| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Vascular and Interventional Radiology |
1 From the Department of Radiology, Division of Interventional Radiology (R.D.L., A.K., R.D.S., T.W.I.C., A.A.P., S.W.S., J.I.M., J.A.S., C.M.T., S.O.T.), and Department of Medicine, Division of Nephrology (R.M.C.), University of Pennsylvania Medical Center, 1 Silverstein, 3400 Spruce St, Philadelphia, PA 19104. Received January 24, 2005; revision requested March 30; revision received July 13; accepted August 11; final version accepted September 1. Supported in part by a grant from MedComp, Harleysville, Pa. Address correspondence to S.O.T. (e-mail: streroto{at}uphs.upenn.edu).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Materials and Methods: The institutional review board granted exemption for this retrospective study, the need for informed consent was waived, and all data collection was in compliance with HIPAA. Patients with hemodialysis access requiring maintenance procedures between 1998 and 2004 and incidentally found to have ipsilateral (
50%) CVS were identified from a departmental database. Thirty-five patients (19 men, 16 women; mean age, 58.7 years) with 38 grafts met inclusion criteria, and 86 venograms were reviewed. CVS was measured by using venograms obtained before and after PTA, if performed. Patients with arm swelling, multiple CVS, indwelling catheters, and stents at the first encounter were excluded. CVS progression was calculated by dividing the change in the degree of stenosis by the time between venographic examinations. Wilcoxon rank sum test was used to evaluate differences in rate of CVS progression between treated and nontreated patients.
Results: Mean degree of CVS before intervention was 71% (range, 50%–100%). Sixty-two percent (53 of 86) of lesions had associated collateral vessels; 28% (24 of 86) of CVSs were not treated. Mean degree of stenosis in this group was 72% (range, 30%–100%); mean progression was –0.08 percentage point per day. No untreated CVS progressed to symptoms, stent placement, or additional CVS. Seventy-two percent (62 of 86) of CVSs were treated with PTA. Mean degree of stenosis in this group was 74% (range, 50%–100%) before and 40% (range, 0%–75%) after treatment; mean progression was 0.21 percentage point per day after treatment (P = .03). Six (8%) of 62 treatments were followed by CVS escalation; one patient developed arm swelling, four required stents, and four developed additional CVS.
Conclusion: PTA of asymptomatic CVS greater than 50% in the setting of hemodialysis access maintenance procedures was associated with more rapid stenosis progression and escalation of lesions, compared with a nontreatment approach.
© RSNA, 2006
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Central venous stenosis (CVS) in the subclavian or brachiocephalic veins ipsilateral to a functioning hemodialysis access or the superior vena cava can also cause graft malfunction, as well as arm, breast, and/or face swelling by limiting the flow of blood leaving the accessed limb (6). PTA of a CVS, supplemented by stent placement as needed, is effective and considered the primary treatment for such lesions owing to the lack of viable and safe surgical options (4,5).
High-grade CVS may exist in asymptomatic patients who do not exhibit the pathognomonic arm, face, or breast swelling (7–9). These lesions are typically found incidentally during declotting procedures concurrent with flow-limiting peripheral venous stenoses (10,11).
While many earlier reports of PTA and/or stent placement for CVS generally described treatment of symptomatic lesions (ie, arm swelling) (12,13), authors of one large series (14) that included CVS stated that lesion significance was "empirically determined" on the basis of the intraprocedure estimate of the degree of stenosis and the extent of collateral vessels. More recent publications and national guidelines have described treatment of all lesions greater than 50% in the absence of arm, face, or breast swelling as long as there is a "clinical indicator of failure," low flow, high pressures, abnormal physical examination of the access, et cetera (4,5,15). In one series, 40% of central venous stents were placed in patients whose clinical indicator was thrombosis, apparently without prior arm swelling or other symptoms (16). However, venous stenoses peripheral to the CVS can cause these "clinical indicators of failure," and the benefits of PTA for asymptomatic (nonedematous) CVS have not been investigated. Thus, the purpose of our study was to retrospectively evaluate the natural history of high-grade (>50%) asymptomatic CVS in hemodialysis patients and the outcome of serial treatment of CVS with PTA.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
Names of patients in whom ipsilateral CVS (subclavian or brachiocephalic) 50% or greater was identified during hemodialysis access maintenance procedures were selected from a quality assurance database by a single reviewer (R.D.L.). Patients with symptoms referable to a central stenosis, such as breast, arm, or facial swelling at initial presentation, were excluded. Other exclusions were made to limit variability, including preexisting stents, extrinsic venous compression (due to difficulty in measurement and tracking intrinsic narrowing) (17), multiple stenoses (to limit variables), noniodinated contrast material use (because of difficulty in accurately measuring lesions with carbon dioxide as the contrast agent), or indwelling catheters at the time of initial venography. Demographic information was collected (R.D.L.), including patient age and sex and access type (graft or fistula) and location (upper arm or forearm).
By using these criteria, 38 grafts in 35 patients (19 men, 16 women; age range, 26.2–82.5; mean, 58.7 years) with asymptomatic CVS were identified. Three patients had two separate dialysis access grafts, one in each arm, during the study period. Eighty-six venograms were obtained in these patients between 1998 and 2004, and these formed the basis of this study. Indications for undergoing fistulography included elevated graft pressure during hemodialysis (10 of 86), inadequate hemodialysis (two of 86), inadequate flow (18 of 86), graft pulsatility (one of 86), thrombosis (45 of 86), and other (10 of 86). Access configurations included an upper arm graft (n = 32), forearm graft (n = 6), upper arm fistula (n = 0), and forearm fistula (n = 0). While patients with either fistulas or grafts were eligible for inclusion, no patients with a fistula happened to meet inclusion criteria. Age of the access was not determined.
Procedures
Central venography was performed by manually injecting 10–20 mL of iodinated contrast material into the venous limb of the graft, with digital subtraction. Images were obtained in a posteroanterior projection at a frame rate of one to three images per second at operator discretion. According to our division's imaging protocol, at least one image with ipsilateral central veins and superior vena cava fully opacified or several images, which are obtained in aggregate, that show all ipsilateral central veins and the superior vena cava are archived.
Treatment of stenoses was performed at the discretion of the operating interventional radiology attending physician. The number of procedures performed by each physician is as follows: A.K., 16; R.D.S., 10; T.W.I.C., nine; A.A.P., five; S.W.S., five; J.I.M, five; J.A.S., four; C.M.T., four; S.O.T., seven; and nonauthor physicians, 21 procedures. Their experience in treating hemodialysis malfunction with PTA varied from 1 to 15 years at the time of the procedure. Pressure gradients were not consistently measured. Patients were divided into two groups: those who received treatment of CVS and those who did not receive treatment. No patients had arm, face, or breast swelling at the time of the procedure. Angioplasty balloons ranging from 8 to 16 mm (mean, 12.6 mm) were used in the treated group. When angioplasty was performed, the treatment goal was less than 30% residual stenosis, per Kidney Disease Outcomes Quality Initiative guidelines (4,5).
Image Evaluation
Venograms were reviewed by a single observer (S.O.T.) with 15 years experience in interpreting such images, independent of data storage and analysis. Lesion location was identified and degree of stenosis was calculated before and after PTA (if performed) by comparing minimal lumen diameter with the diameter of adjacent normal vessels. Other features of central vasculature, including presence of collateral veins and thrombus embolization to the level of the stenosis ("logjamming"), were noted (S.O.T.). All available follow-up venograms and procedure reports were obtained and reviewed (R.D.L., S.O.T.) in the order they became available from the film library. Stenosis evolution, development of collateral vessels, additional stenosis formation, and stent placement were documented. Symptoms of CVS (breast, arm, or face swelling) that developed after initial venography were noted from review of medical records (R.D.L.); no patients had symptoms of CVS at the time of the initial venography. Presence or absence of symptoms was determined while the ipsilateral access was functioning, since central venous lesions may be occult in the absence of ipsilateral functioning grafts or fistulas.
Statistical Analysis
After measurement of the degree of stenosis, the rate of stenosis progression was calculated by dividing the degree of stenosis by the number of elapsed days between venography, if follow-up was performed. Because three patients with bilateral grafts had them in opposite arms, each graft was treated as a separate "patient"; thus, all analyses are based on 38 grafts. In cases of CVS treated with PTA, the degree of stenosis after treatment was used as the baseline for subsequent calculations of stenosis progression.
The nonparametric Wilcoxon rank sum test was used on the rate of stenosis progression data to assess the association between the treatment status and the progression of measured venous stenosis. We chose the Wilcoxon rank sum test because there was one observation that caused the distribution of measured venous stenosis to become skewed. This test does not require a symmetric distribution and is not affected by a few outliers. We also used the Fisher exact test to assess the significance of the association between sex and treatment status. This test is recommended over the Pearson 
2 test when any of the expected cell counts is below zero. A P value less than .05 was considered to indicate a significant difference. The SAS statistical software program (version 9.1; SAS, Cary, NC) was used for this analysis.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
|
Treated Group
Seventy-two percent (62 of 86) of CVSs were treated with PTA. The mean degree of stenosis was 74% (range, 50%–100%) prior to treatment and 40% (range, 0%–75%) after treatment. The mean (standard deviation) rate of stenosis progression was 0.21 (0.48) percentage point per day after treatment, meaning an increase in stenosis had occurred over time. Twelve percent (three of 26) of treated CVSs regressed over time at a mean of –0.48 (0.54) percentage point per day. None remained stable (0 of 26), and 88% (23 of 26) progressed at a mean of 0.50 (1.03) percentage point per day. In 58% (36 of 62) of CVSs, no follow-up venogram was obtained or indicated.
Escalation of CVS
Six (8%) of 62 PTA treatments were followed by escalation of CVS; one patient developed arm swelling, four required stent placement, and four developed additional CVS. No untreated CVS progressed to symptoms, necessity of stent placement, or development of additional CVS during the study period.
Results of the nonparametric Wilcoxon rank sum test suggested that those treated with PTA had a significantly greater rate of stenosis progression than did those not treated with PTA (P = .03). There was no evidence of a sex effect on the rate of stenosis progression (P = .78).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
It is also not clear if an isolated CVS always compromises flow in hemodialysis access circuits. In one study in which the static pressure technique was used to identify significant CVS, these lesions alone were not flow limiting (10). In another study of five patients with elevated ipsilateral hemodialysis graft dynamic pressures without arm swelling, subclavian stenosis alone caused significant elevations in graft pressure that normalized after PTA (9).
In an attempt to standardize care, the Kidney Disease Outcomes Quality Initiative guidelines advocated treatment of all lesions greater than 50%, provided a clinical indicator of failure is present and regardless of concurrent peripheral lesions (4,5). These guideline recommendations were based on a study in which treatment of peripheral stenoses greater than 50% in the absence of clinical indicators did not improve outcome over that in a control group with similar lesions that were not treated (20). However, no such study exists for CVS, which leaves the treating physician to decide whether to treat an asymptomatic CVS greater than 50%. As a result, empiric treatment of central venous lesions without symptoms is commonly practiced, as evidence by recent publications (15,16).
Every patient in this study had a clinical indicator of failure, but all also had peripheral lesions that alone could account for the graft malfunction. The CVS discovered during venography may or may not have contributed. Differences in the treatment threshold regarding asymptomatic CVS among the interventional radiology physicians allowed this retrospective comparison. Data from this cohort suggest that treatment of asymptomatic stenosis may not be indicated; a randomized prospective study is needed to corroborate these findings. In addition to pressure gradients and collateral vessels, other criteria have been suggested to warrant treatment of asymptomatic CVS, including (a) recurrent access failure in the absence of another more peripheral lesion (this was not the case in any of our patients) and (b) logjamming of a clot during thrombectomy procedures (which occurred in only one case). Some interventional radiology physicians prophylactically treat all central lesions prior to thrombectomy procedures to avoid this latter problem, but there is no evidence to support this practice.
In this asymptomatic cohort, 62% (53 of 86) of lesions had collateral vessels. Development of these vessels is a natural response to venous drainage occlusion and, if sufficient to decompress the ipsilateral arm, might actually be an indication not to treat the underlying stenosis. Indeed, one of our most surprising findings was that in 9% (eight of 86) of venograms, complete central venous occlusion was present without arm, breast, or face swelling, which is a strong testament to the ability of collateral vessels to decompress even the abnormally high flow delivered by a graft or fistula.
The appropriate use of prophylactic PTA can reduce thrombosis rates and possibly prolong access life (1–3); but injudicious use of the same technique may accelerate stenosis formation and access failure. This is not surprising as the mechanism of venous PTA is endothelial disruption and intimal stretching. Damage to these sensitive vessel layers can trigger immune reactions, myointimal proliferation, and fibromuscular hyperplasia, processes that together may ultimately accelerate stenosis formation (21). In our study, not only did CVS progress at a greater rate in treated individuals, but PTA also may have triggered adverse events such as new stenosis formation, stent requirement, and progression to symptomatic arm swelling. Along with being detrimental to the long-term patency of the central veins, the treatment of asymptomatic CVS with PTA had a low technical success rate in this cohort. A mean of 40% residual stenosis was left after treatment with PTA despite aggressive use of large high-pressure balloons. While this degree of residual stenosis following PTA is higher than that in other studies in which PTA and/or stent placement for CVS has been described, restenosis is always the rule when this treatment modality is implemented (9,12,13,22).
The mutability of the venous system allows the adaptation of the central veins to hemodialysis-induced high-flow and stenotic lumen diameter reduction. Authors of several studies (6,23) found that up to half of significant subclavian thromboses and stenoses occurring in the context of central venous catheter placement have self-resolved during a 3–4-month period without intervention. Asymptomatic high-grade CVS is often found incidentally during declotting and preoperative venographic procedures; collateral circulation can often effectively shunt blood around these stenoses. Pressure measurements obtained during declotting procedures suggest that distal anastomotic lesions contribute more to pressure gradients than to CVS (10). Findings of these studies imply that high-grade CVS rarely compromises hemodialysis access function and that treatment with PTA may not be a necessary intervention unless symptoms are present, particularly since PTA of CVS may be technically less successful than procedures in smaller, more peripheral vessels (14).
Like any retrospective study, this one had important limitations that warrant consideration. It was possible to obtain complete follow-up in only 71% (25 of 35) of patients (66% of grafts [25 of 38]) since only malfunctioning hemodialysis grafts or symptomatic CVS, which occurred in only one patient, were reimaged. It could be argued that this makes the study biased toward detection of those patients in whom PTA resulted in progression, while those with the best long-term outcomes would not have been detected because they did not require venographic follow-up or hemodialysis graft maintenance procedures. The indication for venography was malfunctioning hemodialysis grafts with peripheral stenosis in all but the one patient who developed symptomatic CVS. Since treatment of peripheral vein stenosis, as well as central venography performed during hemodialysis maintenance procedure, was universal in this cohort, detection of nonsymptomatic CVS should be free of selection bias and dependent only on the functioning of the peripheral vasculature. Second, the relative number of follow-up studies was the same in both groups.
Another important limitation was the difficulty in measuring venous stenosis, especially in the central veins, because of difficulty in determining what constitutes an "adjacent normal vessel". However, any error introduced as a result of this limitation and any limitation resulting from use of a single observer should be the same in both groups. The standard deviations are acceptably low, and the resulting statistical significance reflects this. Clearly, the optimal way to corroborate our findings is a prospective randomized trial. Given the relatively small numbers of patients with CVS now that internal jugular catheterization has replaced subclavian catheterization for temporary dialysis access, such a trial would likely need to encompass several centers. The results of the present study could be used to help design such a trial.
In conclusion, PTA of asymptomatic CVS greater than 50% in the setting of hemodialysis access maintenance procedures was associated with more rapid stenosis progression and escalation of lesions, compared with a nontreatment approach. We believe that, like peripheral lesions, CVS in hemodialysis patients should only be treated in the presence of a clinical indicator such as arm, breast, or facial swelling.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Abbreviations: CVS = central venous stenosis • PTA = percutaneous transluminal angioplasty
See Materials and Methods for pertinent disclosures.
Author contributions: Guarantors of integrity of entire study, J.A.S., S.O.T.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; approval of final version of submitted manuscript, all authors; literature research, R.D.L., S.O.T.; clinical studies, R.D.L., R.M.C., A.K., R.D.S., A.A.P., S.W.S., J.I.M., J.A.S., C.M.T., S.O.T.; statistical analysis, R.D.L., T.W.I.C.; and manuscript editing, R.D.L., R.M.C., T.W.I.C., A.A.P., S.W.S., J.I.M., J.A.S., C.M.T., S.O.T.
|
References |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONReferences |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
