Evaluation of Immature Hemodialysis Arteriovenous Fistulas Based on 3-French Retrograde Micropuncture of Brachial Artery

Transcription

1 Vascular and Interventional Radiology Original Research Yan et al. Arterial Evaluation of Immature Hemodialysis Fistulas Vascular and Interventional Radiology Original Research Yan Yan 1 Michael C. Soulen Richard D. Shlansky-Goldberg S. William Stavropoulos Mandeep Dagli Maxim Itkin Scott O. Trerotola Yan Y, Soulen MC, Shlansky-Goldberg RD, et al. Keywords: angioplasty, arteriovenous fistula, brachial artery, fistula maturation, fistulography, hemodialysis fistula DOI: /AJR Received October 18, 2011; accepted after revision January 27, This project was supported in part by a National Institute of Diabetes and Digestive and Kidney Diseases short-term research grant (grant no. 5T35DK ). S. O. Trerotola receives royalties from Cook Incorporated. 1 All authors: Department of Radiology, Division of Interventional Radiology, University of Pennsylvania Medical Center, 3400 Spruce St, 1 Silverstein, Philadelphia, PA Address correspondence to S. O. Trerotola (streroto@uphs.upenn.edu). AJR 2012; 199: X/12/ American Roentgen Ray Society Evaluation of Immature Hemodialysis Arteriovenous Fistulas Based on 3-French Retrograde Micropuncture of Brachial Artery OBJECTIVE. The objective of our study was to assess outcomes after evaluation of immature hemodialysis arteriovenous fistulas (AVFs) via 3-French brachial artery access and to identify the incidence of arterial and venous puncture site spasm. MATERIALS AND METHODS. One hundred twenty-three outpatients (82 men, 41 women; mean age, 58 years; age range, years) with immature AVF were identified retrospectively in whom diagnostic fistulography was performed via 3-French retrograde brachial artery puncture. Percutaneous transluminal angioplasty was performed via a separate venous puncture during the same visit in 95 patients. Patient age and sex, fistula age and type, and technical success and complications were recorded. Images were reviewed for lesion location, potentially competing vessels, and arterial and venous puncture related spasms. RESULTS. The mean fistula age was 99 days (range, days). There were 49 AVFs in the left forearm; 30 in the left upper arm; 26 in the right forearm; and 18 in the right upper arm. Twenty-eight AVFs were transposed. Angioplasty was technically successful in 81 of 95 patients (85%; mean diameter, 7 mm; range, 4 10 mm). Brachial artery puncture caused no major complication. Arterial spasm occurred in 19 patients (15%) and was severe in one patient. There were two hematomas (1.6%). Venous spasm, ranging from mild (four patients) to occlusive (8 patients), occurred in 38 patients (40%) at the site of venipuncture for intervention. Nitroglycerin (mean, 325 mcg; range, mcg) was used in 26 procedures (21%). Venous spasm was more common with forearm (50%) than upper arm (24%) fistulas (p = 0.02) and with decreasing vein diameter (p = 0.02). CONCLUSION. Evaluation of immature AVFs based on 3-French micropuncture of the brachial artery can be safely performed on an outpatient basis. Spasm is more common in forearm AVFs and in smaller veins. A mong the various types of longterm vascular access for hemodialysis in patients with end-stage renal disease (ESRD), native arteriovenous fistulas (AVF) have been recommended because of their longer patency and fewer complications [1]. Whereas older studies in the literature have shown maturation failure in relatively few fistulas ( 25 30%) [2 4], with an emphasis on more fistulas in the United States by the Centers for Medicare & Medicaid Services Fistula First Initiative, maturation failure has become more common. For example, a large randomized trial investigating the effects of clopidogrel on newly created fistulas showed 60% maturation failure in both groups [5]. Accurate and early evaluation of an immature AVF can facilitate subsequent salvage interventions and can prolong fistula life [6 14]. For example, stenosis in an immature native AVF can be detected early using fistulography and can then be treated with transvenous angioplasty to maintain functional patency in most cases [6 12]. Historically, the preferred method for the diagnosis of an immature AVF is retrograde venous fistulography, but this technique has multiple disadvantages. First, the venipuncture may induce spasm that is difficult to distinguish from stenosis. Second, compression of the draining vein to encourage retrograde contrast delivery may divert blood flow through accessory veins that normally do not affect the fistula [15]. Last, venous-based fistulography occurs in a nonphysiologic state, where flow and steal cannot be accurately assessed [15, 16]. Although ultrasound and other noninvasive imaging techniques have been proposed as alternatives to fistulography, the latter procedure remains the standard of care [17 20]. AJR:199, September

2 Yan et al. At our institution, we routinely use the arterial approach to evaluate immature AVFs through a retrograde puncture of the brachial artery with a 3-French catheter. For this retrospective study, we examined the immediate complications associated with this diagnostic procedure and identified the incidence of spasm with venipuncture of immature AVFs. Materials and Methods Institutional review board approval and informed consent waiver were received for this HIPAA-compliant retrospective study. From our electronic quality assurance and procedure database (Hi-IQ, Conexsys), we identified all patients with ESRD and immature fistulas who had undergone diagnostic fistulography at our institution between May 2002 and May Although definitions of fistula immaturity have evolved over time [1, 5, 6, 10, 21], these patients all had in common a fistula that had not matured adequately to support dialysis, which is perhaps the most widely accepted definition of an immature fistula. The specific inclusion criteria established for this study required ESRD patients to have an immature AVF that had undergone diagnostic fistulography via a 3-French retrograde brachial artery access as the initial evaluation, and the fistulograms needed to be available for review. In patients whose AVF required multiple interventional radiology procedures during the study period, only the first procedure was considered in this retrospective study. This definition of AVF immaturity allowed a larger sample size with a wider range of fistula ages than the definition proposed by the National Kidney Foundation s Kidney Diseases Outcomes Quality Initiative [1] would and reflects our actual practice. During the study period, 265 patients with ESRD and an immature AVF underwent evaluation (Fig. 265 Patients with ESRD and immature fistulas underwent diagnostic fistulography between May 2002 and May 2010 Included 176 patients with reported retrograde brachial artery access using a 3-French catheter 1); 89 patients were excluded from further analysis because they did not meet the primary inclusion criterion of 3-French retrograde brachial artery access as the initial step in the evaluation. Seventy-five of the excluded patients had primary venous access, seven had antegrade artery access, one had radial artery access, two had arterial puncture of indeterminate direction, and four had brachial artery access established with a device other than a 3-French catheter. Of the remaining 176 patients, 53 were excluded because images were not available for review; the fistulographic images were essential to our assessment of the technical outcome and vasospasm associated with 3-French micropuncture brachial arterial diagnostic fistulography. Review of all images and procedural reports was performed for the remaining 123 patients who formed our cohort. The study cohort was composed of 82 men (67%) and 41 women (33%). Their mean age was 58 years (range, years). Diagnostic fistulography using digital subtraction angiography was performed from the arterial inflow to the right atrium via retrograde puncture of the brachial artery at or just above the elbow with a 21-gauge needle followed by placement of a 3-French (1- mm) catheter over a inch mandril wire (inner dilator from Micropuncture Introducer Set, Cook). Real-time ultrasound guidance was used in 63% of the procedures (78/123), whereas palpation alone was used in the remaining procedures depending on operator preference and equipment availability. The exact number of attempts required to achieve access was not recorded in the medical record and was thus unavailable for this retrospective analysis. If diagnostic fistulography revealed stenosis or occlusion, percutaneous transluminal angioplasty (PTA) was then attempted, as described elsewhere [6 12], via a separate venipuncture during the same visit in 95 patients 75 Primary venous access 7 Antegrade artery access Excluded 89 patients 1 Radial artery access or was scheduled for a subsequent visit because of residual renal function (with desire to limit contrast dose) or operator preference. After removal of the 3-French catheter at the end of the procedure, direct pressure was applied to the puncture site or sites until hemostasis was achieved. Patients were discharged after recovery from moderate sedation or after resting 1 hour if not sedated. Data collected included fistula age (days from surgical creation to interventional radiology study), fistula type (anatomic location: right vs left, upper arm vs forearm), completion of two-stage venous transposition, fistula lesion location, largest balloon diameter used in any attempted angioplasty and technical success (defined as presence of a thrill throughout the fistula), potentially competing vessels, and complications from arterial and venous punctures. Table 1 aggregates these data for each fistula location. The largest angioplasty balloon diameter was taken as an indirect measure of fistula vein diameter. The 123 procedures were performed either by 16 interventional radiology attending physicians, all of whom have a certificate of added qualification or are eligible for that certificate, or by fellows and residents under their direct supervision. The attending physicians had between 1 year and more than 20 years of experience in hemodialysis access interventions. One attending physician with more than 20 years experience in hemodialysis access interventions retrospectively reviewed all images. If the written report was incomplete or inconsistent with the intraoperative images, the imaging findings superseded the written report. Particular attention was paid to vasospasm caused by arterial and venous punctures. In every case in which the arterial or venous puncture site was visible, the spasm was subjectively graded as none, mild but not flow-limiting, moderate but not flow-limiting, severe and flow-limiting, or completely occluded. 2 Brachial artery access of unclear direction 4 Brachial artery access established with other device Excluded 53 patients without images 123 Patients in final analysis Fig. 1 Flowchart shows patient selection. ESRD = end-stage renal disease. 684 AJR:199, September 2012

3 TABLE 1: Aggregated Data by Fistula Location Arterial Evaluation of Immature Hemodialysis Fistulas Location of Fistula Data Left Forearm Left Upper Arm Right Forearm Right Upper Arm Total no. of patients (n = 123) Transposed vein (n = 28) Basilic vein (n = 22) Cephalic vein (n = 6) Procedures done with ultrasound guidance (n = 78) Patients with no arterial spasm (n = 104) Patients with no arterial spasm (n = 88) (see text) Patients with unrecorded arterial puncture site counted as no arterial spasm (n = 16) (see text) Patients with arterial spasm from puncture (n = 19) Patients with mild arterial spasm (n = 14) Patients with moderate arterial spasm (n = 4) Patients with severe, flow-limiting arterial spasm (n = 1) Hematomas (n = 2) Fistula age available (n = 101) Mean fistula age (d) Range of fistula ages (d) Vein-based angioplasty (n = 95) Mean of largest balloon diameter (mm) Mode of largest balloon diameter (mm) Range of largest balloon diameter (mm) Venous spasm from puncture (n = 38) Patients with mild venous spasm (n = 4) Patients with moderate venous spasm (n = 16) Patients with severe, flow-limiting venous spasm (n = 10) Patients with completely occlusive venous spasm (n = 8) Patients with thrill after angioplasty (n = 81) Procedures used nitroglycerin (n = 26) Blood vessel with lesion None Basilic vein Cephalic vein Median cubital vein Brachial vein Radial artery Brachial artery Multiple vessels (any combination of the vessels above) Other lesions Lesion type Long segment lesion over entire vessel Inflow lesion Midfistula lesion (Table 1 continues on next page) AJR:199, September

4 Yan et al. TABLE 1: Aggregated Data by Fistula Location (continued) Lesion type (continued) Data Location of Fistula Left Forearm Left Upper Arm Right Forearm Right Upper Arm Outflow lesion Anastomotic lesion Perianastomotic arterial lesion Perianastomotic venous lesion Swing point lesion Terminal arch lesion Diffuse arterial disease Central venous occlusion Occluded fistula Possible competing vessels E A C Fig. 2 Complete diagnostic fistulography of immature right forearm fistula performed via transbrachial 3-French puncture. A, Image from anastomosis and forearm shows cephalic vein stenosis (arrow) just downstream from anastomosis. B, Image from region of elbow shows puncture site with moderate arterial spasm (arrow). Note accessory cephalic vein is not considered competing vessel that needs treatment. C, Image from upper arm shows patent upper arm cephalic, basilic, and brachial veins. D, Image from shoulder region shows mild extrinsic compression of cephalic arch related to arm position. E, Central venous image shows tunneled dialysis catheter in place and mild stenosis (arrow) of right brachiocephalic vein. Note diagnostic image quality in spite of narrow-gauge, arterial-based injection. B D 686 AJR:199, September 2012

5 Arterial Evaluation of Immature Hemodialysis Fistulas The technical success of angioplasty is reported using the guidelines established by the Society of Interventional Radiology (SIR), but clearly fistula maturation is the final endpoint. The clinical outcome of the procedure (i.e., fistula maturation) was not examined in this study; the purpose of the study was to determine the incidence of complications from diagnostic 3-French retrograde arterial access of AVFs and the incidence of spasm from subsequent therapeutic venous access. To the extent that they apply, SIR definitions and reporting standards have been used throughout [22, 23]. Statistical analysis considered all procedures in which a venipuncture site was visible on available images. For some procedures with an imaged venipuncture site, however, the corresponding procedural reports did not completely document details about all attempted therapeutic venous interventions, venous transposition, or patient information. Findings from the images superseded the report, so individual statistical tests used different subsets of procedures depending on the availability of imagebased information. The Fisher exact test was used to assess the association between the occurrence of spasm at the venipuncture site (yes or no) individually with each of the following: sex, race, fistula age, fistula type (upper arm or forearm), venous transposition (yes or no), and largest angioplasty balloon diameter (in millimeters). The Fisher exact test uses a hypergeometric distribution to compute the statistical association between any two categoric variables without assuming any particular form for the underlying sample distribution [24]. This method is preferred when any of the contingency table cell counts is less than 5 [25]. Fig. 3 Venous-based intervention for treatment of venous stenosis causing failure of fistula to mature (same patient and procedure as shown in Fig. 2). A, After placement of 6-French sheath in cephalic vein and percutaneous transluminal angioplasty (PTA), there is no residual stenosis (arrow); there is occlusive spasm at puncture site (arrowhead) that did not respond to nitroglycerine. B, After removal of sheath, spasm (arrow) persists. C, Second venous puncture is made and PTA of area of spasm was performed with resolution. Note new mild spasm near elbow (arrow). Note also complete resolution of brachial artery spasm (arrowhead) when compared with Figure 2B. The Cochran-Mantel-Haenszel test, which is an extension of the chi-square test, was then used to assess the statistical association between venous puncture site spasm and fistula type after stratifying patients by venous transposition [24]. Similarly, the Cochran-Mantel-Haenszel test was used to assess the statistical association between venous puncture site spasm and venous transposition after stratifying patients by fistula type. Results Figures 2 and 3 show a typical example of fine catheter-based fistulography and ensuing PTA of stenosis, including arterial and venous spasms, the latter requiring PTA for resolution. As shown in Table 1, there were 49 left forearm, 30 left upper arm, 26 right forearm, and 18 right upper arm AVFs. In 28 AVFs, the vein was transposed, six with the cephalic vein and 22 with the basilic vein; in one patient, only the first stage of a two-stage basilic vein transposition had been completed at the time of the interventional radiology study. Potentially competing vessels were identified in seven patients; none of these required immediate treatment because we customarily treat stenoses first and if the AVF still fails to mature, we reassess for possible competing veins. PTA was indicated and attempted via a separate venipuncture immediately after arterial diagnostic fistulography in 77% of procedures (95/123), of which 85% (81/95; mean maximum balloon diameter, 7 mm; range, 4 10 mm) were technically successful (thrill present throughout the fistula) [22]. Partial in situ fistula thrombosis from attempted PTA was documented in one case in which no residual thrombus remained after A in situ thrombectomy. The clinical outcome of the fistulas after this intervention is beyond the scope of this study. Nitroglycerin (mean, 325 mcg; range, mcg) was used in 21% of all procedures (26/123) either to prevent spasm or to treat spasm that had occurred; medical records did not distinguish between prophylactic and therapeutic use of nitroglycerin and did not consistently document the route of nitroglycerin administration. Of the 26 patients in whom nitroglycerin was used, seven had documented partial to complete resolution of spasm, most of which were venous, after nitroglycerin administration; 13 records stated only that nitroglycerin was administered; and six records documented vasospasm that did not respond to nitroglycerin. Heparin (mean, 3250 U; range, U) was used in 34% of all procedures (42/123). Operative reports for fistula creation were available for 101 patients, of whom 91 had never had an AVF or graft. Of the 10 patients with a prior long-term dialysis access, five had a failed AVF or graft in the contralateral arm, two had a failed AVF on a different site of the ipsilateral arm, two had a failed AVF on the same site of the ipsilateral arm, and one patient had failed accesses bilaterally. Fistula age was available for 101 patients, whose mean time from fistula creation to diagnostic fistulography was 99 days (range, days). The large range of fistula ages reflects the diversity of our patient population as a major referral center and also reflects real practice in our metropolitan area in the wake of Fistula First recommendations. B C AJR:199, September

6 Yan et al. Retrograde puncture of the brachial artery caused no major complication per SIR definitions [23]. Spasm at the arterial puncture site occurred in 15% of procedures (19/123), of which 14 were mild but not flow-limiting, four were moderate but not flow-limiting, and one was severe and flow-limiting. The single severe and flow-limiting arterial spasm occurred in a 48-year-old woman who had completed only the first stage of two-stage basilic vein transposition. This spasm was not treated because arteriograms showed otherwise normal arterial anatomy with no evidence of an arteriovenous anastomosis. Two weeks after this patient s interventional radiology study, she underwent a right lower extremity arteriovenous loop graft construction without complication using a 4- to 7-mm polytetrafluoroethylene graft connecting the superficial femoral artery and femoral vein. Since then, she was lost to follow-up. There were no records of arm ischemia. A small hematoma requiring no additional management occurred at the brachial artery puncture site in 1.6% of the patients (2/123). One patient s intraoperative images showed a brachial artery with a high bifurcation, in which both bifurcated vessels were occluded, a collateral vessel was present near the elbow, and there was complete occlusion of this brachial artery approximately 10 cm below the catheter puncture site. This occlusion did not respond to nitroglycerin even after waiting for 10 minutes. She did not return for a scheduled follow-up Doppler evaluation of the arterial occlusion. Her medical record suggested that this brachial artery was chronically occluded from long-standing arterial disease possibly related to a prior synthetic upper arm A Fig. 4 Venous puncture induced spasm. A, Transarterial fine-catheter fistulogram shows cephalic vein inflow stenosis (arrow). Note remainder of cephalic vein is free of stenosis. B, After placement of 6-French transvenous sheath and percutaneous transluminal angioplasty (PTA) of stenosis, no significant residual stenosis remains. Note, however, venipuncture-induced spasm downstream from sheath (arrow) could be misinterpreted as stenosis if diagnostic study had been performed via venous puncture (see text). graft and not resulting from the arterial puncture performed for diagnostic fistulography on her forearm AVF. This patient was counted as having no arterial spasm. Spasm occurred at the venipuncture site in 40% of the procedures (38/95) (Fig. 4). In four, venous spasm was mild but not flowlimiting; in 16, moderate but not flow-limiting; in 10, severe and flow-limiting; and in eight, completely occlusive. The Fisher exact test showed no significant association between the occurrence of spasm at the venipuncture site and patient sex (p = 0.86) or fistula age (p = 0.34). There was a statistically significant association between the occurrence of spasm at the venipuncture site and the largest angioplasty balloon used (p = 0.02). The largest angioplasty balloon diameter was taken as an indirect measure of fistula vein diameter. Our data suggest that larger veins were associated with fewer venipuncture site spasms. Specifically, the odds of spasm occurring decreased 51% (95% CI, 68.2% to 25.1%) with each 1-millimeter increase in balloon diameter. Initially, the Fisher exact test revealed a statistically significant, unadjusted association between the occurrence of spasm at the venipuncture site and fistula location (p = 0.02) and venous transposition (p = 0.03). Specifically, 50% of forearm fistulas (29/58) had spasm, whereas only 24% (9/37) of upper arm fistulas had spasm. Similarly, 46% of the fistulas (35/76) with an in situ (i.e., not transposed) vein had spasm, and only 17% of the fistulas (3/18) with a two-stage transposed vein had spasm. On closer examination, however, transposed vein status and fistula location appeared highly confounded (Fisher exact test, p < ): Only 3.5% of the forearm fistulas (2/58) had a transposed vein, whereas 44% of the upper arm fistulas (16/36) had a transposed vein. After adjusting for fistula location, the Cochran-Mantel-Haenszel test showed no statistically significant association between venous transposition and occurrence of spasm (p = 0.18). In summary, fistula location and venous transposition were too highly confounded in our sample so there was not enough statistical power to detect any true association between the occurrence of spasm at the venipuncture site and transposition independent of fistula location. Discussion AVFs are the preferred form of long-term vascular access for hemodialysis in ESRD patients [1]. Early identification of fistulas that are failing to mature and early treatment of underlying causes result in a functional fistula in most patients [6 14], although the long-term viability of such fistulas has been questioned recently [26, 27]. In two prospective studies, endovascular treatment salvaged approximately 83 87% of immature AVFs [10, 12]. A fundamental underpinning of immature AVF salvage is a complete evaluation of the fistula to identify stenoses and other potential causes of failure to mature. Because immature veins may be at risk for spasm with venipuncture (a hypothesis confirmed in this study) and because of the need to evaluate for collaterals that may point to an otherwise occult venous stenosis and for possible competing veins, an ideal means of evaluating such fistulas is brachial artery puncture. Unlike venous-based techniques, a primary arterial approach for the diagnostic evaluation of immature AVFs allows study of the untouched B 688 AJR:199, September 2012

7 Arterial Evaluation of Immature Hemodialysis Fistulas fistula and outflow veins before therapeutic decision making. When Staple [28] first described retrograde (mature) venous fistulography in 1973, he judged that brachial artery puncture carried too much risk of occlusion or aneurysm. In 1994, Goldin and Dacie [29] proposed retrograde puncture of the brachial artery with a 20-gauge needle followed by a 3-French catheter for evaluating a mature AVF, as did Bücker et al. in 1995 [16] and Lui et al. in 2001 [15]. In spite of these reports, there has not been widespread adoption of this approach for mature fistulas, although some have advocated its use [30]. Indeed, the arguments discussed earlier in favor of a transbrachial approach in immature AVFs generally do not apply to mature fistulas. Many, if not most, interventionalists including our group avoid arterial access for hemodialysis access interventions in both native AVFs and grafts whenever possible because frequent transarterial interventions may cause complications and injuries at the puncture site, including pseudoaneurysms, hematomas, and iatrogenic stenosis [15, 30, 31]. Moreover, arterial interventions at many institutions necessitate overnight observation, which negates the outpatient advantage of percutaneous interventions [31]. In contrast to a transarterial approach for both diagnosis and intervention that is, by definition, antegrade in nature, diagnostic-only brachial artery procedures may have fewer complications [15, 16, 31]. For example, in a retrospective study on the safety of diagnostic arteriography for hemodialysis shunts via palpation-guided retrograde puncture of the brachial artery at the elbow with a 22-gauge cannula, Bücker et al. [16] detected a single pseudoaneurysm and no other major complication in 217 procedures. The forms of access studied included both grafts and fistulas, and although not explicitly stated by Bücker et al., all of the fistulas were most likely mature because the concept of maturation assistance had not yet been developed at the time of the Bücker study. In another retrospective study of the safety and efficacy of angiographic evaluation of mature AVFs via ultrasoundguided antegrade puncture of the brachial artery at the elbow with a 20-gauge needle, Lui et al. [15] detected no arterial spasm or occlusion in 208 procedures. We use retrograde puncture of the brachial artery to minimize the risk of arterial dissection and because any intervention to treat the fistula is done via a second venous puncture, not transbrachially for reasons stated earlier [31]. We believe the combination of fine catheter evaluation of the immature fistula followed by transvenous intervention where appropriate reaps the benefits of an arterial approach without the downsides of a purely transarterial approach. We have shown that this approach even with a 3-French (equivalent to a 19-gauge needle) catheter allows complete diagnostic evaluation of the fistula in its untouched state, confirming the observations of Lui et al. [15] who used 20-gauge needles (2.7-French) in mature fistulas. We observed only minor complications from arterial puncture, and this approach did not prolong postprocedural observation time. Our particular attention to vasospasm during image review and our operators 63% usage of ultrasound (thus observing clinically occult hematomas) as well as strict use of SIR definitions may account for our relatively higher reported incidences of arterial spasm (15%) and hematoma (1.6%) than those reported by Bücker et al. [16] (2.3% and 0.2%, respectively) and Lui et al. [15] (0% and 0.9%, respectively). Investigators have shown that 2% lidocaine injection alone before cannulation of the radial artery can cause temporary vasospasm [28]. A key outcome variable in this study was the incidence of venous spasm with puncture of outflow veins in immature AVFs. A detailed literature search using PubMed and search terms including spasm, venipuncture, hemodialysis, dialysis, fistula, and vein failed to identify any descriptions of this phenomenon. We have shown that such spasm is extremely common, occurring in nearly half of all fistula outflow veins punctured. This key observation has important bearing on interpretation of fistulograms obtained without arterial puncture in immature fistulas. The venipunctures (with resulting spasm) in our patients were made for attempted therapeutic interventions, but they were comparable in all but device size to venipunctures that would have been made for purely diagnostic transvenous fistulography. The incidence of venous spasm we observed, which was after placement of a 6-French sheath for the intervention, might not reflect the incidence that would occur with a smaller device, such as a 4-French micropuncture set or 18-gauge IV catheter (our standard for venous-based diagnostic fistulography of mature fistulas and grafts); however, we are unaware of any evidence linking the size of a puncture device to venospasm. It is unlikely that spasm is associated with any long-term detriment to a fistula unless the spasm is so severe that it causes immediate thrombosis of the fistula; we did not observe thrombosis in our series, although spasm did prompt heparin administration relatively often. However, the observed spasm would be difficult to distinguish from true stenosis without comparison with a prior arterial-based fistulogram showing the undisturbed physiologic venous anatomy (as shown in Fig. 4). Misidentification of a spasm as a stenosis might prompt not only unnecessary PTA but also another venipuncture because the site of spasm, being very near the puncture site, would generally be inaccessible to PTA from the initial venous access. Unnecessary PTA of an otherwise normal venous segment could have long-term sequelae to a fistula. Further, spasm changes the flow dynamics of the fistula and may make it difficult or impossible to determine if other veins seen are collaterals or competing vessels, although admittedly the latter are controversial. Our medical records did not specify prophylactic versus therapeutic administration of nitroglycerin; thus, although it is possible that nitroglycerin was used to prevent spasm in some patients but not in others, it is not our practice to administer nitroglycerine before arterial puncture, making this possibility unlikely. A final consideration is that by viewing a fistula in its entirety before venipuncture for intervention, the number of venous punctures needed to complete the intervention might be reduced. Of course, this potential reduction is offset by the addition of the initial arterial puncture. Whether reducing the number of venous punctures (by virtue of spasm or subsequent intimal hyperplasia at the puncture site) would have any long-term benefit requires further study, ideally in a prospective randomized trial. There are several limitations to our study, most of which relate to its retrospective nature. Because of long-term archiving issues, not all studies were available for review, although we have no reason to believe that the dataset we studied was not representative of our entire population. We were able to review archived images only, not entire angiographic runs. The ages of the fistulas were quite variable, including some that had been in place for more than 1 year but that had neither been used nor matured. However, all met the definition of a fistula that had never matured to allow hemodialysis. We have AJR:199, September

8 Yan et al. not sought to show that long-term outcomes, such as maturation rate or fistula survival, are different with our diagnostic approach than with other approaches; that would require a prospective randomized trial. The mean age of the fistulas in our study cohort was 99 days, which is longer than the recommended 6-week time point for immature fistula evaluation by the Vascular Access Guidelines [1]. The reasons for the broad range are multifactorial and beyond the scope of this study; all patients met the broadly accepted definition of an immature fistula, which is one that had not matured to the point where it could be used for hemodialysis [5, 26]. This definition of immaturity allowed a larger sample size, which increased our confidence in the conclusion on the safety of retrograde brachial artery fistulography. Another limitation is that spasm was graded subjectively, albeit consistently, by a single experienced attending physician. It can be argued that spasm might be distinguishable from fixed stenosis by virtue of the spasm s proximity to the puncture site; however, stenosis discovered in mature access is also often at or near puncture sites, in our experience. On another note, details of patients vascular comorbidities at the time of their fistulography were not available, thereby precluding correlation with complications from the interventional radiology arterial puncture. Finally, our single-institution experience may not be generalizable to the broader community. In conclusion, 3-French micropuncture catheter retrograde brachial artery based evaluation of immature AVF can be safely performed on an outpatient basis. In a large subset of patients, this technique provided information about the undisturbed venous anatomy of the fistula, whereas a purely venous-based approach might misinterpret spasm as stenosis. Acknowledgments We thank Jesse Chittams for assistance with statistical analysis and Jeffrey Mondschein for assistance in data collection. References 1. Vascular Access 2006 Work Group. Clinical practice guidelines for vascular access. Am J Kidney Dis 2006; 48(suppl 1):S176 S247 sion. Ann Surg 1985; 202: Miller PE, Tolwani A, Luscy CP, et al. Predictors of adequacy of arteriovenous fistulas in hemodialysis patients. Kidney Int 1999; 56: Rodriguez JA, Armadans L, Ferrer E, et al. The function of permanent vascular access. Nephrol Dial Transplant 2000; 15: Dember LM, Beck GJ, Allon M, et al. Effect of clopidogrel on early failure of arteriovenous fistulas for hemodialysis: a randomized controlled trial. JAMA 2008; 299: Beathard GA, Settle SM, Shields MW. Salvage of the nonfunctioning arteriovenous fistula. Am J Kidney Dis 1999; 33: Turmel-Rodrigues L, Mouton A, Birmele B, et al. Salvage of immature forearm fistulas for haemodialysis by interventional radiology. Nephrol Dial Transplant 2001; 16: Beathard GA, Arnold P, Jackson J, Litchfield T. Aggressive treatment of early fistula failure. Kidney Int 2003; 64: Shin SW, Do YS, Choo SW, Lieu WC, Choo IW. Salvage of immature arteriovenous fistulas with percutaneous transluminal angioplasty. Cardiovasc Intervent Radiol 2005; 28: Nassar GM, Nguyen B, Rhee E, Achkar K. Endovascular treatment of the failing to mature arteriovenous fistula. Clin J Am Soc Nephrol 2006; 1: Clark TW, Cohen RA, Kwak A, et al. Salvage of nonmaturing native fistulas by using angioplasty. Radiology 2007; 242: Manninen HI, Kaukanen E, Makinen K, Karhapaa P. Endovascular salvage of nonmaturing autogenous hemodialysis fistulas: comparison with endovascular therapy of failing mature fistulas. J Vasc Interv Radiol 2008; 19: Faiyaz R, Abreo K, Zaman F, Pervez A, Zibari G, Work J. Salvage of poorly developed arteriovenous fistulae with percutaneous ligation of accessory veins. Am J Kidney Dis 2002; 39: Asif A, Roy-Chaudhury P, Beathard GA. Early arteriovenous fistula failure: a logical proposal for when and how to intervene. Clin J Am Soc Nephrol 2006; 1: Lui KW, Yeow KM, Wan YL, Cheung YC, Ng KK, Tseng JH. Ultrasound guided puncture of the brachial artery for haemodialysis fistula angiography. Nephrol Dial Transplant 2001; 16: Bücker A, Vorwerk D, Günther RW. Transbrachial fine-needle arteriography with special focus on haemodialysis shunt imaging. Nephrol Dial Transplant 1995; 10: venous fistulas. J Vasc Access 2008; 9: Froger CL, Duijm LE, Liem YS, et al. Stenosis detection with MR angiography and digital subtraction angiography in dysfunctional hemodialysis access fistulas and grafts. Radiology 2005; 234: Grogan J, Castilla M, Lozanski L, Griffin A, Loth F, Bassiouny H. Frequency of critical stenosis in primary arteriovenous fistulae before hemodialysis access: should duplex ultrasound surveillance be the standard of care? J Vasc Surg 2005; 41: Doelman C, Duijm LE, Liem YS, et al. Stenosis detection in failing hemodialysis access fistulas and grafts: comparison of color Doppler ultrasonography, contrast-enhanced magnetic resonance angiography, and digital subtraction angiography. J Vasc Surg 2005; 42: [No authors listed]. III: NKF-K/DOQI clinical practice guidelines for vascular access update Am J Kidney Dis 2001; 37(suppl 1):S137 S Gray RJ, Sacks D, Martin LG, Trerotola SO. Reporting standards for percutaneous interventions in dialysis access. J Vasc Interv Radiol 2003; 14:S433 S Aruny JE, Lewis CA, Cardella JF, et al.; Society of Interventional Radiology Technology Assessment Committee. Quality improvement guidelines for percutaneous management of the thrombosed or dysfunctional dialysis access. J Vasc Interv Radiol 2003; 14:S247 S Goetz M. Think outside the box: analysis of categorical data. Arlington, VA: Quintiles, Stokes ME, Davis CS, Koch GG. Categorical data analysis using the SAS system. Cary, NC: SAS Institute, Lee T, Ullah A, Allon M, et al. Decreased cumulative access survival in arteriovenous fistulas requiring interventions to promote maturation. Clin J Am Soc Nephrol 2011; 6: Dember LM. Fistulas first: but can they last? Clin J Am Soc Nephrol 2011; 6: Staple TW. Retrograde venography of subcutaneous arteriovenous fistulas created surgically for hemodialysis. Radiology 1973; 106: Goldin J, Dacie JE. Technical report: use of a 3 French catheter for transbrachial evaluation of surgically-created arteriovenous fistulae in dialysis patients. Clin Radiol 1994; 49: Manninen HI, Kaukanen ET, Ikaheimo R, et al. Brachial arterial access: endovascular treatment of failing Brescia-Cimino hemodialysis fistulas initial success and long-term results. Radiology 2001; 218: Palder SB, Kirkman RL, Whittemore AD, Hakim 17. Rooijens PP, Serafino GP, Vroegindeweij D, et al. 31. Trerotola SO, Turmel-Rodrigues LA. Off the RM, Lazarus JM, Tilney NL. Vascular access for Multi-slice computed tomographic angiography for beaten path: transbrachial approach for native fis- hemodialysis: patency rates and results of revi- stenosis detection in forearm hemodialysis arterio- tula interventions. Radiology 2001; 218: AJR:199, September 2012