Mesenteric duplex scanning: A blinded prospective study

Mesenteric duplex scanning: A blinded prospective study Gregory L. Moneta, M D , R a y m o n d W. Lee, M D , Richard A. Yeager, M D , L l o y d M. Taylor, Jr., M D , and J o h n M. Porter, M D , Portland, Ore.

Purpose: Based on retrospective comparisons of duplex scanning with arteriography of the celiac (CA) and superior mesenteric (SMA) arteries in 34 patients, we previously suggested that an SMA peak systolic velocity of 275 cm/sec or greater or no flow signal and a CA PSV of 200 cm/sec or greater or no flow signal were reliable indicators of a 70% or greater angiographic stenosis of the SMA and CA, respectively. We now report the results of a blinded, prospective study in a larger patient group designed to determine the ability of mesenterie duplex scanning to visualize the CA and SMA and to validate our proposed duplex criteria for splanchnic artery stenosis. Methods: During an 18-month period 100 patients admitted to our vascular surgery service for aortography underwent routine mesenteric artery duplex scanning and lateral abdominal aortography regardless of abdominal symptoms. The lateral aortograms were evaluated to determine the presence or absence of a 70% or greater stenosis in the CA or SMA. Duplex-determined peak systolic velocities from the CA and SMA were recorded without knowledge of the angiographic results. Results: Aortography satisfactorily visualized 100% of the CAs and 99% of the SMAs. Of these, 93% of the SMAs and 83% of the CAs were visualized by duplex. According to the above criteria, duplex sensitivity, specificity, positive predictive value, negative predictive value, and overall accuracy for detection of a 70% or greater SMA stenosis were 92%, 96%, 80%, 99%, and 96% and for a 70% or greater CA stenosis 87%, 80%, 63%, 94%, and 82%. Conclusions: Mesenteric duplex scanning is feasible in the majority of patients. Prospective evaluation of duplex diagnostic criteria for 70% or greater stenosis indicates that mesenteric duplex scanning is sufficiently accurate to be clinically useful as a screening examination to detect SMA and CA stenosis. (J VAse SURG 1993;17:79-86.)

Atherosclerotic occlusion or narrowing of the orifices of the celiac artery (CA) and superior mesenteric artery (SMA) accounts for most cases of chronic intestinal ischemia (CII). Visceral angiography has been the primary diagnostic modality used in the assessment of patients with suspected CII. However, t h e rarity of CII, combined with the reluctance of most physicians to employ visceral angiography early in the evaluation of patients with abdominal pain, has generally resulted in late diagnosis of CII. A recent report describes a mean lag From the Divisionof Vascular Surgery, Oregon Health Sciences Universityand Veterans AffairMedicalCenter, Portland. Supported by a grant from the Department of Veterans Affairs. Presented at the Forty-sixthAnnual Meeting of the Societyfor Vascular Surgery, Chicago,Ill,, June 8-9, 1992. Reprint requests: GregoryL. Moneta,MD, AssociateProfessorof Surgery,Divisionof VascularSurgery(OP-11), OregonHealth Sciences University, 3181 S. W. Sam Jackson Park Rd., Portland, OR 97201. 24/6/41706 0741-5214/93/$1.00 + .10.

time of 18 months from the onset of abdominal symptoms to a confirmed diagnosis of CII.1 Use of mesenteric duplex scanning (MDS) to detect high-grade splanchnic artery stenosis noninvasively in patients with symptoms of CII was initially reported by J~iger et al. 2 in 1984. Since then, a number of investigators have suggested the possibility of developing MDS criteria for the diagnosis of significant CA and superior mesenteric artery (SMA) stenosisf1-6 Recently, we retrospectively correlated hemodynamic data from mesenteric duplex scans of the CA and SMA with visceral angiography in 34 patients with atherosclerotic vascular disease. 5 Based on these data, we postulated that fasting peak systolic velocities (PSVs) were potentially accurate predictors of high-grade splanchnic artery stenosis. Specifically, a PSV of 200 cm/sec or greater or no flow signal in the CA and a PSV of 275 cm/sec or greater or no flow signal in the SMA predicted the presence of a 70% or greater angiographic stenosis or occlusion in the CA 79

80

M o n e t a et al.

and SMA, respectively. This retrospective study was regarded by us as a hypothesis-seeking step in our effort to evaluate the efficacy of MDS. We next designed a blinded study to evaluate the ability of MDS to visualize the splanchnic arteries and to prospectively validate our proposed criteria for detection of 70% or greater SMA and CA stenosis. The results of the blinded, prospective study form the basis for this report.

METHODS Patient selection. Beginning July 1990, patients admitted to the vascular surgery service at the Portland Veteran Affairs Medical Center for aortography and treatment of visceral or peripheral arterial atherosclerotic disease were, whenever possible and without regard to body habitus or presence of abdominal symptoms, evaluated with fasting MDS. Potential study patients therefore included individuals with suspected CII (n = 13), as well as patients with symptomatic peripheral vascular disease (n = 87). Lateral aortography was obtained routinely on all these patients. The study was terminated prospectively after 100 patients had undergone both MDS and lateral aortography. This protocol was approved by the Human Subjects Review Committee of the Oregon Health Sciences University Hospital and the Portland Veterans Affairs Hospital. Informed consent was obtained from all patients. All study patients underwent angiography and MDS within 2 weeks of each other without intervening treatment of the visceral vessels. A 2-week period between studies was allowed because there would be little likelihood of a change in vessel status during this time and it facilitated workups of patients traveling long distances to the Portland Veterans Affairs Hospital. Angiographic technique and evaluation. Aortography was performed by the standard Seldinger technique from either a femoral or axillary approach. Both regular "cut" films of the lateral abdominal aorta and intraarterial digital subtraction studies were used. To limit contrast load, digital subtraction studies were used in the majority of patients without symptoms suggestive of mesenteric ischemia. Cut films were used in all patients in whom the primary indication for angiography was possible mesenteric vascular disease. Angiographic stenosis of the CA and SMA was recorded as percent diameter reduction based on caliper measurements of the width of the contrast column at the point of maximal narrowing compared with the widest portion of the contrast column not obviously involved in poststenotic dila-

Journal of VASCULAR SURGERY

tation. All lateral aortograms were graded without knowledge of the results of MDS examinations. Mesenteric artery duplex scanning technique. MDS was performed by one of two registered vascular technologists with extensive research and visceral duplex experience and without knowledge of the angiographic results. Examinations were performed with a color duplex scanner (Acuson 128; Acuson, Inc., Mountain View, Calif.) with either a 2.5 or 3.5 MHz sector transducer. All patients were examined after an overnight fast in the supine position with the head of the bed elevated 30 degrees. The CA, SMA, and supraceliac aorta were imaged according to a previously described technique. 7 Doppler waveforms were obtained with a measured Doppler angle of less than 70 degrees. Peak systolic velocity (PSV) and end-diastolic velocity (EDV) were recorded from the CA and SMA. MDS was evaluated for its ability to visualize the CA and SMA. CA PSV and SMA PSV were used to categorize duplex studies as indicating a less than 70% stenosis or a 70% to 100% stenosis in the CA and SMA, according to our criteria listed abovefi Data analysis. All data on the 100 patients were coded and entered into a personal computer data base. The duplex results were assessed with angiography used as the standard. The sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of MDS to detect 70% or greater angiographic stenosis in the CA and SMA were calculated. The mean _ SD of the PSV and EDV in the CA and SMA in patients with less than 70% angiographic stenosis were calculated and compared with the corresponding mean velocities in patients with 70% to 100% angiographic stenosis. The nonparametric Wilcoxon two-sample test was used to evaluate significance of the differences in Doppler velocities (CA PSV, CA EDV, SMA PSV, and SMA EDV) between patients with less than 70% anglographic stenosis and patients with 70% or greater angiographic stenosis. RESULTS Patients The median and mean ages of the 100 patients entered into the study were 68 years and 67 + 9.5 years (range 33 to 81 years). All but 12 patients were men. Thirteen patients were undergoing evaluation for possible CII.

Angiography Lateral aortography adequately visualized the splanchnic vessels for determination of CA and SMA

Volume 17 Number 1 j'anuary 1993

stenosis in 100% of the CAs and 99% of the SMAs. A large calcific aortic plaque precluded adequate visualization of the SMA origin in one patient. There were 20 patients with 70% or greater stenosis of the CA, four patients with complete CA occlusion, and 76 patients with less than 70% stenosis of the CA. For the SMA, angiography revealed 13 patients with 70% or greater stenosis, one patient with complete occlusion, and 85 patients with less than 70% stenosis.

Mesenteric duplex scanning Celiac artery. MDS successfially visualized 83 (83%) of 100 angiographically visualized CAs. A technically satisfactory duplex examination demonstrafing CA PSV of 200 cm/sec or greater or no flow signal predicted a 70% to 100% angiographic stenosis with a sensitivity of 87%, specificity of 80%, positive predictive value of 63%, and negative predictive value of 94% (Table I). The accuracy was 82%. One patient with a 70% o r greater CA angiographic stenosis had a technically unsatisfactory duplex examination. If this patient is included in the analysis, MDS identified 83% (20 of 24) of patients with 70% to 100% stenosis in the CA. Mean CA PSV for patients with 70% to 99% angiographic stenosis, including the three patients with false-negative duplex studies (Table I) by our criteria, was 352 _+ 154 cm/sec (range 78 to 600 cm/sec). The mean EDV for this group was 100 + 58 cm/sec (range 23 to 234 cm/sec). CAPSV for patients with less than 70% angiographic stenosis, including the patients with false-positive duplex studies (Table I), was 175 _+ 90 cm/sec (range 52 to 600 cm/sec). The mean EDV for this patient group was 52 _+ 56 cm/sec (range 17 to 454 cm/sec) (Fig. 1). Both PSV and EDV were increased significantly in patients with a 70% to 99% angiographic CA stenosis compared with patients with a less than 70% angiographic stenosis (p = 0.0001). Superior mesenteric artery. Ninety-two of 99 angiographicaUy visualized SMAs were visualized successfully by MDS. A technically satisfactory duplex scan demonstrating SMA PSV of 275 cm/sec or greater or no flow signal predicted a 70% to 100% angiographic stenosis in the SMA with a sensitivity of 92%, specificity of 96%, positive predictive value of 80%, negative predictive value of 99%, and accuracy of 96% (Table II). One of the seven patients with a technically unsatisfactory duplex examination had a 70% or greater stenosis by angiography. Including this patient, MDS successfully identified 12 (86%) of 14 patients with a 70% to 100% stenosis in the SMA.

A/lesenteric duplex scanning

81

Table I. Comparison of MDS and angiography for detection of 70% to 100% angiographic stenosis in the CA (n = 83) Angiographic stenosis (n) Duplex

70%-100%

< 70%

70%-100% < 70%

20 3

12 48

Mean PSV and EDV for patients with 70% to 99% angiographic stenosis in the SMA, including a patient with a false-negative duplex study (Table II), was 395 -+ 143 cm/sec (range 65 to > 600 cm/sec) and 109 _+ 75 cm/sec (range 19 to 228 cm/sec), respectively. Including three patients with falsepositive duplex studies (Table II), mean SMA PSV for patients with less than 70% angiographic stenosis was 167 +- 62 cm/sec (range 53 to 385 cm/sec) and mean EDV was 23 + 13 cm/sec (range 0 to 88 cm/sec) (Fig. 2). Both PSV and EDV were increased significantly in patients with a 70% to 99% angiographic SMA stenosis compared with patients with a less than 70% angiographic stenosis (p = 0.0001).

Patients with suspected CII Six of the 13 patients evaluated for possible CII underwent mesenteric revascularization. All six patients had 70% to 100% SMA angiographic stenosis and four had 70% to 100% CA angiographic stenoses. Duplex scanning identified five of the six high-grade SMA stenoses and three of the four high-grade CA stenosis in these patients.

DISCUSSION The potential of MDS to identify splanchnic artery stenosis was recognized nearly 10 years ago. However, it was only recently that our group proposed the first MDS criteria for detection of 70% or greater CA and SMA stenosis, s Only 2 months after our initial report, Bowersox et al.6 from Dartmouth published a similar retrospective review comparing MDS and angiography and suggested an SMA EDV greater than 45 cm/sec as predictive of a 50% or greater SMA stenosis (sensitivity 100%; specificity 92%). They found that an SMA PSV greater than 300 cm/sec was highly specific (100%) but less sensitive (63%) for detecting a 50% or greater SMA stenosis. No duplex criteria for CA stenosis were identified. Our current study represents the first blinded prospective validation of specific duplex criteria for diagnosis of CA and SMA stenosis. The percentage of patients with high-grade CA

Jo~n~of VASCULAR SURGERY

82 2klone~a et al.

¼

650 600

•

70%

•

< 70%

- 99%

Stenosis

I

Stenosis

550 500

1

450

o~ E

vO ,m O (D >

400-

T

-x

350

•

•i

300

250200-

'

150-

i

100-

500 I

I PSV

t

1

EDV

Fig. 1. Distribution of CA, PSVs and EDVs in patients with less than 70% CA angiographic stenosis (.) and patients with 70% to 99% CA angiographic stenosis (A). Table II. Comparison of MDS and angiography for detection of 70% to 100% angiographic stenosis in the SMA (n = 92) Angiographic stenos# (n) Duplex

70%-100%

< 70%

70%-i00% < 70%

12 1

3 76

Table III. Comparison of retrospectively and prospectively derived sensitivities, specificities, positive predictive values, negative predictive values, and overall accuracies of MDS in detection of 70% or greater angiographic stenosis in the CA and SMA Retrospective Current prospective

studyS(%)

stenosis (24%) identified in this study is consistent with that reported previously.7 The percent of patients with high-grade SMA stenosis (14%), however, is higher than expected in a general vascular surgical population. This undoubtedly reflects the inclusion of 13 patients in this study with suspected CII, six of whom actually underwent SMA revascularization. Our prospective data suggest that the proposed duplex criteria for detection of 70% to 100% angiographic stenosis of the CA and SMA accurately detect high-grade angiographic SMA stenosis. The ability of duplex to detect high-grade CA stenosis, however, is not as good as that for the SMA. These data are quite comparable to those from our retrospective studys (Table III). Mean PSV and EDV for patients with angiographically normal CAs and SMAs were also consistent with the values obtained from our retrospective study (Table IV). It is difficult to compare duplex criteria for SMA stenosis from Bowersox et al.6 directly with our own because their proposed criteria were for detection of a 50% or greater angiographic SMA stenosis,

Sensitivity Specificity Positive predictive value Negative predictive value Accuracy

study (%)

CA

SMA

CA

SMA

75 89 86 80 82

89 92 80 96 91

87 80 63 94 82

92 96 80 99 96

whereas we have sought to evaluate criteria for detection of 70% to 100% angiographic splanchnic artery stenosis. Although we feel most patients with symptomatic CII will have visceral artery stenosis that exceeds 70%, it is interesting to examine our data according to the criteria proposed by Bowersox et al.6 (Tables V and VI). In our patients an SMA PSV of 300 cm/sec or greater or no flow signal predicts a 50% or greater angiographic stenosis of the SMA with a sensitivity of 59%, specificity of 97%, positive predictive value of 87%, negative predictive value of 89%, and accuracy of 88%. These values are comparable to those from the Dartmouth study: sensitivity of 63%, specificity of 100%, and accuracy of 8 6 % . 6 Our sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of MDS to

Volume 17 Number 1 January 1993

Mesenteric duplex scanning

600 550 500 450 400 0 v

350 300

• 70% - 99% Stenosis • < 70% Stenosis

I i

I

I

I

83

i

I

1" 7

250 (D

>

200 150-

100A

50-

l i i,,:,,,..... •

•

0 0

I

I

t

PBV

J

EDV

Fig. 2. Distribution of SMA, PSVs and EDVs in patients with less than 70% SMA angiographic stenosis (.) and patients with 70% to 99% SMA angiographic stenosis (•). Table IV. Comparison of retrospectively and prospectively derived PSVs and EDVs for patients with angiographically normal CA or SMA Retrospective study (cm/sec) CA PSV CA E D V SMA PSV SMA E D V

138 31 163 28

-+ ---+ -+

38 15 59 15

Current prospective study (cm/sec) 162 53 163 22

--+ 89 -+ 73 .+_ 60 -+ 10

predict a 50% or greater SMA stenosis based on an SMA EDV of 45 cm/sec or greater, as suggested by the Dartmouth group, were 48%, 99%, 91%, 86%, and 86%. The specificity of the EDV Dartmouth criteria in our hands is therefore also similar in our prospective study to that obtained retrospectively by Bowersox et al. 6 Our sensitivity of 48%, using an EDV of greater than 45 cm/sec to detect a 50% or greater stenosis, is less than that reported by the Dartmouth group. Overall, our results indicate that PSV is a more consistent and reliable parameter to evaluate visceral artery stenosis. It is important to note that MDS depends primarily on analysis of duplex-derived Doppler velocity waveforms. Doppler interrogation of deep abdominal vessels requires the use of a relatively low-frequency transducer at the inescapable cost of deterioration in B-mode image resolution. 8 The B-mode image is used solely as a guide for placement of the Doppler sample volume. With less than optimal image resolution, erroneous placement of

•Table V. Comparison of MDS and angiography for detection of a 50% or greater angiographic SMA stenosis based on an SMA PSV greater than 300 cm/sec (n = 92) Angiographic stenosis (n) Duplex

50%-100%

< 50%

50%-100% < 50%

13 9

2 68

Table VI. Comparison of MDS and angiography for detection of a 50% or greater angiographic SMA stenosis based on an SMA EDV greater than 45 cm/sec (n = 88) Angiographic stenosis (n) Duplex

50%-I 00%

< 50%

50%-100% < 50%

10 11

1 66

the Doppler sample volume may be a significant source of error. Indeed we were not able to visualize all the mesenteric vessels we attempted to study, usually because of obesity or excessive bowel gas. We suspect the most likely source of a false-positive duplex examination is insonation of the vessel at a Doppler angle greater than that indicated by the software of the duplex scanner. Rizzo et al. 9 have shown that Doppler angles in excess of 70 degrees

84

Journal of VASCULAR SURGERY

AJoneta et al.

produce falsely elevated velocity measurements. It is our practice to maintain an angle of insonation less than 70 degrees to minimize angle-dependent variation in flow velocity. Nevertheless, significant tortuosity or angulation of the visceral vessels, especially the CA, over short distances may preclude accurate insonation of some vessels by even the most experienced technologist. Other sources of error include failure to insonate the precise portion of the vessel containing the lesion and placing the Doppler sample volume erroneously over a collateral artery. Interrogation of the artery distal to a high-grade stenosis may reveal a low PSV and thus be erroneously interpreted as normal. We believe detection of a low PSV, especially if it is less than aortic PSV, mandates a vigilant proximal search to ensure that an area of severe stenosis or occlusion has not been missed. This type of error likely accounted for our false-negative examination in a single patient with high-grade stenosis in both the CA and SMA. Recently it has been suggested that other duplex findings may serve as useful adjuncts in the diagnosis of severe CA or SMA stenosis. We have reported the ability of duplex scanning to demonstrate enlargement of the inferior mesenteric artery and arc of Riolan in two patients with high-grade CA and SMA stenosis) ° Abnormal hepatic artery flow patterns such as retrograde flow, turbulent flow, and diminished pulsatility indexes may be suggestive of CA stenosis, z~ A postprandial study may help differentiate patients with normal SMAs from those in whom a significant SMA lesion was not detected during a fasting examination by systolic velocity criteria.S Failure of SMA velocities to increase 20 to 30 minutes after a test meal may indicate a significant SMA stenosis. This hypothesis, however, remains speculative and unsupported by systematically gathered data. It is critically important to differentiate highgrade mesenteric artery stenosis from the clinical entity of CII. Not all patients who have mesenteric artery obstruction have CII. Symptom-free patients with angiographically proved high-grade stenosis or occlusion of the CA or SMA are well recognized. Duplex detection of high-grade splanchnic artery lesions therefore cannot be used to confirm the clinical entity of CII. However, because the majority of patients with CII do indeed have very high-grade lesions of the splanchnic arteries, ~2 the use of duplex scanning to identify 70% to 100% stenosis in the CA and SMA may be useful in guiding the diagnostic evaluation of patients with postprandial abdominal pain. The above data indicate that MDS performed by experienced technologists can be an useful noninva-

sive technique for detection of high-grade visceral artery stenosis. We believe this diagnostic modality should be used early in the evaluation of patients with abdominal pain consistent with CII. Patients with technically satisfactory duplex examinations and norreal visceral duplex findings (SMA PSV <275 cm/sec and CA PSV < 200 cm/sec) will rarely have 70% or greater visceral artery stenosis (high specificity). Visceral angiography may then be reserved for patients with symptoms suggestive of CII and a positive duplex examination or, infrequently, a nonvisualizing duplex examination. We acknowledge with appreciation the technical expertise of Cary Cummings RN, RVT, and John Caster RN, RVT, in the performance of the mesenteric duplex scanning. We also acknowledge the excellent assistance o f biostaricians Gary Sexton, PhD, and David Wilson, MS, in data management and analysis. REFERENCES

1. Schneider PA, Ehrenfeld WK, Cunningham CG, Reilly LM, Goldstone J, Stoney RJ. Recurrent chronic visceral ischemia. J VASC SURG (in press). 2. J~iger KA, Former GS, Thiele BL, Strandness DE Jr. Noninvasive diagnosis of intestinal angina. J Cfin Ultrasound 1984;12:588-91. 3. Nicholls SC, I(ohler TR, Martin RL, Strandness DE Jr. Use of hemodynamic parameters in the diagnosis of mesenteric insufficiency. J VAsc SUING1986;3:507-10. 4. Flinn WR, Rizzo RJ, Park JS, Sandager GP. Duplex scanning for assessment of mesenteric ischemia. Surg Cfin North Am 1990;70:99-107. 5. Moneta GL, Yeager RA, Dalman R, Antonovic R, Hall LD, Porter ~M. Duplex ultrasound criteria for diagnosis of splanchnic artery stenosis or occlusion. J VAsc SURG 1991; 14:511-20. 6. Bowersox IC, Zwolak RM, Walsh DB, et al. Duplex ultrasonography in the diagnosis of celiac and mesenteric artery occlusive disease. J VASCSURG 1991;14:780-8. 7. Valentine RJ, Martin JD, Myers SI, Rossi MB, Clagett GP. Asymptomatic celiac and superior mesenteric artery stenoses are more prevalent among patients with unsuspected renal artery stenoses. J VASCSURG 1991; 14:195 -9. 8. Moneta GL, Taylor DC, Yeager RA, Porter JM. Duplex ultrasound: apphcafions to intraabdominal vessels. Perspect Vase Surg 1989;2:133-48. 9. Rizzo R1, Sandager G, Asfleford P, et al. Mesenteric flow velocity variations as a function of angle ofinsonation. 1 Vhsc SURG 1990;11:688-94. 10. Moneta GL, Cummings C, Caster 1, Porter JM. Duplex ultrasound demonstration of postprandial mesenteric hyperemia in splanchnic circulation collateral vessels. J Vase Technol 1991;15:37-9. 11. LaBombard FE, Musson A, Bowersox JC, Zwolak RM. Hepatic artery duplex as an adjunct in the evaluation of chronic mesenteric ischemia. J Vasc Technol 1992;16:7-11. 12. Geelkerken RH, van Bockel JH, de Roos WK, Hermans J, Terpstra JL. Chronic mesenteric vascular syndrome: results of reconstructive surgery. Arch Surg 1991;126:1101-6. Submitted lune 12, 1992; accepted Aug. 7, 1992.

Volume 17 Number 1 January 1993

Mesenteric duplex scanning 85

DISCUSSION Dr. D. Eugene Strandness (Seattle, Wash.). You have taken another step in identifying those velocity criteria that may be useful in documenting the status of the superior mesenteric and celiac arteries. Calls for studies of the mesenteric vessels generally fall into two categories: The first is patients who have a clinical syndrome that is confusing and may have as one of the possible causes diseases of the celiac and superior mesenteric arteries. In our experience these are the most frequent referrals, and in most cases they are very simple to deal with because if one of the three major potential inputs to the small bowel is normal, one can rule out disease of the mesenteric arteries as being relevant. The second group of patients includes those who have the classic syndrome of chronic mesenteric ischemia whereby it is important to know the degree of involvement and where. This has also not been a problem in our experience to date. Although I do not disagree with some of the findings, there are several aspects of this study that need to be addressed and answered because they tend to detract from the worth and potential impact of the data for clinical practice. My concerns are as follows: (1) You had a golden opportunity to check on the intraobserver and interobserver variability of the contrast studies but failed to do so. This is a key concern because you have knowingly assumed that the contrast studies were correct, which they may not be. The problem is made even worse because some of the studies were cut film and others were digital. Because we do not know the variability in the reading of these films, we are left with a dilemma as to what the variability of these readings is. We know that in the case of the carotid artery and peripheral arteries, the variability in angiographic interpretation is quite large. (2) For reasons that I cannot understand, you chose a 70% diameter cutoff point as the only meaningful end point. You justify this by saying, you believe that most patients with symptomatic chronic intestinal angina will have visceral artery stenoses that exceed 70%. I am very concerned by this statement because it will undoubtedly lead some people to believe it is always true. You provide no objective data to confirm this impression. In the discussion, another statement appears that is not true and should not be construed as such. It is, "Patients with technically satisfactory duplex examinations and normal visceral duplex findings (SMA PSV < 275 cm/sec and CA PSV < 200 cm/sec) will rarely have a 70% or greater visceral artery stenosis (high specificity)." Again, everything is focused on this greater than 70% stenosis, which you apparently accept as a gold standard. This is potentially a very dangerous statement because there will undoubtedly be abnormal mesenteric arteries with cutoff points that do not reach the levels you consider significant. One must also ask what the range is of velocities for arteries that are completely normal? Finally, I cannot understand why you feel forced to arrive at the magic cutoff point of 70% for clinical significance. In every patient we have seen for chronic

mesenteric ischemia, the lesions in the SMA and CA have been severe, and in some cases the arteries have been occluded. The documentation in such cases has been simple, because they have far exceeded the limits you set. Could you please tell me if you have made any therapeutic decisions based on your cutoff points as outlined in this article? I seriously doubt it. Finally, one must not lose sight of the fact that both the celiac and superior mesenteric arteries can be totally occluded and the patient remain symptom free. This is possible because of the collateral potential that may be available through the hypogastric and inferior mesenteric arteries. Dr. Gregory L. Moneta. Concerning the angiograms, it is true that we did not assess the variability of interpretation of stenoses. That was not the point of our study. There is potential variability in angiographic interpretation, but as we will show in our article on carotid stenosis, there really is not much in predicting a specific level of stenosis when the question is yes or no. In my opinion, a high-quality digital subtraction angiogram is just as good as a conventional film for assessing stenosis. It is clearly not as good for assessing ulcerations, but the focus of this study was stenosis. Why did we use 70% as a cutoff?. First, you have to chose something. As Dr. Strandness is aware, flow in a low-resistance arterial system, which both the celiac artery and the postprandial superior mesenteric artery are, does not reduce markedly until levels ofstenosis are considerably beyond 50%. As Dr. Strandness may not be aware, the Oregon group has one of the largest series of mesenteric revascularizafions. Very few, if any, bypasses have been done for less than 70% lesions. We certainly do not mean to imply that 70% lesions are always associated with chronic intestinal ischemia. We made that point in our previous retrospective study. We made that point in the manuscript that we sent to Dr. Strandness, and we made that point in the presentation. We think this is enough. Regarding velocity ranges in normal arteries, we published the velocity ranges in normal arteries in our previous retrospective article in the JOURNAL OF VASCUEAR SURGERY last fall. I have also published the velocity ranges in normal arteries in another group of patients when I was in Seattle, an article of which Dr. Strandness is the senior author. I did not think that we needed to repeat them in this presentation but they will be included in the final manuscript. We do make clinical decisions based on mesenteric duplex scans. The decision is whether to proceed with angiography. Such decisions however, are not made in a vacuum. The patient's history and clinical course are also factored into this decision. Clearly a patient with a soft history for mesenteric ischemia and a negative duplex study is considered less actively for angiography than a patient with convincing symptoms and a positive duplex study. Dr. William R. Flinn (Chicago, Ill.). In your study,

86 21/Ioneta et al.

you have identified between 13% and 24% of patients with significant occlusive lesions of the mesenteric vessels in i00 angiograms. Most of us would consider this a very high prevalence for the occurrence ofmesenteric artery disease in a random sample of arteriograms, even in a population selected for peripheral atherosclerosis. These observations are, of course, skewed somewhat because i3 patients were evaluated for symptoms suggestive of chronic intestinal ischemia. Were the severe CA-SMA lesions identified found in the patients with symptoms or were there indeed a number of symptom-free patients in whom unsuspected lesions were identified? If a significant number of these latter lesions were seen, what were their recommendations regarding the follow-up and treatment of patients with asymptomatic severe mesenteric artery occlusions? I believe the natural history studies of severe renal artery stenosis have indicated that such lesions may indeed not be benign and may threaten renal parenchymal function. With more specific duplex scan criteria for the identification of asymptomatic mesenteric occlusive disease, should we begin to change our therapeutic approaches to some of these lesions? Dr. Timothy H a r w a r d (Gainesville, Fla.). I continue to be confused by the 70% stenosis cutoff criteria. Dr. Strandness has sufficiently brought that issue out and I will not pursue it further. However, one must remember that noninvasive testing, with Doppler criteria used to predict the degree of anatomic luminal narrowing, is done by evaluating the changes of flow hemodynamics. These tests do not provide any data concerning the clinical significance of arterial stenoses found with duplex scanning. Several investigators are using peak systolic parameters to estimate arterial disease. They have used regression

Journal of VASCULAR SURGERY

analysis to generate linear criteria for objectively predicting the progression of increasing degrees of arterial narrowing. With this thought in mind, I would like to ask at least one question. You did look at linear regression analysis of the data and generate a relationship between velocity data and percent diameter reduction. With this fine, one should be able to predict the degree of luminal narrowing rather than just providing a yes or no answer regarding the presence or absence of a particular percent stenosis. Have you attempted to do linear regression analysis on the velocity data you presented today and then to follow up the patients according to these objective criteria in an effort to delineate the natural history of their disease (i.e., is the disease progressing and if the diseased segment is getting worse, can you correlate the progression with clinical criteria). Dr. Moneta. In response to Dr. Flinn, there certainly were lesions in patients with symptoms. Although I3 patients were evaluated for chronic intestinal ischemia, only six were operated on for that disease in this series. These patients were evaluated in a blinded fashion just like those who were symptom free. I do not believe their presence in the study significantly skewed the data. In response to both Dr. Flinn and Dr. Harward, I do not believe that we know the natural history of mesenteric vascular disease and what its implications are over time. We have not looked at the ability to follow a lesion and predict subsequent worsening of that lesion on the basis of our duplex criteria. In fact, the distribution of peak systolic velocities in the patients who had greater than 70% lesions was really quite wide. We thus far do not have enough patients with serial angiograms and duplex studies to identify whether increasing peak systolic velocities is synonymous with angiographic progression and worsening clinical symptoms.