Intravenous Fluid Resuscitation
Almost all circulatory shock states require large-volume IV fluid replacement, as does severe intravascular volume depletion (eg, due to diarrhea or heatstroke). Intravascular volume deficiency is acutely compensated for by vasoconstriction, followed over hours by migration of fluid from the extravascular compartment to the intravascular compartment, maintaining circulating volume at the expense of total body water. However, this compensation is overwhelmed after major losses.
Choice of resuscitation fluid depends on the cause of the deficit.
Loss of red blood cells diminishes oxygen-carrying capacity. However, the body increases cardiac output to maintain oxygen delivery (DO2) and increases oxygen extraction. These factors provide a safety margin of about 9 times the resting oxygen requirement. Thus, non–oxygen-carrying fluids (eg, crystalloid or colloid solutions) may be used to restore intravascular volume in mild to moderate blood loss. However, in severe hemorrhagic shock, blood products are required. Early administration of plasma and platelets probably helps minimize the dilutional and consumptive coagulopathy that accompanies major hemorrhage. A ratio of 1 unit of plasma for each 1 unit of blood and each 1 unit of platelets is currently recommended (1). When the patient is stable, if the hemoglobin is < 7 g/dL (70 g/L), in the absence of cardiac or cerebral vascular disease, oxygen-carrying capacity should be restored by infusion of additional blood (or in the future by blood substitutes). Patients with active coronary or cerebral vascular disease or ongoing hemorrhage require blood for hemoglobin < 10 g/dL (100 g/L).
Crystalloid solutions for intravascular volume replenishment are typically isotonic (eg, 0.9% saline or Ringer's lactate). Water freely travels outside the vasculature, so as little as 10% of isotonic fluid remains in the intravascular space. With hypotonic fluid (eg, 0.45% saline), even less remains in the vasculature, and, thus, this fluid is not used for resuscitation. Both 0.9% saline and Ringer's lactate are equally effective; Ringer's lactate may be preferred in hemorrhagic shock because it somewhat minimizes acidosis and will not cause hyperchloremia. For patients with acute brain injury, 0.9% saline is preferred. Hypertonic saline is not recommended for resuscitation because the evidence suggests there is no difference in outcome when compared to isotonic fluids.
Colloid solutions (eg, hydroxyethyl starch, albumin, dextrans) are also effective for volume replacement during major hemorrhage. However, colloid solutions offer no major advantage over crystalloid solutions, hydroxyethyl starch increases risk of renal injury, and albumin has been associated with poorer outcomes in patients with traumatic brain injury. Both dextrans and hydroxyethyl starch may adversely affect coagulation when > 1.5 L is given (2).
Blood typically is given as packed red blood cells, which should be cross-matched, but in an urgent situation, 1 to 2 units of type O Rh-negative blood are an acceptable alternative. When > 1 to 2 units are transfused (eg, in major trauma), blood is warmed to 37°C. Patients receiving > 6 units may require replacement of clotting factors with infusion of fresh frozen plasma or cryoprecipitate and platelet transfusion (see also Blood Products).
Blood substitutes are oxygen-carrying fluids that can be hemoglobin-based or perfluorocarbons. Hemoglobin-based fluids may contain free hemoglobin that is liposome-encapsulated or modified (eg, by surface modification or cross-linking with other molecules) to limit renal excretion and toxicity. Because the antigen-bearing red blood cell membrane is not present, these substances do not require cross-matching. They can also be stored > 1 year, providing a more stable source than banked blood. Perfluorocarbons are IV carbon-fluorine emulsions that carry large amounts of oxygen. However, no blood substitutes have yet proved to increase survival and some have significant adverse effects (eg, hypotension). Currently, no blood substitutes are commercially available for use.
Isotonic crystalloid solutions are typically given for intravascular repletion during shock and hypovolemia. Colloid solutions are generally not used. Patients with dehydration and adequate circulatory volume typically have a free water deficit, and hypotonic solutions (eg, 5% dextrose in water, 0.45% saline) are used.
1. Holcomb JB, Tilley BC, Baraniuk S, et al: Transfusion of plasma, platelets, and red blood cells in a 1:1:1 vs a 1:1:2 ratio and mortality in patients with severe trauma: The PROPPR randomized clinical trial. JAMA 313(5):471-482, 2015. doi:10.1001/jama.2015.12.
2. Myburgh JA, Finfer S, Bellomo R, et al: Hydroxyethyl starch or saline for fluid resuscitation in intensive care. N Engl J Med 367(20): 1901-1911, 2012. doi: 10.1056/NEJMoa1209759.
Standard, large (eg, 14- to 16-gauge) peripheral IV catheters are adequate for most fluid resuscitation. With an infusion pump, they typically allow infusion of 1 L of crystalloid in 10 to 15 minutes and 1 unit of packed red blood cells in 20 minutes. For patients at risk of exsanguination, a large (eg, 8.5 French) central venous catheter provides more rapid infusion rates; a pressure infusion device can infuse 1 unit of packed red blood cells in < 5 minutes.
Patients in shock typically require and tolerate infusion at the maximum rate. Adults are given 1 L of crystalloid (20 mL/kg in children) or, in hemorrhagic shock, 5 to 10 mL/kg of colloid or packed red blood cells, and the patient is reassessed. An exception is a patient with cardiogenic shock who typically does not require large volume infusion.
Patients with intravascular volume depletion without shock can receive infusion at a controlled rate, typically 500 mL/hour. Children should have their fluid deficit calculated and replacement given over 24 hours (half in the first 8 hours).
The actual end point of fluid therapy in shock is to optimize tissue perfusion. However, this parameter is not measured directly. Surrogate end points include clinical indicators of end-organ perfusion and measurements of preload.
Adequate end-organ perfusion is best indicated by urine output of > 0.5 to 1 mL/kg/hour. Heart rate, mental status, and capillary refill may be affected by the underlying disease process and are less reliable markers. Because of compensatory vasoconstriction, mean arterial pressure (MAP) is only a rough guideline; organ hypoperfusion may be present despite apparently normal values. An elevated arterial blood lactate level may reflect hypoperfusion and/or continued sympathetic drive from endogenous catecholamine production; however, lactate levels do not decline for several hours after successful resuscitation. The trend of the base deficit can help indicate whether resuscitation is adequate. Other investigational methods such as measurement of sublingual tissue carbon dioxide or near-infrared spectroscopy to measure tissue oxygenation through the skin may also be considered.
Because urine output does not provide a minute-to-minute indication, measures of preload may be helpful in guiding fluid resuscitation for critically ill patients. Central venous pressure (CVP) is the mean pressure in the superior vena cava, reflecting right ventricular end-diastolic pressure or preload. Normal CVP ranges from 2 to 7 mm Hg (3 to 9 cm water). A sick or injured patient with a CVP < 3 mm Hg is presumed to be volume depleted and may be given fluids with relative safety. When the CVP is within the normal range, volume depletion cannot be excluded, and the response to 100- to 200-mL fluid boluses should be assessed; a modest increase in CVP in response to fluid generally indicates hypovolemia. An increase of > 3 to 5 mm Hg in response to a 100-mL fluid bolus suggests limited cardiac reserve. A CVP > 12 to 15 mm Hg casts doubt on hypovolemia as the sole etiology of hypoperfusion, and fluid administration risks fluid overload.
Because CVP may be unreliable in assessing volume status or left ventricular function, pulmonary artery catheterization may be considered for diagnosis or for more precise titration of fluid therapy if there is no cardiovascular improvement after initial therapy. Care must be taken when interpreting filling pressures in patients during mechanical ventilation, particularly when positive end-expiratory pressure (PEEP) levels exceeding 10 cm water are being used or during respiratory distress when pleural pressures fluctuate widely. Measurements are made at the end of expiration, and the transducer is referenced to atrial zero levels (mid chest) and carefully calibrated.
Ultrasonography of the inferior vena cava and right ventricle can provide information on circulating volume status and overall cardiac function. However, interpretation of the images is highly user dependent and can be complicated by the presence of valvular dysfunction and the use of positive pressure ventilation. Widespread use of ultrasonography to guide volume resuscitation requires more study.
Patients with traumatic hemorrhage shock may require a slightly different approach. Experimental and clinical evidence indicates that internal hemorrhage (eg, due to visceral or vascular laceration or crush) may be worsened by resuscitation to normal or supranormal MAP. Thus, some physicians advocate a systolic blood pressure of 80 to 90 mm Hg as the resuscitation end point in such patients pending surgical control of bleeding, unless higher pressure is needed to provide adequate brain perfusion.
After blood loss is controlled, hemoglobin is used to guide the need for further transfusion. A target hemoglobin of 8 to 9 g/dL (80 to 90 g/L) is suggested to minimize the use of blood products. Patients who may have difficulty tolerating moderate anemia (eg, those with coronary or cerebral artery disease) are kept above 30% hematocrit. A higher hematocrit does not improve outcome and, by causing increased blood viscosity, may impair perfusion of capillary beds.
Overly rapid infusion of any type of fluid may precipitate pulmonary edema, acute respiratory distress syndrome, or even a compartment syndrome (eg, abdominal compartment syndrome, extremity compartment syndrome).
Hemodilution resulting from crystalloid infusion is not of itself injurious, although hematocrit must be monitored to note whether threshold values for transfusion are met.
Red blood cell transfusion has a low risk of directly transmitting infection, but in critically ill patients, it seems to cause a slightly higher rate of hospital-acquired infection. This risk may be minimized by using blood < 12 days old; such red blood cells are more plastic and less likely to cause sludging in the microvasculature. Other complications of massive transfusion are discussed elsewhere.