Diagnostic catheter size has been progressively decreased in order to reduce complications (particularly access-site complications) and permit early ambulation after coronary angiography. However, excessive down-sizing can result in poor catheter conformation and poor imaging quality of coronary angiograms (CA). This study randomly compared the accuracy and angiographic quality (QUAL) of CA performed with 4 French (Fr) vs. 6 Fr diagnostic catheters. Injections were done manually using a low-viscosity, non-heated, low-osmolality contrast media (Iomeprol). CAs were performed via the femoral approach using Judkins catheters. Handling, torque, selectively and stability were graded from 1 (excellent) to 4 (unacceptable) by the operator. QUAL was also graded from 1 (unacceptable) to 10 (excellent) by the operator in all patients and by an independent Core laboratory in 50 patients matched for gender and weight. Between January and April 1997, a total of 405 consecutive patients were randomized. Mean age was 63.4 +/- 11.1 years and 79% were male. Clinical characteristics of patients, quality of left coronary catheter and cross-over rates (1.5% with 6 Fr vs. 3.9% with 4 Fr catheters) were similar in both groups. Using the right coronary catheters, the only difference was handling, which was found to be easier with 6 Fr catheters (1.16 +/- 0.55 vs. 1.34 +/- 0.77, respectively; p = 0.007). Similarly, handling difficulty using the pigtail catheter was the only significant difference between the two groups (1.16 +/- 0.50 vs. 1.33 +/- 0.77, respectively; p = 0.009), but no cross-over was necessary in either group. The QUAL of CA was slightly but significantly better with 6 Fr than with 4 Fr catheters but considered non-diagnostic (< 7/10) in 1.4% vs. 6.8% of left CAs (p = NS). Procedural time (21.0 +/- 7.2 minutes vs. 19.0 +/- 8.1 minutes; p = 0.007) was shorter with 4 Fr catheters, but x-ray exposure, compression times and amount of contrast media used were similar. Ambulation was obtained at 2 hours in 15.1% vs. 34.0% of patients (p < 0.001) and at 4 hours in 43.8% vs. 52.4% (p < 0.05), respectively. The incidence of the worst access-site complication (moderate hematoma) was similar (1%) in both groups. CA can be performed using 4 Fr catheters and manual injections of low-viscosity contrast media with acceptable angiographic results in the majority of cases. This is associated with a shorter procedural time and earlier ambulation, and a decreased but acceptable angiographic quality.