If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
The chimney graft (CG) technique was introduced to rescue accidentally covered aortic branches during aortic endovascular repair. It extends the sealing zone. There is concern about “gutter” type I endoleak (EL-I) and about the durability of CGs. The aim of the present report was to analyze the rapidly increasing existing data.
Methods
A search was performed (PRISMA criteria) for all studies of visceral and thoracic/arch chimney grafts. Technical and clinical details and outcome were assessed.
Results
The present review includes 831 patients who underwent EVAR/TEVAR (endovascular aneurysm repair/thoracic endovascular aneurysm repair) with one or more chimney, periscope, or sandwich grafts. For aortic visceral vessels 517 patients received 911 visceral CGs and 314 patients received 364 arch CGs. Most procedures (81% visceral and 69% arch CGs) were elective. Thirty day mortality was 4% for both groups. The rate of early EL-I was 13% (visceral CGs) and 11% (arch CGs). Most EL-I were handled conservatively (observation: 70% for visceral CG and 45% for arch CG). Early CG patency was high (97–99%) and remained high during follow up (median 17 months). Late (after 30 days) EL-I was reported in nine visceral (2%) and 12 arch (4%) CG cases. Few other late complications were reported, but those losing a kidney at the initial repair seemed to have a high risk of requiring permanent hemodialysis.
Conclusion
Increasing amounts of data support the benefit of visceral and arch chimney graft techniques. In particular, the low early mortality and complication rates and high long-term patency seem advantageous; however, the majority of cases have been treated electively, and there is a high risk of bias in all studies. Mid- to long-term data suggest few late complications, except in cases where one renal artery was sacrificed. The CG technique is valuable for complex urgent patients and needs further documentation for other patient groups.
A chimney stent graft preserves vital aortic branches in aortic endovascular repair. The so called gutter endoleak causing a type I endoleak has been a potential burden. There are excellent reviews on this topic; however, during the last year the case series have been several fold larger, including over 800 patients, and with a longer follow up. Whether urgent or semi-urgent patients, not fit for open surgery and unsuitable for fenestrated or branched stent grafts, should be offered this chimney technique is an important question. What can be learned from this increased knowledge? Is the chimney stent graft technique really useful or hazardous?
Introduction
The principle of the chimney graft (CG) technique is similar for visceral and arch branches of the aorta. A bare or covered stent is deployed into a vital aortic side branch parallel to the main aortic stent graft. Thereby, the sealing zone of the aortic stent graft can be extended beyond the origin of the vital side branch. CGs are applied in many types of lesions such as aneurysms, dissections, transections, occlusive disease, and following accidental overstenting of an aortic branch. The first known CG was implanted in 1999 (Fig. 1) to rescue an overstented renal artery (Lönn L, Malina M. Personal communication regarding the first Swedish chimney stent graft used as bailout for accidental renal artery coverage, 1999). That CG remained patent for 13 years, the lifespan of the patient. The first case to be reported in the literature appeared in 2003.
The main difference between thoracic/arch and visceral CGs is that the thoracic CGs can be implanted from the periphery, which is not practical for the visceral branches and has only been applied occasionally for the superior mesenteric artery.
Figure 1Computed tomography follow up of the first known chimney graft (CG). The CG was inserted in 1999 as a bailout procedure to rescue an accidentally overstented right renal artery during endovascular aneurysm repair for a juxtarenal aneurysm. The CG remained patent and the aneurysm was excluded until the death of the patient 13 years post-operatively. (Courtesy of L. Lönn and M. Malina).
Fenestrated and branched stent grafts have to date, not become universally applicable because the grafts are mostly customized and take a long time to manufacture (1–2 months). They also remain costly
and may not fit in tortuous aneurysms or in patients with compromised access. Urgent cases that need to be treated without delay, who are not fit for open repair, and that have an inadequate sealing zone for standard stent graft repair may have no other therapeutic alternative than the CG technique.
Several modifications of the CG technique have been developed. The periscope graft is a distally oriented CG that allows retrograde flow up to an aortic side branch. The periscope graft was first used for preservation of visceral arteries during repair of a ruptured thoraco-abdominal aortic aneurysm,
but it has also been applied to preserve the left subclavian artery in aortic arch repair. The sandwich technique comprises a tubular stent graft that is deployed to create an artificial neck, which serves to implant the CGs. The primary aortic stent graft is subsequently extended by another piece to exclude the entire aneurysm. Thereby, the CGs gets “sandwiched” between the two aortic stent grafts. This technique was initially used to preserve the internal iliac arteries and later for the visceral arteries in thoraco-abdominal aneurysms.
Sandwich technique for aortoiliac aneurysms extending to the internal iliac artery or isolated common/internal iliac artery aneurysms: a new endovascular approach to preserve pelvic circulation.
Maybe the most important reason for this is that CGs are available off the shelf and are less prohibitively expensive than fenestrated and branched stent grafts.
The main concern about CGs has been the risk of EL-I due to the so called gutters. The gutters are channels that may appear between the CG and the main aortic stent graft. Such channels are difficult to seal completely with the current technique and leakage is best avoided by applying long CGs.
Several reports on the CG technique have been published during the last year with mid- to long-term follow up.
Therefore, an updated systematic review and critical analysis of current data for chimney, periscope, and sandwich stent grafts of both visceral and arch branches is warranted.
Methods
Search strategy
This review of the existing reports on CGs (including snorkel, sandwich, and periscope techniques) in endovascular repair of AAAs and TAAs (see electronic supplementary material figure 4) was conducted according to the PRISMA criteria for systematic reviews.
The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate health care interventions: explanation and elaboration.
An electronic search was made using PubMed/MEDLINE, Science Direct, Scopus, Ovid, Embase, and the Cochrane Library. The search included all papers published up to April 2015 (n = 847). Medical subject heading (MeSH) terms used included “chimney stent grafts”, “chimney graft”, “chimney”, “periscope”, “sandwich”, “snorkel”, “thoracic aortic branches revascularization”, “visceral arteries revascularization”, “abdominal”, “thoracic”, “arch”, “thoracic arch aneurysms”, and “thoraco-abdominal”, “thoraco-abdominal types II, III, and IV aneurysm” “juxta- or supra-renal aneurysm”, “aortic repair”, “endovascular repair”, “EVAR”, “TEVAR”.
Study selection
Articles were selected based on their title and abstract (n = 220). Two independent reviewers (A.B.G., B.L.) critically evaluated the papers for eligibility before inclusion. The included articles were scrutinized to identify further relevant studies (n = 5). Articles were excluded if they did not describe a case series, the type of CG technique employed, demographic data, patient outcomes/follow up or if the data were not original (n = 125). Only the most recent report from each center was accepted (excluding 13 series, to avoid duplicate publications).
Extraction of data
The following data were extracted by two reviewers (A.B.G., B.L.) from each report: number of patients, mean age, sex, number of arch or visceral CGs, urgency of treatment, type of aortic lesion treated, aortic aneurysm morphology if applicable (arch, thoracic, thoraco-abdominal, or abdominal), type of stent and stent graft, mean length of follow up, early and late CG patency (based on CTA), early and late type I endoleak (EL-I) rate, early and late type III endoleak, access related complications, 30 day complications and mortality rates, development of end stage chronic renal impairment and permanent dialysis. Early patency and early endoleak was defined as those diagnosed within the first 30 post-operative days.
Quality assessment of included studies
All of the included studies carried a high risk of bias in study design, selection of patients and methodology.
Innovative application of available stent grafts in Japan in aortic aneurysm treatment—significance of innovative debranching and chimney method and coil embolization procedure.
A percutaneous technique for preservation of arch branch patency during thoracic endovascular aortic repair (TEVAR): retrograde catheterization and stenting.
There were no randomized or really controlled studies. None was blinded, or used independent evaluation of a patient's events. The guidelines for reporting data both on aortic and thoracic endovascular therapy did not consider the CG technique and did not include chimney related problems. For example, how should patency of CG be reported? How should EL-I be diagnosed? Are sealed EL-I still risky?
In the quality assessment (A.B.G., B.L.) (Table VI in electronic supplementary material) none of the studies had a high scientific quality and, additionally, the risk of publication bias was high in an area of new and not really established techniques. The risk that catastrophic cases or even case series have not been published must be considered. Thus, in summary, the scientific value of reported studies was low and heterogeneity and bias high, and this must be considered in the analysis of the findings.
Statistical analysis
Only descriptive data are presented, since this PRISMA systematic review is not comparative.
Results
Included studies and patients
A total of 831 visceral and arch CG treated patients were identified. Of these, 17 case series and 38 case presentations with 911 visceral CGs and endovascular aortic stent graft repair (Table 1, Table 2) were identified in 517 patients. Thoracic arch CGs appeared in 18 case series and 23 case reports with 314 patients and 364 CGs for thoracic arch aortic branches (Table 3, Table 4). Table 5 shows the reason/etiology for CG treatment. Tables VII–X in the electronic supplementary material provide data and references, including possible duplicate reports and individual case reports.
Note. Percentage in parenthesis. Duplicate series from the same institution excluded. No. = number; CGs = chimney, periscope, and sandwich grafts; EL-I = type 1endoleak; M = mortality, ? = uncertain data.
Note. Percentage in parenthesis; FU = follow up; CGs = chimney grafts; M = mortality; LRM = lesion related mortality; ms = months; CRI = chronic renal impairment; D = dialysis; EL-I = type I endoleak; EL-III = type III endoleak; ? = uncertain data.
Note. Percentage in parenthesis. Duplicate series from the same institution excluded. CGs = chimney and periscope grafts; EL-I = type I endoleak; M = mortality.
Note. Percentage in parenthesis. Duplicate series from the same institution excluded. FU = follow up; CGs = chimney and periscope grafts); M = mortality; LRM = lesion related mortality; ms = months; CRI = chronic renal impairment; D = dialysis; EL-I = type I endoleak; EL-II = type II endoleak; SCI = spinal cord ischemia; ? = uncertain data.
Urgent repair comprised 96 (19%) of the visceral CGs. This is different from thoracic arch CGs where 98 repairs (31%) were urgent. There is a wide variation in urgent/elective patients between studies (Table 1, Table 3).
Early (30 day) results
Early mortality
In the series the reported 30 day mortality was 19 of 517 (4%) for visceral CG patients (Table 1; range 0–10%) and 14 of 314 (4%) of thoracic arch CG patients (Table 3; range 0–17%).
Technical failures and early patency
Fifteen of 911 visceral CGs (1%) and three of 364 thoracic CGs (1%) were reported as primary CG technical failures. A large proportion of failures were handled with extra-anatomic bypass, although some were “asymptomatic”. The failed branches were inconsistently reported.
The overall early patency rate of the CGs was 97–99% (Table 2, Table 3). Eleven early post-operative occlusions of visceral CGs were reported. One of the three occlusions of the superior mesenteric artery was fatal and two were handled with extra-anatomic bypass. Two of four renal occlusions were symptomatic and were revascularized by open bypass. Another four renal occlusions were poorly defined. The only early “asymptomatic” occlusion of a thoracic/arch CG occurred in the left subclavian artery.
Early EL-I
EL-I was reported in 67 of 517 patients with a visceral CG (13%, range 0–33%) and in 33 of 314 patients with a thoracic CG (11%, range 0–44%) (Table 2, Table 3).
The majority of early EL-I were “only” observed and many of them were reported to seal. Seventy percent of the visceral EL-I were observed (n = 47) and most sealed spontaneously (n = 29?). However, there were two early fatal EL-I related courses (both with ruptured AAA) in this group. Twenty visceral EL-I required treatment mainly consisting of coil embolization/glue (n = 11), re-stenting of CGs (n = 2), and main stent graft elongation (n = 7).
Fifteen (45%) of thoracic arch EL-I were observed and seven were reported to seal. Nine (27%) thoracic EL-I were embolized or glued and five were treated by extension of the aortic stent graft.
Four abdominal cases required open conversion. No abdominal conversion was done for EL-I. Two conversions were due to suprarenal aneurysm expansion: one as a bridging procedure for infection, and one attempted to reduce thrombus volume.
Three thoracic cases required open conversion, all for EL-I (Fig. 2). A reasonable division between elective and urgent reconstructions and the risk of EL-I is not possible (see Fig. 3).
Figure 2Bridging procedure for a mycotic aneurysm. (A) Symptomatic mycotic arch aneurysm involving the brachiocephalic trunk in a septic patient with a history of several malignancies. (B) Semi-urgent thoracic endovascular aneurysm repair with chimneys to brachiocephalic trunk (BT) and left common carotid artery (LCCA) was performed but a minor type I endoleak with faint opacification of the sac persisted. (C) Post-operatively the aneurysm expanded and the leak increased. (D) Embolization of the chimney gutters with thrombin and onyx sealed the leak. The patient improved, infectious signs resolved but the endoleak recurred. Her improved general condition now allowed successful open arch reconstruction. The patient subsequently died from disseminated malignancy with brain metastases.
Figure 3(A) Computed tomography angiography of a 63 year old woman with pneumococcal pneumonia, sepsis, multifocal abscesses and rapidly expanding pseudoaneurysms in the thoraco-abdominal aorta in spite of drainage and antibiotics. (B–D) She was treated endovascularly with the sandwich technique and chimney grafts to the superior mesenteric artery (SMA), right renal artery (RRA) and left renal artery (LRA). The celiac trunk was embolized. The post-operative course was complicated by acalculous cholecystitis but no other complications were noted. (E) The mycotic pseudoaneurysms remain excluded with no signs of infection after 4 years follow up. The visceral chimney grafts (red) are “sandwiched” between the proximal (blue) and distal (yellow) aortic stent grafts.
Access/procedure related complications were inconsistently reported (Table 1, Table 3). Six to 8% of the patients with both visceral and thoracic CGs were reported to have complications that may relate to access/procedure problems (Table 1, Table 3). Such complications include intra-operative stroke (n = 6 for visceral CGs, 1 fatal, n = 6 for arch CGs, 1 fatal), iliac artery problems, bleeding, or pseudoaneurysm at the puncture site. A complication specific to the CG technique is attributable to the frequent use of stiff guidewires with a short soft tip. This type of guidewire is often required to gain sufficient stability for insertion of the CG. Such guidewires are at risk of perforating or dissecting distal branches of the visceral arteries, which may cause either arterial hemorrhage or parenchymal infarction. There were reports on eight retroperitoneal hematomas (2 fatal), four renal hematomas, three intestinal branch bleeding, two intestinal ischemia (1 fatal), and five renal ischemia that could be related to guidewire injuries.
Follow up results
The follow up ranged from 1–27 months (median 17 months) for visceral and 1–28 months (median 16 months) for thoracic CGs (Table 2, Table 4).
Late mortality
The reported overall mortality during follow up differs considerably between the studies and the median is reported to be 14% for visceral (range 0–57%) and 11% for thoracic CG (range 5–68%) treated patients. Procedure related death is reported to be 6–7% for both visceral and thoracic cases.
Late CG patency during follow up
The data are somewhat uncertain since the number of patients at risk is often not clearly stated. During follow up, 24 visceral CGs occluded among an estimated 905 visceral CGs at risk (3%) (Table 2). Two celiac trunk occlusions required emergent bypass, one of them had a fatal outcome. Two of the four occlusions of a superior mesenteric artery were asymptomatic, one was fatal, and data are missing on the fourth case. Two of the 15 renal occlusions were symptomatic, four asymptomatic, and for nine patients data are not clearly reported. Additionally, four occlusions were not defined. Many renal artery occlusions were identified at routine follow up. It proved impossible to extract data on the incidence of developing chronic renal failure requiring dialysis from late renal occlusion in these series (Table 2).
Eight late occlusions occurred in thoracic arch CGs (2%) (Table 4). Two left common carotid artery occlusions required extra-anatomic bypass while the six occlusions of the left subclavian artery were asymptomatic. Overall patency at last follow up was 97%.
Chronic renal impairment
The definition of chronic renal impairment (CRI) differs somewhat in the studies, but a 20% reduction of glomerular filtration rate was commonly used. CRI was reported among 78 patients with a visceral CG (14%). Twenty-two of them (4%) required permanent dialysis (Table 3). Nine patients with a thoracic arch CG (3%) were also reported to have developed CRI and two of them required hemodialysis. A few studies reported on the need to sacrifice one renal artery at primary repair. Twelve of 21 (57%) of the author's patients with a sacrificed kidney required permanent dialysis. Eleven of these were urgent cases.
Late type I or III endoleak
Late EL-I was reported in nine patients (1 fatal) with visceral (2%) and 12 patients with thoracic arch (4%) CGs (Table 2, Table 4). Additionally, there was one EL-III in each group.
Sac dynamics at follow up
A total of 309 of the 435 visceral CG related aneurysms in case series reported sac shrinkage. Shrinkage or stable aneurysm size was seen in 90% (range 71–100%) while sac increase was reported in 11% (range 0–29%). Sac dynamics were not reported in most series with thoracic CGs.
Discussion
The overall aim of a systematic review is to accumulate, evaluate, and report sound scientific information in a predefined process to draw as solid conclusions as possible, ideally raising the level of scientific evidence for the benefit of future patients.
The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate health care interventions: explanation and elaboration.
This process is less difficult when large and well designed studies can be aggregated. Unfortunately, this is not the case with the present systematic review. The fact that well designed studies on CGs were lacking is not surprising given the low number and heterogeneity in each series of these complex cases. Inconsistent definitions, variation in diagnosis and treatment of endoleaks, sac dynamics, graft occlusions, etc. illustrate some of the many limiting factors of this review. Specific reporting guidelines do not exist for CGs. Therefore, all conclusions must be viewed with caution. Nevertheless, there was a compelling need for this review conducted in a predefined and transparent manner.
Several CG reports have been published recently and the present review is based on a collective experience of more than 800 patients with CGs. Reported 30 day mortality is low, patency of CGs is high, and the risk of EL-I is acceptable. Also, the overall mortality during follow up is low in spite of the fact that most patients were categorized as unfit for open repair.
Standard endovascular aneurysm repair (EVAR) is an expanding and safe alternative to open surgery.
International trends in patient selection for elective endovascular aneurysm repair: sicker patients with safer anatomy leading to improved 1 year survival.
More than 60% of elective aneurysm repairs are handled endovascularly in many countries. Procedure related mortality is lower (1.4% for EVAR vs. 4.2% for open repair) but late all cause mortality does not differ (15.8% for EVAR repair vs. 17% for open repair at 4 years).
Both selection bias and the natural risks of the many comorbidities of aneurysmal patients may contribute to this. The aneurysm per se indicates a poor state of health, which affects the patient's long-term outcome. Patients of increasingly great age and with more severe comorbidities were accepted for EVAR after the multicenter studies were published.
EVAR trial participants Endovascular aneurysm repair and outcome in patients unfit for open repair of abdominal aortic aneurysm (EVAR trial 2): randomized controlled trial.
and endovascular skills and training are constantly improving. The indications for EVAR should therefore be extended to increasingly complex aneurysms and patients. Yet, a significant proportion of patients cannot be treated with the currently available endovascular devices. This is the main niche for the CG although some investigators advocate that CGs may actually replace complex stent grafts on many occasions.
It is evident that the CG technique needs to be scrutinized continuously as new data emerge. Fenestrated stent grafts are still associated with lower procedural mortality (2.4%), lower risk for chronic dialysis, and lower incidence of EL-I. This suggests that fenestrated stent grafts should be preferred in elective abdominal cases.
But again, the patient cohort comparability is poor, while the differences in outcome are not significant enough to preclude the usefulness of CGs.
The definitions of urgent/emergent/semi-urgent/elective were often inconsistently defined. No strict guidelines exist. Surprisingly, the majority of CGs both visceral
were used in elective procedures. This may explain the low 30 day mortality and also implies that CGs were selected in many cases where more established endovascular techniques would have been applicable, including the use of complex custom made stent grafts.
It is noteworthy that the favorable results of CGs were achieved in patients who had been disqualified for open repair and/or who had a high American Society of Anesthesiologists ASA score with a high frequency of severe comorbidities. The heterogeneity of existing data is evident from the wide range of overall mortality. Overall mortality of patients with visceral CGs has been reported as 0%,
This raises some doubt about the claim of other investigators that patients were not candidates for open repair. Nevertheless, urgent/emergent repairs do stand out as those that may definitively benefit from CGs.
The frequency of early EL-I and the need to treat it were surprising low for both visceral and thoracic CGs. However, the frequency ranges
Innovative application of available stent grafts in Japan in aortic aneurysm treatment—significance of innovative debranching and chimney method and coil embolization procedure.
The larger diameter required for a brachiocephalic chimney is associated with a higher risk of endoleakage and may partly explain these differences. Yet another potential factor is that endoleaks seem less likely in dissections than in aneurysms. In addition, was the diagnosis in some of the studies based on completion angiography while the patient was still anticoagulated, or was it an early post-operative CT examination? Was the CT done with delayed scans? Such issues may explain some of the discrepancies between the studies.
A large proportion of early EL-I was merely observed and often seem to have sealed spontaneously. Existing data do not allow firm conclusions whether these patients are at increased risk of late rupture. The guidelines do not require follow up data on these patients to be reported. Improvements of the guidelines themselves are needed but will certainly be challenging to achieve.
Access and procedure related problems are often neglected. Such complications can be devastating and their current rate of about 10% is not optimal. Most access complications are puncture related and theoretically possible to avoid. The use of the micro puncture technique and puncturing at a compressible site are useful adjuncts to improve safety. Intra-operative anticoagulation is required to avoid thrombosis/embolization but needs to be monitored to minimize the risk of hemorrhage. Gentle instrumentation in the arch is mandatory (stroke rate 1–2% in this review). Similarly, stiff guidewires may be needed for cannulation of visceral arteries but must be used with uttermost care to avoid complications such as pararenal hematomas and disastrous injuries to the intestinal arteries.
CRI has been a matter of concern for stent grafts with suprarenal fixation and for fenestrated stent grafts. However, recent reports have been reassuring and EVAR seems to have at least equal or better results than open repair.
as 29 out of 46 (57%) of the patients with a visceral CG developed CRI. Intentional sacrifice of one renal artery was particularly associated with CRI (n = 23/51 patients, 45%) and half of these patients required permanent dialysis. The high number of emergent cases and pre-existing renal impairment can at least partially explain this. It may be fully justified to sacrifice a kidney in order to save the patient's life in spite of the high rate of CRI. However, based on the findings here and a few other reports, sacrificing a kidney should be seen as a last resort since the risk of CRI and dialysis is high in this group.
One problem for the CG technique is the weak support from manufacturers to develop dedicated covered stents for CGs. Research and investment have focused on branched and fenestrated stent grafts. None of the available covered stents is intended for CGs and neither flexibility, nor radial force, shape, length, etc., are ideal.
In conclusion, the CG technique is an established and definitively useful technique for emergent cases that are not suitable for the current commercially available branched or fenestrated devices. Cost related issues might also support the use of CGs for emergent and elective cases. The rate of EL-I is “acceptable”, lesion related mortality seems low, and patency of CGs is surprisingly high. However, sacrificing a kidney in an emergent situation is associated with a high risk of dialysis. It is the authors' opinion that this technique should be recommended for urgent cases not fit for open surgery and for those unsuitable for the currently available branched or fenestrated stent grafts. A highly restrictive use is recommended for selected elective cases until more data exists.
Conflict of Interest
None.
Funding
None.
Appendix A. Supplementary material
The following are the supplementary data related to this article:
Sandwich technique for aortoiliac aneurysms extending to the internal iliac artery or isolated common/internal iliac artery aneurysms: a new endovascular approach to preserve pelvic circulation.
The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate health care interventions: explanation and elaboration.
Innovative application of available stent grafts in Japan in aortic aneurysm treatment—significance of innovative debranching and chimney method and coil embolization procedure.
A percutaneous technique for preservation of arch branch patency during thoracic endovascular aortic repair (TEVAR): retrograde catheterization and stenting.
International trends in patient selection for elective endovascular aneurysm repair: sicker patients with safer anatomy leading to improved 1 year survival.
To submit a comment for a journal article, please use the space above and note the following:
We will review submitted comments as soon as possible, striving for within two business days.
This forum is intended for constructive dialogue. Comments that are commercial or promotional in nature, pertain to specific medical cases, are not relevant to the article for which they have been submitted, or are otherwise inappropriate will not be posted.
We require that commenters identify themselves with names and affiliations.
00600-0&id=gr1.jpg){kind=link}
00600-0&id=gr2.jpg){kind=link}
00600-0&id=gr3.jpg){kind=link}