The SVC originates from the confluence of the two innominate veins at the level of the cartilaginous portion of the first right rib. It descends into the anterior portion of the visceral compartment of the mediastinum and enters the right atrium. Its trunk has an average length of 7 cm and transverse diameter of 2 cm. It is adjacent to the thymus gland and right pleura and lung anteriorly; the right lateral tracheal lymphatic chain, pulmonary artery, and superior pulmonary vein posteriorly; the ascending aorta medially; and the right pleura, phrenic nerve, and small superior diaphragmatic vessels laterally.

The caval–atrial junction is within the pericardium. The serous pericardium englobes the anteroexternal surface of the SVC for a length of 2 cm. The sinus node is located along the anterolateral aspect of the junction between the SVC and the right atrium. It is superficial, lying just beneath the epicardial surface in the sulcus terminalis, and measures approximately 15 by 5 by 1.5 mm. The medial area lying between the intrapericardial SVC and ascending aorta includes (a) an extrapericardial region, where the origin of the right main bronchus lies, and (b) the anterior aspect of the pulmonary artery lying behind the SVC and the right orifice of the Thiele sinus.

McIntire and Sykes⁹ described four main collateral routes of the SVC in humans: (a) the azygos venous system (the only collateral draining directly into the posterior surface of the SVC above the right pulmonary artery and main bronchus); (b) the internal thoracic venous system (where the blood pours into the IVC from the internal thoracic vein through the superior and inferior epigastric veins, and external and common iliac veins); (c) the vertebral venous system (where the blood from the sinus venosus and bilateral brachiocephalic veins flows into the intercostal, lumbar, and sacral veins and then pours into the IVC—portions of this blood flow into the internal thoracic veins); and (d) the external thoracic venous system (this is the superficial collateral system where the blood from the subclavian vein and the axillary vein reaches the lateral thoracic vein and then pours into the femoral vein through the thoracoepigastric and superficial epigastric veins) (see Chapter 180).

The first clinical experiences with SVC surgical replacement, by Thomas¹¹ and Salsali,¹⁰ reported almost uniformly a cerebral edema nearly 60 minutes after the interruption of the venous flow. A plausible explanation of this phenomenon was the following: (a) venous stasis at the level of the cephalic territory, (b) problems in the absorption capacity of the cerebral fluid, (c) cellular hypoxia and hypercapnia resulting from the vascular stasis, and (d) modification of the permeability of the cerebral vessels leading to vasogenic cerebral edema. Recently, however, the present authors³ have provided evidence that the hemodynamic repercussions of SVC clamping depend on whether or not the SVC is obstructed.

For patients whose SVC is completely obstructed or tightly stenosed, intraoperative venous clamping results in a negligible hemodynamic compromise, since a functioning collateral venous network already exists and supplements the flow obstruction of the SVC. However, although Masuda and coworkers⁶ showed that 1-hour clamping of a nonobstructed SVC was well tolerated in a nonhuman primate model, this has not been confirmed clinically.³ In these circumstances, venous clamping time beyond 45 minutes is poorly tolerated from a clinical standpoint, and a longer period of time might induce irreversible brain damage, especially after ligation of the azygos vein (almost always necessary during total SVC replacement). The reason behind this is that the reduced return of venous blood flow to the right heart causes a cascade of hemodynamic events leading to reduced cardiac output and cerebral perfusion pressure (the safe
physiologic level of which should be >60 mmHg), as noted by McDowall.⁸

For some patients whose invaded SVC is not obstructed, even a brief venous clamping may trigger a hemodynamic cascade of events, including decreased cardiac inflow and outflow, increased pressure in the cerebral venous system, and alterations of the cerebral arteriovenous gradient, leading to irreversible brain damage and intracranial bleeding. Several technical details may mitigate this hemodynamic instability in the nonobstructed SVC (e.g., pharmacologic agents and fluid implementation, shortening the venous clamping time, and anticoagulation therapy). Intraluminal shunting of the blood from one of the brachiocephalic veins into the right atrium may reduce the hemodynamic consequences of venous clamping in animals, as reported by Gonzales-Fajardo and coworkers,⁴ and in humans, as reported by Warren and colleagues,¹² for at least 35 minutes; unfortunately, the mean clamping time of the SVC during an excision of the lung or of mediastinal tumors is usually longer than this.

The SVC is usually subject to easy obstruction owing to its anatomic site, thin wall, low hemodynamic pressure, and encirclement by chains of lymph nodes draining the entire right thoracic cavity as well as mediastinal tissues. The main indications and contraindications for SVC resection and reconstruction are outlined in Table 179-1. Major elective SVC reconstructive procedures should be limited to mediastinal tumors and, according to McCormack,⁷ to less than 1% of operable patients with right-sided bronchial carcinomas invading the SVC directly or, as we⁵ have noted, to an even lesser extent when the invasion is the result of extension of the tumor from involved superior mediastinal lymph nodes. Palliative SVC procedures are rare and limited to slow-growing diseases like mediastinal primary or secondary fibrosis, SVC thrombosis of unknown etiology, or saccular SVC aneurysms.

Patients most suitable for elective SVC prosthetic reconstruction are limited to those with operable anterior mediastinal tumors or right-sided bronchial carcinomas without hemodynamically significant venous outflow obstruction. The caval reconstruction requires great technical expertise and, although it likely adds significant procedural time, the long-term survival and patency of SVC prosthetic reconstruction appears to exceed the related morbidity and operative mortality. This has recently been pointed out by Yildizely and Dartevelle¹³ in their report of 39 patients with non-small-cell lung carcinoma invading the SVC: median survival time was 19 months, and the actuarial 5- and 10-year survivals were 29.4% and 22.1%, respectively, with a 7.7% early mortality (related to right pneumonectomy).