Early Computed Tomographic Angiography after Endovascular Aneurysm Repair: Worthwhile or Worthless?

Patients and Methods

During a 9-year single-center study period from December 1996 to January 2006, 369 consecutive patients (350 men), with a mean age of 71 years (range 39–92 years), were treated by EVAR for an infrarenal AAA. During this period, two types of stent graft were used: the infrarenal fixating AneuRx stent graft and the transrenal fixating Talent stent graft (both from Medtronic AVE, Santa Rosa, CA). Indications for endovascular surgery were an AAA of at least 45 mm in diameter with suitable aneurysm morphology for endovascular repair. The study excluded patients who underwent endovascular repair for a ruptured AAA, a para-anastomotic aneurysm, or a solitary iliac aneurysm.

Before inclusion, all patients underwent CTA. From the start of the study until November 2000, a single-slice CT scanner was used with a section thickness of 3 mm (Philips Tomoscan SR 7000, Philips Medical Systems, Best, the Netherlands). After November 2000, a multidetector-row CT scanner was used with a section thickness of 1.5 mm (Philips IDT CT scanner, Philips Medical Systems). Nonionic intravenous contrast was used in all patients. The entire CTA database was interpreted by radiologists who frequently interpret those CTAs, with supervision of the EVARt by the three interventional radiologists.

Surgery was performed under general anesthesia by a team consisting of a vascular surgeon and an interventional radiologist. The details of surgery, including the amount of used contrast material, were noted in a prospectively gathered database. The key indices relating to these surgeries were the length of operation, amount of blood loss, technical and clinical success, and EVAR-related in-hospital complications.

Before hospital discharge, each patient underwent dual-phase CTA. After hospital discharge, each patient entered a regular follow-up program consisting of a clinical visit and radiographic surveillance with renewed dual-phase (early and delayed) CTA examinations at 3 and 12 months and annually thereafter. For the current study, all postprocedural and 3-month CTAs were reassessed by one of the interventional radiologists and one of the vascular surgeons to monitor the stent graft position, presence of migration, endoleaks, stent graft thrombosis, aneurysm diameter changes, and thromboembolic complications.

Stent graft thrombosis was defined as a complete thrombosis of the main body of the stent graft or one or both of its legs. Endoleak was classified into types I to IV, as described in 2002 by Veith and colleagues. ⁶ Stent graft migration was defined as a distal stent graft displacement of at least 5 mm compared with the stent graft position on the first postprocedural CTA.

AAA enlargement was defined as an aneurysm growth ≥ 5 mm and AAA shrinkage was defined as an aneurysm shrinkage of ≥ 5 mm; both were compared with the measurements on the former CT scan. The protocol in our institution for the treatment of endoleaks confirmed at the predischarge CT scans was as follows: (a) type I endoleaks: endovascular surgical reintervention with extension cuffs or aortouni-iliac devices was performed when the distance between the lowest renal artery and the proximal part of the stent graft was ≥ 5 mm at the predischarge CT scan, and no elective open conversion was done for type I endoleaks during the first 3 months postprocedure; (b) type II endoleaks: treatment (basically coil embolization) was given only in cases of symptomatic (painful) aneurysms; and (c) type III endoleaks: treatment was always with interposition cuffs.

The medical records of all patients were evaluated to assess EVAR-related complications. All secondary interventions were noted from discharge to a follow-up of 1 month after the 3-month postprocedural CTA. All secondary interventions were evaluated for any association with the findings on the early postprocedural and 3-month CTAs.

The presence of death was verified by checking the digital medical record of the patients and by calling the general practitioner of all treated patients at the end of the study period.

Analyses of data were performed using SPSS version 11.1 software (SPSS Inc, Chicago, IL). Quantitative data were expressed as mean ± standard deviation and range.

Results

During the study period, 369 patients underwent endovascular exclusion of an infrarenal AAA. Of these patients, 291 had both a postprocedural and 3-month follow-up CTA, all of which were included in the study. Most of the 78 excluded patients (n = 51) had their follow-up in other hospitals owing to logistic reasons. These patients had no 3-month follow-up but had 6-month CT scans. In the remaining 27 patients, the predischarge CT-scan was not performed owing to logistic reasons in our own institution. The baseline patient demographics of the 291 included patients are summarized in Table 1. The mean length of operations was 110 minutes (55–245 minutes), and the mean blood loss was 330 mL (80–1,100 mL). In two patients (0.6%), thrombectomy of the superficial femoral artery had to be done during the operation. In Table 2, the detected complications related to the stent graft are summarized for the postprocedural and 3-month CTAs.

The postprocedural CTAs detected 1 stent graft thrombosis and 93 endoleaks (32%), consisting of 7 proximal type I, 1 distal type I, 84 type II, and 1 type III endoleaks. These findings resulted in four secondary procedures (1%). Two type I endoleaks (one distally and one proximally) were treated with extension cuffs within 2 weeks after the initial operation. One interposition cuff was placed for a type III endoleak within 1 week after the initial operation. A patient with a large type II endoleak and an AAA > 10 cm underwent coiling of two paired lumbar arteries, which was not successful. Because of progressive abdominal pain, surgical conversion with explantation of the stent graft was done within 2 months after the initial operation. All secondary procedures were performed in an elective setting.

Six of eight type I endoleaks were managed conservatively. Five patients had a small endoleak with a small (< 55 mm) preoperative AAA diameter. One patient with a small distal type I endoleak, who already had an overstenting of the internal iliac artery on the contralateral side, was treated conservatively.

In the 3-month time interval between the postoperative and 3-month CTA, four acute reinterventions were required because of thromboembolic complications (three Fogarty procedures, one femorofemoral crossover bypass), and one groin infection was explored. These interventions were done irrespective of the findings on the postprocedural CTA. None of the four patients treated for thromboembolic complications had signs of stent graft kinking or obstructed iliac arteries on the predischarge CTA, indicating a higher risk for thromboembolic complications.

Eight patients died within 3 months after EVAR. None of the deaths was related to the AAA or EVAR (four cardiac, two pulmonary, one gastric bleeding, and one carcinoma). Four of eight patients (50%) underwent autopsy, which confirmed the clinical diagnosis. No AAA ruptures occurred in the first 3 months after EVAR.

On the 3-month CTAs, 43 endoleaks (3 proximal type I, 40 type II), 3 stent graft thromboses, and 1 proximal stent graft migration were seen. In two patients, a new type II endoleak was diagnosed compared with the postprocedural CTA.

During 3 and 12 months of follow-up, two of the proximal type I endoleaks were successively treated by a proximal extension cuff. One patient with a type I endoleak had AAA shrinkage compared with the postprocedural CTA and was treated conservatively. Of the 40 patients with a type II endoleak on the 3-month CTA, 37 had a stable AAA (n = 32) or AAA shrinkage (n = 5) and were under a watchful waiting policy. AAA growth was observed on the 3-month CTA in three patients with a type II endoleak. These type II endoleaks were treated successfully by coil embolization (n = 2) or thrombin injection (n = 1).

All patients who needed a secondary intervention were treated in an elective setting. No patients needed an urgent intervention within a 1-month time interval after the 3-month CTA.

Five of 291 included patients had an AAA growth within the first 3 months of follow-up (mean AAA diameter of 47 mm postprocedure to 58 mm at 3 months). Three of these patients had a type II endoleak and two had a type I endoleak on the postprocedural CTA.

No AAAs ruptured during the first 6 months of follow-up. Thirteen patients (5%) had a prolonged hospital stay for intravenous fluid administration to return creatinine clearance to preoperative levels.

Discussion

In the last decade, EVAR has become one of the leading treatment methods for exclusion of infrarenal AAAs. Despite its less invasive character, this technique has unique complications, as described in many reports that have focused on the results of EVAR. ^7–11 Because most of these complications will occur asymptomatically, regular post-EVAR radiographic follow-up is mandatory to detect and treat these complications. The optimal follow-up frequency and the best imaging modality are not fully defined to date. It must be taken into account that new-generation devices will lead to less device-related complications compared with the first-generation devices and less intensive radiologic follow-up might be justified in the near future. For instance, the use of the described AneuRx device has strongly declined.

The EUROSTAR reports advise examinations to start within the first month of follow-up and not during hospital stay, but the protocol in many centers still calls for a postprocedural CTA before hospital discharge. ^3–5

The CTA has long been accepted as the standard as it readily identifies a broad spectrum of postprocedural complications, including endoleak, migration, device fracture, aortic branch occlusion, delivery route dissection or injury, and retroperitoneal hematoma. With the introduction of dynamic magnetic resonance angiography and refined color duplex ultrasonography, the optimal imaging modality might become a matter of debate. ^12–14 However, neither of the competing techniques can assess the full spectrum of abnormalities that CTA is able to examine. In the current study, the early postprocedural follow-up scheme was the subject of interest, and the type of follow-up modality was beyond the scope of our study.

One of the reasons for performing an early postoperative CTA will be to establish the position of the device at the time of deployment to evaluate eventual future migration. However, of the 291 patients we studied, only 1 (0.3%) had proximal migration of > 5 mm in the first 3 months post-EVAR. All others had no or minimal migration of < 2 mm on the 3-month CTA; therefore, the baseline stent graft position can be reliably settled at the 3-month scan. Another important issue is the time interval between the preoperative CT examination and the EVAR procedure itself. Most of the preprocedural CTAs (74%) of the studied patients were performed within 8 weeks prior to the EVAR. Only 11% of the patients had a preprocedural CT scan older than 4 months. In these particular patients, an early postprocedural study might provide a more effective baseline for comparison with the scans during a longer follow-up. This should not be a predischarge CT-scan but might be a 1-month postprocedural scan. An important drawback of the early postoperative CTA is the use of nephrotoxic agents. It is, however, hard to measure the sole influence of the CTA because all patients receive contrast agents during the operation. In our study population, 5% had a prolonged hospital stay for intravenous fluid administration to normalize creatinine clearance, part of which might be caused by the extra amount of agents needed for CTA. Another cause could be the occurrence of thromboembolic renal complications. Although we could not prove to what extent the predischarge CTA deteriorates renal function after EVAR, efforts should be made to minimize the amount of nephrotoxic contrast periprocedure.

In our study, the data of the early postprocedural CTA did not influence postoperative treatment policy in 99% of EVAR patients. Only four patients were treated for EVAR-related complications found on the early postprocedural CTA. Another five (2%) interventions were done in the 3-month interval after successful EVAR, but all were done for clinical symptoms such as critical leg ischemia and groin infection. The early postprocedural CTA was therefore of no clinical value in discovering these postoperative complications.

At the intraoperative angiographies, 80 of 93 (86%) of the endoleaks from the predischarge CT scans were already diagnosed. This includes 6 of 8 type I endoleaks seen on the predischarge CT scans. In case of intraoperatively detected type I endoleaks, these were all judged to be small, with no need for intraoperative conversion to an open procedure or endovascular revision.

More than half of all endoleaks diagnosed on the early CTA resolved within 3 months after EVAR without intervention. This overdiagnosis is a psychological burden for most of these patients. In addition, only two (1%) new endoleaks appeared in the 3-month period, and both were benign type II endoleaks.

Two of eight type I endoleaks were treated in the first postoperative month with endovascular reintervention. Although it is recommended to treat type I endoleaks as they appear in association with the growth of aneurysms, spontaneous sealing of type I endoleaks has been described. ^15,16 In our studied patient population, six of eight type I endoleaks were managed in a conservative manner because the AAA diameter and endoleak in these patients were relatively small. Type I endoleaks resolved spontaneously in three of six conservatively managed patients. More data from other studies are necessary to outline a thorough policy in case of early detected type I endoleaks.

Although the current study shows that the early postoperative CTA did not influence our treatment policy in the first 3 months post-EVAR in most of the 291 treated patients, possible bias has been introduced by the retrospective character of the study. The interpretation of the 3-month CTA could have been influenced by the information available from the predischarge CTA. A prospective randomized trial would be necessary to exclude this bias.

A limitation of this study might be the fact that it cannot set the course of the most ideal follow-up schedule in the first year post-EVAR. For this reason, the frequently used follow-up protocol including 1-, 6-, and 12-month CT scans (three scans) should be prospectively compared with a 3- and 12-month follow-up protocol (two scans). The 1-month scan could be preferred over the 3-month scan for those patients in whom the preoperative CTA is too old for settlement of AAA diameter or if the AAA diameter is so large that any kind of endoleak would propose a major risk for AAA rupture. However, the main benefit of this study is the fact that the predischarge CT scan can safely be abandoned from all possible follow-up protocols.

Last but not least, it is important to note that no patients needed an urgent intervention after the 3-month postoperative scan. If the indication for secondary reintervention was set, it could be performed in an elective setting. Therefore, none of the 99% patients would have been exposed to unwarranted hazards if the early CTA had been skipped.

Conclusions

In 287 (99%) of 291 patients, the postprocedural CTA did not influence our treatment policy in the first 3 months after EVAR. More than half of the early endoleaks were self-limiting, and new endoleaks were seen in only two (< 1%) patients at the 3-month follow-up CTA. Hospital stay was prolonged in 5% of the patients owing to temporary worsening of their creatinine clearance. After an uneventful EVAR procedure, the early postprocedural CTA can be left out.