Breast arterial calcification (BAC), detected on routine mammography, is the calcification of medial arteries. BAC has been suggested to be linked to cardiovascular disease (CVD) risk. A systematic search was done that identified studies examining BAC, CVD risk factors (diabetes, hypertension, dyslipidemia, smoking, obesity), cardiovascular outcomes [stroke, myocardial infarction (MI), heart failure (HF), cardiac mortality], and all-cause mortality. Additionally, an atherosclerotic CVD (ASCVD) composite outcomes including MI, stroke, and cardiac mortality was analyzed. A random-effects model was used to calculate risk ratios (RR) and odds ratio (OR) with 95% confidence intervals (CI). Heterogeneity was assessed with Q values and I2 statistics. 45 studies were included in the final meta-analysis, representing 68,584 women. BAC prevalence was 17.1%. Among cross-sectional studies, BAC was associated with diabetes (OR: 1.97, 95% CI: 1.71–2.27, I2= 70.78%), hypertension (OR: 1.82, 95% CI: 1.52–2.18, I2 = 88.3%), and hyperlipidemia (OR: 1.24, 95% CI: 1.06–1.45, I2 = 76.4%). BAC was negatively associated with smoking (OR: 0.50, 95% CI: 0.41–0.61, I2 = 78.4%). BAC was associated with known CVD (OR: 2.71, 95% CI: 2.13–3.45, I2 = 76.7%). Among cohort studies, BAC was associated with incident stroke (RR: 2.05, 95% CI: 1.58–2.65, I2 = 50.8%), HF (RR: 2.14, 95% CI: 1.38–3.32, I2 = 87.1%), cardiac death (RR: 2.94, 95% CI: 1.32–6.54, I2 = 72.7%), ASCVD (RR: 1.58, 95% CI: 1.23–2.04 I2 = 81.9%) and all-cause mortality (RR: 2.04, 95% CI: 1.08–3.84, I2 = 96.78%). Significant interstudy heterogeneity in this meta-analysis is a limitation on confidence in the pooled results. In conclusion, BAC observed on mammography may serve as a marker for increased CVD risk and mortality in women; however, future research is needed to standardize BAC assessment and confirm its clinical utility in CVD risk stratification.
Cardiovascular disease (CVD) continues to be the leading cause of death among women in the United States, accounting for 420,000 deaths annually. CVD risk estimation tools, such as the pooled cohort equation and PREVENT equation are used to determine the absolute risk of a first CVD event over 10 years. Current CVD risk assessment tools include traditional risk factors of CVD, such as hyperlipidemia, diabetes, hypertension, smoking, and obesity in addition to cardiovascular biomarkers. Medial arterial calcification (MAC), a calcification of vascular smooth muscle cells in the medial layer of arteries, has been found to be associated with vascular stiffness, microvascular dysfunction, heart failure (HF), and CVD events. MAC differs from the atherosclerotic-calcified plaques observed on coronary computed tomography scans which is a later stage calcification of the intimal layer of arteries. A noninvasive mammogram used for breast cancer screening has been shown to detect MAC of medium-sized arteries in the breast, called breast arterial calcification (BAC). BAC is not commonly reported in radiology reports, as there is no established association between BAC and breast cancer. However, BAC may be detectable before the onset of clinical CVD and could represent an important marker of subclinical CVD in women. Therefore, we conducted a systematic review to evaluate the quality of available literature on BAC in relation to CVD risk factors, CVD, and all-cause mortality, and conducted a meta-analysis of the associations between BAC, CVD, and mortality.
Methods
This systematic review was performed using guidelines for reporting systematic reviews outlined in the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) 2020 statement.
Eligibility criteria
Key inclusion criteria for studies included English language publications that reported BAC prevalence on a mammogram and information on CVD risk factors, or CVD events [stroke, myocardial infarction (MI), cardiac mortality], and all-cause mortality. CVD risk factors include obesity, diabetes, smoking, hypertension, and dyslipidemia. Key exclusion criteria included nonhuman studies, review papers, editorials, and any study that does not report either the BAC prevalence or any outcome of interest listed in the inclusion criteria. Studies assessing the association between BAC and CVD based solely on imaging measures, rather than clinical events, were excluded.
Search strategy
The systematic search was conducted in the electronic databases PubMed and Cochrane Register of Clinical Trials. Specific search terms and phrases were used for identifying relevant publications. The list of search terms and MeSH phrases used are provided in Table S1 . A comprehensive literature search was performed to identify studies relevant to CVD prevalence and BAC in the available published literature from the years 1979 to October 1, 2024. Studies that were not found in the search but were cited in previous meta-analyses were also included for eligibility screening.
Study selection
Titles found during the search process were screened for relevance (by author EJC). Next, relevant abstracts and manuscripts were screened for inclusion criteria by 2 independent reviewers (EJC and CMH). Reasons for exclusion were “review article”, and “no outcome of interest (NOOI)”. The number of excluded studies and the reasons for exclusion are presented in the PRISMA flow diagram, Figure 1 . Any disagreements were resolved by a third reviewer (MN).
Flow chart for systematic review of studies according to PRISMA flow diagram. (PRISMA = Preferred Reporting Items for Systematic Reviews and Meta-analyses).
Data extraction
From each study, data extracted included PubMed identification (PMID) (if applicable), title, first author, year published, follow-up time in years (if applicable), study design, country of origin, total number of participants, BAC prevalence, number of participants with and without BAC, average age of participants and the primary outcome/endpoint of each study. Raw data were extracted for CVD risk factors, CVD events, and all-cause mortality. The risk factors extracted were a history of hypertension, type 2 diabetes, hyperlipidemia, smoking cigarettes, and any history of CVD event prior to BAC assessment. CVD events included MI, ischemic stroke, HF, cardiac-related mortality, and all-cause mortality.
Risk of bias assessment
The risk of bias was quantified using a modified Newcastle-Ottawa Quality Assessment Scale (NOS) for cohort and cross-sectional studies. The NOS tool was designed to evaluate the quality of nonrandomized studies in 3 broad categories: the selection of study groups, the comparability of groups, and the ascertainment of the outcome of interest. It uses a “star system” in which a study can earn a “star” or point in each domain of the 3 aforementioned categories to earn a maximum total score of 8. For both cohort and cross-sectional study types, a coding manual defined decision rules for each domain before the risk of bias analysis as used in methods of similar meta-analyses. Each study could receive a maximum of 8 stars or points if its protocol adhered to all predetermined domains. Studies scoring 7 to 8 were considered “low risk of bias,” 5-6 as ‘intermediate risk of bias’, and < 5 as ‘high risk of bias’. Bias ratings were further broken down for each study into the subcategories of selection, comparability, and outcome, as described above.
Data synthesis and statistical analysis
Given the anticipated clinical and methodological heterogeneity among studies, random effects meta-analyses were conducted to estimate pooled risk ratios (RR) and odds ratios (OR) with 95% confidence intervals (CI) using the Mantel-Haenszel method. RR were used to report the association between BAC, CVD events, and all-cause mortality in the cohort studies while OR were used to report the association between BAC, CVD risk factors, and a history of CVD among the cross-sectional studies. An atherosclerotic CVD (ASCVD) composite outcome including MI, stroke, and cardiac mortality was also analyzed. Heterogeneity was assessed using Q statistics and by I 2 statistics. A point-by-point sensitivity analysis was performed to assess the robustness of the findings and ensure the findings were sensitive to the exclusion of any one study. Publication bias was assessed by visual inspection of funnel plots and by performing Egger’s test, where an intercept, 95% CI, standard error, and 2-sided p-value were calculated, when ≥10 studies were available for a given outcome. For outcomes with fewer than 10 studies, formal assessment of publication bias was not performed due to the limited power and reliability of these methods. All statistical analyses were conducted using Comprehensive Meta-Analysis software. A 2-sided alpha value of 0.05 is considered statistically significant. The p-values along with the corresponding 95% CI for pooled effect estimates are reported.
Results
Study selection
A total of 348 unique articles were found through PubMed, Cochrane, and other resources. After screening titles and abstracts as described above, 78 articles were retrieved for full-text analysis for eligibility determination. After thorough review, 33 articles were excluded due to having outcomes not relevant to the review, being duplicate studies, or lacking full-text availability in English. The final number of studies included in the meta-analysis was 45. The PRISMA 2020 flow diagram that summarizes the literature search is summarized in Figure 1 .
Study characteristics
The 45 included studies comprised 68,538 women; 56,801 without BAC and 11,737 with BAC (overall study prevalence was approximately 17.1%). The average age of participants was 60.52 years with a standard deviation of 6.1 years. The studies included 5 prospective cohort studies, 9 retrospective cohort studies, 1 case control study and 30 retrospective cross-sectional studies. The raw data was extracted from a study if it provided the data for the particular risk factor or outcome.
Quality assessment
According to the modified NOS guide, most studies contained some elements that contributed to a risk of bias associated with the findings. Overall, 22.2% of studies had low risk of bias, 26.7% had intermediate risk of bias, and 51.1% had high risk of bias ( Table S1 ). Across cohort studies, the most common areas containing limitations were that studies had inadequate follow-up time or ascertained their endpoints with surveys that could lead to recall bias (66%). Across cross-sectional studies, a common area of bias was age differences in the exposed and control groups leading to possibility of confounding bias. Additionally, some studies were at increased risk of bias due to patient populations not being representative of a typical population since many included women who required invasive coronary angiograms due to a pre-existing cardiac complaint. A summary of these findings is found in Figure 2 .
Summary of categorical bias across (a) cohort and (b) cross-sectional studies. Studies graded in the categories of selection bias, comparability bias and outcome bias according to the Newcastle Ottawa Scale. Studies were assigned either high bias risk, possible bias risk or low bias risk for selection and outcome categories. Studies were assigned either low bias risk or high bias risk for comparability.
BAC and traditional cardiovascular risk factors
A total of 45 cross-sectional studies were included in the meta-analysis assessing BAC and traditional cardiovascular risk factors (diabetes n = 43, hypertension n = 39, smoking n = 36, hyperlipidemia n = 31, BMI n = 17). There was a statistically significant association between BAC presence and hypertension (5,202/10,325 [50.4%] in BAC group versus 14,329/44,654 [32.1%] in control group; OR: 1.82, 95% CI: [1.52– 2.18], p < 0.0001, Q value = 342.48, P value <0.0001, I 2: 88.32%), with no evidence of publication bias based on a symmetric funnel plot (Egger’s test P = 0.44, Y-int = −0.51), and sensitivity analysis confirmed a stable effect size in 1-study removed analysis ( Supp. Fig. S2 A-C). There was a statistically significant association between BAC and diabetes (2,329/11,748 [19.8%] in BAC group versus 4,830/56,836 [8.5%]; OR: 1.97, 95%CI: [1.71– 2.27], p < 0.0001, Q value = 150.55, P value <0.001, I 2: 70.78%), with no evidence of publication bias based on a symmetric funnel plot (Egger’s test P = 0.14, Y-int = 0.64), and sensitivity analysis confirming the robustness of the findings, as the effect size remained stable in one-study-removed analyses ( Supp. Figure S3 A-C.)
There was a statistically significant inverse association between BAC presence and smoking history (1,294/10,538 [12.3%] in BAC group versus 12,844/52,081 [24.6%] in control group; (OR: 0.50, 95% CI: [0.41– 0.61], p < 0.0001, Q value = 166.95, I 2: 78.44%), with no evidence of publication bias based on a symmetric funnel plot (Egger’s test p = 0.056, Y-int = −0.997), and sensitivity analysis confirms stable effect size in one-study removed analysis ( Supp. Fig. S4 A-C). There was a statistically significant association between BAC presence and hyperlipidemia (4,471/8,886 [50.3%] in BAC group versus 11,123/27,586 [40.3%] in control group; OR: 1.24 95% CI: [1.06– 1.45], p = 0.009, Q value = 131.40, I 2: 76.41%), with no evidence of publication bias based on a symmetric funnel plot (Egger’s test p = 0.058, Y-int = −0.92), and sensitivity analysis confirms stable effect size in one-study removed analysis ( Supp. Fig. S5 A-C).
There were no overall statistically significant associations found between BAC and BMI (mean BMI= 28.3 kg/m 2 in BAC group versus 27.6 kg/m 2 in control group; pooled raw mean difference= 0.402, 95% CI: [−0.09– 0.898], p = 0.112, Q value = 95.59, I 2: 84.31%), with no evidence of publication bias based on a symmetric funnel plot (Egger’s test p = 0.99, Y-int = −0.008), and sensitivity analysis confirms stable effect size in one-study removed analysis ( Supp. Fig. S6 A-C).
BAC and history of known cardiovascular disease
There were 21 cross sectional studies eligible to assess for association between BAC and history of known cardiovascular disease including 40,441 women. The mean age was 60.1 years. BAC was associated with past history of known cardiovascular disease (1,084/7,309 [14.8%] in BAC group versus 2,152/33,132 [6.5%] in control group; OR: 2.71 95% CI: [2.13– 3.45], Q value = 96.49, p value = <0.0001, I 2 = 76.7%), with no evidence of publication bias based on a symmetric funnel plot (Egger’s test p = 0.96, Y-int = 0.044), and sensitivity analysis confirms stable effect size in one-study removed analysis ( Supp. Fig. S7 A-C).
BAC and cardiovascular disease and mortality
A total of 9 cohort studies were identified that assessed BAC and CVD and/or mortality comprising 46,958 women total with an average age of 61 years. The average follow-up time was 10.2 years. There was no overall statistically significant association between BAC presence and development of myocardial infarction (104/6,265 [1.7%] in BAC group versus 324/30,995 [1.0%] in control group; RR: 1.54, 95%CI: [0.89–2.66], p = 0.122, q value = 19.87, I 2: 74.84%), with no evidence of publication bias based on a symmetric funnel plot (Egger’s test p = 0.64, Y-int = −1.29). However, the sensitivity analysis demonstrates an unstable effect size in the one-study removed analysis ( Supp. Fig. S8 A-C) in which the removal of the study performed by Penugonda et al. (2010) would result in an overall statistically significant association between BAC and MI (RR: 2.02, 95% CI: [1.47– 2.78], p < 0.0001). This is likely attributable to the study’s highly selected cohort of symptomatic women with angina or abnormal stress tests, in whom the prevalence of coronary artery disease (CAD) was already high.
There was a statistically significant association between BAC presence and development of ischemic stroke (235/6,142 [3.8%] in BAC group versus 1,520/30,887 [4.9%] in control group; RR: 2.05, 95%CI: [1.58—2.65], p < 0.0001, Q-value = 8.14, I 2 = 50.83%), with no evidence of publication bias based on a symmetric funnel plot (Egger’s test p = 0.83, Y-int = 0.33), and sensitivity analysis confirms stable effect size in one-study removed analysis ( Supp. Fig. S9 A-C).
There was a statistically significant association between BAC presence and development of congestive heart failure (288/6,009 [4.8%] in BAC group versus 1,266/29,978 [4.2%] in control group; RR: 2.14, 95%CI: [1.38—3.32], p = 0.001, Q value = 23.27, I 2: 87.11%), with no evidence of publication bias based on a symmetric funnel plot (Egger’s test p = 0.54, Y-int = −3.26), and sensitivity analysis confirms stable effect size in one-study removed analysis ( Supp. Fig. S10 A-C). There was a statistically significant association between BAC presence and development of composite ASCVD (485/6,629 [7.3%] in BAC group versus 2,107/32,294 [6.5%] in control group; RR: 1.58, 95%CI: [1.23—2.04], p < 0.0001, Q value = 38.76, I 2 =81.94%), with no evidence of publication bias based on a symmetric funnel plot (Egger’s test p = 0.30, Y-int = −2.43), and sensitivity analysis confirms stable effect size in one-study removed analysis ( Supp. Fig. S11 A-C).
There was a statistically significant association between BAC presence and development of cardiac-related mortality (121/2,573 [4.7%] in BAC group versus 503/14,826 [3.4%] in control group; RR: 2.94, 95%CI: [1.32—6.54], p = 0.008, Q value = 7.33, I 2: 72.72%), with evidence of possible publication bias based on a mildly asymmetric funnel plot shifting to the left (Egger’s test p = 0.033, Y-int = 2.51), and sensitivity analysis confirms stable effect size in one-study removed analysis ( Supp. Fig. S12 A-C). There was a statistically significant association between BAC presence and development of all-cause mortality (588/6,796 [8.7%] in BAC group versus 1,849/28,695 [6.4%] in control group; RR: 2.04, 95%CI: [1.08—3.84], p = 0.027, Q value = 92.24, p = 0.027, I 2: 96.75%), with no evidence of publication bias based on a symmetric funnel plot (Egger’s test p = 0.96, Y-int = 0.35), and sensitivity analysis demonstrated that removal of 1 study would cause a decrease in overall effect size which may indicate some level of bias ( Supp. Fig. S1 3 A-C).
Discussion
This study highlights the potential role of BAC as a marker of cardiovascular risk in women. Specifically, the presence of BAC was associated with traditional cardiovascular risk factors, including a history of hypertension, diabetes, hyperlipidemia, and known cardiovascular disease, but not with BMI. Notably, an inverse association with a smoking history was observed. BAC was associated with the development of adverse outcomes, such as ischemic stroke, ASCVD, HF, cardiac death, and all-cause mortality. Although no significant association was observed between BAC and the incidence of MI in the primary analysis, sensitivity testing demonstrated that the exclusion of a single study resulted in a significant positive association. The study by Penugonda et al. was conducted exclusively among symptomatic women referred for coronary angiography, introducing selection bias and limiting generalizability. Given the high baseline prevalence of coronary artery disease in this population, the incremental predictive value of BAC was likely underestimated, which may explain the attenuation of the pooled association when this study was included. This observation further supports the overall conclusion that presence of BAC may serve as an indicator of increased risk for future CVD events. Given that BAC can be identified during routine mammography, it may represent a promising, noninvasive, sex-specific risk stratification biomarker for ASCVD in women, without additional cost or radiation exposure.
Coronary artery calcification (CAC) is a marker of atherosclerotic plaque burden and is quantified using the CAC score via noncontrast cardiac computed tomography and has been identified as a CVD risk marker that is additive to and independent of traditional risk factors included in the Framingham risk score (FRS). However, CAC and BAC represent distinct vascular calcification processes with different pathophysiological mechanisms. CAC represents intimal layer calcification, and the distribution and amount of calcification do not directly reflect plaque size. In contrast, BAC represents calcification within the medial layer of arteries rather than intimal atherosclerosis. Calcium deposition in the tunica media leads to arterial stiffening without necessarily causing luminal narrowing or plaque rupture. . BAC may predict CAC. This was evaluated in the Joint BrEast CAncer & CardiOvascular ScreeniNg (BEACON) study, which revealed BAC’s high specificity of 92% in detecting CAC, indicating the association of BAC presence with a high likelihood of underlying coronary calcification. BAC’s diagnostic accuracy for predicting elevated CAC is highest in middle-aged women under 60 years old, thus supporting the possible use of BAC as a cardiovascular risk marker in middle-aged women. However, it remains unknown if BAC is additive to and independent of the coronary artery calcium score. Prospective studies will need to determine the role of BAC as an independent CVD risk marker and screening tool.
Of note, BAC has been assessed by multiple methods. Most studies included in this analysis report radiologist findings of BAC as either present or absent on mammograms in a binary (yes or no) method which does not provide information on the extent or severity of calcification. semiquantitative grading methods have been described in the literature and typically involve assigning a score based on the number of calcified vessels, length of calcification, or degree of vessel involvement. Both a 4-point scale (none, mild, moderate, severe) and a 12-point scale have been described for this grading. BAC mass (mg) can also be determined by densitometry. Additionally, machine learning algorithms have been developed to automate the detection and quantification of BAC. Cheng et al. formulated an automated algorithm using calcification and “vesselness” as the 2 underlying cues within a multiseeded tracking scheme. This scheme formulates sampling pathways to describe the topology of calcium deposits within the vasculature. The performance of their algorithm reached 92.6 ± 2.2% sensitivity and 83.9 ± 3.6% specificity. Wang et al. proposed a deep convolutional neural network (CNN) that differentiates between BAC and non-BAC pixels. This was compared to the manual detection of BAC and reached a performance comparable to that of radiologists.
Our results are consistent with previous, smaller scale meta-analyses that have identified associations between BAC and cardiovascular risk factors, such as age, hypertension, CAD, hypercholesterolemia, and diabetes mellitus, with an inverse association observed with smoking. ,, These analyses also reported that BAC is associated with an increased risk of cardiac death, cardiovascular disease events, stroke, peripheral vascular disease, and HF. , BAC has also been identified as an independent and strong predictor (OR > 22) of high CAC scores among women without known CAD. Finally, among women aged 40 to 79 years, BAC presence was found approximately 9 years on average before the development of CAC, thereby suggesting a role in predicting the development of CAC.
This study has several notable strengths. This is the largest meta-analysis of BAC and CVD to date and includes multiple methods of BAC detection. Interventional or randomized controlled studies are not feasible for this subject, so only observational studies can be conducted and included in this meta-analysis. Nevertheless, all available literature on this topic is represented, and the risks of selection and outcome bias were generally low across most of the included studies. However, there are limitations of this study which should be considered. Significant interstudy heterogeneity was observed as reflected by the high I² values across several analyses. This reduces confidence in the results and reflects the variability in the available literature on this topic. However, sensitivity analysis was performed to assess the robustness of the findings and to explore the impact of significant heterogeneity among the studies. The lack of standardized BAC grading and quantification, ranging from binary presence/absence to semiquantitative and continuous scales, limits the comparability of BAC. In addition, there was variability in outcome assessment. Some studies relied on self-reported endpoints, introducing potential recall or ascertainment biases ( Table 1 ).
Table 1
Summary of study characteristics for included studies
| Author, Year | Country of origin | Follow-up time (years) | Study design | Total # of Participants | n control group | n BAC group | Control Average Age ± SD | BAC Average Age ± SD | Method for BAC Assessment | Main Findings/Primary Outcome |
|---|---|---|---|---|---|---|---|---|---|---|
| Kemmeren et al. | Netherlands | 16 | Prospective cohort | 12084 | 10977 | 1107 | 57.5 | 60.1 | Blinded radiologist | BAC is associated with cardiovascular related mortality in women above age 50. |
| Iribarren et al. | USA | 25 | Prospective cohort | 12761 | 12337 | 424 | 56 | 66 | Blinded radiologist | BAC is associated with CVD events and ASCVD risk factors. |
| Schnatz, et al. | USA | 5 | Prospective cohort | 1454 | 1247 | 207 | 54.3 ± 11.1 | 68.7 ± 10.7 | Blinded radiologist | BAC is associated with increased risk of CAD and stroke. |
| Iribarren, et al. | USA | 5 | Prospective cohort | 5059 | 3721 | 1338 | 65.2 ± 4.2 | 67.1 ± 4.8 | Rigorously validated densitometry method | BAC is associated with increased ASCVD and global CVD. |
| Nudy, et al. | USA | 10 | Prospective cohort | 1039 | 925 | 114 | 54.1 ± 10.4 | 66.8 ± 9.7 | Blinded radiologist | BAC is associated with increased risk of CAD and stroke. |
| Abou-Hassan, et al. | USA | 3.5 | Retrospective cohort | 202 | 84 | 118 | 53.9 ± 12.3 | 61.5 ± 11.4 | Blinded radiologist | BAC is strongly associated with peripheral artery disease in women with end stage renal disease. |
| Soran, et al. | USA | 7.5 | Retrospective cohort | 602 | 441 | 161 | 53.9 ± 8.8 | 61.9 ± 9.8 | Blinded radiologist | No association found between left-sided breast radiation therapy or BAC with subsequent cardiac events. |
| Karm, et al. | USA | 2 | Retrospective cohort | 198 | 116 | 82 | 61 | 72 | Blinded radiologist | Presence of BAC, smoking and HLD are all associated with risk of CAD. |
| Galiano, et al. | Spain | 23 | Retrospective cohort | 256 | 128 | 128 | 59.5 | 59 | Blinded radiologist | BAC is associated with cardiovascular events and mortality. |
| Lee et al. | Australia | 9 | Retrospective cohort | 1020 | 836 | 184 | 59.3 ± 6.9 | 65.1 ± 5.5 | Blinded radiologist | BAC is associated with increased ASCVD risk but not independent of known CVD risk factors. |
| Allen et al. | USA | 6 | Retrospective cohort | 18092 | 13869 | 4223 | 54.2 ± 10 | 65.2 ± 11.6 | Artificial intelligence generated score | BAC is independently associated with morality and CVD when quantifying BAC with artificial intelligence. |
| Bae et al. | South Korea | N/A | Case-control | 202 | 101 | 101 | 58.9 ± 7 | 58.9 ± 7 | Blinded radiologist | Association of BAC and increased risk of the metabolic syndrome and CAD in women above age 40. |
| Kelly et al. | Ireland | N/A | Retrospective cohort | 104 | 90 | 14 | 58.87 ± | 59.4 | Blinded radiologist | BAC is associated with increased severity of CAD in symptomatic patients who underwent invasive coronary angiogram. |
| Pudil et al. | Czech Republic | N/A | Retrospective cohort | 163 | 129 | 34 | 67 ± 12 | 73 ± 13 | Assessed independently by 2 reviewers (non-radiologist) | BAC is associated with more severe forms CAD as determined by cardiac catheterization. |
| Penugonda et al. | USA | N/A | Retrospective cohort | 94 | 37 | 57 | 63.2 ± 10.9 | 68.9 ± 9.7 | Blinded radiologist | No association of BAC with ASCVD risk factors, CAD, or MI in patients who have undergone cardiac catheterization. |
| Baum et al. | USA | N/A | Cross sectional | 319 | 282 | 37 | 50 | 65 | Reviewer not specified | Presence of BAC is associated with diabetes. |
| Crystal et al. | Israel | N/A | Cross-sectional | 865 | 713 | 152 | 54 | 65 | Blinded radiologist | BAC association with increased ASCVD risk and CVD risk factors. |
| Henkin et al. | Israel | N/A | Cross-sectional | 319 | 188 | 131 | 59.9 | 64.3 | Blinded radiologist | Lack of association between BAC and CAD on angiography. |
| Maas et al. | Netherlands | N/A | Cross-sectional | 600 | 462 | 138 | 66.5 | 70.4 | Non-radiologist reviewer | BAC is associated with known ASCVD risk factors and with increased parity in postmenopausal women. |
| Cetin et al. | Turkey | N/A | Cross-sectional | 2400 | 2181 | 219 | N/A | N/A | Reviewer not specified | BAC is associated with hypertension and diabetes. |
| Kataoka et al. | United Kingdom | N/A | Cross-sectional | 1590 | 1336 | 254 | 62.7 | 65.8 | Non-blinded radiologist | Presence of BAC is associated with CAD when adjusted for age. |
| Taskin et al. | Turkey | N/A | Cross-sectional | 985 | 500 | 485 | N/A | N/A | Non-blinded radiologist | Presence of BAC is correlated with advancing age and other ASCVD risk factors in post-menopausal women. |
| Fiuza et al. | Brazil | N/A | Cross-sectional | 85 | 44 | 41 | N/A | N/A | Blinded radiologist | BAC is associated with CAD observed with coronary angiography. |
| Topal et al. | Turkey | N/A | Cross-sectional | 123 | 74 | 49 | 52.34 | 63.85 | Blinded radiologist | BAC is correlated with advancing age and presence of CAD. |
| Rotter et al. | USA | N/A | Cross-sectional | 1919 | 1651 | 268 | 54 | 70 | Blinded radiologist | BAC is associated with CVD risk factors and cardiovascular morbidity. |
| Dale et al. | USA | N/A | Cross-sectional | 1000 | 839 | 161 | 56 | 70 | Blinded radiologist | BAC is associated with history of CAD and diabetes. |
| Ferreira et al. | Brazil | N/A | Cross-sectional | 307 | 281 | 26 | 55.1 ± 6.9 | 57.4 ± 6.4 | Blinded radiologist | BAC is associated with CVD in post-menopausal women independent of other risk factors. |
| Zgheib et al. | USA | N/A | Cross-sectional | 172 | 115 | 57 | 60.4 ± 11.1 | 72 ± 9.8 | Blinded radiologist | No correlation between BAC and CAD observed with invasive coronary angiogram. |
| Sarrafzadegann et al. | Iran | N/A | Cross-sectional | 84 | 78 | 6 | 43.2 | 46.6 | Blinded radiologist | No association between BAC and CAD observed with invasive coronary angiogram in pre-menopausal women. |
| Sedighi et al. | Iran | N/A | Cross-sectional | 204 | 125 | 79 | 58.9 ± 7.8 | 60.6 ± 7.2 | Blinded radiologist | BAC is associated with increased risk of carotid atherosclerosis and thicker carotid intima-media thickness. |
| Hekimoglu et al. | Turkey | N/A | Cross-sectional | 55 | 32 | 23 | 59.4 ± 8.9 | 68 ± 6.5 | Blinded radiologist | BAC is associated with the presence and severity of CAD seen on coronary angiogram. |
| Yildiz et al. | Turkey | N/A | Cross-sectional | 310 | 205 | 105 | 53.4 ± 6.2 | 60.7 ± 9.4 | Non-blinded radiologist | BAC is associated with hypertension, diabetes and metabolic syndrome in post-menopausal women. |
| Margolies et al. | USA | N/A | Cross-sectional | 292 | 168 | 124 | 57.7 ± 9.5 | 66.5 ± 10.3 | Blinded radiologist | BAC is associated with severity of CAC seen on CT and other CVD risk factors. |
| Chadashvili et al. | USA | N/A | Cross-sectional | 126 | 94 | 32 | 61 | 56 | Blinded radiologist | BAC is associated with known CVD risk factors and coronary artery calcium score >11. |
| Ronzani et al. | Brazil | N/A | Cross-sectional | 312 | 241 | 71 | 59.3 ± 6.5 | 54.8 ± 7.3 | Blinded radiologist | BAC is associated with advancing age, hypertension, diabetes, CKD and low GFR. |
| Fathala et al. | Saudi Arabia | N/A | Cross-sectional | 435 | 177 | 258 | 55 ± 7 | 61 ± 8 | Non-radiologist reviewer | Presence or severity of BAC does not predict MI on stress MPS. |
| Fathala et al. | Saudi Arabia | N/A | Cross-sectional | 307 | 165 | 142 | 54 ± 7.5 | 59 ± 8 | Non-radiologist reviewer | BAC is associated with age, hypertension, CKD, and CAC seen on CT but not with dyslipidemia, smoking or family history of CAD. |
| Yildiz et al. | Turkey | N/A | Cross-sectional | 132 | 66 | 66 | 54.09 ± 10.8 | 54.09 ± 11 | Blinded radiologist | BAC is associated with increased 10-year fatal CVD risk quantified by SCORE risk system. |
| Soylu et al. | Turkey | N/A | Cross-sectional | 404 | 281 | 123 | 54.5 ± 9.1 | 67.9 ± 9.9 | Blinded radiologist | BAC is associated with CAC and aortic calcifications observed on CT. |
| McLenachan et al. | United Kingdom | N/A | Cross-sectional | 405 | 312 | 93 | 57 ± 8 | 63 ± 7 | Blinded radiologist | BAC is associated with presence and severity of CAD observed with invasive coronary angiogram. |
| Fathala et al. | Saudi Arabia | N/A | Cross-sectional | 202 | 81 | 121 | 58.5 ± 5.9 | 62.4 ± 7.5 | Blinded radiologist | BAC was not predictive of CAD based on angiography, but BAC was associated with age and known CVD risk factors. |
| Shobeiri et al. | Iran | N/A | Cross-sectional | 600 | 539 | 61 | 55.32 ± 6.57 | 58.59 ± 7.82 | Blinded radiologist | BAC is associated with increased rate of CVD. |
| Goel et al. | Australia | N/A | Cross-sectional | 153 | 116 | 37 | 58 ± 10 | 70 ± 8 | Blinded radiologist | BAC and epicardial adipose tissue volume are associated with CVD risk factors and CAD but are not associated with each other. |
| Oksul et al. | Turkey | N/A | Cross-sectional | 320 | 273 | 47 | 56.7 ± 7.9 | 64.7 ± 7.8 | Blinded radiologist | BAC is associated with an increased CAC score observed on CT. |
| Gardinalli-Filho et al. | Brazil | N/A | Cross-sectional | 183 | 144 | 39 | 59.6 ± 10 | 68.2 ± 9.6 | Blinded radiologist | No association between BAC and severity of CAD observed with invasive coronary angiogram. |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
