Journal of Vascular Surgery
Volume 50, Issue 5 , Pages 1012-1018, November 2009

Endoleak after endovascular aneurysm repair: Duplex ultrasound imaging is better than computed tomography at determining the need for intervention

Presented at Thirty-third Annual Meeting of Southern Association of Vascular Surgery, Tucson, Ariz, Jan 14-17, 2009.

Division of Vascular Surgery, Eastern Virginia Medical School, Norfolk Va

Received 14 April 2009; accepted 8 June 2009.

Article Outline

Objective

Color duplex ultrasound (CDU) imaging is a noninvasive alternative to computed tomography (CT) for the detection of endoleak. This study compared CT and CDU imaging in the detection of endoleaks requiring intervention after endovascular aneurysm repair (EVAR).

Methods

All EVARs performed at our institution from 1996 to 2007 were retrospectively reviewed. CDU and CT scans ≤3 months were paired and the presence of an endoleak and its type were recorded. Clinical follow-up was reviewed and interventions for endoleak were recorded. Interventions were performed for type I, for type II with sac enlargement, and for type III endoleaks. The first analysis of clinical test outcomes used the findings of CT scan as a gold standard and the second used the findings at time of intervention as a gold standard.

Results

During the time period reviewed, 496 patients underwent EVAR, and 236 of these had CDU and CT follow-up studies paired ≤3 months of each other. Mean follow-up was 17 months (range, <1-111 months). We reviewed 944 studies or 472 pairs. Eighteen patients (7.6%) required intervention for 19 endoleaks: six type I, 11 type II, and two type III. Early endoleak (≤1 month) requiring reintervention was detected in 1 vs late endoleak (mean, 28 months; range, 0.6-88 months) in 18. All type I and III endoleaks were treated with endovascular cuff or limb extension placement. Three type II endoleaks were treated with open ligation, and coil or glue embolization was used in eight. CDU imaging detected endoleaks requiring intervention in 89% of cases, whereas CT detected endoleak in 58% (P < .05). The ability to correctly identify the type of endoleak as confirmed at time of intervention was 74% with CDU imaging vs 42% by CT (P < .05). CDU, for the detection of endoleak requiring intervention, had a sensitivity of 90%, specificity of 81%, negative predictive value (NPV) of 99%, and positive predictive value (PPV) of 16%, while CT had a sensitivity of 58%, specificity of 87%, NPV of 98%, and PPV of 15%.

Conclusions

CDU imaging has a high sensitivity in detecting endoleaks requiring intervention, is better at identifying the type of endoleak, and is an excellent test for graft surveillance after endovascular aneurysm repair. Compared with CT scan, CDU imaging in our experience is the preferred test on which to base an intervention for endoleak.

 

Endovascular aneurysm repair (EVAR) has become the main treatment of abdominal aortic aneurysms (AAA) during the last decade. Although EVAR has shown decreased morbidity and mortality compared with open AAA repairs, the procedure has also been associated with different rates and types of postoperative complications such as endoleaks.1, 2, 3 These complications can result either in reintervention or in the need for closer surveillance. Hence, long-term surveillance after EVAR has become paramount to identify these complications and allow for secondary interventions to maintain an optimal outcome after EVAR.

A contrast-enhanced computed tomography (CT) scan has been the gold standard for postoperative surveillance of EVAR. CT surveillance, however, is associated with the cumulative risks of contrast nephrotoxicity, radiation exposure, and cost.4 Recent studies have shown an increased risk of radiation-induced cancer after repeated exposures from CT scans.5, 6, 7, 8 On the other hand, surveillance with color duplex ultrasound (CDU) imaging does not have either of these risks. Although CDU imaging has been used for preoperative AAA surveillance, its role with EVAR surveillance is still debated.9 There have been conflicting studies on its efficacy compared with CT scan.10, 11, 12 The ultimate success of a screening test is to accurately predict the disease process and guide appropriate interventions.

We reviewed all articles evaluating CDU surveillance for endografts using CT scan as a gold standard. We performed a retrospective review of our experience with CDU and CT surveillance to detect endoleaks and whether this correctly correlated with the actual findings at the time of reintervention after EVAR. We compared both imaging studies to a common measure, the endoleak detected at the time of intervention.

Back to Article Outline

Patients and methods 

A retrospective review was performed on patients with AAAs who underwent elective treatment with EVAR from July 11, 1996, through March 31, 2007. Only patients with paired imaging studies (CDU and CT) ≤3 months of each other were included. The analysis excluded patients with symptomatic or ruptured AAA and isolated iliac aneurysms. Patient demographics, clinic notes, preoperative CT/CDU scans, operative reports, types of EVAR devices, CT and CDU surveillance studies were reviewed after approval by the Institutional Review Board.

CDU or CT examinations within a week of each other were scheduled at 1, 3, 6, and 12 months, and annually thereafter. We adhered to our CDU imaging protocol previously reported by Sato et al.12 After an overnight fast, patients were placed supine and studied with a low frequency (range, 2.5-5 MHz) curved array, phased array, or mechanical sector, and pulsed Doppler scan transducer. The endograft, proximal and distal fixation points, and the AAA sac were imaged in B-mode. The size of the AAA sac was measured.

The CDU scan was performed in sagittal and transverse views to evaluate the AAA sac for the presence of flow outside the graft. Perigraft leaks were identified when reproducible, pulsatile color Doppler scan flow images were present outside the graft and within the AAA sac. Focus was directed at the superior and inferior stent attachment, the anterior mid-AAA sac (inferior mesenteric artery), and the posterior mid-AAA sac (lumbar arteries). CDU scan evidence of an endoleak required the identification of perigraft Doppler scan signals with color flow and was confirmed with spectral analysis and mapping of blood flow pattern (Fig).

All CDU examinations occurred in a peripheral vascular laboratory accredited by the Intersocietal Commission for the Accreditation of Vascular Laboratories and were performed by vascular technicians with RVTS certification after being proctored by senior technicians. The results were read by vascular surgeons.

CT scan surveillance was performed using a GE Lightspeed plus 16-slice scanner (GE Healthcare, Waukesha, Wis) using 2.5-mm acquisition slice. Omnipaque 350 contrast (120 mL; GE Healthcare) was injected at a rate of 4 to 5 mL/s using SmartPrep software (GE Healthcare). The arterial phase acquisition was obtained by an average delay of 25 seconds after injection. The delayed phase was obtained at 70 seconds after completion of first scan. CT scan reconstruction used a 0.625-mm format.

Interpretation of CT scan results was performed by radiology, whereas vascular surgery interpreted CDU results. Vascular surgeons made clinical decisions by reviewing both imaging modalities and the patient's clinical findings. Endoleaks were categorized as type I, type II, type III or indeterminate. The CDU examination was considered inadequate if the endograft graft was poorly or incompletely seen secondary to patient habitus or obscured by bowel gas. In this study, only adequate CDU studies were examined. Studies were adequate if they visualized (1) the AAA residual sac, (2) the proximal and distal fixation points, and (3) the endograft and bilateral limbs. An internal review of CDU for EVAR surveillance revealed a 1% rate of inadequate studies.

The types of reinterventions, findings at reintervention, and the paired imaging studies obtained ≤3 months before reintervention were evaluated. The type of endoleak reported at the time of intervention was compared with the predicted type of endoleak by imaging studies. The type of endoleak found at the time of intervention was determined by angiography or open surgical findings. Early endoleaks and reinterventions were defined by detection or intervention ≤30 days of EVAR deployment. Late endoleaks and reinterventions were defined by detection or intervention after 30 days of EVAR deployment.

Commercially available and investigational devices were used during the study period. They included 160 AneuRx (Medtronic, Santa Rosa, Calif), 55 Ancure/EVT (Guidant, Indianapolis, Ind), 13 Zenith (Cook, Indianapolis, Ind), 5 Powerlink (Endologix, Irvine, Calif), 2 Excluder (W. L. Gore & Associates, Flagstaff, Ariz), and 1 Quantum (Cordis, New Brunswick, NJ).

Data were analyzed using XLStat software (Addinsoft USA, New York, NY). Continuous data are expressed as mean with standard deviation and were compared using the t test. Noncontinuous data are expressed as percentages and were compared by using the z test comparison for proportions. P < .05 was considered statistically significant. CDU/CT surveillance analysis was performed in two segments. First, CDU/CT were analyzed in a two-by-two grid for sensitivity, specificity, negative predictive value (NPV), and positive predictive value (PPV). CT scan was used as a gold standard. The second analysis (reintervention analysis) used the type of endoleak detected at the time of reintervention as the gold standard for both CDU imaging and CT.

Back to Article Outline

Results 

Surveillance analysis 

From July 11, 1996, through March 31, 2007, 496 consecutive patients underwent an EVAR procedure. Paired CDU and CT imaging studies were obtained in 236 patients, of which 202 (86%) were men. The mean age at the time EVAR was 72 years (range, 51-90 years). The study population compromised 211 whites (89%), 20 African Americans (8%), 4 Asians (2%), and 1 Hispanic.

The analysis reviewed 472 paired imaging studies or 944 studies. The mean interval between CDU and CT scans was 18 days (range, 0-90 days), and 33% of paired studies were performed ≤4 days of each other. A CT scan was obtained before CDU scan 69% of the time (n = 325), a CDU study was obtained before the CT scan 15% of the time (n = 71), and both studies were obtained on the same day 16% of the time (n = 76). There were no statistically significant differences in these intervals in patients with endoleaks vs those without endoleaks, nor in patients requiring an intervention vs those without an intervention. The mean follow-up time was 17 months (ranged, <1-111 months).

Overall, 137 endoleaks (29% of paired studies) were detected by CT or CDU scan. There were 8 type I endoleaks, 123 type II endoleaks, 4 type III endoleaks, and 2 indeterminate endoleaks (one detected by CDU only and the other detected by CT only). During the surveillance, 91 patients (39%) were found to have an endoleak, comprising 25 endoleaks ≤30 days (18%) compared with 112 late endoleaks (82%). CDU imaging detected 110 endoleaks compared with 75 by CT, which was significant (P < .01). There was moderate agreement between the two tests for detection of endoleak by association analysis (κ = 0.42).

The CDU sensitivity, specificity, PPV, and NPV for overall endoleak detection were 64%, 84%, 44%, and 93% (Table I). CDU imaging showed a trend toward better detection of type I endoleaks compared with CT scan (CDU, 7 of 8 vs CT, 4 of 8; P = .10). CDU and CT performed equally well for type III endoleak detection (CDU, 3 of 4 vs CT, 4 of 4; P = .29). The CDU sensitivity, specificity, PPV, and NPV for type I endoleaks were 75%, 99%, 43%, respectively, and for type III endoleaks, 99% and 75%, 100%, 100%, and 99%. CDU performed better than CT at detection of type II endoleaks (CDU, 99 of 123 vs CT, 66 of 123; P < .05). CDU sensitivity, specificity, PPV, and NPV for type II endoleaks were 64%, 86%, 42%, and 94%, respectively.

Table I. Overall endoleak detection of color duplex ultrasound vs computed tomography scana
Color duplex ultrasound
PositiveNegativeTotals
CT scan
Positive482775
Negative62335397
Totals110362472

CT, computed tomography.

aSensitivity, 64%; specificity, 84%; positive predictive value, 44%; negative predictive value; 93%.

Reintervention analysis 

To better compare the clinical utility of these two modalities, the endoleak confirmed at time of reintervention was used as the gold standard to compare CDU imaging and CT. Eighteen patients (7.6%) required interventions for 19 endoleaks, comprising six type I, 11 type II, and two type III. One early (≤1 month) endoleak reintervention occurred vs 18 late (mean, 28 months; range, 0.6-88 months) endoleak reinterventions. All type I and III endoleaks were treated with endovascular cuff or limb extension placement. Type II endoleaks underwent intervention if there was aneurysmal sac enlargement or large aneurysmal sac with failure to regress. Three type II endoleaks were treated with open ligation, whereas coil or glue embolization was used in eight.

CDU detected an endoleak requiring intervention in 89% of cases, whereas CT detected endoleak in only 58% (P < .05). The ability to correctly identify the type of endoleak as confirmed at time of intervention was 74% with CDU compared with 42% by CT (P < .05). CDU missed one type I and one type II endoleaks. CDU also incorrectly detected type II endoleaks for one actual type I and two type III endoleaks. CT scan missed two type I and six type II endoleaks and incorrectly detected type II endoleaks for one actual type I endoleak and two type III endoleaks. For the detection of endoleak requiring intervention, CDU imaging had a sensitivity of 90%, specificity of 81%, NPV of 99%, and PPV 16%, whereas CT had a sensitivity of 58%, specificity of 87%, NPV of 98%, and PPV of 15% (Table II, A and B).

Table II. Overall endoleak detection at time of reintervention for (A) color duplex ultrasound imaging and (B) computed tomography scan
InterventionTotals
YesNo
A
CDUa
Positive1788105
Negative2365367
Totals19453472
B
CT scanb
Positive116172
Negative8392400
Totals19453472

CDU, Color duplex ultrasonography; CT, computed tomography.

aSensitivity, 90%; specificity, 81%; positive predictive value, 16%; negative predictive value, 99%.

bSensitivity, 58%; specificity, 87%; positive predictive value, 15%; negative predictive value, 98%.

Back to Article Outline

Discussion 

The goal of surveillance or screening tests is to not miss a diagnosis or complication; that is, have a high sensitivity and NPV. The crux of this assessment lies in the chosen gold standard. For AAA screening and surveillance, CDU imaging has been the modality of choice.9 For EVAR surveillance, a contrast CT scan as a comparison standard has been an arbitrary choice. Studies comparing CDU with CT scan for endoleak detection have produced mixed results. Sato et al12 and d'Audiffret et al13 both showed that CDU is an excellent screening test for endoleaks, with sensitivities of 97% and 96%, respectively. Elkouri et al,10 however, found poor results for CDU endoleak detection, with a sensitivity of 25% and specificity of 89%. Most studies comparing CDU with CT for endoleak detection, however, show modest sensitivities of 52% to 81% and good NPV of 86% to 95%.14, 15, 16, 17, 18 This study also shows a modest sensitivity of 64%, but an excellent NPV of 93%. Our results almost mirror the meta-analysis results of Sun,19 which showed an overall sensitivity, specificity, PPV, and NPV value for endoleaks of 66%, 93%, 76%, and 90%, respectively.

Although CDU sensitivity for overall endoleak detection appears adequate at best, the decrease in sensitivity occurred for type II endoleaks. CDU sensitivity for type I and III endoleak detection is excellent. AbuRahma et al18 found a similar pattern of worse outcomes for type II endoleak detection vs type I. This decreased sensitivity can be partially attributed to the statistics using CT scan as the gold standard. If CDU were used as the gold standard, then CT would have a sensitivity of 44%. This study also showed that CDU detected more type II endoleaks than CT by a ratio of 1.5:1. Other studies have confirmed the ability of CDU to better detect type II endoleaks than CT, with ratios varying up to 2:1.11, 12, 13 Some studies have increased detection of type II endoleaks with ultrasound contrast.20, 21 Although contrast-enhanced CDU imaging may aid in decreasing missed type II endoleaks, it increases the amount of false-positive results, may be too sensitive, and increases the cost of the study, the time of the study, and the invasiveness of the study because patients require an intravenous catheter. Finally, these studies have been in small cohorts, and this experience has not been used more broadly to truly assess its efficacy.

Overall, it appears that CDU provides an excellent ability to detect type I and type III endoleaks, which are the most worrisome endoleaks that require reintervention. Our results also show that CT scans fail to show 50% of type I endoleaks. We believe this results from confusion of distal type I endoleaks for type II endoleaks with a static image from a CT scan, whereas CDU can provide a more dynamic picture illustrating the higher velocity flow associated with type I endoleaks. On the other hand, several studies have shown that both imaging modalities may miss type II endoleaks. Hence, neither CT nor CDU imaging will detect all type II endoleaks, nor do they need to detect all endoleaks. Fortunately, the fate of type II endoleaks usually follows a benign and often sporadic course. In addition, this fate is intertwined with aneurysmal sac size, which may ultimately serve as the main marker for EVAR surveillance. Therefore, type II endoleak detection may be less important than aneurysmal sac size.

The reintervention analysis is an attempt to compare CDU with CT by using a shared gold standard of endoleak confirmation at the time of reintervention. These results clearly show that CDU imaging not only performs better for endoleak detection than CT but also for identification of the type of endoleak. This analysis, however, was performed on a small subset of patients.

Which test is better? Neither test is or will be the perfect screening test for EVAR surveillance. CDU imaging does have limitations:

•It is operator-dependent, which may lead to disparate findings between the different technicians.

•Suboptimal examinations due to body habitus or bowel gas can compromise results. An internal review of our CDU examinations during 1 year showed about 1% of studies were inadequate; however, this will vary according to experience and patient population.

•Availability and time commitment restricts its broader use; whereas CT scanners are ubiquitous and fast.

•The ability of CDU to detect endograft migration is currently unknown.

The limitations of CDU must be weighed with its benefits. CDU offers no exposure to radiation, intravenous contrast, and cheaper cost. On the other hand, CT scan surveillance is more expensive and exposes patients to cumulative risks of radiation and intravenous contrast. Recent studies have raised awareness of the increased risk of cancer after repeated exposures from CT scans.5, 6, 7, 8 We have obtained both CT scans and CDU imaging for EVAR surveillance for several years. This provided a way to validate the use of CDU. Over time this practice has evolved, with some patients getting only CT scan, others CDU only, and still others obtaining both.

Our study has several limitations. First, its retrospective design contributed to incomplete data collection for some preprocedural and postprocedural variables. A retrospective study can also introduce selection bias into patient subsets. Our use of CDU, however, has been extensive and has never been used selectively.

Second, there was an average of 18 days between the CDU and CT scans. Although this time frame between tests is relatively small, the behavior of endoleaks, recurring and sealing, can be variable. This could lead to overestimation or underestimation of endoleak detection by each modality. Ideally, all scans would be perfectly paired each day and every patient would follow our protocol. In clinical practice, however, scheduling conflicts arise that limit the ability to obtain both imaging studies in the same day. This is retrospective study in which own surveillance protocol was followed as best as possible.

Third, a wide variety of endografts were used during the study. The variability of performance between the different endografts may have confounded the outcomes. This, however, reflects a true clinical practice. And the ability of CDU imaging to monitor a variety of endografts despite their different characteristics reaffirms its utility in EVAR surveillance.

Back to Article Outline

Conclusion 

Duplex ultrasound imaging has a high sensitivity and negative predictive value in detecting endoleaks requiring intervention, is better at identifying the type of endoleak, and is an excellent test for graft surveillance after endovascular aneurysm repair. Compared with CT scan, our experience shows DU is the preferred test on which to base an intervention for endoleak.

Back to Article Outline

Author contributions 


Conception and design: GCS, JP

Analysis and interpretation: GCS, JP

Data collection: GCS, CS

Writing the article: GCS

Critical revision of the article: GCS, CS, GKS, FP, JP

Final approval of the article: GCS, GKS, FP, JP

Statistical analysis: GCS, CS

Obtained funding: Not applicable

Overall responsibility: GS, JP

Back to Article Outline

References 

  1. EVAR trial participants. Endovascular aneurysm repair versus open repair in patients with abdominal aortic aneurysm (EVAR trial 1). Lancet. 2005;365:2179–2186
  2. Blankensteijn JD, de Jong SE, Prinssen M, van der Ham AC, Buth J, von Stertenburg SMM, et al. Two-year outcomes after conventional or endovascular repair of abdominal aortic aneurysm. N Engl J Med. 2005;352:2398–2405
  3. Jones JE, Atkins MD, Brewster DC, Chung TK, Kwolek CJ, LaMuraglia , et al. Persistent type 2 endoleak after endovascular repair of abdominal aneurysm is associated with adverse late outcomes. J Vasc Surg. 2007;46:1–8
  4. Prinssen M, Wixon CL, Buskens E, Blankensteijn JD. Surveillance after endovascular aneurysm repair: diagnostics, complications, and associated costs. Ann Vasc Surg. 2004;18:421–427
  5. Brenner DJ, Elliston CD. Estimated radiation risks potentially associated with full-body CT screening. Radiology. 2004;232:735–738
  6. Brenner DJ, Hall EJ. Computed tomography–an increasing source of radiation exposure. N Engl J Med. 2007;357:2277–2284
  7. De Jong PA, Mayod JR, Golmohammadi K, Nakano Y, Leqin MH, Tiddens H, et al. Estimation of cancer mortality associated with repetitive computed tomography scanning. Am J Respir Crit Care Med. 2006;173:199–203
  8. Einstein AJ, Henzlova MJ, Rajagopalan S. Estimating risk of cancer associated with radiation exposure from 64-slice computed tomography coronary angiography. JAMA. 2007;298:317–323
  9. Thomas PR, Shaw JC, Ashton HA, Kay DN, Scott RA. Accuracy of ultrasound in a screening programme for abdominal aortic aneurysm. J Med Screen. 1994;1:3–6
  10. Elkouri S, Panneton JM, Andrews JC, Lewis BD, McKusick MA, Noel AA, et al. Computed tomography and ultrasound in follow-up of patients after endovascular repair of abdominal aortic aneurysm. Ann Vasc Surg. 2004;18:271–279
  11. Parent FN, Meier GH, Godziachvili V, LeSar CJ, Parker FM, Carter KA, et al. The incidence and natural history of type I and type II endoleak: a 5-year follow-up assessment with color duplex ultrasound scan. J Vasc Surg. 2002;35:474–481
  12. Sato DT, Goff CD, Gregory RT, Robinson KD, Carter KA, Herts BR, et al. Endoleak after aortic stent graft repair: Diagnosis by color duplex ultrasound scan versus computed tomography scan. J Vasc Surg. 1998;28:657–663
  13. d'Audiffret A, Desgranges P, Kobeiter DH, Becquemin JP. Follow-up evaluation of endoluminally treated abdominal aortic aneurysms with duplex ultrasonography: Validation with computed tomography. J Vasc Surg. 2001;33:42–50
  14. Wolf YG, Johnson BI, Hill BB, Rubin GD, Fogarty TJ, Zarins CK. Duplex ultrasound scanning versus computed tomographic angiography for postoperative evaluation of endovascular abdominal aortic aneurysm repair. J Vasc Surg. 2000;32:1142–1148
  15. Collins JT, Boros MJ, Combs K. Ultrasound surveillance of endovascular aneurysm repair: A safe modality versus computed tomography. Ann Vasc Surg. 2007;21:671–675
  16. Golzarian J, Murgo S, Dussaussois L. Evaluation of abdominal aortic aneurysm after endoluminal treatment: comparison of color duplex sonography with biphasic helical CT. Am J Roentgenol. 2002;178:623–628
  17. Raman KG, Missig-Carol N, Richardson T. Color-flow duplex ultrasound scan versus computed tomographic scan in the surveillance of endovascular aneurysm repair. J Vasc Surg. 2003;38:645–651
  18. AbuRahma AF, Welch CA, Mullins BB, Dyer B. Computed tomography versus color duplex ultrasound for surveillance of abdominal aortic stent-grafts. J Endovasc Ther. 2005;12:568–573
  19. Sun Z. Diagnostic value of color duplex ultrasonography in the follow-up of endovascular repair of abdominal aortic aneurysm. J Vasc Interv Radiol. 2006;17:759–764
  20. Henao EA, Hodge MD, Felkai DD, McCollum CH, Noon GP, Lin PH, et al. Contrast-enhanced duplex surveillance after endovascular abdominal aortic aneurysm repair: improved efficacy using a continuous infusion technique. J Vasc Surg. 2006;43:259–264
  21. Giannoni MF, Fanelli F, Citone M, Acconcia MC, Speziale F, Gosetti B. Contrast ultrasound imaging: the best method to detect type II endoleak during endovascular aneurysm repair follow-up. Interactive Cardiovasc Thoracic Surg. 2007;6:359–362

 Competition of interest: none.

PII: S0741-5214(09)01331-7

doi:10.1016/j.jvs.2009.06.021

Journal of Vascular Surgery
Volume 50, Issue 5 , Pages 1012-1018, November 2009

Article Tools

* Image Usage & Resolution