OBJECTIVE Our purpose was to determine the effects of apnea during end-inspiration compared with free breathing on timing examinations performed to optimize gadolinium-enhanced 3D MR angiography (MRA) of the thoracic aorta and arch vessels. METHODS Thirty patients referred for gadolinium-enhanced 3D MRA of the thoracic aorta and branch vessels underwent two timing examinations: one performed during free breathing and one during apnea at end-inspiration to replicate more closely the respiratory pattern used to obtain 3D MRA. For each, axial images at the level of the proximal neck were acquired every 2 s for 40 s, during which time 1 ml of gadolinium contrast agent followed by 20 ml of saline was infused at 2 ml/s. The time to peak arterial enhancement (Ta), time to first jugular venous enhancement (Tj), and arterio-venous window (time from peak arterial enhancement to first jugular venous enhancement; AV) were compared for the two examinations in each patient. RESULTS Overall there was no statistically significant difference in Ta, Tj, or AV between examinations performed during free breathing and apnea in end-inspiration, although a trend to delayed circulation times was observed with apnea (p = 0.2-0.3). In five patients (17%), the difference in Ta between free breathing and apnea was 4 s; in three patients (10%), the difference was 6 s. CONCLUSIONS Circulation times determined during apnea at end-inspiration may differ from those obtained during free breathing by as much as 6 s in an individual patient. This difference may account for inappropriately timed gadolinium-enhanced MR angiograms performed with timing examinations, especially when very short acquisition times and low doses of gadolinium (20 ml) are used.