Advertisement

Association Between Device Type and Type IIIb Endoleaks Following Thoracic Endovascular Aortic Repair

Open AccessPublished:October 05, 2022DOI:https://doi.org/10.1016/j.ejvs.2022.10.005

      Objective

      Endoleaks following thoracic endovascular aortic repair (TEVAR) can lead to the need for re-intervention and aortic rupture. Given the recent recall of a specific thoracic stent graft type due to concerns about type IIIb endoleak (T3bE), in which blood leaks through the stent graft fabric, the aim was to characterise patterns of T3bEs among available TEVAR devices.

      Methods

      Reports related to thoracic stent grafts in the Food and Drug Administration’s (FDA) Manufacturer and User Facility Device Experience (MAUDE) database between 2010 and 2020 were reviewed. Proportional reporting ratios (PRRs) and chi squared tests were used to assess for the presence of a signal of association between device type and T3bE. A PRR > 2 and chi squared value > 4 with three or more reports overall constituted a signal of association. The institutional database of patients undergoing TEVAR for thoracic aneurysms since 2002 was also queried for T3bE.

      Results

      There were 7 328 MAUDE reports available for analysis. When analysing T3bEs in the MAUDE database, the lowest PRR was 0.36 (95% confidence interval [CI] 0 – 1.03 in Gore CTAG, nine T3bE among 1 328 reports) and the highest was 2.07 (95% CI 1.72 – 2.42 in Medtronic Valiant, 64 T3bE among 2 520 reports). The T3bE chi squared value for Medtronic Valiant was 17.3. The relationship between Medtronic Valiant and T3bE was ascertainable by MAUDE data as early as 2013. Among 542 TEVARs for an aneurysm indication at Duke University Hospital since 2002, there were eight T3bEs – all in devices with sutured on graft material.

      Conclusion

      The Medtronic Valiant device met criteria for association with type IIIb endoleaks in the FDA’s MAUDE database and met those criteria as early as 2013. A possible relationship between woven graft fabric and T3bEs is supported by the observation that all types of T3bE that occurred among a large number of TEVARs at the institution followed placement of grafts with sutured on woven fabric.

      Keywords

      A retrospective review of the Food and Drug Administration’s Manufacturer and User Facility Device Experience and institutional aortic surgery databases revealed a proportional reporting ratio-based signal of association between the Medtronic Valiant device and type IIIb endoleaks, which was ascertainable as early as 2013. Other devices with sutured on graft material also had numerically higher proportional reporting ratios, consistent with the institutional data. Given the time lag between signal development and Medtronic Valiant Navion recall, serial monitoring of device related reports with in depth investigation in the presence of a signal may help surgeons make informed device related decisions and help prevent patient harm.

      Introduction

      An estimated 8 000 patients undergo thoracic endovascular aortic repair (TEVAR) yearly in the United States for repair of aortic dissections or aneurysmal disease.
      • Kalesan B.
      • Cheng T.W.
      • Farber A.
      • Zuo Y.
      • Kalish J.A.
      • Jones D.W.
      • et al.
      Readmissions after thoracic endovascular aortic repair.
      Endoleaks occur following upwards of 10% of TEVARs and can lead to ongoing aortic expansion, need for re-intervention, and aortic rupture.
      • Biancari F.
      • Mariscalco G.
      • Mariani S.
      • Saari P.
      • Satta J.
      • Juvonen T.
      Endovascular treatment of degenerative aneurysms involving only the descending thoracic aorta: systematic review and meta-analysis.
      • Vaaramaki S.
      • Suominen V.
      • Pimenoff G.
      • Saarinen J.
      • Uurto I.
      • Salenius J.
      Long-term experience of endovascular repair for thoracic aortic aneurysms and dissections.
      • Kamman A.V.
      • Jonker F.H.
      • Nauta F.J.
      • Trimarchi S.
      • Moll F.L.
      • van Herwaarden J.A.
      A review of follow-up outcomes after elective endovascular repair of degenerative thoracic aortic aneurysms.
      Although type I and II endoleaks are related to clinician technique and patient anatomy, the aetiology of type III endoleaks may be related to the stent graft itself. This is especially true for type IIIb endoleaks in which blood passes through a defect in the graft fabric, as opposed to type IIIa endoleaks, in which blood passes through different components of an endovascular reconstruction (such as two overlapping stent grafts). Therefore, a better understanding of graft related type IIIb endoleak risk could help guide stent graft choice in situations in which multiple stent graft types are viable options. Though type III endoleaks are relatively uncommon, affecting < 2% of patients following TEVAR, they disproportionately lead to re-intervention, accounting for 14% of endoleak related re-interventions.
      • Biancari F.
      • Mariscalco G.
      • Mariani S.
      • Saari P.
      • Satta J.
      • Juvonen T.
      Endovascular treatment of degenerative aneurysms involving only the descending thoracic aorta: systematic review and meta-analysis.
      ,
      • Nozdrzykowski M.
      • Luehr M.
      • Garbade J.
      • Schmidt A.
      • Leontyev S.
      • Misfeld M.
      • et al.
      Outcomes of secondary procedures after primary thoracic endovascular aortic repair.
      Recently, the Medtronic Valiant Navion device (Medtronic Vascular, Santa Rosa, CA) was recalled because of concerns about stent fractures and endoleaks (specifically type III endoleaks). The recall was issued in February 2021 on the basis of three stent fractures, seven stent ring enlargements, and four type IIIb endoleaks. Subsequent independent core laboratory review of 404 patients’ imaging yielded eight type IIIb endoleaks. Medtronic now recommends computed tomography angiography (CTA) every six months for patients previously implanted with a Medtronic Valiant Navion graft.
      Food and Drug Administration
      Medtronic recalls Valiant Navion Thoracic Stent Graft System due to risk of stent fractures and type III endoleaks.
      ,

      de Jesus Hernandez E. UPDATED Patient management recommendations Medtronic Valiant Navion Thoracic Stent Graft System global voluntary product recall. Available at: https://www.medtronic.com/content/dam/medtronic-com/global/HCP/Documents/valiant-navion-urgent-medical-device-recall-may-2021-update.docx [Accessed 11 October 2022].

      Given the clinical implications of endoleaks and the potential association between stent graft type and type IIIb endoleak, the aim was to characterise patterns of type IIIb endoleaks among TEVAR device types available in the United States. Because type IIIb endoleaks are not frequent enough to make statistical inferences about device-related patterns in single institutional datasets, the Food and Drug Administration’s (FDA’s) Manufacturer and User Facility Device Experience (MAUDE) database was used. Data from the institution were also reviewed to further explore the clinical context surrounding type IIIb endoleak occurrence.

      Materials and methods

      This was a retrospective review of data collected as part of (a) the FDA MAUDE database and (b) the institutional aortic surgery database. The former consists of publicly available, de-identified information and is not subject to institutional review board approval; the use of the latter data source was approved by the Duke University (Durham, NC) institutional review board. No patient consent was needed for this analysis.

      Data sources and curation

      The MAUDE database, maintained by the FDA, is a compendium of reports of adverse medical device events. It has been used previously to characterise adverse events associated with thoracic endografts.
      • Mansukhani N.A.
      • Haleem M.S.
      • Eskandari M.K.
      Thoracic endovascular aortic repair adverse events reported in the Food And Drug Administration Manufacturer And User Facility Device Experience Database.
      Manufacturers, importers, and device user facilities (e.g., hospitals, ambulatory surgical centres) are mandated to report events in which devices caused or contributed to death or serious injury, or have malfunctioned in a way that could contribute to death or serious injury in the future.
      Food and Drug Administration
      Mandatory reporting requirements: Manufacturers, importers and device user facilities.
      Voluntary reports can also be made by people such as healthcare professionals, patients, or consumers. MAUDE stores data for 10 years and can be queried via an online interface at https://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfMAUDE/search.CFM#fn1.
      For this analysis, MAUDE was queried by device type following a previously defined search strategy.
      • Mansukhani N.A.
      • Haleem M.S.
      • Eskandari M.K.
      Thoracic endovascular aortic repair adverse events reported in the Food And Drug Administration Manufacturer And User Facility Device Experience Database.
      All reports pertaining to the device types of interest (Bolton/Terumo RELAY (Terumo Aortic, Sunrise, FL), Cook Zenith TX2, Cook Zenith Alpha (Cook Medical, Bloomington, IN), Gore TAG, Gore CTAG (W.L. Gore and Associates, Flagstaff, AZ), Medtronic Valiant (including Valiant Navion during the years that it was commercially available from 2018 to 2020), and Medtronic Talent during the last 11 years (1 January 2010 – 31 December 2020) were downloaded.
      MAUDE reports contain categorisations of the event type (injury, death, or device malfunction) and the device problem (sticking, leak or splash, material erosion, etc.) as well as free text elements. The categorisations were not used because they were inconsistently applied (including frequent use of the “insufficient information” category) and did not contain adequate detail to categorise endoleaks. Reporters are encouraged to include all information known about the event in the free text portion, but no specific details are absolutely required.
      Food and Drug Administration
      Mandatory reporting requirements: Manufacturers, importers and device user facilities.
      A single author (E.H.W.) reviewed all reports that contained the word “endoleak” in the free text. Reports were categorised into type IIIa, type IIIb, and type III undetermined subtypes based on the report either specifically categorising the endoleak as such or based on the described endoleak characteristics. Therefore, descriptions of blood leaking between graft components were considered type IIIa endoleaks while descriptions of blood leaking through graft fabric were considered type IIIb endoleaks (and unclear descriptions were considered type III undetermined).
      Reports detailing non-thoracic uses of the devices and duplicate reports were excluded. Reports were considered to detail non-thoracic use if they described implantation of the device in the abdominal aorta or other non-thoracic aorta vessel. Duplicate reports were identified on the basis of identical wording and spelling; when multiple reports about the same event were submitted over time in reflection of changes in the clinical course, only the most complete report was used. Because manufacturers, importers, and device user facilities are required to report to the FDA when they learn that any of their devices have malfunctioned, some reports consisted of abstracts or results from published case reports or series, which often aggregated numerous device types. These reports of published case series were therefore also excluded unless they included the specific number of type III endoleak events in relation to specific device types. It was not possible to exclude reports on the basis of indication for stent graft implantation, as indication was not required to be reported and was variably present.
      The institutional aortic surgery database was queried for all patients undergoing TEVAR for any indication since the inception of the database (2002), yielding 688 patients. Of these, 87 underwent TEVAR for acute dissection without aneurysmal degeneration, 33 for aortic injury, 16 for penetrating ulcer, and 10 for management of complications of prior TEVARs, leaving 542 patients undergoing TEVAR for an aneurysm indication for analysis. Patients with heritable thoracic aortic disease (HTAD) were not excluded because diagnostic sensitivity for HTAD has changed significantly over the duration of the aortic database and excluding patients with HTAD would therefore introduce bias. Patients undergo CTA imaging and clinical follow up at one, six, and 12 months, and yearly thereafter (with deviations from this surveillance regimen as necessary) as per the previously described institutional protocol.
      • Hughes G.C.
      • Daneshmand M.A.
      • Swaminathan M.
      • Nienaber J.J.
      • Bush E.L.
      • Husain A.H.
      • et al.
      “Real world” thoracic endografting: results with the Gore TAG device 2 years after U.S. FDA approval.
      The database contains information about patient demographics, comorbidities, index and subsequent aortic procedures, and complications.

      Statistical analysis

      The MAUDE data and the institutional data were analysed separately. The number of MAUDE reports was described by device by year of report receipt and by device overall. Because a significant proportion of type III endoleaks could not be definitively subcategorised into type IIIa or type IIIb, the MAUDE data were analysed using two parallel approaches to capture the relationship between device type and type IIIb endoleak occurrence in the most comprehensive and the most conservative manners possible. In the comprehensive approach, all type III undetermined endoleaks were assumed to be type IIIb (the endpoint was a combination of type IIIb and type III undetermined endoleaks, together called “non-a type III endoleaks”), and in the conservative approach none of the type III undetermined endoleaks was assumed to be type IIIb (the endpoint was type IIIb endoleaks only). This was done in order to convey the extent to which inadequately detailed reports (i.e., type III undetermined) might affect the perceived relationship between device type and type IIIb endoleak occurrence, which is important because prior research suggests that type IIIb endoleaks may be especially difficult to diagnose definitively on imaging vs. in the course of operative re-intervention.
      • Fujimura N.
      • Ichihashi S.
      • Matsubara K.
      • Shibutani S.
      • Harada H.
      • Obara H.
      • et al.
      Type IIIb endoleak is not extremely rare and may be underdiagnosed after endovascular aneurysm repair.
      The proportions of reports related to non-a type III endoleaks and type IIIb endoleaks were calculated by device type. Because MAUDE does not represent the actual volume of TEVARs done nationally, the primary analyses consisted of proportional reporting ratios (PRRs), a technique previously used in the MAUDE database.
      • Evans S.J.
      • Waller P.C.
      • Davis S.
      Use of proportional reporting ratios (PRRs) for signal generation from spontaneous adverse drug reaction reports.
      ,
      • Aggarwal N.
      • Chow R.
      Real world adverse events of interspinous spacers using manufacturer and user facility device experience data.
      First described in 2001 for use in the United Kingdom’s adverse drug reaction database, comparison of PRRs between device (or drug) types facilitates identification of signals of adverse events, which should guide further investigation as clinically indicated. For this analysis, each device’s PRR was calculated by dividing the proportion of that device’s reports related to non-a type III endoleaks (or type IIIb endoleaks) by the proportion of reports related to non-a type III endoleaks (or type IIIb endoleaks) among all other device types. The PRR is an indicator of how much more (or less) frequently a specific type of endoleak appeared in reports for a given device compared with all the other types of devices combined. For instance, a PRR of 2 would indicate that a specific type of endoleak showed up twice as often in reports for one device compared with others. The presence of a signal for device association with non-a type III endoleaks or type IIIb endoleaks was determined following criteria set by Evans et al. (three or more cases, PRR of at least two, and chi squared of at least four).
      • Evans S.J.
      • Waller P.C.
      • Davis S.
      Use of proportional reporting ratios (PRRs) for signal generation from spontaneous adverse drug reaction reports.
      For the purposes of this analysis, chi squared without Yates’ correction was used due to no expected counts < 5. Both the PRR itself and the chi squared value are included in the criteria for a signal of association because the chi squared value adds information about the magnitude of deviation from the expected number of non-a type III and type IIIb related reports. This depends not only on the size of the difference (as indicated by the PRR), but the number of reports and the number of specific endoleak types identified. Therefore, the chi squared value may be large with either a high or low PRR, indicating a greater magnitude of deviation from the expected value.
      For any device meeting the criteria set by Evans et al. to determine the presence of a signal for device association with endoleaks, an analysis was conducted to determine when, chronologically, the device would have met the criteria if the devices had been evaluated on a yearly basis with the cumulative data available. This analysis was conducted using the conservative endpoint of only type IIIb endoleaks and consisted of calculating the number of type IIIb endoleaks, the PRR, and the chi squared for the device of interest on a yearly basis starting from the year in which at least three type IIIb endoleak reports had accumulated.
      The primary outcome for analysis of the local institutional data was type IIIb endoleak. Presence of type IIIb endoleak was determined on the basis of CTA imaging findings and intra-operative findings for those patients requiring re-operation. Because of the small number of type IIIb endoleaks at the institution, institutional data were reported descriptively without inferential statistics. Data were collected in Microsoft Excel and analysed in R.

      Results

      Manufacturer and User Facility Device Experience reports and analysis

      There were 7 328 reports overall from MAUDE included for analysis, with totals per device ranging from 302 (Cook Zenith Alpha) to 2 520 (Medtronic Valiant; Table 1). Two devices had reports available throughout all 11 years of the study period (Medtronic Talent and Gore TAG), with yearly volume of reports per device fluctuating as new devices entered the market (Supplementary Figure 1).
      Table 1Type III endoleak subtype reports by device in the Food and Drug Administration’s Manufacturer and User Facility Device Experience (MAUDE) database between 2010 and 2020
      Device typeYear of first reportTotal reportsType III endoleak reports (% of total device reports)
      Type IIIa endoleaksNon-a type III endoleaks
      Type IIIb endoleaksType III undetermined endoleaks
      Medtronic Talent201099175 (7.6)18 (1.8)18 (1.8)
      Gore TAG20101 32831 (2.3)9 (0.7)28 (2.1)
      Medtronic Valiant20112 52043 (1.7)64 (2.5)31 (1.2)
      Cook TX220113894 (1.0)8 (2.1)1 (0.3)
      Gore CTAG20121 47943 (2.9)11 (0.7)8 (0.5)
      Bolton RELAY20123194 (1.3)8 (2.5)4 (1.3)
      Cook Alpha20163024 (1.3)5 (1.7)4 (1.3)
      Data are presented as n (%).
      All devices had more than three non-a type III endoleaks and more than three type IIIb endoleaks reported (Table 1). The percentage of non-a type III endoleaks that could specifically be identified as type IIIb endoleaks ranged from 22.0% (Gore CTAG) to 88.9% (Cook Zenith TX2).
      In the comprehensive analysis, the proportion of reports related to non-a type III endoleaks ranged from 1.28% (Gore CTAG) to > 3.6% (3.63% for Medtronic Talent, 3.76% for Bolton RELAY, and 3.77% for Medtronic Valiant; Table 2). The lowest PRR was 0.38 (95% confidence interval [CI] 0 – 0.85 in Gore CTAG). No device met the Evans criteria for non-a type III endoleaks, as the chi squared value for the Medtronic Valiant was 8.74 but the PRR was 1.49 (95% CI 1.22 – 1.75 in Medtronic Valiant).
      Table 2Comprehensive analysis of proportional reporting ratios to assess associations between device type and non-a type III endoleaks based on the reports in the Food and Drug Administration’s Manufacturer and User Facility Device Experience (MAUDE) database between 2010 and 2020
      Device typeYear of first reportTotal reportsNon-a type III endoleaksPercentage of reports related to non-a type III endoleaksPRR (95% CI)Chi squared
      Medtronic Talent2010991363.631.27 (0.92–1.62)1.80
      Gore TAG20101 328372.790.93 (0.58–1.28)0.17
      Medtronic Valiant20112 520953.771.49 (1.22–1.75)8.74
      Cook TX2201138992.310.77 (0.11–1.43)0.60
      Gore CTAG20121 479191.280.38 (0–0.85)18.1
      Bolton RELAY2012319123.761.29 (0.71–1.86)0.74
      Cook Alpha201630292.981.00 (0.35–1.66)<0.01
      Data are presented as n, unless stated otherwise. PRR = proportional reporting ratio; CI = confidence interval; TEVAR = thoracic endovascular aortic repair.
      In the conservative analysis, the proportion of reports related to type IIIb endoleaks ranged from 0.7% (Gore CTAG) to > 2.5% (2.51% for Bolton RELAY, and 2.54% for Medtronic Valiant; Table 3). The lowest PRR was 0.36 (95% CI 0 – 1.03 in Gore TAG). The Medtronic Valiant met the Evans criteria with a PRR of 2.07 (95% CI 1.72 – 2.42) and a chi squared value of 17.3.
      Table 3Conservative analysis of proportional reporting ratios to assess associations between device type and type IIIb endoleaks based on the reports in the Food and Drug Administration’s Manufacturer and User Facility Device Experience (MAUDE) database between 2010 and 2020
      Device typeYear of first reportTotal reportsType IIIb endoleaksReports related to type IIIb endoleaks – %PRR (95% CI)Chi squared
      Medtronic Talent2010991181.821.10 (0.60–1.59)0.13
      Gore TAG20101 32890.680.36 (0–1.03)9.84
      Medtronic Valiant20112 520642.542.07 (1.72–2.42)17.3
      Cook TX2201138982.051.24 (0.53–1.95)0.36
      Gore CTAG20121 479110.740.39 (0–1.01)9.81
      Bolton RELAY201231982.511.53 (0.82–2.24)1.39
      Cook Alpha201630251.660.99 (0.10–1.87)<0.01
      Data are presented as n, unless stated otherwise. PRR = proportional reporting ratio; CI = confidence interval.
      When the cumulative type IIIb endoleak reports were analysed chronologically for Medtronic Valiant, which met the Evans et al. criteria for a signal of association, there were three or more type IIIb endoleak related reports available for analysis starting in 2012 (Table 4). As of 2013, when there were 10 type IIIb endoleak related reports, Medtronic Valiant met the Evans et al. criteria for a signal of association between device type and type IIIb endoleak, with a PRR of 2.64 and a chi squared of 5.60. In the seven years from the end of 2013 till the end of 2020, an additional 54 type IIIb endoleak related reports were filed.
      Table 4Medtronic Valiant type IIIb endoleak related reports, proportional reporting ratio (PRR), and chi squared by year based on the reports in the Food and Drug Administration’s Manufacturer and User Facility Device Experience (MAUDE) database between 2010 and 2020
      YearCumulative reportsCumulative type IIIb endoleak related reportsPRRChi squared
      201224132.201.32
      2013454102.645.60
      2014761172.819.06
      20151 174293.1116.1
      20161 421322.4813.1
      20171 674372.3814.2
      20181 987442.2114.2
      20192 255522.0013.0
      20202 520642.0717.3
      Data are presented as n, unless stated otherwise. PRR = proportional reporting ratio.

      Duke institutional data

      Among 542 TEVARs done at the institution for aneurysms or aneurysmal degeneration of chronic dissections since 2002, 232 were done in female patients (42.8%). The median age at the time of TEVAR was 69 years (interquartile range [IQR] 59, 76). Nearly one third were done in Black patients (n = 165, 30.4%). Most patients had hypertension (n = 496, 91.5%) and hyperlipidaemia (n = 361, 66.6%); coronary artery disease (n = 165, 30.4%), chronic obstructive pulmonary disease (n = 164, 30.3%), and peripheral arterial disease (n = 143, 26.4%) were also common. Among the institutional TEVARs, 361 used Gore devices, 85 Cook Zenith TX2, 57 Medtronic Talent, 32 Medtronic Valiant, 15 Cook Zenith Alpha, and seven Bolton RELAY devices (multiple device types were used in some cases). The median number of devices implanted per patient was two (IQR 1, 3). Median follow up was 44.7 months (IQR 13.1, 81.1) and 38 (7.0%) patients had less than one month of follow up. There were eight type IIIb endoleaks (1.2% incidence), which developed between 617 and 4 046 days following initial endograft placement (Table 5). Two followed TEVAR with a Medtronic Talent, two with a Cook Zenith TX2, two with a Medtronic Valiant, one with a Cook Zenith Alpha, and one with a Bolton RELAY device. One led to aneurysm rupture and three to symptomatic expansion; six required relining and the other two are being surveilled (Fig. 1).
      Table 5Type IIIb endoleaks over 542 thoracic endovascular aortic repairs (TEVARs) for aneurysm indication at Duke University Hospital, Durham, NC, USA
      Patient numberAge, sex at time of TEVARIndicationEndograft typeGraft components – nTime between index TEVAR and type IIIb endoleak – dType IIIb presentationRe-intervention required
      168, MHybrid TAAAMedtronic Valiant3617Symptomatic expansionRelining
      269, MHybrid TAAAZenith TX25908Aneurysm expansionRelining
      357, FDTAMedtronic Talent21 841Symptomatic expansionRelining
      476, MDTABolton RELAY11 898Aneurysm expansionSurveillance
      564, MHybrid TAAAMedtronic Valiant22 671Symptomatic expansionRelining
      658, MDTAZenith Alpha22 839Aneurysm expansionSurveillance
      769, FHybrid TAAAMedtronic Talent22 987Aneurysm expansionRelining
      832, MDTAZenith TX224 046Aneurysm ruptureRelining
      DTA = descending thoracic aneurysm; TAAA = thoraco-abdominal aortic aneurysm.
      Figure 1
      Figure 1Example of Medtronic Talent stent graft with multiple wire frame fractures and type IIIb endoleak shown in computed tomography angiography images from patient number 7 (). (A, B) 3D reconstruction images demonstrating multiple wire fractures of the endograft frame (arrows) with associated type IIIb endoleak. (C) Axial image demonstrating re-pressurised and markedly expanded aneurysm sac.

      Discussion

      Type IIIb endoleaks are a source of aortic expansion, re-intervention, and death that may potentially be reduced if specific device types can be identified as more prone to developing type IIIb endoleak. In this analysis of more than 7 000 MAUDE reports, there appeared to be an association between the Medtronic Valiant device and type IIIb endoleaks (following PRR criteria). Moreover, the Medtronic Valiant device met PRR criteria seven years before the Medtronic Valiant Navion was recalled due to concerns about type IIIb endoleaks. In addition, the PRRs for the Bolton RELAY device were numerically higher than the other devices in both the comprehensive and conservative analyses. This pattern was also borne out in examination of TEVAR data from the local institution, in which type IIIb endoleaks were disproportionately observed among Medtronic Talent, Medtronic Valiant, Cook Zenith TX2, Cook Zenith Alpha, and Bolton RELAY devices despite the relative infrequency of their use compared with Gore devices.
      The Medtronic Valiant Navion graft was recalled in early 2021 due to concerns about stent fracture and type III endoleak.
      Food and Drug Administration
      Medtronic recalls Valiant Navion Thoracic Stent Graft System due to risk of stent fractures and type III endoleaks.
      The recall notice and subsequent follow up communication reported only eight type IIIb endoleaks in relation to the issue, which were discovered through review of clinical trial and commercial patients, yet the analysis of MAUDE yielded an apparent 64 type IIIb endoleaks among the Medtronic Valiant device reports (which included both Captivia and Navion devices).
      Food and Drug Administration
      Medtronic recalls Valiant Navion Thoracic Stent Graft System due to risk of stent fractures and type III endoleaks.
      ,

      de Jesus Hernandez E. UPDATED Patient management recommendations Medtronic Valiant Navion Thoracic Stent Graft System global voluntary product recall. Available at: https://www.medtronic.com/content/dam/medtronic-com/global/HCP/Documents/valiant-navion-urgent-medical-device-recall-may-2021-update.docx [Accessed 11 October 2022].

      The recall notice did not describe the criteria that were used to determine which events were considered “device related”, but the large number of Medtronic Valiant reports related to type IIIb endoleaks in MAUDE raises questions about whether the recall might have been issued earlier. Further investigation at the time at which the signal of association between device type and type IIIb endoleak developed in 2013 could have prevented subsequent type IIIb endoleaks, re-interventions, and possibly deaths. While manufacturers and certain medical facilities are mandated to report qualifying events, the process by which reports are analysed for the presence of problematic patterns requires further scrutiny. In fact, the FDA does not document their review processes for MAUDE data and details are opaque concerning the frequency, criteria, and consequences of report reviews.
      Given the concerns about delayed identification of a device related issue raised by this analysis, it is important to highlight the relatively elevated proportions of MAUDE reports related to type IIIb endoleaks and non-a type III endoleaks for the Bolton RELAY device. Though lower than the Medtronic Valiant and the Bolton RELAY devices, the Cook Zenith TX2 device also had a relatively elevated PRR in the conservative analysis compared with the Gore devices. Although elevated PRRs are not cause for recall in and of themselves, they may suggest devices deserving of increased scrutiny. At the same time, the PRRs associated with Gore devices were relatively lower, with concomitant large chi squared values suggesting a relatively lower occurrence of type IIIb endoleaks vs. what would be expected based on the other devices. Of interest, the Medtronic, Bolton, and Cook devices are made of woven polyester fabric sutured to the metal stent and all are available in some conformations with bare stent portions proximally, distally, or on both ends. In contrast, the Gore devices are made of expanded polytetrafluoroethylene bonded rather than stitched to the stent and do not have bare stents. It is therefore likely that something about the combination of the fabric used (woven polyester), the method of fixation (suturing), and possibly the presence of bare metal stents might predispose the Medtronic, Bolton, and Cook devices to type IIIb endoleak formation. In fact, Medtronic speculated that the type IIIb endoleaks seen with Valiant Navion grafts may be due to “loss of suture integrity, which could lead to separation of the longitudinal seam of the stent graft or stent ring detachment from the surface of the graft fabric.”

      de Jesus Hernandez E. UPDATED Patient management recommendations Medtronic Valiant Navion Thoracic Stent Graft System global voluntary product recall. Available at: https://www.medtronic.com/content/dam/medtronic-com/global/HCP/Documents/valiant-navion-urgent-medical-device-recall-may-2021-update.docx [Accessed 11 October 2022].

      ,
      • Verzini F.
      • Cieri E.
      • Kahlberg A.
      • Sternbach Y.
      • Heijmen R.
      • Ouriel K.
      • et al.
      A preliminary analysis of late structural failures of the Navion stent graft in the treatment of descending thoracic aortic aneurysms.
      A mechanical failure analysis of explanted second generation thoracic endografts by Bussmann et al. found that most damage was related to either compression (presumably of the endograft within the delivery system) or to cyclical movements among various graft components and between the graft and the aortic wall.
      • Bussmann A.
      • Heim F.
      • Delay C.
      • Girsowicz E.
      • Del Tatto B.
      • Dion D.
      • et al.
      Textile aging characterization on new generations of explanted commercial endoprostheses: a preliminary study.
      Unfortunately, this analysis did not compare wear patterns between graft types. Of note, though the Cook devices are similar in overall construction to the Medtronic Valiant and the Bolton RELAY devices, they use specifically Dacron branded fabric, which may account for the relatively lower type IIIb or non-a type III endoleak PRRs compared with the Medtronic Valiant and the Bolton RELAY devices. This hypothesis is supported by further work done by the same group showing that the extent of damage caused by the fabric infolding necessary for device storage and delivery differs significantly between monofilament and multifilament fabric, with less extensive damage with multifilament fabric (though damage was incurred in both cases).
      • Chenesseau B.
      • Heim F.
      • Pidancier C.
      • Lejay A.
      • Thaveau F.
      • Georg Y.
      • et al.
      How compression inside a delivery system can degrade the cover of aortic endografts.
      The Medtronic and Bolton devices use monofilament fabric while Cook devices use multifilament fabric.
      Though this rationale explaining the observed results is hypothetical and the Medtronic Valiant Navion stent graft has been recalled, the potential that other devices (especially those featuring woven fabric sutured to a stent) are contributing to potentially avoidable aortic expansion and rupture bears further investigation. This may take the form of mechanistic ex vivo studies of wear patterns under stress over time or analyses of large databases such as the Society of Thoracic Surgeons Adult Cardiac Surgery database or the Society for Vascular Surgery Vascular Quality Initiative. Confirming that certain stent graft types are more prone to type IIIb endoleaks does not necessarily mean that those devices should be withdrawn from the market. Instead, understanding the relationship between device type, type IIIb endoleak development, and clinical outcomes will facilitate better and more concrete balancing of the risks and benefits of using certain device types. For instance, there may be situations in which the elevated risk of type IIIb endoleak is outweighed by other benefits of specific devices. Patients implanted with higher risk devices for type IIIb endoleaks may need to undergo more frequent surveillance imaging, as now recommended by Medtronic for patients with Medtronic Valiant Navion devices. Regardless of how higher risk devices are approached, graft related decision making should ideally be informed by full and up to date knowledge about the risks associated with various stent graft types, suggesting that ongoing monitoring of databases such as MAUDE may be necessary.
      This analysis has several limitations. Though the MAUDE dataset is large, it does not allow for analyses of the relative frequency of stent graft type use that would enable direct device to device comparisons. Similarly, though manufacturers and certain facilities are mandated to report events, the enforcement mechanism for that mandated reporting is unclear and it is likely that the actual incidence of type IIIb endoleaks is higher than reported. However, it is unlikely that one device would be affected by this more than any other, so the conclusions of the MAUDE analysis remain valid. Also, the level of detail varied widely between reports, sometimes rendering the distinction between different type III endoleak subtypes difficult or impossible to ascertain. This can be seen in the fact that the percentage of type III endoleaks that could be subcategorised definitively varied widely between device types. This issue was addressed by repeating the analysis once assuming that all type III undetermined endoleaks were type IIIb endoleaks, and once assuming none were type IIIb endoleaks. This difficulty subcategorising reports was also problematic for the Medtronic devices, of which two subcategories of the Valiant brand existed at the same time (Navion and Captivia). Unfortunately, nearly a quarter of Valiant reports during the three years during which both were on the market were not subcategorisable, requiring combination of both subtypes. Third, given the variable level of detail in the reports, the theoretical possibility exists that some reports were unrecognised duplicates of each other. This variable level of detail also limited the ability to exclude or adjust for patient or procedural characteristics such as indication for implantation or anatomic characteristics. Similarly, the MAUDE analyses could not be clustered by institutions or clinicians, and so it is possible that institution or clinician level characteristics could distort these findings. Finally, type IIIb endoleak occurred too infrequently in the institutional data to conduct inferential statistics, and the conclusions drawn from it may therefore be subject to confounding.

      Conclusions

      In more than 7 000 reports in the MAUDE dataset, the Medtronic Valiant device had a signal of association with type IIIb endoleaks based on number of reports, PRR, and chi squared values. Chronological analysis suggests that this signal could have been uncovered seven years prior to the recently issued recall. In addition, the finding of numerically higher PRRs among the device types sharing certain manufacturing characteristics (stitched on woven polyester fabrics with bare metal components) raises the concern that other TEVAR stent graft types could also be associated with type IIIb endoleaks. Further investigations in datasets that allow direct comparisons between devices on the basis of frequency of use and or ex vivo studies of device resilience to mechanical stresses are needed to confirm and contextualise the findings in this analysis.

      Conflict of interest

      None.

      Funding

      E.H.W. receives support from the National Heart, Lung, And Blood Institute of the National Institutes of Health under Award Number F32HL151181. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

      Appendix A. Supplementary data

      The following is the Supplementary data to this article:
      Supplementary Figure S1

      References

        • Kalesan B.
        • Cheng T.W.
        • Farber A.
        • Zuo Y.
        • Kalish J.A.
        • Jones D.W.
        • et al.
        Readmissions after thoracic endovascular aortic repair.
        J Vasc Surg. 2018; 68: 372-382
        • Biancari F.
        • Mariscalco G.
        • Mariani S.
        • Saari P.
        • Satta J.
        • Juvonen T.
        Endovascular treatment of degenerative aneurysms involving only the descending thoracic aorta: systematic review and meta-analysis.
        J Endovasc Ther. 2016; 23: 387-392
        • Vaaramaki S.
        • Suominen V.
        • Pimenoff G.
        • Saarinen J.
        • Uurto I.
        • Salenius J.
        Long-term experience of endovascular repair for thoracic aortic aneurysms and dissections.
        Vasc Endovascular Surg. 2016; 50: 335-342
        • Kamman A.V.
        • Jonker F.H.
        • Nauta F.J.
        • Trimarchi S.
        • Moll F.L.
        • van Herwaarden J.A.
        A review of follow-up outcomes after elective endovascular repair of degenerative thoracic aortic aneurysms.
        Vascular. 2016; 24: 208-216
        • Nozdrzykowski M.
        • Luehr M.
        • Garbade J.
        • Schmidt A.
        • Leontyev S.
        • Misfeld M.
        • et al.
        Outcomes of secondary procedures after primary thoracic endovascular aortic repair.
        Eur J Cardiothorac Surg. 2016; 49: 770-777
        • Food and Drug Administration
        Medtronic recalls Valiant Navion Thoracic Stent Graft System due to risk of stent fractures and type III endoleaks.
        2021 (Available at:)
      1. de Jesus Hernandez E. UPDATED Patient management recommendations Medtronic Valiant Navion Thoracic Stent Graft System global voluntary product recall. Available at: https://www.medtronic.com/content/dam/medtronic-com/global/HCP/Documents/valiant-navion-urgent-medical-device-recall-may-2021-update.docx [Accessed 11 October 2022].

        • Mansukhani N.A.
        • Haleem M.S.
        • Eskandari M.K.
        Thoracic endovascular aortic repair adverse events reported in the Food And Drug Administration Manufacturer And User Facility Device Experience Database.
        Med Devices (Auckl). 2019; 12: 461-467
        • Food and Drug Administration
        Mandatory reporting requirements: Manufacturers, importers and device user facilities.
        2020 (Available from)
        • Hughes G.C.
        • Daneshmand M.A.
        • Swaminathan M.
        • Nienaber J.J.
        • Bush E.L.
        • Husain A.H.
        • et al.
        “Real world” thoracic endografting: results with the Gore TAG device 2 years after U.S. FDA approval.
        Ann Thorac Surg. 2008; 86: 1530-1537
        • Fujimura N.
        • Ichihashi S.
        • Matsubara K.
        • Shibutani S.
        • Harada H.
        • Obara H.
        • et al.
        Type IIIb endoleak is not extremely rare and may be underdiagnosed after endovascular aneurysm repair.
        J Vasc Interv Radiol. 2019; 30: 1393-1399
        • Evans S.J.
        • Waller P.C.
        • Davis S.
        Use of proportional reporting ratios (PRRs) for signal generation from spontaneous adverse drug reaction reports.
        Pharmacoepidemiol Drug Saf. 2001; 10: 483-486
        • Aggarwal N.
        • Chow R.
        Real world adverse events of interspinous spacers using manufacturer and user facility device experience data.
        Anesth Pain Med (Seoul). 2021; 16: 177-183
        • Verzini F.
        • Cieri E.
        • Kahlberg A.
        • Sternbach Y.
        • Heijmen R.
        • Ouriel K.
        • et al.
        A preliminary analysis of late structural failures of the Navion stent graft in the treatment of descending thoracic aortic aneurysms.
        J Vasc Surg. 2021; 74: 1124-1134
        • Bussmann A.
        • Heim F.
        • Delay C.
        • Girsowicz E.
        • Del Tatto B.
        • Dion D.
        • et al.
        Textile aging characterization on new generations of explanted commercial endoprostheses: a preliminary study.
        Eur J Vasc Endovasc Surg. 2017; 54: 378-386
        • Chenesseau B.
        • Heim F.
        • Pidancier C.
        • Lejay A.
        • Thaveau F.
        • Georg Y.
        • et al.
        How compression inside a delivery system can degrade the cover of aortic endografts.
        Ann Vasc Surg. 2018; 51: 150-159

      Comments

      Commenting Guidelines

      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.
      • Comments must be in compliance with our Terms & Conditions.
      • Comments are not peer-reviewed.