| JOURNAL HOME | CME HOME | THIS MONTH | PAST ISSUES | ETOC | COLLECTIONS |
| AUTHORS | REVIEWERS | EDITORIAL BOARD | FEEDBACK | RSS | HELP |
| ||||||||||||||
| ||||||||||||||
|
|
|
|
||||||||||||
|
||||||||||||||
Department of Intensive Care Medicine, University Hospital Bern, Bern, Switzerland
Address correspondence to Stephan M. Jakob, MD, PhD, Department of Intensive Care Medicine, University Hospital, Inselspital, CH-3010 Bern, Switzerland. Address e-mail to stephan.jakob{at}insel.ch
Establishing appropriate tissue oxygenation is a fundamental component of intensive care treatment. This includes optimal oxygenation, intravascular fluid replacement with or without red blood cells, administration of vasoactive drugs, and minimizing oxygen consumption. Although there is evidence that inadequate respiratory (1) and fluid (2,3) management, including administration of red blood cells (4), is associated with organ dysfunction and increased mortality, increasing oxygen delivery does not uniformly result in a better outcome (5,6). Patient groups who may profit are high-risk surgical patients (7–9) and patients with early sepsis (10). Conversely, administration of crystalloids (11,12), colloids (13,14), and erythrocytes (4,15) can harm patients for different reasons. Although the risks associated with red blood cell administration, mainly infection and anaphylaxis, are both increasingly rare and well known, the risks of crystalloid and colloid administration are probably underestimated. One of these risks is edema formation after administration of large amounts of crystalloids in conditions with capillary leakage (16). Unfortunately, blood volume cannot be reliably assessed clinically (17,18) and there are not many substitute variables which consistently predict the response to fluid administration (19–21).
Vasoactive drugs can also harm the patient. For instance, dopamine may interfere with hormonal homeostasis (22), and dopamine and phenylephrine can impair hepatic metabolic functions (23,24). In addition, vasoactive drugs can also cause harm as a result of their main effects; e.g., ß-adrenergic drugs increase cardiac oxygen demands and can induce or worsen cardiac dysfunction (25), and vasoconstrictive drugs may lead to impaired tissue perfusion and overt ischemia (26,27).
When the positive effects of vasoactive drugs (re-establishing adequate tissue oxygenation) outweigh the adverse effects, the latter are usually accepted. This general principle implies that both intended and adverse effects can be measured, which is not necessarily the case. Substitutes for organ functions such as urinary output do not reliably assess true organ function, and drugs, in this case diuretics, may improve the variable without changing the underlying dysfunction. Moreover, such substitutes may not reflect subtle changes or they may react with a certain delay, and some organs may be affected earlier than others from impaired oxygen supply.
Another possibility to assess adequate tissue oxygen delivery is the measurement of global, indirect variables of tissue oxygenation, such as lactate or base excess, or direct variables of oxygen transport, such as cardiac output or mixed venous oxygen saturation.
Although the assessment of intended effects of a therapy can be difficult, the detection of side effects of fluid and drug administration can be even more trying. They can resemble symptoms of the disease they are being used to treat, or they can simply be hidden for a prolonged period of time (e.g., non-occlusive mesenteric ischemia)(28).
In this issue of Anesthesia & Analgesia, Lee et al. (29) present four critically ill patients who developed blindness as a result of ischemic optic neuropathy (ION) within 1 mo of each other. Although ION has been described perioperatively, especially in patients after prone position during spine surgery, it is a rare complication in intensive care unit (ICU) patients. Although all of the patients described by Lee et al. had known risk factors for ION, such as hypotension, venous congestion, large blood loss, and use of vasoconstrictors, in all but one of them visual abnormalities developed before any operative procedures.
The occurrence of four cases of ION within 1 mo suggests that there might be additional risk factors or an unusual combination of several known risk factors. The four patients suffered from pancreatitis and hypovolemia (one patient), multiple fractures (two patients), and intraabdominal sepsis (one patient). All were men between 50 and 60 yr of age and all of them had a history of alcohol abuse. All had systolic blood pressure <70 mm Hg during the initial presentation, were acidotic with some degree of hypoxia (lowest PaO2, 49–56 mm Hg), and had mild or moderate anemia (lowest hematocrit between 0.23 and 0.34). For correction of acidosis and low arterial blood pressure, the patients received 22–30 L of crystalloids and blood products within 18–24 h and continuous infusions of between 2 and 4 vasoconstrictive and inotropic drugs, including vasopressin up to 0.09 U/min, norepinephrine up to 0.55 µg · kg–1 · min–1, and dobutamine up to 20 g · kg–1 · min–1. The patients were mechanically ventilated with a maximal positive end-expiratory pressure of 25 mm Hg. Their central venous pressures increased to 22, 24, 40, and 41 mm Hg, respectively. All patients developed one or several organ failures within a few days, including acute respiratory distress syndrome, abdominal compartment syndrome, and acute tubular necrosis. The authors discuss the possibility that the combination of hypotension, hypoxia, high levels of positive end-expiratory pressure, infusion of large amounts of fluid, and prolonged administration of multiple vasoconstrictive drugs, including vasopressin, may explain the strikingly frequent incidence of ION in their ICU population. They conclude that "it is difficult to make recommendations for these patients, without risking their death, regarding alternative therapy to avoid the devastating visual outcome."
Recently, guidelines for the treatment of patients with sepsis have been published (30). In these patients, recommendations for fluid therapy and vasoactive drugs reach the evidence level B (supported by one level I investigation). Target mean arterial blood pressure of 65 mm Hg and central venous pressure of 8–12 mm Hg or 12–15 mm Hg in mechanically ventilated patients or patients with increased abdominal pressure are recommended.
What is the current practice? A closer look at the literature on the management of patients with septic shock reveals target mean arterial blood pressures of 60–80 mm Hg, but effective arterial blood pressures of 80–95 mm Hg (31–34). The same is true for the filling pressures. Lee et al. attempted to reach "reasonable" values for mean arterial blood pressure (70 mm Hg) and filling pressure (central venous pressure, 10–14 mm Hg). But what should be done when the "reasonable" targets can only be reached with an "unreasonably" large amount of volume or pharmacological support? Is it reasonable to administer 30 L of crystalloids and blood products within 24 h to a patient who is not bleeding excessively? Is it reasonable to combine four different vasoactive drugs to treat shock? Or would it be more reasonable to accept lower pressures once a significant amount of fluid and vasoactive support had been infused?
Admittedly, in sepsis or severe inflammatory states, it is often mandatory to infuse large amounts of fluids to improve microcirculation. For the clinical assessment of microcirculation, skin color and temperature, especially of the limbs, and capillary refill time are often used. The reason why the microcirculation should improve in response to fluid administration is an improved cardiac performance by correction of intravascular fluid depletion resulting from capillary leakage. Unfortunately, a significant amount of the infused fluids leaves the intravascular compartment rapidly when capillary leakage is present. One could speculate that once tissue edema is large enough, the Starling forces would correct the situation by increasing extravascular hydrostatic pressure and at the same time decrease extravascular oncotic pressure by dilution. To counteract the Starling forces more rapidly, some advocate the use of hyperosmolar NaCl infusions or albumin. However, in conditions with increased vascular permeability, leakage of albumin into the extravascular space can aggravate edema formation (16,35). It has recently been demonstrated that using large amounts of fluids to resuscitate trauma victims, as compared with normal resuscitation, significantly increases the incidence of intraabdominal hypertension, abdominal compartment syndrome, multiple organ failure, and death (36). Hence it seems reasonable, and even mandatory, to limit fluid administration when the microcirculation cannot be improved.
This implies frequent assessment of cardiac performance and microcirculation, especially in response to fluid and vasoconstrictive drugs and consequently with a physician at the bedside. Vasoconstriction, where it occurs, by definition does not open up the microcirculation. In a situation of impaired microcirculation, a target arterial blood pressure of 70 mm Hg may not be adequate and filling pressures may not help guide the treatment because the compliance of the vascular system is low and the target can only be reached at the expense of extremely large amounts of infused fluids. Thus administration of vasoconstrictors seems to be most reasonable to compensate for excessive vasodilatation, once the microcirculation is "open."
In recent years, the use of small-dose vasopressin in the treatment of vasodilatory shock and after cardiac arrest has generated interest. Vasopressin has been advocated to support other vasoconstrictors when the target arterial blood pressure cannot be reached. However, even at small doses, vasopressin impairs mesenteric perfusion (37), and I would be extremely reluctant to administer it as long as this effect cannot be monitored in the clinical setting.
Recently, the association between treatments and outcomes was evaluated in a large cohort of critically ill patients and published under the title "More interventions do not necessarily improve outcome in critically ill patients" (38). Lee et al. argue that "as ICU therapy advances, the cost of improved survival may be increased morbidity and end organ sequelae." The cases they present illustrate the need for continuous and careful review of old and new approaches in the treatment of critically ill patients, with regard to rationale and desired and undesirable effects. "Advanced" therapy should result in a better patient outcome. Unfortunately, complications such as those described by Lee et al. in this issue of Anesthesia & Analgesia may markedly tilt the balance between benefit and harm
 |
References |
|---|
 Top References |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
