IV contrast material for abdominal CT: comparison of three methods of administration.

Despite a decade of experience, there is still no consensus as to the optimal IV contrast regimen for use in combined abdominal and pelvic CT scanning. In order to determine which regimen is most effective, 90 patients undergoing CT were prospectively randomized into one of three groups, depending on the method by which IV contrast material was administered: (1) a single bolus (150 ml or 175 ml, depending on the patient's weight) started when scans were made at the level of the dome of the diaphragm; (2) a split bolus delivered by means of a power injector, with the first bolus (100 or 125 ml) given when scans were made at the level of the dome of the diaphragm and the second bolus (50 ml) given when scans were made at the level of the iliac crest; (3) an initial hand-delivered bolus (100 or 125 ml) given when scans were made at the level of diaphragm, followed by rapid IV drip infusion of 50 ml throughout the remainder of the study. Quantitative comparison of pre- and postcontrast scans was performed at two levels: at the mid-liver to assess hepatic enhancement and 1 cm above the sacrosciatic notch to assess pelvic vascular enhancement. The single bolus provided better mean liver enhancement (46 H) than did either the split-bolus (36 H) or the bolus-drip (32 H) method (p less than .05). The last two methods achieved a sufficient aortocaval difference (greater than 10 H) to allow for evaluation of the liver in the nonequilibrium phase of contrast enhancement in which lesion detection is thought to be optimal. Mean enhancement of pelvic vessels was significantly better with the split bolus (arterial enhancement of 56 H, venous enhancement of 47 H) than with the single bolus (34 H, 31 H) or bolus-drip infusion (38 H, 35 H) (p less than .05). We conclude that the split-bolus method is optimal for routine combined abdominal and pelvic CT scanning. The bolus-drip method is the least effective method for administering the contrast material.