Detection of celiac axis and superior mesenteric artery occlusive disease with of abdominal duplex scanning

Transcription

1 Detection of celiac axis and superior mesenteric artery occlusive disease with of abdominal duplex scanning use Timothy R. S. Harward, MD, Sheila Smith, RVT, and James M. Seeger, MD, Gainesville, Fla. Purpose: Detection of mesenteric arterial insufficiency is clinically difficult, and diagnosis frequently requires arteriography. Advances in duplex scanning make this an ideal technique to noninvasively screen patients for chronic mesenteric arterial occlusive disease. However, the accuracy of mesenteric duplex scanning compared with arteriography remains unclear. This study will clearly define the accuracy of abdominal duplex scanning for detection of mesenteric arterial insufficiency. Methods: The mesenteric duplex scans of 38 patients obtained over a 4-year period were reviewed and compared with lateral aortograms to clarify this issue. Optimal peak systolic frequency (PSF) for predicting less than 50% or 50% or greater stenoses of the superior mesenteric artery (SMA) and celiac axis (CA) were determined from receiver-operating characteristic curves. Results: In the SMA a PSF of 4.5 khz was 96% sensitive (24/25), 92% specific (12/13), and 95% accurate (36/38) at predicting stenoses less than 50% or 50% or greater. For arteries with stenoses 50% to 99%, regression analysis demonstrated excellent linear correlation between percent stenosis and PSF (r = 0.89). In the CA a PSF of 4.0 khz had a sensitivity of 100% (30/30), a specificity of 88% (7/8), and an accuracy of 97% (37/38). Again, for arteries with stenoses 50% to 99%, an excellent linear correlation existed between PSF and percent stenosis (r = 0.86). All total arterial occlusions (14) were correctly identified. In all, mesenteric arterial duplex scanning was 96% accurate for predicting SMA and CA stenoses/occlusions. Conclusions: Abdominal duplex scanning is a noninvasive technique that accurately detects total occlusions and objectively quantitates SMA and CA arterial stenoses. (J VASC SURG 1993;17: ) Historically it has been difficult to make a clinical diagnosis of mesenteric ischemia. A classic clinical presentation consists of progressive postprandial pain, weight loss, change in bowel habits, and an epigastric bruit. 1"3 However, this classic clinical picture is not always present, even in patients with significant chronic mesenteric arterial insufficiency. 3s Several other common problems have similar clinical presentations leading the physician to pursue other diagnostic procedures before obtaining a mesenteric arteriogram to define the degree ofmesenteric arterial From the Section of Vascular Surgery, University of Florida College of Medicine, Gainesville. Reprint requests: Timothy R. S. Harward, MD, Section of Vascular Surgery, University of Florida College of Medicine, P.O. Box , JHMHC, Gainesville, FL. Copyright 1993 by The Society for Vascular Surgery and International Society for Cardiovascular Surgery, North American Chapter /93/$ /1/ occlusion present. In addition, physicians are often reluctant to have the elderly, often debilitated patient who commonly is admitted with these symptoms undergo arteriography because of the small but real risk of significant complications. As a result of this type of practice, Stoney 6 noted that the average time needed to make a diagnosis of significant mesenteric arterial occlusive disease in his patients with chronic mesenteric ischemia was 18 months. The celiac axis (CA) and the superior mesenteric artery (SMA) can be readily examined with abdominal duplex scanning because of their location anterior to the aorta. Pulsed-wave Doppler signals are obtained from the mesenteric arteries, and noninvasive detection of significant arterial occlusive disease should be possible on the basis of changes seen in these Doppler signals. This method would allow detection or exclusion of mesenteric arterial disease early in the evaluation of patients suspected of having

2 Volume 17, Number 4 Harward, Smith, and Seeger 739 chronic mesenteric ischemia. However, establishment of diagnostic criteria for the identification of significant mesenteric arterial occlusive disease with use of duplex scanning, and verification of the accuracy of these criteria compared with arteriography haw: been difficult because the prevalence of mesenteric arterial occlusive disease and ischemia in the general population is low. In this report the results of mesenteric arterial duplex scanning are compared with findings on lateral aortograms in a relatively large group of patients with significant nesenteric arterial occlusive disease so that diagnostic criteria fi3r identifying and objective quantifying of CA and SMA disease could be developed. PATIENTS IN THE STUDY AND METHODS Each patient underwent a standard prestudy preparation protocol. This included a clear liquid dinner the evening before the study followed by "nothing by mouth" for the 8 to 12 hours before the study. To decrease gastric gas content, each patient was given 80 mg of simethicone 15 minutes before starting the examination. With the patient in the relaxed supine position, the suprelrenal aorta, CA, and SMA were visualized in the sagitral plane (Fig. 1). The Doppler sample volume was positioned in the lumen of the artery, ~nd frequency spectrum versus time tracings were recorded.. Special attention was paid to maintaining the angle of insonation to the axis of the artery between 45 and 60 degrees. The initial 15 patients were examined with a 3.5 MHz transducer for real-time B-mode imaging combined with a 3.0 MHz pulsed-wave Doppler transducer for insonation (Ultramark-8, Advanced Technology Laboratories, Bothell, Wash.), and the final 23 patients were examined with a 2.5 MHz transducer for both real-time B-mode imaging and pulsed-wave colorflow Doppler insonation (Ultramark-9, Advanced Technology Laboratories). The Doppler sample volume was maintained at 1.5 mm 3 except when the artery was not well seen on B-mode imaging. In the latter situation the sample volume was increased to locate the arterial blood flow, after which attempts were made to decrease the sample volume back to 1.5 mm 3. The CA was visualized and insonated from the aortic takeoff to its bifurcation into the splenic and common hepatic arteries. In like manner the SMA was visualized and insonated from its origin distally for 5 to 6 cm. Data collected were peak systolic frequency (PSF) and end-diastolic frequency (EDF) Fig. 1. Deep abdominal B-mode ultrasonic image and schematic likeness ofsuprarenat aorta (AO), CA, and SMA. Opacity seen on posterior wall of aorta is a hemodynamically insignificant fibrous plaque. (Fig. 2). Inability to locate an audible Doppler signal while generating an image of the aorta and the appropriate mesenteric artery was interpreted to represent total arterial occlusion. Each duplex scan was subjectively graded as to clarity of the B-mode image and quality of the Doppler signal. An excellent study demonstrated clear visualization of the artery bifurcation off the aorta with clean Doppler tracings. A good study allowed accurate demonstration of the arterial anatomy but the walls of the artery were not clearly seen; however, the Doppler signal remained dean. A marginal study occurred when arterial anatomy was not clearly visualized but the course of the artery could be followed with an adequate Doppler signal. A poor

3 IOURNAL OF VASCULAR SURGERY 740 Harward, Smith, and Seeger April 1993 Fig. 2. Comparison of a lateral aortogram, /e~, with the pulsed-wave Doppler tracings obtained from CA and SMA. CA stenosis is > 95%, with a peak systolic frequency of > 12,000 khz. SMA stenosis is 90%, with a peak systolic frequency of > 10,500 khz. study occurred when the artery could not be visualized and an adequate Doppler signal could not be obtained. Transfemoral anteroposterior and lateral aortograms were done in the standard fashion with use of the Seldinger technique. Luminal narrowing of the CA and SMA was graded objectively on each lateral aortogram without direct knowledge of the final duplex scan results. Percent stenosis was objectively calculated by comparing the smallest luminal diameter, measured by caliper, with the nearest downstream luminal diameter where the opposing walls of the artery were parallel and free of disease. Statistical analysis Peak systolic frequency and EDF data collected from each artery examined were compared with percent stenosis. Receiver-operating characteristic (ROC) curves were constructed from Doppler frequency data to determine optimal cutoff values to differentiate between less than 50% and 50% or greater stenosis of the CA and SMA. The accuracy of this pulsed-wave Doppler data for detecting significant mesenteric arterial occlusive disease (_>50% stenosis) as demonstrated on lateral aortography Was determined with 2 2 contingency tables and linear regression analysis. 7 Reproducibility Five healthy subjects underwent mesenteric duplex scanning on three separate sittings. Each individual underwent preexamination preparation identical to the symptomatic patient population preparation. Pulsed-wave Doppler frequency data from the CA and SMA were collected, and average coefficients of variation for repeated measure of the resting values were calculated. 7 RESULTS Between July 1987 and August 1991, 38 patients (15 men, 23 women) were evaluated with both mesenteric arterial duplex scanning and lateral aortography. This group represented 59% (38/64) of the patients at three separate institutions undergoing mesenteric arterial duplex scanning under the supervision of the primary author. The mesenteric duplex scan was done as an outpatient procedure in 15 patients, whereas 23 studies were done on in-hospital

4 Volume 17, NLtmber 4 Harward, Smith, and Seeger 741 patients. Initially, seven studies (18%) (3 outpatient, 4 inpatient) were unsatisfactory because of excessive bowel gas and were repeated the following day, with achievement of an adequate study. Once a technically adequate examination was obtained, 17 studies (45%) were graded as excellent, 14 (37%) as good, and 7 (18%) were marginal. One of the seven marginal studies was due to patient obesity, three were associated with moderate interference of the ultrasound signal by an incisional scar, and three were associated with the presence of a large aortic ~.2eurysm. Mesenteric duplex scans were done to investigate (1) symptoms of abdominal pain, weight loss, diarrhea, and a concomitant epigastric bruit in 26 patients,(69%), (2) mesenteric arterial involvement ofa juxtarenal(5)/thoracoabdominal(2) aneurysm in seven patients (18%), (3) asymptomatic epigastric bruits in three patients (8%), or (4) the study was done as part of the preoperative evaluation before a distal splenorenal shunt in two patients (5%). Fourteen of the 26 symptomatic patients underwent surgical bypass of obstructed mesenteric arteries, one underwent attempted balloon angioplasty of an SMA stenosis,,one died after repair of a symptomatic aortic anemysrn, eight were followed medically, and two were lost to follow-up. Of the eight patients followed medically, surgical intervention was not pursued '-'cause,of patient refusal in one case, associated severe radiation enteritis in two cases, severe concomitant cardiac dysfunction in two, and abdominal symptoms without associated weight loss in three. Lateral aortograms demonstrated a 50% to 99% diameter reduction in 41 arteries (26 CA, 15 SMA). Total arterial occlusion was present in 14 arteries (5 CA, 9 SMA). Seven of the remaining 21 arteries were diseased but had less than 50% diameter reduction, and 14 arteries were classified as normal. Correlation of arteriographic findings with clinical presentation showed the presence of a stenosis 50% or greater in both the CA and SMA of 20 symptomatic and 3 asymptornatic patients. A single artery was 50% or nore stenotic in five symptomatic and five asymptomatic patients, whereas one symptomatic and four asymptomatic patients were free of significant disease in both arteries. Accuracy of peak systolic frequency Superior mesenteric artery. The optimal PSF value to differentiate less than 50% from 50% or greater stenoses calculated from ROC curves was 4.5 khz. With use of this criterion, mesenteric duplex scanning was 95% accurate compared with arteriog- Table I. Statistical evaluation of PSF criteria for diagnosing mesenteric arterial occlusive disease CA SMA Overall Sensitivity 100% 96% 98% Specificity 88% 92% 91% Accuracy 98% 95% 96% PPV 97% 96% 96% NPV 100% 92% 95% PPV, Positive predictive value; NPV, negative predictive value. raphy (Table I). The positive predictive value was 96%, whereas the negative predictive value was 92%. All nine total occlusions were correctly identified. The two incorrect duplex scan interpretations occurred in arteries with arteriographic stenoses between 45% and 55%. A direct, linear relationship was seen between PSF and arteriographically measured stenoses 50% to 99% (r = 0.89,p < ) (Fig. 3). Celiac axis. The optimal PSI: value to differentiate less than 50% from 50% or greater stenoses calculated from ROC curves was 4.0 khz. This duplex scanning criterion was 97% accurate compared with arteriography (Table I). The positive and negative predictive values were 97% and 100%, respectively, and all five total occlusions were identiffed. One duplex scan with a false-positive outcome occurred in a CA that was measured by arteriography to have a stenosis of 45%. Again, a direct linear relationship existed between PSI: and alxeriographically measured stenoses 50% to 99% (r = 0.86, p < ) (Fig. 4). Accuracy of end-diastolic frequency Superior mesenteric artery. The optimal EDI: value to differentiate less than 50% from 50% or greater stenoses calculated from ROC curves was 0.4 khz. This value was 92% accurate when compared with arteriography (Table II). There were two incorrect interpretations, and both were examinations with false-positive results associated with a suprarenal aortic aneurysm. Linear regression analysis produced a correlation coefficient of 0.74 (2 < o.oool). Celiac axis. By ROC curve calculations, the optimal EDI: value to differentiate less than 50% from 50% or greater CA stenoses was 1.2 khz. This value was 90% accurate when compared with arteriography. There were three incorrect interpretations (1 false positive, 2 false negative). The two with false-negative results occurred in an obese patient and

5 742 Harward, Smith, and Seeger April Superior Mesenteric Artery 10 PSF (khz) Y X r I % STENOSIS Fig. 3. Linear regression analysis of pulsed-wave Doppler PSF values obtained in the SMA versus percent stenosis when degree of diameter reduction was 50%; excluded from this analysis were nine totally occluded SMAs. 15 Celiac Axis 10 PSF (khz) I I Y X r ' lbo % STENOSIS Fig. 4. Linear regression analysis of the pulsed-wave Doppler PSF values obtained from CA versus percent stenosis when the degree of diameter reduction was >- 50%; excluded from this analysis were five totally occluded CAs. in a patient with a large aortic aneurysm, whereas the study with false-positive results occurred in a patient examined just before a distal splenorenal shunt. Linear regression analysis produced a correlation coefficient of 0.73 (p < 0.005). Reproducibility Mean PSF in the SMA of healthy volunteers was 2.6 khz (range, 1.5 to 3.5), whereas the mean PSF in the CA was 3.2 khz (range, 2.5 to 3.8). Average coefficient of variation for repeated measures of the

6 Volume 17, Number 4 Harward, Smith, and Seeger 743 resting PSF values in the SMA equaled 8.05%, whereas the average coefficient of variation for CA values was 3.72%. DISCUSSION The diagnosis of chronic mesenteric arterial occlusive disease has been an enigma to physicians for decades. In the past the only objective diagnostic test available to identify occlusive disease in the visceral arteries was the lateral aortogram. However, use of arteriography in patients thought potentially to have aesenteric arterial occlusive disease was often delayed because the test was expensive, invasive, and risky in elderly patients most likely to have this problem. This scenario was especially true in patients who did not have with the classic signs and symptoms of postprandial abdominal pain, weight loss, change in bowel habits, and an epigastric bruit. Recent advances in the technology of B-mode ultrasound, pulsed-wave Doppler and computer software have allowed development of duplex scanners that: can produce images of the mesenteric arteries and provide concomitant Doppler spectral analysis of blood flow in these arteries. Initially several reports appeared evaluating normal resting blood flow patterns in mesenteric arteries and physiologic changes in this blood flow seen in response to oral consumption of a standard test meal How- ~-er, few reports exist that evaluate blood flow patterns in mesenteric arteries with significant stenoses. In 1984 J~iger et al.13 reported a case study of one patient with symptoms of mesenteric arterial insufficiency found on mesenteric duplex scanning to have markedly abnormal Doppler flow patterns (peak systolic w:locities > 300 cm/sec) in the CA and SMA. Arteriography verified high-grade stenoses in both arteries. Two years later, Nicholls et al.~4 attempted to define arterial pulsed-wave Doppler curve characteristics diagnostic for significant mesenteric arterial stenoses. However, only four patients were in the study, arid definitive conclusions could not be reached. Recently Moneta et al.~5 compared the 'esults of mesenteric duplex scanning with the findings on lateral aortography in 34 patients. They found that peak systolic velocities 275 cm/sec or greater arid 200 cm/sec or greater defined 70% or greater stenosis of the SMA and CA, respectively. The calculated accuracies of these predictors were 91% and 82%, respectively. In a similar manner Bowersox et al) 6 correlated velocity measurements with arterial occlusive disease and found that an end-diastolic velocity greater than 45 cm/sec was 95% accurate in discriminating between a less than Table II. Statistical evaluation of EDF criteria for diagnosing mesenteric arterial occlusive disease CA SMA Overall Sensitivity 91% 100% 94% Specificity 88% 82% 84% Accuracy 90% 92% 91% PPV 95% 87% 92% NPV 78% 100% 89% 50% and 50% or greater stenosis of the SMA; however, they were not able m identify a velocity parameter that would discriminate between minimal or severe stenosis in the CA. In the present study with PSF data, the accm'acy of the duplex examination for defining stenoses less than 50% or 50% or greater was excellent ( > 95%) for both the SMA and CA. It is more important to note that the relationship of increasing PSF was directly proportional to increasing arterial percent stenosis in the CA and SMA such that the degree of stenosis between 50% and 99% was accurately predicted. In addition, all 14 total occlusions were easily identified. The use of EDF data did not increase the accuracy of the examination, although the trends were similar to those seen with use ofpsf. A possible explanation for the lesser accuracy of EDF criteria might be that the diagnostic window for these criteria is small, making precise placement of the pulsed-wave Doppler sample volume into the centerstream of flow exiting a critical stenosis more crucial for collecting accurate EDF data than for obtaining accurate PSF data. In fact, 18% of the mesenteric examinations in this study did not have a clear image of the CA and SMA, and without clear anatomic detail, accurate positioning of the pulsed-wave Doppler sample volume can be diffficult. Therefore these marginal studies may have had a significant negative impact on EDF data collection, leading to less accurate results than seen with PSF data. The quality and reproducibility of mesenteric duplex scanning were also examined in this study. Eighteen percent of the patients required a second examination because of technical troubles at the initial setting; however, the final quality of the examination was good to excellent in 82% of the patients. Also, in healthy subjects the variability of the frequency shift paraaneters used in this study were small, suggesting excellent reproducibility. In fact, the pooled coefficients of variations for repeated measures in the SMA were very similar to those published by Qamar et al.9 Thus in light of the

7 744 Harward, Smith, and Seeger April 1993 accuracy of this examination for detecting significant CA and SMA flow changes, this variability appears satisfactory for clinical use. Two possible explanations exist for the results obtained in this report. First, no time limits were imposed on study performance so that each examination was continued until the examiner was confident of the accuracy of the data collected. As mentioned by Moneta et al.,~s the anatomy of this area can be complex and occasionally confusing. Rushing the performance of this examination could well lead to acquisition of inaccurate data. Second, this study predicts arterial abnormalities with use of frequency data collected while attempting to maintain a fixed constant Doppler angle of 45 degrees to 60 degrees relative to the axis of the artery. Phillips et al.~7 have suggested that this method will provide more meaningful results than velocity measurements because of intrinsic problems that occur in estimating flow velocity with present day instrumentation. To obtain flow velocity data with currently available instrumentation, two assumptions are required: (1) blood flow is uniform along the axis of the artery, and (2) the direction of the velocity vector is unidirectional. However, arterial blood flow is nonaxial, ~s and velocity is a three-dimensional vector with quantities of magnitude and direction. The standard duplex scanners in use today are equipped with only a single transducer; therefore only one directional component of the three-dimensional flow velocity vector can be evaluated, and the other two components remain unknown. Without complete evaluation of all components of the velocity vector, the Doppler angle of the ultrasound beam as it relates to the actual velocity vector is unknown. Without accurate knowledge of this angle, the conversion of Doppler frequency shift data to velocity measurements is very unreliable. In fact, a 10-degree to 20-degree variation in the angle of the ultrasound beam with respect to nonaxial blood flow can create a 50% to 100% error in estimated flow velocity measurements. 17 Rizzo et al.19 verified this clinically, showing that as the angle ofinsonation increases over 70 degrees, the calculated velocity values became more unreliable, increasing as much as 120% over control values obtained with a 60-degree Doppler angle. Like velocity measurements, frequency measurements also have intrinsic problems. They depend on (1) the angle of the ultrasound beam with respect to the axis of the artery and (2) the frequency of the emitted Doppler signal. 17 However, if one attempts to maintain the ultrasound signal at a constant fixed angle with respect to the axis of the artery, the issue of angle variability is markedly lessened. Also, since the transmitted frequency is important in determining the frequency shift value, each examination should be performed with the same probe so that the transmitted ultrasound signal remains consistent. In this study the pulsed-wave Doppler frequency changed from 3.0 MHz to 2.5 MHz because of an upgrade in equipment to color-flow instrumentation. The difference in transmitted frequency of these two probes was small and did not cause a significz~7~. - difference in the final generated criteria. However, if a greater difference in transmitted frequency had occurred, a significant discrepancy in the final predictive criteria would have been expected. Still, with use of the approach described in this study, changes in frequency shift data appear to accurately represent variations in blood flow patterns caused by arterial abnormalities without significant variations in data calculation techniques. The linear correlation obtained between frequency shift data and percent stenosis in this study seems to strongly justify this approach in our laboratory. Until three-dimensional duplex scanners are available, accurate determination of the Doppler angle is impossible. Without a reliable and reproducible method to determine the angle of insonation of all velocity vector components, calculation of velocj~ magnitude via the Doppler equation will remain ohly a rough estimates Given the accuracy and relative ease ofmesentetic duplex scanning in comparison with other deep abdominal examinations (e.g., renal artery scanning), it is tempting to suggest that mesentetic duplex scanning will greatly aid in deciphering the physiologic significance of varying degrees of mesentetic arterial occlusive disease in patients with abdominal complaints. However, mesenteric duplex scanning can only predict the presence of anatomic disease in the CA and SMA. Because an extensive arterial collateral network exists between the CA, SMA, and inferior mesenteric artery that varies from patient t~7 patient, mesenteric duplex scanning can provide no insight into the quantity of blood flow below which intestinal ischemia will occur. Therefore as suggested by others, Is'16 mesenteric duplex scanning should be used only as a diagnostic screening test for patients with clinical signs/symptoms suggestive of mesenteric arterial occlusive disease or chronic intestinal ischemia or both. On the basis of results of this study, we have generated the following diagnostic guidelines. First,

8 Volume 17, Number 4 Harward, Smith, and Seeger 745 an SMA PSF less than 4.5 khz and a CA PSF less than 4.0 khz predict an arterial stenosis less than 50%. It is not possible to accurately differentiate between the normal artery and the diseased artery with stenosis less than 50% because of the extensive overlap :seen between peak systolic data obtained from healthy volunteers and those seen from patients with arterial stenoses less than 50%. Second, an SMA PSF 4.5 khz or greater and a CA PSF 4.0 khz or greater both predict arterial stenosis 50% or greater. In this setting the linear regression line generated q~om the above data is used to predict the percentage diameter reduction (50% to 99%) of the respective artery. Finally, if an appropriate arterial pulsed-wave Doppler signal is absent, regardless of whether or not an adequate image can be obtained of the artery, the artery is predicted to be totally occluded. In conclusion, duplex scanning of the mesenteric arterial system provides an accurate method to determine the severity of CA and SMA disease. However, mesenteric duplex scanning can only predict the presence of an anatomic lesion, and correlation with clinical symptoms is necessary to determine the physiologic consequences of altered intestinal blood flow. Regardless, the accuracy of this examination means that the mesenteric circulation can be ~:eliably studied in symptomatic patients, which should decrease the delay in diagnosing this,~nusual but important problem. REFERENCES 1. Bergan JJ, Yao JST. Chronic mesenteric ischemia. In: Rutherford RB, ed. Vascular surgery. 3rd ed. Philadelphia: WB Sannders, 1989: Dunphy JE, Abdominal pain of vascular origin. Am J Med 1936;192: Laufman H, Nora PF, Mittelpunkt AI. Mesenteric blood vessels: advances in surgery and physiology. Arch Surg 1964;88: Morris GC, Crawford ES, Cooley DA, DeBakey ME. Revascularization of the celiac and superior mesenteric arteries. Arch Surg 1962;84: Reul GJ, Wukasch DC, Sandiford FM, ChiariUo L, Hallman GL, Cooley DE. Surgical treatment of abdominal angina: review of 25 patients. Surgery 1974;75: Stoney RJ. In discussion: Moneta GL, Yeager R_A, Dalman R, Antonovic R, Hall LD, Porter JiM. Duplex ultrasound criteria for diagnosis ofsplanchnic artery stenosis of occlusion. J VASC SURG 1991;14: Daniel WW. Biostatistics: a foundation for analysis in the health sciences. 3rd ed. New York: John Wiley & Sons, 1983: Qamar MI, Read AE, Skidmore R, Evans JM, Williamson RCN. Transcutaneous Doppler ultrasound measurements of celiac axis blood flow in man. Br J Surg 1985;72: Qamar MI, Read AE, Skidmore R, Evans JM, Wefts PNT. Transcutaneous Doppler ultrasound measurement of superior mesenteric artery blood flow in man. Gut 1986;27: J~iger K, Bollinger A, Valli C, Ammann R. Measurement of mesenteric blood flow by duplex scanning. J VAsc SURG 1986;3: Lilly MP, Harward TRS, Flinn WR, et al. Duplex ultrasound measurement of changes in mesenteric flow velocity with pharmacologic and physiologic alteration of intestinal blood flow in man. J VASC SURG 1989;9: Moneta GL, Taylor DC, Helton WS, et al. Duplex ultrasound measurement of postprandial intestinal blood flow-: effect of meal composition. Gastroenterology 1988;95: J~iger KA, Former GS, Thiele BL, Strandness DE. Noninvasire diagnosis of intestinal angina. J Clin Ultrasound 1984; 12: Nicholls SC, Kohler TR, Martin R_L, Strandness DE. Use of hemodynamic parameters in the diagnosis of mesenteric insufficiency. J VASC SURG 1986;3: Moneta GL, Yeager RA, Dalman R, Antonovic R, Hall RD, Porter JM. Duplex ultrasound criteria for diagnosis of splanchnic artery stenosis or occlusion. J VASC SURe 1991; 14: Bowersox JC, Zwolak RM, Walsh DB, et al. Duplex ultrasonography in the diagnosis of celiac and mesenteric artery occlusive disease. J VAse Suing 1991;14: Phillips DJ, Beach KW, Primozich J, Strandness DE. Should results of ultrasound Doppler studies be reported in units of frequency or velocity? Ultrasound Med Bio11989; 15: Strandness DE, Sumner DS. Physics of arterial blood flow. In: Strandness DE, Sumner DS, eds. Hemodynamics for surgeons. New York: Grime and Stratton, 1975:73-95, 19. Rizzo RJ, Sandager G, Astleford P, et al. Mesenteric flow velocity variations as a function of angle of insonation. J VAsc SURG 1990;11: Submitted Dec. 17, 1991; accepted July 22, 1992.