Intermediate Follow-Up after Endovascular Aneurysm Repair: Can We Forgo CT Scanning in Certain Patients?
Article Outline
Current recommendations for follow-up after endovascular repair of abdominal aortic aneurysms (EVAR) include yearly computed tomographic (CT) scans after the first year. We hypothesize that this is unnecessary for patients who have aneurysm sacs that are stable or shrinking at 1 year and no evidence of endoleak. To explore this hypothesis, we reviewed the records of all patients undergoing EVAR at our institution who were implanted with grafts that are currently commercially available and had a minimum of 18 months' follow-up. Of 415 patients who underwent EVAR over an 8-year period, 93 met the entry criteria. At a mean follow-up of approximately 3 years, secondary interventions were required in 13%, 39%, and 25% of patients undergoing EVAR with Zenith, AneuRx, and Excluder devices, respectively, and secondary interventions after the first year were required in 3%, 22%, and 8% of such grafts, respectively. Seventy-one patients (76%) had aneurysm sacs that were stable or shrinking at 1 year and no endoleak. Only two of these patients subsequently required reintervention. Both patients had AneuRx grafts, and both problems could have easily been identified without CT scanning. Our data support the hypothesis that patients who meet these criteria at 1 year are unlikely to have problems that cannot be identified by ultrasound and/or clinical evaluation alone and, thus, that CT scans are not necessary after this point, especially in patients with Zenith or reengineered Excluder devices.
Introduction
Endovascular repair of abdominal aortic aneurysms (EVAR), introduced into general clinical practice in the mid-1990s, is out of its infancy and well into an early, robust adolescence. Like any teenager, however, we still do not quite know how things are going to turn out. EVAR has a reputation, largely based on experiences with early-generation devices, that unpredictable late problems can occur.1, 2, 3
Current recommendations therefore are to obtain computed tomographic (CT) scans with contrast at yearly intervals essentially for the lifetime of the graft and patient.4 This creates, however, two problems: persistently increasing total cost of repair (including discomfort and trouble for the patient) and renal damage due to the cumulative contrast load.5, 6, 7 Our experience is that there is a very low incidence of new problems in patients who are doing well in the short term. This is based largely on the preferential use of later-generation devices incorporating suprarenal fixation. In addition, all such problems can be identified by noting either the size of the aneurysm or the presence of major endoleak, along with being aware of the patient's clinical situation. In other words, a good ultrasound and clinical visit will identify all such problems in patients who have done well in the short term.
We specifically hypothesize that in patients whose sacs are shrinking and who have no endoleak at 1 year following EVAR, no problems will arise that could not have been identified by ultrasound and clinical exam alone (with the caveat that specific device-related characteristics may exist). In other words, we believe that CT scan is not needed in patients receiving Gore or Cook grafts after the first year if these criteria are met.
Methods
All patients undergoing EVAR at the University of Rochester Medical Center from August 7, 1997, to July 20, 2006, were identified. Several groups were excluded as not relevant to this hypothesis: those with grafts placed as part of a research protocol and not currently commercially available and those without both testing at 1 year and at least 6 months' follow-up after that. Data were derived from three sources: our prospective EVAR database, inpatient hospital records, and outpatient clinic follow-up records. Choice of grafts was surgeon- and time-dependent. We observed a shift from AneuRx (Medtronic, Sunnyvale, CA) to Zenith (Cook, Bloomington, IN) devices as our primary graft of choice as time passed, primarily due to our perception of better intermediate-range outcome. Apart from rare cases where an Excluder (Gore, Flagstaff, AZ) was chosen for small iliac vessels, choice of graft was not made on the basis of anatomy, nor was a different type of graft used after an unsuccessful attempt at using another. We also noticed an increasingly liberal willingness to accept suboptimal proximal anatomy when using a Zenith device. Reintervention strategies were consistent throughout this time frame and, although subjective, were based on sac size, duration and type of leak, and clinical status of the patient but not on the type of device implanted.
Recognizing that variability exists, measurements were accepted as accurate to the nearest millimeter assuming that errors will cancel out. When both CT and ultrasound measurements were available, short-axis CT measurements were preferentially used. If a sac size was within 1 mm either way of the original size, it was considered to be no different. Hypertension and chronic obstructive pulmonary disease (COPD) were defined simply by a diagnosis of such documented in the admission or consultation note and family history by a first- or second-degree relative with a documented aneurysm.
Our specific hypothesis is that sacs that are smaller (by ≥1mm) than at the time of repair and have no demonstrable endoleak by CT or ultrasound at the 1-year exam have no late problems; these patients were defined as “optimal” 1-year outcomes for the purposes of this investigation. All others were defined as “suboptimal” 1-year outcomes. Note that the suboptimal group includes patients whose sacs were within 1 mm either way of the starting point (no change), those whose sacs were larger by more than 1 mm (enlarging), and those with any endoleak by any modality, regardless of sac size.
Statistical analyses were performed using Statview 5.0.1 (SAS Institute, Cary, NC). Continuous variables are expressed as means ± 1 standard deviation (with ranges given, when appropriate), analyzed by means of t-tests or analysis of variance depending on the number of groups, and continuous variables as percentages, analyzed by means of chi-squared and Fisher's exact testing depending on cell size.
Results
Four hundred fifteen EVARs were performed at the University of the Rochester between August 1997 and July 2006. Ten of these were done for reasons other than aortic aneurysms, 75 patients had received grafts that are not currently commercially available, 135 either died or were lost to follow-up (patients or charts) prior to 1 year, seven did not have testing performed at 1 year, and 95, currently doing well, did not have sufficient follow-up after the 1-year visit to be meaningfully analyzed here (Fig. 1). This left a total of 93 patients with grafts that are currently commercially available, who were imaged at 1 year after implantation, and who had significant follow-up after this to analyze.
Fig. 1
Exclusion criteria algorithm for this study. Our goal was to include only those patients whose experiences would be clinically relevant to current practice.
Demographics are presented in Table I. Our cohort was overwhelmingly male and, for reasons that are obscure, Caucasian. Forty-three percent of the patients had a family history noted with at least one first- or second-degree relative with an aneurysm, 35% were current smokers or had quit within the decade, and 22% had documented COPD.
Table I. Initial demographics
| Unspecified | ||
|---|---|---|
| Age (years, range) | 71 ± 17 (50-92) | |
| Male gender | 80 (86%) | |
| Race | 8 | |
 Caucasian | 83 (98%) | |
| African American | 1 | |
| Hispanic | 1 | |
| Risk factors | ||
| Family history | 32 (43%) | 19 |
| Ever smoked | 69 (80%) | 7 |
| Current | 13 (15%) | |
| Quit within 10 years | 17 (20%) | |
| Former | 39 (45%) | |
| Hypertension | 60 (67%) | 4 |
| COPD | 19 (22%) | 6 |
At the time of repair, mean aneurysm size was 5.6 cm (range 4.3-9.0) (Table II). Sixty-eight percent of repairs (n = 63) were performed using the Zenith graft, 19% (n = 18) using AneuRx, and 13% (n = 12) using Excluder. Twenty-nine percent had an endoleak at the close of the procedure, the large majority of which (85%) were type II. There was one leak felt to be type I despite proximal cuff placement (AneuRx), although on later imaging this was found to be a type II leak and successfully coiled. There were also two type IV and one “uncharacterized” endoleaks, all of which spontaneously resolved by their 1-month CT scans. Median length of stay was 3 days.
Table II. Status at time of EVAR (n = 93)
| Size at EVAR (cm) | 5.6 ± 0.8 (range 4.3-9.0) |
|---|---|
| Device (n [%], mean size [cm], endoleaks at close [%]) | |
| Cook Zenith | 63 (68%), 5.9, 15 (24%) |
| Medtronic AneuRx | 18 (19%), 5.9, 7 (39%) |
| Gore Excluder | 12 (13%), 5.4, 5 (42%) |
| Endoleaks at close of procedure | 27 (29% of cases) |
| Type II | 23 (85% of endoleaks) |
| Length of stay (days) | 3.5 ± 2.7 (range 1-22, median 3) |
Eighteen patients (19%) required a total of 32 secondary reintervention procedures at some time following repair, at a mean follow-up of 32 ± 17 months (Table III). The rate of secondary intervention differed significantly between graft types (p < 0.04), with that after use of Zenith being lowest at 13% and that after use of AneuRx being highest at 39% (p < 0.02); the rate after use of Excluder was intermediate but not significantly different from either. In addition, the need for secondary intervention after use of Zenith and Gore was higher in the first year (75% and 66% of all those needing intervention, respectively). After 1 year secondary intervention was highest after AneuRx graft placement (4 of 7, p < 0.02). Again, the rate of secondary interventions after the first year was significantly different between the Zenith and AneuRx groups (p < 0.01), while that after use of Excluder was not significantly different from either. Interventions needed included 21 coil embolizations for type II endoleak, five additional distal extensions, four femoral-femoral bypasses for limb occlusion, one explant, and one conversion to an aortouni-iliac configuration.
Table III. Need for secondary intervention (n = 93)
| All | Within 1 year | After 1 year | |
|---|---|---|---|
| Zenith (n = 63) | 8 (13%) | 6 (10%) | 2 (3%) |
| AneuRx (n = 18) | ∗7 (39%) | 3 (17%) | ∗4 (22%) |
| Excluder (n = 12) | 3 (25%) | 2 (17%) | 1 (8%) |
| Total | 18 (19%) | 11 (12%) | 7 (7%) |
| p | <0.04 overall | NS | <0.02 overall |
| <0.02 vs. Zenith | <0.01 vs. Zenith |
∗Aneurx was different than Zenith. |
At 1 year, 79%, 56%, and 83% of patients undergoing repair using Zenith, AneuRx, and Excluder grafts had shrunk by more than 1 mm and had no evidence of endoleak (defined as optimal 1-year outcomes with respect to this algorithm), respectively (nonsignificant, Table IV). Overall, about three-quarters of EVARs met the criteria for optimal outcome at 1 year. Patterns of suboptimal outcome, however, were different between devices (p < 0.04). After use of Zenith, nine (of 11) sacs were the same size and only 6% of sacs had an endoleak. After AneuRx, however, three of five suboptimal sacs were larger, with only two being unchanged, and endoleaks existed in 28% of cases (p < 0.02 vs. Zenith). Excluder, again, had results intermediate between the two (one sac larger, the other without change, and 8% endoleaks).
Table IV. CT findings at 1-year follow-up (n = 93)
| Optimal | Suboptimal | ||||
|---|---|---|---|---|---|
| Follow-up (months) | Smaller and no endoleak | Same size | Larger | Endoleak | |
| Zenith (n = 63) | 31.4 ± 18 | 51 (81%) | 9 | 2 | 4 (6.3%) |
| AneuRx (n = 18) | 37.3 ± 16 | 10 (56%) | 2 | 3 | 5 (27.8%) |
| Excluder (n = 12) | 28.5 ± 8 | 10 (83%) | 1 | 1 | 1 (8.3%) |
| Overall (n = 93) | 32.2 ± 17 | 71 (76%) | 22 (24%) | ||
| 12 | 6 | 10 (10.8%) | |||
| p | 0.31 | 0.22 | 0.04 overall | ||
| 0.02 vs. Zenith | |||||
Intermediate results are shown in Table V. Mean follow-up was the least after use of Excluder and greatest after AneuRx, but these differences did not approach significance. Notably, only two patients, both with AneuRx grafts who met our criteria for 1-year optimal outcome with regard to this algorithm, had late problems. One had a proximal dislodgement resulting in a type I endoleak and expansion to 7.0 cm at 24 months, and the second had limb occlusion with an ischemic leg at 13 months.
Table V. Long-term results (n = 93)
| Outcome at 1 year (n) | Outcome after 1 year (n) | |
|---|---|---|
| Optimal | 71 | 69 no problems (all sacs smaller, no endoleaks) |
| 1 (AneuRx): proximal dislodgement, expanded to 7 cm at 24 months | ||
| 1 (AneuRx): limb occlusion at 13 months | ||
| Suboptimal | 22 | |
| Zenith | 13 | |
| Same size (n = 9) | ||
| Larger (n = 2) | ||
| Smaller (9.0-8.6) at 18 months, no leak | ||
| Same size (4.5-4.5) at 24 months, no leak | ||
| Endoleaks (n = 4) | ||
| All type II: 2 smaller, 2 same size | ||
| AneuRx | 7 | |
| Same size (n = 2) | ||
| Larger (n = 3) | ||
| Smaller (7.2-4.7) at 60 months, no leak | ||
| Larger (5.8-6.6) at 60 months, no leak | ||
| Larger (6.5-6.9) at 60 months, no leak | ||
| Endoleaks (n = 5) | ||
| 4 type II: 2 smaller, 2 same size | ||
| 1 type I: larger | ||
| Excluder | 2 | |
| Same size (n = 1) | ||
| Larger (n = 1) | ||
| Distal extension at 13 months, larger (5.5-6.0) at 36 months, no leak | ||
| Endoleak (n = 1) | ||
| Type II: larger |
Of the 27 endoleaks seen at the close of operation, 23 had disappeared at 1 year and four (15%) persisted. Two of these (one Excluder, one Zenith) spontaneously resolved, and the other two (both AneuRx) required coil embolization. Of the 66 with no leak at operation, 60 remained without leak and six new leaks were found at 1 year. One of these (AneuRx) spontaneously resolved, and five (two AneuRx, two Zenith, and one Excluder) required intervention. Of the 83 patients without endoleak at 1 year, only one (AneuRx) developed a later (type I) leak.
Discussion
We hypothesize that patients with a shrinking sac and no evidence of endoleak 1 year after EVAR will have negligible risk of late problems and that any problems that do occur can be picked up by ultrasound determination of sac size and history and physical examination alone, especially if they have been repaired using a second-generation Excluder or Zenith device. In other words, if these criteria are met, further CT scans are not needed and follow-up can be by means of ultrasound and office visits only. Our data, although with a smaller number of patients and shorter follow-up than are probably needed for “proof,” are consistent with this hypothesis. Three-quarters of the total patient group met our criteria of a shrinking sac and absence of endoleak, and of these, only two had subsequent problems: one graft dislodgement with dramatic enlargement and another with limb occlusion. Both of these problems were easily identifiable by means of ultrasound, history, and physical; and both occurred in patients with AneuRx grafts.
The recommendation that lifelong CT scans be obtained was based on early uncertainty of long-term behavior of the aorta. Proximal cuff dilation is seen after open repair,8 and experience with early-generation devices (Ancure, AneuRx, and the original Excluder)2, 3 suggested that rates of proximal dislodgement, persistent endoleak, and ongoing need for reintervention are high.2, 3, 4, 9
Later-generation endografts, such as Zenith (proximal suprarenal barbs and modularity) and revised Excluder (low-permeability polytetrafluoroethylene), are associated with significantly lower rates of failure and reintervention.1, 2 For example, in one recent series fewer than 9% of patients with a Zenith device needed a second procedure at any point.10 In our experience, the overall reintervention rates in patients with Zenith and revised Excluder devices are 13% and 25%, respectively, but only 3% and 8%, respectively, after the first year.
Late problems with endovascular repairs fall into two major categories: endoleak with enlargement and limb occlusion. Limb occlusion is easily picked up with history and physical, and a CT scan is not needed for diagnosis. Endoleak is of little consequence as long as the sac is shrinking, so size alone can be followed.11 Although the sensitivity of ultrasound for endoleak identification is highly variable, it can measure sac size with reasonable accuracy.12, 13 Even if a type II endoleak is seen, the patient is usually observed as long as the sac is not enlarging. In other words, the size of the residual sac matters more than the presence of a type II endoleak. Limb occlusion, of course, is identified clinically (and, if asymptomatic, does not require therapy).
Our algorithm should be quite evident but is graphically illustrated in Figure 2 for reference. If at 1 year after EVAR, especially if the patient has a Zenith or modern Excluder in place, the sac is clearly shrinking and no endoleak is seen, the patient undergoes ultrasonic evaluation by a trained technician and office evaluation on a yearly basis. If at any time the sac is found to be enlarging, CT scan is performed. If an endoleak is seen but the sac is not enlarging, required work-up varies. We suggest that if the sac is definitely shrinking the patient can continue to be followed by ultrasound, but CT scanning is probably the best option if the sac is stable (and certainly if larger). If an ultrasound is nondiagnostic because of obesity, gas, or lack of a suitable window, CT should be performed; but this can be done without contrast simply to assess size. Finally, issues such as limb occlusion are dealt with according to accepted clinical principles. We suggest that patients with AneuRx grafts continue to undergo CT scanning because of the higher risk of late proximal dislodgement and type I endoleak, but even patients with early-generation Excluders with endotension almost always manifest with sac enlargement and thus can undergo ultrasound alone if it is shrinking.
Fig. 2
Overall suggested algorithm for choosing between CT scanning and ultrasound alone (both combined with clinical evaluation) based on status 1 year after EVAR.
Our data obviously apply to the first 3 years or so only of follow-up. While the increasing use of Zenith devices later in the time frame of this study may induce learning-curve bias, we are also treating more difficult proximal anatomy with this device. We are currently applying this algorithm clinically and will continue to share our experiences as follow-up lengthens. It is hard to postulate that a sac that has shrunk to “encase” the graft, as illustrated in Figure 3, could ever cause a problem. Our 1-year cut-off is, as well, fairly arbitrary; and more experience will be needed to refine the precise determinants of safety or risk. For example, is it status at 1 year exactly that predicts outcome or some other time frame? If a patient meets these criteria at any point in time, is he or she “home free”? Just what constitutes “shrinkage,” and how does the interplay of accuracy and systemic error for both CT scanning and ultrasound affect prediction of problems? Finally, if size and clinical status alone are all that matter (along with the observation that different devices have different late failure profiles and the opinion that endoleaks significantly major enough to cause problems should be identifiable on ultrasonic examination), is CT at any point ever required, or can follow-up be performed completely by means of ultrasound alone?
Fig. 3
A patient 40 months following EVAR using a Zenith device. A Cut at prior largest diameter of original aneurysm (was 5.3, now 3.0 mm). B Cut after limb bifurcation.
Conclusions
Our experience supports the hypothesis that CT scanning is not needed after EVAR if patients demonstrate a shrinking sac and absence of endoleak, especially if repair was performed using a later-generation Excluder or Zenith device. Approximately three-quarters of patients meet these criteria at 1 year, and no such patient has had any problem which could not have been picked up by ultrasound and clinical exam only. Adoption of this algorithm would save money and time and reduce the risk of cumulative contrast nephrotoxicity.
References
- . Mid- and long-term device migration after endovascular abdominal aortic aneurysm repair: a comparison of AneuRx and Zenith endografts. J Vasc Surg. 2005;42:392–401
- Sac behavior after aneurysm treatment with the Gore Excluder low-permeability aortic endoprosthesis: 12-month comparison to the original Excluder device. J Vasc Surg. 2006;44:694–700
- . Long-term results of endovascular abdominal aortic aneurysm treatment with the first generation of commercially available stent grafts. Arch Surg. 2007;142:33–41
- ACC/AHA 2005 practice guidelines for the management of patients with peripheral arterial disease (lower extremity, renal mesenteric and abdominal aortic). Circulation. 2006;113(11):e463–e654
- Follow-up costs increase the cost disparity between endovascular and open abdominal aortic aneurysm repair. J Vasc Surg. 2005;42:912–918
- . Acute renal failure. N Engl J Med. 1996;334:1448–1460
- . Intravascular contrast media: adverse reactions. AJR Am J Roentgenol. 1998;149:665–670
- . Fate of the proximal aortic cuff: implications for endovascular repair. J Vasc Surg. 1997;26:492–501
- Gore Excluder bifurcated endoprosthesis annual clinical update. 2006, www.goremedical.com.
- Long-term outcome and reintervention after endovascular abdominal aortic aneurysm repair using the Zenith stent graft. J Vasc Surg. 2007;(in press)
- Realistic expectations for patients with stent-graft treatment of abdominal aortic aneurysms. Results of a European multicentre registry. Eur J Vasc Endovasc Surg. 1999;17:507–516
- Computed tomography versus color duplex ultrasound for surveillance of abdominal aortic stent grafts. J Endovasc Ther. 2005;12:568–573
- Color duplex ultrasonography is insensitive for the detection of endoleak after aortic endografting: a systematic review. J Endovasc Ther. 2005;12:296–305
Presented at the Seventeenth Annual Winter Meeting of the Peripheral Vascular Surgery Society, Steamboat Springs, CO, January 26-28, 2007.
PII: S0890-5096(07)00287-7
doi:10.1016/j.avsg.2007.07.015
© 2007 Annals of Vascular Surgery Inc. Published by Elsevier Inc All rights reserved.
