Low-trauma fracture increases 12-month incidence of cardiovascular disease for adults with cerebral palsy

Abstract

Individuals with cerebral palsy (CP) have poor skeletal and cardiovascular health; however, no studies have examined if skeletal fragility enhances cardiovascular disease (CVD) risk for this population. The purpose of this study was to determine whether adults with CP have higher 12-month CVD incidence following a low-trauma fracture compared to adults without CP. Data, from the Optum Clinformatics® Data Mart, were extracted from adults (18+ years) that sustained a low-trauma fracture between 01/01/2012–12/31/2016. The primary outcome measure was incident CVD within 12 months following a low-trauma fracture. Cox proportional hazards regression models were used to compare 12-month incident CVD with adjustment for sociodemographics and chronic disease comorbidities. Mean age (SD) at baseline was 54.7 (18.9) for adults with CP (n=1,025, 43.3% men) and 60.4 (19.7) for adults without CP (n=460,504, 33.7% men). During the follow-up, 121 adults with CP (11.8%, mean age [SD]=63.9 [16.3]) and 45,330 adults without CP (9.8%, mean age [SD]=74.5 [11.9]) developed CVD. In the fully adjusted model, adults with CP had higher 12-month post-fracture CVD incidence (hazard ratio [HR]=1.63; 95% confidence interval [CI]=1.37–1.95). When the outcome was stratified by CVD subtype, adults with CP had higher 12-month post-fracture incidence of ischemic heart disease (HR=1.45; 95% CI=1.09–1.92), heart failure (HR=1.68; 95% CI=1.22–2.31), and cerebrovascular disease (HR=1.96; 95% CI=1.54–2.50). Study findings suggest that among adults with CP, low-trauma fracture may enhance 12-month CVD incidence compared to adults without CP.

Introduction

Cerebral palsy (CP) is the most common pediatric physical disability¹ caused by damage to or malformation of the developing brain. Secondary complications of CP include neurological and neuromuscular impairments, preventing optimal fulfillment of motor functional capacity and mechanical loading.^{2, 3} As a result, children with CP are predisposed to inadequate development of musculoskeletal mass, architecture, and metabolic health throughout growth, regardless of condition severity,^2–5 leading to a heightened susceptibility for low-trauma fracture for this pediatric population.^{6, 7} The secondary complications of skeletally fragility among individuals with CP persists throughout the lifespan. Young adults (18–30 years) with CP are approximately ten times more likely to have a musculoskeletal disease (e.g., osteoporosis, osteoarthritis) compared to young adults without CP.⁸ Further, the prevalence of osteoporosis increases throughout the lifespan to about 26% for adults with CP >50 years of age,⁹ which is 2.5 times higher than the general population of adults >50 years of age.¹⁰ This increases risk for fracture, which is more than two-fold higher among young and middle-aged adults (18–64 years) with vs. without CP, even after accounting for osteoporosis and other confounding factors.¹¹

Despite documented evidence showing a concerning skeletally fragile phenotype, little is known about the contribution of skeletal fragility to health and survival outcomes for adults with CP. This is important because adults with CP also have increased risk for high burden chronic diseases across physiological systems (e.g., neural,^12–14 kidney, liver¹²): especially cardiovascular disease (CVD)^{8, 12} and CVD-related mortality.^{15, 16} CVD is a leading contributor to disease burden globally and in North America,¹⁷ and was responsible for 17.6 million deaths worldwide in 2016.¹⁸ Among other skeletally vulnerable populations (e.g., elderly, postmenopausal women), fracture is a major cause of morbidity,¹⁹ poor quality of life,²⁰ and premature mortality,^{21, 22} and is implicated in the pathogenesis of CVD.¹⁹

In the general population, the temporal sequence of developing skeletal fragility and CVD can go both ways, in that CVD can increase fracture risk²³ and skeletal fragility can increase CVD risk.¹⁹ For individuals with CP, poor musculoskeletal health²⁴ typically precedes the development of CVD risk factors²⁵ in children. In adults with CP, the magnitude of the disease disparity compared to adults without CP is far greater for skeletal fragility than CVD.^{8, 12} It is possible that skeletal fragility initiates or exacerbates the development of CVD among adults with CP through inadequate mechanical loading, a unique physiological environment (i.e., biological factors released from active and inactive musculoskeletal and adipose tissue), and/or low physical activity. However, this has not been adequately investigated. Understanding if skeletal fragility increases CVD risk among adults with CP could lead to improved post-fracture healthcare management and reduced CVD burden.^{12, 15} Therefore, the primary objective of this study was to compare CVD incidence in the 12 months following a low-trauma fracture between adults with vs. without CP. We hypothesized that adults with CP would have a higher post-fracture 12-month CVD incidence compared to adults without CP, even after adjusting for pre-fracture chronic diseases.

Methods

This study is a retrospective cohort study with Level 3 evidence. Since data are de-identified, the university Institutional Review Board approved this study as non-regulated.

Data source

Clinformatics® Data Mart Database (OptumInsightTM, Eden Prairie, MN, USA) is a U.S. nationwide de-identified single private payer administrative claims database. Data from 2011 to 2017 were extracted. This claims-based data includes all health service utilization (e.g., inpatient, emergency department) for each individual throughout enrollment. To maintain patient confidentiality, researchers leveraging this database are allowed either the Date of Death or Socioeconomic Status table. The current investigation was developed under a larger project in which the Date of Death table was leveraged. Therefore, some information regarding socioeconomic status (i.e., income, education) were not available.

Sample selection

All medical conditions (e.g., CP, fracture, CVD) were identified using the International Classification of Diseases, Ninth and Tenth Revision (ICD-9 and ICD-10), Clinical Modification codes to account for the shift in reporting codes on October 1^st, 2015. Information regarding how diagnoses were made or by whom (e.g., primary care physician) is not available in claims data.

Adults ≥18 years of age that had at least one claim in any position for a low-trauma fracture between 2012 to 2016 of the vertebral column, hip (including proximal femur), non-proximal femur, tibia/fibula, humerus, ulna/radius, or unspecified location were identified. Low-trauma fracture was defined as a fracture without trauma codes (e.g., motor vehicle accident) 7 days before to 7 days after the index fracture date, as guided by previous studies.^{26, 27} The single claim-based definition has excellent accuracy (up to 98% positive predictive value) for identifying fractures, which is similar or better than other algorithms (e.g., 2+ claims).²⁸

Individuals were included if they had at least 12 full months of enrollment in a health plan prior to their low-trauma fracture index date. This was done for two reasons. First, it is not possible to determine lifetime history of fracture from claims data and many individuals can have breaks in their health plan enrollment, making longitudinal research designs with long follow-up periods challenging. Therefore, this 12-month pre-fracture period was used to exclude individuals that had any fracture to ascertain data from individuals with a relatively new fracture. Second, this 12-month pre-fracture period was used to sequester baseline chronic disease data.²⁹ Individuals were excluded if they had at least one claim for the outcome measure within the 12 months prior to the low-trauma fracture index date.

Individuals with CP were identified by at least one claim which covered all diagnostic subtypes (e.g., spasticity, quadriplegic, athetoid). Unfortunately, information regarding the severity of CP using common clinical measures (e.g., gross motor function classification system) are not available in claims data, and more than 70% had “other” or “unspecified” CP.¹² This limits the current study as stratification or statistical adjustment for the clinical subtypes of CP is not possible. The comparison group included individuals with no claims for CP. The single claim-based definition has good accuracy for identifying pediatric-onset conditions using claims data with 99% sensitivity and a positive predictive value of 79%.³⁰

Outcome measure

Post-fracture incidence of CVD was defined by at least one claim in any position for ischemic heart disease, heart failure, or cerebrovascular disease between 14 and 365 days after the index date of low-trauma fracture. These diseases were selected as they are serious, life-threatening conditions and are among the major contributors to global CVD burden¹⁷ and mortality.¹⁸ The latent period of 2 weeks was selected to omit individuals that were diagnosed with CVD around the time of fracture, which may have been due to medical screening brought on by hospitalization from the fracture event. In these cases, the CVD may have been present prior to the fracture. Ischemic heart disease included angina pectoris, myocardial infarction, other acute ischemic heart disease (e.g., coronary thrombosis), and chronic ischemic heart disease (e.g., atherosclerosis of coronary arteries, aneurysm). Cerebrovascular disease included intracranial hemorrhage, cerebral infarction, occlusion and stenosis of precerebral and cerebral arteries, and “other” cerebrovascular disease.

A single claim was used to identify CVD given the short study follow-up period of up to 12 months. Using at least one claim in any position has shown excellent accuracy for identifying ischemic heart disease, heart failure, and cerebrovascular disease (96–98% positive predictive value) in claims-based data.³¹

Covariates

Covariates were selected based on their relevance to adults with CP, fracture, and CVD, as well as availability in the administrative claims database. Sociodemographic variables included age, sex, race, and U.S. region of residence (West, Midwest, South, and Northeast). Baseline chronic diseases were identified by at least two different claim days (i.e., same claim over two or more different days). This claims-based definition improves accuracy of identifying individuals over a single claim. The date of the first claim had to be within 12 months prior to the index date of low-trauma fracture. Chronic diseases included hypertension (i.e., medical diagnosis or a single claim for a prescription fill for antihypertension medication: diuretics, renin inhibitors, renin-angiotensin-aldosterone system inhibitor, beta-adrenergic blocking agents, or calcium-channel blocking agents), diabetes (i.e., type 1 or type 2), and chronic obstructive pulmonary diseases (i.e., chronic bronchitis, emphysema, and “other” chronic obstructive pulmonary diseases).

Statistical analysis

Baseline descriptive characteristics for the entire sample and for the sample that developed a CVD during the 12-month follow-up were summarized. Group differences between adults with and without CP were examined using independent t-tests for continuous variables and Chi-square tests for categorical variables.

Cox proportional hazards regression models were developed to compare 12-month post-fracture CVD incidence between adults with and without CP. Unadjusted models were not examined as we anticipated that CVD would be developed substantially earlier for adults with CP.⁸ Model 1 adjusted for age. Model 2 adjusted for age, sex, and U.S. region. Lastly, model 3 adjusted for the variables in model 2 and chronic diseases. Individuals were right censored if they died or discontinued enrollment within the follow-up period or did not develop CVD at the end of the follow-up period.

Separate Cox proportional hazards regression models were also developed for each CVD subtype, adjusting for sociodemographic covariates. CP group by age and sex interactions were examined. Where significant, additional analyses were performed after stratifying by the appropriate demographic variable. The main effect of CP group (CP vs. without CP) was interpreted.

Cox proportional hazards regression models did not adjust for race to limit bias because of the extent of missingness/unknown (~15%). We therefore conducted sensitivity analyses that further adjusted all models for race among individuals with complete data.

We considered statistical significance as P≤0.05 (two-tailed) and effect estimates were reported as hazard ratios (HR) with 95% confidence intervals (CI). Analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC, USA).

Results

Baseline descriptive characteristics of adults with (n=1,025) and without (n=460,504) CP that sustained a low-trauma fracture are presented in Table 1. Compared to adults without CP, adults with CP were 5.7 years younger on average and were more likely to be male (43.3% vs. 33.7%) (both, P<0.001). Over the 12-month study period, 121 adults with CP (11.8%) and 45,330 adults without CP (9.8%) developed CVD. Among those that developed CVD (Table 2), adults with CP were 10.6 years younger on average and were more likely to be male (44.6% vs. 29.7%) (both, P<0.001).

Table 1.

Baseline descriptive characteristics of privately insured adults with and without cerebral palsy (CP) that sustained a low-trauma fracture.

	With CP (n=1,025)	Without CP (n=460,504)
	No. (%)	No. (%)	P value
Demographic characteristics
Age, mean (SD)	54.7 (18.9)	60.4 (19.7)	<0.001
Sex			<0.001
Women	581 (56.7)	305,126 (66.3)
Men	444 (43.3)	155,378 (33.7)
Race			0.018
White	676 (66.0)	309,777 (67.3)
Black	96 (9.4)	32,405 (7.0)
Hispanic	75 (7.3)	38,226 (8.3)
Asian	19 (1.9)	11,997 (2.6)
Unknown/missing	159 (15.5)	68,099 (14.8)
US region			0.635
West	301 (29.4)	129,845 (28.2)
Midwest	249 (24.3)	112,095 (24.3)
South	367 (35.8)	172,873 (37.5)
Northeast	108 (10.5)	45,691 (9.9)
1 st Fracture characteristics
Fracture distribution			<0.001
Unspecified location	24 (2.3)	19,228 (4.2)
Vertebral column	233 (22.7)	105,330 (22.9)
Hip	217 (21.3)	64,024 (13.9)
Femur, non-proximal	64 (6.2)	11,676 (2.5)
Tibia/fibula	277 (27.0)	118,549 (25.7)
Humerus	102 (10.0)	45,524 (9.9)
Ulna/radius	108 (10.5)	96,173 (20.9)
Chronic disease comorbidities
Hypertension	437 (42.6)	204,571 (44.4)	0.250
Diabetes	132 (12.9)	68,484 (14.9)	0.073
Chronic obstructive pulmonary diseasepulmonary disease	92 (9.0)	36173 (7.9)	0.183

Table 2.

Baseline descriptive characteristics of privately insured adults with and without cerebral palsy (CP) that developed cardiovascular disease during follow-up.

	With CP (n=121)	Without CP (n=45,330)
	No. (%)	No. (%)	P value
Demographic characteristics
Age, mean (SD)	63.9 (16.3)	74.5 (11.9)	<0.001
Sex			<0.001
Women	67 (55.4)	31,864 (70.3)
Men	54 (44.6)	13,466 (29.7)
Race			0.294
White	76 (62.8)	29,907 (66.0)
Black	13 (10.7)	3,419 (7.5)
Hispanic	12 (9.9)	3,766 (8.3)
Asian	0 (0)	1,015 (2.2)
Unknown/missing	20 (16.5)	7,223 (15.9)
US region			0.022
West	29 (24.0)	14,142 (31.2)
Midwest	23 (19.0)	9,872 (21.8)
South	46 (38.0)	16,330 (36.0)
Northeast	23 (19.0)	4,986 (11.0)
1stFracture characteristics
Fracture distribution			0.540
Unspecified location	2 (1.7)	1,337 (3.0)
Vertebral column	42 (34.7)	15,265 (33.7)
Hip	33 (27.3)	10,604 (23.4)
Femur, non-proximal	4 (3.3)	1,158 (2.6)
Tibia/fibula	21 (17.4)	6,655 (14.7)
Humerus	9 (7.4)	4,607 (10.2)
Ulna/radius	10 (8.3)	5,704 (12.6)
Chronic disease comorbidities
Hypertension	79 (65.3)	32,286 (71.2)	0.150
Diabetes	22 (18.2)	11,714 (25.8)	0.055
Chronic obstructive pulmonary disease	18 (14.9)	7,568 (16.7)	0.592

The adjusted HRs for 12-month post-fracture CVD incidence between adults with and without CP is presented in Table 3. The 12-month incidence of CVD was higher for adults with vs. without CP after adjusting for the variables in model 1 (HR=1.70; 95% CI=1.42–2.03), model 2 (HR=1.68; 95% CI=1.41–2.01), and model 3 (HR=1.63; 95% CI=1.37–1.95). When the outcome was stratified by CVD subtype, adults with CP had higher incidence of ischemic heart disease (model 1, HR=1.47; 95% CI=1.11–1.95: model 2, HR=1.45; 95% CI=1.09–1.92), heart failure (model 1, HR=1.69; 95% CI=1.23–2.33: model 2, HR=1.68; 95% CI=1.22–2.31), and cerebrovascular disease (model 1, HR=1.98; 95% CI=1.55–2.53: model 2, HR=1.96; 95% CI=1.54–2.50).

Table 3.

Cox proportional hazards regression to compare 12-month post-fracture cardiovascular disease incidence between adults with and without cerebral palsy (CP; n=461,529).

	Model 1	Model 2	Model 3
	HR (95% CI)	HR (95% CI)	HR (95% CI)
Any cardiovascular disease
Without CP	Reference	Reference	Reference
With CP	1.70 (1.42, 2.03)	1.68 (1.41, 2.01)	1.63 (1.37, 1.95)
Ischemic heart diseases
Without CP	Reference	Reference	*
With CP	1.47 (1.11, 1.95)	1.45 (1.09, 1.92)
Heart failure
Without CP	Reference	Reference	*
With CP	1.69 (1.23, 2.33)	1.68 (1.22, 2.31)
Cerebrovascular diseases
Without CP	Reference	Reference	*
With CP	1.98 (1.55, 2.53)	1.96 (1.54, 2.50)

HR, hazard ratio; CI, confidence interval. Model 1: age. Model 2: age, sex, and US region. Model 3: age, sex, race, US region, and baseline hypertension, diabetes, and chronic obstructive pulmonary diseases.

^*Sample size for CP group insufficient for further adjustment: n=48 for ischemic heart disease; n=38 for heart failure; n=65 for cerebrovascular disease.

The CP group by sex interaction was not significant (P=0.119), but the CP group by age interaction was significant (P<0.001). We therefore stratified age by <65 years and ≥65 years to enhance model parsimony. The 12-month incidence of CVD was higher for adults with vs. without CP after adjusting for the variables in model 2 for the <65 year age group (HR=2.39; 95% CI=1.84–3.10) and the ≥65 year age group (HR=1.30; 95% CI=1.01–1.66) (Table 4).

Table 4.

Cox proportional hazards regression to compare 12-month post-fracture cardiovascular disease incidence between adults with and without cerebral palsy (CP) stratified by age.

	<65 years of age (n=237,534)	≥65 years of age (n=223,995)
	HR (95% CI)	HR (95% CI)
Any cardiovascular disease
Without CP	Reference	Reference
With CP	2.39 (1.84, 3.10)	1.30 (1.01, 1.66)

HR, hazard ratio; CI, confidence interval. Model: age, sex, and US region.

The results of the sensitivity analysis (n=393,271) that further adjusted for race were largely consistent with the primary analyses. The conclusion was not changed when the outcome was CVD (model 3 + race, HR=1.59; 95% CI=1.31–1.93), heart failure (model 2 + race, HR=1.66; 95% CI=1.17–2.34), or cerebrovascular disease (model 2 + race, HR=1.91; 95% CI=1.46–2.49), or CVD for <65 year age group (HR=2.42; 95% CI=1.83–3.19). However, ischemic heart disease was no longer significantly higher for adults with CP (model 2 + race, HR=1.34; 95% CI=0.97–1.84; P=0.073) and CVD was no longer significantly higher for ≥65 year age group (HR=1.20; 95% CI=0.91–1.57; P=0.204).

Discussion

The chief finding of this investigation is that privately insured adults with CP had higher 12-month CVD incidence following a low-trauma fracture compared to adults without CP. This was evident even after accounting for chronic disease comorbidities that are associated with CVD. In conjunction with previous work highlighting skeletal fragility,^{8, 9, 11} our findings corroborate the need for earlier strategies post-fracture to prevent CVD (e.g., medications, stricter treatment for CVD risk factors) among adults with CP. Previous studies in non-CP populations provide evidence that fracture screening,³² better detection of skeletal fragility,³³ and exercise and pharmaceutical interventions^{32, 34} have the potential to prevent fractures and reduce fracture burden by limiting the post-fracture disease sequela. Whether preventing fracture or improving post-fracture healthcare management would reduce CVD burden for adults with CP requires further investigation.

Fractures are a high-burden condition that increases risk for premature mortality.³⁵ While the cause and effect has yet to be unraveled, sustaining a low-trauma fracture may exert its effect on premature mortality directly or indirectly³⁶ through post-fracture development of chronic diseases,³⁷ especially CVD.¹⁹ In a cohort study from England, Ryan et al.¹⁵ reported that the median age of death for adults with CP was 40 years with a 3-fold increased risk for CVD-related death. To date, no studies have examined if skeletal fragility enhances CVD risk or mortality for this underserved adult population. This is important because individuals with CP first manifest skeletal health problems prior to developing CVD risk factors. In addition to their underdeveloped and structurally weak musculoskeletal system,^{2, 4} children with CP exhibit excess intramuscular and bone marrow fat infiltration,² which is an additional signal for the pathogenesis of CVD.⁵ While the relationship between skeletal fragility and CVD can be bidirectional,^{19, 23} it is plausible that poor skeletal health development and preservation throughout the lifespan initiates or exacerbates CVD risk among adults with CP. The current study found that low-trauma fracture accelerated CVD incidence among adults with CP, as compared to adults without CP. However, the research design did not allow for investigating the broad contribution that low-trauma fracture has on CVD risk, since inclusion criteria was limited to individuals that sustained a low-trauma fracture and did not have a CVD claim in the year prior to the fracture event. Further research is needed to determine the extent that skeletal fragility impacts CVD risk and its related sequela for the entire CP population over the lifespan, and if this temporal sequence of disease development differs based on CP characteristics (e.g., severity, type of CP) or comorbidities.

Mechanisms relating poor bone and cardiovascular health have yet to be fully elucidated, but may involve alterations in signaling pathways common to both bone remodeling and arterial calcification, low-grade inflammation, or a shared set of risk factors (e.g., low estrogen levels, low physical activity, obesity).³⁸ For example, arteriosclerosis is theorized to occur more readily in the blood vessels of bone vs. those in soft tissue³⁹ and ossification (i.e., mineralization and calcification) of bone marrow blood vessels has been observed in rodents and humans.⁴⁰ Age-related increase in bone marrow blood vessel ossification is accompanied by vascular rarefaction,^{40, 41} augmented marrow fat,^40–42 reduced vasodilator capacity,^{43, 44} a lower number of red blood cells and percent of circulating lymphocytes,⁴⁵ and diminished skeletal blood flow.^{42, 43} On a more intimate level, blood vessels and bone cells can regulate the activity of each other via secretion and exchange of paracrine factors. For example, a pro-inflammatory marrow microenvironment is suspected to diminish the vasodilator capacity of bone blood vessels.⁴⁴ In addition, the skeleton is now recognized as an endocrine organ, capable of regulating a variety of systemic physiological functions.⁴⁶

Individuals with CP demonstrate several age-related phenomena, including increased bone marrow fat infiltration,² premature mortality,¹⁵ and development of CVD at a significantly younger age⁸ compared to individuals without CP (current investigation: ~65 vs. ~74 years old). Thus, it is plausible to assume they suffer from several of the pathologies associated with the aging bone vascular network. How well these bone vascular alterations extend to or are reflective of the cardiovascular system in general remains to be elucidated. Important questions to address would be whether individuals with CP have pre-clinical cardiovascular impairments and fitness levels that easily transition into a clinical status subsequent to low-trauma fracture; i.e., during a rehabilitation period which may be accompanied by increased sedentary behavior. On the other hand, does the in vitro milieu associated with CP alter the endocrine function of the skeletal system in such a manner that predisposes to increased CVD risk subsequent to low-trauma fracture? These questions require further investigation.

It is necessary to point out that the CVD incidence post-fracture in this study may be underestimating the true extent of the problem. This is because study findings were derived from a nationwide private insurance database, which likely represents the higher functioning, less skeletally fragile, and healthier segment of the CP population. In our previous studies, we reported that the prevalence of osteoporosis, a robust and independent predictor of low-trauma fracture, was higher among privately insured young and middle-aged adults (18–64 years) with CP (n=5,555; osteoporosis prevalence, 5.5%)⁴⁷ compared to adults without CP (1.3%), but substantially lower than what we have previously published from a clinical-based sample of adults with CP (n=1,395; osteoporosis prevalence, ~10% for 18–40 year olds, ~15% for 41–50 year olds, ~26% for >50 year olds).⁹ In that clinical-based sample,⁹ over half had moderate to severe forms of CP. We have reported that the prevalence of musculoskeletal morbidities among adults with moderate to severe forms of CP is close to double compared to adults with milder forms of CP.⁸ Further research is warranted that leverages data that better represents the severity spectrum of the CP population.

The limitations of this study must be discussed. First, as previously noted, the sample may reflect a healthier and less skeletally fragile sector of the CP population. Moreover, there may be a “survivor” effect that may help to explain why the HR was only modestly higher among adults ≥65 years of age with vs. without CP. Adults with CP have a lower life expectancy, especially when the CP condition is more severe,⁴⁸ and individuals covered by private insurance tend to have less medical needs and complications compared to individuals covered by public insurance (e.g., Medicare). It is therefore possible that there is a bigger public health issue than what the current data can capture. Study conclusions should be considered within the scope of this particular population of privately-insured adults with CP. Second, due to the observational design, results are subject to bias from unmeasured confounding. In order to estimate the extent of unmeasured confounding, we computed e-values, which measures the minimum strength of association needed to fully explain away a specific exposure-outcome association, conditional on the set of covariates.⁴⁹ We used model 3 when CVD was examined as the outcome and model 2 when the CVD subtypes were examined as the outcome. The e-value (lower 95% CI) needed to fully explain away the effect for the group variable (CP vs. without CP) was 2.64 (2.08) for CVD, 2.26 (1.40) for ischemic heart disease, 2.75 (1.74) for heart failure, and 3.33 (2.45) for cerebrovascular disease. Given the large e-values, it appears unlikely that unmeasured confounding largely biased effect estimates for the exposure variable. Third, this study was underpowered to examine differences in post-fracture CVD incidence for adults with CP by sex, race, and different fracture locations. Studies of the general population have shown sex³⁷ and racial⁵⁰ differences for post-fracture complications. Fracture location may also differentially associate with post-fracture complications for adults with CP. The lifetime burden of surgeries, altered mechanical loading, unhealthful aging, and skeletal fragility may lead to site-specific vulnerability to fractures,⁶ which may in turn impact health and survival outcomes. Fourth, the follow-up period of 12 months was relatively short. Future studies are needed that employ a longer follow-up period and a more comprehensive assessment of post-fracture complications.

In conclusion, we report for the first time that privately insured adults with CP have increased incidence of CVD within 12 months following a low-trauma fracture compared to adults without CP, despite an age difference of nearly a decade for those who developed CVD. Further, higher CVD incidence was evident after accounting for pre-fracture chronic diseases. Given the robust findings and the implications for clinical practice, conducting similar analyses in non-privately insured populations is warranted as it may reveal larger disparities than what was captured in this study. Future studies are needed to identify biological mechanisms linking skeletal fragility with CVD specific to the population with CP.