Percutaneous Creation of Arteriovenous Shunts

Abstract

Reductions in systemic vascular resistance and increased cardiac output are thought to help maintain tissue oxygen delivery during hypoxia. Here we discuss the rationale for creating arteriovenous (AV) shunts in humans and we describe novel techniques for percutaneous creation of AV shunts in patients. Although large AV shunts have been associated with cardiac failure, we believe that the creation of a moderate arteriovenous fistula (AVF) might benefit selected patients with respiratory disease. Here we study the acute effects of an AVF on mixed venous oxygen content and arterial oxygenation during hypoxia in pigs. Closure of the AVF resulted in a 15% reduction in cardiac output.

Introduction

In the 1960s, Holman and colleagues at Stanford described a series of experiments revealing that small-to-moderate fistulae with rigid borders typically remain stationary in size without a deleterious effect on hemodynamics.^1,2 In contrast, fistulae with a distensible rim progressively dilate, leading ultimately to cardiac dilatation and, in some cases, cardiac failure.³ Generally, except for vascular access in dialysis patients, the presence of an arteriovenous (AV) shunt promotes discussion about the merits and risks of surgical closure. Recently, we have explored the concept that the creation of a moderate-sized AV shunt might benefit selected patients with cardiac and respiratory disease. In this paper we will first explain the rationale behind the therapeutic potential of AV shunts; then, we will describe novel techniques for the percutaneous creation of AV shunts; and finally, we hope to balance the benefits against possible risks associated with AV shunts.

The rationale for creating AV shunts to improve exercise capacity. Since exercise performance is related to cardiac output, our rationale for creating AV shunts to improve exercise capacity is based on a triad of physiology principles. (Q is cardiac output, P is the mean arterial pressure minus the central venous pressure, SVR is systemic vascular resistance, HR is heart rate, SV is stroke volume, VO2 is oxygen consumption, CaO2 is the oxygen content of arterial blood, and CvO2 is the oxygen content of mixed venous blood).

Equation 1: Q liters per minute-1 = dP Pascals / SVR Dynes.

Equation 2: Q liters. minute-1 = HR beats per minute -1 x SV liters. beat -1.

Equation 3: Q liters. minute -1 = VO2 grams per minute-1 / CaO2 – CvO2 (grams. liter -1).

In engineering terms, an AV fistula is a low-resistance circuit allowing flow from a high-pressure arterial system to a low-pressure venous system, thus bypassing the higher resistance capillary beds. Creation of an AV shunt lowers systemic vascular resistance (and to a lesser extent arterial diastolic blood pressure), resulting in an increased cardiac output (Equation 1). A chronic increase in cardiac output increases stroke volume, thus allowing an amplified cardiac output response to changes in heart rate (Equation 2). An amplified cardiac output response to exercise-induced tachycardia allows for greater oxygen delivery and consumption, which is essential for exercising muscles (Equation 3). Maximal cardiac output is closely related to maximal oxygen consumption, which in turn is related to maximum exercise capacity.

In a hypothetical example, a 40-year-old, 70 kg, man has a resting heart rate of 70 beats per minute (bpm)-1 and a stroke volume of 70 ml, and a cardiac output of 4.9 liters per min (lpm).-1 During maximal exercise his heart rate increases to a maximum of 180 bpm-1 (220 minus age bpm-1), and his cardiac output increases to approximately 12.6 lpm-1 After the creation of an AV fistula, his resting heart rate remains approximately 70 bpm.-1 His stroke volume will have increased to 85 ml, so that his resting cardiac output will be 5.95 lpm-1 (approximately 1 lpm-1 will flow through the AV shunt, so that the effective cardiac output to the body at rest will remain at approximately 4.9 lpm;-1 and the pulmonary blood flow will effectively increase by approximately 1 lpm-1). However, during maximal exercise when his heart rate increases to 180 bpm,-1 his cardiac output will be 15.3 lpm.-1 Since approximately 1 lpm-1 will still flow through the AV shunt, the effective cardiac output to the body would be 14.3 lpm-1: an increase of 1.7 lpm-1 over the maximum of 12.6 lpm-1 before the creation of the AV shunt. This will allow for an increased oxygen consumption of 250 ml per min-1 (since the ratio of change in cardiac output to change in oxygen consumption is generally about 6:1).

The normal cardiovascular response to hypoventilation is to increase cardiac output and reduce systemic vascular resistance — related in part to the effects of hypoxemia and hypercarbia on the autonomic nervous system.⁴ The net result is to increase cardiac output and hence tissue oxygen delivery to compensate for the hypoventilation. We believe that a moderate AV fistula might replicate or even amplify this response and thus benefit patients with respiratory disease. Interestingly, the increase in cardiac output in these circumstances does not lead to a measurable increase in cardiac oxygen consumption because the workload of the contracting heart is reduced because of a drop in afterload.⁵ A controlled afterload reduction (approximately 10 mmHg reduction in diastolic pressure) could theoretically increase cardiac power output, another index that is related to exercise capacity. Since the physiologic effect size of a fistula is closely related to its size, we have recently studied the effect of a moderate-sized fistula in young pigs. We created a side-to-side anastomosis between the iliac artery and iliac vein in five young pigs (60 kg body mass). The size of the fistula ranged from 8–12 mm in length and 3–4 mm in diameter. The flow through the shunt (as measured by Doppler ultrasound) ranged from 0.8 to 1.3 lpm-1 For closure of the fistula, a balloon catheter was inserted into the femoral artery and advanced under fluoroscopic guidance to the site of the fistula and then inflated to occlude blood flow through the fistula. Predictably, an open shunt (compared to closed) was associated with a 15% (10 mmHg) reduction in mean systemic arterial pressure and a 15% (2 lpm-1) increase in cardiac output. Under hypoxic conditions, an open shunt was associated with a 10% increase in arterial oxygen saturation, and a 20% increase in mixed venous oxygen saturation. We believe, therefore, that creation of a moderate AV fistula leads to significant changes in cardiovascular and oxygenation indices and allows for the recirculation of blood through the lungs.

Novel techniques for the percutaneous creation of AV shunts. Based on the above, we estimate that an AV shunt in the iliofemoral region with a diameter of 5 mm will allow blood flow of between 0.8 and 1.5 lpm-1 In order to create such a shunt, we have developed and tested a system that provides access between the iliac artery and the vein (via a novel crossing needle, delivery system and vascular implant). Using this system, it is possible to deploy a novel device that is positioned between the artery and its accompanying vein. This device is a metallic, self-expanding nitinol coupler that maintains a fistula between the artery and vein with a lumen of approximately 5 mm in diameter. We have already successfully deployed this device using percutaneous techniques in patients with end-stage chronic obstructive pulmonary disease — patients who have few therapeutic options for their profound problem of oxygen transport during exercise.

The risks of AV shunt creation. The theoretical risk associated with AV shunt creation is related both to shunt size and location. In this paper, we will not discuss intracerebral or intrapulmonary AV shunts because these carry complications that are specific to their anatomical location. In selected patients with cardiovascular and pulmonary disease, we propose to create arteriovenous fistulas (AVF) in the iliofemoral region. The hazards of systemic AV shunts have been reported in three ways: (1) case reports that describe the development of large AV or aortocaval fistulas after trauma or aneurysm rupture leading to high output cardiac failure;^6–10 (2) a body of literature that describes the complication rates of AVF that are created for vascular access in hemodialysis patients;^11–15 and (3) papers that describe the change in physiology after creation of AVF in animal models.^16,17 AVF reliably increase cardiac output, cardiac stroke volume,⁵ and reduces diastolic blood pressure. Although millions of patients have had AVF created (primarily for vascular access in hemodialysis patients), few develop complications. Creation of an AVF, at least in the population of hemodialysis patients, appears to be acceptable and safe.¹⁸ The risk of cardiovascular events and cardiac failure are not significantly increased in hemodialysis patients with AVF that allow approximately 800 ml per min-1 of blood flow.¹⁸ In theory, any increase in cardiac output via creation of an AV shunt could increase the relative percentage of blood flow that perfuses poorly ventilated regions of the lungs and thus worsen hypoxemia via worsening ventilation perfusion (V/Q) inequality). This phenomenon has been widely reported in several models of pulmonary edema and lung injury.^19–21 Such an effect, however, is believed to be more dependent on the level of oxygenation than the absolute change in cardiac output. Changes in V/Q inequality with cardiac output are great in hyperoxia, but several fold less in normoxia and hypoxia. In one study of lung injury in dogs, for example, relatively large increases in V/Q inequality (> 10% per liter of cardiac output increase) are seen when an AVF is opened under conditions of hyperoxia, while far less ventilation perfusion inequality (only 2% per liter of cardiac output increase) occurs when an AVF is opened under conditions of hypoxia.²¹ Thus, in the setting of respiratory or cardiac disease, any potential increase in V/Q inequality might be outweighed by the increase in mixed venous oxygen saturation as a result of the AVF. Indeed, prior studies have demonstrated higher arterial oxygen saturation and paO2^5,22 in healthy animals with an AVF, when compared to animals without a fistula. Large AVF can cause large increases in cardiac output, resulting in the development of congestive heart failure. The balance between volume overload and urinary sodium excretion in this clinical situation is determined by the balance between vasoconstrictor neurohormonal systems (the renin-angiotensin system, sympathetic nervous system [SNS], the endothelin system, and arginine vasopressin) and compensatory activation of vasodilating systems: atrial natriuretic peptide (ANP) and nitric oxide (NO). In patients who develop heart failure, enhanced activities of the sodium-retaining systems are thought to overcome the effects of the natriuretic systems, leading to sodium and water retention.^23,24 Since AVF increase cardiac output and venous return, the creation of AV shunts might be thought to lead to adverse outcomes — especially in patients with very severe COPD who at a high risk of cardiovascular and peripheral vascular disease^,25,26 and pulmonary arterial hypertension.²⁷

It is reassuring that a large clinical experience with two groups of patients exists that suggests AV shunts do not significantly lead to pulmonary vascular disease: (1) patients with congenital left-to-right shunts and (2) patients who have an AVF created for vascular access (hemodialysis patients). Patients with left-to-right shunts (e.g., congenital ventricular septal defects and atrial septal defects) are only at risk of developing PAH when the size of the shunt is large.^28,29 Patients with large ventricular septal defects who survive more than 20 years usually have pulmonary hypertension with associated right ventricular failure.²⁸ Similarly, in contrast to large atrial septal defects, a small atrial septal defect (Conclusion Although the formation of an AVF after blunt or penetrating trauma has been associated with the development of cardiac failure, we are currently exploring the concept that the creation of an AV shunt might have therapeutic potential. When planning the percutaneous creation of an AVF, we consider it wise to assess the procedure-related challenges, which vary between individual subjects (including vascular anomalies, vessel calcification, and bleeding disorders), in addition to the physiologic challenge of having a permanent AV fistula in situ. We believe that a moderate AV fistula will be well tolerated by most, and will provide benefit in selected patients. Several physiologic effects of AV shunt creation, including controlled reductions in diastolic pressure, increases in stroke volume, and increases in maximum cardiac output, might benefit patients with reduced exercise tolerance, particularly patients in whom oxygen delivery is a key limiting factor. The use of an AVF as a therapy merits further study.