Shock (hemodynamic failure) is ubiquitous in the modern intensive care unit (ICU). Venodilation, transudation of fluid from the vascular space into the interstitium and increased insensible losses will all result in hypovolemia in the course of patients with sepsis. Absolute hypovolemia is defined as a reduction of total circulating blood volume, while relative hypovolemia is an inadequate distribution of blood volume between the central and peripheral compartments.
Early goal-directed therapy emphasizes aggressive fluid resuscitation of septic patients during the initial 6 h of presentation. Persistent hypotension after initial fluid resuscitation is common and poses the dilemma of whether the patient should receive additional fluid boluses, positive inotropic agent or a vasopressor. Persistent signs of organ hypoperfusion such as oliguria make timely decision making crucial. While a number of technologies including pulse counter analysis, transpulmonary thermodilution, and bioreatance have all shown promise in the evaluation of volume status of septic patients, bedside ultrasonography has already established itself as a useful tool for evaluating cardiac function. Applying the same echocardiographic techniques to dynamically assess the physiological response to spontaneous or mechanical ventilation, bedside maneuvers and the response to therapeutic interventions will likely become a cornerstone of hemodynamic monitoring in the modern ICU. This chapter reviews the utility of echocardiography for identification of the volume-responsive patient with hemodynamic failure.
Benefits and Pitfalls of Fluid Resuscitation
When hypovolemia (either absolute or relative) is present, fluid resuscitation will provide benefit to the patient by increasing venous return, left ventricular diastolic volume, cardiac output, arterial blood pressure, and ultimately tissue perfusion. The rapidity with which euvolemia is reestablished may be a decisive factor in the eventual outcome. That being said, there is an increasing body of evidence suggesting that fluid resuscitation is not without serious and possibly lethal complications. Those complications may be related to preexisting conditions such as systolic or diastolic heart failure, acute cor pulmonale (ACP), or the development of sepsis-related cardiac dysfunction. In patients with ACP, volume resuscitation may be particularly harmful as it may cause further right ventricular (RV) enlargement and left ventricle (LV) compression, thus worsening the shock state (see Chapter 11 and Videos 10-1 and 10-2). ACP can be readily recognized by the clinician with basic-level echocardiography skills. Extravasation of prescribed fluids may result in worsening of acute respiratory distress syndrome (ARDS) and prolonged mechanical ventilation. Anemia and clotting disorders occur with hemodilution. Excessive fluid resuscitation can be positively correlated with increased mortality in the ICU. Given the risk to benefit ratio of volume expansion, the key question is whether the patient would benefit from additional fluid boluses. It is essential to make this determination as clinical studies have repeatedly demonstrated that only about 50% of hemodynamically unstable ICU patients are volume responsive (see definitions below).
Fluid Challenge Versus Volume Responsiveness
Previously this key question was answered by administering a “fluid challenge” of 30 mL/kg of crystalloid solution and the patients’ clinical (blood pressure, heart rate, urine output) and hemodynamic response (central venous pressure [CVP], pulmonary artery occlusion pressure [PAOP]) to the challenge was evaluated. Importantly, because a fluid challenge has to be given to assess volume responsiveness and hypervolemia is associated with significant complications, it is possible, that the increases in mortality associated with invasive hemodynamic monitoring may be attributed to this approach. Given the increased mortality associated with excessive fluid resuscitation it seems prudent to predict the response to a fluid bolus prior to administering the bolus; a concept known as volume responsiveness.
The standard definition of volume responsiveness is a >15% increase in cardiac output in response to volume expansion. Although the volume of the fluid bolus has not been well standardized, a volume of between 500 mL and 1000 mL of crystalloid solution has been most studied. One or more baseline hemodynamic parameters are measured and evaluated for the ability to discriminate between responders and nonresponders.
A static parameter is measured under a single-ventricular loading condition and is presumed to reliably estimate the preload of the right ventricle (RV), LV, or both ventricles (Figure 10-1). This estimation is used to evaluate the probability of responsiveness to ventricular filling, by assuming that a lower preload increases the probability of a response to volume expansion. Several static parameters of ventricular preload have been used in the ICU; some are based on direct pressure measurements, while others use echocardiographic indices.
(A) The Frank–Starling curve indicating a patient with preload responsiveness; the increase of preload is followed by a significant increase of stroke volume, indicating that the patient is in the ascending part of the Frank–Starling curve. (B) A patient without preload responsiveness; the increase of preload is not followed by a significant increase of stroke volume, indicating that the patient is in the horizontal part of the Frank–Starling curve.
The traditional approach to fluid resuscitation consists of measuring a pressure parameter such as the CVP or pulmonary artery occlusion pressure (PAOP) together with a cardiac output determination. The clinician would then prescribe a “fluid challenge” and reassess the above-mentioned parameters. This approach has been largely discredited by the data suggesting a poor or no correlation between the CVP or PAOP and volume responsiveness as well as intravascular volume. Nevertheless, the vast majority of intensivists still utilize the CVP to assess volume status and the major critical care societies advocate for CVP as a measure of successful fluid resuscitation. Multiple studies have demonstrated that the response to a fluid challenge even in healthy volunteers cannot be predicted by either the CVP or PAOP. In a study by Kumar et al. in healthy subjects, static indices of ventricular preload (CVP, PAOP, left ventricular end diastolic volume [LVEDV] index, right ventricular end diastolic volume [RVEDV] index) and cardiac performance indices (cardiac index, stroke volume [SV] index) were measured before and after 3 L of normal saline loading. In this study, there was no correlation between baseline static pressure parameters and changes in the cardiac performance indices (cardiac index, SV index) after fluid loading. Similarly, there was no correlation between changes in the CVP and PAOP and changes in cardiac performance. A meta-analysis by Coudray et al. reviewed five studies on a mixed population of spontaneously breathing critically ill patients and demonstrated the absence of a correlation between the initial PAOP and the response to a crystalloid infusion (an average of 1 L).
Noninvasive echocardiography has advantages over pressure-derived parameters particular those obtained from CVP catheters or pulmonary artery catheterization. Transthoracic echocardiography (TTE) is preferred; however, in certain circumstances transesophageal echocardiography (TEE) may be required. The CVP and PAOP (left atrial pressure) can be approximated by echocardiography. In spontaneously breathing patients, there is a fairly good correlation between the size of the inferior vena cava (IVC) and the CVP. However, Feissel et al. demonstrated that the absolute IVC size failed to predict fluid responsiveness in patients with septic shock.
Left atrial pressure (left ventricular end diastolic pressure [LVEDP], PAOP) estimates involving the use of Doppler mitral flow, E/A ratio, pulmonary venous flow, tissue Doppler (E/e′ ratio) or colored coded Doppler (E/Vp ratio) may be helpful. While beyond the expertise level of most American intensivists, the estimated left atrial pressure (LVEDP, PAOP) can be estimated, utilizing formulas: LAP = 1.9 + 1.24 × E/e′ (mmHg) or LAP = E/e′ + 4 (mmHg) as a part of a comprehensive echocardiographic examination performed by an experienced operator. However, it is also worth noting that the PAOP fails to predict volume responsiveness whether measured directly or by echocardiography. The RV and LV diastolic diameter or area has been used as a measure of preload. However, Tavernier et al. and Feissel et al. have demonstrated that LV size (left ventricular end-diastolic area [LVEDA]) are not useful predictors of fluid responsiveness in patients on mechanical ventilation, unless the LV is very small and hyperkinetic (Figure 10-2 and Videos 10-3 and 10-4). A meta-analysis by Marik et al. demonstrated the failure of the LVEDA to predict volume responsiveness in mechanically ventilated patients.
Generally speaking, static parameters appear to be poor predictors of volume responsiveness except in patients with relatively obvious hypovolemia, which is an uncommon event in the modern ICU. It can be concluded that standard static indices of preload are not useful in predicting volume responsiveness in ICU patients. This observation may be due to dynamic changes in left (LV) and to a lesser degree RV compliance, making the diastolic pressure–volume relationship nonlinear, unpredictable and perhaps subject to change during resuscitation itself. Systolic left ventricular function is also a subject to change in critically ill, both in those, with or without pre-existent cardiac disease. Vieillard–Baron and coauthors demonstrated the development of systolic left ventricular dysfunction in 60% of patient with septic shock. Changing left ventricular function makes it difficult to predict the position of the patient on his/her Frank–Starling curve. It is even difficult to estimate which family of Frank–Starling relationships should utilized to predict fluid responsiveness (Figure 10-1). Furthermore, the development of acute RV failure (ACP), particularly in patients receiving mechanical ventilation with high plateau pressures (>27 cm H2O) further confounds the issue (see Chapter 11). Unrecognized acute RV failure can mimic hypovolemia hemodynamically but would not respond or even get worse with volume expansion (Video 10-5). Dynamic hemodynamic parameters offer the intensivists the best opportunity of predicting response to fluid resuscitation.