CardioVascular and Interventional Radiology

, Volume 28, Issue 3, pp 372–376

Balloon Dilation of the Superior Vena Cava (SVC) Resulting in SVC Rupture and Pericardial Tamponade: A Case Report and Brief Review


    • Department of RadiologyMemorial Sloan Kettering Cancer Center
  • George I. Getrajdman
    • Department of RadiologyMemorial Sloan Kettering Cancer Center
Case Report

DOI: 10.1007/s00270-004-0001-0

Cite this article as:
Brown, K.T. & Getrajdman, G.I. Cardiovasc Intervent Radiol (2005) 28: 372. doi:10.1007/s00270-004-0001-0


Stent placement is an accepted primary treatment for SVC syndrome. Balloon dilation is frequently performed prior to stent placement. Although various stent-related hemorrhagic complications have been reported, as well as reports of iatrogenic catheter and guidewire perforations, there has been only one previous report of balloon dilation–related SVC rupture. We report a second case, including the clinical scenario, in the hope that should this complication occur, it might be recognized quickly and treated successfully.


Vena cava, interventional procedureSuperior vena cava syndromeVena cava, stenosis or obstructionAngioplasty complication

Malignant tumors in the chest may encase, narrow, and ultimately occlude the superior vena cava (SVC). Depending on how rapidly the narrowing progresses, patients with such tumors may present with SVC syndrome. This syndrome includes some or all of the following symptoms: facial, head, neck, chest and upper extremity swelling, shortness of breath, and headache. Historically, SVC syndrome secondary to malignancy has been treated with chemotherapy and/or radiation. As interventional radiologic techniques evolved, attempts were made to treat benign SVC syndrome by dilating the narrowed or occluded area with balloons. Because of the rigid nature of malignant tumors, this was not an effective method of treating malignant SVC syndrome. When metallic endoprostheses were developed, these stents were used and found to provide rapid and effective palliation of symptoms [13] with few complications. Most recently, stent placement has been advocated as the initial method of treatment [46] of malignant SVC syndrome.

Balloon dilation alone had also been used to treat benign central vein stenoses. As interventional radiologists became more involved with dialysis graft salvage, central vein stenoses related to myointimal hyperplasia in hemodialysis patients were discovered and treated. As in the case of malignant SVC obstruction, the results of balloon dilation alone for benign SVC strictures have been disappointing [7, 8], and these patients began to be treated with metallic endovascular stents [3, 9], with or without balloon dilation.

When balloon dilation of large, intrathoracic veins was initially undertaken, some operators expressed concerns regarding the possibility of rupture. As a body of literature developed, rupture of these major veins came to be viewed primarily as a theoretical possibility, with balloon “venoplasty” actually proving quite safe. In 1996, Complications in Diagnostic Imaging and Interventional Radiology stated that although overdistension of a thin-walled vein with an angioplasty balloon could theoretically result in perforation, “the only reported case of such transmural vascular injury was not immediately evident either radiologically or clinically, but was noted at autopsy in which the cause of death was not attributable to the stenting” [10]. In 1997 Funaki et al. did report two episodes of subclavian vein rupture at the time of dialysis graft thrombolysis that were successfully treated by placement of uncovered metallic stents [11]. In this report, no mention was made of hemodynamic instability, development of hemothorax, or other serious hemorrhagic sequelae, either before or after stent placement.

We report a case in which balloon dilation of tumor-related SVC obstruction, prior to placement of a self-expanding metallic stent, resulted in SVC rupture into the pericardium, pericardial tamponade, and death.

Case Report

A 65-year-old woman presented to a local physician with a 2 month history of worsening shortness of breath. In the previous 3–4 weeks she had also noted puffiness of her face, bilateral upper extremity swelling, as well as prominence of her anterior chest. Her shortness of breath had progressed to the point of dyspnea on moderate exertion. A chest radiograph demonstrated mediastinal fullness, and a chest CT without intravenous contrast revealed a right hilar mass, with right upper lobe atelectasis. Intravenous prednisone was begun, and the patient experienced significant improvement in the facial and upper extremity swelling. She was discharged 4 days later on oral prednisone and admitted to our hospital 4 days after that for further management and to initiate treatment. The patient underwent a contrast-enhanced CT scan of the chest that confirmed the right hilar mass and better defined a segment of narrowing involving the SVC resulting from compression by the tumor (Fig. 1A) just above the junction of the SVC with the right atrium (Fig. 1B), and identified a left adrenal mass. A CT-guided needle biopsy of the lung mass performed the day after admission diagnosed squamous cell carcinoma, and SVC stent placement was requested prior to initiating treatment.
Figure 1

A CT scan demonstrating the right paramediastinal mass, with compression of the SVC (arrow). B CT scan at the level where normal-caliber SVC (*) enters the right atrium.

The patient was given midazolam and meperidine for conscious sedation. The right neck was sterilely prepared and draped, and using ultrasound guidance, a 21 gauge needle was used to puncture the right internal jugular vein. A micropuncture introducer (Cook, Bloomington, IN) was placed into the internal jugular vein, and Iohexol 300 (Amersham Health, Princeton, NJ) was injected by hand while a digital run was performed. The introducer was exchanged for an 8 Fr vascular sheath, through which a 5 Fr directional catheter and a hydrophilic angled guidewire were used, with road-mapping guidance, to cross the 4 cm segment of high-grade narrowing that had been demonstrated in the SVC just above the right atrium (Fig. 2A). The catheter was advanced into the right atrium and injected with contrast to confirm intra-atrial positioning (Fig. 2B). The catheter and guidewire were manipulated into the inferior vena cava, and a second digital run was performed through the vascular sheath (Fig. 2C). The 8 Fr sheath and directional catheter were exchanged over an Amplatz Extra Stiff guidewire (Boston Scientific, Natick, MA) for an 11 Fr sheath, through which a 16 mm PE-MT low-pressure balloon (Boston Scientific, Natick, MA) was used to predilate the narrowed segment of SVC. During the brief balloon inflation, the patient appeared to experience discomfort, as she raised her hand up to her chest while the balloon was being inflated. The balloon was deflated and removed, and a postdilation digital run was performed, injecting the vascular sheath with Iohexol 300. A 16 mm by 4 cm Wallstent (Boston Scientific, Natick, MA) was advanced over the guidewire and deployed across the narrowed area. The vascular sheath was injected with contrast, and excellent flow was noted through the stent, with restoration of a good luminal diameter. Fluoroscopically, however, there seemed to be diminished motion of the cardiac silhouette compared with normal.
Fig.  2

A Image from the venogram demonstrates a segment of narrowing of the SVC (dots). Open black arrow, innominate vein; open white arrow, azygous arch. B Image obtained during concomitant injection of contrast material into the sheath above the short segment narrowing and an end-hole catheter in the high right atrium (outlined by small dots) that was used to cross the narrowed area. C Venogram during contrast injection of the sheath, after the catheter and guidewire had been advanced into the inferior vena cava, before balloon dilation of the SVC. No extravasation was evident at this point in procedure.

The vital signs at that time were normal, including the heart rate, as recorded electronically, and the blood pressure. In retrospect, there was noted to be slightly diminished amplitude of the ECG complex. The patient was found to be pallid and nonresponsive, without palpable pulses. The oxygen saturation was dropping, and cardiopulmonary resuscitation (CPR) was initiated. Romazicon, 0.2 mg, was administered intravenously while a code was called. Primary entities included in the differential diagnosis at that point included oversedation, pulmonary embolus, or acute myocardial infarction. Once resuscitative measures had been assumed by the code team with anesthesia support, the images obtained during the procedure were reviewed at the workstation. At that time, reviewing both the subtracted and unsubtracted images, contrast material outside the confines of the vascular system was appreciated. In particular, contrast could be seen outlining the ascending aorta (Fig. 3). The possibility of cardiac tamponade secondary to hemopericardium was entertained, and an ultrasound examination confirmed the presence of fluid within the pericardial sac. An 8.5 Fr catheter was percutaneously placed from a sub-xyphoid approach, and blood was evacuated from the pericardial sac. Repeat ultrasound examination demonstrated persistent or recurrent fluid within the pericardium, and a second catheter was placed. At that point it became obvious that further resuscitative measures were unlikely to be effective, and CPR was discontinued.
Fig. 3

Image obtained immediately following balloon dilation of the SVC with a 16 mm balloon. Note the contrast material outlining the ascending aorta (dots), and visible outside the confines of the right atrium.

At autopsy, a tear of the intrapericardial portion of the SVC was identified, through which the Wallstent was visible. There was evidence of squamous cell tumor infiltrating the SVC, except on the anterior aspect of the intrapericardial SVC, where the tear occurred. There was an estimated 300 cm3 of blood in the pericardial sac, despite placement of the drainage catheters.


Malignant obstruction of the SVC is most commonly seen in patients with lung cancer who have advanced disease and limited life expectancy. Since the use of metal stents to treat SVC occlusion was first reported [12] in 1986, these stents have become an accepted method of treating SVC syndrome, and recently have been advocated as the first treatment option in symptomatic patients with SVC obstruction secondary to malignancy [4, 5]. Chemotherapy and radiation therapy had formed the mainstay of treatment in the past. Unfortunately, both these methods have side effects, and may take 3–4 weeks to produce results. In a series of 52 patients with SVC syndrome treated with stents by Lanciego et al. [5], complete (80%) or partial (20%) resolution of symptoms was seen within 72 hr, and in many cases relief was virtually immediate. Since SVC stenting has been considered safe by most authors [1, 6, 9, 10], and since this patient population has limited life expectancy, it makes sense to treat them in a manner that can be expected to result in the most expedient improvement of their symptoms, with very little time in hospital.

In our patient, SVC rupture was clearly caused at the time of balloon inflation, prior to stent placement. The SVC was not occluded, but rather there was a short segment of severe narrowing, with an easily negotiated patent channel. Once this segment had been traversed, the guidewire was advanced into the inferior vena cava, and at no time was that guidewire pulled back to, or above, the level of the right atrium. Having crossed the narrowed channel uneventfully, a digital run was performed prior to balloon dilation that, even in retrospect, demonstrates contrast material only within the confines of the vascular system. Although the patient appeared to experience discomfort during the balloon inflation, this is commonly seen when balloons are used to dilate tubular structures within the body. We have seen this previously, it has been reported [1, 2] during SVC dilation, and we did not think it unusual. The patient’s subsequent clinical deterioration, however, was dramatic, and temporally related to the balloon inflation, and contrast material was unequivocally demonstrated within the pericardial sac on the digital run obtained just prior to stent deployment.

In the two most recent articles advocating stent placement as the initial method of treating malignant SVC syndrome, one group routinely dilated the lesions prior to stenting [5], the other group did not [6]. Even in the group that did not routinely predilate lesions, 13 of 52 patients subsequently required balloon dilation, and in these patients balloons 1 mm larger than the “normal” vessel diameter were chosen. The authors typically used stents 14–16 mm in diameter, and presumably used similar-sized balloons. Six of their patients were dilated when their stents failed to expand to greater than 50% of the original vessel diameter. In all the other reports cited [13, 6, 7, 9] balloons were used to predilate the area to be stented. Balloon dilation was felt to be important for several reasons. Many authors commented on the value of predilation for accurate localization of the stenosis [1, 2, 3, 6] and estimation of the length of vessel involved [1]. In addition, some authors [1, 3, 6, 10] felt balloon dilation delineates the “toughness” of the stricture.

We do not routinely dilate self-expanding stents when such stents are placed within the biliary tree, preferring instead to leave a catheter across the stent, and wait 24 hr for the stent to expand on its own. This spares the patient the discomfort of the dilation, while the covering catheter allows us to preserve access to the biliary tree in the event that the stent expands asymmetrically and a second stent is required. However, it has been our experience that sluggish flow resulting from an inadequate lumen obtained at the time of SVC stent placement can lead to thrombosis within the stent, necessitating thrombolysis. The presence of brain metastases, thrombocytopenia, or other hemorrhagic problems that occur in many patients with advanced lung cancer often makes the administration of thrombolytics, or even anticoagulants, risky. Since it is impossible to know ahead of time how resistant a narrowed or occluded region might be, there is no way of predicting which stents will expand adequately. If only a narrow channel is achieved after stent deployment, thrombus formation within the stent can be virtually instantaneous. Assuming thrombus does not form, but the stent is felt to be inadequately expanded, post-placement dilation can be performed. Unfortunately, advancing a balloon through a narrowed stent can cause the stent to be displaced. In addition, we often place venous access devices at the time of SVC stenting. The catheter of such a device can further compromise an inadequately expanded stent lumen, again, with resultant thrombosis. For these reasons, we continue to predilate, although we now use balloons only 10–12 mm in diameter so as to ensure stent expansion adequate to avoid thrombosis while, it is hoped, minimizing the risk of rupture.

Once predilation of the SVC has been decided upon, many methods of balloon sizing have been proposed. When dilating arteries, balloons are chosen to approximate the size of the vessel, or in some cases to overdilate by up to 10%. When SVC dilation is undertaken, various schemes have been used to choose a balloon. Balloons of 12–20 mm were used to predilate the SVC in the series cited above [13, 79], and at least one author has said “valvuloplasty balloons with a diameter larger than 2 cm can be used” for the SVC [1]. Another author dilated all patients with a 12 mm balloon [2], and one author chose a balloon 1 mm smaller than the nominal diameter of the completely expanded stent [9]. We typically have chosen our balloon to approximate the diameter of the stent we intend to use, which has been chosen to be slightly smaller than the predicted or measured normal diameter of the SVC. In the patient who forms the basis of this report, the SVC measured 20 mm below the confluence of the brachiocephalic veins, above the region of SVC narrowed by the tumor.

Until this time, although it has been thought conceivable that overdistension with a balloon could result in vessel rupture, this possibility has been said to “have never been seen or reported” in clinical practice [1, 10]. In the report by Funaki et al. of two episodes of subclavian vein rupture that occurred while treating central vein stenoses at the time of dialysis graft thrombolysis, “rupture” was defined as “extravasation into the soft tissues accompanied by vessel wall irregularity and poor antegrade flow” [11]. These episodes were successfully treated by placement of uncovered metallic stents, and no mention was made of hemodynamic instability, development of hemothorax, or other serious hemorrhagic sequelae, either before or after stent placement.

It had not been our practice to include the possibility of death from this cause when obtaining informed consent for SVC stent placement. Eight reports of hemopericardium or cardiac tamponade associated with SVC stenting were found on review of the literature. In three of the seven reports hemopericardium was clearly related to perforation by catheters or guidewires, but not balloon inflation, prior to stent placement. In one case there was clearly extravascular passage of the guidewire when the occlusion was traversed [13], another occurred after stent placement, when a temporary pacing wire was being inserted to treat third degree heart block that developed during stenting [14], and the third was cursorily reported as being caused by “iatrogenic SVC perforation” without further details [10]. Four cases of hemopericardium followed stent placement and occurred 15 min to 6 months after the procedure [1517]. Thus, none of these six were related to balloon inflation. Only one case of SVC rupture clearly related to balloon dilation [19] occurred in a patient being treated for presumed benign SVC stenosis, when a 10 mm balloon was used to predilate the stenosis prior to stent placement. Fortunately, the complication was recognized and successfully treated with a Wallgraft (Boston Scientific, Natick, MA). This report was made in a publication that may not have as widespread an audience as the clinical relevance of the report warrants. Given our recent experience, it seemed worthwhile engaging a larger audience.

In our patient it is likely that the asymmetric nature of tumor ingrowth into the wall of the SVC, with the anterior wall remaining relatively spared, led to rupture of this thinner, less rigid portion during balloon inflation. Unfortunately, in many cases of malignant SVC syndrome it is impossible to predict exactly how the tumor surrounds or encases the SVC. As the current case illustrates, one may not assume it is symmetrical and therefore subject to concentric distribution of force during balloon inflation. For this reason, the possibility of SVC rupture will be included in our informed consent process in the future. Although we continue to balloon-dilate prior to SVC stent placement, we now undersize the balloon. Using balloons of approximately 10 mm in diameter we hope will ensure adequate flow, avoiding thrombus formation while minimizing the possibility of vessel rupture. Should SVC rupture occur, it is possible that treatment with a covered stent might be life-saving, as reported by Burket [19], assuming that the tear is easily visualized, limited to the SVC, and does not extend significantly into the right atrium.

Copyright information

© Springer Science+Business Media, Inc. 2005