SAVS Main Site  |  Past & Future Meetings
Southern Association For Vascular Surgery

Back to 2022 Posters


The association between device type and type IIIb endoleaks following TEVAR
Elizabeth H Weissler, Bradley G. Hammill, Jenna L. Armstrong, Andrew M Vekstein, Parth Chodavadia, Chandler Long, Matthew Roe, G. Chad Hughes
Duke University, Durham, NC

Introduction: Endoleaks following thoracic endovascular aortic repair (TEVAR) can lead to need for reintervention and aortic rupture. In the setting of a recent recall of a specific thoracic stent graft type due to concerns about type IIIb endoleak, we sought to characterize patterns of type IIIb endoleaks among available TEVAR devices.
Methods: The Food and Drug Administration’s Manufacturer and User Facility Device Experience (MAUDE) database was queried for all reports related to each thoracic stent graft type between 2010 and 2020. Reports concerning endoleaks were categorized by type, including type IIIa, type IIIb, and type III-undetermined. Proportional reporting ratios (PRR) and chi-squared tests were used to assess for the presence of a signal of association between device type and type IIIb endoleaks or “non-a” type III endoleaks (consisting of type IIIb and type III-undetermined endoleaks). Following standard practice, a PRR >2 and chi-squared value >4 with ≥3 reports overall constituted a signal of association. For devices meeting the criteria for a signal of association, a chronological analysis was conducted to ascertain when there would have been adequate data for the signal to have been identified.
Results: There were 7,328 MAUDE reports concerning 7 graft types for analysis. The Medtronic Valiant™ device had the most related reports, at 2,520; the Cook Alpha® device had the fewest related reports at 302. When analyzing non-a type III endoleaks, the lowest PRR was 0.38 (95%CI 0-0.85 in Gore CTAG®) and the highest was 1.49 (95%CI 1.22-1.75 in Medtronic Valiant™). No devices met the PRR criteria when non-a type III endoleak was used as the endpoint. When analyzing type IIIb endoleaks, the lowest PRR was 0.36 (95%CI 0-1.03 in Gore CTAG®) and the highest was 3.07 (95%CI 2.75-3.93 in Medtronic Valiant™). The type IIIb endoleak chi-squared value for Medtronic Valiant™ was 17.3, meeting criteria for an association. Regardless of the endpoint used, the Gore devices had the lowest PRRs while the Cook, Medtronic, and Bolton devices had numerically greater PRRs. The Medtronic Valiant™ device met PRR criteria for association in 2013, 7 years prior to the recall being issued.
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 off number of reports, PRR, and chi-squared values. Chronological analysis suggests that this signal could have been uncovered 7 years prior to the recently-issued device 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 investigation in a dataset that allows direcehweisslert 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 contextualize the findings in this analysis.

Non-a type III endoleaksType IIIb endoleaks
Device typeProportion of reports related to non-a type III endoleaksPRR (95% CI)X2 valueProportion of reports related to type IIIb endoleaksPRR (95% CI)X2 value
Medtronic Talent0.03631.27 (0.92-1.62)1.800.01821.10 (0.60-1.59)0.13
Gore TAG0.02790.93 (0.58-1.28)0.170.00680.36 (0-1.03)9.84
Medtronic Valiant0.03771.49 (1.22-1.75)8.740.02543.07 (2.75-3.93)17.3
Cook TX20.02310.77 (0.11-1.43)0.600.02051.24 (0.53-1.95)0.36
Gore CTAG0.01280.38 (0-0.85)18.10.00740.39 (0-1.01)9.81
Bolton RELAY0.03761.29 (0.71-1.86)0.740.02511.53 (0.82-2.24)1.39
Cook Alpha0.02981.00 (0.35-1.66)<0.010.01660.99 (0.10-1.87)<0.01
PRR: Proportional reporting ratio (>2 considered to be a signal of association), X2: chi-squared (>4 considered to be a signal of association)


Back to 2022 Posters