Background Patients with substantially impaired kidney function and peripheral arterial disease (PAD) underwent comparative CO2-based depiction of the pelvic arteries (PAs). Purpose To evaluate the feasibility and diagnostic performance of CO2-based C-arm computed tomography (CACT) and compare its depiction of PAs with CO2-digital subtraction angiography (DSA). Material and Methods Fifteen patients (10 men, mean age 70 ± 11 years) with PAD received CO2-DSA and CO2-CACT of the PAs, depicted from the aorta to femoral arteries. These were divided into nine segments (135 in total) and graded by two independent readers for image quality (IQ; 1 = sufficient, 2 = minimal impairments, 3 = insufficient, 4 = outside field of view) and subsequent stenosis grading (SG; grade 1: normal to grade 4: occlusion), under exclusion of all segments with insufficient IQ. Inter-observer and inter-modality agreement calculation and subsequent consensus reading were performed and correlated to a standard of reference (StOR), representing a modality consensus. Results Of 135 segments, 117 showed sufficient IQ, excluding 18 segments (10 CACT, 8 DSA). Inter-observer agreement for IQ and consecutive SG demonstrated good to excellent agreement: IQDSA: κ = 0.83, IQCACT: κ = 0.76; StenosisDSA: κ = 0.71, StenosisCACT: κ = 0.84. Inter-modality agreement for SG lay at κ = 0.76 and κ = 0.65, respectively. More stenoses could be detected by CACT, and analysis of pooled consensus values of SG in CACTcons versus StOR showed an excellent agreement (κ = 0.96) that proved considerably higher than the moderate agreement between consensus values in DSAcons versus StOR (κ = 0.43). Conclusion CO2-CACT proved feasible, and has the potential to optimize angiographic work-up of PAD in patients with contraindications for other contrast media.