Letter by Fillinger Regarding Article, “Predictors of Abdominal Aortic Aneurysm Sac Enlargement After Endovascular Repair”
To the Editor:
I read with interest the article by Schanzer et al.1 I am currently analyzing the M2S database as the co-Principal Investigator on the Characterization of Human Aortic Anatomy Project with Dr Tina Morrison of the US Food and Drug Administration. In our opinion, several issues may confound the interpretation of findings by Schanzer and colleagues.
The M2S database is not a consecutive, observational series. By survey, less than one third of users send consecutive patient series to M2S. Most send selected cases that are unusual, difficult, or known to have suboptimal outcomes. How can one draw conclusions about the percentage that meet the anatomic instructions for use in the real world?
A large percentage of normal anatomy is purposely excluded. M2S does not release sponsored data. More than 4000 endovascular aortic aneurysm repairs meeting the instructions-for-use criteria in the database were excluded from the dataset sent to Schanzer et al. Thus, the database reported by Schanzer et al excludes 30% of the actual cohort meeting their criteria, and approximately half of all cases meeting the device instructions-for-use criteria. Most of these cases were commercially available devices in postmarket studies and registries in the real world.
The database is not solely abdominal aortic aneurysms (AAAs). M2S labels all aortoiliac reconstructions as “AAA.” Examining actual computed tomography (CT) scans and records (with institutional consent) revealed that 15% of cases labeled AAA and >3 cm were not, in fact, managed for AAA, but rather pseudoaneurysm, dissection, ulcer, thoracoabdominal aneurysm, etc. The article states “31013 CT scans were reviewed.” By whom? The authors had access to measurements, but not the CT images, so how could they evaluate the primary pathology or the actual percentage of small aortic aneurysms in the database?
Measurement methods in this study were not reported in natural history studies. Maximum sac diameter in the M2S database is measured orthogonal to the aneurysm lumen, not on a 2D axial cross section, which is on average 4 mm larger. Thus, at 5.9 cm mean diameter, most AAAs did meet the standard criteria. Schanzer et al did not have access to the CT images, 2D measurements, rate of expansion, or symptom status. Noting points 1 to 4, the percentage of cases meeting the standard threshold for operation cannot be calculated.
Cases were submitted or withdrawn based on the outcome being studied. By survey, most users submit postoperative CT scans to M2S only when the aneurysm is not shrinking by diameter, or they stop submitting CT scans if the aneurysm shrinks. This leads to a high percentage of cases with expanding aneurysms, and limits available follow-up. This study had 61% of postoperative scan data at 1 year, and 4% at 5 years, and patients were withdrawn based on the studied outcome. How is life table analysis or conclusions about “percent enlarging” possible? Figure 2A to 2C show nearly the same rate of sac enlargement whether the aneurysm was treated “on-label” or “off-label,” probably because of these issues. Also, a lower percentage of centers routinely submitted consecutive postoperative cases during later years, explaining the results in Figure 2D. Figure 2A to 2C shows off-label use cannot explain the large difference in the curves in Figure 2D. Off-label use was less common during the later period (Table 4). Therefore, on-label cases must have much worse expansion rates in the later period also, which makes no sense. Please explain.
Sac expansion was not calculated from the first postoperative CT (per Society for Vascular Surgery reporting standards), which inflates the expansion rate relative to standard reporting methods.
I agree with the authors that this is an important area of research, and we are trying to address all of these issues with the Characterization of Human Aortic Anatomy Project in collaboration with the US Food and Drug Administration.
Mark Fillinger, MD
Dartmouth-Hitchcock Medical Center
Dr Fillinger has consulted for Cook, WL Gore, Medtronic, Endologix, and M2S, receiving <$10 000 annually, the majority going to the Dartmouth-Hitchcock Medical Center rather than to Dr Fillinger.
- © 2012 American Heart Association, Inc.