**1. Introduction**

According to the American Heart Association, "cardiovascular disease is the leading global cause of death", accounting for more than 17.6 million deaths in 2016, a number that is expected to grow to more than 23.6 million by 2030 [1] In the event of acute coronary syndrome, percutaneous coronary intervention (PCI) has been shown to be beneficial on outcome [2]. The beneficial effect of PCI on the course of chronic stable coronary artery disease (CAD) has, so far, not been proven ye<sup>t</sup> [3]. A recently published randomized controlled trial among patients with stable, single-vessel CAD, the so called ORBITA trial (e.g., Objective Randomised Blinded Investigation With Optimal Medical Therapy of Angioplasty in Stable Angina) [4], found that PCI of the stenotic lesion did not prolong exercise time by more than the effect of a sham procedure during the short observation period of six weeks. The new aspect of the ORBITA trial was a methodological one, that is, the use of a sham control group of patients undergoing the invasive coronary procedure, but not the actual PCI. The importance of a sham control group in interventional procedures is pivotal, especially in a population with a high level of suffering [5,6]. After all, it is known that the placebo effect can cause significant clinical improvements (e.g., an increased exercise duration of >90 s [7]). Coronary artery bypass grafting (CABG), on the other hand, has been found superior to PCI with respect to all-cause or cardiovascular mortality [8].

The number of patients with incomplete revascularization as well as so-called "no-option"-patients (i.e., patients without options for PCI or CABG still suffering from symptoms of CAD despite optimal medical therapy) is on the rise. It is estimated, that 30,000–50,000 new patients are affected in continental Europe per year [9] and Williams et al. reported a prevalence of 25.8% of incomplete revascularization in patients with CAD [10]. Apart from the limited quality of life, these patients also have a higher mortality at three years than patients with complete revascularization [10].

Accordingly, new therapeutic approaches are required. Because of the known survival benefit of patients with a well-developed coronary collateral circulation [11,12], interventions aiming at the promotion of coronary collaterals are a promising strategy. Coronary collaterals represent pre-existing inter-arterial anastomoses and as such are the natural counterpart of surgically created bypasses. To this end, biochemical (e.g., intracoronary vascular-endothelial growth factor or intravenous granulocyte-macrophage colony-stimulating factor) as well as biophysical (e.g., external counterpulsation) approaches have been evaluated for the promotion of those collaterals.

The aim of this review is to describe basic principles of the coronary collateral circulation, its extracardiac anastomoses as well as different therapeutic approaches, especially that of stimulating extracardiac coronary supply via permanent occlusion of the internal mammary arteries.

### **2. Basic Principles of the Human Coronary Collateral Circulation**

### *2.1. Coronary Collateral Circulation*

The development of the cardiovascular system during embryogenesis occurs by vasculogenesis, a process defined as "the de novo formation of blood vessels from endothelial precursor cells" [13]. Directed by the concentration of local messenger substance, endothelial precursor cells sprout out and start forming a dense vascular network with multiple anastomoses. The density of this network is at its peak in neonates and declines subsequently by physiological regression, a process called pruning [14–16].

Nevertheless, it has been hypothesized early on and tested that the coronary anastomoses of the neonate do not vanish completely but some collaterals rather recede in calibre. This concept has been decisively advanced by the findings of the Scottish pathologist W.F. Fulton, who found "numerous anastomoses in all normal hearts" by using a vascular overlay detecting technique with radiographic contrast medium containing uniform particles sized 0.5–2.0 μm to visualize even small arteries [17].

Interestingly, with changing vascular pressure- and resistance conditions, it is possible to recruit these receded arterial anastomoses. This process is often seen during the course of CAD with development of a pressure gradient across a stenotic lesion, which itself induces augmented flow in preformed arterial anastomoses and finally, structural augmentation of these collateral arteries (arteriogenesis). Accordingly, the prevalence of functional coronary anastomoses depends on the presence of CAD and is highest in chronic total coronary occlusions [16].

Coronary collaterals in patients without coronary atherosclerosis range in calibre between 10–200 μm; collateral arteries of patients with CAD are approximately four times bigger (100–800 μm) [17]. This observation is in accordance with an experimental rabbit model, where occlusion of the femoral artery increased the lumen diameter of pre-existent arterioles four- to fivefold [18]. "At the same time, the growth in structural size goes along with a decreasing number of collateral arteries, a process called pruning. Pathophysiologically and in the sense of the Hagen Poiseuille law, pruning may be interpreted as a way of effectively reducing vascular resistance to collateral flow" [13].

### *2.2. Extracardiac Coronary Supply*

Apart from inter-coronary arterial anastomoses, the human coronary arterial circulation is supplied by several extracardiac anastomoses, also called the non-coronary collateral myocardial blood flow (NCCMBF) [19]. Hence, the heart receives additional blood from the arteries of surrounding structures [20–24]. Most of the extracardiac anastomoses originate from arteries, which supply the pericardium [21] and these arteries are typically located at the sites of pericardial reflections (e.g., the entry of the caval veins or the exit of the grea<sup>t</sup> arteries) [22]. Thus, a well-known extracardiac anastomosis connects the right internal mammary artery (IMA, also called internal thoracic artery) to the right coronary artery via the pericardiacophrenic branch and the sinus node artery [25] (Figure 1). This extracardiac coronary supply can also develop after coronary bypass surgery as shown exemplary in Figure 2 [22].

**Figure 1.** Angiographic demonstration of extracardiac coronary supply. (**A**) Posterior-anterior projection of the right internal mammary artery (IMA, marked by \*) and its connection to the right coronary artery via the pericardiacophrenic branch (marked by +). (**B**) Lateral projection using the same markers. Noteworthy, additional branches of the IMA (marked by #) heading towards the heart.

**Figure 2.** Angiographic demonstration of extracardiac coronary supply after coronary artery bypass surgery. (**A**) Posterior-anterior projection of the left internal mammary artery bypass (marked by a \*) on the left anterior descending coronary artery (LAD, marked by a #). Upstream of the bypass anastomosis, retrograde filling of the LAD is incomplete revealing coronary occlusion, which triggered the arteriogenesis of the pericardiacophrenic branch (marked by a +) (**B**) Lateral projection using the same markers revealing the connection of the pericardiacophrenic branch with the third diagonal branch (marked by III).

Most commonly, NCCMBF originates from the bronchial or the internal mammary arteries [22]. Bjork et al. showed a prevalence for bronchial-coronary-anastomoses of more than 20% by reviewing 200 coronary angiographies [26]. According to this observation, most of the anastomoses connect to the left circumflex artery (LCX) and demonstrate poor blood flow. However, blood flow within an anastomosis between two arterial beds depends on the respective vascular resistances. Thus, a constant decrease of vascular resistance in one arterial bed causes an increased blood flow to it with associated arteriogenesis. Consequently and depending on the underlying pathology, bronchial-to-coronary (e.g., in the case of a chronic occluded coronary artery [27]) as well as coronary-to-bronchial anastomoses (e.g., during chronic pulmonary diseases [28]) have been described.

Additional evidence for extracardiac anastomoses comes from the work of Hudson et al., who, by injecting ink into the coronary arteries, demonstrated anastomoses with anterior mediastinal, phrenic and intercostal arteries as well as with esophageal arterial branches of the aorta [21].

NCCMBF has also been increasingly recognized by cardiac surgeons as they discovered that anastomotic blood flow can dilute, and thus, be a potential hazard to cardioplegia [23]. To quantify this phenomenon, several studies have been conducted with reported values of anastomotic perfusion ranging between 3.4 to 14 mL/100 g/min [29,30] during cardiopulmonary bypass with cross-clamping of the aorta.

### *2.3. Quantitative Evaluation of the Coronary Collateral Circulation*

The first in vivo functional coronary collateral measurements were conducted in the 1970s, showing a direct relation between "angiographic appearance and functional performance of coronary collaterals during bypass surgery" [31]. Rentrop et al. proposed a transluminal coronary angioplasty approach, which divided the appearance of coronary collaterals in four groups (0 = no collateral filling from the contralateral vessel to 3 = "complete filling of the epicardial segmen<sup>t</sup> of the artery") [32]. Unfortunately, the method is only qualitative and evaluation of extracardiac collaterals is not feasible.

Thereafter a method for quantitative coronary collateral function assessment based on coronary occlusive pressure measurements was introduced. The so called collateral flow index (CFI) [33,34] "is the ratio between mean coronary occlusive and aortic pressure both subtracted by central venous pressure as obtained during a 1-min proximal coronary balloon occlusion" [33] (Figure 3). The method is accepted as the reference method for functional collateral assessment in patients with chronic stable CAD [35,36]. In terms of sufficient collateral blood supply, it has been demonstrated that a CFI of >0.20–0.25 is related to absent signs of ischemia on the intracoronary electrocardiogram (i.c.ECG) during this 1-min coronary artery balloon occlusion [37,38].

CFI has also been determined in patients with angiographically normal coronary arteries, revealing functional collateral arteries "to the extent, that one fifth to one quarter of them (i.e., the patients without coronary stenoses) do not show signs of myocardial ischemia during the brief vascular occlusions" [39]. Those findings of functional sufficient collaterals even in the absence of CAD support the above mentioned pathoanatomic observations [17], that coronary anastomoses calibre remain functional to a considerable degree.

### **3. Angiogenesis and Arteriogenesis**

To understand the different therapeutic approaches for promoting the coronary collateral circulation, it is crucial to differentiate between two basic physiologic principles, that is, angiogenesis and arteriogenesis.
