Age-Related Factors Associated with Hip Fracture Risk

Abstract

Objectives:

Advancing age is a powerful risk factor for hip fracture. The biological mechanisms through which aging impacts hip fracture risk have not been well studied.

Methods:

Biological factors associated with “advancing age” that help to explain how aging is associated with hip fracture risk are reviewed. The findings are based on analyses of the Cardiovascular Health Study, an ongoing observational study of adults ages ≥65 years with 25 years of follow up.

Results:

Five age-related factors were found to be significantly associated with hip fracture risk: (1) microvascular disease of the kidney (albuminuria and / or elevated urine albumin to creatinine ratio) and of the brain (abnormal white matter disease on brain MRI); (2) increased serum levels of carboxymethyl-lysine (CML), an advanced glycation end-product that reflects glycation and oxidative stress; (3) reduced parasympathetic tone, as derived from 24-hour Holter monitoring; (4) carotid artery atherosclerosis in the absence of clinical cardiovascular disease; and (5) increased trans-fatty acid levels in the blood. Each of these factors was associated with a 10–25%. increased risk of fracture. These associations were independent of traditional risk factors for hip fracture.

Conclusion:

Several factors associated with older age help to explain how “aging” may be associated with hip fracture risk. These same factors may also explain the high risk for mortality following hip fracture.

INTRODUCTION

Hip fractures are among the most serious complications of osteoporosis. Of all the factors known to be associated with osteoporosis and the risk of hip fractures, none is stronger than “advancing age”. Until age 70 years, osteoporotic hip fractures are rare; after age 70 years, the incidence increases exponentially [1]. Hip fractures are a watershed in the lives of the elderly. They signify an immediate and long-term increase in risk of infirmity and mortality, as well as loss of independence and reduced quality of life [2]. The reasons for this are not certain.

The term “advancing age” is not specific and does not inform the determinants by which “aging” is biologically associated with impaired bone health. While loss of bone mineral density is age-related – especially in women – it is not a sensitive or specific risk factor for hip fracture. Most hip fractures occur in people with non-osteoporotic bone density. [1–6]. Prior studies of the “aging skeleton” have focused on “classical” risk factors: (1) reduced sex hormone levels; (2) unhealthy lifestyle practices, e.g., smoking, excess consumption of alcohol, malnutrition; (3) the effects of loss of bone loading from muscle weakness; and (4) the presence of chronic illnesses, e.g., heart failure, liver failure. Not all these factors are age-related.

Here several biological, non-traditional characteristics of “aging” that put an older person at risk for hip fracture are discussed. These include markers of microvascular disease; a serum measure of protein glycation and oxidative stress; cardiovascular autonomic nervous system imbalance; and markers of subclinical cardiovascular disease. In addition, a food category that is related to hip fracture risk is described. It is no longer part of the American diet, but it was widely consumed until recently, to which today’s elderly population was exposed.

The data for these studies derive from the Cardiovascular Health Study (CHS), a prospective observational study of 5888 US adults, Caucasian and African American, ages ≥65 years [7]. The study began in 1989/1990 and actively followed participants for more than 10 years. Since then, it has passively followed participants. Information on hospitalization was collected every 6 months through June 2015. Identification of hip fractures was from ICD-9 codes from the discharge abstract. All participants were extensively phenotyped at baseline and during follow up, producing a data set rich in factors related to cardiovascular disease (CVD) risk and other infirmities of old age.

Selected baseline characteristics of the cohort categorized by sex and race are shown in the Table in the Appendix.

1. MICROVASCULAR DISEASE

Increasing age is associated with an increased prevalence of microvascular disorders. Examples include albuminuria (>29.9 mg albumin/g creatinine) and white matter hyperintensities (WMH) on brain MRIs. Albuminuria is present in more than 20% of non-diabetic individuals the age >69 years and in 30–40% of those with diabetes [8]. WMH represent cerebral small blood vessel ischemic changes. In CHS any grade of WMH (from little to extensive) was present in 87% of the cohort [9]. Interestingly, people with albuminuria have a high prevalence of WMH [10, 11], suggestive of a systemic disorder.

Hip fractures occur in cancellous (or spongy) bone. Such bone is characterized by a honeycomb pattern of trabeculae which enclose spaces through which runs the osseous microvasculature. It is here, in specialized niches, that endothelial cells (EC) that line the vessel walls, help to regulate the coupling of osteoclast and osteoblast function that result in bone formation [12]. Perturbations of this environment - possibly as part of a systemic disorder of the microvasculature - could then lead to impaired bone formation and to increased fracture risk. One way to test this hypothesis indirectly is to examine whether hip fracture risk is related to the presence of other microvascular disorders.

To do so the association of albuminuria with hip fracture risk was examined in a sub-cohort of CHS participants [13]. There were 98 fractures among men (7.7%) during a median of 8.04 years of follow-up and 215 fractures among women (11.7%) during a median of 9.50 years of follow-up. The incidence rates of hip fracture for men and women, with and without albuminuria, were 1.43 and 0.93/100 person-years, and 1.84 and 1.33/100 person-years, respectively. For each gender, the log rank test showed significantly different curves based on albuminuria status (p=0.02 men; p=0.04 for women). FIGURE 1

Figure 1:

Kaplan Meier plot of the proportion of men and women from the Cardiovascular Health Study without hip fractures categorized by the presence or absence of albuminuria.

In unadjusted Cox proportional hazard models, a doubling of urine albumin to creatinine ratio (UACR; mg/g) from time of urine collection to incident hip fracture was associated with hip fracture risk in men and women (hazard ratio (HR), 1.13 (95 % CI, 1.02–1.27); HR, 1.15 (95 % CI, 1.05–1.25), respectively). With added adjustments, a doubling of UACR remained significantly associated with hip fracture risk in women (HR, 1.12 (95 % CI, 1.001, 1.26)), but not in men (HR, 1.02 (95 % CI, 0.89, 1.17)).

In a related study,[14] the association of UACR and the degree of WMH with bone mineral density (BMD) were examined. Regression models, accounting for potentially confounding factors, showed AWMD was associated with lower hip, spine, and total body BMD in women (β −3.08 to −4.53; p < 0.01 for all) and lower hip and total body BMD in men (β −2.90 to −4.24; p = 0.01 – 0.03). UACR was associated with lower hip (β −3.37; p =0.05) and total body (β −3.21; p = 0.02) BMD in men. The associations of AWMD and UACR with BMD persisted with mutual adjustment and appeared to be additive to each other.

Taken together these findings are consistent with the notion that hip fractures are associated with other microvascular disorders which, in turn, could be related to a systemic disorder of the small blood vessels. Corroborating evidence for this hypothesis comes from the Maastricht Study which has an extensive array of microvascular tests in its cohort and has found microvascular disorders to be related with one another. [15, 16]. The association of osteoporosis with microvascular disease should also be seen in the context of new information regarding how the microvasculature plays a role in bone accretion. Osseous EC that express high levels of the cell surface markers CD31 and endomucin [called type H EC] form capillary columns in the metaphysis of growing bones. They produce chemical factors that stimulate bone accretion. With cessation of growth and the reduction of the metaphysis, there is bone marrow expansion, and type H vessel columns become sinusoidal vessels with low surface molecule expression [type L EC].

These EC are not associated with bone accretion. The aging process is associated with further diminishment of the number of H capillaries [12, 17].

2. CARBOXYMETHYL-LYSINE LEVELS

Biological aging is associated with non-enzymatic glycosylation of long-lasting proteins, such as collagen [18]. In contrast to enzymatic cross-linking of collagen fibrils, which orients collagen fibrils and contributes to their tensile strength, non-enzymatic glycosylation of collagen weakens collagen’s mechanical properties [19]. This is because glycosylation products – termed advanced glycation end products (AGEs) – activate the gene NF-kB and its downstream inflammatory proteins that can disrupt collagen’s tensile character. In vitro cell culture studies also show that non-enzymatic glycosylation of collagen makes it resistant to osteoclastic bone resorption and decreases osteoblast differentiation and proliferation [20]. Excess amounts of “aged” bone results from these processes. Finally, AGEs may inhibit osteoblast function and bone growth.[21]

Carboxy-methyl-lysine (CML) is the major AGE epitope recognized by antibodies prepared against AGE proteins. A CHS sub-cohort was examined to see whether serum levels of CML are related to hip fracture risk (median age 78.0 years) [22]. Over a median follow-up of 9.2 years, there were 348 hip fractures. The rate of fracture increased with higher quartiles of CML levels (0.94, 1.34, 1.18, and 1.69 / 100 participant-years) FIGURE 2. The unadjusted HR for hip fracture risk per 1-SD (189 ng/ml) increment in serum CML levels was 1.27 (1.15, 1.40). Adjustment for factors associated with osteoporosis (including doubling of urine albumin per gram of creatinine) moderately attenuated this association (1.18 [1.06, 1.32]; p=0.004). Men and women had similar adjusted HR for hip fracture in fully adjusted models (1.23 [1.09, 1.38] and 1.25 [1.10, 1.42], respectively). There were also no significant differences in HR between participants with or without diabetes (1.18 [1.07, 1.30] and 1.22 [1.09, 1.37], respectively). The advanced age of the cohort possibly had more impact on CML levels than the presence of DM. Finally, adjustment for hip bone mineral density did not attenuate the association of CML levels with hip fracture risk, suggesting that fracture risk in association with CML was through impaired bone quality.

Figure 2:

Kaplan Meier plot of the proportion of Cardiovascular Health Study participants free of hip fractures categorized by quartiles of carboxymethyl-lysine.

In support of these findings, the Rotterdam Study reported that a higher dietary intake of foods with CML was linearly associated with vertebral fractures (odds ratio, OR = 1.16, 95% CI (1.02–1.32) and a similar but non-significant trend with major osteoporotic fractures (OR = 1.12 (0.98–1.27).[23] Further information regarding AGEs and bone - mostly laboratory based - is summarized in the article by Asadipooya [24].

3. CARDIOVASCULAR AUTONOMIC NEUROPATHY

With aging there is a shift of autonomic nervous system function away from parasympathetic (PNS) dominance to sympathetic dominance (SNS) [25]. In younger people, the main regulator of the cardiovascular system is the PNS. In contrast, SNS tone dominates in older people owing to reduced PNS tone (not from increased SNS tone). The healthy heart shows a great deal of HRV in response to hormonal, diurnal, and psychological stressors that occur over the course of a 24-hour day to maintain homeostasis. Healthy individuals who exercise (and have low HRV during exercise) and who sleep well (with low heart rate and high HRV) have a wide distribution of HRV. The opposite holds true for unhealthy individuals who do not exercise or who sleep poorly. An overall measure of HRV is the standard deviation of the N-N interval (SDNN) – i.e., the standard deviation of the differences between normal-to-normal heart beats

In CHS 1299 participants underwent a 24-hour Holter monitor.[26] During a mean follow-up time of 14.7 years, there were 190 incident hip fractures (14.6%): 144 among 714 women (1.31 [1.06, 1.61] fractures per 100 person years) and 46 among the 585 men (0.62 [0.43, 0.90] fractures per 100 person years). A one standard deviation higher level of SDNN was associated with a 20% reduced risk of hip fracture (HR 0.80 [0.65, 0.98]; p=0.03) in women, but not in men. Increased SDNN values, in men and women, were associated with better health (less diabetes, lower weight, and lower BP) and lower levels of inflammation proteins (CRP and IL6), suggesting that conserved PSS function is a marker of overall health.

We note that hypertension (HTN) is an independent risk factor osteoporosis and for osteoporotic fractures [27]. The development of HTN is linked to the transition from PNS to SNS autonomic control of cardiovascular function with aging [28]. This finding serves to further underscore the importance of cardiovascular autonomic function for bone health.

Another aspect of autonomic nervous system together with central nervous system (CNS) control over bone health are clock genes. Bone remodeling is under strict control of the biological clock and disruption of circadian (i.e., 24-hour) rhythms, as exemplified by night shift work, is associated with more osteoporosis and fractures. As another example, osteocalcin, a product of osteoblasts, shows a diurnal rhythm being higher at nighttime compared with daytime, suggesting that osteoblast activity (and bone formation) is highest during the resting phase. For more information the reader is referred to the article by Winter et al. [29]

4. SUB-CLINICAL VASCULAR DISEASE

The prevalence of clinical cardiovascular disease (CVD) – e.g., myocardial infarctions, strokes, claudication – increases with age, as does the prevalence of osteoporosis. Both disorders are age-related. Hence, it is not surprising that clinical CVD is related to and a risk factor for osteoporosis and vice versa [30]. Clinical CVD is preceded by subclinical atherosclerosis – the thickening of the intima media of major blood vessels, plaque formation, and vessel stenosis in the absence of clinical disease. A CHS sub-cohort was examined to determine if the association of hip fracture with vascular disease begins prior to the appearance of clinical disease.

There were 3385 participants (mean age 74.7 years) with a median time to fracture of 12.1 years, who underwent baseline carotid artery and aortic wall ultrasound scanning and ankle brachial blood pressure index (ABI) measurements [31]. There were 494 hip fractures during follow-up. Among persons without any clinical CVD, an average 1 standard-deviation increase of a composite score of maximal common and internal carotid artery intimal medial thickness was associated with an increased risk of hip fracture [(HR 1.18 [1.04, 1.35]). (This risk was approximately 50% of the risk of a CVD event for the same carotid factor [HR 1.34 (1.26, 1.42) (p<0.001)]). Aortic wall thickness and ABI were not significantly associated with hip fracture risk or BMD. FIGURE 3. Based on the association of subclinical carotid vascular disease with hip fracture risk, it is concluded that vascular health, even in its early stages, is linked to bone health.

Figure 3:

Hazards ratios for hip fracture per 1 unit increase in the ratio (index) of the ankle to brachial blood pressures; per one average standard deviation increase of carotid artery intimal medial thickness [IMT] (i.e., standard deviations away from the mean IMT variable); and per 1 mm increase in both walls of the aorta.

Model 1: adjusted for age, sex and race, current smoking + alcohol (0 ref, 1–7, >7 drinks).

Model 2: Model 1 + weight (kg) + diabetes+ hypertension+ eGFR cystatin + estrogen (females) + history of falling in the year before baseline

ABI - ankle brachial index: CD — carotid intimal medial thickness: AWT - aortic wall thickness.

5. NON-ESTERIFIED FATTY ACID LEVELS

The final category of non-traditional risk factors for hip fractures is a food additive that was commonly ingested until not so long ago. Plasma non-esterified fatty acids (NEFA) are carbon chains of varying lengths (12–24 carbons) and degrees of saturation (0–6 double bonds) that are primarily derived from lipolysis of triacylglycerols in the abdominal fat. They are transported to peripheral tissue for energy utilization. They are characterized as: saturated fatty acids (SFA), monosaturated and polyunsaturated fatty acids (MUFA, PUFA), and total trans fatty acids (TFA). Elevated serum NEFAs have been associated with inflammation, oxidative stress, and insulin resistance [32], all of which are related to hip fracture risk [22]. Hence, it is plausible that NEFA levels could be associated with fracture risk. NEFA levels were measured in 2139 non-diabetic CHS participants [33]. The median follow up was 11.1 years during which there were 303 incident hip fractures. Total NEFA and each subclass tended to be associated with higher risk of hip fracture, but only total TFA reached statistical significance (HR 1.17 [95% CI, 1.04, 1.31; p=0.01] per one standard deviation higher level) with full adjustment for cofactors. FIGURE 4.

Figure 4:

Hazards ratios for incident hip fracture, per one standard deviation of each NEFA class adjusted for cofactors.

The Cox regression models are adjusted for age, gender, black race, clinic of participation, highest attained level of education, smoking status (current, never, former), alcohol use (none, < 7 drinks per week, > 7 drinks per week), weight, height, cholesterol, presence of hypertension, HTN medications, SBP, DBP, previous Ml and/or stroke, heart failure, serum albumin, eGFR based on cystatin C level, and degree of physical activity (none, moderate, and strenuous)

NEFA — non esterified Fatty acid SFA - saturated fatty acid MUFA - monounsaturated fatty acid

PUFA — polyunsaturated fatty acid TFA — trans fatty acid

TFA come exclusively from the diet, unlike other NEFA [34]. The time of sample collection in the CHS was in the early 1990s, prior to the recognition of the detrimental effects of TFA and their removal from the US food supply. It follows that the amount of TFA detected in the CHS is likely much higher than TFA levels today. Nonetheless, a study using NHANES data showed only a 50–58% reduction in TFA levels in the US population from 2000 to 2010 [35]. Higher food insecurity was associated with higher serum TFA levels in another study as late as 2010, [36] possibly making individuals in lower socio-economic classes more susceptible to hip fracture. The mechanism by which TFA can be associated with hip fracture were not studied in CHS since it is an observational study. It is noted, however, that bone cells have receptors for blood NEFA.[37].

SUMMARY

The current review suggests that hip fracture risk occurs in the context of systemic processes that are associated with “aging” FIGURE 5: microvascular disease, glycation and oxidative stress, a shift of autonomic nervous system function to a dominant sympathetic tone, and subclinical macrovascular disease. In addition, a food additive that was commonly used in the past (to which today’s elderly were exposed) was found to increase hip fracture risk, emphasizing the impact of diet on bone health. The relative impact of each of these factors on hip fracture risk is modest (10–25%), but together they may help to explain the deleterious association of “aging” on bone health independent of classical risk factors for osteoporosis and expand our understanding of the etiology of hip fractures.

Figure 5:

Summary of factors associated with hip fracture risk in older adults: 1. microvascular disease, 2. carboxymethyl-lysine, 3. cardiovascular autonomic neuropathy, 4. subclinical carotid artery disease, and 5. the ingestion of trans fatty acids lead to an increased risk of hip fracture.

These same factors may also help to explain the high morbidity and mortality that follow a hip fracture.

HIGHLIGHTS.

Advanced age is the strongest risk factor for osteoporotic hip fractures.

Osteoporotic hip factors are associated with high short-term and long-term mortality and risk of loss of independence. The reasons for this are unclear.

Here several disorders which associate with biological aging are described which in turn increase the risk of hip fractures, independent of traditional risk factors for osteoporosis.

These factors include the presence of several types of microvascular diseases; elevated levels of advanced glycation end products – markers of oxidative stress and inflammation; reduced cardiovascular autonomic parasympathetic tone; and the presence of subclinical vascular disease in the carotid arteries.

Ingestion of trans fatty acids derived from the diet is also associated with hip fracture risk

Clinical Relevance.

Osteoporotic hip fractures occur in the context of age-related disorders such as the presence of microvascular diseases, oxidative stress, cardiovascular autonomic neuropathy, and subclinical vascular disease, Ingestion of dietary trans fatty acids also contribute to the risk of hip fractures.