Intravenous Fluid Management

The general principles of fluid management for neurosurgical anesthesia are (1) maintenance of normovolemia and (2) avoidance of a reduction in serum osmolarity. The first principle is a derivative of the concept presented in the section "Management of Blood Pressure" that it is in general ideal to maintain normal MAP in patients undergoing most neurosurgical procedures and neurosurgical critical care.^[139] Maintaining normovolemia is simply one element of maintaining normal MAP. The second principle is a derivative of the oft-repeated observation that lowering serum osmolarity results in edema of both normal and abnormal brain. Administering fluids that provide free water (i.e., fluids that do not have sufficient non-glucose-containing solutes to render them iso-osmolar with respect to blood) will lower serum osmolarity if the amount of free water administered is in excess of that required to maintain ongoing free water loss. Half-normal saline is probably a reasonable choice for the traditional maintenance fluid allowance. However, fluids administered to replace blood and third-space loss (i.e., iso-osmolar losses) should be more nearly iso-osmolar with respect to plasma (295 mOsm/L). Normal saline and lactated Ringer's solution (LR) are often used. Normal saline is in fact slightly hyperosmolar, 308 mOsm/L. It has the disadvantage that in large volumes it can cause hyperchloremic metabolic acidosis.^[140] The physiologic significance of this acidosis, which involves the extracellular but not the intracellular fluid space, is unclear. At a minimum, it has the potential to confuse the diagnostic picture when acidosis is present. As a result, many clinicians use predominately LR. Although LR (273 mOsm/L) is in theory not ideal for either maintenance fluids or replacement of losses individually, it serves as an entirely reasonable compromise for meeting both needs simultaneously and is very suitable in most instances. However, it is a hypoosmolar fluid, and in healthy experimental animals, it is possible to reduce serum osmolarity and produce cerebral edema with a large volume of LR.^[141] Therefore, in the setting of large-volume fluid administration, such as significant blood loss and multiple trauma, it is our practice to alternate, liter by liter, LR and normal saline.

The crystalloid versus colloid discussion is a recurrent one that usually arises in the context of a head injury victim. Despite numerous fervent beliefs regarding this issue, there has in fact been only a single demonstration that a reduction in colloid oncotic pressure in the absence of a change in osmolarity can actually contribute to augmentation of cerebral edema in the setting of experimental head injury.^[142] The transcapillary membrane pressure gradients that can be produced by a reduction in colloid oncotic pressure are in fact very small by comparison to those created by changes in serum osmolarity. Nonetheless, it appears that these small gradients do, probably in the setting of a blood-brain barrier injury of intermediate severity, have the potential to augment edema. It seems reasonable, then, to select a fluid administration pattern that in addition to maintaining normal serum osmolarity, will prevent substantial reductions in colloid oncotic pressure. For the large majority of elective craniotomies, administration of colloid solutions will not be required. However, in situations requiring substantial volume administration (multiple trauma, aneurysm rupture, cerebral venous sinus laceration, fluid administration to support filling pressure during barbiturate coma), a combination of isotonic crystalloid and colloid may be appropriate.

Which colloidal solutions should be used? Albumin is a reasonable choice. Dextran-containing solutions are generally avoided because of their effects on platelet function. The various starch-containing solutions should be used cautiously in neurosurgery because in addition to a dilutional reduction of coagulation factors, they interfere directly with both platelets and the factor VIII complex.^{[143] [144] [145] [146]} Note that the literature addressing the coagulation effects of starches should be read with the understanding that the effects on coagulation are proportional to the average molecular weight of the starch preparation. Unfortunately, the preparations in use in North America are those with the greatest molecular weights. Accordingly, although hetastarch solutions will be used in neuroanesthesia, the clinician should respect the manufacturer's recommended dose restrictions and use additional restraint in situations with other reasons for impairment of the coagulation mechanism. Several reported instances of bleeding in neurosurgical patients have been attributed to hetastarch administration. However, all of them have involved circumstances in which the manufacturer's recommended limit of 20 mL/kg/day

2144

Substantial current interest has been generated in the use of hypertonic fluids, in particular, in the resuscitation of a multiple-trauma victim. ^[149] It appears unlikely that the ICP effect of these fluids on the brain is substantially different from an equiosmolar exposure to mannitol.^{[150] [151]} Accordingly, when decisions about the appropriateness and relevance of these fluids in the resuscitation process are eventually made, it seems unlikely that these decisions should be made on the basis of specific cerebral effects. It seems more likely that the final judgment regarding this class of fluids will be made on the basis of their effects on the systemic circulation.^[150] In general, an intervention that has the advantage of more effectively restoring systemic hemodynamics is likely to be advantageous to the injured brain. That assertion is made with the proviso that sustained hyperosmolarity (e.g., >320 mOsm/L) caused by any fluid has the potential to result in rebound swelling of the brain.