Purpose: To identify preoperative anatomical aortic characteristics that predict seal failures after endovascular aneurysm sealing (EVAS) and compare the incidence of events experienced by patients treated within vs outside the instructions for use (IFU). Methods: Of 355 patients treated with the Nellix EndoVascular Aneurysm Sealing System (generation 3SQ+) at 3 high-volume centers from March 2013 to December 2015, 94 patients were excluded, leaving 261 patients (mean age 76±8 years; 229 men) for regression analysis. Of these, 83 (31.8%) suffered one or more of the following events: distal migration ⩾5 mm of one or both stent frames, any endoleak, and/or aneurysm growth >5 mm. Anatomical characteristics were determined on preoperative computed tomography (CT) scans. Patients were divided into 3 groups: treated within the original IFU (n=166), outside the original IFU (n=95), and within the 2016 revised IFU (n=46). Categorical data are presented as the median (interquartile range Q1, Q3). Results: Neck diameter was significantly larger in the any-event cohort vs the control cohort [23.7 mm (21.7, 26.3) vs 23.0 mm (20.9, 25.2) mm, p=0.022]. Neck length was significantly shorter in the any-event cohort [15.0 mm (10.0, 22.5) vs 19.0 mm (10.0, 21.8), p=0.006]. Maximum abdominal aortic aneurysm (AAA) diameter and the ratio between the maximum AAA diameter and lumen diameter in the any-event group were significantly larger than the control group (p=0.041 and p=0.002, respectively). Regression analysis showed aortic neck diameter (p=0.006), neck length (p=0.001), and the diameter ratio (p=0.011) as significant predictors of any event. In the comparison of events to IFU status, 52 (31.3%) of 166 patients in the inside the original IFU group suffered an event compared to 13 (28.3%) of 46 patients inside the 2016 IFU group (p=0.690). Conclusion: Large neck diameter, short aortic neck length, and the ratio between the maximum AAA and lumen diameters are preoperative anatomical predictors of the occurrence of migration (⩾5 mm), any endoleak, and/or aneurysm growth (>5 mm) after EVAS. Even under the refined 2016 IFU, more than a quarter of patients suffered from an event. Improvements in the device seem to be necessary before this technique can be implemented on a large scale in endovascular AAA repair.