The clinical significance of retrograde collateral arterial perfusion of abdominal aortic aneurysms after endovascular repair (type II endoleak) has not been completely characterized. In this study a canine model was used to analyze intra-aneurysmal pressure, thrombus histologic characteristics, endoleak patency, and radiographic appearance of type II endoleaks originating from single and multiple aneurysm side branches. Prosthetic aneurysms with an intraluminal solid-state strain-gauge pressure transducer were created in the infrarenal aorta of 14 mongrel dogs. A single collateral side branch was reimplanted in 4 animals, multiple side branches were reimplanted in 6 animals, and no side branches were reimplanted in 4 control animals. Intra-aneurysmal and systemic pressure was measured for 60 to 90 days after creation of the type II endoleak. Endoleak patency and flow were assessed with duplex ultrasound scanning and cine-magnetic resonance angiography. Histologic analysis of the intra-aneurysmal thrombus was also performed. Stent-graft exclusion reduced intra-aneurysmal pressure significantly in all animals, as compared with systemic pressure (P <. 001). All intra-aneurysmal pressure values are indexed to the systemic pressure, and are represented as a percentage of the simultaneously obtained systemic pressure, which has a value of 1.0. Type II endoleaks originating from multiple side branches exhibited significantly increased intra-aneurysmal systolic pressure, mean pressure, and pulse pressure, as compared with endoleaks derived from either a single side branch (systolic pressure: multiple, 0.70 ± 0.28 vs single, 0.50 ± 0.19; P <. 001; mean pressure: multiple, 0.78 ± 0.23 vs single, 0.59 ± 0.22, P <. 001; pulse pressure: multiple, 0.41 ± 0.25 vs single, 0.17 ± 0.15, P <. 001) or excluded control aneurysms that had no side branches and no endoleak (systolic pressure, 0.17 ± 0.09; mean pressure, 0.14 ± 0.10; pulse pressure, 0.098 ± 0.08; P <. 001). Cine-magnetic resonance angiograms and duplex ultrasound scans documented persistent patency of multiple branch endoleaks up to the time of euthanasia. In contrast, single side branch endoleaks thrombosed within 3 days (P <. 001). Thrombus in the aneurysm sac in close proximity to the endoleak contained intact red blood cells and limited fibrin. Thrombus distant from the endoleak demonstrated extensive fibrin deposition and degraded red blood cells. The canine model may be used to reliably measure intra-aneurysmal pressure in the presence of patent and thrombosed type II endoleaks. In this model 2 or more side branches are necessary to maintain persistent patency of type II endoleaks. These endoleaks are associated with significantly elevated intra-aneurysmal pressure, that is, 70% to 80% of systemic pressure. These results suggest that persistent type II endoleaks have clinical significance. Endoleaks originating from retrograde flow in the side branch vessels of the aneurysm generate significant levels of intra-aneurysmal pressure, that is, 70% to 80% of systemic pressure. At least 2 patent side branch vessels appear to be necessary to cause persistent patency of type II endoleak in the canine model. Further studies will be necessary to enable more complete characterization of retrograde endoleaks and to extend these findings to allow clinical application. However, these results suggest that persistently patent type II endoleaks are of clinical significance and may require more intensive follow-up or intervention.