Endovascular repair of descending thoracic aortic aneurysms—a mid-term report from the Global Registry for Endovascular Aortic Treatment (GREAT)

Abstract

 

Open in new tabDownload slide

OBJECTIVES

The aim of this study was to evaluate the short- to mid-term outcomes of descending thoracic aortic aneurysm (DTAA) repair from the Gore Global Registry for Endovascular Aortic Treatment (GREAT).

METHODS

This is a multicentre sponsored prospective observational cohort registry. The study population comprised those treated for DTAA receiving GORE thoracic aortic devices for DTAA repair between August 2010 and October 2016. Major primary outcomes were early and late survival, freedom from aorta-related mortality and freedom from aorta-related reintervention.

RESULTS

There were 180 (58.1%) males and 130 (41.9%) females: the mean age was 70 ± 11 years (range 18–92). The median maximum DTAA diameter was 60 mm (interquartile range 54–68.8). Technical success was achieved in all patients. Operative mortality, as well as immediate conversion to open repair, was never observed. At the 30-day window, mortality occurred in 4 (1.3%) patients, neurological events occurred in 4 (1.3%) patients (transient ischaemic attacks/stroke n = 3, paraplegia n = 1) and the reintervention rate was 4.5% (n = 14). Estimated survival was 95.6% [95% confidence interval (CI) 92.6–97.4] at 6 months, 92.7% (95% CI 89.1–95.2) at 1 year and 57.3% (95% CI 48.5–65.1) at 5 years. Freedom from aorta-related mortality was 98.3% (95% CI 96.1–99.3) at 6 months, 98.3% (95% CI 96.1–99.3) at 1 year and 92.2% (95% CI 83.4–96.4) at 5 years. Freedom from thoracic endovascular aortic repair (TEVAR)-related reintervention at 5 years was 87.2% (95% CI 81.2–91.4).

CONCLUSIONS

TEVAR for DTAAs using GORE thoracic aortic devices is associated with a low rate of device-related reinterventions and is effective at preventing aorta-related mortality for up to 5 years of follow-up.

Clinical registration number

NCT number: NCT01658787.

Subject collection

161, 164.

INTRODUCTION

Thoracic endovascular aortic repair (TEVAR) has been proven to have substantial early benefits in terms of perioperative morbidity and mortality and is at least as effective in preventing aortic-related death (ARM) as open surgical repair for descending thoracic aortic aneurysms (DTAA) [1–4]. At longer follow-up terms, the reported rates of reinterventions, endoleaks, late aortic ruptures, conversions to open surgical repair and other aortic complications have been generally low. However, they are not negligible and require careful monitoring, because they have a negative impact on survival [5, 6]. Current data available in the literature are either controlled trial or single-centre studies; however, registries may be a valuable tool for collecting data for specific devices [7]. This report aims to describe the outcomes of a recent international, multicentre cohort of DTAAs treated with the GORE (W.L. Gore & Associates, Flagstaff, AZ, USA) thoracic aortic stent graft (SG).

METHODS

Patients cohort

The GORE Global Registry for Endovascular Aortic Treatment (GREAT) registry design and specifications have been described in previous publications [8, 9]. The registry was designed as a multicentre, sponsored, prospectively collected registry of consecutive patients who were treated with GORE endovascular aortic products (NCT number: NCT01658787). It was conducted according to the Declaration of Helsinki and the International Conference on Harmonization and Good Clinical Practice guidelines and approved by the ethics committee or institutional review board of each participating centre (protocol no. 0038114, 6 November 2013). Data were acquired and then given to the authors by the sponsor. For this specific analysis we identified all cases of TEVAR performed for intact, non-ruptured DTAA [10]. Between August 2010 and October 2016, a total of 5013 patients were enrolled and 318 (6.3%) DTAAs were identified. Eight (2.5%) patients were excluded because of incomplete data; therefore, 310 (6.2%) DTAAs form the cohort for the final analysis.

Operative options

Although GREAT does not mandate treatment guidelines in the protocol, introducer sheath and device selection were recommended to be made according to the instruction for use (IFU) of the manufacturer. The choice of the proximal landing zone for the SG deployment, as well as the type of approach to the access vessel, was left to individual surgeon. A commonly accepted primary indication for the use of a conduit was the presence of bilateral small iliac arteries (e.g. smaller than the sheath diameter) or those with a combination of extensive calcification and/or adverse tortuosity. More than 1 access method could be reported in the same subject. In general, thoracic SG oversizing was optimized according to the type of device implanted, taking into account the underlying aortic disease: however, oversizing was the maximum according to the manufacturer’s IFU based on the diameter chosen for the thoracic device. Ballooning of the overlapping zones was generally performed when >1 SG was used.

Recommended patient follow-up and imaging were recommended to follow the product IFU and the method of follow-up was at the discretion of the treating physician. Typically, endograft imaging was performed at 1 and 6 months and then annually thereafter, unless at the discretion of the treating physician it was required more frequently. According to the protocol, follow-up is for 10 years or until the time of the patient’s death. If a patient does not return to the site for follow-up evaluation, we request that the site, at a minimum, contact the patient or the patient’s next of kin/representative to encourage follow-up visits and determine survival.

Follow-up and definitions

Morphological characteristics and outcomes were defined according to the Society for Vascular Surgery (SVS) ad hoc committee on TEVAR reporting standards [11]. The procedure was considered urgent when the intervention was performed ≤ 24 h from hospital admission and the first diagnosis of the DTAA, especially in symptomatic patients. Aortic segments were classified with the corresponding anatomical landmarks. Participating centres were required to enter only those adverse events that met the International Organization of Standardization definition of serious (ISO 14155, https://www.iso.org/obp/ui/#iso:std:iso:14155:ed-2:v1:en). Follow-up was performed according to the schedule of each participating centre. Specifically for this study, primary outcomes were early and late survival, freedom from aorta-related mortality (ARM) and freedom from aorta-related reintervention. Missing data exist and were expected during follow-up because, per the GREAT protocol, there is no mandated follow-up schedule. Participating sites are encouraged to follow the IFU follow-up guidelines for the device implanted. In addition, the protocol allows patients to miss a visit and not be discontinued because the intent is to see if patients will come back at a later date, though sites are encouraged to follow up with their patients.

Data collection, processing and analysis

Collected data were recorded on a web-based electronic report form (iMedidata, Medidata Worldwide Solutions, Inc., New York, NY, USA) to ensure reliability and secure authentication and traceability [12]. Data management was performed by the Gore Clinical Research Department (W.L. Gore & Associates). All collected data were reviewed and if missing or inconsistent data were detected, relevant queries were posed to the investigators for resolution. Monitoring visits were performed at each enrolment site to verify necessary study documents, including signed informed consent for each patient. Consistency between electronically imported data and source documents was also examined. Categorical variables are expressed as number and percentage (%). Continuous variables were presented with mean ± standard deviation (SD) or median and interquartile ranges (IQR), based on the data distribution. Data distribution was assessed by means of visual plotting and by means of the Shapiro–Wilk test for normality. Clinical and technical time-to-events (ARM and freedom from TEVAR-related reintervention) outcomes during follow-up were estimated by Kaplan–Meier analysis with Kaplan–Meier survival confidence intervals (CIs). For each outcome (all-cause mortality at 2 years and any reintervention at 2 years), a series of Cox proportional hazards models were built. All those without an event by 2 years were censored at 2 years. Modelling could not be completed for ARM because there were only 7 events in 2 years. Univariable models were built for key demographics (age, race, gender, weight, medical history) as well as anatomical/procedural (lesion diameter, proximal neck length, branch vessel procedure, access vessel technique and site) covariates. The proportional hazards models added time to each parameter to test the hazards assumption for each predictor. For the variables for which this interaction was significant (P  ≤  0.10), that specific interaction term was included in each univariate model to adjust for the impact of time. Significant interactions (P ≤  0.1) with time were left in final univariable models, and non-significant interactions were removed. For multivariable models and the Cox proportional hazards model, backward selection with 0.1 stay criteria was used to create a multivariable model to predict each outcome. Variables from univariable models that had a P-value ≤ 0.1 and their interactions with time (if interaction P ≤  0.1 in univariable model) were provided as options in the backward selection model. Any variables from univariable models that had a P-value ≤ 0.1 but % missing >15% were not provided as options for the final model. All reported P-values were two-sided; P-value < 0.05 was considered significant. All data were analysed using statistical SAS software, version 9.4 of the SAS System for Windows (Copyright 2002–2008 by SAS Institute Inc., Cary, NC, USA).

RESULTS

Baseline data

There were 180 (58.1%) males and 130 (41.9%) females: the mean age was 70 ± 11 years (range 18–92). Demographic data, medical history and preoperative imaging findings of the entire cohort are shown in Table 1. TEVAR was performed as a primary procedure in 253 (81.6%) patients, 35 (11.3%) as a reintervention after a previous endovascular procedure and 22 (7.1%) as a reintervention after an open surgical procedure. The median maximum DTAA diameter was 60 mm (IQR 54–68.8). The median length of the proximal aortic neck was 30 mm (IQR 20–57.5). The median diameter of the maximum proximal landing zone diameter was 32.1 mm (IQR 29–36.1), and 30 mm (IQR 27–33.3) was the median minimum proximal landing zone diameter. The femoral artery was the most (n = 287, 92.6%) used access vessel for TEVAR: 89 (28.7%) patients were treated fully percutaneously, while a retroperitoneal/abdominal exposure was used to approach the iliac artery (n = 26, 8.4%) or the infrarenal aorta (n = 2, 0.6%). The SG was deployed in ‘zone 1’ in 17 (6.5%), ‘zone 2’ in 38 (14.6%), ‘zone 3’ in 141 (54%) and ‘zone 4’ in 65 (24.9%) patients. There were a total of 560 devices implanted within the 310 GREAT subjects included in this specific subset, with a median of 2.0 SGs used per subject. Specifically, 266 subjects had the Conformable GORE TAG Thoracic SG only, 31 subjects had the GORE TAG Thoracic SG only and 11 subjects had a combination of the 2 devices. Table 2 shows the additional branch vessel procedures stratified by proximal landing zone.

Early outcomes (< 30 days)

Operative mortality and immediate conversion to open repair were never observed. There was 1 procedural stroke (0.3%) and 1 spinal cord injury (0.3%). Of the 30 (10%) subjects in the subset who had no previous left subclavian artery revascularization, no neurological events (i.e. stroke, paraplegia/paraparesis) were recorded, as defined by the GREAT registry. The median hospital stay was 6 days (IQR 3–10). At the 30-day window, 4 patients had died (1.3%; 95% CI 0.0035–0.0327): Causes of death were shock liver (n = 1), acute mesenteric ischaemia (n = 1), stroke (n = 1) and unknown in 1 patient. New onset of neurological events in the form of stroke/transient ischaemic attack or paraplegia occurred in 4 (1.3%; 95% CI 0.0035–0.0327) patients. The reintervention rate was 4.5% (n = 14; 95% CI 0.0249–0.0746); cause for and type of reintervention is reported in Table 3.

Late (≥30 days) outcomes

There were 42/84 (50%) subjects who completed any follow-up within the 5-year follow-up window. Of those 84 subjects, 29/84 (34.5%) had imaging completed within the 5-year follow-up window and 5/84 (6.0%) subjects were lost due to death/discontinuation. Estimated survival was 95.6% (95% CI 92.6–97.4) at 6 months, 92.7% (95% CI 89.1–95.2) at 1 year and 57.3% (95% CI 48.5–65.1) at 5 years: after a univariate screen, a multivariable model at 2 years identified renal insufficiency [P = 0.027; hazard ratio (HR) 2.02], peripheral arterial obstructive disease (P = 0.006; HR 39.4), previous stroke (P = 0.033; HR 2.35) and congestive heart failure (P = 0.044; HR 2.31) as predictors of late mortality. Estimated ARM was 98.3% (95% CI 96.1–99.3) at 6 months, 98.3% (95% CI 96.1–99.3) at 1 year and 92.2% (95% CI 83.4–96.4) at 5 years as reported in Figure 1. Aortic rupture was documented in 1 (0.3%) patient: It was a rupture of a juxtarenal abdominal aortic aneurysm that occurred 6 months after TEVAR and was fatal. The type 1a endoleak rate through the 2-year follow-up was 1.6% (n = 5). Table 4 shows the different types of endoleaks at the different time windows of follow-up. None of the patients with endoleaks had a lesion increase >5 mm reported at any time during their follow-up. Migration occurred in only 1 (0.3%) patient at 1 month after TEVAR, while fracture or collapse was never detected. Overall, estimated freedom from TEVAR-related reintervention was 97.3% (95% CI 94.6–98.6) at 6 months, 95.0% (95% CI 91.7–97.0) at 1 year and 87.2% (95% CI 81.2–91.4) at 5 years (Figure 2). Late reinterventions are reported in Table 5. A multivariable model at 2 years identified an aortic branch vessel procedure (P = 0.009; HR 2.39) and chronic obstructive disease (P = 0.007; HR 4.38) to be associated with the necessity of aortic reintervention during the follow-up.

DISCUSSION

Although randomized clinical trials represent the benchmark method to perform research studies, prospective registries are a valid alternative research method to obtain prompt ‘real-world’ data on safety, efficacy and durability of a specific treatment with newer techniques or for rare pathologies [7]. Defining a target population is the crucial first step of a registry-based study to determine the reproducibility of the study results [13]. The most important aspects of this analysis are that it is a ‘real-world’, very recent, large cohort of TEVAR for DTAA.

Several pivotal studies comparing TEVAR versus open surgical repair of DTAA already demonstrated the effectiveness of TEVAR in patients with suitable DTAA [2–4]. However, pivotal cohorts may not have been reflective of daily practice because strict trial settings may differ from ‘real-world’ device performance. In fact, results coming from large administrative data sets reported similar results between TEVAR and open surgical repair [1, 5, 6, 14]. Indeed, the GREAT registry offers a substantial opportunity for data analysis. First, it includes different clinical scenarios. Second, being a worldwide registry means it has recruited as many people from different countries and latitudes as possible, thus making the results closer to the ‘real-world’ practice. In this context, the analysis of the GREAT registry DTAA cohort showed the safety and efficacy of TEVAR at mid-term follow-up, whose values for ARM, all-cause mortality and freedom from TEVAR-related reintervention were close, if not better, to those reported in the potentially ‘muffled’ cohorts of the pivotal trials [2–4].

One of the major outcomes against which to test the effectiveness of TEVAR is ARM during the follow-up, namely the protection that TEVAR confers against aortic-related death [10]. Freedom from ARM in this GREAT cohort was 92.2% at 5 years of follow-up, which is in agreement with the estimate of between 94% and 97% in the pivotal trials [2–4]. These data compare well also with the corresponding data coming from the more ideal scenarios of the GORE TAG trial [2]. Such satisfactory results may be explained both by the increased surgeon experience with TEVAR, which has probably had a positive influence on the outcomes of the current registry, and also by the more conformable design of the device. This latter aspect may have played an important role in containing the incidence of type 1a endoleak at 1.6%, an acceptably low rate in this realistic experience, which included nearly 20% of DTAA requiring ‘zones 1–2’ aortic arch TEVAR [3].

Reported concerns of TEVAR for DTAAs have been that the early mortality benefit may be lost due to an excess of late all-cause mortality. An important caveat to this observation is that there are few long-term follow-up data available for patients who have undergone TEVAR. The GREAT registry still has not reached the 10-year follow-up outcome analysis, which is the main goal for this large real-world registry. However, this preliminary 5-year estimation shows that currently there are no TEVAR-related significant predictors for late mortality; rather only preoperative comorbidities had a significant role in predicting mortality. Furthermore, the 92% freedom from ARM at 5 years is satisfactory. Although thoracic aortic pathology may represent a sign of severe systemic disease, these data also highlight the safety and efficacy of TEVAR during the follow-up.

Secondary aortic interventions have been described commonly in TEVAR series, usually with a higher prevalence compared to reintervention after open repair. In this analysis, within the 13% aortic reintervention estimated at 2 years, 84% was managed endovascularly and did not cause additional mortality. This observation underlines the fact that there is a significant association between additional procedures involving the branch vessels, usually performed in an anatomic configuration not included in the IFU, and the need for aortic reintervention. In addition, it needs to be noted that treating lesions that required additional branch vessel procedures was often challenging, whereas isolated TEVAR can be an effective and lasting procedure.

Limitations

The present analysis has major limitations. First, the non-randomized design of the registry can result in incomplete data and limited variable definitions. Second, there is sampling bias because patients undergoing total open repair were not included for comparison, but this reflects the careful selection process for these types of procedures. Third, the present study includes few patients with a follow-up evaluation at 5 years, although differences in follow-up schemes and reintervention protocols exist among the participating centres. Lastly, analyses were exploratory in nature and there was no prespecified plan to adjust for multiple subgroup comparisons or meaningful multivariate analysis also due to the few adverse events determined by the good outcomes with TEVAR.

CONCLUSIONS

This analysis of the global controlled GREAT registry shows that TEVAR for DTAA using GORE TAG and Conformable GORE TAG SGs can be performed with a low ARM and SG-related reintervention rate at 5 years. Longer follow-up is needed to evaluate the durability of these outcomes.

Presented at the 33rd Annual Meeting of the European Society for Vascular Surgery, Hamburg, Germany, 24–27 September 2019.

ACKNOWLEDGEMENTS

The authors are grateful to Hillary Alberta for the data acquisition and management and to Lily Ray for the statistical support.

Funding

This paper is financially unfunded but received data managing and technological support for the final analysis.

Conflict of interest: Santi Trimarchi and Ali Azzizzadeh are consultants and Gabriele Piffaretti is a lecturer for W.L. Gore^® and recipient of research grants to his institution from them. All other authors declared no conflict of interest.

Author contributions

Viviana Grassi: Resources. Santi Trimarchi: Writing—review & editing. Fred Weaver: Resources. Hector W.L. de Beaufort: Formal analysis; Methodology. Ali Azzizzadeh: Resources. Gilbert R. Upchurch Jr: Writing—review & editing. Gabriele Piffaretti: Conceptualization; Data curation; Formal analysis; Methodology; Project administration; Resources; Software; Supervision; Writing—original draft. Chiara Lomazzi: Conceptualization; Methodology; Resources.

Reviewer information

European Journal of Cardio-Thoracic Surgery thanks Luca Bertoglio, Roman Gottardi, Mario Lescan and the other, anonymous reviewer(s) for their contribution to the peer review process of this article.

REFERENCES

ABBREVIATIONS

 
• ARM

  Aorta-related mortality

 
• CI

  Confidence interval

 
• DTAA

  Descending thoracic aortic aneurysm

 
• GREAT

  Global Registry for Endovascular Aortic Treatment

 
• HR

  Hazard ratio

 
• IFU

  Instruction for use

 
• IQR

  Interquartile range

 
• TEVAR

  Thoracic endovascular aortic repair