INTRODUCTION — Perioperative maintenance of adequate intravascular volume status is important to achieve optimal outcomes after surgery, but there are controversies regarding both composition and volume of intraoperative fluid therapy. This topic will review derangements and monitoring of intravascular volume status in this setting, as well as strategies for choosing appropriate composition, amount, and timing of intraoperative fluid administration.

Severe intravascular volume depletion in surgical patients is discussed in other topics. (See "Intraoperative management of shock in adults", section on 'Hypovolemic shock management' and "Massive blood transfusion".)

Routine management of maintenance and replacement fluids in nonsurgical settings is discussed separately. (See "Maintenance and replacement fluid therapy in adults".)

CAUSES OF INTRAVASCULAR VOLUME DERANGEMENTS

Preoperative factors

●Preoperative fasting overnight for approximately 10 hours does not significantly reduce intravascular volume [1,2]. Nevertheless, preoperative dehydration should be avoided by limiting the period of fasting [3], and encouraging patients to consume clear oral liquids up to two hours before surgery. (See "Preoperative fasting in adults".)

●Mechanical bowel preparation may be associated with fluid loss from the gastrointestinal tract, which may reduce preoperative intravascular volume.

●Disorders such as bowel obstruction or pancreatitis may cause intravascular volume loss due to inflammation and interstitial edema. (See "Etiology, clinical manifestations, and diagnosis of volume depletion in adults", section on 'Etiology'.)

●Ongoing bleeding typically requires surgical hemostasis in order to allow adequate volume repletion. (See "Initial management of NON-hemorrhagic shock in adult trauma".)

Anesthesia-related factors

●Most anesthetic and adjuvant drugs cause dose-dependent vasodilation and myocardial depression that may lead to hypotension [4,5]. Avoiding unnecessarily deep anesthesia avoids hypotension, which in turn avoids otherwise unnecessary fluid administration. (See "Hemodynamic management during anesthesia in adults", section on 'Selection and dosing of anesthetic agents'.)

Hypotension that persists after reduction in anesthetic depth (if appropriate) and administration of fluid to replace known surgical losses can be treated with intravenous vasopressor/inotropic agents such as phenylephrine. (See 'Restrictive (zero-balance) strategy' below and "Hemodynamic management during anesthesia in adults", section on 'Vasopressor and positive inotropic agents'.)

●Sympathetic blockade during neuraxial anesthesia can result in relative hypovolemia due to increased venous capacitance and dilation of arteriolar resistance vessels, with resultant hypotension. Rather than preloading with IV fluid, which is a traditional practice, use of a vasopressor such as phenylephrine or norepinephrine is preferred.  

Surgery-related factors

●Hemorrhage.

●Coagulopathy due to hemodilution and/or hypothermia, which aggravates blood loss.

●Decreased venous return due to:

•Abdominal insufflation during laparoscopy.  

•Compression of the inferior vena cava or other major veins (eg, portal vein).

●Positive pressure ventilation with large tidal volumes, recruitment maneuvers, or positive end-expiratory pressure (ie, ventilation with a high level of continuous positive airway pressure to reverse atelectasis) during mechanical ventilation. (See "Mechanical ventilation during anesthesia in adults", section on 'Recruitment maneuvers'.)

●Prolonged operative time, particularly with an open abdominal cavity, may eventually lead to increased bowel edema and sequestration of fluid [6]. However, during short less invasive procedures there is minimal evaporative or insensible fluid loss from exposed body cavities or wounds and minimal sequestration of fluids in tissues (<1 mL/kg per hour) [7].

●Factors preventing an early transition from intravenous to oral fluid therapy within 24 hours of surgery [3].

CONSEQUENCES OF INTRAVASCULAR VOLUME DERANGEMENTS — Maintenance of intravascular euvolemia throughout the perioperative period is ideal. Both hypovolemia and hypervolemia are associated with increased postoperative morbidity [8-11].

Hypovolemia — Absolute or relative hypovolemia is common in the perioperative period due to preoperative dehydration, vasodilation caused by anesthetic and adjuvant drugs, and surgical bleeding. (See 'Causes of intravascular volume derangements' above.)

Preoperative hypovolemia may increase the risk of significant decreases in blood pressure during induction of anesthesia [12]. Persistent hypovolemia leads to low cardiac output and decreased tissue perfusion and, if severe, can lead to shock and multiorgan failure [13,14].

Hypervolemia — The most common cause of perioperative hypervolemia is retention of fluid administered during surgery. Clinically significant postoperative fluid retention (ie, weight gain >10 percent above preoperative baseline) has been associated with increased morbidity, length of stay in the intensive care unit, and mortality [8,15]. In critically ill surgical patients, this association may be spurious since fluid may be administered to treat hemodynamic instability (from increased vascular permeability) or bleeding, factors that are independently associated with poor outcome. However, tissue edema has independent deleterious effects on various organ systems, even in patients who remain hemodynamically stable [16]. These include:

●Respiratory effects – Increased extravascular fluid in the lung tissue can impair oxygen exchange and increase risk for postoperative respiratory failure and pneumonia [16]. Some patients develop frank pulmonary edema, particularly those with a history of heart failure [17]. (See "Overview of the management of postoperative pulmonary complications", section on 'Pulmonary edema'.)

●Gastrointestinal effects – Increased extracellular fluid in the bowel can lead sequentially to gastrointestinal edema, decreased gastrointestinal motility, and possibly ileus [18]. In patients undergoing bowel surgery, intestinal edema can increase tension at bowel anastomoses and may contribute to anastomotic dehiscence [16].

Occasionally, massive fluid resuscitation may be associated with acute ascites [19]. Ascites and bowel edema can contribute to development of abdominal compartment syndrome [20]. (See "Abdominal compartment syndrome in adults".)

●Coagulation – Excess intravascular fluid may dilute clotting factors, which can contribute to or cause coagulation abnormalities. (See "Massive blood transfusion", section on 'Alterations in hemostasis'.)

●Wound healing – Marked tissue edema can impair wound healing. (See "Basic principles of wound healing".)

MONITORING INTRAVASCULAR VOLUME STATUS

Goals and challenges — The purpose of monitoring intravascular volume status is to guide fluid administration in order to maintain tissue perfusion. Hypovolemia typically reduces tissue perfusion. However, tissue edema can also reduce tissue perfusion (eg, excessive volume administration or compensatory fluid retention in a patient with severe myocardial dysfunction).

Intraoperative monitoring with standard hemodynamic measurements (eg, noninvasive blood pressure [BP] and heart rate [HR]) and in selected cases, with one or more invasive monitors of dynamic hemodynamic parameters, can be used to predict fluid responsiveness (ie, increase in stroke volume following intravenous fluid administration (figure 1)) [21]. Laboratory values may be obtained, but these do not immediately reflect acute changes in volume status. Other clinical methods routinely used to assess volume status in the awake state are not available in the anesthetized patient (eg, thirst, postural dizziness, lethargy, confusion).

Regardless of the monitors employed, intraoperative determination of intravascular volume status is challenging because of continuously changing cardiovascular responses to anesthetic drugs, variable surgical volume losses that are difficult to quantify, and a preoperative volume status that may be suboptimal or unknown.

Static (traditional) parameters — Static parameters (eg, measurements of BP, HR, urine output [UO], and central venous pressure [CVP]) have been traditionally used to provide supplemental data regarding intravascular volume status (see "Monitoring during anesthesia", section on 'Monitoring modalities'). However, sole use of these parameters to guide fluid therapy may result in either hypovolemia or hypervolemia, and significant intraoperative reduction in tissue perfusion may not be recognized even with continuous monitoring of these static parameters due to the following limitations [13,14,22]:

●Blood pressure and heart rate – BP and HR responses to changes in intravascular volume status are not predictable in individual patients.

For example, a healthy young patient with a subclinical hypovolemic state often has normal BP and HR because the stress response to surgery activates the sympathetic nervous system and the renin-angiotensin system, with release of vasoconstrictor hormones that increase BP. Although the resulting peripheral vasoconstriction favors maintenance of adequate perfusion to the heart and the brain, perfusion to other organs such as the kidneys, gastrointestinal tract, and skin may be reduced. Furthermore, general or neuraxial anesthesia may blunt these compensatory vasoconstrictive responses to decreased perfusion.

Another example is a patient treated with beta-blockers who may not manifest tachycardia in response to hypovolemia. Thus, changes in volume status would not be detected by changes in HR.

●Central venous pressure – CVP and pulmonary artery occlusion pressure are sometimes used to provide supplemental data regarding intravascular volume status, but these parameters are inaccurate surrogates to determine cardiac preload, are poor predictors of fluid responsiveness [23], and do not detect or predict impending pulmonary edema indicative of hypervolemia [17,21,24-30].

●Urine output – Oliguria (UO <0.5 mL/kg per hour) is a commonly used indicator of hypovolemia. However, in patients undergoing anesthesia and surgery, oliguria alone is not a sufficient indication for fluid administration. Inhalation anesthetics, as well as surgical stress, may reduce UO. If the patient is actually euvolemic, administration of fluid to treat oliguria may lead to fluid overload [4,5].

Traditional thresholds for intraoperative oliguria do not predict acute kidney injury (AKI) [11,31-33]. In a 2016 systematic review and meta-analysis of 28 trials conducted in surgical and critically ill patients, less renal dysfunction was noted in patients receiving goal-directed fluid therapy without use of oliguria to guide fluid administration compared with patients who received conventional fluid management that targeted oliguria reversal (odds ratio [OR] 0.45, 95% CI 0.34-0.61) [33]. Different thresholds for UO were examined in a large retrospective study of 3560 patients undergoing major abdominal surgery [34]. Intraoperative UO <0.3 mL/kg per hour was associated with increased risk for AKI (OR 2.65, 95% CI 1.77-3.97) compared with UO 0.3 to 0.5 mL/kg per hour or higher. Another retrospective study in patients undergoing major surgery noted that duration of intraoperative oliguria (defined as UO <0.5 mL/kg) longer than 120 minutes was associated with increased risk for AKI compared with propensity-score matched patients with a shorter duration of oliguria (OR 2.2, 95% CI 1.8-2.7) [35]. Similarly, in a post hoc analysis of data from a trial of restrictive versus liberal fluid therapy (see 'Our approach to fluid management' below) [36], intraoperative UO <0.5 mL/kg per hour was associated with increased risk of AKI (risk ratio [RR] 1.38, 95% CI 1.14-1.44), although more severe oliguria with UO <0.3 mL/kg per hour did not increase risk further, and most patients with oliguria did not develop AKI [37]. Taken together, these data suggest that traditional targets that attempt to continuously maintain UO >0.5 mL/kg per hour are not warranted, but sustained oliguria may be associated with increased risk of renal injury, particularly if <0.3 mL/kg per hour [34,38,39].

●Mixed venous oxygen saturation – Measurements intended to track global oxygen (O₂) delivery have limited utility to guide fluid therapy [40]. These include measurements of mixed venous O₂ saturation (SvO₂; or central venous O₂ saturation [ScvO₂]), which may be obtained intermittently from blood gas analysis or continuously using a fiberoptic catheter. Although SvO₂ and ScvO₂ are proportional to cardiac output, tissue perfusion, and tissue O₂ delivery, these measurements are also inversely proportional to tissue O₂ consumption and do not reflect changes in tissue perfusion during the perioperative period when O₂ consumption varies [40,41]. (See "Oxygen delivery and consumption".)

Dynamic hemodynamic parameters — Dynamic hemodynamic indices are used to guide goal-directed fluid therapy in patients undergoing major invasive surgery, particularly high-risk patients with large expected blood losses and fluid shifts [3]. Dynamic parameters provide superior assessment of response to a fluid challenge (ie, volume responsiveness) compared with traditional static parameters [25,40,42-46]. (See 'Goal-directed fluid therapy' below.)

Indices based on respiratory variation (arterial pressure waveform) — Variations in the arterial pressure waveform that occur during respiration (eg, pulse pressure variation [PPV], stroke volume variation [SVV], systolic blood pressure variation [SPV], or change in inferior vena cava diameter) can be observed or measured to assess responses to fluid challenges (figure 1) [46-50]. During controlled mechanical ventilation, inspiration increases intrathoracic pressure, which reduces venous return, right ventricular (RV) filling volume, and right and left ventricular stroke volume (SV). The opposite effects occur during expiration. These changes in venous return lead to variations in SV, as well as pulse pressure and systolic BP (figure 2), if arterial vasomotor tone and cardiac function remain constant [49]. Normal respiratory variations in these dynamic parameters are <10 percent [51]. Greater variations suggest fluid responsiveness and the likely need for fluid administration (figure 1) [23].

Although hemodynamic indices of respiratory variation can be computed (manually or automatically), visual estimation may be adequate to guide fluid therapy. In one study, determination of the need for a fluid bolus based on visual estimation of SPV in the arterial waveform was compared with computed values, with only 1 percent of treatment decisions based on visual estimation being incorrect (figure 1) [47].

Each of the dynamic indices based on respiratory variation has advantages and disadvantages, with limitations in sensitivity and specificity (table 1) [52-54]. In a 2018 systematic review (5017 patients; 68 studies), the pooled area under the curve (AUC) for PPV was 0.86 (95% CI 0.80-0.92), with a sensitivity of 80 percent (95% CI 74-85 percent) and specificity of 83 percent (95% CI 73-91 percent) [54]. In that study, the pooled AUC for SVV was 0.87 (95% CI 0.81-0.93), with a sensitivity of 82 percent (95% CI 75-89 percent) and specificity of 77 percent (95% CI 71-82 percent). Furthermore, these indices are not useful during open chest procedures [55,56], during mechanical ventilation with low tidal volumes <8 mL/kg or high positive end-expiratory pressure >15 cm H₂O, or in patients who have elevated intraabdominal pressure, cardiac arrhythmias, RV failure, or a requirement for vasoactive infusions [50,57,58]. Thus, although superior to static parameters for assessment of fluid responsiveness (see 'Static (traditional) parameters' above), indices based on respiratory variation are interpreted with caution and should consider the clinical setting.

Stroke volume estimates — SV derived by using esophageal Doppler technology or arterial wave form analysis have been used to guide fluid therapy in an attempt to maintain intravascular volume [46,59-61]. Such devices may be useful when indices based on respiratory variation in the intra-arterial waveform cannot be used. (See 'Indices based on respiratory variation (arterial pressure waveform)' above.)

Left ventricular size estimates (transesophageal echocardiography) — With transesophageal echocardiography (TEE), intravascular volume status can be quickly assessed in the transgastric midpapillary short-axis view by qualitative visual assessment of LV cavity size. Underfilling of the left ventricle caused by acute hypovolemia is easily recognized in a patient with hyperdynamic systolic function and decreased end-diastolic and end-systolic LV cavity dimensions (movie 1). Also, quantitative measurements of the internal diameter or cross-sectional area of the LV at end-diastole can be made (image 1 and image 2 and table 2) [62,63]. Changes from baseline (normovolemia) are monitored using these qualitative and/or quantitative assessments. (See "Intraoperative transesophageal echocardiography for noncardiac surgery", section on 'Assessment of left ventricular volume'.).

Noninvasive technologies — Noninvasive commercially available technologies that measure cardiac output or assess fluid responsiveness have been studied (eg, pulse wave transit time, pulse contour analysis, carbon dioxide rebreathing, and thoracic electrical bioimpedance or bioreactance devices) [64]. A 2017 meta-analysis concluded that the percentage error of these devices for measurements of cardiac output is high compared with standard thermodilution techniques [65].

Measurement of laboratory values — Absent an ischemic insult to a specific organ, increased serum lactate levels or lactic acidosis on sequential arterial blood gases can be an important indicator of reduced global tissue perfusion. However, these laboratory values do not provide information regarding contemporaneous clinical intravascular volume status since they are measured intermittently and do not immediately reflect acute changes. (See "Etiology, clinical manifestations, and diagnosis of volume depletion in adults", section on 'Clinical manifestations'.)

CHOOSING FLUID: CRYSTALLOID, COLLOID, OR BLOOD

Crystalloid solutions — Crystalloids are solutions of electrolytes and sterile water that may be isotonic, hypotonic, or hypertonic with respect to plasma. Balanced electrolyte solutions (also termed buffered crystalloid solutions) that have an electrolyte composition similar to plasma with the addition of a buffer (eg, lactate) are most widely used [66]. Examples include Ringer's lactate (also termed Hartmann's solution) or Plasmalyte.

We typically select a balanced electrolyte crystalloid solution for routine perioperative fluid administration in order to maintain normovolemia and/or replace lost blood. Crystalloids are administered on a 1.5:1.0 volume basis until a transfusion threshold is met. We avoid dextrose-containing solutions due to the putative adverse effects of hyperglycemia [67].

We avoid administration of a large volume of normal saline (NS; 0.9 percent), as this has been associated with hyperchloremic acidosis [68-81]. Risk for other adverse outcomes, particularly acute kidney injury (AKI), has been associated with NS in several observational and randomized studies [69,70,82-85]. However, results are not consistent in patients who do not receive large volumes of NS or are not critically ill [86,87]. A meta-analysis of randomized trials that included 1096 surgical patients noted that perioperative administration of buffered (ie, balanced) electrolyte solutions was associated with a lower incidence of minor metabolic derangements, particularly mild metabolic acidosis compared with NS, but did not demonstrate an effect on mortality or renal replacement therapy (RRT); however, most of these patients were not critically ill [68]. Similarly, another meta-analysis of randomized trials in more than 3700 unselected critically ill or perioperative adult patients did not find a significant reduction in mortality (odds ratio [OR] 0.90, 95% CI 0.69-1.17) or incidence of RRT (OR 1.12, 95% CI 0.80-1.58) with administration of balanced electrolyte solutions rather than NS; however, most of these patients received a low fluid volume [88].

Colloid solutions — Colloids are human plasma derivatives (eg, human albumin, fresh frozen plasma [FFP]) or semisynthetic preparations (eg, hydroxyethyl starch [HES], gelatins) [89]. Colloids may be dissolved in isotonic saline or in a solution with a balanced electrolyte concentration similar to plasma [89].

There is little evidence that colloid solutions are superior to balanced electrolyte crystalloid solutions [90-94]. In one study, closed loop administration of balanced HES for elective major open abdominal surgery produced lower morbidity on postoperative day two and better disability-free survival at follow up one year after surgery, compared with balanced crystalloid [93,94]. Some clinicians prefer to use colloids in selected patients or situations in attempts to expand microvascular volume with minimal capillary leakage, thereby minimizing edema formation and the total quantity of administered fluid [95]. For example, during blood loss, colloids may be administered on a 1.0:1.0 volume basis until a transfusion threshold is met [95,96]. (See "Intraoperative transfusion of blood products in adults", section on 'Red blood cells'.)

Albumin — Human serum albumin is available in both 5 and 25 percent solutions. In some parts of the world, human serum albumin is available as 4 and 20 percent solutions. Human albumin 5 percent has a volume effect (ie, the percent of fluid administered that remains intravascular) of 70 percent, while albumin 25 percent solution is isosmotic with plasma (table 3). Human albumin is pasteurized and does not transmit any known infectious diseases. (See "Plasma derivatives and recombinant DNA-produced coagulation factors", section on 'Albumin'.)

Albumin is more expensive than other solutions, and may not be safer or more efficacious than synthetic colloids (eg, HES) or balanced crystalloid solutions [86,90,97-99]. (See 'Hydroxyethyl starches' below and 'Crystalloid solutions' above.) However, administration of 20 percent albumin during the intraoperative period during noncardiac surgery caused a long-lasting plasma volume expansion lasting into the postoperative period [100,101]. In one study, the intravascular half-life was 9.1 (5.7 to 11.2) hours [100]. A pilot study of administration of 20 percent albumin in the postoperative period after cardiac surgery resulted in less positive fluid balance and volume of fluid boluses, as well as decreased dosing of a vasopressor (norepinephrine) [102].

Hydroxyethyl starches — We do not use HES solutions for fluid therapy in surgical patients. HES solutions are synthetic colloids, identified by three numbers corresponding to concentration, molecular weight, and molar substitution (ie, the average hydroxyethyl groups per one glucose unit) (table 3). As an example, Hespan is HES 6 percent (600/0.75) with a volume effect of 100 percent and a high molar substitution of 0.75. Risk of HES-induced renal toxicity may be influenced by the molar substitution level in the specific product, with lower risk for low substituted HES products [103]. Due to concerns regarding renal toxicity and effects on hemostasis, administration of HES solutions is restricted in Europe and North America [86].

In a randomized trial of more than 800 patients undergoing major abdominal surgery, administration of a low molar-substituted HES 130/0.4 solution versus 0.9 percent saline to provide goal-directed fluid therapy (see 'Goal-directed fluid therapy' below) did not reduce the composite outcome of death or major postoperative complications, and may have had harmful effects (eg, increased of postoperative AKI [relative risk 1.34, 95% CI 1.00-1.80]) [104]. Similarly, a 2018 systematic review in critically ill patients with medical and surgical diagnoses noted a higher incidence of RRT in those receiving HES solutions compared with those receiving crystalloids (risk ratio [RR] 1.30, 95% CI 1.14-1.48; 8527 participants, nine studies) [90]. Also, a 2010 systematic review in mixed surgical and nonsurgical patient populations found an overall increased risk of author-defined kidney failure in surgical and medical patients receiving HES solutions (including some who received highly substituted HES products) compared with those receiving various other types of fluid therapy [105]. Furthermore, administration of HES solutions in cardiac surgical patients increases risk of postoperative bleeding and transfusion compared with albumin [106]. However, data are not consistent. Some randomized and observational studies have noted no differences in risk for AKI or other serious postoperative complications in patients receiving goal-directed fluid therapy with HES compared with other types of fluids (eg, balanced crystalloid solution, 5 percent albumin) [92,99,107-109].  

Since HES products impair platelet reactivity and decrease circulating plasma concentrations of coagulation factor VIII and von Willebrand factor, administration may result in weakening of clot formation and more transfusions of blood products including FFP, cryoprecipitate, and platelets compared with other solutions [90,110-112]. A 2018 systematic review in critically ill patients noted a higher incidence of transfusion in those receiving starch solutions compared with those receiving crystalloids (RR 1.19, 95% CI 1.02-1.39; 1917 participants, eight studies) [90]. HES products with low molar substitution (eg, pentastarch and tetrastarch) may have less effect on hemostasis [113]. One meta-analysis compared HES solutions with low (0.4) versus a somewhat higher (0.5) molar substitution, noting significantly less blood loss (404 mL) and red blood cell (RBC) transfusion (137 mL) in patients receiving the low substituted product (seven trials; 449 patients) [114].

Gelatins — Gelatins are not used in the United States because of their short duration of action (two to three hours) due to rapid excretion in the urine, possible effects on coagulation, and a relatively high incidence of anaphylaxis [86,89,115]. (See "Perioperative anaphylaxis: Clinical manifestations, etiology, and management", section on 'Colloids and plasma expanders'.)

Gelatins are used in some countries because they are inexpensive and have a volume effect of 70 to 80 percent (table 3) [116].

Blood

Transfusion of red blood cells — RBCs are used to replace intraoperative blood loss when a transfusion threshold is met, as discussed separately. (See "Intraoperative transfusion of blood products in adults", section on 'Red blood cells'.)

Transfusion of plasma derivatives — Decisions regarding transfusion of plasma derivatives of human blood (eg, FFP, cryoprecipitate) are based on estimates of the amount of current and expected ongoing blood loss and evidence of intractable microvascular bleeding indicating abnormal hemostasis, ideally with confirmation by diagnostic test results. These decisions are discussed separately. (See "Intraoperative transfusion of blood products in adults", section on 'Plasma' and "Intraoperative transfusion of blood products in adults", section on 'Cryoprecipitate'.)

OUR APPROACH TO FLUID MANAGEMENT — Our intraoperative fluid management strategy is based on the invasiveness of the surgical procedure, as well as whether invasive monitoring of dynamic hemodynamic parameters is employed. Expected blood loss and likelihood of nonhemorrhagic fluid shifts (eg, from open body cavities and wounds) are factors in decision-making regarding selection of noninvasive and/or invasive monitors for a planned surgical procedure (see 'Dynamic hemodynamic parameters' above). Other patient-specific and procedure-specific factors affecting fluid management include the patient's comorbidities (eg, anemia) and planned postoperative disposition (eg, home, ambulatory hospital ward, critical care unit).

Minimally/moderately invasive surgery — For most adult patients undergoing relatively brief minimally or moderately invasive surgery with planned early postoperative ambulation, we administer 1 to 2 L of a balanced electrolyte solution if the procedure does not incur significant fluid shifts or blood loss. This 1 to 2 L of fluid is typically administered during surgery, over a period of 30 minutes to two hours. Such empiric but limited fluid administration for less invasive surgery in ambulatory patients addresses the mild dehydration caused by preoperative fasting and is associated with less postoperative nausea and vomiting, as well as less postoperative pain [117,118]. A smaller fluid volume is appropriate in patients with a history of heart failure or chronic obstructive pulmonary disease.

Major invasive surgery — For adult patients undergoing major invasive surgical procedures, we employ a restrictive, zero-balance approach that minimizes fluid administration, or a goal-directed approach with fluid administration to achieve a pre-specified goal [8]. Evidence suggests that either a restrictive or goal-directed approach is associated with a lower incidence of perioperative morbidities (and possibly mortality) compared with traditional liberal or fixed-volume approaches that may result in fluid overload [10,13,14,22,59,60,119-122]. (See 'Avoid traditional liberal or fixed-volume approaches' below.)

We use a restrictive (zero-balance) approach to fluid therapy for invasive major surgery with expected blood loss <500 mL if there is no invasive monitoring of dynamic hemodynamic parameters via an intra-arterial catheter, esophageal Doppler probe, or transesophageal echocardiography (TEE) probe (see 'Restrictive (zero-balance) strategy' below). A potential disadvantage of this approach is that hypovolemia may not be clinically appreciated. Also, if hypotension occurs, it may be difficult to determine the etiology (eg, surgical volume losses or other causes such as cardiovascular responses to anesthetic agents).

For these reasons, we typically select a goal-directed approach for high-risk patients undergoing a major surgical procedure, particularly if significant blood losses (eg, >500 mL) and/or fluid shifts are expected. A goal-directed approach necessitates use of one or more invasive dynamic hemodynamic parameters, as noted below (see 'Goal-directed fluid therapy' below). Final decisions regarding selection of noninvasive versus invasive monitoring are also based on procedure-specific and patient-specific factors, as discussed in separate UpToDate topics.

Restrictive (zero-balance) strategy — With a restrictive zero-balance approach, only the fluid that is lost during surgery is replaced, including the following strategies [14]:

●During the intraoperative period, patients receive a balanced electrolyte crystalloid solution administered at a rate of 1 to 3 mL/kg per hour to replace sensible and insensible losses [13].

●For blood loss, additional fluid may be administered. Studies suggest that the optimal crystalloid-to-blood volume ratio is approximately 1.5:1.0, and that the optimal colloid-to-blood ratio is 1:1, until a threshold for red blood cell (RBC) transfusion is reached [123-126].

●Patients do not receive "preloading" of crystalloids prior to a neuraxial block or induction of general anesthesia.

●We avoid replacement of nonanatomic "third space" losses, since evidence suggests that this practice has no benefit, and may cause morbidity [6,127,128].

●We avoid extremely deep anesthesia (eg, bispectral index values <40) that may result in hypotension. If necessary, vasopressor agents such as phenylephrine or ephedrine may be employed to treat hypotension caused by administration of anesthetic agents and/or neuraxial block [129,130]. (See "Accidental awareness after general anesthesia", section on 'Brain monitoring' and "Hemodynamic management during anesthesia in adults", section on 'Hypotension: Prevention and treatment'.)

●Administration of a total volume of balanced electrolyte solutions that modestly exceeds zero fluid balance is appropriate in patients with evidence of hypovolemia [131].

In most studies, restrictive fluid therapy has been associated with better outcomes than traditional liberal or fixed-volume approaches for major elective surgical procedures [13,14,22,86,132,133]. However, variability in study design has resulted in inconsistent results. A 2012 meta-analysis found that standard or liberal approaches to fluid therapy in patients undergoing major abdominal procedures resulted in a higher risk for pneumonia (risk ratio [RR] 2.2, 95% CI 1.0-4.5) and pulmonary edema (RR 3.8, 95% CI 1.1-13), as well as longer hospital stay (mean difference two days, 95% CI 0.5-3.4), compared with a restrictive approach to fluid therapy (1160 patients; 12 randomized trials) [134]. In a subsequently published randomized trial of 3000 patients undergoing major abdominal surgery, a restrictive (zero-balance) fluid regimen was associated with a higher rate of acute kidney injury (AKI) compared with a liberal fluid regimen (8.6 versus 5.0 percent; RR 1.71, 95% CI 1.29-2.27) [36]. A limitation of this study was its pragmatic design in that perioperative care was not standardized and there was a wide variation in the anesthetic and analgesic techniques, including use of epidural analgesia, variable intraoperative hemodynamic management, and variable postoperative care. The total fluid volume administered during and up to 24 hours after surgery was 3.7 versus 6.1 L in the restrictive and liberal groups, respectively. These volumes were lower than traditional liberal fluid strategy totals, and balanced electrolyte solutions were used in both groups [131]. Similarly, a restrictive approach was associated with AKI in an observational trial of 769 patients undergoing cystectomy (odds ratio [OR] 0.79, 95% CI 0.68-0.91) [135]. Conversely, a randomized trial in 166 patients undergoing radical cystectomy employed a restrictive approach (1 mL/kg/h) combined with low-dose norepinephrine infusion during the initial portion of the surgery subsequently followed by hydration (3 mL/kg/h) during the last part of the surgical procedure [136]. This so-called "restrictive deferred hydration" approach was associated with fewer complications (RR 0.70, 95% CI 0.55-0.88) and a lower median duration of hospital stay (15 days, range 11 to 27) than a liberal approach with 6 mL/kg/h (17 days, range 11 to 95).

Taken together, these studies suggest that administration of a total volume of balanced electrolyte solutions that modestly exceeds zero fluid balance is not harmful [131]. Also, evidence suggesting hypovolemia should be recognized and appropriately treated with fluid, as noted below (see 'Goal-directed fluid therapy' below). However, excessive perioperative administration of intravenous fluid, which was common in traditional liberal or fixed-volume approaches to fluid therapy, should be avoided [10,86]. (See 'Avoid traditional liberal or fixed-volume approaches' below.)

Goal-directed fluid therapy — We typically select a goal-directed approach to fluid therapy (GDT) in patients undergoing major invasive surgery with expected blood loss >500 mL and/or other significant perioperative fluid shifts. With this approach, we ensure that intravascular volume status is optimal before adding vasopressor therapy to achieve optimal blood pressure [3].

While GDT appears superior to traditional liberal or fixed-volume approaches, there are limited data comparing GDT to the restrictive approach described above [134,137]. One disadvantage of GDT is that it requires invasive monitoring of dynamic hemodynamic parameters [3,14,59,60,120]. We consider the following factors in selection of monitoring modalities:

●In most patients undergoing major surgery, we use the intra-arterial waveform tracing for automated measurements of pulse pressure variations (PPV) or stroke volume variation (SVV), or visually estimated or manually calculated PPV or systolic pressure variations (SPV), in order to determine responses to fluid boluses (typically 250 mL increments) (figure 1 and figure 2).

●For high-risk patients undergoing a surgical procedure with expected blood loss >1000 mL, significant nonhemorrhagic fluid losses, and/or likely prolonged duration, we typically use a commercially available device that provides automated calculation of PPV, SVV, or SPV by analyzing the intra-arterial waveform tracing to assess responses to fluid challenges. An alternative is use of an esophageal Doppler device to estimate stroke volume (SV) [64]. TEE is another option, allowing visual qualitative evaluation or quantitative measurements of the left ventricular (LV) cavity size to monitor fluid responsiveness (movie 1 and image 1 and image 2 and table 2). (See 'Dynamic hemodynamic parameters' above.)

With a GDT approach, fluid is administered to achieve a pre-specified goal:

●If respiratory variations in the arterial pressure waveform (PPV or SPV) are >10 to 15 percent, then the patient is assumed to be fluid responsive and we administer fluid boluses of a balanced electrolyte crystalloid solution (typically in 250 mL increments) [23,51,138]. Once change in the monitored dynamic parameter is <10 percent, fluid administration is stopped to avoid hypervolemia.

●If SV estimates are used for dynamic hemodynamic monitoring, the typical goal of therapy is to achieve and maintain optimal intravascular volume with maximum SV. The new SV value after fluid administration has resulted in <10 percent change is recorded as the new baseline goal value (representing the maximum SV to be maintained). (See 'Indices based on respiratory variation (arterial pressure waveform)' above.).

●If TEE is employed, hypovolemic and hypervolemic states can be quickly assessed by visual qualitative evaluation or quantitative measurements of LV cavity size. Fluid administration is stopped once normovolemia has been achieved (movie 1 and image 1 and image 2 and table 2). (See 'Left ventricular size estimates (transesophageal echocardiography)' above.)

Although most studies evaluating GDT have used boluses of colloid fluid [139,140], we typically select a balanced crystalloid solution for fluid boluses. Randomized trials in patients undergoing elective abdominal surgery have found little difference in postoperative complications or any clinical benefit with use of a 6 percent hydroxyethyl starch (HES) colloid solution for fluid boluses compared with a balanced crystalloid solution to provide GDT [92,94,123]. (See 'Choosing fluid: Crystalloid, colloid, or blood' above.)

GDT appears to be superior to traditional fixed-volume or liberal fluid approaches (see 'Avoid traditional liberal or fixed-volume approaches' below). In one meta-analysis, fluid administration with a GDT approach was associated with improved clinical outcomes, including lower risk of mortality (odds ratio [OR] 0.66, 95% CI 0.50-0.87), pneumonia (OR 0.69, 95% CI 0.51-0.92), acute kidney injury (OR 0.73, 95% CI 0.58-0.92), wound infection (OR 0.48, 95% CI 0.37-0.63), and shorter length of hospital stay (-0.9 days; 95% CI -1.3 to -0.5 days), compared with standard fluid management according to protocol (eg, maintenance of mean arterial pressure >65 mmHg) or the discretion of treating clinicians (11,659 patients; 95 randomized trials) [137]. Other randomized trials and meta-analyses have noted similar results with GDT compared with standard care (eg, lower risk of respiratory, renal, wound, and gastrointestinal complications, with shorter time to hospital discharge) [60,134,141-146]. While most studies reported lower complication rates with GDT, numerous limitations should be noted. These include clinical heterogeneity among trials with differing definitions for GDT, lack of well-defined endpoints, different types of fluid therapy, different devices used to monitor dynamic hemodynamic parameters, variations in management of the control group, different types of surgery, and small sample sizes [14,64,147-149]. In addition, most studies have included only limited information regarding anesthetic techniques and perioperative surgical care [147].

Whether GDT is superior to a restrictive fluid strategy is less certain. In a 2019 meta-analysis of randomized trials in patients undergoing major noncardiac surgery, very low-certainty evidence suggested that restrictive fluid therapy does not affect the risk of complications (RR 1.61, 95% CI 0.78-3.34; five studies; 484 participants), but may increase the risk of all-cause mortality (risk difference [RD] 0.03, 95% CI 0.00-0.06) [140]. Furthermore, GDT does not appear to offer additional benefits over a restrictive fluid approach for patients managed with protocols to achieve enhanced recovery after surgery (ERAS), [150-153]. The likely reason is that ERAS protocols implement multiple processes that each reduce the risk of perioperative fluid imbalances (eg, avoidance of preoperative dehydration, use of an intraoperative restrictive fluid approach, emphasis on early postoperative alimentation and ambulation). One meta-analysis assessing GDT in this setting found no benefit (or harm) with its use [143]. (See "Anesthetic management for enhanced recovery after major surgery (ERAS) in adults", section on 'Fluid management'.)

Clinical questions regarding GDT that remain unanswered include which types of patients are most likely to benefit, timing of the use of GDT (preoperative, intraoperative, and/or postoperative), which outcome measures or endpoints are optimal, which combinations of therapies constitute the best approach (eg, fluids with or without vasopressors or inotropic agents), and how long the regimen should be maintained during the postoperative period.

Avoid traditional liberal or fixed-volume approaches — We do not use traditional liberal or fixed-volume approaches to fluid therapy in patients undergoing any type of surgery, because these strategies result in administration of a large volume of crystalloid solution, with increased incidence or severity of tissue edema and associated adverse outcomes [13,14,16,22,59,60,119-122,154-156].

Traditional fixed-volume approaches were based upon predetermined calculations that included administration of fluid to account for presumed preoperative deficits, as well as intraoperative blood and urinary losses. In addition, fluid was typically administered to compensate for calculated non-anatomic "third space" fluid losses during the surgical procedure. This practice is inappropriate, as it has been well-established that such third space losses do not exist [6,127,128]. Furthermore, fixed volume regimens specified replacement of initial blood loss with crystalloid volume that was three times the amount of lost blood. For example, if blood loss was estimated to be 500 to 1000 mL, then 1500 to 3000 mL of crystalloid was typically administered. This calculation is not supported by available data. Rather, optimal volume ratios to compensate for lost blood are estimated to be 1.5:1.0 for crystalloid and 1.0:1.0 for colloid, as discussed above [123-125] (see 'Crystalloid solutions' above and 'Colloid solutions' above). Similarly, large volumes of fluid are unnecessary as preloading before a neuraxial block.

The overall result of use of traditional fixed volume calculations was administration of a large total volume of crystalloid solution and a high likelihood of increased perioperative tissue edema. For these reasons, such approaches have been abandoned.

SUMMARY AND RECOMMENDATIONS

●Maintenance of tissue perfusion by maintaining euvolemia is the goal for intraoperative monitoring of intravascular fluid volume and administration of intraoperative fluid therapy. Both hypovolemia and hypervolemia are associated with postoperative morbidity. (See 'Consequences of intravascular volume derangements' above.)

●Absolute or relative hypovolemia is common in the perioperative period due to preoperative dehydration, vasodilation caused by anesthetic and adjuvant drugs, and surgical bleeding. The most common cause of perioperative tissue edema is retention of fluid administered during surgery. (See 'Causes of intravascular volume derangements' above.)

●While physiological parameters such as blood pressure (BP), heart rate (HR), central venous pressure (CVP), and urine output (UO) are monitored during surgery, significant intraoperative reduction in tissue perfusion may not be recognized even with continuous monitoring of these static parameters. (See 'Static (traditional) parameters' above.)

●Monitoring of dynamic hemodynamic parameters based on respiratory variation uses the intra-arterial waveform tracing for automated analysis or visually estimated or manually calculated pulse pressure variations (PPV), stroke volume variation (SVV), or systolic pressure variations (SPV) (figure 1 and figure 2), or estimates of stroke volume (SV) in order to determine responses to fluid boluses (typically administered in 250 mL increments). An alternative is use of transesophageal echocardiography (TEE) to recognize hypovolemic and hypervolemic states by visual qualitative assessment or quantitative measurements of left ventricular (LV) cavity size (movie 1 and image 1 and image 2 and table 2). (See 'Dynamic hemodynamic parameters' above.)

●We suggest the use of balanced electrolyte solutions (eg, Ringer's lactate, Plasmalyte), rather than normal saline or colloid as the standard intravenous fluid to maintain or replace intravascular volume in surgical patients (Grade 2C). In particular, we avoid hydroxyethyl starch (HES) colloid solutions due to possible harmful effects (eg, increased risk of postoperative acute kidney injury). We typically select a balanced crystalloid solution to be administered at 1 to 3 mL/kg per hour to replace sensible and insensible losses, and for fluid boluses (typically 250 mL) to optimize intravascular volume. (See 'Colloid solutions' above and 'Crystalloid solutions' above.)

●We replace initial blood loss with either crystalloid administered in a volume that is 1.5 times the amount of lost blood, or colloid administered on a 1.0:1.0 volume basis, until a transfusion threshold is met. (See 'Choosing fluid: Crystalloid, colloid, or blood' above.)

●For most adult patients undergoing minimally or moderately invasive surgical procedures with planned early postoperative ambulation, we administer 1 to 2 L of a balanced electrolyte solution to provide adequate intravascular hydration. (See 'Minimally/moderately invasive surgery' above.)

●For major invasive surgical procedures, we suggest a restrictive (zero-balance) or goal-directed therapy (GDT) approach to fluid therapy, rather than traditional liberal or fixed-volume approaches (Grade 2B). We choose between these strategies based on anticipated significant blood loss (eg, >500 mL) and/or other significant perioperative fluid shifts. Both the restrictive and GDT strategies appear to be associated with reduced morbidity compared with more liberal fluid administration. (See 'Major invasive surgery' above.)

●A restrictive (zero-balance) approach to fluid therapy replaces only fluid lost during the procedure. (See 'Restrictive (zero-balance) strategy' above.)

●A GDT approach to fluid therapy requires invasive monitoring of dynamic hemodynamic parameters (ie, intra-arterial catheter). (See 'Goal-directed fluid therapy' above and 'Dynamic hemodynamic parameters' above.)

•For most patients, we use visually estimated SPV in the intra-arterial waveform tracing to determine responsiveness to fluid boluses (typically 250 mL) (figure 1).

•For high-risk patients undergoing a surgical procedure with expected blood loss >1000 mL, significant nonhemorrhagic fluid shifts, and/or likely prolonged duration, we typically select a commercially available device to determine responses to fluid challenges (eg, devices that provide automated calculations of respiratory variations in systolic BP, pulse pressure, or SV in the intra-arterial waveform tracing, esophageal Doppler device to estimate SV, TEE to assess LV cavity size).