Comparison of crystalloids and colloids for hemodilution in a model of focal cerebral ischemia.

Forty dogs were subjected to 6 hours of occlusion of the left internal carotid and middle cerebral arteries. They were divided into two "hemodilution groups" of 13 dogs each and a control "nonhemodiluted group" of 14 dogs. Thirty minutes after arterial occlusion, isovolemic hemodilution was performed by phlebotomy and infusions of low-molecular weight (MW) dextran in one group and of lactated Ringer's solution in the other group. The animals were sacrificed 1 week after temporary arterial occlusion. Hemodilution reduced the hematocrit to a level of 33% to 34%, which lasted throughout the week in both groups. After hemodilution there was a very significant reduction in blood viscosity, plasma total protein content, and fibrinogen levels in both groups in the acute stage; these levels gradually returned to baseline by the end of the week. In the group with lactated Ringer's solution hemodilution, both osmotic and oncotic pressures were decreased by hemodilution in the acute stage. In the control and low-MW dextran groups, osmotic and oncotic pressure remained unaltered throughout the week. Hemodilution resulted in a slight decrease in mean arterial blood pressure in all groups in the acute stage, but there were no significant changes in central venous, pulmonary arterial, or pulmonary wedge pressures. During the week of study, there were no differences in the cardiac index and total blood volume between the groups, and no significant changes in hematological parameters with the exception of a slight increase in bleeding time immediately after hemodilution with low-MW dextran. Daily neurological assessment showed consistently poorer condition during the first 5 days in the group with lactated Ringer's solution compared to either the control group or the group receiving low-MW dextran. Based on Mann-Whitney U-testing, the infarct volume of the lactated Ringer's solution recipients, expressed as a percentage of the total volume of that hemisphere (median 15.7%, range 6.6% to 25.2%) was significantly larger than that of the group receiving low-MW dextran (median 2.2%, range 0% to 15.8%) and that of the control group (median 11.9%, range 0% to 39.9%). The results indicate that, in this model, hemodilution with colloids was beneficial, whereas hemodilution with crystalloids was deleterious. It is likely that the decrease in oncotic pressure observed after hemodilution with lactated Ringer's solution is one of the most important reasons for its detrimental effect.