Abstract
Background
Hip 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.
Methods
The 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.
Results
Average 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.
Conclusion
Severe 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.
Keywords: Hip Fracture, Women, Stroke, Glasgow Outcome Scale
Background and Purpose
Fractures are a serious post-stroke complication and lead to increased morbidity and poorer functional outcomes.(1) About 3–6% of stroke survivors experience a fracture in the first-year following stroke.(2–6) Stroke survivors are at two to four times higher risk of experiencing a hip fracture compared to those who have not had a stroke.(3, 7, 8)
Stroke survivors are a heterogeneous group with some having minimal or no disability while others need assistance in activities of daily living (26%).(9, 10) Few studies on hip fracture risk after stroke account for post-stroke functional status and, there is variability in the reported results.(7, 11–13) Research suggests that bone mineral density decreases on the paretic side after stroke and, that these changes are more marked with increasing disability.(3, 14, 15) There may be a role for preventive therapies post-stroke for individuals at an increased risk of fractures.
In this study, we describe the risk of hip fractures after stroke among the participants of the Women’s Health Initiative (WHI). We estimated this risk across three disability levels measured at the time of discharge. We examined models that accounted for the competing risk of death since stroke survivors have a high subsequent mortality. The WHI represents a unique data source for this study due to detailed baseline data on all participants, carefully adjudicated stroke and hip fracture outcomes during prospective follow-up, and data on medication use, specifically those related to bone health.
Methods
Deidentified data and materials have been made publicly available at the NHLBI BioLINCC and can be accessed at https://biolincc.nhlbi.nih.gov/search/.
Study population
The WHI is a long running study of 161,808 postmenopausal women enrolled between the ages of 50–79 years. The trial has an observational component (OS, n=93,676) and a clinical trial component (CT, n=68,132). Data from both CT and OS participants are included in our analytic sample.
The design and recruitment for the WHI has been described previously.(16) The institutional review boards of each participating institution approved the study protocol. Each study participant provided written informed consent. Briefly, the WHI enrolled participants from 1993 through 1998. The main study completed follow up in 2005. Subsequently, participants were invited into the Extension Study 1 (Ext1) through 2010.
The WHI collected extensive baseline data including participant demographics, diet, physiological and physical measurements, medical history, and psychosocial and personal habits. A total of 6,771 women had a physician adjudicated stroke event. Each adjudication form included data on disability at discharge collected using the GOS (Glasgow Outcome Scale), a standard measure of outcome after brain injury.(17) We excluded those who had a GOS status of vegetative (n=42), deceased (n=660) or missing/unknown (n=862), had stroke in Ext2 (n=545), and did not survive more than 7-days (n=22), resulting in a total of 4,640 stroke survivors in our study sample (Figure 1).(18)
Figure 1.
Strengthening the Reporting of Observational studies in Epidemiology (STROBE) Diagram of Analytic Cohort with inclusion and exclusion criteria. Women’s Health Initiative (1993–2010).
Stroke ascertainment
The WHI defined stroke as a rapid onset of neurologic deficit lasting more than 24 hours, of vascular etiology and, without evidence of other causes. Participants self-reported stroke at annual study visits.(19, 20) Neurologists centrally adjudicated all strokes requiring hospital stays using medical record-based criteria set out by the WHI.(18) Only one participant in our study sample reported a stroke prior to the WHI study. We only included strokes that occurred during the WHI study period in our analysis. For participants with multiple stroke events during the WHI follow up, only the first adjudicated event is included in our analyses. We included stroke events occurring from the beginning of the study (1993–1998) through Ext1 (2010).
Hip fracture ascertainment
Hip fracture was self-reported via annual questionnaires. Study physicians blinded to clinical trial and observational arm assignments centrally adjudicated all self-reported hip fractures using radiological reports and/or hospital discharge summaries.(18) Adjudication of hip fractures ended after Ext1 (2010). Hence, ascertainment of hip fractures for this study ended at Ext1 (2010). Only the first adjudicated hip fracture was used for this analysis. We have 19–24 years of follow-up data on current participants, with hip fracture ascertainment and adjudication through Ext1 or 2010. Hip fracture was the only fracture variable that was adjudicated and therefore, fractures involving other body parts were not considered.
Covariates
We used the following covariates in our analysis: age, race/ethnicity (white vs non-white race/ethnicity), Fracture Risk Assessment Tool (FRAX)-predicted 10-year hip fracture risk without bone density information (BMD) (version 3.0, https://www.sheffield.ac.uk/FRAX), (21) previous osteoporosis diagnosis, GOS status (good recovery, moderately disabled, severely disabled), hormone therapy use (HT) (never, current, former user), depressive symptoms score (Center for Epidemiologic Studies Depression Scale (CES-D)),(22) diabetes history, smoking status (current, former, never), calcium supplements or osteoporosis medication use, and enrollment in one of three clinical trial arms (hormone therapy, calcium/vitamin D supplementation, and dietary modification).
We defined osteoporosis medication as use of an approved US Food and Drug Administration medication during the study period (1993–2010) up until the time of stroke including: bisphosphonates (alendronate, etidronate, ibandronate, risedronate, and zoledronic acid), calcitonin, raloxifene, and teriparatide. Denosumab was not approved until 2010 and is therefore not included for this analysis. Use of osteoporosis medication prior to stroke is likely a marker of having osteoporosis. FRAX without BMD predicts the 10-year absolute risk of hip fracture based on clinical risk factors, including age, parental hip fracture, and previous fractures.(21) The FRAX without BMD risk score was taken at the baseline assessment (1993–1998). The FRAX risk score was dichotomized to <3% as low/medium risk and ≥3% as high risk for 10-year hip fracture risk, according to the National Osteoporosis Foundation treatment recommendations regarding treatment of osteopenia.(23)
GOS was determined by discharge medical records review by trained vascular neurologists.(24) Participants were classified into six categories: good recovery, moderately disabled, severely disabled, vegetative survival, dead, or unable to categorize. The WHI adjudication manual defined the GOS categories as follows: Good recovery: Participants could lead a full and independent life with or without minimal neurological deficit. Moderately disabled: participants who had neurological or intellectual impairment but were independent. Participants with severe impairment entering acute rehabilitation were classified as moderately disabled if the medical record provided information that the expectation was of recovery to an independent functional level. This takes into account post-stroke recovery. Severely disabled: participants who were conscious but dependent on others to get through daily activities. Vegetative survival: Participants who had no obvious cortical functioning.
History of osteoporosis diagnosis, HT, and the clinical trial enrollments were stratified in analyses.
Statistical Analysis
We summarized participant characteristics using measures of centrality and dispersion (mean, standard deviation) for continuous variables and frequency measures for categorical variables. We examined the distribution of hip fracture by post-stroke disability status. Mean imputation was used for those who were missing a depressive symptoms score (CES-D) (n=156). Less than 1% of the sample were missing the remaining covariates and therefore these were not imputed. We calculated the incidence of hip fractures as a function of time among those with a stroke and disability status of good recovery (GOS=1), moderate disability (GOS=2), and severe disability (GOS=3). In order to account for the substantial post-stroke mortality, we performed a secondary analysis accounting for the competing risk of death. Cox proportional hazards models estimated the effect of disability on the risk of post-stroke hip fracture. Time in days from stroke to fracture, death, or end of follow-up was calculated. Day of stroke and hip fracture were based on day of diagnosis. Survival free from hip fracture post-stroke was estimated using Kaplan-Meier curves and Fine and Gray (competing risk) method.(25) Cox proportional hazard (PH) regression models and Fine and Gray competing risk models were included in the analyses of this paper. Both regression analyses were included to assist in understanding the complete effect of the risk factors on the incidence and rate of outcome.(26) Cox PH models answer an etiological question and censors all events that are not the outcome of interest.(26, 27) Competing risk models answer a predicted prognosis question and keeps individuals at-risk until they experience the outcome of interest.(26, 27) Individuals remain at-risk even after experiencing the competing event.(27) Log rank tests were used to compare distributions of multiple variables including GOS, age categories of <65 years and ≥ 65 years, and stroke type. Cumulative risk and incidence of hip fracture were summarized at 30-days, 1-year, and 5-years post-stroke.
Models were run unadjusted and then adjusted for age, history of diabetes, race/ethnicity, baseline depressive symptoms score, history of HT, baseline FRAX risk score, history of osteoporosis diagnosis, post-stroke disability status (GOS), calcium supplements or osteoporosis medication use, smoking status, and the clinical trial enrollments. Schoenfeld Residual plots assessed the proportional hazard assumption. There was a modest deviation for GOS. We explored multiplicative interactions of GOS or FRAX score by covariates.
We censored participants at death, loss to follow-up, or end of Extension 1, whichever came first, for the non-competing risk model. In the competing risk model, we censored participants at loss to follow-up or the end of Extension 1. We performed a sensitivity analysis with the WHI observational study participants alone using Cox Proportional Hazard Regression and Fine and Gray Competing Risk Regression.
A p-value <0.05 was considered statistically significant. Statistical analyses were conducted using SAS, version 9.4 (SAS Institute Inc, Cary, NC).
Results
Study Sample
The study sample consisted of 4,640 stroke survivors with a GOS status of 1–3 and surviving more than 1-week post-stroke. Table 1 summarizes the baseline characteristics. The mean age at the time of stroke was 74.6±7.2 years. The mean age at the time of hip fracture was 77.3±6.2 years. Forty one percent of participants were discharged with good recovery, 30% with moderate disability, and 29% with severe disability. The majority of stroke survivors (84%) were white. Eighty-eight percent of strokes were ischemic and 11.8% were hemorrhagic. Thirteen cases did not have a stroke type classification (0.3%).
Table 1.
Baseline characteristics of 4640 postmenopausal women with stroke and post-stroke hip fracture within 5-years after stroke in the Women’s Health Initiative Study (1993–2010).
| All (n=4,640) | Hip Fracture (n=124) | |
|---|---|---|
| Age, mean (SD) | 74.56 (7.21) | 77.29 (6.24) |
| ≤65 years, (%) | 489 (10.54) | 5 (4.03) |
| 65–74 years, (%) | 1822 (39.27) | 32 (25.81) |
| ≥75 years, (%) | 2329 (50.19) | 87 (70.16) |
| White, n(%) | 3910 (84.27) | 118 (95.16) |
| BMI, kg/m2, mean, n(%) | 28.31 (5.93) | 26.98 (5.01) |
| Normal | 1459 (31.44) | 43 (34.68) |
| Overweight/Obese | 3139 (67.65) | 79 (63.71) |
| Glasgow Outcome Scale, mean, n(%) | ||
| Good Recovery | 1917 (41.31) | 45 (36.29) |
| Moderately Disabled | 1400 (30.17) | 34 (27.42) |
| Severely Disabled | 1323 (28.51) | 45 (36.29) |
| Stroke Type | ||
| Ischemic | 4082 (87.97) | 111 (89.52) |
| Hemorrhagic/Other | 545 (11.75) | 13 (10.48) |
| FRAX-Hip Risk Score, mean (SD) | 3.40 (4.39) | 6.26 (6.36) |
| ≥3.0% | 1719 (37.05) | 80 (64.52) |
| Reported Osteoporosis Diagnosis, n(%) | 1451 (31.27) | 46 (37.10) |
| Parental Hip Fracture | 573 (12.35) | 22 (17.74) |
SD: Standard Deviation
BMI: Body Mass Index; FRAX: Fracture Risk Assessment Tool
Mean follow-up was 3.1±1.8 years (range: 8–1,826 days). Hip fractures occurred in 2.7% of our study sample (124/4,640). Hip fractures by GOS status occurred in 2.3% (45/1917), 2.4% (34/1400), and 3.4% (45/1323) among good recovery, moderately disabled, and severely disabled, respectively. Death occurred in 23.3% (1079/4640) of our sample before the end of Ext1 or the occurrence of a hip fracture. Deaths by GOS status were as follows: good recovery 15.8% 302/1917), moderately disabled 21.1% (295/1400), and severely disabled 36.4% (482/1323).
Cumulative Risk of Hip Fracture
Cumulative risk of hip fracture, without the competing risk of death, is summarized in Table 2. Overall, cumulative risk of hip fracture was 0.07% at 30 days, 1% at 1 year, and 4% at 5 years. Survival-free from hip fracture differed by disability status (p=0.003) and older age (p=0.002), but not stroke type (p=0.8) (Table 2). Incorporating the competing risk of post-stroke mortality into the overall model slightly lessened the incidence at 30 days (0.06%), 1 year (0.9%), and 5 years (3%) (Table 3). Only age category remained significantly associated with fracture risk at 5-years, after accounting for death (p=0.006, Table 3).
Table 2.
Cumulative risk of hip fracture (n=4,640) 30-days, 1-year and 5-years after stroke by stroke type, Glasgow Outcome Scale, and age group in the Women’s Health Initiative Study (1993–2010).
| 30 days | 1 year | 5 years | |
|---|---|---|---|
| All Stroke | 0.07% | 0.99% | 3.97% |
| Ischemic Stroke | 0.07% | 1.01% | 3.99% |
| Other Stroke | 0.00% | 0.82% | 4.00% |
| Good Recovery | 0.00% | 0.50% | 3.28% |
| Moderate Disability | 0.07% | 0.78% | 3.58% |
| Severe Disability | 0.16% | 2.05% | 5.72%* |
| <65 years | 0.00% | 0.43% | 1.16% |
| ≥65 years | 0.07% | 1.06% | 4.44%* |
logrank: p<0.05
Table 3.
Cumulative incidence of hip fracture with the competing risk of post-stroke mortality 30-days, 1-year, and 5-years after stroke by stroke type, Glasgow Outcome Scale, and age group in the Women’s Health Initiative Study (1993–2010; n=4,640).
| 30 days | 1 year | 5 years | |
|---|---|---|---|
| Overall Stroke | 0.06% | 0.92% | 3.33% |
| Ischemic Stroke | 0.07% | 0.94% | 3.36% |
| Other Stroke | 0.00% | 0.76% | 3.17% |
| Good Recovery | 0.00% | 0.49% | 2.96% |
| Moderate Disability | 0.07% | 0.74% | 3.04% |
| Severe Disability | 0.15% | 1.74% | 4.18% |
| <65 years | 0.00% | 0.42% | 1.10% |
| ≥ 65 years | 0.07% | 0.98% | 3.65%* |
logrank: p<0.05
Predictors of post-stroke hip fracture with and without competing risk of death
The results of the Cox proportional hazard survival models and the competing risk regression models are in Table 4. Participants with missing covariate information were excluded (n=129), leaving 4,511 stroke survivors available for analyses. Severe disability (Hazard Ratio (HR): 2.1 (95% CI: 1.4–3.2)), but not moderate disability (HR: 1.1 (95% CI: 0.7–1.7)), was significantly associated with an increase in risk of hip fracture compared to a good recovery status. High FRAX without BMD (≥3% risk) compared to low FRAX without BMD-predicted hip fracture risk (<3% risk) (HR: 1.9 (95%CI: 1.2–3.0)) and increasing age (HR: 1.1 (95%CI: 1.0–1.1)) were associated with an increased risk of hip fracture. Non-white race (HR: 0.4 (95%CI: 0.2–1.0)) was significantly associated with a decrease in hip fracture risk.
Table 4.
Cox Proportional Hazard (1) and Fine and Gray Competing Risk (2) Models of the predictors of post-stroke hip fracture risk among postmenopausal stroke survivors in the Women’s Health Initiative (1993–2010).
| Variable | Model 1 Adjusted HR | 95% CI | Model 2 Adjusted HR | 95% CI |
|---|---|---|---|---|
| GOS Category | ||||
| Good Recovery | Reference | - | Reference | - |
| Moderate Disability | 1.07 | 0.67–1.70 | 1.04 | 0.66–1.64 |
| Severe Disability | 2.10* | 1.37–3.23 | 1.51 | 0.98–2.33 |
| Age, years | 1.06* | 1.02–1.10 | 1.03 | 1.00–1.06 |
| Race/ethnicity | ||||
| White | Reference | - | Reference | - |
| Non-White | 0.41* | 0.18–0.96 | 0.42* | 0.18–0.99 |
| FRAX without BMD Risk Score | ||||
| <3%, low risk | Reference | - | Reference | - |
| ≥3%, high risk | 1.87* | 1.18–2.98 | 2.05* | 1.25–3.36 |
| Calcium/Osteoporosis medication use | 1.26 | 0.85–1.86 | 1.22 | 0.82–1.80 |
| Depressive Symptoms Score (CES-D) | 0.92 | 0.14–6.03 | 0.80 | 0.14–4.60 |
| History of Diabetes | 0.97 | 0.60–1.58 | 0.97 | 0.60–1.58 |
| Smoking Status | ||||
| Never | Reference | - | Reference | - |
| Former | 1.10 | 0.76–1.61 | 1.07 | 0.73–1.56 |
| Current | 1.28 | 0.61–2.67 | 1.06 | 0.52–2.16 |
Model 1: Cox Proportional Hazard Model
Model 2: Fine and Gray Competing Risk Model
HR: Hazards Ratio; CI: Confidence Interval
GOS: Glasgow Outcome Scale; FRAX: Fracture Risk Assessment Tool; BMD: bone mineral density; CES-D: Center for Epidemiological Studies Depression Scale
p<0.05
In the competing risk model, the association was similar and not significant for moderate disability (HR: 1.0 (95%CI: 0.7–1.6)) and somewhat attenuated and did not reach level of significance for severe disability (HR 1.5 (95%CI: 1.0–2.3)) after we accounted for mortality. However high FRAX risk score (HR: 2.1 (95%CI: 1.3–3.4)) and non-white race/ethnicity (HR: 0.4 (95%CI: 0.2–1.0)) remained significant. The association between age and hip fracture was attenuated and not significant after accounting for post-stroke mortality (HR: 1.0 (95%CI: 1.0–1.1)).
There was a significant interaction (p<0.05) between age and GOS status in both the competing risk (p=0.0002) and Cox proportional hazards models (p=0.01). Stratified results by mean age (≤74.56 vs >74.56) are described in Supplemental Table I. In the OS-only analysis, severely disabled and high FRAX hip fracture probability were significantly associated with increased risk of post-stroke hip fracture in the Cox proportional hazard model and the Fine and Gray competing risk model (Supplemental Table II).
Discussion
The risk of hip fracture in female stroke survivors in the WHI was 1% at 1-year and 4% at 5-years. These are comparable to results from other prospective studies on stroke survivors and retrospective cohort and claims studies.(2, 3, 7, 28, 29) Among stroke survivors, severe disability and a high FRAX-predicted hip fracture risk were associated with a higher risk of subsequent hip fracture. However, only the FRAX-predicted hip fracture risk association remained significant after we accounted for post-stroke mortality in the competing risk model.
Post-stroke fracture risk is associated with a decrease in bone mineral density and an increased susceptibility to falls. The affected side is most impacted by this bone loss due to loss of mobility resulting in a decrease in skeletal loading and an increase in osteoclast resorption rate. Approximately 12–17% of bone mineral density on the affected side is lost within 1 year after stroke.(3, 11, 14, 15, 30–32) There appears to be little variability in bone mineral density on the unaffected side.(2, 11, 15, 30–32) Additionally, post-menopausal women may have lower bone mineral density.(33, 34)
Studies have reported an inverse correlation between the change in bone density and the functional status of the stroke survivor.(15, 35–37) Worse functional status or decreased mobility is associated with a greater bone density loss.(15, 35–37) Our study adds to this literature by separately quantifying this risk by the degree of post-stroke disability status, and adjusting for FRAX-predicted hip fracture risk, calcium supplement and osteoporosis medication use, and prior osteoporosis diagnosis among an all-female cohort. Few studies have incorporated a post-stroke disability status within risk models.(2, 7, 11–13) To our knowledge, none has used the GOS measurement. However, moderate to high correlations between the modified Rankin Scale (mRS) and GOS have been reported.(38, 39) We found that those with severe disability, i.e., a higher GOS status, had the highest risk of hip fractures, but the association was attenuated and did not reach significance after accounting for the competing risk of mortality. Previous studies with both male and female participants, report no significant association between hip fracture risk and stroke severity; however, there was an association between severity with overall fracture risk.(2, 11)
Our study adds to the current literature by accounting for osteoporosis-related variables, such as medication use and pre-stroke FRAX risk. We showed that a high FRAX risk was associated with an increased risk for a post-stroke hip fracture, even after considering the competing risk of mortality. Currently, very few stroke survivors are evaluated and treated for fracture and/or osteoporosis prevention upon discharge.(40)
A recent large study (n=16,581) reported that stroke survivors were infrequently screened and treated for osteoporosis with only 5% undergoing bone mineral density screening.(40) While the American Heart/Stroke Association guidelines discuss fall prevention in the context of stroke, they adopt the recommendation of the US Preventive task force (USPTF) for fracture preventions in the general population. (41) Current USPTF guidelines do not include stroke in their risk stratification, though literature indicates that stroke survivors have a significant risk of fractures.(2–6, 11, 42) Fracture risk assessment tools, including FRAX, can predict 10-year fracture risk with and without BMD. Our results suggest that stroke units can consider the incorporation of osteoporosis screening into care pathways. The use of formal tools such as FRAX scores and subsequent Endocrinology follow-up based on risk assessment may be part of the pathway.
Our results found a significant interaction between age and GOS status. Specifically, younger post-menopausal women who were discharged as severely disabled were at a significantly higher risk of having a post-stroke hip fracture. There was no significant difference among older women by GOS status. This is consistent with other studies that found younger stroke survivors at increased risk of hip fracture(8, 30); however, these studies did not adjust for a disability status. Future prospective studies to explore different predictors between younger and older post-menopausal stroke survivors and risk of hip fracture will be useful.
While not statistically significant, there was an increased risk of hip fracture among those who had a history of taking calcium supplements or osteoporosis medication before the time of stroke. This may be due to medication use being a marker of osteoporosis or higher fracture risk before the index stroke event. Our results may point to reverse causation. Hence a prospective assessment whether increased screening and treatment significantly changes fracture risk is warranted.
Lastly, being of non-white race was protective of hip fracture in both models. This result is consistent with previous post-stroke(2) and general population hip fracture studies(43) and may be due to differences in bone mineral densities,(44) bone size and geometry.
We acknowledge the following limitations. Information on the side of hip fracture and the location of the neurological deficit was unavailable. Additionally, there was no information whether the fracture resulted from a fall. Falls could result from worse post-stroke disability, and hence mediate the increased risk after stroke. We were able to account for fracture probability through the FRAX-predicted hip fracture risk. We did not have any specific data on bone mineral density before and after the indexed stroke. Therefore, we could not assess post-stroke bone loss.
We included data from both clinical trial arms and the observation study arm and addressed any effect of the clinical trial treatment on our results through stratification. We acknowledge that disability status was assessed at discharge and post-stroke recovery could change participant functional status. This may be the case particularly for those discharged to acute rehabilitation. While this is reasonable based on the typical trajectory of post-stroke recovery, the WHI adjudication instructions took this into account. Specifically, if the medical records of patients discharged to acute rehabilitation provided information that they were expected to function independently after discharge from rehabilitation, this was taken into account in the disability classification. Our results indicate a separation between the good recovery vs. severely disabled category in hip fracture risk but not between the good recovery and the moderately disabled category. This suggests that the expected lack of functional independence may be key in predicting hip fracture risk. Lastly, we acknowledge that this study is not a representative sample of all women and therefore, cannot be generalized to men or the general population.
Conclusion
Our study of female stroke survivors estimates the risk of post-stroke hip fracture in a large cohort of prospectively followed patients. A post-stroke fracture assessment tool such as FRAX risk score may be beneficial in identifying those at high risk of fractures post-stroke.
Supplementary Material
Acknowledgements
The authors thank the WHI investigators and staff for their dedication, and the study participants for making the program possible. A full listing of WHI investigators can be found at: https://www.whi.org/researchers/Documents%20%20Write%20a%20Paper/WHI%20Investigator%20Long%20List.pdf
Sources of Funding:
The WHI program is funded by the National Heart, Lung, and Blood Institute, National Institutes of Health, U.S. Department of Health and Human Services through contracts HHSN268201600018C, HHSN268201600001C, HHSN268201600002C, HHSN268201600003C, and HHSN268201600004C.
C.A.N. was supported by the National Heart, Lung, and Blood Institute of the National Institutes of Health under Award Number T32HL007779. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
Footnotes
Disclosures
KL receives consultation income from the Ohio State University for continued research using WHI Study data.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
References
- 1.Benjamin EJ, Virani SS, Callaway CW, et al. Heart Disease and Stroke Statistics-2018 Update: A Report From the American Heart Association. Circulation 2018;137(12):e67–e492. [DOI] [PubMed] [Google Scholar]
- 2.Lisabeth LD, Morgenstern LB, Wing JJ, et al. Poststroke fractures in a bi-ethnic community. J Stroke Cerebrovasc Dis 2012;21(6):471–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Ramnemark A, Nyberg L, Borssén B, Olsson T, Gustafson Y. Fractures after stroke. Osteoporos Int 1998;8(1):92–5. [DOI] [PubMed] [Google Scholar]
- 4.Kerse N, Parag V, Feigin VL, et al. Falls after stroke: results from the Auckland Regional Community Stroke (ARCOS) Study, 2002 to 2003. Stroke 2008;39(6):1890–3. [DOI] [PubMed] [Google Scholar]
- 5.Callaly EL, Ni Chroinin D, Hannon N, et al. Falls and fractures 2 years after acute stroke: the North Dublin Population Stroke Study. Age Ageing 2015;44(5):882–6. [DOI] [PubMed] [Google Scholar]
- 6.Whitson HE, Pieper CF, Sanders L, Horner RD, Duncan PW, Lyles KW. Adding injury to insult: fracture risk after stroke in veterans. J Am Geriatr Soc 2006;54(7):1082–8. [DOI] [PubMed] [Google Scholar]
- 7.Dennis MS, Lo KM, McDowall M, West T. Fractures after stroke: frequency, types, and associations. Stroke 2002;33(3):728–34. [DOI] [PubMed] [Google Scholar]
- 8.Pouwels S, Lalmohamed A, Leufkens B, et al. Risk of hip/femur fracture after stroke: a population-based case-control study. Stroke 2009;40(10):3281–5. [DOI] [PubMed] [Google Scholar]
- 9.Kelly-Hayes M, Beiser A, Kase CS, Scaramucci A, D’Agostino RB, Wolf PA. The influence of gender and age on disability following ischemic stroke: the Framingham study. J Stroke Cerebrovasc Dis 2003;12(3):119–26. [DOI] [PubMed] [Google Scholar]
- 10.Benjamin EJ, Blaha MJ, Chiuve SE, et al. Heart Disease and Stroke Statistics-2017 Update: A Report From the American Heart Association. Circulation 2017;135(10):e146–e603. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Kapral MK, Fang J, Alibhai SM, et al. Risk of fractures after stroke: Results from the Ontario Stroke Registry. Neurology 2017;88(1):57–64. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Melton LJ, Brown RD, Achenbach SJ, O’Fallon WM, Whisnant JP. Long-Term fracture risk following ischemic stroke: a population-based study. Osteoporos Int 2001;12(11):980–6. [DOI] [PubMed] [Google Scholar]
- 13.Benzinger P, Rapp K, König HH, et al. Risk of osteoporotic fractures following stroke in older persons. Osteoporos Int 2015;26(4):1341–9. [DOI] [PubMed] [Google Scholar]
- 14.Marsden J, Gibson LM, Lightbody CE, Sharma AK, Siddiqi M, Watkins C. Can early onset bone loss be effectively managed in post-stroke patients? An integrative review of the evidence. Age Ageing 2008;37(2):142–50. [DOI] [PubMed] [Google Scholar]
- 15.Jørgensen L, Jacobsen BK. Functional status of the paretic arm affects the loss of bone mineral in the proximal humerus after stroke: a 1-year prospective study. Calcif Tissue Int 2001;68(1):11–5. [DOI] [PubMed] [Google Scholar]
- 16.Design of the Women’s Health Initiative Clinical Trial and Observational Study. Controlled Clinical Trials 1998;19(1):61–109. [DOI] [PubMed] [Google Scholar]
- 17.Albanese MA, Clarke WR, Adams HP, Woolson RF. Ensuring reliability of outcome measures in multicenter clinical trials of treatments for acute ischemic stroke. The program developed for the Trial of Org 10172 in Acute Stroke Treatment (TOAST). Stroke 1994;25(9):1746–51. [DOI] [PubMed] [Google Scholar]
- 18.Curb JD, McTiernan A, Heckbert SR, et al. Outcomes ascertainment and adjudication methods in the women’s health initiative. Annals of Epidemiology 2003;13(9):S122–S8. [DOI] [PubMed] [Google Scholar]
- 19.Seth A, Mossavar-Rahmani Y, Kamensky V, et al. Potassium intake and risk of stroke in women with hypertension and nonhypertension in the Women’s Health Initiative. Stroke 2014;45(10):2874–80. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Chen JC, Brunner RL, Ren H, et al. Sleep duration and risk of ischemic stroke in postmenopausal women. Stroke 2008;39(12):3185–92. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Kanis JA, Johnell O, Oden A, Johansson H, McCloskey E. FRAX and the assessment of fracture probability in men and women from the UK. Osteoporos Int 2008;19(4):385–97. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Radloff LS. The CES-D Scale: A Self-Report Depression Scale for Research in the General Population. Applied Psychological Measurement 1977;1(3):385–401. [Google Scholar]
- 23.Cosman F, de Beur SJ, LeBoff MS, et al. Clinician’s Guide to Prevention and Treatment of Osteoporosis. Osteoporos Int 2014;25(10):2359–81. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Outcomes: Cardiovascular Outcomes. WHI Study Manual. Version: 10/30/08 ed: Women’s Health Iniative;66–7. [Google Scholar]
- 25.Fine JP, Gray RJ. A Proportional Hazards Model for the Subdistribution of a Competing Risk. Journal of the American Statistical Association 1999;94(446):496–509. [Google Scholar]
- 26.Austin PC, Lee DS, Fine JP. Introduction to the Analysis of Survival Data in the Presence of Competing Risks. Circulation 2016;133(6):601–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Noordzij M, Leffondré K, van Stralen KJ, Zoccali C, Dekker FW, Jager KJ. When do we need competing risks methods for survival analysis in nephrology? Nephrol Dial Transplant 2013;28(11):2670–7. [DOI] [PubMed] [Google Scholar]
- 28.Brown DL, Morgenstern LB, Majersik JJ, Kleerekoper M, Lisabeth LD. Risk of fractures after stroke. Cerebrovasc Dis 2008;25(1–2):95–9. [DOI] [PubMed] [Google Scholar]
- 29.Lin HL, Lin HC, Tseng YF, et al. Hip fracture after first-ever stroke: a population-based study. Acta Neurol Scand 2015;131(3):158–63. [DOI] [PubMed] [Google Scholar]
- 30.Kanis J, Oden A, Johnell O. Acute and long-term increase in fracture risk after hospitalization for stroke. Stroke 2001;32(3):702–6. [DOI] [PubMed] [Google Scholar]
- 31.Ramnemark A, Nyberg L, Lorentzon R, Englund U, Gustafson Y. Progressive hemiosteoporosis on the paretic side and increased bone mineral density in the nonparetic arm the first year after severe stroke. Osteoporos Int 1999;9(3):269–75. [DOI] [PubMed] [Google Scholar]
- 32.Levendoglu F, Ugurlu H, Gürbilek M, Akkurt E, Karagözolu E. Increased bone resorption in the proximal femur in patients with hemiplegia. Am J Phys Med Rehabil 2004;83(11):835–41. [DOI] [PubMed] [Google Scholar]
- 33.Bainbridge NJ, Davie MW, Haddaway MJ. Bone loss after stroke over 52 weeks at os calcis: influence of sex, mobility and relation to bone density at other sites. Age Ageing 2006;35(2):127–32. [DOI] [PubMed] [Google Scholar]
- 34.del Puente A, Pappone N, Mandes MG, Mantova D, Scarpa R, Oriente P. Determinants of bone mineral density in immobilization: a study on hemiplegic patients. Osteoporos Int 1996;6(1):50–4. [DOI] [PubMed] [Google Scholar]
- 35.Lazoura O, Groumas N, Antoniadou E, et al. Bone mineral density alterations in upper and lower extremities 12 months after stroke measured by peripheral quantitative computed tomography and DXA. J Clin Densitom 2008;11(4):511–7. [DOI] [PubMed] [Google Scholar]
- 36.Pang MY, Eng JJ. Muscle strength is a determinant of bone mineral content in the hemiparetic upper extremity: implications for stroke rehabilitation. Bone 2005;37(1):103–11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Beaupre GS, Lew HL. Bone-density changes after stroke. Am J Phys Med Rehabil 2006;85(5):464–72. [DOI] [PubMed] [Google Scholar]
- 38.Lees KR, Bath PM, Schellinger PD, et al. Contemporary outcome measures in acute stroke research: choice of primary outcome measure. Stroke 2012;43(4):1163–70. [DOI] [PubMed] [Google Scholar]
- 39.Kim DH, Haney CL, Van Ginhoven G. Utility of outcome measures after treatment for intracranial aneurysms: a prospective trial involving 520 patients. Stroke 2005;36(4):792–6. [DOI] [PubMed] [Google Scholar]
- 40.Kapoor E, Austin PC, Alibhai SMH, et al. Screening and Treatment for Osteoporosis After Stroke. Stroke 2019;50(6):1564–6. [DOI] [PubMed] [Google Scholar]
- 41.Winstein CJ, Stein J, Arena R, et al. Guidelines for Adult Stroke Rehabilitation and Recovery: A Guideline for Healthcare Professionals From the American Heart Association/American Stroke Association. Stroke 2016;47(6):e98–e169. [DOI] [PubMed] [Google Scholar]
- 42.Screening for osteoporosis: U.S. preventive services task force recommendation statement. Ann Intern Med 2011;154(5):356–64. [DOI] [PubMed] [Google Scholar]
- 43.Silverman SL, Madison RE. Decreased incidence of hip fracture in Hispanics, Asians, and blacks: California Hospital Discharge Data. Am J Public Health 1988;78(11):1482–3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Looker AC, Orwoll ES, Johnston CC, et al. Prevalence of low femoral bone density in older U.S. adults from NHANES III. J Bone Miner Res 1997;12(11):1761–8. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.