The purpose of this study was to determine whether intrasac spectral Doppler flow velocities can predict whether or not a type II endoleak will spontaneously seal and to relate intrasac flow to preoperative branch vessel anatomy. Between October 1996 and June 2002, 265 patients with abdominal aortic aneurysms underwent endovascular repair. Patients with less than 24 months of follow-up and type I endoleaks were excluded. Type II endoleaks were confirmed with duplex scan and computed tomographic angiography. Two groups were identified: 14 patients with sealed endoleaks (<6 months) without intervention and 16 patients with persistent endoleaks greater than 6 months and without resolution. Spectral Doppler flow velocities were recorded from endoleaks within the aneurysm sac. The two groups were similar in age, demographics, and aneurysm morphology. The mean follow-up times were 29.9 +/- 7.9 months for sealed endoleaks and 30.2 +/- 8.6 months for persistent endoleaks (P = not significant). Spectral Doppler velocities were significantly lower in patients with sealed endoleaks compared with persistent endoleaks (75.5 +/- 78.8 cm/s versus 138.2 +/- 36.2 cm/s; P <.01). Patients with sealed endoleaks and low (<100 cm/s) intrasac Doppler velocities had significantly fewer patent inferior mesenteric arteries (43% versus 81%; P <.01), a smaller inferior mesenteric artery (5.6 +/- 1.8 mm versus 7.2 +/- 1.3 mm; P <.01), and fewer paired lumbar arteries (1.3 +/- 0.8 versus 2.4 +/- 0.6; P <.0001) compared with those with persistent endoleaks and high (>100 cm/s) intrasac flow velocities. Three patients with sealed endoleaks had Doppler velocities of 200 cm/s or greater. However, the diameter of the inferior mesenteric artery in these patients was 4 mm or less with no visualized lumbar arteries before surgery. Aneurysm diameter(-4.6 +/- 5.6 mm) and volume (-0.9 +/- 45.2 mL) decreased in patients with sealed endoleaks. Aneurysm diameter (1.8 +/- 4.9 mm) and volume (18.5 +/- 33.9 mL) increased slightly in patients with persistent endoleaks (P <.05). No ruptures or conversions occurred in any patient. Secondary interventions to treat type II endoleaks were unsuccessful in six of 16 patients (38%) with persistent endoleaks. Intrasac Doppler velocities can be used to predict whether a type II endoleak will spontaneously seal. High-velocity type II endoleaks are related to preoperative large branch vessel diameter and number and are resistant to endovascular treatment.