Bleeding rates necessary for detecting acute gastrointestinal bleeding with technetium-99m-labeled red blood cells in an experimental model.

Proponents of [99mTc]sulfur colloid for GI bleeding studies argue that, although labeled red blood cells are useful for intermittent bleeding, they are not capable of detecting low bleeding rates. Studies of dogs with experimental GI bleeding have indicated bleeding rates of 0.05 ml/min can be detected with [99mTc]sulfur colloid. Since similar data in the dog model were unavailable for 99mTc-labeled red blood cells, we undertook this study. To simulate lower GI bleeding, catheters were inserted into the bowel lumen. Each dog's blood was labeled with 99mTc using an in vitro technique. Venous blood was then withdrawn and re-infused into the lumen of the bowel using a Harvard pump. Fourteen dogs were studied, ten receiving a bleeding rate from 4.6-0.02 ml/min in the descending colon and four with proximal jejunal bleeds of 0.20-0.02 ml/min. Bleeding rates of 4.6-0.2 ml/min were detected within 10 min in the colon and bleeding rates as low as 0.04 ml/min were seen by 55 min. Slower bleeding rates were not detected. Similar findings were noted for proximal jejunal bleeds. Based on the time of appearance, a minimum volume of approximately 2-3 ml labeled blood was necessary to detect bleeding. We conclude that 99mTc-labeled RBCs are sensitive for low bleeding rates in the dog model. The rates are comparable to those described for [99mTc]sulfur colloid in this experimental setting. The time of appearance of activity is related to the bleeding rate.