Reduced Bone Mineral Density Is Associated with Breast Arterial Calcification

Abstract

Background: Arterial calcification, a marker of atherosclerosis, results from a complex process of biomineralization resembling bone formation. Breast arterial calcification (BAC) has been associated with angiographic and clinical cardiovascular disease. The purpose of this study was to determine the association between reduced bone mineral density (BMD) and BAC, which may share a common pathophysiology.

Methods: We conducted a retrospective study of 228 women (55% Hispanic, mean age 64 ±10 yr) who had both mammography and BMD evaluation at Columbia University Medical Center from 2001–2003. Each mammogram was reviewed for the presence of BAC using standardized methods. BMD was measured using dual-energy x-ray absorptiometry and categorized as normal, low bone density (osteopenia), or osteoporosis as defined by the World Health Organization. Univariate and multivariate logistic regression analyses were performed to evaluate the association between reduced BMD and BAC.

Results: The prevalence of BAC, low bone density (osteopenia), and osteoporosis was 39, 42, and 29%, respectively. Women with BAC were significantly more likely to be older, Hispanic, and postmenopausal and have osteoporosis as compared with women without BAC. In age-adjusted analyses, women with BAC were more likely to have reduced BMD (odds ratio 3.0, P < 0.01) as compared with women without BAC. Furthermore, osteoporosis was strongly associated with the presence of BAC (odds ratio 3.5, P < 0.01).

Conclusion: These data suggest that osteoporosis and arterial calcification are strongly and independently correlated. Reduced BMD may identify women at risk of vascular disease.

Reduced bone mineral density and increased breast arterial calcification have each been associated with atherosclerotic cardiovascular disease, but a direct link between the two has not been established. This correlation of the two may have implications for using mammography to define risk of cardiovascular disease.

Cardiovascular disease (CVD) and osteoporosis are significant causes of morbidity and mortality among older women and may coexist in many women (1). Recent research and several prospective studies have documented an association between reduced bone mineral density (BMD) and clinical CVD, including acute stroke, angiographically documented coronary artery disease, and increased cardiovascular mortality (2,3,4,5). In addition, reduced BMD has been independently associated with surrogate markers of subclinical vascular disease including aortic artery calcification, coronary artery calcium score, and carotid artery atherosclerosis. This suggests that results of BMD testing may identify women at risk of future clinical cardiovascular events (6,7,8,9).

Breast arterial calcification (BAC) is the deposition of calcium within the media of arteries supplying breast tissue. Readily identifiable on routine screening mammography, its presence has been associated with numerous cardiovascular risk factors including diabetes and hypertension (10,11,12). In a recently published study of a large multiethnic cohort, women with BAC had significantly increased risk of coronary heart disease, ischemic stroke, and heart failure over a median follow-up of 25 yr (11). Accumulating data suggest that the mammogram may be an informative way to identify women at increased cardiovascular risk.

To our knowledge, there are no published studies that have explored the association between BAC and BMD among women. The primary purpose of this study was to determine whether there is an association between the presence of BAC on routine screening mammography and reduced BMD among white and Hispanic women.

Subjects and Methods

Study subjects

This retrospective study was conducted at the New York Presbyterian Hospital-Columbia University Medical Center. The study cohort consisted of 228 women of white or Hispanic race/ethnicity who had been referred for dual-energy x-ray absorptiometry (DXA) scans by a physician and underwent routine screening mammography from 2001–2003. Reasons for referral for DXA scans included but were not limited to postmenopausal screening, premature ovarian failure, early hysterectomy with bilateral salpingoophorectomy, chronic steroid use, previous fracture, or history of osteoporosis. Subjects were excluded if they were younger than 40 yr of age or if their mammograms were found to have a lesion suspicious for breast cancer. Demographic and clinical patient data including age, race/ethnicity, menopause, diabetes mellitus, and hypertension were ascertained from medical records. This study was approved by the Western Institutional Review Board and the Columbia University Medical Center Institutional Review Board.

BAC and BMD measurements

Mammograms were conducted in accordance with established breast cancer screening protocols and obtained in two standard projections, craniocaudal and mediolateral oblique, for each breast (13). A standard film-screen technique was used on three mammographic units (Lorad M-IV Platinum, Hologic, Danbury, CT; Senographe DMR-plus and Senographe 800T from GE Medical Systems, Milwaukee, WI). One of us (S.J.S.), blinded to the subject’s clinical data, reviewed each mammogram for the presence of arterial calcifications using the standardized methods described by Kemmeren et al. (12) and validated by Reddy et al. (14). If calcifications were present on the right, left, or both projections of the breast, the mammogram was categorized as BAC positive.

Bone density was measured in accordance with established osteoporosis screening guidelines and assessed in the lumbar spine (L1–L4) and total hip using DXA on a Hologic QDR-4500 bone densitometer (Hologic, Waltham, MA). BMD results were categorized into three groups as defined by the World Health Organization as normal (T-score > −1.0 sd), low bone density or osteopenia (T-score between −1.0 and −2.50 sd), or osteoporosis (T-score < −2.5 sd) (15). If low bone density (osteopenia) or osteoporosis was present in the lumbar spine, total hip, or both, the BMD was classified based on the lowest T-score. In our study, reduced bone density was defined as having low bone density (osteopenia) or osteoporosis.

Statistical analysis

The data were stratified by the presence of BAC. Comparisons were made using the χ² test for categorical data and the Student’s t test for continuous data.

A logistic regression model was constructed for each variable, adjusted for age. A multivariate logistic regression model, including all variables, was also created to explore the independent associations with BAC.

Statistical significance was set at P < 0.05, and data were analyzed using SAS version 8.2 (SAS Institute, Inc., Cary, NC).

Results

Characteristics of the study cohort are shown in Table 1. The mean age of the women, ranging from 40–88 yr, was 64.0 ± 10.0 yr. Of the 228 women, 126 (55%) were Hispanic and 90 (39%) had BAC. The prevalence of low bone density (osteopenia) and osteoporosis was 42 and 29%, respectively. Women with BAC were significantly more likely to be older, Hispanic, and postmenopausal and have osteoporosis compared with women without BAC (Table 2). In age-adjusted analysis, the presence of BAC was significantly associated with Hispanic ethnicity (P < 0.001), menopause (P = 0.04), osteopenia (P < 0.01), and osteoporosis (P = 0.006).

Table 1.

Prevalence of risk factors among study cohort (total n = 228)

Risk factor	Mean ± sd or n (%)
Age (yr)	64 ± 10.0
Hispanic	126 (55)
Menopause	87 (87)
BAC positive	90 (39)
Low bone density (osteopenia)	95 (42)
Osteoporosis	65 (29)
Diabetes mellitusa	35 (17)
Hypertensiona	110 (54)

^aTotal n = 205 due to missing data on 23 patients. 

Table 2.

Multivariate adjusted odds ratios (OR) and 95% confidence intervals (CI) for predictors of BAC

Variable	OR	95% CI
Age (10 yr)	1.5	1.0–2.2
Hispanic	3.1	1.5–6.3
Menopause	4.1	0.8–21
Diabetes mellitusa	1.6	0.7–3.7
Hypertensiona	0.9	0.5–1.7
Bone density
Low bone density (osteopenia)	2.7	1.1–6.8
Osteoporosis	4.4	1.6–12

^aTotal n = 205 due to missing data on 23 patients. 

Discussion

We have documented an independent association between reduced BMD and BAC, a relationship that remained significant after adjustment for age and other potential confounders. Specifically, we found that osteoporosis is highly correlated with the presence of BAC.

Our results are consistent with previous studies investigating the association between osteoporosis and clinical CVD. Among postmenopausal women, reduced BMD has been shown to be a risk factor for acute stroke and increased cardiovascular mortality later in life (2,3,4,5). Jørgensen et al. (2) found that women in the lowest quartile of BMD had a 4.8 times higher risk of stroke compared with women in the highest quartile. Longitudinal data from the Framingham Study revealed that women with the lowest bone mineral content had the highest incidence of coronary heart disease over a 30-yr follow-up (4). Marcovitz et al. (5) was the first to report an independent correlation between reduced BMD and angiographically documented coronary artery stenosis and multivessel coronary disease.

In addition to clinical cardiovascular events, reduced bone density has also been associated with surrogate markers of CVD, including aortic artery calcification and carotid artery atherosclerosis (6,7,8,9). Schulz et al. (6) found a significant age-independent association between bone density and the degree of aortic calcification measured on computed tomography in a large group of healthy postmenopausal women. Uyama et al. (9) demonstrated a significant relationship between severity of atherosclerosis measured on carotid artery ultrasound and reduced bone density.

Several studies have suggested that BAC is an independent predictor of cardiovascular mortality (10,11,12). Kemmeren et al. (12) concluded that BAC represented a significant risk factor for cardiovascular mortality in women over 50 yr based on data gathered from a large breast cancer screening trial of 12,239 women conducted in The Netherlands. In a 25-yr prospective study of 12,761 ethnically diverse women, Iribarren et al. (11) reported that after adjustment for multiple potential confounders, BAC was associated with a 1.32-fold increased risk of coronary heart disease, a 1.41-fold increased risk of ischemic stroke, and 1.52-fold increased risk of heart failure. In our study, Hispanic race/ethnicity was found to be a significant independent predictor of BAC. Reddy et al. (14) were the first to report that Hispanic women had a significantly higher prevalence of BAC compared with whites of similar age. This finding is in line with the high prevalence of cardiovascular risk factors among Hispanics. If BAC is a marker for future CVD, then our data suggest reduced BMD may also be a marker, suggesting a shared pathophysiology.

Basic research and animal studies have shown arterial calcification is the result of a biologically active and highly regulated process of calcium deposition that resembles processes seen in the metabolism of bone tissue itself (16,17,18,19,20). Molecular studies indicate that factors and specific proteins involved in osteogenesis are also expressed in calcified atherosclerotic vascular lesions (19). For example, mice deficient in a γ-carboxylated matrix protein, osteocalcin, develop extensive vascular calcification resulting in premature death from CVD (19). In addition, osteoprotegerin, an inhibitor of bone resorption, may also play a critical role. Mice deficient in osteoprotegerin develop severe osteoporosis and medial arterial calcification of the aorta (20). Whether the association between osteoporosis and arterial calcification is causal has yet to be determined.

There are limitations to our study. Using a cross-sectional design, subjects were selected in a retrospective manner from patients who underwent both routine screening mammography and BMD evaluation from 2001–2003. At present, universal screening for osteoporosis in women under the age of 65 yr is not recommended. Thus, women are referred for BMD evaluation based on the presence of certain risk factors. This may represent a selection bias, potentially accounting for the higher prevalence of low bone density (osteopenia) and osteoporosis in our study cohort than the general population. Thus, our results may not be generalized to the population at large. Demographic and clinical data were ascertained from medical records and by self-report of a physician’s diagnosis of the condition. The study design precluded the systematic determination of some clinical data, in particular the diagnosis of diabetes mellitus and hypertension in 23 study participants. Objective measurements of bone density and BAC, however, were applied to all participants. Our study was limited to white and Hispanic woman; therefore, the results may not be generalized to other ethnicities. Nevertheless, the significance of the correlation may well help to identify women who may need further evaluation.

The paradigm of CVD as a categorical diagnosis has been replaced by the growing recognition that there is a continuum of risk. It is important to identify women in preclinical stages of vascular disease so optimal preventive treatments can be offered. The possibility that routine screening mammography can also be used to identify women at increased risk for osteoporosis and CVD is appealing, but future research still needs to further explore the relationship between vascular calcification and bone loss.

Our data, thus, add to a growing body of evidence that links arterial calcification, a measure of CVD, to bone loss in older women. Strategies for the early identification of women at risk for CVD may also be considered for testing to slow the progression of osteoporosis. In the future, effective secondary preventive and treatment interventions may substantially reduce the risk of these two very common diseases in older women.