Abstract
Objectives
To evaluate outcomes after endovascular treatment of abdominal aortic aneurysms (AAA) involving the renovisceral arteries and to compare outcomes after fenestrated/branched endovascular aortic repair (f/b-EVAR), chimney/periscope EVAR (ch-EVAR), and bailout ch-EVAR.
Methods
A retrospective multicenter study including all patients with AAA involving the renovisceral segment, treated with f/b-EVAR, ch-EVAR, or bailout ch-EVAR, between 1 January 2005 and 30 June 2015, in three Swedish vascular centers. Patient charts were reviewed for data. Renovisceral stent graft patency was assessed on follow-up CT. Mortality was cross-checked against the Swedish Population Registry. Bailout ch-EVAR was defined as a perioperative decision of renovisceral endografting, as the artery was accidentally covered, or as the aneurysm neck sealing zone was considered inadequate.
Results
Of the 99 identified patients (76 men; mean age 74 years (range 58–89 years)), 68 underwent f/b-EVAR, 18 ch-EVAR, and 13 bailout ch-EVAR. Follow-up lasted for a median of 3.2 years (Q1, Q3 (2.1, 4.7 years)). Elective surgery comprised 87.9% (n = 87) of the cases. Six patients died within 30 days, and the 30-day mortality after elective surgery was 4.6% (95% CI, 1.3%–11.4%) overall, 1.6% after f/b-EVAR (95% CI, 0.0%–11.4%), 15.4% after ch-EVAR (95% CI, 1.9%–45.4%), and 10.0% (95% CI, 0.3%–44.5%) after bailout ch-EVAR. During follow-up, there were 16 secondary interventions, of which 75% (n = 12) were performed within six months after the primary intervention. Compared with f/b-EVAR, ch-EVAR was associated with a higher degree of type 1 endoleaks (1.5% vs. 22.2%, P = 0.001) and re-interventions during follow-up (13.2% vs. 33.3%, P = 0.046). The overall assisted target vessel patency was 96.1% (95% CI, 91.7%–98.6%) at one year and 95.2% (95% CI, 89.2%–98.4%) at two years.
Conclusions
Results after EVAR involving endografting of renovisceral arteries from three centers in Sweden with medium volumes are consistent with results previously reported from centers with larger volumes.
Introduction
Since the introduction of endovascular aneurysm repair (EVAR) in the early 1990s,1,2 it has today become a common treatment for standard abdominal aortic aneurysms (AAA). Endovascular treatment of aortic disease involving the renal and visceral arteries could still, however, considered being under development. Novel techniques include fenestrated3–5 and branched stent grafting (f/b-EVAR), 6 as well as solutions with parallel stent graft techniques, i.e. chimneys7,8 and periscopes (ch-EVAR). 9 Reports on the results of these treatments are mostly limited to single-center studies with short-term follow-up. In a systematic literature review by Katsargyris et.al., 21 it was concluded that f/b-EVAR was associated with low operative morbidity and mortality and is a valid treatment option in both low- and high-risk patients. Furthermore, in a recent propensity-matched comparison, similar early outcomes between f/b-EVAR and open AAA repair of pararenal and paravisceral aortic aneurysms were suggested. 10 Due to limited reports on ch-EVAR, Katsargyris et al. suggested that it should be considered only in acute poor surgical risk patients, as a bailout in case of unintentional renal artery coverage, or in elective poor surgical cases that are not suitable for f/b-EVAR. However, in the PERICLES Registry on ch-EVAR treatment, consisting of 119 patient centers in the United States and 398 in Europe, data suggest comparable results after ch-EVAR as with those previously reported for f/b-EVAR. 11
The aim of this study was to report overall short- and midterm outcome after EVAR involving the renovisceral arteries and to evaluate specific performance of f/b-EVAR, planned ch-EVAR, and ch-EVAR when used as a bailout procedure, from three vascular centers in Sweden, and to compare the results with previous data published from centers with larger volumes.
Methods
All consecutive patients with abdominal aortic aneurysm treated by endovascular means including stenting/endografting of any of the superior mesenteric artery, celiac artery, or the renal artery/ies (i.e. renovisceral arteries), between January 2005 and June 2015 in three vascular surgical referral centers in central Sweden (Stockholm South General Hospital, Karolinska University Hospital in Stockholm, and Uppsala University Hospital) were included. The volume of AAA repairs in each center was 65, 92, respectively, 45 in 2017, of which 60% to 83% were endovascular repairs. 12 The three centers could, thus, be defined as medium volume centers. 13 Patients were excluded if the aneurysm was a result of infection (n = 2) or dissection (n = 2), where the indication for intervention was endoleakage after a previous EVAR procedure (n = 6) or pseudo-aneurysm (n = 1), if the disease also involved the thoracic aorta (n = 32), if the Nellix Endo Vascular Aneurysm Sealing System was used (n = 3 (as this system was based on a different technique not comparable with standard stent grafts)) or if the patient previously had undergone renovisceral endografting as part of EVAR (n = 2), and if the stent grafts were custom-made by the physician in the operating room (n = 2). Data were collected retrospectively through circumstantial evaluation of the electronic patient records including computerized tomography angiography (CTA) imaging by two or more clinical specialists (vascular surgeons or interventional radiologists) in each of the three respective centers included. Assessment of AAA morphology, and target artery orientation, was essential for pre-operative planning and choice of treatment method. Generally, indication for treatment was an aneurysm size ≥55 mm, symptomatic disease, signs of threatening rupture, or rupture. In some patients with complex aortic anatomy and comorbid diseases where per- and postoperative risks were considered higher than the risk of rupture, a larger diameter was set as a threshold for treatment. Procedures were performed on a routine basis in angio suites, or in hybrid surgical theatres, with fixed imaging systems (Siemens or Philips). General anesthesia was used in a majority of the cases. Bilateral access through femoral arteries was generally performed; otherwise, additional access was gained through cut down to subclavian/axillary arteries, or by puncture of the brachial artery. Digital subtraction angiography with a calibrated catheter was performed in all patients. Overlay of CTA images in real-life fluoroscopy images (fusion-technique) was used in most cases from 2014 and forward. Subgroups were created, divided by emergency grade of the intervention (elective or emergent (acute/urgent (treatment within seven days))), and by type of intervention performed (f/b-EVAR, ch-EVAR, or bailout ch-EVAR). Bailout ch-EVAR was defined as a ch-EVAR stent graft were renovisceral endografting was not pre-operatively planned but PER-OPERATIVELY DECIDED TO BE performed either as the renovisceral arteries were accidentally covered during deployment of the main body graft, or as the aneurysm neck sealing zone was considered inadequate.
A target vessel was defined according to a report on fenestrated EVAR by Scurr and McWilliams, 14 as a “vessel potentially covered by the stent graft if not for a deliberate mechanism of preservation”. The primary endpoint was overall survival, while secondary endpoints were clinical complications, renovisceral stent graft patency, and secondary intervention. Primary patency was defined as freedom from loss of a target vessel and secondary intervention, while assisted patency was defined as freedom from loss of a target vessel, regardless if the patient underwent secondary intervention or not. Loss of a target vessel was defined as a complete occlusion of the target vessel main stem. The revised Lund-Malmö equations were used to estimate glomerular filtration rate (eGFR) based on overall creatinine levels. 15 The study was approved by the Regional Ethical Review Board, Stockholm, Sweden.
Follow-up
All in-hospital events were collected through patient records. After discharge, all patients entered a specified follow-up schedule, with clinical and laboratory examination, including CTAs. Post-operative CTA controls were performed after four to eight weeks, at one year, and annually thereafter, unless the scans did not mandate a more frequent follow-up schedule, or secondary intervention.
Statistical analysis
Characteristics and complications are reported in absolute numbers and proportions. Differences in characteristics and postoperative complications between groups were analyzed with the Pearson Chi-square test for proportion, Student’s t-test for means, and Wilcoxon rank sum test for skewed or non-normally distributed continuous data. Survival and secondary intervention were both subject to Kaplan–Meier analysis and log rank comparison among elective cases. In Kaplan–Meier analyses, follow-up for survival was terminated four years after date of index intervention, as the number of patients after four years was less than 10. In analyses of secondary intervention, follow-up time was censored at the date of last CTA control. Standard errors in percent were calculated using the Greenwood formula.
Statistical significance of difference in proportion of patent renovisceral arteries at 30 days, 1 year, and 2 years, between groups, was obtained by the use of Pearson Chi square test. Statistical analyses were performed with Stata version 12.1 (StataCorp, TX, USA). All statistical tests were two-sided, and P values <0.05 were considered statistically significant.
Results
During the 10 years of inclusion, 149 patients underwent endografting/stenting of any renovisceral artery as part of an endovascular treatment for AAA, of whom 99 were eligible for inclusion in this study (n = 68 f/bEVAR, n = 18 ch-EVAR, respectively, n = 13 with bailout ch-EVAR). Baseline characteristics among all patients, and by type of intervention, are displayed in Table 1. All f/b-EVAR patients were treated with stent grafts produced by a single manufacturer (Cook Medical, Bloomington, IN, USA), while ch-EVAR patients were treated with stent grafts produced by three different manufacturers (Cook Medical, Bloomington, IN, USA (ch-EVAR, n = 5; bailout ch-EVAR, n = 9); Medtronic, Dublin, Ireland (ch-EVAR, n = 12; bailout ch-EVAR, n = 2); and W. L. Gore & Associates, Flagstaff, AZ (ch-EVAR, n = 1; bailout ch-EVAR, n = 2)). Over the duration of inclusion, there was a trend towards an increasing number of target vessels when comparing the first third of treated patients with the last (median before 2011, 2 vs. 3 after December 2013; P = 0.023; Z=–2.27). ch-EVAR was more commonly used during acute or urgent surgery, while f/b-EVAR treatment more often involved endografting/stenting of the celiac trunk and the superior mesenteric artery. The most common stent for renovisceral artery endografting was balloon-expandable covered stents through femoral access during f/b-EVAR and subclavian/brachial access during ch-EVAR, respectively, and balloon-expandable bare stents through femoral access during bailout ch-EVAR.
Baseline characteristics of the 99 patients, in total and by type of intervention.
Note: Values are means (SD) unless otherwise indicated. Numbers may not add up due to missing values.
ASA-class: American Society of Anesthesiologists Physical Status Class; BMI: body mass index; CA: celiac artery; eGFR: glomerular filtration rate estimated with the revised Lund-Malmo equation; EVAR: endovascular aortic repair (f/b-EVAR, fenestrated/branched; ch-EVAR, chimney/periscope); Q1 and Q3, quartile 1 and 3; SD: standard deviation; SMA: superior mesenteric artery.
P-values were obtained with Pearson Chi-square test for differences in proportion, Students t-test for differences in age and maximal aortic diameter, and Wilcoxon rank sum test for difference in length of hospital stay.
aP vs. ch-EVAR <0.05.
bP vs. f/b-EVAR <0.05.
cBrachial puncture site also included.
Per- and post-operative mortality
The rates of per- and postoperative mortality and complications are displayed in Table 2. Of those who underwent an elective intervention, four (4.5% (95% CI, 1.3%–11.4%)) died within 30 days: one related to myocardial infarction, one multi organ failure, and two due to bleeding from heparin-induced thrombocytopenia. Corresponding 30-day mortality for emergent repair was 16.7% (n = 2; one multiorgan failure after a ruptured AAA and one myocardial infarction (95% CI, 2.1%–48.4%)). The 30-day mortality rate seemed to be lower among those treated with f/b-EVAR (1.5% (95% CI, 0.1%–7.9%)) compared with those treated with planned ch-EVAR (16.7% (95% CI, 3.6%–41.4%; P = 0.006)) and bailout ch-EVAR (15.4% (95% CI, 1.9%–45.4%; P = 0.015)), while no difference was observed between the two ch-EVAR groups (P = 0.92). The 30-day mortality rates among those treated in an elective setting were 1.5% (95% CI, 0.1% – 8.4%) after f/b-EVAR (n = 1/64), 15.4% (95% CI, 1.9%–45.4%) after ch-EVAR (n = 2/13), respectively, and 10% (95% CI, 0.3%–44.5%) after bailout ch-EVAR (n = 1/10). Corresponding rates in the acute/urgent setting were 0% (95% CI, 0–60%) after f/b-EVAR (n = 0/4), 20% (95% CI, 0.5%–71.6%) after ch-EVAR (n = 1/5), respectively, and 33.3% (95% CI, 0.8% – 90.6%) after bailout ch-EVAR (n = 1/3). There was no indication of a difference in 30-day mortality rate during study inclusion (data not shown).
Postoperative complications in the 99 patients, in total and by type of intervention.
EVAR: endovascular aortic repair (f/b-EVAR, fenestrated/branched; ch-EVAR, chimney/periscope); Q1 and Q3: quartile 1 and 3. P-values were obtained with Pearson Chi-square test for differences in proportion, and Wilcoxon rank sum test for difference in length of hospital stay. There were no statistically significant differences between planned ch-EVAR and bailout ch-EVAR.
aP vs. f/b-EVAR <0.05.
bRenal complication includes partial and total infarction.
cInfection includes urinary tract infection, wound infection, septic shock, pneumonia, graft infection, and other infection.
dOther bleeding includes gastric ulcer, hematoma, and hematuria.
eThromboembolic event includes embolectomy and graft thrombus.
fOther complications include gluteal claudication, leg compartment, pseudo aneurysm, and multiorgan failure.
The one-year mortality was 11.5% (n = 10; 95% CI, 5.6% – 20.1%) in elective cases, respectively, and 25% (n = 3; 95% CI, 5.5% – 57.2%) in emergent cases. Figure 1 shows the Kaplan–Meier plot of survival the first two years after elective primary intervention among all patients. In total, 34 deaths were observed during follow-up (median survival time, 3.2 (interquartile range, 2.1 to 4.7) years).
Freedom from death of all causes after elective complex endovascular abdominal aortic aneurysm repair involving the renovisceral arteries performed in three Swedish vascular centers. The standard error at 48 months was 8.2%.
Other post-operative complications
In total, 45.5% (n = 46) experienced any of the registered complications within 30 days. There was no statistically significant difference in overall complication rate between elective and acute/urgent surgery (43.7% vs. 66.7%, respectively; P = 0.13). However, the rate of any complications seemed to be higher in the ch-EVAR group (77.8%) compared with the f/b-EVAR group (35.3%; P = 0.001), but not compared with the bailout ch-EVAR group (61.5%; P = 0.08). It was more common with puncture site complications where cut-down was necessary to stop the bleeding after ch-EVAR compared with f/b-EVAR, and the median admission to hospital was longer. Some of the more serious adverse events, such as transient spinal ischemia, stroke/TIA, myocardial infarction, retroperitoneal bleeding, and thromboembolic events were only observed in the f/b-EVAR group. During follow-up, it was more common with type 1 endoleaks after ch-EVAR (22.2%, n = 4), compared with f/b-EVAR (1.4%, n = 1; P = 0.001).
Target vessel patency and secondary interventions
In the 99 patients, 167 renovisceral arteries were treated using f/b-EVAR (129 renals, 31 SMA, and 7 CA), 33 with ch-EVAR (28 renals, 5 SMA), respectively, and 15 renal arteries with bailout ch-EVAR. Table 3 displays the target vessel patency. The overall primary patency at two years of follow-up was 82.9%, while the assisted patency was 95.2%. There were no statistically significant differences in primary patency between the three different types of intervention. The assisted patency differed between the groups and was higher in the f/b-EVAR group compared with the ch-EVAR group at 30 days/in-hospital (98.1% vs. 88.5%, respectively; P = 0.010), and during follow-up >2 years (96.4% vs. 78.6%, respectively; P = 0.010). There was no statistically significant difference in primary or assisted patency comparing bailout ch-EVAR with ch-EVAR, or f/b-EVAR.
Patency during follow-up, per renovisceral artery intervened with (n = 215 in the 99 patients).
Note: Primary patency was defined as freedom from loss of a target vessel and secondary intervention, while assisted patency was defined as freedom from loss of a target vessel, regardless whether secondary intervention was performed or not. EVAR indicates endovascular aortic repair (f/b-EVAR, fenestrated/branched; ch-EVAR, chimney/periscope). P-values were obtained with Pearson Chi-square test for differences in proportion of patent target arteries at the given time point. There were no statistically significant differences between planned ch-EVAR and bailout ch-EVAR.
aProportion among those alive at the end of each given time-point or end of follow-up.
bMean follow-up until last CT-control among those alive after two years was 3.6 (±1.6) years.
cP vs. f/b-EVAR <0.05.
Overall, 16 patients underwent secondary interventions during follow-up. Indication for re-intervention was stenosis or occlusion of the target vessels (n = 8), endoleak (n = 7), and one due to bleeding. During follow-up, two patients experienced stent graft migration: one f/b-EVAR patient who never underwent a re-intervention and one ch-EVAR patient who underwent three re-interventions due to type 1a endoleak. The majority of secondary interventions (75%, n = 12 out of 16) were performed during the first six months after primary intervention, although some were observed up to three years postoperatively. Displayed in Figure 2 is the Kaplan–Meier plot of probability of secondary intervention after elective primary intervention.
Probability of secondary intervention after elective complex endovascular abdominal aortic aneurysm repair involving the renovisceral arteries performed in three Swedish vascular centers. The standard error at 48 months was 5.3%.
The proportion of secondary interventions over follow-up was higher after ch-EVAR compared with f/b-EVAR (33.3% vs. 13.2%, respectively; P = 0.046).
Discussion
In this retrospective tri-center study of 99 patients treated by endovascular means for AAA involving renovisceral arteries, the 30-day mortality and target vessel patency were comparable with previously published results reported from centers with larger volumes.
Strengths of the current study involve the inclusion of patients from three separate centers of vascular surgery in Sweden, and the comprehensive chart review of pre-operative characteristics and post-operative complications. Furthermore, due to the national population register in Sweden, there was no loss to follow-up.
Per- and post-operative mortality
The 30-day mortality rate of 1.4% in the f/b-EVAR group in this study is well comparable to the most recent reports of fenestrated and branched endovascular repair for complex aortic aneurysms ranging from 1% to 8%.4,5,16–19 The 30-day mortality rate after planned ch-EVAR was 16.7%, respectively, and 15.4% for bailout ch-EVAR, which is high compared with the 4% (range 0% to 10%) reported in a recent systematic review of 17 studies involving 517 patients with renovisceral arteries stented with ch-EVAR. 20 The 95% confidence intervals of ch-EVAR treatment did, however, range from 1.9% to 45.4% in this material due to the small patient material. Given the retrospective design of this study, and the heterogeneity of patients in the three subgroups of treatment, observed differences are most likely explained by differences in patient characteristics and emergency grade of the procedure, rather than the techniques themselves. Among electively treated cases, no significant difference in mortality rate was observed in this study. Furthermore, the reason for choosing one endovascular method over the other was not available, and since delivery-time for custom-made stent grafts for f/b-EVAR could be up to >12 weeks, it is likely that patients were selected for ch-EVAR since treatment needed to be expedited and custom-made stent grafts could not be awaited. Moreover, the experience of each respective center was relatively small in comparison with large volume centers, and procedures were most likely performed by multiple operators, which further may have increased the heterogeneity between groups.
Other post-operative complications and patency
We observed a relatively high degree of per- and postoperative complications in general, while ch-EVAR was associated with a higher rate of complications than f/b-EVAR. The rate of complications did not differ significantly between elective and acute/urgent treatment. It was more common with puncture site complications where cut-down was necessary to stop the bleeding after ch-EVAR compared with f/b-EVAR, while the median admission to hospital was longer. However, some of the more serious adverse events, such as transient spinal ischemia, stroke/TIA, myocardial infarction, retroperitoneal bleeding, and thromboembolic events were only observed in the f/b-EVAR group.
In a study by Roy et al, retrospectively investigating 173 FEVAR patients, 27.2% had a post-operative complication which prolonged hospital stay, whereas the proportion of patients with any of the registered complications in this material was 34.2%. 5 As previously mentioned, it is hard to compare these studies due to the heterogeneous patient material. Compared with ch-EVAR, f/b-EVAR has previously been associated with lower perioperative mortality, lower risk of chronic dialysis, and a lower incidence of endoleak type 1. 21 We also observed an increased risk of type 1a endoleak, and an increased risk of secondary intervention, after ch-EVAR, most likely explained by leak through the “gutters”.
The assisted patency observed in this study corresponds to the 90% to 97% previously reported for f/b-EVAR,5,17,22,23 respectively, the 85% to 100% reported for renovisceral ch-EVAR.11,20,24 A possible explanation for the lower assisted patency observed after ch-EVAR might be the inaccessibility for revascularization from subclavian/brachial access, resulting in a postponed decision to re-intervene on these patients. Given that endovascular treatment of complex aortic disease is still evolving, learning curves and technical advancements may have affected the outcomes over the 10-year inclusion period. We did, however, not observe such trends when stratifying the analyses over follow-up time, despite an increase in the number of vessels treated over time. Yet again, the heterogeneity and sample sizes in this material do not allow for proper comparisons between these three groups, and results should be interpreted with caution.
In conclusion, the results after f/b-EVAR, ch-EVAR and bailout ch-EVAR from three centers in Sweden with medium volumes are consistent with the results previously reported from higher volume centers.
Footnotes
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.