Author + information
- Received August 3, 2006
- Revision received September 29, 2006
- Accepted October 23, 2006
- Published online March 6, 2007.
- Aphrodite Tzifa, MD, MRCPCH⁎,†,
- Audrey C. Marshall, MD⁎,
- Doff B. McElhinney, MD⁎,
- James E. Lock, MD, FACC⁎ and
- Robert L. Geggel, MD, FACC⁎,⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. Robert L. Geggel, Children’s Hospital, Department of Cardiology, 300 Longwood Avenue, Boston, Massachusetts 02115.
Objectives The purpose of this research was to investigate the causes and symptoms of superior vena cava (SVC) obstruction or occlusion and report on the long-term results of transcatheter therapy.
Background Information on transcatheter therapy for SVC obstruction is limited.
Methods Superior vena cava catheterization interventions between August 1984 and April 2006 were reviewed. Patients were divided into 2 subgroups depending on whether or not they had previously undergone congenital cardiac surgery.
Results Sixty-three patients with median age of 3.7 years (range 1 month to 42 years) and weight of 13.3 kg (range 3 to 114 kg) were treated. Fifty patients (79%) were symptomatic, although only 50% had symptoms suggestive of SVC obstruction. Superior vena cava syndrome was more common in the non-cardiac surgical group (52% vs. 10%, p = 0.001). The mean gradient and SVC diameter improved from 10.8 ± 5.8 mm Hg to 2.6 ± 2.2 mm Hg (p < 0.001) and 3.1 ± 2.7 mm to 9.1 ± 3.8 mm, respectively (p < 0.001). The obstruction was adequately relieved in all 36 patients receiving stents and in 21 of 27 patients (78%) who had balloon dilation alone. Complications occurred in 12 patients (19%), all of whom had previously undergone cardiac surgery; 10 of these patients were successfully treated in the catheterization laboratory. Freedom from re-intervention did not differ between patients undergoing balloon dilation or stent implantation, but was longer in patients age >5 years at the time of intervention.
Conclusions Superior vena cava-related symptoms occur in only 50% of patients with hemodynamically significant SVC obstruction. Endovascular therapy is successful in relieving the stenosis and associated symptoms with good long-term results.
The prevalence of superior vena cava (SVC) obstruction in pediatric and young adult patients is largely unknown. Causal factors have only been identified in small series or case reports (1–5). Long segment SVC obstruction has been related to surgical injury, thrombosis, infection, chronic indwelling catheters or devices, and external compression. Recently, the potential for transcatheter recanalization of obstructed or occluded venous structures has revived awareness of, and interest in, SVC obstruction at our institution. Concurrently, pediatric practices such as wider application of extracorporeal membrane oxygenation (ECMO), more frequent use of central venous catheters in small patients, increased implantation of transvenous pacemakers and defibrillators (6), and a resurgence of the Mustard and Senning procedures (in the context of double-switch operations) may be increasing the numbers of patients at risk for SVC obstruction. The frequency with which SVC obstruction is now diagnosed at this institution has allowed us to investigate, for the first time, the causes and clinical sequelae of such obstruction, and the results of transcatheter therapy. We reviewed our entire 22-year experience of transcatheter interventions for obstructed SVC segments to try to determine and describe: 1) the most frequent underlying causes of SVC obstruction; 2) the clinical signs and symptoms associated with it; and 3) the short- and longer-term results of transcatheter therapy.
Medical records, cardiac catheterization data, and angiograms were reviewed for all patients with biventricular hearts who underwent transcatheter treatment of SVC obstruction at Boston Children’s Hospital between August 1984 and April 2006, in accordance with a protocol approved by the Children’s Hospital Committee on Clinical Investigation. Patients with palliated univentricular congenital heart disease and SVC obstruction in the setting of a cavopulmonary anastomosis or Fontan connection were excluded.
Catheterization and SVC intervention
The technique of balloon angioplasty and stenting of venous channels has been described previously (7,8). Femoral venous access was used in the majority of patients. After angiography and hemodynamic measurements, the diameters of the SVC at its narrowest point and above the stenosis were recorded, along with the length of the obstruction.
In patients with complete SVC occlusion, a jugular or subclavian venous line was often placed to guide SVC recanalization using the stiff end of a wire or a transseptal needle. In other cases, angiography through peripheral intravenous catheters in the upper extremities localized the course of the veins upstream from a total occlusion.
Although technical details varied, the general interventional approach is summarized below. The SVC obstruction was initially crossed and dilated with a balloon measuring approximately 250% to 300% of the stenotic diameter, but not exceeding the diameter of the adjacent unobstructed SVC. If a waist formed on the balloon at low pressure (1 to 2 atm), further inflation was carried out until the waist resolved or until the rated burst pressure of the balloon was reached. Full inflation of the balloon with no apparent waist was interpreted as vessel wall stretching without induction of a therapeutic medial tear (7). In such situations, a larger balloon was then used. When the stenosis was severe (total occlusion or diameter <2 mm), initial dilating balloons were typically 4 mm. In some instances, an externalized veno-venous wire rail (femoral-internal jugular or femoral-subclavian) was created to provide stability during stenting. Palmaz long or XL stents (Johnson & Johnson, Piscataway, New Jersey) were used until June 2002, when they were replaced by Palmaz Genesis (Cordis, Johnson & Johnson) or pre-mounted stents (Genesis pre-mounted, Cordis, Johnson & Johnson). Before 2002 and in larger patients since 2002, stents were hand-crimped onto the appropriate size balloon and usually introduced via a long sheath. Relief of obstruction was considered successful if the vessel diameter increased by at least 50%, or the gradient across the lesion decreased by at least 50%.
All patients who underwent SVC dilation with stenting were treated with intravenous heparin until therapy with aspirin (with or without clopidogrel) had been initiated. Short- and long-term antiplatelet/anticoagulant therapy for patients after simple balloon dilation without stenting was decided on an individual basis.
Outcomes assessed included acute and chronic relief of SVC obstruction and related symptoms, survival, freedom from SVC re-intervention, and adverse events. Independent variables analyzed for association with outcomes included age, year of surgery, balloon dilation versus stenting, prior cardiac surgery, complete SVC occlusion, type/category of SVC obstruction, and presence and type of symptoms. For acute outcomes, variables of interest included acute efficacy and adverse events. Time-dependent outcomes (survival, freedom from re-intervention) were assessed with the Kaplan-Meier product limit method. Factors associated with time-dependent outcomes were analyzed using Cox proportional hazards regression. Factors significant by univariable Cox regression analysis were entered into a multivariable model using forward stepwise entry. For comparison of continuous variables between groups, independent samples t test or the Wilcoxon rank sum test were used, and for comparison of categorical variables Fisher exact test or chi-square analysis was performed. Values are expressed as mean ± SD or median (range).
Sixty-three patients underwent transcatheter treatment for SVC obstruction. The median age at intervention was 3.7 years (1 month to 42 years), and the median weight was 13.3 kg (range 3 to 114 kg). Twenty-two (35%) patients were infants, 23 (36%) were 1 to 17 years of age, and 18 (29%) were ≥18 years of age.
Underlying causes/types of SVC obstruction are summarized in Table 1. Despite the extremely varied causes seen, the patients were divided into 2 general groups for purposes of data analysis and discussion: 1) postcardiac surgical group (n = 41; Mustard/Senning procedures, partial anomalous pulmonary venous return repair, cardiac transplantation, arterial switch operation); and 2) noncardiac surgical group (n = 22; postchronic line instrumentation or ECMO and 1 patient with extrinsic compression by a neurofibroma). Patients in the cardiac surgical group were older than those in the non-cardiac surgical group (median 8.1 years, range 1 month to 42 years vs. median 10 months, range 1 month to 39 years, p = 0.001).
Although some type of presenting symptom was recorded at the time of catheterization in 79% of patients (50 of 63), only one-half of patients (n = 32) had symptoms suggestive of SVC obstruction (Table 2). Of particular interest are 4 patients who presented with frequent chest infections. An additional 7 patients presented with primarily respiratory symptoms in the absence of pleural effusions. Patients in the non-cardiac surgical group were more likely to present with typical SVC syndrome (52% vs. 10%, p = 0.001) as evidenced by swelling affecting the upper body, upper extremities, neck, and face. Almost all ECMO patients (11 of 12, 92%) were symptomatic, predominantly with SVC-related symptoms (SVC syndrome [n = 5] and pleural effusions [n = 4]), while 2 patients presented with pulmonary hypertension.
Thirteen asymptomatic patients underwent SVC intervention for placement of transvenous pacing leads (n = 2) or for improvement of hemodynamically significant obstruction diagnosed incidentally at the time of catheterization (n = 4) or follow-up echocardiography (n = 7).
Catheterization and interventions
Sixty patients with successful transcatheter procedures had markedly elevated baseline SVC pressures, with a mean gradient of 10.8 mm Hg from the SVC to the right atrium. Three patients had failed endovascular interventions and required subsequent surgery, due to catheter-related complications in 2 (SVC tear with hemothorax or stent embolization) and failure to improve the hemodynamic condition in 1 unstable postoperative Senning patient.
The SVC was completely occluded and required recanalization in 19 of 63 patients (30%) (Fig. 1). Total SVC occlusion was more common in the nonsurgical group (50% vs. 20%, p = 0.03). Of 12 ECMO patients, 75% were found to have total SVC occlusion. After recanalization of totally occluded vessels, 16 patients (84%) had stent placement, and in all but 3 patients, the recanalized segment was predilated before stenting.
Twenty-seven patients underwent balloon dilation alone. The maximum balloon-to-vessel diameter ratio was 3.8 ± 2.1, and the largest effective balloon diameter (11.4 ± 4.9 mm) was similar to the SVC diameter above the narrowing (11.2 ± 4.1 mm). A total of 49 stents were placed in 36 patients, 10 of whom were infants. In 26 patients, the SVC was predilated before stent deployment.
Acute results of intervention are summarized in Table 3. Adequate relief of obstruction was observed in 21 of the 27 patients (78%) who underwent balloon dilation alone and in all 36 patients who underwent stent implantation. In 30 of the 32 patients with SVC-related symptoms, symptomatic relief was observed after the procedure before hospital discharge.
All 11 patients who underwent intervention for the purpose of placement of transvenous pacing were successfully treated, 8 with stent implantation and 3 with balloon angioplasty alone. Stent implantation was feasible in 3 patients who had existing pacemaker leads traversing the narrowed or occluded SVC segment. Two of these patients had a new transvenous pacing lead inserted through the stent soon after SVC recanalization (Fig. 2), and in 1 patient, the existing pacemaker lead remained intact and functioned normally.
Complications (SVC tear, n = 6; posterior right atrial perforation, n = 2; and stent malposition or embolization, n = 4) occurred in 12 patients (12 of 63, 19%), 10 of whom were managed in the cardiac catheterization laboratory. All 12 of the acute complications occurred in patients who had undergone prior cardiac surgery including 8 of 24 patients who had previously undergone a Mustard or Senning operation. One patient with an unconfined SVC tear and hemothorax and a second patient with an embolized stent required surgery. Patients who underwent balloon dilation alone had higher incidence of SVC tears or posterior right atrial perforation (n = 6) than patients undergoing stent implantation (n = 2). Adverse events in patients who underwent recanalization for complete SVC occlusion included SVC tears (n = 2) and stent displacement during implantation (n = 2). Episodes of pulmonary or systemic thromboembolism were not encountered.
Follow-up information was available in 62 patients (99%) at a median of 2 years (1 week to 17 years). There were 12 deaths, all from causes unrelated to catheterization, at a median of 1.5 months after SVC intervention (1 day to 2 years). Factors associated with worse survival by univariable Cox regression analysis included age <5 years at the time of SVC intervention (p = 0.01) and no prior history of congenital cardiac surgery (noncardiac surgical group; p = 0.008), which were strongly associated with one another (p = 0.002). Survival among noncardiac surgical patients was significantly worse than among those in the cardiac surgery group (76% at 1 month and 59% at 1 and 5 years vs. 98% at 1 month, 95% at 1 year, and 91% at 5 and 10 years, p = 0.007). However, by multivariable Cox regression, only age <5 years (survival 80% at 1 month, 68% at 1 year, and 63% at 5 years compared with 96% at 1, 5, and 10 years among patients >5 years of age) remained a significant predictor of worse survival. Year of surgery was not associated with survival or with re-intervention-free survival.
During follow-up, 18 patients underwent surgery (n = 2) or catheterization (n = 16) resulting in a total of 27 SVC re-interventions between 1 day and 16 years after the initial procedure. Survival free from re-intervention was 90% at 1 month, 79% at 1 year, 53% at 5 years, and 35% at 10 years (Fig. 3). Freedom from re-intervention did not differ between patients who had undergone stent implantation or balloon dilation alone at the initial procedure. By univariable Cox analysis, patients >5 years of age at the time of intervention (p < 0.001) and those in the cardiac surgical group (p = 0.003) had significantly longer freedom from re-intervention. By multivariable Cox analysis, only age >5 years at the time of intervention was independently associated with longer freedom from SVC re-intervention (p < 0.001) (Fig. 4).
A total of 25 transcatheter re-interventions was performed in 16 patients. Six of 27 patients (22%) who underwent balloon dilation at the initial intervention had re-intervention, including stent implantation in 5. Ten of 35 stented patients (27%) underwent re-intervention, 6 of whom had neo-intimal proliferation within the stent (Fig. 5), and 4 of whom had balloon dilation of the existing stent to match somatic growth. Hemodynamic and anatomic changes at re-intervention are summarized in Table 3. Complications at re-intervention occurred in 2 patients (stent embolization with surgical stent retrieval and SVC tear with no hemodynamic consequences).
Although the literature regarding transcatheter therapy for SVC obstruction is limited to small case reports, the feasibility and acute efficacy of both balloon dilation and stent placement in this setting are generally accepted (7,8,10–16). As a result, we are now able to examine a series of 63 patients who presented to the catheterization laboratory for treatment.
In this study, we reviewed our 22-year experience of patients who underwent endovascular therapy for SVC occlusion or obstruction. There appeared to be 2 general groupings of causative factors. Two-thirds of our patients had undergone prior cardiac surgery. Superior vena cava obstruction in these patients is credibly related to the reconstructed anatomy, the presence of surgical anastomoses, or even caval cannulation sites for cardiopulmonary bypass. Of the remaining patients, 12 had been treated with ECMO. Prior reports estimate an incidence of SVC obstruction after ECMO to be 10%, and it has been previously recognized that less than 50% of these patients present with SVC syndrome (17,18). The design of this study precludes an estimate of the incidence of SVC obstruction at our institution though, based on this prior information and on our own findings, we believe that SVC obstruction may continue to be under-diagnosed. Almost all of the patients in our series who had previously been on ECMO were symptomatic at the time of presentation to the catheterization laboratory. Because SVC obstruction was frequently asymptomatic in the overall group of patients, and some patients presented with symptoms other than “classic” SVC syndrome, we speculate that there is an occult incidence of SVC obstruction in the post-ECMO population.
Our experience shows, in the largest series to date, that transcatheter therapy, whether with balloon dilation or stenting, is effective for relief of SVC obstruction and recanalization of SVC occlusion. Among our patients, mean SVC pressure decreased from 18 to 12 mm Hg after intervention, and SVC anatomy was improved. Follow-up evaluations showed a survival with freedom from re-intervention to be 79% at 1 year. In contrast with some prior reports (19,20), balloon dilation resulted in lasting effect in some patients, with a restenosis rate of 22%. The finding that stent implantation was associated with a restenosis rate of 27% is consistent with the findings of other investigators (21–23).
Procedural complications occurred only in patients who had undergone prior cardiac surgery, mostly in patients who had previously undergone Mustard or Senning operation. The reported incidence of systemic venous baffle obstruction after atrial switch operation ranges from 0% to 20% (24–26), and endovascular therapy has been reported either in isolated case reports or in small case series (2,3,5,10,11,14,16). In our cohort, patients with Mustard/Senning procedures experienced relief of SVC obstruction after transcatheter therapy, but were at higher risk for procedural complications than other patient groups. The higher complication rate related to dilation of the SVC-right atrial junction in these patients may be due to the highly variable compliance of this region, which can include atrial myocardial wall, suture material, pericardium, and/or prosthetic material.
Patients in the noncardiac surgical group had no procedural complications, despite having a higher incidence of complete occlusion and the need for recanalization, often with transseptal needles, and being substantially younger and smaller. Complete SVC occlusion was not associated with higher complication risk. This apparently low procedural risk in the noncardiac surgical group, coupled with the acute efficacy of the procedure, supports a recommendation to pursue diagnosis and treatment of SVC obstruction in these patients.
Superior vena cava occlusion or obstruction can be successfully relieved by balloon dilation or stent implantation in patients with a variety of underlying conditions. Technical success and efficacy in relieving associated symptoms and facilitating transvenous pacing are high. Symptomatic relief achieved at the time of intervention persists at long-term follow-up. The procedure was associated with no mortality and approximately 20% morbidity. Most of the complications can be dealt with in the cardiac catheterization laboratory. Patients who have undergone prior cardiac surgery are at higher risk for complications, while the incidence of SVC tears is higher in the balloon dilation group. Rate of re-intervention was higher in patients <5 years of age at the time of the initial intervention (9).
- Abbreviations and Acronyms
- extracorporeal membrane oxygenation
- superior vena cava
- Received August 3, 2006.
- Revision received September 29, 2006.
- Accepted October 23, 2006.
- American College of Cardiology Foundation
- Lock J.E.,
- Bass J.L.,
- Casteneda-Zuniga W.,
- Fuhrman B.P.,
- Rashkind W.J.,
- Lucas R.V. Jr.
- O’Laughlin M.P.,
- Perry S.B.,
- Lock J.E.,
- Mullins C.E.
- Bu’Lock F.A.,
- Tometzki A.J.P.,
- Kitchiner D.J.,
- Arnold R.,
- Peart I.,
- Walsh K.P.
- Mullins C.E.,
- O’Laughlin M.P.,
- Vick G.W.
- Ward C.J.,
- Mullins C.E.,
- Nihill M.R.,
- Grifka R.G.,
- Vick G.W. 3rd.
- McMahon C.J.,
- El-Said H.G.,
- Grifka R.G.,
- Fraley J.K.,
- Nihill M.R.,
- Mullins C.E.
- Ing F.F.,
- Grifka R.G.,
- Nihill M.R.,
- Mullins C.E.
- Cheung Y.F.,
- Sanatani S.,
- Leung M.P.,
- Human D.G.,
- Chau A.K.,
- Culham J.A.