Crystalloids versus Colloids

Much controversy exists about the role of crystalloids and colloids in fluid therapy. Proponents of colloid fluid point out that resuscitation with crystalloid solution dilutes the plasma proteins, with a subsequent reduction of plasma oncotic pressure resulting in fluid filtration from the intravascular to the interstitial compartment and the development of interstitial pulmonary edema. Proponents of crystalloid solutions have argued that albumin molecules normally enter the pulmonary interstitial compartment freely and then are cleared through the lymphatic system returning to the systemic circulation. Additional albumin should merely increase the albumin pool cleared by the lymphatics. A review of the literature by Moss and Gould^[117] confirmed that all unflawed clinical and experimental studies showed that isotonic solutions are effective plasma volume expanders for resuscitation without the addition of a variety of colloid fluids. The additional cost and potential risks of colloids compared with crystalloids is another argument against colloid administration.

Colloids used in the United States include albumin, hydroxyethyl starch (hetastarch), and dextran. Because the molecules are large, colloids usually do not cross capillary membranes and remain intravascular. Distribution of fluid throughout the body is represented by the Starling-Landis equation:

J_v = K_h A ([P_MV − P_T ] − δ[COP_MV − COP_T ])

In this equation, J_v represents the net volume of fluid moving across the capillary wall per unit of time, expressed as cubic micrometers per minute; K_h is the hydraulic conductivity for water, which is the fluid permeability of the capillary wall, expressed as cubic micrometers per minute per square micrometer of capillary surface area per 1 mm Hg pressure difference. The value of K_h increases up to fourfold from the arterial to the venous end of a typical capillary. P_MV is the capillary hydrostatic pressure; P_T is the tissue hydrostatic pressure; A is the capillary surface area; and δ is the reflection coefficient for plasma proteins. This coefficient is necessary because the plasma proteins are slightly permeable to the microvascular wall, preventing full expression of the two colloid osmotic pressures. When δ is 0, molecules freely cross the membrane; when δ is 1, molecules cannot cross the membrane. Typical δ values for plasma proteins in the microvasculature exceed 0.9 in most organs, and these values remain constant but can be decreased significantly by pathophysiologic processes, including hypoxia, inflammation, and tissue injury. COP_MV is the colloid oncotic pressure, and COP_T is the colloid oncotic pressure in the tissue.

The hydrostatic and colloid pressure differences across capillary walls (i.e., Starling forces) cause movement of water and dissolved solutes into the interstitial spaces. These movements play a minor role with regard to tissue nutrition relative to simple diffusion. The reflection coefficient that expresses the ability of the semipermeable membrane to prevent movement of a solute varies greatly among tissues. The lungs are moderately permeable relative to other organs, and during pathophysiologic processes such as surgical trauma, the reflection coefficient may further change to alter capillary permeability, resulting in increased capillary permeability or leak. In this setting, colloids move more easily into the interstitium and increase interstitial edema.

With leakage of colloid molecules into the interstitial space, further swelling of tissues occurs because of the unfavorable oncotic pressure gradient, and these molecules are removed by the lymphatic system. Removal of colloids requires longer periods than for crystalloids and is a significant problem in burn and major surgical patients. A well-known meta-analysis by Velanovich^[118] of mortality from eight studies concluded that trauma patients should be resuscitated with crystalloid solutions, whereas colloids were more effective in nonseptic, non-traumatic, elective surgical patients.

Another factor that must be considered is the effect of resuscitation fluid on coagulation. It is a well-known phenomenon that after trauma, patients may suffer from low levels of circulating coagulation factors. This deficiency may lead to transfusion of fresh frozen plasma to restore deficient factors. The clotting cascade is an extremely complex string of events, with many factors affecting the progression to clot. Studies using a thromboelastogram to measure trauma patients' clotting status have indicated that these patients may be hypercoaguable.^[119] Another study using thromboelastographic profiles showed that patients undergoing major surgery who were given intraoperative lactated Ringer's solution were also hypercoagulable when compared with baseline.^[120] These studies suggest that the type of fluid used for volume resuscitation may play a role in the coagulopathy often seen in trauma patients. However, definitive studies are still

1787