Fluid Management in the Complicated Obstetric Patient

A variety of physiologic cardiovascular changes occur during pregnancy, and these must be considered when managing volume status in complicated obstetric patients. Beginning in early pregnancy, total body water slowly increases by 6 to 8 L due to retention of an additional 500 to 900 mEq of sodium.^1,2,3 This leads to a steady rise in plasma volume throughout the first two trimesters and into the early third, with a plateau at approximately 32 weeks.⁴ In a singleton pregnancy at term, the plasma volume is nearly 50% greater than that seen in the nonpregnant individual.⁵ A smaller but parallel increase occurs in red cell mass, with a resultant physiologic fall in hemoglobin concentration.⁴

Maternal cardiac output starts to increase at approximately 10 weeks and plateaus near the end of the second trimester at levels 30% to 50% higher than nonpregnant values.^6,7,8,9,10 This results from an increase in both stroke volume and heart rate. In the third trimester, heart rate reaches a peak of 10 to 15 beats/min over base-line.¹⁰ Systolic and diastolic blood pressures fall throughout the first two trimesters, reaching a nadir at 24 to 28 weeks before increasing to nonpregnant levels at term.¹¹ The systolic pressure decreases an average of 5 to 10 mmHg, while diastolic pressure falls 10 to 15 mmHg.¹² Blood pressure and cardiac output may be further affected by maternal posture. Late in pregnancy, the gravid uterus can mechanically obstruct the aorta and vena cava while supine.^13,14 In addition, the changes in cardiac output and blood pressure result in an initial decrease in systemic vascular resistance (SVR), followed by a rise toward nonpregnant values near term.¹¹

Colloid oncotic pressure (COP) is another important variable affected by pregnancy. Both plasma and interstitial COP fall throughout gestation, the latter decreasing to a greater extent.¹⁵ There is an accompanying increase in capillary hydrostatic pressure.¹⁶ An increase in hydrostatic pressure or decrease in plasma COP may overcome the delicate balance and favor edema formation in late pregnancy. After delivery, there is a further fall in plasma COP, reaching a nadir between 6 and 16 hours and returning toward intrapartum levels after 24 hours.^17,18

The described physiologic changes seen in pregnancy affect both the assessment of a pregnant patient's volume status as well as subsequent treatment. In this chapter, discussion of fluid management will be limited to pregnancies complicated by hemorrhage, pregnancy-induced hypertension, or septic shock.

Hemorrhage commonly complicates childbirth and remains a major cause of maternal mortality.¹⁹ Blood transfusion secondary to hemorrhage during gestation is relatively common, occurring in 1% to 2% of pregnancies.^20,21 Although uterine atony is the leading cause of significant bleeding requiring transfusion, retained placenta, trauma, placenta previa, and abruptio placentae are also important etiologies.²⁰ The problem of hemorrhage in pregnancy is further aggravated by the notorious inaccuracy of clinical determinations of blood loss. In cases of postpartum hemorrhage, visual estimations can be in error by as much as 50% when compared to quantitative measures.^22,23,24 This results in a large underestimation of the prevalence of postpartum hemorrhage.²⁵ Additionally, normal vital signs do not necessarily preclude the presence of significant blood loss.²⁶ The above factors may lead to a delay in diagnosis with an associated delay in therapy. The initial step in appropriate management is maintaining a high index of suspicion.

The previously described normal physiologic changes that occur during gestation allow most pregnant patients to tolerate the inevitable blood loss associated with delivery. In early hemorrhage, vascular tone increases, as do heart rate and myocardial contractility, to improve oxygen delivery. Cardiac output is redistributed, selectively maintaining perfusion to the adrenal glands, brain, and heart at the expense of other organs, including the uterus. Before delivery, this shunting may lead to fetal hypoxia and distress. With continued blood loss and delayed or inadequate resuscitation, secondary changes occur in the microcirculation. Initially, interstitial fluid enters capillary beds. Later, capillary endothelial damage occurs, resulting in an increased permeability and leakage of fluid back into the interstitial space. Finally organ ischemia and cell death result.²⁷

The primary goal of therapy is to restore and maintain tissue oxygen delivery. This begins with aggressive replacement of intravascular volume. At the same time, supplemental oxygen should be added and efforts begun to control the cause of blood loss. Crystalloid solutions are most commonly used in initial fluid resuscitation. They are inexpensive, readily available, and often already infusing as a routine. Lactated Ringer's and 0.9% sodium chloride (normal saline) are the two most common crystalloid solutions. They distribute primarily throughout the extracellular space, expanding both the intravascular and interstitial compartments. With the infusion of 1 L of lactated Ringer's solution, 200 mL will remain in the vasculature while 700 mL enters the interstitium.²⁸ Roughly 3 L of crystalloid are required for each liter of blood loss. Lactated Ringer's has the advantage of containing small quantities of additional electrolytes and lactate (Table 1). The lactate is converted to bicarbonate by the liver. In theory, the bicarbonate may buffer some of the lactic acidosis produced from poor perfusion and offset expansion acidosis resulting from dilution of existing buffers. Lactated Ringer's is recommended by the American College of Surgeons as the initial fluid for resuscitation. Normal saline is a secondary choice because of the potential for hyperchloremic acidosis.²⁹

TABLE 1. Composition and Properties of Crystalloid Solutions

*Also contains K⁺ (4 mEq); Ca²⁺ (3 mEq); and lactate (28 mEq), which is converted to HCO₃^-.

Colloid solutions may be used as an alternative or an adjunct to crystalloids. Colloids are solutions containing large-molecular-weight substances. Their distribution is limited primarily to the intravascular space; thus, intravascular volume is expanded with little increase in the interstitial volume. All of these solutions have been associated with anaphylaxis, but the incidence is low (less than 0.04%). Characteristics of these preparations are described in Table 2. Albumin in a 5% solution is the colloid most commonly used for volume expansion. Adverse effects seen with albumin include a decrease in free calcium levels, which may be due in part to citrate binding. There is also a decrease in platelet aggregation as well as dilution of clotting factors, which may lead to prolongation of the prothrombin time (PT) and partial thromboplastin time (PTT). Preparation of the 5% albumin kills HIV and hepatitis B and C viruses, thus eliminating the risk of infection with these agents. Plasma protein fraction consists primarily of albumin with lesser amounts of alpha- and beta-globulins. Hetastarch is derived from cornstarch and contains molecules of varying sizes. The plasma expanding effects are long lasting. Its use may be associated with a rise in amylase levels during the first 24 hours. An increased bleeding time may be seen as a result of increased fibrinolysis, decreased platelet adhesion, and decreased factor VIII activity. Dextrans consist of high- and low-molecular-weight preparations: dextran 70 and dextran 40, respectively. These solutions have been associated with bleeding due to decreased platelet adhesion and dilution of clotting factors.³⁰

TABLE 2. Characteristics of Colloid Solutions

Considerable controversy exists as to whether crystalloids or colloids are the optimal fluid for initial resuscitation in hypovolemic shock. Crystalloids require considerably larger infused volume than colloids to obtain a comparable level of volume expansion. Theoretically this excess volume distributed within the interstitial space may impair oxygen transport to cells.^31,32 In addition, the duration of the effect is also shorter with crystalloid solutions. In an animal model, lactated Ringer's solution was able to restore but not maintain cardiac output after severe hemorrhage.³³ A comparison of 0.9% saline, 5% albumin, and 6% hetastarch in a prospective randomized trial found that saline infusion required two to four times the volume of the other solutions to obtain the same effects.³⁴ This same study and others have shown that the saline solution also decreased COP while albumin and hetastarch increased the COP.^34,35 The concern with decreasing COP combined with increasing hydrostatic pressure associated with volume replacement is that there will be a decrease in the gradient between COP and pulmonary capillary wedge pressure (PCWP). This increases the risk of developing pulmonary edema.^36,37,38 Pulmonary lymphatic flow is very efficient in normal lungs and provides a protective mechanism against pulmonary edema in this setting.³¹ Use of colloid solutions does not eliminate the risk of pulmonary edema, however. In albumin-resuscitated patients, Lucas and associates^39,40 reported a prolonged need for ventilatory support compared to those receiving crystalloids. When capillary permeability is increased, the high-molecular-weight substances may leak out of the vasculature, increase the interstitial COP, and lead to worsened and prolonged pulmonary edema.⁴¹

Renal function is also a management concern in patients being treated for hypovolemic shock. Both animal and human data suggest that crystalloids are better at preserving renal function than colloids. Renal blood flow is increased with colloid solutions; however, this does not necessarily increase urine output or protect against acute tubular necrosis.³⁹ Using a baboon model, Siegel and colleagues⁴² compared lactated Ringer's alone to a combination of lactated Ringer's and 5% albumin. Those animals treated with the crystalloid alone showed rapid return of urine output and renal function, whereas those treated with the colloid-crystalloid combination exhibited a delay in improvement. Lucas and colleagues^39,40 found that renal function was adversely affected in humans as well. There was a decrease in glomerular filtration rate and urine output, as well as an increased incidence of acute renal failure, in those patients who were resuscitated with albumin compared to crystalloids. There was also a significant delay in diuresis of excess fluid. Despite the longstanding controversy surrounding the choice between colloid and crystalloid solutions for fluid resuscitation, there has been no clear and convincing evidence that overall morbidity or mortality is reduced with the use of one solution over the other.^31,43,44

A variety of hypertonic saline solutions alone or in combination with colloids have been investigated as an alternative to both crystalloids and colloids in the resuscitation of patients in hemorrhagic shock. Animal experiments have identified several beneficial effects of hypertonic saline when compared to isotonic crystalloid solutions. In a pig model, improved hemodynamic values can be obtained with significantly smaller volumes.⁴⁵ In sheep, cardiac output, plasma volume, and urine output were significantly increased, whereas systemic and pulmonary vascular resistance were decreased.^46,47 Survival was increased in dogs subjected to severe hemorrhagic shock. Hemodynamically, cardiac output was increased while heart rate decreased. Mean arterial pressure (MAP) and mesenteric blood flow also increased. Despite a significant lack of change in hematocrit or plasma volume in these animals with resuscitation, there was no fixed acid accumulation. This suggests that a direct effect of the hypertonic saline solution is to increase myocardial contractility and to dilate resistance vessels.⁴⁸

Caution in extrapolating animal data to humans regarding the use of hypertonic saline solutions for resuscitation has been urged. Concerns have been expressed regarding decreased intracranial pressure and increased cerebral blood flow leading to intracranial hemorrhage, in addition to complications associated with hypernatremia and hyperosmolality.⁴⁹ Nevertheless, beneficial effects have been demonstrated in humans. In an investigation of postoperative cardiac bypass patients comparing hypertonic and normal saline solutions, smaller volumes of the hypertonic solution were required to obtain equivalent hemodynamic measurements, as well as oxygen delivery and consumption. There was a greater amount of chest drainage in the normal saline patients, while hypernatremia and hyperosmolality occurred in some of the hypertonic saline patients, leading to discontinuation of the infusion.⁵⁰ Potential fetal effects are unknown, as there are no reports on the use of hypertonic saline in the pregnant patient.

In severe hemorrhage, the goal of restoring oxygen delivery to the tissues is only partially achieved by initial intravascular volume replacement and improved perfusion. Significant blood loss resulting in symptomatic decreased oxygen-carrying capacity requires blood transfusion.^51,52 Blood component therapy has largely eliminated the use of whole blood in hypovolemic shock. Reasons for the move away from whole blood include decreased availability due to a short shelf life (less than 7 days old) and inactivity of platelets, factor V, and factor VIII after 24 hours.⁵¹ In addition, by separating blood into its components, 1 unit can benefit several patients. Packed red blood cells are the component of choice for the treatment of hypovolemic shock due to hemorrhage.⁵¹ Each unit can be expected to increase the hemoglobin by approximately 1 g/dL.⁵²

Concerns about the morbidity associated with blood transfusions, particularly the acquired immunodeficiency syndrome, have stimulated education and research efforts aimed at the effective and efficient use of blood products.^53,54,55 As previously noted, packed red blood cells should be given when oxygen-carrying capacity is inadequate. They should not be used for volume expansion, enhancement of wound healing, or when time permits use of a hematinic.⁵² There is no arbitrary hemoglobin threshold that should automatically trigger a transfusion. Traditionally a hemoglobin of greater than 10 g/dL has been considered optimal in compromised patients. This is based on the assumption that arterial oxygen content should exceed myocardial oxygen extraction.⁵⁶ Clinical experience does not appear to support this assumption. In an otherwise healthy person, cardiac output does not dramatically increase until the hemoglobin falls below 7 g/dL. Animal data would suggest that wound healing is not impaired until the hematocrit decreases to 15%.⁵³ Dietrich and co-workers⁵⁷ found no further improvement in volume-resuscitated shock patients with an increase in hemoglobin above 8.3 g/dL. In a pilot study, the Canadian Critical Care Trials Group⁵⁸ compared the outcome of critically ill patients randomly assigned to receive transfusions to maintain hemoglobin in the range of 10 to 12 g/dL or 7 to 9 g/dL. They found no difference between the two groups in mortality or the development of organ dysfunction.

Fresh frozen plasma (FFP) is the plasma portion of a unit of whole blood frozen within 8 hours of collection. It contains plasma protein and clotting factors and is reserved for patients with a demonstrated clotting deficiency.^51,59 FFP is reserved for those patients with bleeding and prolongation of the PT and PTT. In the acute setting, when it may not be practical to wait for laboratory results to confirm a coagulopathy, a red-topped tube may be taped to the wall during the resuscitation and observed for clot formation. A clot should form within 5 minutes. In the absence of active bleeding, prolongation of PT and PTT values less than 1.5 times control rarely requires treatment. FFP may also be of benefit in the treatment of thrombotic thrombocytopenic purpura and hemolytic uremic syndrome. FFP should not be used for volume expansion. It also should not be used prophylactically with red cell transfusions. In the case of massive blood transfusion, where replacement of more than one blood volume has taken place over a few hours in the presence of continued bleeding, FFP may be used.⁵⁹

For isolated hypofibrinogenemia, cryoprecipitate should be used rather than FFP.^51,54 Cryoprecipitate is the insoluble portion of FFP obtained when thawing between 1°C and 6°C. It contains factors VIII and XIII, von Willebrand's factor, and fibrinogen. Because cryoprecipitate contains concentrated fibrinogen, it also may be beneficial in cases of consumptive coagulopathy, where elevation of fibrinogen cannot be achieved quickly enough with FFP. The goal of therapy is to obtain a fibrinogen level greater than 100 mg/dL.⁵⁹

Platelet use should also be limited to well-established indications. In the bleeding patient, transfusion is indicated when the platelet count falls below 50,000/mm³. As FFP, platelets should not be given prophylactically with massive blood replacement.^51,60 Dilutional thrombocytopenia occurs, but after replacement of one blood volume, 35% to 40% of platelets are still present. The patient's clinical condition and serial platelet counts should guide platelet therapy. Without other complicating factors, most patients will not develop bleeding even with replacement of one to two blood volumes.⁶¹ If bleeding does occur in this setting, however, it is most likely related to thrombocytopenia rather than loss of coagulation factors. Prophylactic transfusions are appropriate preoperatively when the platelet count is less than 50,000/mm³. In nonoperative situations where the count falls below 10,000/mm³, prophylactic transfusions are indicated because of the increased risk of serious spontaneous bleeding involving the gastrointestinal tract or central nervous system.⁵⁹ Platelet counts can be expected to rise by approximately 5000 to 10,000/mm³ with each unit of platelet concentrate transfused.^52,59

Transfusion-related morbidity can be significant, and it is one of the main reasons behind efforts to promote the effective use of blood products in this country. Currently, although the risk of transfusion-transmitted infection is low, it is not zero. New and improved screening tests and procedures have continued to increase the safety of the blood supply. Much of the remaining risk is due to the “window period” between infection and development of detectable antibodies. Table 3 describes the per-unit incidence of transfusion-related infections and the duration of the window period.^62,63,64,65 Other complications include acute or delayed hemolytic reactions, alloimmunization, and allergic and febrile reactions.^51,52

TABLE 3. Per-unit Risk of Transfusion-Related Infections and the “Window Period” Where an Early Infection May Go Undetected

HBV = hepatitis B virus; HCV = hepatitis C virus; HIV = human immunodeficiency virus; HTLV = human T-cell lymphotropic virus.
(Adapted from Sloand EM, Pitt E, Klein HG: Safety of the blood supply. JAMA 274:1368, 1995)

Fluid management for the patient with pregnancy-induced hypertension (PIH) presents a challenge for the obstetrician. This is especially true for the subset of patients with preeclampsia. The etiology and pathophysiology of PIH are incompletely understood.^66,67 As a result, considerable controversy exists regarding optimal therapy for most aspects of PIH. Fluid therapy is no exception. In general, PIH is characterized by a failure to achieve the normal pregnancy-associated plasma volume expansion. Later in gestation, this is combined with manifestations of excess interstitial fluid. The paradox of intravascular volume depletion combined with increased extracellular volume leads to divergent opinions regarding optimal fluid management. Volume replacement has been advocated by some^68,69 because of the intravascular hypovolemia, whereas others^70,71 have recommended volume restriction and even diuretics because of the extracellular fluid excess.

We suggest that management plans be tailored to the clinical situation. In uncomplicated PIH, a crystalloid infusion limited to between 75 and 125 mL/hour throughout the intrapartum and early postpartum period (12 to 24 hours) is appropriate.^72,73 Clinical resolution of the disease will begin in most patients during this time with a brisk, spontaneous diuresis. In more complicated cases, or when other therapeutic interventions are planned, the patient's intravascular volume status becomes an important management consideration. Oliguria, pulmonary edema, antihypertensive therapy, and regional analgesia or anesthesia are clinical settings where meticulous fluid management is important.

In the presence of oliguria, defined as urine output of less than 30 mL/hour over a 2-hour period, an initial fluid challenge of 500 to 1000 mL of crystalloid is indicated.⁷⁴ If adequate urine output does not result, further therapy should be guided by careful assessment of oxygenation status and, if compromised, by invasive hemodynamic monitoring. Complications such as pulmonary, cerebral, or laryngeal edema can result from overzealous volume infusion.^75,76,77 Using the pulmonary artery catheter in oliguric patients who failed to respond to an initial fluid challenge, Clark and associates⁷⁴ defined three hemodynamic subsets based on left ventricular function, PCWP, and SVR (Table 4). The first group exhibited a relative intravascular volume depletion and responded to further crystalloid infusion. The hemodynamic profile of the second group was not one of intravascular volume depletion, and the postulated mechanism for oliguria was selective renal arteriospasm. Improvement in this group was seen with vasodilator therapy. In this clinical scenario, consideration could be given to low-dose dopamine, which causes selective renal artery vasodilation. A single patient constituted the final group, and this case was characterized by intravascular fluid overload, decreased left ventricular function, and vasospasm. Treatment consisted of fluid restriction and vasodilator therapy. This group may alternatively benefit from diuretic therapy.

TABLE 4. Hemodynamic Subsets of Oliguric Preeclamptic Patients

PCWP = pulmonary capillary wedge pressure; SVR = systemic vascular resistance.

In the future, the pulse oximeter may serve as an alternative to invasive hemodynamic monitoring in some oliguric patients. We are currently comparing changes in oxygen saturation measured by pulse oximetry to central hemodynamic parameters obtained with the pulmonary artery catheter after fluid bolus.

In patients requiring volume expansion, colloids would seem to be an ideal choice, considering the known early postpartum decrease in colloid osmotic pressure¹⁸ and the further decrease noted with the use of intrapartum intravenous fluids.⁷⁸ However, Kirshon and colleagues⁷³ found no benefit in fetal outcome by using colloids to raise COP unless there was a very prolonged negative COP-PCWP gradient. They did note an increased requirement for diuretic therapy in the group receiving colloids due to an elevated PCWP.

Pulmonary edema associated with PIH occurs after delivery in 70% to 80% of cases.^79,80 Postpartum mobilization of interstitial fluids contributes to the occurrence of pulmonary edema by increasing PCWP.^70,81 The associated increase in hydrostatic pressures, decrease in COP, and increase in vascular permeability resulting from endothelial damage⁶⁶ all combine to increase the risk of pulmonary edema. Management consists of supplemental oxygen, fluid restriction, diuretics, and possibly vasodilators to facilitate afterload reduction. Use of the pulmonary artery catheter in extremely complicated cases to aid therapy should be encouraged.

Antihypertensive agents are often necessary during the peripartum course of PIH. Intravascular hypovolemia can complicate the use of these medications. Intrapartum volume expansion alone has been reported to decrease peripheral resistance and increase cardiac output, whereas MAP remains unchanged or even decreases.⁸² A bolus injection of hydralazine is commonly used for acute blood pressure control during the intrapartum period. It effectively lowers MAP and SVR while increasing cardiac output, but individual patient responses are quite variable. Volume loading before hydralazine therapy may prevent an abrupt and precipitous drop in blood pressure.⁸³ This is particularly important antepartum, when a significant decrease in blood pressure can lead to decreased uterine perfusion and fetal hypoxia. Similar beneficial effects were seen when volume expansion was performed before administration of intravenous nitroglycerin in patients with severe PIH.⁶⁸ With pretherapy hydration, decreases in PCWP, cardiac index, and oxygen delivery were prevented, and the degree of decrease in MAP was reduced.

The use of regional analgesia or anesthesia in PIH patients requires similar considerations concerning volume status as those required by antihypertensive medications. In uncomplicated pregnancy, the sympathetic blockade causes vasodilation. Without adequate rehydration, significant hypotension can occur.⁶⁷ The intravascular hypovolemia present in PIH can compound the sympathetic block, resulting in a profound drop in blood pressure. An associated decrease in uteroplacental perfusion occurs, and fetal well-being may be adversely affected. Attempts to resolve the hypotension with fluid boluses may lead to overhydration, placing the mother at risk of developing pulmonary or cerebral edema, particularly in the postpartum period. In the past, the potential for these complications had led some to advise the avoidance of epidural anesthesia in severe preeclampsia.^84,85 These same sources have more recently softened their positions and have suggested that cautious use of epidural anesthesia in severe preeclampsia by experienced operators is appropriate, as increasing evidence of its safety has been reported.^67,86 Others emphatically state that when properly used regional analgesia or anesthesia is safe and should be considered the method of choice except when a coagulopathy exists.^87,88

Several beneficial fetal and maternal effects of epidural anesthesia have been demonstrated. Intravillous blood flow is improved.⁸⁹ MAP can be appropriately reduced without a significant change in cardiac index, pulmonary or systemic vascular resistance, central venous pressure, or PCWP.⁹⁰ This may be due, in part, to a reduction in maternal catecholamine levels.⁹¹ In vaginal deliveries, Moore and colleagues⁹² found no difference in maternal hypotensive episodes or neonatal outcome between those receiving labor epidural or local anesthesia. In the same report comparing epidural to general anesthesia for cesarean section patients, again there was no significant difference in hypotensive episodes, but there was a significantly higher systolic pressure associated with intubation. There was also greater blood pressure variation. Neonatal outcome was better in the epidural group; however, much of this could be attributed to the fetal condition before anesthesia. All of these patients were placed in a left-lateral-tilt position to prevent aortocaval compression, and 500 to 1000 mL of lactated Ringer's solution was administered before epidural placement to reduce the risk of hypotension.

Fluid management in the pregnant patient with PIH is complex and should be tailored to the specific clinical situation. Careful fluid therapy in which intake and output are balanced is adequate for most uncomplicated cases. When complications arise or therapeutic interventions that lower blood pressure are contemplated, volume expansion may be desirable. In particularly difficult cases, pulmonary artery catheterization may help guide fluid therapy.

Infections complicating pregnancy and the puerperium are common events. The incidence of bacteremia associated with these infections is low and reported at less than 1%. Septic shock may occur in up to 5% of these bacteremic patients. Mortality rates of 20% to 50% are seen with septic shock.⁹³Table 5 lists the most common infections associated with septic shock in the obstetric patient.⁹³ Specific data related to septic shock in pregnancy are limited and consist primarily of experiments using various animal models.^93,94 Nevertheless, prompt empiric initiation of antibiotics and appropriate fluid therapy are cornerstones in management of the pregnant patient, as they are in the nonpregnant population.

TABLE 5. Infections Leading to Septic Shock and Associated Frequency in the Obstetric Patient

(Data from Septic shock. ACOG Technical Bulletin No. 204, April 1995.)

Early septic shock is characterized by several complex pathophysiologic derangements. Initially, cardiac output is increased due to a rise in heart rate. SVR decreases, often precipitously, overwhelming the change in cardiac output.^93,94,95,96 This results in intravascular hypovolemia with poor oxygen delivery. Increased capillary permeability can compound fluid loss from the vasculature. Tissue oxygenation is further compromised by a decreased ability to extract oxygen.⁹⁶ Cardiac dysfunction, demonstrated by left ventricular dilation and decreased ejection fraction, combined with leaky capillaries complicates attempts to improve volume status, while increasing the risk of pulmonary edema.

The immediate therapeutic goal is to improve oxygen delivery and maintain perfusion to vital organs.⁹⁷ Aggressive fluid replacement is the mainstay of early treatment and is aimed at eliminating hypovolemia, improving hypotension, and maintaining or increasing cardiac output. Often, large volumes of fluid are required.⁹⁸ Hypotension and cardiac function may not completely respond to fluid alone. It then becomes important to determine volume status objectively in order to prevent fluid overload, as well as to evaluate the need for vasopressor and inotropic agents. The flow-directed pulmonary artery catheter provides central hemodynamic data necessary to guide these decisions. An overview of the information derived from right heart catheterization is provided in the final section of this chapter.

With the pulmonary artery catheter in place, the PCWP can be observed while fluid challenges are given. A PCWP pressure ranging from 12 to 16 mmHg is considered optimal.⁹³ In a group of patients with septic and hypovolemic shock, Packman and Rackow⁹⁹ found no improvement in stroke volume index, left ventricular stroke work index, or cardiac index when the PCWP was increased from 12 to 16 mmHg, suggesting that left heart filling pressures need not exceed 12 mmHg during resuscitation.

One method of achieving adequate volume replacement is based on the “7–3 rule.” An infusion of 10 mL/min of crystalloid solution is given over a 15-minute period. A rise in PCWP of 7 mmHg or more would slow the infusion rate to maintenance levels, whereas a rise of less than 3 mmHg would prompt a repeat bolus. Further fluid bolus with an intermediate rise of 3 to 7 mmHg would be based on the PCWP and other clinical and hemodynamic indicators.⁹³

Another method is repeated 500-mL crystalloid boluses over a 10-minute interval until the desired PCWP is reached. Red blood cell transfusions should be used as necessary based on hemoglobin levels. If hypotension persists after appropriate fluid resuscitation has been achieved, vasopressors should be added. Many of these drugs also have positive inotropic activity, which may improve cardiac function and elevate a depressed cardiac output. A description of the more common agents is provided in Table 6.¹⁰⁰

TABLE 6. Characteristics of Common Vasoactive Drugs

*Listed in order of increasing vasopressor activity and decreasing inotropic activity.
(Adapted from Billups NF, Billups SM: American Drug Index. St. Louis, Facts and Comparisons, 1996)

Arguments surrounding the choice between crystalloid and colloid solutions have previously been presented and vary little in the septic shock patient. Concerns over enhanced interstitial fluid flux are heightened in the setting of sepsis because of increased capillary permeability. The higher volume requirements and decrease in COP seen with crystalloids and passage of the large colloid molecules out of the vasculature could both theoretically lead to the development of edema, particularly pulmonary edema. Clinically, colloids compared to crystalloids have been associated with a decreased incidence of pulmonary edema in older patients.³⁴ In younger patients typically seen in obstetric practice, these two solutions appear to be equally effective in treating the hypovolemia associated with severe sepsis.⁹⁸

After acute resuscitation, recent interest has focused on improving outcome in nonpregnant critically ill patients, including those in septic shock, by attempting to elevate specific cardiovascular parameters to “supranormal levels.” Various investigators have attempted to manipulate one or more of the following indices: cardiac index,^{101,102,103,104} oxygen delivery,^{101,102,103,105} oxygen extraction,^101,103 and mixed venous oxygen saturation.¹⁰⁵ Initial therapy consisted of volume expansion with crystalloid and/or colloid solutions. Blood was used as indicated. If therapeutic aims were not reached with fluids alone, inotropic agents (dopamine and/or dobutamine) were added. Vasopressors or vasodilators were then added if further augmentation of the designated indices was required. Despite aggressive therapy, the desired supranormal indices were often not obtained. Outcome results comparing attempted supranormal values with conventional therapy varied among reports. Two studies^101,102 identified a decrease in morbidity and mortality. Two found no difference in outcome (including the largest of the trials, which included more than 10,000 subjects).^104,105 The final report suggested that therapy aimed at attaining supranormal values may even be harmful.¹⁰³

Despite the controversy associated with many aspects of the management of septic patients, it is clear that adequate volume expansion is an essential component of therapy. Fluid management should be guided by central hemodynamic monitoring in the critically ill patient because of the many and varied pathophysiologic alterations that can confuse and compound the clinical picture.

Fluid management in most obstetric conditions consists of a peripheral venous catheter with monitoring of vital signs and fluid balance. In the complicated obstetric patient, however, the flow-directed pulmonary artery catheter can be a valuable clinical aid in the assessment and management of volume status. In the early 1970s the Swan-Ganz catheter moved from the research laboratory into medical and surgical intensive care units to facilitate monitoring of the critically ill patient.¹⁰⁵ During the the following decade, use of the pulmonary artery catheter increased in the critically ill obstetric patient.⁸⁸ Common indications for use in obstetrics are presented in Table 7.^{88,106,107,108}

TABLE 7. Indications for the Use of the Pulmonary Artery Catheter in Obstetric Patients

  Indication
  Massive blood loss complicated by

  Respiratory compromise
  Oliguria after clinically appropriate fluid replacement

  Pregnancy-induced hypertension complicated by

  Persistent oliguria unresponsive to fluid challenge
  Pulmonary edema of uncertain etiology or unresponsive to conventional therapy

  Septic shock
  Adult respiratory distress syndrome
  Cardiac disease intrapartum or intraoperatively

  New York Heart Association Class III or IV
  Ischemic heart disease
  Congestive heart failure

Insertion and use of the pulmonary artery catheter requires access to the central venous circulation. This can be obtained through either peripheral (external jugular, antecubital) or central (internal jugular, subclavian, femoral) veins. The antecubital and femoral sites make proper placement technically more difficult. These locations may be preferable in patients with abnormal coagulation parameters in order to prevent an intrathoracic bleed. Operator experience and preference ultimately dictate where access is obtained.¹⁰⁹ The right internal jugular vein is the most commonly selected site in obstetric patients.¹⁰⁷ This approach reduces the risk of pneumothorax, allows easier hemostasis should arterial puncture occur, and avoids the technical difficulties associated with more distal approaches.

Before insertion, the catheter is flushed to remove air, the balloon is tested, and a pressure transducer is attached. The pulmonary artery catheter is then advanced through the venous sheath under continuous electrocardiographic monitoring to observe for arrhythmias. Once it has been advanced into the thorax (10 to 15 cm with internal jugular or subclavian approach), the balloon is inflated and the catheter is rapidly advanced, allowing blood flow to carry it into the pulmonary artery. Proper advancement and positioning is determined by observing the characteristic pressure waveforms as the balloon passes from the right atrium into the right ventricle, then into the pulmonary artery and into the wedged position. Typically the desired location is reached 40 to 50 cm from an internal jugular insertion site.

When the pulmonary artery catheter is properly placed, central venous pressure can be measured from the proximal port. Pulmonary artery pressures and PCWP can be obtained from the distal port. Cardiac output calculations can be performed by injection of chilled crystalloid solutions. Use of a fiberoptic catheter allows continuous measurement of mixed venous oxygenation, which can be used to determine oxygen delivery and extraction. Comparison of normal values for cardiovascular and respiratory parameters between nonpregnant and pregnant patients is presented in Table 8.¹¹⁰

TABLE 8. Normal Values for Cardiorespiratory Parameters of Term-Pregnant and Nonpregnant Patients

(Modified from Clark SL, Cotton DB, Lee W et al: Central hemodynamic assessment of normal term pregnancy. Am J Obstet Gynecol 161:1439, 1989)

Although the safe use of the pulmonary artery catheter has been documented in obstetric patients, the operator must be aware of potential complications. In a series of 75 pulmonary artery catheterizations in obstetric and gynecologic patients reported by Clark and co-workers,¹⁰⁷ ventricular ectopy developed in 27% of patients, arterial puncture in 8%, and superficial cellulitis in 3%; one patient (less than 1%) had a pneumothorax. Other complications that can occur include venous thrombosis, systemic infection, pulmonary infarction, and balloon rupture. In an autopsy series reported by Raad and associates,¹¹¹ 38% of catheterized veins had mural thrombi. Right atrial mural thrombi were identified in 6% of patients, and another 6% had nonbacterial thrombotic endocarditis. Those patients with thrombosis had a much higher incidence of septicemia. Less common complications include pulmonary artery rupture, cardiac tissue or valvular injury, complete heart block, and catheter knotting.¹¹²

Fluid therapy in the complicated obstetric patient requires knowledge and application of the normal physiologic changes found in pregnancy. An understanding of the pathophysiologic alterations associated with conditions likely to require fluid manipulation in the pregnant patient is also necessary. It is important to know the characteristics of the products available for volume expansion as well as their advantages and disadvantages. It is also necessary to be familiar with the drugs that may be used in conjunction with fluid therapy in specific clinical situations. Finally, invasive hemodynamic monitoring can be a valuable tool to aid in assessing the patient's volume status and in achieving the final therapeutic goals.