Abstract
Objectives
Explantation of both endovascular endovascular aneurysm repair and open aortic grafts is a procedure associated with high peri-operative risk, and the current study sought to determine the outcomes and trends over time in these patients.
Methods
This study examined data from all patients undergoing explant of an aortic graft (both open and endovascular) between January 2004 and December 2020 at a single centre. Variables analysed included comorbidities, duration to and indication for explantation, type of revascularization, in-hospital complications and mortality, duration of hospital and ICU stay, and out-patient mortality.
Results
Of 688 open and 1352 EVARs performed, 46 patients underwent 48 explants. Five were open grafts and 43 were endografts, equating to an explant rate of 0.73% of open and 3.18% EVARs. Average time to explant was 70 months, with patients presenting electively having a significantly longer duration to representation than those presenting emergently (51 vs 44 months, p=0.003). Indication for explant was endoleak in 70%, infection in 23%, and occlusion in 6%. Of the endoleaks, 61% of were Type 1, 22% Type II, 11% Type IV, and 6% Type V. On representation, 17 patients (35%) were symptomatic. Overall mortality rate was 8.3%, with a trend for higher mortality in emergent than elective presentations (11.8 vs 6.5%, p=0.55). There was no change in explant rate over time.
Conclusions
Elective aortic graft explantation is associated with low mortality, despite its complexity and patient comorbidities. Patients presenting with symptoms suffered higher mortality and a longer post-operative course, suggesting that aortic graft explantation should be considered sooner rather than later, rather than persisting with repeated endovascular management.
Introduction
Endovascular aneurysm repair (EVAR) is considered the optimal method of management for abdominal aortic aneurysms (AAAs) in certain patient populations,1,2 due to its lower mortality at short-term follow-up.3-5 This results in over 70% of AAA repairs being performed with EVAR, 6 although there are still significant numbers of open AAA repairs. Explant of an EVAR is required in two to four percent of patients,7,8 with the primary indication for explantation being endoleaks (up to 90% of cases) and infection. 7 A previous Australian study demonstrated 0.9% of open, and 1.4% of endovascular aortic repairs experienced subsequent infection, 9 which is significant given the number of aortic repairs performed annually.
Explantation performed for management of endoleaks involves in situ aortic reconstruction using prosthetic. In contrast, explantation in the setting of infection requires resection of all prosthetic material, wide local debridement, and either in situ reconstruction (a neo-aorto-iliac system using femoral vein is recommended as first line by the European Society of Vascular Surgery, 10 but may include rifampicin-soaked prosthetic), or an extra-anatomic reconstruction such as axillo-bifemoral bypass. Given the increased complexity of excision and reconstruction for infective indications, explantation in these settings has a mortality rate of 18–30%.11,12 In contrast, elective EVAR explantation for endoleak has a mortality rate of less than three percent, although such procedures are still associated with a significant morbidity, and prolonged hospital stay. 13
The current study sought to assess epidemiology, morbidity, mortality, outcomes and trends over time of aortic explants after both open and endovascular repair at a single institution.
Materials and methods
A retrospective analysis of patient records at our tertiary referral centre was performed, examining data from all patients undergoing explant of an aortic graft (both open and endovascular) between January 2004 and December 2020, initially inserted for aneurysmal disease. Our institution services a population of one million people. Patients were identified from a prospectively collected vascular surgery database. Demographic, surgical and clinical variables were collected. Independent variables included age, sex, aneurysm size, comorbidities, indication for initial repair and whether this was open or endovascular, type of graft, duration to explantation, and indication for explantation. Outcome variables included type of reconstruction, in-hospital complications, in-hospital mortality, rates of return to theatre, acute kidney injury and duration of hospital and ICU stay. Acute kidney injury was defined using the RIFLE criteria, 14 and graft infection was defined using the MAGIC criteria. 15 Standard post-operative follow-up at the study institution is performed at 2–4 weeks, 3–6 months, then 12 months post-operatively, followed by annually, and infective signs and symptoms are assessed at each visit.
We also noted the total number of open and endovascular aneurysm repairs over the study period, to supply the denominator when determining incidence of explantation. Correspondence and follow-up radiology from consultant vascular surgeons’ rooms was obtained to determine length of follow-up, and long-term outcomes. This study was approved by the Local Health District Ethics Review Committee.
Independent T tests were used to compare means between groups for continuous variables, and categorical variables were compared using Pearson chi-squared tests. Linear regression was used to assess change in rates of number of explants performed over time. A P value of 0.05 was considered to be significant. All analyses were performed using SPSS 24.
Results
Demographics
Over the study period, 688 open aneurysm repairs and 1352 EVARs were performed, with 48 aortic explants in 46 patients. There were two patients who underwent explant and aortic reconstruction (one tube and one bifurcated reconstruction) for endoleak, who then went on to require explant for infective indications. Five explants were open grafts (12%), and 43 were endografts, which equates to an explant rate of 0.73% of open repairs, and 3.18% of EVARs (overall rate of 2.25%). There was no change in number of explants per year over the study period (Figure 1). Demographic data for these patients is demonstrated in Table 1. Average age was 67 years, 92% were male, and 82% were current or ex-smokers. Six initial aortic repairs (15%) were performed at another institution. Number of explants performed per year corrected for number of graft insertions, by year of insertion. Baseline characteristics for patients undergoing initial open or endovascular aortic repair during the study period. Results are presented as means with standard deviations. Categorical variables are expressed as raw number, with percentage of the total in brackets, as some data were not available for all patients.
Peri-operative characteristics for patients undergoing initial open or endovascular (EVAR) aortic repair during the study period. Results are presented as means with standard deviations. Categorical variables are expressed as raw number, with percentage of the total in brackets. * p < 0.05. ICU = intensive care unit. £ Supra-celiac or supra-renal clamp includes open repairs only.
Indications for explant
Peri- and intra-operative characteristics for patients undergoing aortic explant during the study period. Results are presented as means with standard deviations. Categorical variables are expressed as raw number, with percentage of the total in brackets. SMA = superior mesenteric artery.
Peri-operative details
The trans-peritoneal approach was utilised in 40 patients (84%; Figure 2). 80% of patients had endografts with supra-renal fixation, requiring a supra-celiac clamp in 44% of cases (duration range of 3–45 min) and a supra-renal clamp in 17% of cases (duration range of 3–60 min). 15% of patients had some graft material retained, all patients with non-infective aetiology. Average surgical duration was 452 min (range 210–940 min). Endovascular graft post-explantation, which had undergone multiple endovascular revisions.
Two patients underwent revascularisation of celiac, superior mesenteric and bilateral renal arteries. One of these was performed via Carrell patch of celiac and SMA onto aortic prosthetic graft, and 6 mm PTFE bypass to bilateral renals. The other visceral and renal vascularisation was performed via a prosthetic branch graft. Six patients underwent renal revascularisation, performed via 6 mm PTFE bypasses from aortic prosthetic graft or its limbs. Seven patients (15%) underwent extra-anatomic lower limb revascularisation, of which six were axillo-bifemoral bypasses, and one bilateral axillo-femoral bypass.
Post-operative characteristics for patients undergoing aortic explant during the study period. Results are presented as means with standard deviations. Categorical variables are expressed as raw number, with percentage of the total in brackets. * p < 0.05. ICU = intensive care unit.
Aortic graft infections
Microbiology of patients requiring explant for aortic graft infections. Note four patients had multi-organism infections.
Long-term outcomes
21 of the surviving 44 patients (47%) followed up with their vascular surgeon post discharge, with an average duration of follow-up of 45 months (range 1–143 months). There was mortality information regarding three patients who died in the follow-up period from non-aortic causes, one from cardiac failure (131 months post-discharge), one from chronic lymphocytic leukaemia (7 months) and one from respiratory failure (33 months).
Discussion
Low morbidity in elective explantation
This is the seventh largest case series on aortic explants to date, and demonstrates the feasibility of aortic graft explantation. Explantation was associated with low peri-operative mortality when performed electively, and good survival at follow-up. The observed mortality rate of eight percent was low, considering the complexity of aortic graft explantation in often comorbid patients. This rate compares favourably with recent series.13,16 Unsurprisingly, mortality was higher in patients presenting emergently, as has been previously demonstrated.13,16,17 As such, we agree with the suggestion that aortic graft explantation should be considered early (if not suitable for fenestrated or branched endografting), before these patients then present with either symptoms or rupture, 16 which is associated with markedly higher mortality. For example, in the case of ongoing endoleak, the option of late open conversion should always be considered rather than persisting with repeated unsuccessful endovascular interventions. While an initial EVAR may have been performed rather than an open repair due to perceived patient frailty, we noted that patients were only 5 years older at representation (on average), with a mean age of 72 years. A high proportion of these patients then went on to survive their explant, suggesting morbidity and mortality of explantation is not as high as expected. This sentiment is echoed by Scali et al., who demonstrated a mortality rate equal to that in elective juxta-renal aneurysm repairs, and suggested that an open option ‘should be considered before pursuing complex endovascular remediation of EVAR failures’. 2
Change in graft technology
The EVAR explant rate of three percent is similar to that seen by two large reviews.17,18 Although not seen in the current study due to low rates of early conversion, a reduction in rates of early conversion has been seen over time, 18 which is not surprising given the improvement in endograft technology. There was a high proportion of patients undergoing explant for Type 1a endoleak in the current cohort. A large proportion of the endografts was first generation, so likely lacked the active proximal fixation that is associated with reduced migration. 19 It is likely that in 2021 patients with Type 1a endoleaks would be offered a fenestrated cuff, while an explant was previously the only definitive option for these patients. However, a recent review demonstrated that while associated with higher 30-day morbidity, EVAR explantation demonstrates equivalent 1-year survival with fenestrated cuff for Type 1a endoleaks, 20 suggesting that explantation is a viable method of management of these patients. Total number of grafts precluded comparison between suprarenal and infrarenal fixation regarding rates of Type 1a endoleak. Further, Cook grafts were the predominant type of graft inserted before 2005, so higher rates of explant in this graft were not considered significant. Regarding graft fabric, a significant reduction in aortic sac growth was noted with changes in graft composition 21 in the Excluder device. 11% of endoleaks in this study were Type IV, which would no longer be seen due to changes in graft material composition. Interestingly, there was no change in explants over time in this series. Only five of the current grafts were inserted before the year 2000, and it is possible that a higher rate of explants (both on-table and delayed) would have been observed due to both learning curve and early technology.
Role for partial graft explantation
The current series demonstrated low rates of retained graft material when compared to others (15 vs 55%). 13 The recent study from Chastant et al. 13 advocated for a policy of partial graft retention if possible, resulting in reduced operative time and blood loss, due to reduced need for dissection. This may be of special importance in the presence of supra-renal fixation, where control of the supra-renal segment will require supra-celiac clamping, and either extensive supra-renal dissection or distal traction, 22 both of which risk significant aortic injuries of the visceral segment. When compared to either supra- or infra-renal clamping and transection of the graft after the first covered stent, morbidity of total endograft removal is much higher, but several factors need to be considered to prevent damage form remaining stent struts. 16 As such, consideration should be given to leaving some graft in situ, as long as there is no concern regarding infection. The reason for the low rate of graft retention in the current series is unclear, but may be due to almost 40% of the patients requiring explant for either a Type Ia endoleak, or infection
Impact of supra-celiac and supra-renal clamping
The current series observed a significant (61%) rate of supra-celiac or supra-renal clamping, especially when compared to similar series (35% 16 to 37% 13 ). This may be related to the high rate of total graft excision, discussed above. Of interest, all patients in the current study who suffered acute renal failure had supra-celiac or supra-renal clamping, especially when compared to the series by Ben Abdallah et al., 16 which saw a low rate of supra-celiac or supra-renal clamping, and had no patients suffer acute renal failure. There are several possibilities for this, most likely that the ischaemic insult from supra-celiac or supra-renal clamping caused temporary renal impairment. Further, there is risk of damage to renal arteries when explanting devices with suprarenal fixation. The recent study of 1808 patients undergoing open AAA repair in the United States demonstrated increased mortality with supra-celiac clamping, and increased renal dysfunction with supra-renal clamping, but only in patients with baseline renal impairment. 23 We agree with the authors’ suggestion that proximal clamp level be kept as low as possible, which again argues for consideration of graft retention where possible. However, there is always the possibility that patients requiring clamping above the level of the renal arteries were more likely to be undergoing more complex interventions.
Infection rate
Rate of aortic infections in this study was 0.7% for open repairs and 0.25% for EVARs, which is less than half that seen previously in Australia 9 and internationally 24 . No reason for this low rate of infection could be identified. There was only one death in the cohort, which is possibly due to explants only being offered to patients with low operative risk, given the high mortality associated with aortic graft explant for infection (with mortality rates previously seen to be 30–40%12,25). The in-hospital morbidity rate was over three-times higher in patients undergoing explant for infective indications than those with endoleaks or occlusions. This has been noted previously, 8 and is likely related to both the infective pathology, but also the complexity of the surgical intervention. 13 Unsurprisingly, these patients then had markedly longer ICU and hospitals stays. Interestingly, a recent series demonstrated no difference in morbidity or mortality between infective and non-infective indications, although there was an increase in rates of surgical revision, as well as surgical duration and transfusion requirement. 13 As in our series, the majority of these patients underwent extra-anatomic reconstruction with axillo-femoral bypasses, and none of these suffered subsequent graft infection, limb loss, nor stump blowout. This is in contrast to that seen previously, with axillo-femoral bypass suggested to be associated with higher rates of graft thrombosis, limb loss and stump blowout than in situ reconstructions. 26 Recent recommendations suggest a neo-aorto-iliac system for reconstruction post aortic graft infection, 10 which may see more reconstructions performed this way in the future. The current study found that patients with infected grafts presented much earlier than patients requiring intervention for endoleaks (2.5 vs 6 years). A previous Australian study found 70% of patients with infected aortic grafts presented within 12 months of the index procedure, 9 suggesting that onset of aortic graft infection is relatively rapid, and all patients should be monitored closely for infective signs at each post-operative visit after aortic repair. Further, the vascular surgeon should impress on the patient and their general practitioner to be aware of signs and symptoms of graft infection, and to have a low threshold for early specialist review.
Follow-up rates post explantation
We noted poor follow-up in patients after aortic graft explant, with attendance at follow-up of 47%. This is even lower than the compliance with follow-up after EVAR at our institution, which was previously noted to be 77%. 27 This is also lower than that seen previously in explant patients, 16 and may be related to distance from treating centre (25% of the patients at our institution are from rural and regional areas), but a greater attempt at ensuring compliance with follow-up should be offered.
Limitations
There are some limitations of this study, namely its retrospective nature and single-centre design, and the poor long-term follow-up. The limited long-term follow-up requires the out-patient survival be interpreted with caution, as it is not unreasonable to think that the most compliant patients (that is, those that attended follow-up) were most likely to have improved long-term survival due to engagement with the healthcare system. Further, there is heterogeneity resulting from patient selection and operating surgeons, and given the small numbers in this study, larger studies are required before detailed inferences can be made. It is also possible that complications occurring after discharge were treated at other institutions (and therefore not included in our analysis), and deaths were missed, which may result in an under-estimation of the out-patient morbidity and mortality.
Conclusions
Elective aortic graft explantation is associated with low mortality, despite its complexity and patient comorbidities. Patients presenting with symptoms suffered higher mortality and a longer post-operative course, suggesting that aortic graft explantation should be considered early, especially given the low mortality rate.
Footnotes
Acknowledgements
Thanks to Alison Burnett and Jana Pinkova for their assistance with data collection
Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
