Current Concepts in Perioperative Fluid Management

Although most aspects of perioperative fluid management have remained relatively constant for more than two decades, new concepts and practices are currently evolving. Until the mid-to-late 1960s, the prevailing strategy of perioperative fluid management was rigid restriction (1). After Shires and colleagues (2,3) convincingly demonstrated that major surgery and trauma were associated with substantial intravascular fluid deficits, perioperative volume replacement became much less restrictive. In the 1970s, attention focused on the still-unresolved crystalloid/colloid controversy. Currently, investigators are applying kinetic analysis, using principles similar to those of pharmacokinetics, to better define responses to perioperative fluid administration (4). Development of new colloid solutions, in terms of the characteristics of both the colloids and their diluents, promises to reduce problems associated with existing formulations. Research into the components of conventional crystalloid solutions may even prompt revisions in the composition of these common anesthetic tools. This review will focus on several areas of research that are changing clinical practice, including the following: 1.The kinetics of plasma volume expansion produced by IV fluids, 2.Recent developments in colloid solutions, and 3.Specific components of available fluids. The Kinetics of Plasma Volume Expansion Produced by IV Fluids Principles Conventional prediction of plasma volume expansion (PVE) after fluid infusion assumes that retained fluid (infused fluid minus excreted fluid) is distributed across body fluid spaces based on the volumes of physiologic fluid spaces (Table 1) and the Starling equilibrium, which governs the distribution of isonatremic, noncolloid fluids between plasma volume (PV) and the interstitial fluid (the two components of extracellular volume [ECV]). The Starling equilibrium is defined as follows: MATHwhere Q = fluid filtration, k = capillary filtration coefficient (conductivity of water), A = the area of the capillary membrane, Pc = capillary hydrostatic pressure, Pi = interstitial hydrostatic pressure, σ = reflection coefficient for albumin, Πi = interstitial colloid oncotic pressure, and Πc = capillary colloid oncotic pressure. The surprisingly small proportion (< 0.5) of osmotic activity contributed by plasma proteins is essential in determining the equilibrium between interstitial fluid volume (IFV) and PV. If serum osmolality is normal, total osmotic pressure exceeds 5400 mm Hg, only 24 mm Hg of which is colloid osmotic (oncotic) pressure.Table 1: Fluid Distribution Volumes The following equation predicts the static effects of fluid infusion on PVE:MATH where Vd = distribution volume (Table 1). For example, how much would PVE increase as a consequence of infusion of 500 mL of 5 dextrose in water? Because the infused volume is sodium-free, the Vd is TBW (equal to 60 of total body weight). Therefore, PVE = 500 mL × (3/42) or 36 mL. Infusion of 500 mL of lactated Ringer’s solution (LRS) or 0.9 saline, for which Vd = ECV (20 of total body weight), would result in PVE = 500 mL × (3/14) = 107 mL. If membrane permeability is normal, fluids containing colloids such as albumin, de-tran, or hydro-yethyl starch (HES) preferentially e-pand PV rather than IFV or ICV. Each gram of intravascular albumin holds 14 to 15 mL of water in the PV (5). Therefore, 500 mL of 5 albumin, containing 25 g albumin, would expand PV by approximately 375 mL. Hyperosmotic fluids, e.g., 7.5 saline, increase venous return primarily by translocating intracellular fluid into PV (6). However, hypertonic saline infusions also transiently increase ECV through osmotic attraction of fluid from the IFV into the PV. Addition of hyperoncotic colloid to the hypernatremic fluid will increase the PV increment. Kinetic analysis of PVE replaces static assumptions with dynamic data and serves the same purposes as pharmacokinetic analysis of drug concentrations, e.g., estimating peak effects and rates of clearance. Both the effects of a bolus of fluid on PVE and the rates of infusion necessary to maintain any given level of plasma dilution can be predicted by kinetic modeling (7). However, because infusion of fluid does not introduce a novel substance that results in a measurable concentration, the effects of fluid infusion on PVE must be inferred from changes in the concentrations of other variables. Svensen and Hahn (4) evaluated three endogenous tracers—blood water concentration, serum albumin concentration, and total hemoglobin (Hb)—in eight volunteers who received bolus infusions of acetated Ringer’s solution, 6 dextran 70, or 7.5 saline. Hb concentration provided simple, reproducible measurements (Fig. 1).Figure 1: The blood hemoglobin (B-hemoglobin, closed circles line), blood water (B-water, open circles line), and serum albumin (S-albumin, X line) concentrations used as markers to indicate the dilution of the plasma volume during IV infusion of 25 mL/kg of Ringer’s solution over 30 min in eight male volunteers. Data are mean values, and correction for blood sampling was not made. Note that in normovolemic, unanesthetized volunteers a substantial bolus of isotonic crystalloid (equivalent to 1750 mL in a 70-kg person) produces a modest peak in volume expansion that rapidly decreases. Modified from Svensén et al. (4) with permission of the publisher.After infusing the test fluids, Svensen and Hahn (4) used nonlinear regression of fluid-induced changes in Hb concentration to separate mathematically a clearance curve into one-volume and two-volume-of-fluid-space (VOFS) models (Fig. 2). Although these compartments do not correspond precisely to anatomic or physiologic spaces (much as pharmacologic volumes of distribution may not correspond to a precise physiologic space), kinetic analysis of crystalloid solutions is consistent with distribution into both PV and ISV. Certainly the most striking aspect of these data was the small proportion of crystalloid remaining in the vascular tree after equilibration.Figure 2: Schematic drawing of the kinetic model used to calculate the size of the body fluid spaces expanded by IV infusions of fluid in humans. Data are fitted to a one-volume or two-volume-of-fluid-space (VOFS) model. The assumptions underlying the one-compartment VOFS model (top) are as follows: 1) during fluid infusion, fluid enters an expandable space of volume v at a constant rate k i; 2) the expandable fluid space has a target volume V, which the body strives to maintain; 3) volume v changes by fluid being eliminated from the fluid space at a basal rate, k b (perspiration and basal diuresis), and at a controlled rate. The controlled rate is proportional by a constant k r to the relative deviation of v from the target volume V. The assumptions behind the two-compartment VOFS model (bottom) are similar: 1) during fluid infusion, fluid enters an expandable space of volume v 1 at a constant rate k i; 2) there is a secondary expandable fluid space of volume v 2 exchanging fluid with the primary fluid space; 3) volume v 1 changes through exchange with the secondary fluid space and as a result of fluid being eliminated from the primary fluid space at a basal rate, k b (perspiration and basal diuresis), and at a controlled rate; 4) the primary and secondary fluid spaces have target volumes V 1and V 2, which the system strives to maintain by acting on the controlled elimination mechanism k r, which is proportional to the relative deviation from the target volume of the primary fluid space, and by acting on the fluid exchange mechanism; 5) the net rate of fluid exchange between the two spaces is proportional to the difference in relative deviations from the target volumes by a constant k t. From Svensén et al. (4) with permission of the publisher.Ultimately, kinetic analysis should be useful in addressing some of the shortcomings of current perioperative fluid management as kinetic analysis further defines the interaction between the kinetics of fluid administration and the effects of surgery and trauma on surgical fluid requirements. For instance, after mild hemorrhage in conscious volunteers, a greater proportion of infused isotonic crystalloid remained in the vascular tree than after fluid infusion in normovolemic volunteers (Fig. 3) (8), perhaps because transcapillary flux of IF augments fluid retention.Figure 3: Superimposed model-predicted dilution curves for the central (left) and peripheral (right) fluid spaces when male volunteers given Ringer’s solution by IV infusion were normovolemic and when 450 mL and 900 mL blood had been withdrawn. The graphs are based on mean values obtained when the two-volume model was justified statistically. Modified from Drobin and Hahn (8) with permission of the publisher.Surgical trauma appears to result in acute sequestration of IF. In otherwise healthy patients who received sodium-free fluid during open upper abdominal surgery, ECV decreased nearly 2 L and glomerular filtration rate acutely declined by 13; in contrast, patients who received LRS maintained ECV and increased glomerular filtration rate by 10 (9). During the first ten days after resuscitation from multisystem trauma, patients demonstrated an increase in total body weight and a 55 increase in IFV (more than 5.0 L in a 70-kg patient) (10). Accumulated fluid mobilizes and returns to the PV, most commonly on the third postoperative day. If the cardiovascular and renal systems are unable to compensate, hypervolemia and pulmonary edema may occur. Although current perioperative fluid management usually avoids hypovolemia, no tools are currently available to permit precise matching of fluid administration to fluid needs. Although this may result in inadequate fluid administration, an equally important problem is e-cessive perioperative fluid infusion. Arieff (11) reported 13 episodes of fatal postoperative pulmonary edema, apparently related to large volumes of perioperative fluid. He also estimated, based on a 1-yr retrospective review of patients undergoing major surgery at two university medical centers, an annual incidence of 8,000 to 74,000 cases of postoperative pulmonary edema in the United States (11). Most troubling was the observation that 2.6 of patients without important comorbidities developed pulmonary edema and, of these, 3.9 died (11). Another troubling aspect of current fluid management is that the rapidity with which fluid is administered may influence immune function. In mice subjected to surgical trauma and shock, slower fluid resuscitation (over 120 min) was associated with more rapid restoration of normal immune function than resuscitation over 30 or 60 min (12). These studies, combined with clinical e-perience, suggest the following conclusions: 1.We can’t accurately evaluate blood volume, 2.We can’t accurately evaluate tissue perfusion, 3.We can’t accurately identify fluid overload, 4.We can’t accurately identify hypovolemia, and 5.We can’t accurately define the correct rate of fluid resuscitation. The lack of a simple, objective method of defining and achieving adequate fluid resuscitation represents an ongoing deficiency in perioperative fluid management. One concept that attempts to address this lack is restoring a target level of systemic oxygen delivery (DO2), based on evidence that unrecognized, subclinical tissue hypoperfusion results in postoperative complications (e.g., acute renal failure, hepatic failure, and sepsis) and that these complications can be reduced by DO2 enhancement. In a single term, DO2 combines cardiac output (Q) and arterial oxygen content (CaO2), according to the equation:MATH where the factor 10 corrects CaO2, in mL O2/dL, to mL O2/L. In high-risk surgical patients who survive, average Q and DO2 are greater than in those who succumb to critical illness (13). Survival is strongly associated with a DO2 ≥ 600 mL O2 · min-1 · m−2(13) (equivalent to a cardiac index of 3.0 L · min-1 · m-2, a Hb concentration of 14 g/dL, and 98 oxyhemoglobin saturation). In attempting to increase DO2, several principles should be kept in mind. First, the use of crystalloid or colloid fluids to increase Q will also Hb the net on DO2 will on the increase in Q or the in Hb concentration primarily Hb concentration with blood will result in a in the net must be infusion of necessary to DO2 effects on tissue and may have effects on other such as or to be to DO2 evidence is at this the in the of perioperative management et al. DO2 ≥ 600 mL O2 · · min-1 for resuscitation of trauma patients based on data better However, et al. in a of trauma patients at the same that trauma patients to conventional management or of DO2 ≥ 600 mL O2 · · min-1 not in patients in the conventional management who DO2 ≥ 600 mL O2 · · min-1 and patients in the who that level had a much better than patients from who to that (Table 2). in the who to the had a perhaps some of in 2: and are that used to increase DO2 to may be a of several may increase the of et al. no increased of during of oxygen delivery for Another is that may function. For example, attempts to increase systemic oxygen with not in hepatic is the observation that the of used to cardiac output may influence et al. that with was associated with complications and than conventional or using both reduced the rate in with However, of two of decreased in a of patients undergoing major abdominal surgery Although to in a in a of patients a conventional management to of DO2 using volume expansion and in patients than 60 of reported better in the other than DO2 be et al. that a for was associated with that a decreased developed days was an of was not the of patients with a with conventional The of when used in with of blood A review by and a review of the and physiologic aspects of Development The peak of between of crystalloid and those of colloid more than The was several by two of by the and of which in which colloid and crystalloid fluid management were that colloid was associated with an increase in of and that their review not the use of that the analysis was in because of the of the that were The focused on albumin, 30 clinical that The relative of was greater in and patients who received albumin, with of to the of the of the The in of crystalloid and colloid fluids have in the in of is better of renal and rapid of the vascular tree the large volumes necessary to adequate intravascular the effects of tissue edema, and the for pulmonary edema by serum in of colloid fluids greater in terms of intravascular volume for any given volume of infused fluid and more the vascular colloids glomerular filtration rate with colloid with with of and and more hydrostatic pulmonary edema the most are the more of intravascular volume colloid infusion in of major fluid such as surgery, and the of crystalloid fluids for most of the of colloid fluids in the composition or of formulations. these review of the used to solutions is is a of obtained from Because is rapidly by in the is reduced by hydro-yethyl for at three and solutions are by three average of and For example, an solution would have an average weight of the will a of of a of of the will have and as as will be important clinical associated with of solutions changes in and For was associated with in changes than or The of also may important The solution commonly used in the United States is in 0.9 saline and, when infused in large is associated with an solution in a solution including should be less to this and may less on perhaps because of the of a modest of in the is that the composition of solutions should be much given the and of formulations. research is necessary to define the clinical and of the of IV Fluids fluids have components that three and received attention The of concentration is based on the effects of changing serum osmolality on water as as a of other effects of hypertonic resuscitation fluids. to lactated Ringer’s solution as a for appears to pharmacologic effects that may be in the concentrations in 0.9 saline, has been associated with that is Because the normal is to small changes in serum greater osmotic pressure across the capillary than do relatively large changes in serum concentrations For instance, an increase of 5 in serum would increase osmolality by 10 or mm Hg of osmotic pressure. In contrast, the osmotic pressure by a normal serum concentration across the would be only approximately mm Hg (Table 3: of in the effects of changes in colloid osmotic pressure and serum on water or pressure have been in with normal in models of and in humans. In that reduced plasma to reduce plasma osmolality by 13 · = increased water content and in contrast, to reduce oncotic pressure from to mm Hg produced no in colloid osmotic pressure also not increase water after Because the the influence on water of changes in serum solutions lactated Ringer’s are more to increase water content than 0.9 saline or colloids in 0.9 saline. However, et al. demonstrated that colloid osmotic pressure influence water after perhaps because by secondary the solutions acutely reduce water and, to reduce In a in 3.0 saline decreased 0.9 saline had no is to that the effects on a of interstitial and rapidly intracellular volume after hypertonic In with hypertonic solutions reduced and decreased water in normal tissue In with and shock, resuscitation with hypertonic saline also blood and oxygen delivery In with bolus fluid resuscitation from was associated with fluid infusion, given as necessary to maintain cardiac further Data in subjected to fluid and mild hemorrhage to any in or blood when the resuscitation fluid was 3.0 saline In was with hypertonic saline demonstrated a increase in data also suggest the of after hypertonic resuscitation from and is important to the of models to the clinical that the acute effects of rapid administration of fluids of usually in those that at rates of administration usually et al. the effects on water of a of fluids administered to for after in edema were between that were fluid and those infused with 25 5 dextrose in saline, or an isotonic plasma have evaluated hypertonic solutions for resuscitation. et al. mL of LRS to 7.5 saline in dextran for resuscitation of trauma patients in blood pressure was mm Although there was no difference in in the of patients with of of those who received only of the patients who received LRS = In a et al. (Table 4) evaluated the effects of mL of with and without 6 and dextran for the resuscitation of trauma patients A small of patients with without anatomic to most from resuscitation with 7.5 saline hypertonic solutions have been used for resuscitation of patients with et al. used hypertonic lactated saline concentration or LRS to blood mm Hg or output during the first 5 days of Although the hypertonic more to reduce the was in that and the in was more et al. to hypertonic saline concentration or LRS for the first days after The hypertonic saline to maintain mm Hg and had and of were investigators also have reported with the use of hypertonic saline solutions for in patients effects of also have received reported physiologic changes associated with hypertonic fluids in the of sequestration of in the of and vascular permeability the of hypertonic resuscitation to have been acute in serum to without central which rapid correction of appears to be most after correction of and has not been in clinical of hypertonic resuscitation. to have no pharmacologic effects other than as for of appears to important effects on at in In from shock, infusion is associated with increased in both the and in LRS was associated with immune (12). However, LRS in to 0.9 saline in with hemorrhage further is necessary to define and rates of delivery of crystalloid fluids. of the more of fluid in the has been the that 0.9 saline, a of fluid produces as a several suggest that is a consequence of rapid replacement or expansion of extracellular volume with fluid containing no If the fluid (e.g., the is more than the fluid in concentrations that normal (4). Fluid management has rapidly in the three are to systemic in the of patients undergoing However, important to be the of complications of current fluid and the of fluid in a of clinical Shires Fluid in Shires changes in extracellular fluid associated with major surgical Svensén Hahn Volume kinetics of solution, dextran 70, and hypertonic saline in male volunteers. of colloids in fluid The of in fluid in resuscitation of Hahn Svensén Plasma dilution and the rate of infusion of Ringer’s Drobin Hahn Volume kinetics of Ringer’s solution in volunteers. et al. fluid following and correction with Ringer’s a et al. changes and of colloid osmotic pressure on the distribution of body Arieff postoperative pulmonary and et al. rates of fluid resuscitation or influence immune responses after et al. of values of as in high-risk surgical et al. of values of cardiac oxygen and oxygen as resuscitation in et al. of resuscitation of normal or A et al. of systemic oxygen delivery in the of et al. of and during oxygen a et al. with and volume effects of systemic oxygen on hepatic in et al. the of major controlled of of oxygen et al. of on after major abdominal a controlled et al. of oxygen delivery on and in high-risk surgical incidence and in measurements of blood concentrations and in patients with a review of and clinical Fluid resuscitation with colloid or crystalloid solutions in a review of albumin administration in review of controlled et al. In effects of starch solutions and lactated solution on using et al. a plasma for large volume use in major a clinical The acute effects of changes in plasma osmolality and oncotic pressure. in colloid oncotic pressure does not increase edema following in The of the of colloid oncotic pressure, with and without of on of the after fluid in the 1: distribution and of saline pressure in after volume of in to hypertonic of a hypertonic lactated Ringer’s solution on pressure and water content in a model of et al. blood following resuscitation from with hypertonic influence of a et al. effects of fluid resuscitation in the of an et al. saline does not oxygen delivery after and mild hemorrhage in et al. in after resuscitation with hypertonic solutions from by an et al. edema and following trauma in the no from large volumes of isotonic or hypertonic fluids, with or without 7.5 for resuscitation of trauma patients undergoing et al. A for resuscitation of patients with 7.5 the of dextran et al. saline resuscitation of patients with a clinical et al. A and controlled of fluid management in with lactated Ringer’s solution hypertonic saline. et al. of hypertonic saline in the of in of hypertonic saline solutions in of edema and et al. saline of is et al. effects of hypertonic resuscitation on the of following et al. saline resuscitation by permeability increase and plasma volume by was by hypertonic saline with or without et al. hypertonic infusion for the of as a factor in et al. with lactated solution in with Ringer’s is to normal saline in a model of hemorrhage and resuscitation. et al. The of in the and of associated V, et al. changes by 5 albumin 6 solution in patients undergoing acute normovolemic a following or albumin in healthy volunteers. is a consequence of infusion of 0.9 saline. associated with perioperative saline dilution or Research