Duplex scan characteristics of bypass grafts to mesenteric arteries

Patients and methods 

All of the operative mesenteric (celiac, superior mesenteric, inferior mesenteric) arterial revascularization procedures performed from January 1995 to December 2005 were identified from a vascular surgery registry at Oregon Health & Science University and were retrospectively reviewed. Hospital medical records and vascular laboratory data were recorded into a database. This information included the symptoms at presentation, patient age, gender, medical comorbidities, graft patency, and freedom from recurrent symptoms. Medical comorbidities assessed included smoking status and a history of peripheral arterial disease (PAD), coronary artery disease (CAD), hypertension, cerebrovascular disease (CVD), renal failure/insufficiency, hypercholesterolemia, diabetes mellitus, and chronic obstructive pulmonary disease (COPD). Patient mortality was derived from the medical records and the Social Security Death Index.

Operative data included the configuration of the bypass graft (antegrade from the supraceliac aorta, retrograde from the infrarenal aorta or from an iliac artery), single vs bifurcated graft, prosthetic vs great saphenous vein graft, and prosthetic graft diameter (6 mm vs 7 mm). Patients were excluded from the analysis if a graft was not used (eg, SMA thrombectomy, SMA transposition, transaortic endarterectomy) or if a postoperative graft surveillance was not performed. Patients who underwent aortic debranching before stent graft placement for thoracoabdominal aneurysms were also excluded.

Vascular laboratory data included graft patency or occlusion. If a graft was patent, the peak systolic velocities (PSV) were routinely measured at the inflow artery, proximal anastomosis, mid graft, distal anastomosis, and distal recipient artery. The vascular laboratory protocol was to perform the surveillance duplex after the patient had been fasting for at least 4 hours before the study. If a mesenteric graft surveillance was not technically feasible, the study was repeated at a later date, with the patient in reverse Trendelenburg position or with the administration of oral simethicone. Graft events (defined as either occlusion or need for revision of the mesenteric bypass) were reviewed to determine if any duplex scan findings were premonitory of a bypass graft occlusion.

Various graft configurations were compared, including antegrade vs retrograde bypasses, single vs bifurcated, graft type (6 mm prosthetic vs 7 mm prosthetic vs vein), and distal target vessel (celiac/hepatic vs SMA). To determine whether mesenteric graft velocities changed over time, we identified all patients who had multiple surveillance ultrasound scans, with the last duplex scan occurring at least 1year after the initial index scan.

The data were collected into an Access database (Microsoft, Redmond, Wash), and analyzed using statistical software SPSS 14 (SPSS Inc, Chicago, Ill). PSV measurements were expressed as the average ± standard deviation. The effects of various graft configurations on the PSV at the inflow, proximal anastomosis, mid graft, distal anastomosis, and outflow artery were compared using one-way analysis of variance (ANOVA). In patients who had multiple surveillance ultrasound scans at least 1 year apart, graft velocity changes over time were assessed with one-way ANOVA. A value of P < .05 was accepted as significant.

Of all the velocities measured with mesenteric graft surveillance, we postulated that the mid-graft PSV was the velocity (dependent variable) most likely to be affected by demographic factors and graft configuration. These demographic factors and graft characteristics also were assessed for their effects on mid-graft PSV with univariate and multivariate linear regression analysis. Independent sample t tests were used to assess binary variables, with one-way ANOVA used to assess variables with more than two categories. Variables with a univariate P < .2 were entered into a multivariate linear regression model with mean mid-graft PSV as the dependent variable. Long-term graft patency was determined with life-table analysis.5

Results 

During the study period, 86 mesenteric revascularization procedures were performed, including 77 bypass grafts, eight SMA embolectomies, and one SMA endarterectomy. Excluded from the study were 34 of the 77 bypass grafts that did not receive a postoperative graft flow examination. Fasting postoperative surveillance duplex scans were reviewed in 38 patients who underwent 43 mesenteric bypass procedures, and these form the basis of this study. We analyzed 167 duplex examinations, with a mean of 4.5 scans per patient (range, 1 to 14). Approximately 95 % of the surveillance scans were technically successful. The remaining patients were all successfully restudied after administering oral simethicone with the patient in reverse Trendelenburg position.

Procedures included single-vessel bypasses to the SMA in 28 patients, single-vessel bypasses to the celiac artery in 3, bifurcated grafts to the SMA and celiac artery in 5, and 2 patients received a bifurcated graft to the SMA and left renal artery. Ultrasound surveillance data were available for bypass grafts to 43 mesenteric arteries. Nine patients had mesenteric grafts originating from the supraceliac aorta, 10 from the infrarenal aorta, and 19 from a common iliac artery. The recipient arteries included eight grafts to the celiac or hepatic arteries and 35 grafts to the superior mesenteric artery. Of the bypass grafts to the superior mesenteric artery, 13 were constructed with 6-mm polyester, 13 with 7-mm polyester, and eight with autogenous great saphenous vein. The only patient with an 8-mm polyester mesenteric bypass was excluded from the graft diameter analysis. Thirty-one bypass grafts were single-vessel revascularizations, and 12 grafts were part of a bifurcated configuration.

Comparison of graft configurations 

Antegrade (supraceliac) and retrograde (infrarenal aortic/iliac artery) graft configurations were compared. The mean velocities are listed in Table I. The inflow artery PSV at the supraceliac aorta was significantly lower than the inflow velocity for the retrograde bypasses (infrarenal aorta or a common iliac artery; P < .05). The elevated native iliac artery velocity accounted for this difference. The average mid-graft PSVs were consistent at 147 ± 52 cm/s for the supraceliac inflow, 152 ± 71 cm/s for the infrarenal aortic inflow, and 154 ± 35 cm/s for the common iliac inflow (P = .99). No significant velocity differences were identified between single-vessel and bifurcated graft configurations (single-vessel mid-graft mean velocity, 150 ± 42 cm/s vs bifurcated mid graft, 155 ± 56 cm/s, P = NS). Nor were there any differences when comparing SMA outflow vs hepatic/celiac outflow: the mean SMA mid-graft PSV was 149 ± 42 cm/s vs an hepatic/celiac mid-graft PSV of 160 ± 78 cm/s (P = NS). The 6-mm and 7-mm diameter polyester grafts and the autogenous saphenous vein grafts (all terminating at the SMA) were also compared. The mean velocities are listed in Table II. ANOVA found an insignificant trend toward increased velocities for the 6-mm grafts compared with the 7-mm grafts (P > .05).

Table I. Mean peak systolic velocities at various locations along the mesenteric bypass graft, categorized by inflow source⁎

	Inflow source	N	Inflow artery	Proximal anastomosis	Mid graft	Distal anastomosis	Outflow artery
	Supraceliac aorta	9	97±70†	141±84	147±52	139±57	158±71
	Retrograde
	Infrarenal aorta	10	99±42	182±77	152±71	162±73	189±170
	Iliac artery	19	172±77	212±71	154±35	169±50	164±56

⁎ All velocities are expressed as centimeters per second ± standard deviation. † The antegrade inflow peak systolic velocity was significantly lower than the retrograde velocities (one-way analysis of variance, P < .05).

Table II. Mean peak systolic velocities at various locations along the superior mesenteric artery bypass graft, categorized by graft material and size⁎

	Graft size/material	N	Inflow artery	Proximal anastomosis	Mid graft	Distal anastomosis	Outflow artery
	Polyester
	6 mm	13	139±70	214±84	163±52	172±57	161±55
	7 mm	13	135±42	169±77	138±71	166±73	207±140
	Saphenous	8	132±77	181±71	160±35	144±50	124±42

⁎ All velocities are expressed as centimeters per second ± standard deviation. No significant difference between groups (one-way analysis of variance, P = NS).

Comparison of velocities over time 

Eighteen patients had multiple surveillance duplex scans at least 1 year after their initial postoperative scan (Table III). The mean time for the first postoperative duplex was 3.8 months, and the mean interval between the index scan and the latest scan was 38 months. No significant difference was found between the PSV measurements at the index scan vs the latest follow-up scan (ANOVA, P > .05).

Table III. The graft surveillance mean peak systolic flow at the initial postoperative duplex and at the last available examination for 18 patients⁎

	Scan (n = 18)	Inflow artery	Proximal anastomosis	Mid graft	Distal anastomosis	Outflow artery
	Index	134±69	178±81	146±30	185±46	144±42
	Follow-up	164±94	192±71	156±54	169±44	138±43

⁎ The mean interval between studies was 38 months. All velocities are expressed as centimeters per second ± standard deviation. No significant difference between groups (one-way analysis of variance, P = NS).

Linear regression for associations with mid-graft peak systolic velocity 

Univariate analysis was performed to determine if any of the demographic variables or graft configurations had any effect on the average mid-graft PSV (Table IV). Of the demographic variables tested, age, gender, smoking status, and a history of CAD, hypertension, CVD, renal failure/insufficiency, hypercholesterolemia, diabetes mellitus, and COPD were highly nonsignificant (P > .2). Peripheral arterial disease was the only demographic variable included in the multivariate model owing to a mild association with the average mid-graft velocity (P = .147).

Table IV. Univariate and multivariate linear regression analysis for associations between demographic and graft variables and mid graft peak systolic velocity

	Characteristics	Analysis univariate	Multivariate
	Demographic
	Gender	0.42
	Coronary disease	0.511
	COPD	0.424
	Renal failure/insufficiency	0.882
	Cerebrovascular disease	0.482
	Peripheral arterial disease	0.147	0.223
	Diabetes mellitus	0.646
	Hypertension	0.647
	Smoking	0.833
	Graft
	Antegrade vs retrograde	0.716
	Distal target	0.96
	Prosthetic vs vein	0.824
	Prosthetic (6 mm vs 7 mm)⁎	0.03	0.061
	Inflow artery	0.933
	Graft type	0.266

⁎ Smaller prosthetic graft size was significantly associated with higher peak systolic velocity on univariate analysis (P = .030), but with a trend toward significance with multivariate analysis (P = .061).

Univariate analysis of the graft characteristics identified that prosthetic graft diameter was the only significant predictor of the mid-graft PSV (P = .03), and it was entered into the multivariate model. Factors not found to be predictive of the mean mid-graft velocity in the univariate analysis included prosthetic vs vein grafts, proximal and distal anastomotic sites, and antegrade vs retrograde graft orientation. In the multivariate linear regression analysis, neither a history of PAD nor prosthetic graft size was found to be an independent predictor of average mid-graft velocity, although there was a trend toward significantly higher mid-graft PSVs in the 6-mm grafts compared with the 7-mm grafts (P = .061).

By life-table analysis, 84% of grafts were patent at 5 years. The patency curve is shown in the Fig 1. Three graft events (1 graft revision, 2 graft occlusions) occurred during the follow-up of the 43 mesenteric bypass grafts. One revision was performed to correct a severe stenosis at the proximal anastomosis of an iliac-to-SMA autogenous vein graft. The velocity at the proximal anastomosis exceeded 400 cm/s, and a severe stenosis was confirmed with angiography. This graft was transposed to the infrarenal aorta, and subsequent surveillance duplex scans showed the velocities had normalized.

View Large Image Download to PowerPoint

Fig. Life-table primary patency of the mesentric bypass grafts was 84% at 5 years.

In the two patients with mesenteric graft occlusion, the velocities in the preceding surveillance duplex scans demonstrated no premonitory findings that would suggest impending thrombosis. The first occluded graft (7-mm polyester) had a retrograde configuration (infrarenal aortic inflow) and became occluded at 42 months. The patient became symptomatic with recurrent postprandial abdominal pain and a 13-pound weight loss. She underwent a repeat bypass to the SMA. The second occluded graft (saphenous vein conduit) also had a retrograde configuration (common iliac inflow) and occluded 1 month after surgery. Worsening abdominal pain and ischemic strictures of the small intestine developed and the patient required a small-bowel resection and stent placement in the inferior mesenteric artery origin.

No postoperative duplex surveillance was done on 34 of the 77 mesenteric bypass procedures. Twenty-one patients died before receiving a mesenteric duplex scan, with four dying ≤30 days after the bypass procedure from multiorgan system failure. Three of the four patients presented with acute mesenteric ischemia, and one presented with chronic mesenteric symptoms. Thirteen patients did not return for follow-up. Many of these patients resided in more remote geographic locations, preferring to have follow-up at their referring institutions.

Discussion 

Mesenteric duplex scan criteria for the diagnosis of mesenteric artery stenosis are well established.6, 7, 8, 9 This modality has become a widely used tool, aiding in the evaluation of patients with a myriad of abdominal symptoms. At our institution, a celiac PSV of >200 cm/s and an SMA PSV of >275 cm/s are indicators of >70% stenoses. Surveillance duplex evaluations of mesenteric artery bypass grafts also are performed routinely. McMillan et al3 have demonstrated that clinical follow-up alone, assessing for recurrence of abdominal symptoms, has very limited sensitivity of as low as 33% in the detection of mesenteric graft occlusion. They suggest that postoperative duplex scanning can provide a safe objective measure of mesenteric bypass patency.3

The duplex scan characteristics have not, to our knowledge, been objectively assessed until now. Various institutions, including our own, have reported their techniques for performing mesenteric revascularization.1, 2, 3, 4, 10, 11, 12, 13 The choices include antegrade vs retrograde orientation, prosthetic graft vs saphenous vein vs femoral vein, single-vessel vs multivessel revascularization, and prosthetic graft diameter. Kansal et al14 compared antegrade and retrograde bypass configurations and identified a higher rate of minor complications and 30-day mortality in patients who underwent antegrade mesenteric artery bypass procedures. Objective patency determination using duplex scanning was only available in a small percentage of patients, however.

This current study evaluated duplex scan characteristics of various mesenteric graft configurations. When antegrade and retrograde bypass grafts were compared, the only difference in velocities occurred at the inflow artery (supraceliac aorta vs infrarenal aorta or iliac artery). The mean inflow velocities at the supraceliac aorta and infrarenal aorta were not dissimilar (97 cm/s and 99 cm/s, respectively); however, the infrarenal aorta and iliac artery inflow sources were treated as a group (both retrograde bypasses). The mean inflow velocities for the retrograde bypasses were significantly higher than for the antegrade bypasses, likely owing to the elevated iliac inflow velocities (172 cm/s). In addition, 19 bypass grafts had iliac inflow sources, equal to the number of supraceliac and infrarenal aortic grafts combined (Table I)

The inverse relationship between smaller diameter bypass grafts and higher mid-graft velocities seems intuitive, and some of our data support this. When proximal anastomotic, mid graft, and distal anastomotic velocities between 6-mm and 7-mm diameter grafts were compared, an insignificant increase was found with the 6-mm graft velocities. In our univariate analysis, the average mid-graft PSV was significantly higher with smaller diameter prosthetic grafts.

The multivariate analysis, however, found only a trend toward significance (P = 0.061). Our inability to demonstrate a statistically significant difference with the multivariate regression was likely a result of the number of patients available for analysis. Only 26 grafts were compared, 13 with 6 mm, and 13 with 7 mm diameter grafts. The effect of graft diameter could not be evaluated in the vein graft patients owing to the obvious biologic variability in the diameter of the saphenous vein.

Because this was a retrospective study, accurate measurements of the saphenous inner diameter were not available. We also could not identify any significant velocity differences between grafts to the celiac axis vs the SMA; however, our series included only eight celiac/hepatic artery bypasses, which limited our ability to identify a true difference.

We also demonstrated that mesenteric graft surveillance velocities generally remain stable over time. This does not equate to a recommendation against further surveillance of mesenteric bypass grafts. If graft velocities are expected to remain stable over time, then the finding of a significant velocity increase on subsequent studies should suggest the presence of a graft stenosis.

Our current practice is to perform mesenteric graft surveillance scans every 6 months. The mean PSV at the proximal anastomosis, mid graft, or distal anastomosis for most of the graft configurations in our study was 140 to 200 cm/s. As a result of these data, we have started to decrease the interval between follow-up scans, or obtain secondary imaging (computed tomography angiography or conventional angiography) if the PSV reaches ≥300 cm/s. We are also obtaining secondary imaging if the PSV decreases below approximately 50 cm/s.

Further validation will be needed to determine whether this is an appropriate threshold for more testing. Only three patients in our study received a conventional angiogram to confirm the presence of one graft stenosis and two graft occlusions. A limitation with our current study was the inability to correlate the surveillance duplex scans with another imaging modality.

The life-table primary graft patency in this series of patients was 84% at 5 years. This agrees well with other studies quoting mesenteric graft primary patency rates of 69% to 91%.4, 10, 11 Among the 38 patients in our study, seven deaths occurred during the 10-year study period. Only one death occurred ≤30 days after surgery, for a perioperative mortality of only 2.6%; however, prior analysis of our mesenteric bypass data indicated that the perioperative mortality was as high as 24% in patients with acute mesenteric ischemia.10 It is likely that some of these patients were excluded from the current study, many never having received a surveillance duplex; in fact, 34 of the 77 total mesenteric bypass procedures remained unstudied. Of these, 21 were because the patients died before receiving the surveillance duplex, and the rest were lost to follow-up. Many patients residing in remote geographic locations chose to have their postoperative follow-up closer to home.

A limitation with our current study may be the large percentage of single-vessel revascularizations in this series. We have previously reported that SMA revascularization alone appears to be an effective and durable procedure for the treatment of acute and chronic mesenteric ischemia.10, 15 In those studies, we reviewed the clinical results of 50 bypass grafts to the SMA alone. Most of the patients in our current study, 28 of 38, also had revascularization of the SMA alone. Our preference for performing single-vessel mesenteric revascularization may have made it more difficult to identify a true difference in the velocities between single-vessel and bifurcated grafts because there were far fewer bifurcated grafts.

This study was not able to identify any premonitory duplex findings that would predict imminent graft thrombosis. In essence, this series described the velocity characteristics of mesenteric bypass grafts that generally do well. Only three graft events occurred in the 43 mesenteric bypasses. Two graft thromboses occurred in patients whose prior mesenteric duplex scans seemed unremarkable. As with arterial reconstructions in other locations, these patients may have had other predisposing factors (eg, hypercoagulable condition) leading to graft occlusion. Symptoms consistent with recurrent mesenteric ischemia developed in both patients and resolved with mesenteric artery revascularization. An increased proximal anastomotic velocity consistent with a severe stenosis developed in the third patient. The stenosis was corrected and the mesenteric duplex velocities normalized afterward. The ability of the duplex scan to detect these velocity alterations supports the continued use of surveillance monitoring protocols in patients with mesenteric bypass grafts.

As a retrospective study, a surveillance follow-up for all of the 77 mesenteric bypass procedures was not available and 34 went unstudied. It is possible that a more complete follow-up may have uncovered other duplex characteristics that could be more predictive of imminent graft thrombosis.

Conclusions 

The results of this study indicate that postoperative duplex surveillance can be used to assess the patency of bypass grafts to mesenteric arteries. Although the inflow arterial PSV may be higher for retrograde bypasses, the anastomotic and mid-graft velocities are not significantly affected by the orientation of the graft. In addition, the PSV can be expected to remain relatively stable on repeat duplex examination.

Author contributions