Introduction

Myocardial ischemia and acute myocardial infarction are late-stage manifestations of atherosclerotic coronary artery disease (CAD). Atherosclerosis is a ubiquitous chronic systemic disease beginning early in life, affecting arteries throughout the body, long before significant stenoses or occlusions appear. Strategies for primary prevention very early in the disease cascade prior to the advent of life-threatening events are needed, and could be based on identification of pre-clinical atherosclerosis.1,2,3 We previously found multi-vessel bed atherosclerosis to be common in ancient human mummies from Africa, Asia, and North and South America, despite their non-western lifestyle and a dearth of traditional risk factors.4 The detection of arterial calcification in systemic arteries outside the coronary circulation is well recognized, not as a risk factor for atherosclerosis, but rather a prima facie diagnostic feature of the disease, often in a pre-clinical state when more individualized and intensive preventive treatment could be employed.2,4,5,6,7,8

In order to assess the prevalence of pre-clinical atherosclerosis in the vascular beds of a contemporary Egyptian population, whole-body CT scans were performed in a cohort of individuals without known CAD and no evidence of myocardial ischemia on MPI.

Methods

Study Population

One hundred and fifty-four consecutive outpatients without known CAD, whose SPECT MPI, performed for standard clinical indications, was normal, were recruited at an outpatient imaging center in Cairo, Egypt (Alfa Scan) from April 2014 to April 2015 to undergo a whole-body CT scan for research purposes. Informed consent of each study subject was obtained prior to enrollment.

Myocardial Perfusion Imaging

A total of 151 (98%) patients underwent symptom-limited treadmill exercise stress according to the standard Bruce protocol, while 3 (2%) patients underwent dipyridamole stress. Termination of the stress test was performed according to recommendations in the updated guidelines of exercise testing.9

A total of 152/154 subjects underwent same-day stress-only MPI imaging in accordance with ASNC 2016 guidelines.9 The remaining two patients underwent two-day technetium-99m sestamibi gated SPECT MPI. All images were acquired with a dual-head Philips JetStream camera (Philips Healthcare, Milpitas, California) with low-energy, high-resolution collimators. All radionuclide images and associated data were processed according to the standard protocols.10,11 Myocardial perfusion images were interpreted using the standard 17-segment model.9,10,11

At least two experienced nuclear cardiology specialists interpreted each segment semi-quantitatively using 0 for normal uptake, 1, 2, and 3 for mild, moderate, and severe reduction of uptake, respectively, and 4 for absent uptake. Summed stress, summed difference, and summed rest scores (SSS, SDS, and SRS) were calculated. High-risk markers were reported, including increased lung-to-heart ratio, transient left ventricular cavity dilatation, and abnormal regional and global wall motion abnormalities. Normal SPECT MPI studies were defined as follows: ejection fraction above 50%, no left ventricular dilatation, and normal myocardial perfusion images (SSS < 4 and SDS = 0).

CT Imaging Protocol

CT scanning was performed on either a Toshiba Aquilion One 320 (Toshiba America Medical Systems, Tustin, California) (N = 96) or a Philips ICT 256 (Philips Healthcare, Bothell, Washington) (N = 58). Gated coronary CT scans were performed with 2.5 cm slice thickness, 50% overlap, 120 kV, 50 to 100 mA, and 90 to 180 mAs tube current, automatically selected by Toshiba Sure Exposure Version V4.74ER016 or Philips iDose Version V3. Low-dose, whole-body CT scanning (from mandible to mid-calf) was obtained with a spiral acquisition, 1 mm slice thickness, 128 × 0.625 mm collimation, and pitch .899, kV 100 to 120. CAC was quantified using the Agatston scoring method with a Vitrea Workstation, Version 4 (Toshiba Medical Systems Europe, The Netherlands) and was confirmed when needed in OsiriX, Version 3.9 (Pixmeo, Switzerland) 64-bit image processing software.12 Thoracic aorta (above the diaphragm, TAo), abdominal aorta (below the diaphragm, AAo), and iliofemoral (ILF) calcium scores were also performed with OsiriX. A Philips Extended Brilliance Workspace CT workstation (Philips Healthcare, Bothell, Washington) was used for calcium scoring of the carotids. TAo, AAo, and ILF calcium were scored according to previously described methods and using coronary calcium scoring software.4,5,6,13 Hounsfield units ≥130 were used as the calcification threshold. Total body calcium (TBC) is the sum of calcium scores in the five vascular beds. The AAo and ILF calcifications of one CT scan could not be quantified because of inadequate image quality due to patient factors.

Statistical Approach

Results were collected, tabulated, and statistically analyzed using SPSS version 17 (Softonic International, San Francisco, California). Differences between continuous variables and categorical variables were tested with unpaired t tests and χ2 tests. For inter-group comparison, an ANOVA test was used. Binary logistic regression was used to determine the strength of the relationship between one dependent variable and a series of independent variables. Coronary (CA) and carotid (CR) vascular beds were defined as event-related vascular beds. A P value of <0.05 was considered statistically significant.

Results

The study population is characterized in Table 1. Mean age was 53.0 ± 10.6 years (range 22 to 79 years), 108 (70%) were male, and 91/154 (59%) complained of chest pain. Traditional risk factors of dyslipidemia, hypertension, tobacco use, and diabetes were prevalent. Patients with known CAD were excluded. Mean exercise duration (Bruce protocol) was 7.7 ± 2.6 minutes with a range of 1 to 16.6 minutes. Twenty-eight of 151 patients (19%) had an abnormal exercise ECG response and 5 (3%) complained of chest pain during exercise testing. CT mean radiation dose was 6.2 ± 1.8 millisieverts.

Table 1 Baseline characteristics of the study population (N = 154)
Full size table

Subclinical atherosclerosis was common. Of the 154 patients, 115 (75%) had ≥1 vascular bed calcification involvement. This was observed in 81/108 (75%) of men and 34/46 (74%) of women. The ILF and the AAo were the most frequently involved beds with calcifications increasing in frequency in an ascending “south to north fashion” in both sexes. The CR and CA were least affected, as illustrated in Figure 1. TBC score increased with age in an exponential fashion in men and women (Figure 2A, B). Beginning at age 50, the rate of rise was greater in women than men. The occurrence of calcification in all vascular beds also increased with increasing age. However, there was a substantial delay in the development of calcification in the CR and CA beds relative to the other beds (Figure 3). In these “event-related” arterial beds, the appearance of calcifications lagged the AAo and ILF beds by ≈10 years (Table 2; Figure 3).

Figure 1
figure 1

Adapted with permission from Laclaustra et al19

Prevalence of arterial calcification in five vascular beds among men and women.

Full size image
Figure 2
figure 2

(A) Mean total body calcium score among study population across age groups. (B) Mean total body calcium score among men and women across all study age groups. P < 0.001 between age groups

Full size image
Figure 3
figure 3

Mean calcium score (AU) in five vascular beds of interest across all age groups. Values are presented as means. P < 0.001 between age groups

Full size image
Table 2 Prevalence and mean calcium score of five vascular beds across age groups in the total study population, men, and women
Full size table

The calcium scores in the TAo, AAo, and ILF beds were considerably higher than those in the CA and CR beds (Table 2). Upon reaching age 70, almost all arterial beds were involved. The number of vascular beds involved increased in a progressive fashion with increasing age for both sexes (Online Appendix Figure 1).

Figure 4A demonstrates the prevalence of calcification in the four non-CA vascular beds based on the presence or absence of CAC and Figure 4B demonstrates the mean calcium score similarly stratified. Even when the coronary calcium score was 0, calcification was frequently present in other vascular beds with the ILF bed most often involved. The frequency of calcification in patients with CAC = 0 was 42% ILF, 36% AAo, 29% TAo, and 7% for CR beds (Figure 4A). A total of 44/83 (53%) of patients with a CAC = 0 had calcifications in other beds. The mean TBC for patients with CAC = 0 was 831 AU (Figure 4B).

Figure 4
figure 4

(A) Prevalence of calcification in four study arterial beds based on the presence of coronary artery calcium (CAC). (B) Mean calcium score in four arterial beds based on the presence of coronary artery calcium. Values are mean ± SE. P < 0.001 between CAC = 0 and CAC > 0

Full size image

Online Appendix Figures 2A and B demonstrate that patients with calcium in the CA and CR arteries frequently had calcification in non-event-related beds and that mean calcification scores were much higher in these beds compared to patients without calcium in the event-related beds.

Lastly, the mean calcium score of four vascular beds similarly increased when stratified by increasing degree of CAC as shown in Figure 5, P < 0.001.

Figure 5
figure 5

Mean calcium score within four arterial beds based on the degree of coronary artery calcium (CAC). Values are presented as mean

Full size image

Discussion

This study showed that patients without known CAD and a normal MPI frequently have subclinical atherosclerosis on whole-body CT in the form of arterial calcification. Even in patients without CAC, atherosclerosis was frequent in other vascular beds. Atherosclerosis was most frequent in the ILF bed and progressed with age in an ascending “south to north fashion” to involve the aorta, coronary, and carotid arteries predominantly later in life. Our findings are consistent with those of the PESA study, in which subclinical atherosclerosis, as assessed by 2- and 3-dimensional ultrasound, was frequent in an asymptomatic middle-aged cohort not undergoing MPI. The prevalence of atherosclerosis in the PESA population was comparable to this Egyptian cohort, and most common in the ILF bed (44%), followed by the CR (31%) and AAo (25%) beds, with CAC present in 18%.14 Our data extend the work of Allison and colleagues in which increasing age was a consistent predictor of atherosclerotic disease as assessed by whole-body CT across all beds.15

The distribution and progression of atherosclerosis in autopsy studies was described in the 1950s by Roberts and Wilkins.16,17,18 Atherosclerosis of the common iliac arteries was observed as early as the second to third decades of life.16,17,18

The recent Aragon Workers’ Health Study by Laclaustra et al also suggests that early atherosclerotic involvement of the femoral and iliac arteries, as assessed by ultrasound, might be used as an early, sensitive diagnostic marker for pre-clinical atherosclerosis.19 These investigators observed that femoral subclinical atherosclerosis was associated with a higher risk of CAC than was carotid atherosclerosis. Area under the receiver operating characteristic curve for prediction of positive (≥1) or high (≥300) CAC were 0.706 and 0.810, respectively, for risk factors and femoral plaques. This was statistically significant compared to prediction using traditional risk factors.19

Currently, non-contrast thoracic CT is increasingly used to detect CAC among patients referred for MPI.20 Patients with normal MPI and a CAC of 0 have a better intermediate-term prognosis compared to patients with normal MPI and CAC > 0.13,21,22,23,24,25 The discovery of pre-clinical CAD offers the opportunity to intervene with preventive strategies.26,27,28,29,30

The current study, PESA Study, and the Aragon Workers’ Health studies have demonstrated the frequent presence of atherosclerosis in extracoronary vessel beds in individuals without CAC. Thus, should patients with normal MPI and asymptomatic patients with atherosclerosis in extracoronary beds, but without CAC, be considered for early preventive therapy?14,19,31,32 We currently treat asymptomatic patients using a non-specific risk factor model. This is often implemented using calculators such as the Framingham risk calculator or application-based ACC/AHA 10-year and lifetime risk factor calculator for atherosclerotic cardiovascular disease.33 A patient-centered paradigm based on the presence of atherosclerosis could prove to be more accurate than a population-based risk factor model.34 If such an approach was used, the findings of this study and the PESA and Aragon Workers’ Health studies suggest that the arteries with the earliest onset of atherosclerosis, the ILF and AAo arteries, would be prime targets for assessment.14,19

Limitations

This study is a single-center study involving a selected population of a moderate number of symptomatic patients referred for MPI. It thus represents patients referred for MPI rather than the general population. Also, while all MPI exams were normal in the patients in our study, MPI does not have 100% sensitivity to detect ischemia. Thus, some of our patients may indeed have had symptomatic obstructive CAD.35 Calcium scoring is also an indirect surrogate for atherosclerotic disease and clinical outcomes are not available in our particular patient cohort. While the study raises the possibility that utility of CT as a modality for calcium scoring outside of the heart may be useful for the early diagnosis of atherosclerosis, this approach involves radiation exposure and the challenge of incidental findings.36 Without further work, it is not possible to evaluate the risk/benefit ratio of such a strategy. Other non-CT approaches for diagnosing extracardiac, pre-clinical atherosclerosis might be developed and more fully evaluated in such a strategy. For example, ultrasonography of the iliofemoral arteries, as performed by the PESA and Aragon Workers’ Health study investigators, or potential novel techniques might be considered.

Conclusions

Quantitative assessment of calcification in different vascular beds demonstrates that many patients without known CAD and normal myocardial perfusion, even those with CAC = 0, often have significant atherosclerosis in extracoronary vessel beds. Atherosclerosis appeared first in the iliofemoral bed and then, in an ascending, “south to north fashion,” to other arterial beds. ILF calcification correlates strongly with coronary atherosclerosis. Imaging the extracoronary vascular beds provides a sensitive method for identifying pre-clinical atherosclerosis, prior to the development of CAC. The evaluation of atherosclerosis prior to its onset in the coronary arteries may lead to better understanding of the progression and natural history of atherosclerosis in its long, pre-clinical phase. It also raises the potential for a more patient-centered approach to prevention compared to the traditional use of CAC or targeting preventive measures based on more non-specific risk factors.

New Knowledge Gained

Subclinical extracoronary atherosclerosis was identified in many patients with normal MPI and without coronary calcium. Atherosclerosis most frequently appeared first in iliofemoral bed and typically progressed in an ascending fashion. Assessment of extracoronary calcification creates the opportunity to consider a more patient-specific paradigm than traditional risk factor evaluation models. Further investigation of the utility and practicality of imaging of such a paradigm is warranted.