INTRODUCTION

Traditional repair of disease of the aortic arch requires sternotomy or thoracotomy, cardiopulmonary bypass (CPB) and circulatory arrest.  These procedures are associated with significant morbidity and mortality even in the modern era of improved surgical techniques. 1  Total endovascular repair, using branched and fenestrated endografts, are not mainstream, are limited to few institutions and the long-term results are uncertain. 2,3  Hybrid interventions require an extra-anatomical bypass to the supra-aortic arch vessels, and the insertion of an endovascular graft during the same surgery or as a second intervention.  These surgeries do not require CPB and may be associated with improved mortality and morbidity. 4,5  There are, however, several aspects, which remain uncertain or controversial:  the durability of repair;  the long-term patency of bypasses to the supra-aortic trunks;  the need to revascularize the left subclavian artery (LSA);  the type of subclavian artery revascularization, carotid subclavian bypass (CSB) vs. subclavian carotid transposition (SCT);  the timing of debranching and endovascular repair, simultaneous or staged; and the length of the landing zones. 

The aim of this work is twofold, to report the results of a cohort study of hybrid interventions with a predefined strategy of a staged endovascular repair and a 30 mm proximal landing zone; and to provide an evidence-based review of the literature on the topic.

PATIENTS AND METHODS

Cohort study 

This is a retrospective cohort study of prospectively collected data of consecutive patients operated upon by a single surgeon (CSC) in tertiary Institutions.  Demographics, indications for intervention, risk factors, complications, operative details and follow-up data were collected and recorded in a database.  Ethics approval for the study was obtained in the institutions involved.  

To evaluate the operative risk of mortality and morbidity, we used the American Society of Anaesthesiologists (ASA) classification, the Revised Cardiac Risk Index (RCRI), and the Gupta Respiratory Risk Calculator (RRC). 6–8

Debranching procedures.  Three types of debranching procedures were performed:  SCT; carotid-carotid subclavian bypass (CCSB); and total debranching of supra-aortic trunks (TDST).  We already described elsewhere our technique of SCT. 9  The CCSB is performed through a horizontal incision at the base of the neck extended over the left supraclavicular region.  The common carotid arteries are exposed below the omohyoid muscles, and the LSA and its branches are dissected after mobilization of the phrenic nerve and transection of the scalenus anterior muscle.  A Dacron® graft of 8 mm in diameter is anastomosed in end-to-side to the right common carotid artery with a ‘T’ configuration and tunneled behind the strap muscles and in front of the trachea.  The left common carotid artery is transected, the distal stump secured with sutures and the proximal end is anastomosed in end-to-side to the superior aspect of the Dacron® graft.  The graft is then tunneled behind the left sternocleidomastoid muscle and the jugular vein and anastomosed to the LSA either in end-to-end or in end-to-side (Figure 1).  The TDST is done through a ministernotomy. A partial occluding clamp is applied to the right antero-lateral aspect of the ascending aorta and an off-the-shelf trifurcated Dacron® graft 12x8x8 mm in diameter is anastomosed in end-to-side.  The brachiocephalic, carotid or subclavian distal anastomoses are then performed in end-to-end.  Occasionally, the one to the LSA is done in end-to-side (Figure 2).

Endovascular technique.  Using a 64-slice computerized tomography scanner and post-processing with the Aquarius workstation v. 3.5, TeraRecon Inc. (San Mateo, California), we obtained centre line measurements of length and diameters of proximal and distal landing zones.  For the total length of aorta to be covered, we used manual readjustments of the center line to the anticipated path of the endovascular stent. 10

Repair was done using Zenith TX2 and Zenith Alpha endovascular thoracic stent grafts.  The general strategy included a proximal and distal landing zone of at least 30 mm, oversizing of 20-25% of the diameter, adventitia-to-adventitia, of the aortic landing zone, and when more than one aortic endovascular graft was planned, we aimed for an overlap of at least 50 mm.  

The thoracic aorta was subdivided in segments according to the SVS reporting standards for thoracic endovascular repair. 11

Endovascular surgery was done under general anaesthesia with transoesophageal echocardiography monitoring, and no adenosine or rapid pacing arrest.  A cerebrospinal fluid (CSF) drainage was inserted in elective cases when we anticipated covering the thoracic aorta distal to the level of T10. 12  All procedures were performed using a C-arm fluoroscopy unit OEM 9900 (GE Health Care Technologies, Salt Lake City, UT) and since 2010, a Ziehm Vision RFD (Ziehm Imaging GmbH, Nurnberg, Deutschand).  Follow up of the endovascular repair was done with computerized tomography at 1 month, at 6 months and yearly thereafter.

Review of the literature

We sought to provide an evidence-based summary of published reports in the English language.  The Medline (National Library of Medicine, Bethesda, MA) database was searched from 1990 to May 2020, using PubMed (National Center for Biotechnology Information, 8600 Rockville Pike, Bethesda, MD 20894) and the search strategy outlined in the Appendix.  We also searched the references of the published relevant and review papers.

All studies reporting on a hybrid procedures and the outcome patency and mortality were included.  Studies reporting on fewer than 10 cases or addressing specifically the treatment of acute type A and type B dissections, pathology of the ascending aorta, aortic coarctation, traumatic disease, thoracoabdominal aneurysms and those involving the celiac artery, fenestrated, branched endograft, and chimney techniques were excluded.  We made an effort to exclude duplicate reporting from the same surgical group.  Finally, studies were excluded if, after retrieving the abstract, the entire publication could not be obtained electronically or after request to the author.

A review of titles and abstracts identified potentially relevant studies, which were then retrieved so that the final decision on relevance was made on the full text.  Studies were reviewed in duplicate (CSC, AA) for relevance and validity (type of study, direction of inquiry - retrospective or prospective -, documentation of demographics, description of interventions, and assessment of outcomes).  Data extraction was done in duplicate (CSC, AA).  Disagreements were resolved by means of consensus (CSC, SC, AA).  

Descriptive data included demographics, indication for surgery, operative techniques, interval between open and endovascular repair, and use of CSF drainage.  The main outcomes were 30-day and long-term mortality.  Secondary outcomes were paraplegia, strokes, cardio-respiratory complications, patency of bypasses, endoleaks, length of follow-up, peripheral nerve and local arterial injuries, and lymph leaks.

Statistical analysis

In the cohort study, categorical variables are expressed as distribution of absolute frequencies, cumulative results and proportions; continuous variables as mean, standard deviation (SD) and 95% Confidence Interval (95% CI).  Contingency tables with Fisher's exact test were performed to evaluate associations between categorical variables.  To determine whether there was a difference between groups, we used ANOVA or Kruskal-Wallis rank test, after checking assumptions.  Post hoc analyses were done using Bonferroni correction.  Survival analysis was calculated from the date of debranching surgery to death or last follow-up and was estimated using the Kaplan-Meier method with Log-Rank test.  A two-tailed P value ≤ 0.05 was considered significant.  Statistical analysis was performed using SAS 9.4 (SAS Institute Inc., Cary, NC, USA.). 

For the review, results were summarized using the generic inverse variance function of OpenMetaAnalyst 13, which allows to weigh the contribution of individual studies according to the sample size.  When a zero value in one cell affected computation, the Haldane correction was applied, replacing the zero value with 0.5. 14  Heterogeneity was evaluated using Cochran Q and I2 test.  In consideration of anticipated clinical and statistical heterogeneity and of the inverse-variance model used, results were summarized using DerSimonian-Laird random effects models. 15  Metanalyses were conducted for all types of hybrid procedures considered together and subgroup analyses for the different types.

RESULTS

Cohort study

Patients and surgical strategy.  From May 2001 to September 2019, 51 patients underwent hybrid reconstructions for disease of the thoracic aorta involving at least one of the supra-aortic trunks.  Thirty-five patients were male (69%).  The average age was 69 ± 10 years (95% CI 66 - 72), and the diameter of the aneurism 6 ± 0.6 cm (95% CI 5.8 - 6.2), without differences between men and women.  Comorbidities included hypertension 88%, smoking habit 73%, chronic obstructive pulmonary disease 22%, previous myocardial infarction (MI) 25%, and diabetes mellitus 22%.  The average serum creatinine level was 1.1 ± 0.3 (95% CI 1.02 - 1.18) (min 0.6 - max 1.9).  Table I describes the results of the ASA class, RCRI and RRC risk indices in the entire population and stratified by type of surgery.  

Indications for surgery were:  41 (80%) degenerative aneurysms; 4 (8%) type B aortic dissections;  2 (4%) congenital aortopathies, one a variant of Ehlers–Danlos disease and one a familial thoracic aneurysm syndrome;  1 (2%), type 1 endoleak after repair of an aneurysm done at another institution;  1 (2%), Kommerell’s aneurysm with dysphagia lusoria;  and 2 (4%) ruptured distal arch aneurysms.

A SCT was performed in 20 patients (39%), a CCSB in 15 (29%), and a TDST in 16 (31%).  In one patient, a dominant left vertebral artery originating from the aortic arch was implanted onto the LSA at the time of a left SCT.  In two patients, the diameter of the landing zone, in Z0 and Z5 respectively, was considered too large for a stable seal.  In both cases, the diameter of the aorta was reduced with a banding procedure.

The overall surgical operating time was 145 ± 73 min (95% CI 125 - 165):  181 ± 82 (95% CI 141 - 221) for the TDST, 122 ± 38 (95% CI 103 - 141) for the CCSB, and 90 ± 35 min (95% CI 73 - 107) for SCT group, respectively.  A statistically significant difference was found between the TDST and the SCT groups (P=0.01).  In patients undergoing TDST the average blood loss was 300 ± 118 ml.  In the other types of debranching procedures, blood losses were trivial.

In two patients with a ruptured arch aneurism, debranching and endovascular procedures were done during the same operation.  In the others, the endovascular repair was done after debranching, at a second surgical sitting.  The interval between the first and the second procedure was 44 ± 41 days (95% CI 33 - 55) without differences between the three types of surgery (P = 0.9).  None of the aneurysms ruptured during the interval between the first and the second surgery.

A CSF drainage was inserted preoperatively in 18 patients in whom we planned a landing zone distal to T10 (35%).  The length of the proximal and distal landing zones were 36 ± 7 mm (95% CI 34 - 38) and 43 ± 7 mm (95% CI 41 - 45), respectively.  The endovascular graft landed distally, in zone Z4 in 32 patients (63%), in Z5 in 19 (37%).

Early mortality and morbidity.  The hospital-stay in the entire cohort was 6 ± 3 days (95% CI 5 - 7):  7 ± 4 (95% CI 5 - 9) for the TDST, 5 ± 1 (95% CI 5 - 6) for the CCSB, and 5 ± 1 (95% CI 5 - 6) days for the SCT group, respectively.  All 16 patients who received a TDST procedure were recovered in intensive care unit (ICU), average stay 3 ± 3 days (95% CI 1.5 - 4.5);  14 patients in the SCT and CCSB groups were recovered in ICU for an average of 1 ± 1 day (95% CI 0.6 - 1.5) because they had a CSF drainage.

One patient in the SCT group died of MI within 30 days of surgery (2%), and two (4%) experienced a non-fatal MI, one after CCSB and one after TDST.  In the TDST group, one patient (2%) suffered a thrombosis of the brachiocephalic vein and a pulmonary embolism; one (2%) a recurrent laryngeal injury, and one (2%) a pneumothorax.  In the SCT group there were two lymphoceles (4%) and two superficial wound infections (4%) treated conservatively.  One patient with a ruptured arch aneurysm, treated with simultaneous debranching and endovascular stenting without CSF drainage, suffered a partial spinal cord ischemia and a CT scan showed multiple inner brain emboli.  There were no patients intubated for more than 24 hours in the entire cohort. 

Using the 2010 Reporting Standards for Thoracic Endovascular Aortic Repair (TEVAR) guidelines, there were no primary endovascular technical failures in the entire series. 11  Stent-grafts were deployed successfully in all patients, with no surgical conversions.  The primary success rate was 100%. 

Late mortality and morbidity.  Patients were followed for a mean 63 ± 32 months (95% CI 54 - 72).  Figure 3 and Figure 4 report the cumulative survival estimates for the entire cohort of patients and for the three subgroups.  Long term survival appears to be better for patients undergoing SCT and CCSB compared with those undergoing TDST, but the difference did not reach statistical difference.  

On follow up there was no graft failure over 98 arteries at risk, and 24 patients died, none of known aortic rupture: 11 (22%) cardiac complications, 3 (6%) cancer, 2 (4%) cerebrovascular complications, 1 (2%) renal failure, 3 (6%) old age, and 4 (8%) unknown causes. 

One patient operated in another institution presented with a vertebrobasilar stroke 13 months after total debranching and endovascular repair of an arch aneurism.  A CT scan showed the thrombosis of a carotid-subclavian bypass and a type 1A endoleak.  We performed a SCT and the insertion of a proximal stent to treat the endoleak.  These surgeries were uneventful.  After 9 months, the same patient suffered an occlusion of the bypass from the ascending aorta to the left carotid artery, which was associate to a left hemispheric transient ischaemic attack (TIA).  This was treated with an axillo-axilary bypass to revascularize the left carotid through the left SCT, which was still patent. 

Three late endoleaks were identified:  one type 1B and one type 3 endoleak occurring in the same patient at 16 and 24 months after the original surgery, and treated with the insertion of a distal stent with landing in zone in Z5 and a bridging stent, respectively; and another type 1B endoleak at 8 months of surgery, treated with a stent extension landing in Z5.

Review of the literature

Search results, relevance and validity.  The search strategy yielded 9312 articles.  After reviewing titles and abstract, 336 were considered potentially relevant based of the predefined inclusion and exclusion criteria.  After retrieving the full articles, 41 were considered relevant and valid. 5,16,25–34,17,35–44,18,45–54,19,55,20–24 (Table II). With the present cohort, a total of 42 studies were included in the analysis.

Interobserver reliability for judgement of relevance was good, with a Cohen's kappa statistic of 0.79.  An adequate description of patient demographics was present in 40 studies (93%), and an adequate description of the indications for surgery and of the interventions were present in 37 (88%) and 35 (81%), respectively.  Interobserver agreement for decisions related to methodologic quality was excellent for the description of study design (kappa, 0.88), good for description of patient demographics (kappa, 0.67), and poor for the description of surgical intervention (kappa, 0.30).  Only retrospective cohort studies were identified and included.  It was generally not clear whether consecutive patients were reported. 

Demographics, indications, and surgical strategy.  Overall there were 2078 hybrid procedures reported:  253 TDST, 108 TDST without LSA revascularization, 325 CCSB, 68 CCB, 906 CSB, 293 SCT, and 125 supra-aortic trunks with a non-specified revascularization.  Demographics and indications for surgery are reported in Table III.

Debranching was done before the endovascular repair in 1044 (50%, 95% CI 49 - 54), at the same time in 989 (46%, 95% CI 44 - 49), and the interval was unknown in 45 (2%, 95% CI 1 - 3).  Fourteen studies reported the interval between debranching and endovascular repair on a total of 790 patients and the delay between the two procedures was 20 days (95% CI 16 - 26).  There was one rupture between the open and the endovascular repair over 1044 patients (0.09%).

In 18 studies (42%, 95% CI 27 -58) a CSF drainage was used selectively based upon various parameters:  level of distal landing of the endovascular stent, though this was not specified; pre-existing infrarenal aortic aneurism repair; and occlusion of the LSA.  Only 13, however, stated the specific number of patients treated with this adjunct, 345 over 1027 (34%, 95% CI 30 - 37).

Seven studies reported the length of the proximal landing zone, which was 28 ± 9 mm and only one the distal landing zone, which was of 38 mm.

Early and late mortality and morbidity.  Average 30-day mortality was 5% (95% CI 3 - 6).  The relative risk of death appeared greater in the TDST group, 2.3 (95% CI 0.8 - 5.9) (P=0.08), with an Absolute Risk Increase (ARI) of 8.7 % (95% CI 8.5 - 9) (P=0.06), but did not reach statistical significance.

Among people who died, causes of postoperative mortality were:  cardiac 10% (95% CI 5 - 16);  cerebrovascular 20% (95% CI 13 - 28);  hemorrhagic complications 16% (95% CI 9 - 23);  sepsis or multiorgan failure 18% (95% CI 12 - 26);  other complications 19% (95% CI 12 - 27);  and unknown 8% (95% CI 4 - 14).  Other complications included: aorto-bronchial or tracheal fistulas, perforation of the left ventricle by the wire, rupture of the iliac artery, transfusion related acute lung injury, pneumonia, retrograde dissection, pulmonary embolus, respiratory failure, acute mesenteric ischemia, hepatic failure, and aortic rupture.

Average follow-up for the entire population was 28 months (95% CI 24 - 31) and mortality at 1, 3, and 5 years was 10 % (95% CI 7 - 12), 19% (95% CI 14 - 24), and 31% (95% CI 23 - 40), respectively.  Figure 5 illustrates 30-day, 1, 3 and 5-year mortality for all hybrid procedures considered together.  There was no difference in mortality among the different subgroups.

Most authors did not report the cause of long-term mortality (77 %, 95% CI 73 - 81).  Where this was described, it  was:  cardiac 9% (95% CI 7-12), cerebrovascular 2% (95% CI 1 - 4), aneurism rupture 1% (95% CI 0.4 - 3), sepsis and/or multiorgan disfunction 6% (95% CI 4 - 8), cancer 2% (95% CI 1 - 4), and other causes 2% (95% CI 1 -4).

Among the post-operative neurologic outcomes, TIA were reported in 3% of patients (95% CI 0 - 6), and strokes in 5% (95% CI 4 – 6).  Most authors did not describe the type of stroke, among those who did, hemispheric e vertebro-basilar strokes were similarly distributed.

Paraplegia occurred in 2% of patients (95% CI 2 - 3), though the distal extent of the endograft was rarely described, and it was permanent in 56% of them (95% CI 39 - 72).  Peripheral nerve injuries were:  phrenic nerve, 2% (95% CI 2 - 14); sympathetic chain causing Horner Syndrome, 2% (95% CI 0 - 4); and recurrent laryngeal nerve causing vocal cord paralysis, 3% (95% CI 2 - 4).

Other post-operative complications included:  cardiac, 6% (95% CI 2 - 11);  respiratory failure, 6% (95% CI 3 - 8);  pulmonary embolism 3% (95% CI 0 - 6);  retrograde aortic dissection after endovascular repair, 2% (95% CI 1 - 3);  upper limb ischemia, 1.5% (95% CI 0 - 4);  lymph leak, 2 % (95% CI 0.5 - 3); and femoral artery or iliac artery injury, 3 % (95% CI 2 – 5).

Patency of debranching and endoleaks.  Over 2078 patients, 24 supra-aortic trunks reconstructions failed 1% (95% CI 0 - 1.7) on follow up.  Table IV reports the failed reconstructions stratified by patient, number of arteries at risk, and type of surgery.  Failure rate was greater in the CSB group compared with SCT (P=0.05).

Early endoleaks, defined as those occurring during surgery, treated during the same hospital stay, or deemed that they were not treatable or necessary to be treated, were reported from 2% 41,47 to 43% 45 of patients with a pooled average of 4% (95% CI 3 - 5):  type I, 3% (95% CI 2 - 4);  and type II, 0.6% (95% CI 0.3 - 1).  Type III endoleaks were reported only in 3 patients.

Late endoleaks, defined as those occurring after discharge and during follow up, were reported from 2% 35,47,56 to 38% 21,45,57 of patients with a pooled average of 5% (95% CI 4 -7):  type I, 2% (95% CI 2 - 3); type II, 2% (95% CI 1 - 3); and type III, 0.5% (95% CI 0 - 1).

Table 1

ASA class and risk scores for the entire cohort and stratified by type of surgery

Table 2

Flow diagram: Review of the literature

Table 3

Review of the literature: Demographics, and indications for surgery

Table 4

Review of the literature: Failed reconstructions, stratified by patient, number of arteries at risk, and type of surgery

Figure 1

Carotid-carotid-subclavian bypass: 3-D surface rendering reconstruction of post-operative CT scan

Figure 2

Total debranching supra-aortic trunks: a trifurcated Dacron graft 12x8x8 mm was used

Figure 3

Survival of the entire cohort of patients undergoing hybrid procedures.

Figure 4

Survival of patients undergoing hybrid procedures stratified according to the type of debranching

Figure 5

Pooled estimate of mortality for all hybrid procedures, random effect models: A) 30-day mortality

Figure 5b

Pooled estimate of mortality for all hybrid procedures, random effect models: B) 1-year mortality

Figure 5c

Pooled estimate of mortality for all hybrid procedures, random effect models: C) 3-year mortality

Figure 5d

Pooled estimate of mortality for all hybrid procedures, random effect models: D) 5-year mortality

DISCUSSION

Our work suggests that hybrid surgeries are durable, effective, with low mortality and morbidity, and good long-term survival.  The perioperative mortality in our cohort (2%) compares well with the one reported in the review (5%), and long-term survival was also similar.  The trend for cervical reconstructions to have lower perioperative mortality and better long-term survival than TDST, did not reach statistical significance.  After surgery we had no late aneurysm rupture, and this event was rare in the literature (1%).  Failure of extra-anatomic reconstructions is uncommon.  It occurred in 1.7% of CSB compared to none in the SCT group (P=0.05), confirming our previous work on the better patency rate of the latter procedures. 9  A CSB, however, is necessary in the presence of a coronary bypass from a left internal mammary artery.  We had no thrombosis of grafts in the CCSB group, and a late occlusion of 1.5% was reported in the literature. 58,59  We prefer to place the graft low in the neck and in a pretracheal position because it is a simple and effective technique, which also prevents dysphagia, an underreported complication. 60  We report the thrombosis of one bypass in a TDST done in another institution where a bifurcated graft was used.  In the literature occlusion of at least on graft was reported in 2 patients (0.6%), and 0.2% of all arteries at risk.  For the TDST, we favour a trifurcated Dacron® graft 12x8x8 mm because the body of an off-the-shelf bifurcated graft has a larger diameter and requires a greater length of ascending aorta for anastomosis.  Moreover, the hood of this anastomosis has often a higher profile, increasing the risk of compression from surrounding structures. 

Among the indices we used to evaluate the preoperative risk of morbidity and mortality, the RCRI, which anticipated an event rate of 5.8%, was predictive of the real incidence of postoperative cardiac complications.  Since there was only one post-operative death in our cohort, the ASA class overestimated mortality.  This may be attributed to inconsistency in the allocation of patients within each ASA class;  the small sample size of our population;  and the derivation of risk, for each ASA class from open traditional surgery.

Debranching through cervical operations and ministernotomies seems to interfere less on respiratory mechanics and function compared with open thoracotomies.  Indeed, the Gupta RRC, with an average predicted risk of 9.8%, overestimated respiratory complications: there was no respiratory failure and only one pulmonary embolus and one pneumothorax in our cohort. 

In patients undergoing hybrid surgeries, neurologic accidents may occur in the cerebral territory and in the spinal cord.  The risk of spinal cord ischemia increases with the extent of the thoracic aorta covered by the stent. 61,62  On the basis of three reviews 12,63,64 supporting the benefits of CSF drainage to prevent paraplegia, we used it in 18 patients (35%) in whom coverage of the aorta distal to the T10 level was planned.  This is consistent with the results of the review where it was used selectively in 34% (95% CI 30 - 37) of patients.

Peripheral nerve injuries (phrenic nerve, sympathetic chain, recurrent laryngeal nerve) were reported in 2 - 3 % and lymph leaks in 2 % of the studies included in the review.  We had only one recurrent laryngeal injury and two lymphoceles.  The treatment of lymphoceles is controversial, but it appears that up to 70% may be treated conservatively 65,66 as it occurred in our patients.

Only 50% of studies in the literature used a staged approach to the hybrid procedure with an average interval of 20 days (95% CI 16 - 26), which is significantly shorter than ours (44 days, 95% CI 33 - 55).  While we suggest performing the endovascular procedure within 2 or 3 weeks after debranching, the reason for the delay included, medical conditions, patient’s preferences, administrative reasons, or availability of the endovascular stent grafts.  There were, however, no deaths between the first and second operation, while in the literature, this event is rare with an incidence of 0.09%.  Taken together these findings suggest that, within the confine of a stable aneurisms of 6 cm, a delay of 1-2 months is safe with respect to the possibility of rupture of the aneurism. 

In planning the endovascular treatment, the length of the proximal and distal landing zone is critical for the outcome of surgery.  The literature and guidelines of the European Association for Cardio-Thoracic Surgery recommend a landing zone of 20 mm in healthy aorta, both proximally and distally. 67  Multivariate analyses showed that a short proximal landing zone is the main risk factor for persistent or new developing type I endoleak. 68  Moreover, angulations of the aorta require a greater length of landing zone to secure a proximal seal 69,70 because they are associated with an increase in shear stresses and risk of migration of the endograft. 71  We and other authors 68 believe that a 20 mm proximal landing zone may not be adequate and a 30 mm is a more suitable length.  In the literature, the average length of the proximal landing zone was 28 ± 9 mm and the incidence of late endoleaks, varied considerably from 2% to 38%.  In our work the average proximal landing zone was 36 ± 7 mm (95% CI 34 - 38) which provided a stable repair over time:  there were no type 1A endoleaks and two type 1B (3.9%) on long term follow up. 

The strength of our work is the prospective design, the number of patients, the predefined strategy, and the long term follow up.  The limitations are those of all non-randomized control trials.  Moreover, the interpretation of results from the review of the literature is limited by the quality and the heterogeneity of the studies included:  they were all retrospective, it was not always clear if patients were consecutive, patency of the extra-anatomic bypasses was not always confirmed with objective tests, and no studies included an assessment of clinical outcomes by an independent observer.  Potential biases, therefore, may be introduced at a number of levels, weakening the inferences that may be drawn.  However, theoretical constructs, the results of mortality, morbidity, and long term follow up support the overall conclusions on the validity of hybrid procedures for the treatment of thoracic and aortic arch pathology.  The findings of this work may be used by physicians to choose the best strategy of treatment based on the results of hybrid procedures, and patients’ conditions and preferences.

Randomized controlled trials are the true experimental design, which might provide the answer to the question if, given similar patients and anatomic features, arch and thoracic pathology may be best treated with open or hybrid techniques.  However, these trials may be practically challenging or require time.  In the interim, research should focus on some aspect of hybrid procedures which are not yet clear or not readily available in the current studies:  long term follow up;  length of landing zones and their relations to the incidence of endoleaks;  use of adjuncts to prevent spinal cord ischemia and relation of this complication to the distal landing of the stent;  and type, number and length of overlap of stent grafts used.

CONCLUSIONS

Our cohort study, using a predefined strategy of debranching of supra aortic trunks and delayed endovascular repair with a proximal landing zone of 30 mm, and the review of the literature, show that hybrid interventions for the treatment of thoracic aortic pathology are durable procedures with low perioperative and late mortality.