgow Outcome Scale of good recovery, moderately disabled, or severely disabled and survived more than 7 days post-stroke. Hip fracture-free status was compared across disability levels. Secondary analysis examined hip fracture risk while accounting for competing risk of death. ResultsAverage age at time of stroke was 74.6±7.2 years; 84.3% were white. There were 124 hip fractures among 4,640 stroke survivors over a mean follow-up time of 3.1±1.8 years. Mortality rate was 23.3%. Severe disability at discharge (Hazard Ratio (HR): 2.1 (95% Confidence Interval (CI): 1.4-3.2), but not moderate disability (HR: 1.1 (95%CI: 0.7-1.7), was significantly associated with an increased risk of hip fracture compared to good recovery status. This association was attenuated after accounting for mortality. White race, increasing age and higher Fracture Risk Assessment Tool (FRAX)-predicted hip fracture risk (without bone density information) were associated with an increased hip fracture risk. After accounting for mortality, higher FRAX risk and white race remained significant. ConclusionSevere disability after stroke and a higher FRAX risk score were associated with risk of subsequent hip fracture. After accounting for mortality, only the FRAX risk score remained significant. The FRAX risk score appears to identify stroke survivors at high risk of fractures. Our results suggest that stroke units can consider the incorporation of osteoporosis screening into care pathways.