Coronary CT angiography (CTA) is currently considered a reliable method to exclude obstructive coronary artery disease (CAD) before valvular heart surgery in patients with low pretest probability. However, its role in excluding obstructive CAD before transcatheter aortic valve implantation (TAVI) is less well established. Single-center retrospective study where patients with severe symptomatic aortic stenosis underwent both CTA and invasive coronary angiography (ICA) as part of TAVI planning. CTA exams were conducted on a 64-slice dual source scanner, with a median interval of 45 days to ICA (IQR 25-75 [13-82]). In both tests, obstructive CAD was defined as a ≥50% stenosis in an epicardial vessel ≥2 mm diameter. Per-patient, per-vessel and per-proximal segment analyses were conducted, excluding and including non-evaluable segments. The study included 200 patients (120 women, mean age 83 ± 6 years). The prevalence of obstructive CAD on ICA was 35.5% (n = 71). On a per-patient analysis (assuming non-evaluable segments as stenotic), CTA showed sensitivity of 100% (95% CI, 95-100%), specificity of 42% (95% CI, 33-51%), and positive and negative predictive values of 48% (95% CI, 44-51%) and 100% (95% CI, 92-100%), respectively. CTA was able to exclude obstructive CAD in 54 patients (27%), in whom ICA could have been safely withheld. Despite the high rate of inconclusive tests, pre-procedural CTA is able to safely exclude obstructive CAD in a significant proportion of patients undergoing TAVI, possibly avoiding the need for ICA in roughly one quarter of the cases.