Date of Award

January 2014

Document Type

Open Access Thesis

Degree Name

Master of Public Health (MPH)


School of Public Health

First Advisor

Shuangge S. Ma


Background: Endoleak has been reported as a major complication among patients who have undergone endovascular repair of their abdominal aortic aneurysm (AAA). Several studies made efforts towards identifying the risk factors for the development of endoleak through various studies and data analyses. Among all the findings, one result--smoking significantly lowers a patient's risk for an endoleak--has been confirmed multiple times. But such a result appears to be contradictory to what researchers have anticipated; yet the contradictory has not been closely studied.

Methods: Data on endoleaks from the Department of Veterans' Affairs Open Versus Endovascular Repair (OVER) randomized controlled trial was used, and 419 male subjects with smoking histories were included in the analysis. A series of logistic regression models and propensity score models was constructed with the baseline and follow-up variables. Multiple imputation techniques were utilized to minimize the impact of missing data and to improve analytical robustness. In addition, a simulation study was also undertaken to better evaluate the models above.

Results: About half (5/12) of the logistic regression models supported the significant effect on endoleak of smoking in the model, and a smaller proportion (3/18) of propensity score models indicated that smoking was a significant factor for endoleak. Missing values had an important impact on the results. Although smoking's effects were not significant for little more than half of the models, odds ratio of developing endoleak for current smokers were always less than 1 compared with non-current smokers, which may be clinically meaningful. Results from simulation studies suggest that clinical trials with larger sample size might be necessary to reach a definitive conclusion on this topic.


This is an Open Access Thesis.

Open Access

This Article is Open Access