after treatment of Renal duplex sonography renovascular disease

Renal duplex sonography renovascular disease after treatment of Dudley A. Hudspeth, MD, Kimberley J. Hansen, MD, Scott W. Reavis, RVT, Susan M. Start, RN, Richard G. Appel, MD, and Richard H. Dean, MD, Winston-Salem, N. C. Purpose: To define the value of renal duplex sonography (RDS) to detect the presence of critical renal artery (RA) stenosis or occlusion after surgical repair or percutaneous transluminal balloon angioplasty (PTRA), we retrospectively reviewed our recent 71-month experience. Methods: From January 1987 through November 1992, 272 patients underwent 279 operative RA repairs and 35 patients underwent PTRA. Three hundred twenty-five RDS examinations were performed in 176 patients after operative intervention or PTRA during the study period. Forty-one of these patients had conventional angiography providing 61 RA for RDS comparison, and these data form the basis of this analysis. Twenty-four women and 17 men (mean age 57 years) underwent 44 operative RA repairs or 17 PTRA for correction of atherosclerotic disease (51 arteries) or fibromuscular dysplasia (10 arteries). Before their renovascular procedure each patient had significant hypertension (mean 193/106 mm Hg). RDS after surgery or PTRA was technically complete for all 61 RA. Results: Compared with angiography RDS correctly identified 47 of 48 repairs with less than 60% RA stenosis, 7 of 11 repairs with 60% to 99% stenosis, and 2 renal artery occlusions, providing a 69% sensitivity rate, 98% specificity rate, 90% positive predictive value, and a 92% negative predictive value. These results were adversely affected by branch RA disease, which accounted for three of four false-negative RDS study results. For 50 kidneys undergoing correction of main RA disease, RDS demonstrated an 89% sensitivity rate, 98% specificity rate, and 96% overall accuracy. RDS results were equivalent for both surgical and PTRA treatment. Conclusions: From this experience we conclude that RDS is useful for anatomic evaluation after surgical RA repair or PTRA. A negative RDS result excludes stenosis or occlusion of a main RA reconstruction but does not exclude significant branch level disease. (J VASC StTRG 1993;18:381-90.) Confirmation of the anatomic success of renal revascularization is assumed on the basis of a favorable blood pressure response in many centers. The fallacy of this assumption, however, has been documented repeatedly by previous reports. 1-3 Those recorded experiences have clearly documented that a favorable blood pressure response can occur after From the Division of Surgical Sciences and the Department of Medicine of the Bowman Gray School of Medicine, Winston- Salem. Presented at the Seventeenth Annual Meeting of the Southern Association for Vascular Surgery, Fort Lauderdale, Fla., Jan. 28-31, 1993. Reprint requests: Kimberley J. Hansen, MD, Department of General Surgery, Bowman Gray School of Medicine, Medical Center Blvd., Winston-Salem, NC 27157. Copyright 1993 by The Society for Vascular Surgery and International Society for Cardiovascular Surgery, North American Chapter. 0741-5214/93/$1.00 +.10 24/6/48841 graft thrombosis and renal infarction, the equivalent of nephrectomy. Alternatively, a favorable blood pressure response can occur when the intervention alters renal perfusion sufficiently to shut down the renin-angiotensin cause of hypertension yet leaves residual lesions that represent only partial success or even a complete technical failure (Fig. 1). For these reasons anatomic delineation with follow-up renal arteriography has remained the traditional "gold standard" for confirmation of the results after renal revascularization procedures. Unfortunately confirmation of the results of renal revascularization procedures by arteriographic survey is seldom undertaken in general clinical practice. Reasons frequently given for this lack of arteriographic study are the risk ofpotentiauy adverse effects on renal function, the risk of vessel wall injury, the financial cost of arteriography, and the associated 381

382 Hudspeth et al. September 1993 Fig. 1. Preoperative (left) and postoperative (right) arteriogram of42-year-old manwith severe hypertension caused by fibromuscular dysplasia of mid left RA. Although aortorenal graft thrombosed, patient was cured of hypertension as consequence of"patch effect" of graft's distal anastomosis, which can be seen along mid inferior aspect of grafted (right arterw'gram) RA. patient discomfort. Recognizing these objections to the liberal application of arteriography yet understanding the errors inherent in the use of clinical outcome as a gauge of technical success, we have continued to explore alternate methods for examination of the renal vasculature. Our previous experience 4,S and that of others 6"7 with renal duplex sonography (RDS) have demonstrated that the technique can be a valuable screening test to identify correctable forms of renovascular disease. Similarly this technique has proven to be useful in the intraoperative detection of technical defects, s On the basis of these experiences, we have explored the use of RDS as a method to define the technical results of interventional treatment of renovascular occlusive disease. This report summarizes our experience with postprocedure use of RDS and evaluates its accuracy in comparison to arteriography as a method for detection of technical defects and residual lesions after both operative intervention and percutaneous transluminal balloon angioplasty (PTRA). PATIENTS AND METHODS Patients. During the period from November 1, 1988 through November 30, 1992, 1722 RDS examinations were performed in the Clinical Vascular Laboratory of our medical center. Three hundred twenty-five of these studies have been performed on 176 patients during follow-up to evaluate the success of renal revascularization by either operation or PTRA. Forty-one of these patients also had conven- tional cut-film or digital subtraction arteriography for evaluation of 61 renal arteries (RA) that had been submitted to interventional management. To avoid bias all patients for whom arteriography and RDS results were available for review were included in the study regardless of clinical outcome, the results of arteriography or RDS, or the sequence or interval between studies (Table I). Study subjects include 24 women and 17 men who ranged in age from 26 to 77 years (mean age 57 years). The comparative data from the RDS and arteriography on these 61 RA form the basis of this report. Operative management had been used in 30 patients to manage 44 RA lesions, and PTRA had been used in 11 patients to correct 17 lesions. Fifty-one of the lesions had been due to atherosclerosis, and 10 lesions had been caused by fibromuscular dysplasia. The techniques used for renal revascularization in the 30 patients who underwent operation are summarized in Table II. Fifty-two procedures were performed for correction of main RA lesions, whereas nine were performed for reconstruction of the branch RA. Arteriography. The Seldinger technique of transfemoral aortography was used for all studies. Flush anterior-posterior cut-film aortography was performed in 38 patients, and digital subtraction arteriography was used in three patients. Arteriography was performed from 1 to 196 weeks (mean 30 weeks, median 2 weeks) after operation and within 0 to 170 weeks (mean 34.6 weeks, median 7.4 weeks)

Volume 18, Number 3 Hudspeth et al. 383 Table I. Comparative data from RDS and angiography Time interval Patient no. Treatment (days) RDS Result Angiography 1 PTRA 12 NEG NEG 2 PTRA 3 NEG NEG 224 POS POS 3 RATEA 19 NEG POS 4 RAB 52 NEG NEG 5 RAB 118 NEG NEG 6 RATEA 0 NEG NEG 7 PTRA 1 POS POS 8 RATEA 690 NEG NEG 9 RAB 70 NEG NEG 10 RAB 362 NEG NEG 11 RAB 940 NEG POS 12 RAB 1 NEG NEG 13 PTRA 7 NEG NEG 14 PTRA 160 NEG NEG 15 PTRA 180 POS POS 16 RAB 2 POS POS PTRA 15 NEG NEG 17 RAB 680 NEG NEG 18 PTRA 1 NEG NEG 19 RATEA 16 NEG POS 20 RAB 11 POS POS 21 RAB 6 NEG POS 22 RAB 345 NEG NEG 23 RATEA 330 NEG NEG 24 PTRA 60 NEG NEG 25 PTRA 45 NEG NEG 26 RAB 85 NEG NEG 27 RAB 80 NEG NEG 28 PTRA 48 NEG NEG 29 RATEA 0 POS POS 30 RAB 40 NEG NEG 31 RAB 360 NEG NEG 32 RATEA 90 NEG NEG 33 RAB 1193 NEG NEG 34 RAB 770 NEG NEG 35 RAB 26 NEG NEG 36 PTRA 74 NEG NEG 37 RAB 21 NEG NEG 38 RAB 707 NEG NEG 39 RAB 15 NEG NEG 40 RAB 10 NEG NEG 41 RAB 20 POS POS RAB, Renal artery bypass; RATTM, renal artery thromboendarterectomy. of the RDS study to which it was compared (Table I). PTRA arteriograms (14 completion studies and two follow-up arteriograms obtained 6 and 8 months later) were used for comparison with RDS to evaluate the 17 arteries managed by that technique. Subsequent RDS was performed from 0 to 26 weeks (mean 7 weeks, median 3.5 weeks) after PTRA and the accompanying completion arteriogram. Residual lesions were estimated from the aortogram in 5% diameter-reducing increments by two independent observers without knowledge of the results of the RDS. Angiographic stenosis was measured independently by two vascular surgeons. Table II. Method of RA reconstruction (n = 61) Aortorenal bypass graft Saphenous vein PTFE RA reimplantation RA TEA No TEA RA TEA Transrenal with patch PTFE Dacron Transaortic PTRA Totals PTFE, Polytetrafluoroethylene; TEA, thromboendarterectomy. 24 2 l0 26 14 17

384 Hudspeth et al. September 1993 Fig. 2. Tubular aortorenal saphenous vein graft stenosis 8 months after repair of inflammatory AAA and bilateral RA bypass. RDS spectra from right RA bypass demonstrate increased RA-PSV (A: RA-PSV > 2.0 m/sec) and poststenotic turbulence (B) with near normal distal RA spectra (C). Table III. RDS criteria Defect Criteria < 60% diameter-reducing RA stenosis RA-PSV from entire RA < 2.0 m/sec -> 60% diameter-reducing stenosis Focal RA-PSV _> 2.0 m/sec and distal turbulent velocity waveform Occlusion No Doppler-shifted signal obtained from an imaged RA Inadequate study for interpretation Failure to obtain Doppler samples from entire main RA Vessel lumen was approximated by visual examination of the contrast stream. When focal stenoses were present, they were compared with the maximal diameter of nondilated normal artery in continuity with the diseased vessel segment. Longer tubular stenoses were compared with the most normalappearing graft/main RA segment (Fig. 2). When the estimated degree of residual or recurrent stenosis differed by more than 15% between the two independent observers, the diameter reduction was estimated by a third independent observer. Observer results were then averaged for each arteriogram. Renal duplex sonography. RDS studies were performed with either an Ultramark-8 or an Ultramark-9 Ultrasound System (Advanced Technology Laboratories, Bothell, Wash.) with either a 3.0 MHz mechanical long-focus probe or a 2.25 MHz phased-array probe with Doppler color-flow capability. Patients fasted overnight and received 10 mg bisacodyl by mouth to minimize bowel gas interference. With the subject supine, the ultrasound probe was positioned in the abdominal midline 2 to 3 cm inferior to the xiphoid process. First a sagittal B-scan image was obtained of the upper abdominal aorta at the level of the superior mesenteric artery. The left renal vein was identified, and the probe was then rotated 90 degrees to obtain a B-scan image of the aorta and proximal superior mesenteric artery in cross section. With the left renal vein used as a visual reference, each RA origin was identified on B-scan or by Doppler signal. Doppler velocity wave forms were obtained from multiple sites in both RA (approximately 10 sites per artery) from aortic origin to renal hilum. During main RA Doppler interrogation, the angle of insonation was maintained at 60 degrees or less for calculation of peak systolic velocity (RA- PSV). Intrarenal Doppler-shifted signals were also

Volume 18, Number 3 Hudspeth et al. 385 Table IV. Arteriographic results for 61 kidneys "6 No. arteries/grafks Stenosis 36 0% ~, 12 < 60% 11 >60% to 99%,~ 2 Occluded ee ~, ",:,, i I i i 0 1 l I obtained from regions of the interlobar and arcuate arteries. This process was then repeated from a flank approach with the patient in the left and right decubitus positions; images and Doppler signals were obtained from each RA and kidney. From this flank approach, length, width, and thickness measurements of each kidney were recorded from the B-scan image. When measured Doppler spectral values differed between values obtained by supine or flank approaches, values obtained with the least angle of insonation were recorded. B-scan images and spectra from the FFT spectrum analyzer (Advanced Technologies Laboratories) were recorded on videotape and hard copy processor for independent review. On the basis of this examination and collection of both Doppler and B-mode ultrasound data during the RDS study, Table III summarizes the criteria used for interpretation of the study. Statistical analysis. Patient demographic data, RDS results, and angiographic interpretation were statistically analyzed. Descriptive statistics (means and standard deviations of continuous data; frequencies and relative frequencies of categoric data) were tabulated, and these data were evaluated to verify that assumptions of statistical tests were met. The relationship between RA-PSV and angiographic data was assessed by a scatterplot (Fig. 3). Confidence limits for comparative analysis estimates (such as sensitivity, specificity) were found by use of standard errors that were adjusted for within-subject correlations. 4 RESULTS The results of arteriography of the 61 RA evaluated in this study are summarized in Table IV. Thirty-six arteries and grafts were considered normal, 12 (20%) had less than a 60% diameter-reducing stenosis, 11 vessels (18%) had 60% to 99% diameterreducing lesions, and 2 reconstructions (3%) were totally occluded (Fig. 3). Among the 11 vessels with 60% to 99% stenoses, one lesion was located at the RA ostium or aortic-graft anastomosis, one stenosis was at the distal graft anastomotic level, five lesions 0 20 40 60 80 RA Angiographic Stenosis (%) Fig. 3. Scatter plot of RA-PSV in m/see versus percent RA stenosis as defined by arteriography. were located in the nonostial portion of the main RA, and four branch-level stenoses were present. Renal duplex sonography. There were no technically inadequate RDS examinations among patients included in this study. RDS identified 51 RA as normal or less than 60% RA stenosis. Eight RA were interpreted to have stenoses greater than or equal to 60%, and two total RA occlusions were identified. Among negative study results (51), mean RA-PSV was 1.24 _+ 0.35 m/see. Among positive study results with 60% to 99% stenosis (8), mean RA-PSV was 2.36 + 1.47 m/see. Mean RA enddiastolic ratio from negative RDS study results was 0.26 + 0.09 and 0.43 + 0.16 for positive study results. The results of RDS in 61 vessels examined are summarized and compared with the results of arteriography in Table V. Comparative analysis. The results of RDS correctly correlated with arteriography in 47 of 48 RA free of significant disease. In the one false-positive RDS result, the RA-PSV was 2.3 m/see. In this case the RDS preceded angiography by 15 days. Seven of the 11 RA with 60% to 99% lesions were correctly identified by RDS. Three of the four false-negative RDS study results occurred in patients with only branch level RA disease present at arteriography (Fig. 4). The fourth false-negative RDS study result occurred in a patient who had undergone a bilateral RA thromboendarterectomy who had a 60% diameter-reducing stenosis identified by arteriography on one side. RDS performed 11 months previously showed an RA-PSV of 1.3 m/see but no focal sites of turbulence and was interpreted as a normal study result. Finally Fig. 3 provides a scatterplot comparing RA-PSV derived from RDS to the percent stenosis as measured by arteriography in the entire group of 61 vessels. Table VI summarizes the comparative results of 100

386 Hudspeth et al. September 1993 Fig. 4. This 48-year-old woman had severe hypertension and bilateral RA fibromuscular dysplasia. (A) She had branch level disease noted in her left RA distribution (arrow) before bilateral saphenous vein aortorenal bypass. (B) Although she remains cured of hypertension, persistent left RA lower pole branch lesion (arrow) is apparent on follow-up arteriography. This lesion was not detected by RDS. Table V. Resuks of duplex scanning versus arteriography Arteriography > 60% to 99% RDS Normal < 60% stenosis stenosis Occluded Totals < 60% stenosis 35 12 4 0 51-60% to 99% stenosis 1 0 7 0 8 Occluded 0 0 0 2 2 Total 36 12 11 2 61 RDS and arteriography in this study. As shown in the table RDS provided an overall 69% sensitivity rate, 98% specificity rate, 90% positive predictive value, 92% negative predictive value, and a diagnostic accuracy of 92% when compared with arteriography. When one limits the comparative evaluation of the accuracy of RDS to patients who had main RA lesions repaired without branch disease (50 vessels) one sees that RDS had a 96% diagnostic accuracy when compared with arteriography. In contrast, RDS had only a 73% diagnostic accuracy when compared with arteriography in the evaluation of 11 cases of branch RA disease. Finally the RDS results agreed with completion and follow-up angiography after PTRA in all RA treated by this method. DISCUSSION This report describes our experience with the use of follow-up RDS to determine its value for the assessment of the success of renal revascularization procedures. Its 69% sensitivity rate, 98% specificity rate, and 92% diagnostic accuracy in comparison to arteriography argues that it can be a clinically useful tool to clarify the status of such procedures during the follow-up period. Nevertheless there are several aspects of its use and our analysis of its value that warrant comment. It is important to underscore that our analysis is based on an uncontrolled retrospective comparison of the results of RDS and arteriography. Neither study was acquired at a prospectively planned time after intervention. Although most of the RDS and arteriographic examinations in this study were performed within 2 months of each other, in some instances several months elapsed between the two studies. Five RDS examination results disagreed with arreriography results. The single false-positive RDS study result was followed by arteriography within 2

Volume 18, Number 3 Hudspeth et al. 387 Table VI. Comparative analysis of parameter estimates and their 95% confidence intervals 95% Confidence Group No. Measure Estimate interval All procedures 61 kidneys Sensitivity 0.69 (0.61, 0.77) Specificity 0.98 (0.96, 1.00) PPV 0.90 (0.85, 0.95) NPV 0.92 (0.88, 0.96) Accuracy 0.92 (0.88, 0.96) Kidneys with main RA dis- 50 kidneys Sensitivity 0.89 (0.82, 0.96) ease procedure Specificity 0.98 (0.95, 1.00) PPV 0.89 (0.82, 0.96) NPV 0.98 (0.95, 1.00) Accuracy 0.96 (0.91, 1.00) Kidneys with branch RA 11 kidneys Sensitivity 0.25 (0.15, 0.35) disease PPV, Positive predictive value; NPV, negative predictive value. ~Confidence interval cannot be calculated as standard error of mean is zero. Specificity 1.00 ~ PPV 1.00" NPV 0.7 (0.59, 0.81) Accuracy 0.73 (0.63, 0.83) weeks. That patient underwent right aortorenal saphenous vein bypass and contralateral nephrectomy and had no angiographic stenosis. Three of the four false-negative results represented branch level disease noted on follow-up angiography performed 1, 2, and 156 weeks after operation. RDS was performed 134 weeks and 10 weeks after arteriography and 3 weeks before arteriography, respectively. RDS results could not be related to disease progression in these cases. The fourth false-negative RDS study result was obtained 11.5 months before arteriography, which demonstrated a 60% stenosis at a distal RA endarterectomy site. In this case, a negative RDS result preceded arteriography by nearly i year and may have allowed disease progression. Although the length of time between examinations is a shortcoming of this retrospective report, such studies were included to avoid bias. The 41 patients available for analysis in this report represent a small and nonrandomly selected sample of our experience with both renal revascularization techniques and postoperative RDS. Included in the 41 patients are several techniques of renal revascularization, each procedure being represented by relatively few numbers. This is especially pertinent to the use of RDS in postoperative follow-up of complex branch RA reconstructions. Whereas RDS was associated with a 73% diagnostic accuracy in cases of branch level disease, the sensitivity rates and confidence limits of this observation are quite poor. Our overall reported experience with RDS suggests that it is not a reliable tool to assess the presence of branch disease, either before or during operation. 4,5 The analysis of the validity of postoperative RDS after branch repairs or in other cases of branch RA disease further strengthens those findings. For this reason we recommend that arteriography is still required to exclude postoperative residual or recurrent branch RA disease. The results of RDS in comparison with arteriography after PTRA are encouraging. There was complete agreement between RDS and arteriography in the group of 17 vessels that were interrogated after PTRA. Notably most of the RDS studies were performed many days after PTRA and the associated intimal disruption and self-limited dissection. The success of RDS for assessment of PTRA results is supported by the report ofmewissen et al. s In their report they found that postangioplasty duplex scanning of dilated femoropopliteal sites was even superior to arteriography in predicting outcome. Our results compare favorably with other pubfished results of RDS use for postintervention follow-up. In the report by Eidt et al. 9 they recorded an overall diagnostic accuracy of 86%, sensitivity rate of 80%, and a specificity rate of 87%. Technically satisfactory scans, however, were only obtainable in 84% of their studies. In this regard the use of RDS in the early postoperative period may be impossible. We have noted increased difficulty with RDS in the first week after operation because of the presence of increased bowel gas and tissue edema. Because the RDS studies used in this comparative analysis of our results were all performed after at least 13 days after

388 Hudspeth et al. September 1993 Clinical RDS Indications for Surveillance Following Renovascular Intervention Good clinical outcome following intervention to main renal artery RDS at 6 months and 12 months post procedure / \ / \ Renal Aogiography Annual RDS I Renal angiography Clinical hypertension Multiple renal arteries Branch level disease 1 Renal Angiography Worsening renal insufficiency with main RA repair I RDS \ \ Continued surveillance Fig. 5. Algorithm for clinical RDS use after surgical RA revascularization or PTRA. surgery, this early factor negating the value of RDS was not pertinent to our results. Mthough blurring of tissue planes from the retroperitoneal dissection may decrease the clarity of the B-scan image, this scarring and obliteration of retroperitoneal planes has not adversely affected the results of RDS when studies are performed at least 1 week after operation. In addition, preoperative screening and intraoperative RDS frequently precede postprocedure RDS at our institution and are performed by a few technicians who often perform all three examinations. That experience combined with a specific knowledge of the surgical reconstruction used or PTRA location have resulted in a high rate of technically complete studies. From the results of this study our recommendations regarding the use of RDS after intervention for treatment of renovascular disease appears in Fig. 5. In this algorithm emphasis is placed on serial RDS during the first year after surgery or PTRA. Surveil- lance studies are then recommended annually or more frequently if hypertension or renal function worsens. The requirement of angiography despite a negative RDS examination result is dependent on the clinical indication for study. If RDS after intervention is obtained in evaluation of hypertension in combination with renal insufficiency (i.e., global renal ischemia), angiography is not required after negative RDS resuks because main RA stenosis or occlusion is reliably excluded. However, angiography is recommended for evaluation of severe hypertension alone, particularly after branch RA repair, because disease at this level is not accurately evaluated by RDS. In summary, we conclude that RDS can be a useful tool for anatomic evaluation after surgical repair or PTRA of main RA occlusive disease. Nevertheless its value is heavily technician dependent, and its success is nontransferrable among vascular laboratories. Therefore one should be cautioned to determine the validity of RDS studies in their own laboratory and only rely on its use on the basis of the results of such an analysis. REFERENCES 1. Dean RH, Wilson JP, Burko H, Foster JH. Saphenous vein aortorenal bypass grafts: serial arteriographic study. Ann Surg i974;i80:469-78. 2. Foster JH, Dean RH, Pinkerton JA, Rhamy RK. Ten years experience with the surgical management of renovascular hypertension. Ann Surg I973;177:755-66. 3. Ernst CB, Stanley JC, Marshall FF, Fry WJ. Autogenous saphenous vein aortorenal grafts: a ten-year experience. Arch Surg 1972;105:855-64. 4. Hansen KJ, Tribble RW, Reavis SW, et al. Renal duplex sonography: evaluation of clinical utility. }" VAse SURe 1990;12:227-36. 5. Hansen KJ, O'Neil EA, Reavis SW, Craven TE, Plonk Jr, G.W., Dean R_H. Intraoperative duplex sonography during renal artery reconstruction. J VAse SURG 1991;14:364-74. 6. Kohler TR, Zierler RE, Martin RL, et al. Noninvasive diagnosis of renal artery stenosis by ultrasonic duplex scanning. I Vase SURG 1986;4:450-6. 7. Rittgers SE, Norris CS, Barnes RW. Detection of renal artery stenosis: experimental and clinical analysis of velocity waveforms. Ukrasound Med Biol 1985;11:523-31. 8. Mewissen MW, Kinney EV, Bandyk DF, et al. The role of duplex scanning versus angiography in predicting outcome after balloon angioplasty in the femoropopliteal artery. J Vasc SUI<G 1992;15:865-6. 9. Eidt IF, Fry RE, Clagett GP, Fisher Jr, D.F., Alway C, Fry WJ. Postoperative follow-up of renal artery reconstruction with duplex ultrasound. J VASC SURG 1988;8:667-73. Submitted Feb. 10, 1993; accepted May 19, 1993.

Volume 18, Number 3 Hudspeth et al. 389 DISCUSSION Dr, John F. Eidt (Little Rock, Ark.). Several points should be emphasized before widespread application of duplex scanning in the postoperative evaluation of RA reconstructions. First, this is a relatively small, retrospective study with the obvious limitations inherent in such a study. Second, there is considerable bias in the selection of patients for arteriography because not all patients underwent arteriography. Third, there was substantial variability in the time between operation and the performance of duplex scanning and between duplex scanning and corresponding arteriography, which injects an additional degree of uncertainty to the comparison between arteriography and duplex scanning. Fourth, it is obvious from your results and from others that multiple RA and branch vessel disease greatly diminishes the utility of duplex scanning. Finally I congratulate your noninvasive vascular technologists for achieving a very high rate of technically successful examination results. It is apparent that intraabdominal duplex scanning is highly technician dependent, and the transferability of the technique cannot be assured. I agree with your recommendation that each vascular laboratory should establish accuracy data for themselves. Have you determined the intraobserver and interobserver reliability of this technique in comparison to arteriography? Did you find that the renal end-diastolic blood flow improved in patients after revascularization or was the end-diastolic blood flow relatively fixed? In the absence of a highly competent vascular technician, would you recommend postoperative arteriography in the evaluation of symptom-free patients after RA reconstruction? Did you observe any technical defects on arteriography that did not meet your strict criteria of a greater than 60% stenosis, such as kinks or flaps? Dr. Dudley A. Hudspeth. We included all the patients through this retrospective experience of almost six years to try to avoid bias, and there was at times considerable difference in the time between when arteriography and duplex scanning were performed. It is true that multiple or polar RA are difficult to detect by this technique. We have shown here that branch RA are difficult to detect. In fact, for the overall series of more than 1700 examinations, the technically incomplete rate at our institution was 4% for RA, and among this selected group of 325 examinations in 176 patients, the technically incomplete rate was 2%. There is a limited number of technicians at our institution performing the examination, and they have often performed screening surface examinations on patients and then have gone to the operating room and assisted with the technical assessment of the grafts in the operating room with intraoperative studies; therefore they were more familiar with the RA anatomy of those patients. We have found the combination of B-mode ultra- sonography and duplex scanning to be very useful in demonstrating kinks or other graft abnormalities in the operating room. Again these examinations were directed at detecting occlusive disease, and the lesions that have been demonstrated on arteriography in this patient series were all for occlusive disease. We believe that this study offers an excellent noninvasive method for surveillance of RA grafts, especially during the first year after a procedure where, if there is a change in the patient's clinical status, blood pressure, or renal fianction, RDS may allow possible intervention in a deteriorating situation before the kidney becomes unsalvageable. We believe that it has a special applicability in that kind of setting; otherwise, it is basically equivalent to arteriography for detecting main RA lesions of that sort. Dr. Timothy R.S. Harward (Gainesville, Fla.). I was a little confused about the exact method used in your study. It sounded like you only performed intraoperative ultrasonography. Did you also perform postoperative arteriography before the patient was discharged from the hospital? At the University of Florida, we have taken a little different approach. At our institution, CO2 arteriography is a very prominent part of our armamentarium for evaluation of patients with RA problems. After RA bypass, we perform CO2 arteriography before all of our patients go home. That way, we know whether any technical defects are present. It is well known that approximately 8% to 10% of RA reconstructions will occlude in the early postoperative time period, and this is believed to be due to technical problems. We have found that using duplex scanning is very good for examining the grafts but very poor for looking at distal RA. Were most distal artery problems found in people with fibromuscallar dysplasia? You had 11 patients who had fibromuscular dysplasia. It is well known that RA with this problem frequently have distal arterial stenoses that are not treatable by surgery. How did this affect the results of your postoperative duplex scanning? Dr. Hudspeth. The example that I showed of a failure to detect a lesion was in the case of fibromuscular dysplasia. The other branch lesions were atherosclerotic. Our policy is that most patients who present for evaluation of renovascnlar hypertension undergo a surface screening examination before they undergo intervention. Almost 95% of patients have undergone intraoperative duplex scanning. Almost 10% of those patients were found to have hemodynamically significant lesions determined on B-mode ultrasonography and confirmed by Doppler scanning and underwent operative correction. In that sense it is very useful. We have not routinely performed arteriography on patients before they go home. But ifa patient presents with recurrent hypertension, we rely on arteriography to assess

390 Hudspeth et al. September 1993 that situation and do not believe that duplex scanning could do an adequate job of detecting that type of lesion. Dr. S. Timothy String (Mobile, Ala.). As the technique is refined, it has great potential for screening for renovascular hypertension. We have been working with this technique in our vascular laboratory over the past several years. Technicians find that they can easily visualize the RA ostium, but beyond several centimeters toward the hilum they have great difficulty. What percentage of renal arteries are you able to visualize in their entirety to the hilum? What is the time frame that this examination requires of your technicians! Finally, you must have some experience with celiac and superior mesenteric artery visualization; consequently have you developed any parameters for these measurements? After the scan is obtained after the patients have fasted, do you feed the patients for a repeat study? Dr. Hudspeth, The typical patient preparation currently at our institution is that they fast for 12 hours before a procedure. Deep abdominal duplex scanning is very different from surface studies of the carotid or peripheral vessels in the leg. We believe that one should budget at least 1 hour for early examinations early in your experience, and it may take longer than that. The series that have reported very low rates of technical completion have been those in which they have limited the examination to 20 minutes so early longer times are necessary. It also has been helpful at our institution for a physician to work with the technician in reviewing technique with the first large group of patients to help to confirm the anatomy and to review that in comparison with angiography on an ongoing basis, which we routinely do. Good patient preparation and cooperation is important for imaging the arteries more distally from the renal artery ostium, as is having a small footprint probe. We believe that is very valuable, especially in looking from the flank and in finding the renal hilum in the distal part of the main RA, examination from the flank on the left side allows the left kidney to provide a very good acoustic window and on the right side the liver does that similarly, which can improve the detection rate of examination and also reduce the angle of insonation and make that examination result more reliable.