The purpose of this investigation was to determine the variability of electron-beam computed tomography (EBCT) measurement of coronary calcification by using two commonly employed image-acquisition protocols and to compare this variability with that of a new image-acquisition protocol. We performed three EBCT scans within 15 minutes on each of 324 consecutive high-risk, asymptomatic subjects participating in an ongoing epidemiologic research investigation. Subjects were divided into three groups: group 1 (n = 175) received two scans with a standard 20-slice, 3 mm slice thickness image-acquisition protocol and a third scan with the new 20-slice, 6 mm slice thickness protocol; group 2 (n = 77) received two scans with a new 6 mm slice thickness image-acquisition protocol and a third scan with a standard 20-slice, 3 mm slice thickness protocol; group 3 (n = 72) received two scans with a 30-slice, 3 mm slice thickness image-acquisition protocol and a third scan with a 20-slice, 6 mm slice thickness protocol. Calcium score, calcium mass estimate, and calcium volume estimate were determined for each scan. We compared retest variability of calcium measurements for each of the three image-acquisition protocols. The variability of the new 6 mm slice thickness protocol was significantly lower than that of either the 20-slice 3 mm slice thickness protocol (p = 0.009) or the 30-slice 3 mm slice thickness protocol (p = 0.02) for measurement of calcium score, mass, or volume. Retest reproducibility for all three image-acquisition protocols was low; however, the 20-slice 6 mm slice thickness protocol gave reproducibility superior to either of the 3 mm slice thickness protocols. Variability of absolute indices increased, and variability of relative indices decreased as the value of these indices increased. EBCT is not sufficiently reproducible to allow serial quantitation of coronary calcium in individual patients over relatively short periods (< 2 years).