Chapter 20. The Pathophysiology of the Circulation in Critical Illness

• Left ventricular (LV) stroke volume (SV) creates arterial pulse pressure (PP) by distending conducting vessels during systole, and systemic vascular resistance (SVR) preserves diastolic pressure (DP) by preventing SV from flowing through arterioles during diastole.
• This coupling of ventricular and vascular elements allows rapid clinical separation of hypotensive patients into those with increased SV and cardiac output (Q̇T) demonstrating bounding pulses with large PP, low DP, and warm digits (low SVR, high Q̇T hypotension, or septic shock) from those who demonstrate thready pulses with small PP and cool digits signaling low SV and Q̇T with increased SVR, as in cardiogenic or hypovolemic shock.
• LV pumping function is described by relating LV end-DP as estimated by pulmonary wedge pressure (Ppw) to SV; LV dysfunction is signaled by increased Ppw and decreased SV and may be due to systolic or diastolic dysfunction.
• Systolic dysfunction, or decreased contractility, connotes increased LV end-systolic volume for a given LV end-systolic pressure that is approximately the mean blood pressure (BP); common causes of acute systolic dysfunction in critical illness are myocardial ischemia, hypoxia, acidosis, sepsis, intercurrent negative inotropic drugs (β or calcium blockers), and acute-on-chronic systolic dysfunction in cardiomyopathies.
• Diastolic dysfunction connotes decreased LV end-diastolic volume despite increased Ppw because the heart cannot fill normally; common causes of diastolic dysfunction in critical illness are pericardial tamponade or constriction, positive end-expiratory pressure (PEEP), or other causes of increased pleural pressure as in pneumothorax, pleural effusion, or abdominal distention, ventricular interdependence in acute right heart syndromes, and chronic LV stiffness as in LV hypertrophy.
• Early differentiation between diastolic and systolic dysfunctions in critical illness is aided by a questioning approach and dynamic imaging such as echocardiography; this avoids inappropriate and ineffective therapy for the wrong etiology of LV dysfunction.
• Venous return (VR) to the right atrium is controlled by mechanical characteristics of the systemic vessels (unstressed volume, vascular capacitance, vascular volume); together these determine the mean systemic pressure responsible for driving VR back to the right atrium (Pra) through the resistance to VR.
• For a given Pms, VR increases as Pra decreases to define the VR curve of the circulation, whereas SV and Q̇T from the heart increase as Pra and preload increase to define the cardiac function curve that intersects the VR curve at a unique value of Pra where VR = Q̇T.
• When Q̇T is insufficient, volume infusion, baroreceptors, or metabolic receptors can increase mean systemic pressure to increase VR and Pra; this effect is mimicked by venoconstricting drugs such as dopamine or epinephrine; alternatively, VR can be increased by positive inotropic (dobutamine) or afterload-reducing (sodium nitroprusside) drugs that decrease Pra by enhancing cardiac function.
• In hypovolemic shock, hemostasis and volume resuscitation are essential, whereas arteriolar constricting agents such as norepinephrine may be used briefly to provide a window of higher BP; in septic shock, considerable volume resuscitation is needed due to nitric oxide–mediated venodilation and decreased Pms, positive inotropic agents such as dobutamine treat the myocardial depression, and arterial vasoconstrictors such as norepinephrine may be needed to maintain BP despite high Q̇T due to very low SVR.
• In cardiogenic shock, preload reduction (morphine, nitroglycerin, furosemide) is effected by venodilation and decreased Pms, but VR often increases because cardiac function improves to decrease Pra; arterial dilating drugs often increase Q̇T from the injured LV, so BP may even increase despite impaired contractility and Q̇T without increasing myocardial O2 consumption when heart rate does not increase.
• Early airway control and continuous mechanical ventilation decrease oxygen consumption and prevent respiratory acidosis but may decrease VR further in hypovolemic patients by raising pleural pressure and Pra; in cardiogenic shock, continuous mechanical ventilation and PEEP have less effect on VR and may increase Q̇T by decreasing LV afterload.
• Cardiogenic and low-pressure pulmonary edema are decreased by decreasing Ppw, and Q̇T and oxygen delivery can be maintained at low Ppw with vasoactive drugs and blood transfusion; arterial oxygenation can be supported with PEEP without decreasing Q̇T and oxygen delivery by affecting the least PEEP and tidal volumes that achieve 90% O2 saturation of an adequate hematocrit on a nontoxic fraction of inspired O2 without profound acidosis.

This chapter reviews several essential concepts of normal cardiovascular function as a basis for approaching and correcting disturbed circulation in critical illness. It begins with a discussion of left ventricular (LV) pumping function and an approach to ventricular dysfunction. Then follows a review of the mechanisms by which the venous return (VR) to the heart is controlled by the systemic vessels as a basis for diagnosis and treatment of hypoperfusion states. The pulmonary circulation and factors governing lung liquid flux are described through measurements obtained by right heart catheterization to provide an approach to treating pulmonary edema without compromising adequate peripheral perfusion. Along this discussion pathway, common mechanical interactions between respiration and circulation are highlighted as a basis for understanding the cardiovascular diseases discussed in the following chapters in this section and in the next section on pulmonary disorders in critical illness.

A primary role of the cardiovascular system is to deliver energy sources from the gut and liver and oxygen from the lungs to all systemic organ systems for their aerobic metabolism; effluent from these tissues removes the waste products of metabolism and delivers them to the lungs, kidney, and liver for excretion. This process is facilitated by return of the entire circulation through the lungs, where CO2 is eliminated and O2 is taken up to arterialize the blood. As depicted in Fig. 20-1, this central circulation is located within the thoracic cavity; movement of gas between the atmosphere and the alveolar space is done by the respiratory muscles, especially the diaphragm, depicted as a piston at the floor of the thoracic cavity. Beyond effecting ventilation to permit pulmonary gas exchange, spontaneous movement of the piston decreases the pleural pressure (Ppl), which approximates the pressure on the outside of extraalveolar vessels including the right and left heart (depicted as chambers labeled Pra [right atrial pressure] and Pla [left atrial pressure]); changes in alveolar pressure (PA) affect pressures within alveolar vessels. Once the blood leaves the lung and enters the left heart, the ventricular pumping function ejects blood into the stiff, high-resistance arterial circulation to perfuse the systemic capillary beds, where O2 is consumed and CO2 is taken up before the venous blood returns to the right heart through the large-volume, very compliant, low-resistance venous circuit.1,2

Ventricular–Vascular Coupling

Figure 20-2 illustrates the typical events in the two phases of ventricular activity: active contraction (systole) and relaxation (diastole). In diastole, the left ventricle fills through the open mitral valve from the left atrium while the aortic valve is closed. After electrical stimulation and contraction of the left ventricle in systole, a stroke volume (SV) is ejected into the proximal arterial chamber. Because more blood is being ejected than runs off through the peripheral resistance located in the distal arterioles, the arterial walls are distended outward, thus raising the pressure (P) in inverse proportion to the capacitance (C = ΔV/ΔP) of the walls of the larger arteries proximal to the resistance vessels. As the ventricle's volume (V) decreases, its ability to generate pressure decreases, as dictated by the length-dependent activation of actin and myosin cross-bridges in the cardiac myocytes,3 until ventricular systolic pressure (SP) falls below the simultaneous arterial pressure. Then the aortic valve closes, as indicated by the dicrotic notch on the arterial pressure pulse (Fig. 20-3). As the ventricle relaxes, the mitral valve opens, and the ventricle fills along its diastolic volume-pressure (V-P) curve; this diastolic filling is aided by atrial contraction and by suction by the ventricle relaxing from its low end-contraction volume. During diastole, the part of the SV stored in the distended arterial bed continues to run off through the peripheral resistance, associated with a progressive decrease in arterial pressure until the next contraction.

This ventricular–vascular coupling acts as a hydraulic filter to convert the intermittent ejection of an SV into a continuous organ flow, the cardiac output (Q̇T). It also decreases the work of the heart by allowing some of the energy imparted by the systolic ventricle to the blood to be stored in distention of the arterial chamber and returned to the circulation by continued flow while the ventricle is resting in diastole. The systemic vascular resistance (SVR) is also essential to the control of blood pressure (BP = Q̇T × SVR) and to the distribution of blood flow among, or within, organs according to tissue needs. For example, in conditions of hypovolemic or cardiogenic shock, reflex sympathetic output constricts arterioles, especially in the mesenteric, renal, and skin vascular beds, to preserve flow to vital organs such as the heart and brain by maintaining aortic BP; alternatively, during pyrexia and tachypnea, the increased metabolic need demands a high Q̇T associated with dilation of the coronary and respiratory muscle vessels induced by accumulation of anaerobic metabolites (e.g., adenosine, H+, and K+) to maintain adequate O2 flow to these vital pumps.

Clinical evaluation of the cardiovascular system in patients with critical illness is much aided by interpretation of the diastolic pressure (DP), the pulse pressure (PP), and indices of SVR such as the rate of color return to the nail bed after releasing pressure on the fingernail and digital temperature. For example, a hypotensive patient with a heart rate (HR) of 110 beats/min, an SP/DP of 100/40 mm Hg, and warm extremities with good color return to the nail bed has a high Q̇T and a low SVR (see Fig. 20-3). This is because the large PP (60 mm Hg) signals a large SV, which, when multiplied by the increased HR, produces increased Q̇T, and the low DP indicates rapid peripheral runoff through low SVR confirmed by digital examination and low mean BP (60 mm Hg) in the face of the high Q̇T. In contrast, a second hypotensive patient with the same HR and mean BP but an SP/DP of 80/65 mm Hg and cold extremities with very slow return of color to the nail bed has a low Q̇T with increased SVR indicated by the small PP (hence, low SV and Q̇T), preserved DP, and constricted digital vessels. As indicated in the lower left panel of Fig. 20-3, this low Q̇T hypotension is cardiogenic when the central venous pressure (CVP) is high and hypovolemic when the CVP is low. Of course, the relation between PP and SV is not quantitative, because it is proportioned by an unknown constant—the vascular capacitance. Nevertheless, in a given critically ill patient whose vascular capacitance changes minimally in a course of acute interventions, a change in PP is the earliest indicator of a change in SV.

The Starling Curve of the Heart

Figure 20-4 presents Starling relations of the heart.2,3 On the abscissa is plotted Pla, which approximates the filling pressure of the left ventricle. On the ordinate is plotted SV (in milliliters); this volume ejected per heartbeat is one measure of ventricular output. Another measure of output can be expressed by multiplying stroke volume by the pressure developed during each beat to obtain stroke work (SV × [BP − Pla]). As filling pressure of the heart increases, the ejected volume and the work done by the heart increase in a curvilinear manner; at higher filling pressures, there is less increase in SV per increase in Pla than at lower values of Pla. On the continuous Starling curve shown in Fig. 20-4B, the normal Pla (10 mm Hg) is associated with a normal SV (75 mL) calculated from Q̇T (6.0 L/min) divided by HR (80 beats/min). When hypovolemia decreases the Pla to 5 mm Hg, SV decreases to 40 mL, thus decreasing Q̇T; if therapeutic expansion of the circulating volume increases Pla to 20 mm Hg, SV increases to 100 mL, thus increasing Q̇T above normal. These relations comprise a common framework for understanding ventricular function in critical illness. A shift up and to the left of the Starling curve generally indicates enhanced ventricular function, with greater SV for a given filling pressure (see interrupted curve in Fig. 20-4). Conversely, a shift down and to the right with reduced SV at a given filling pressure (see dotted Starling curve through points A and B) indicates depressed ventricular function. The lower SV for a Pla of 10 mm Hg at point A could be due to reduced contractility, to increased BP allowing the same stroke work to eject a smaller SV at point A, or to a stiffer ventricle allowing a smaller LV end-diastolic volume (LVEDV) at a Pla of 10 mm Hg. This variety of mechanisms to explain the same data is a limitation of the analysis of hemodynamics by the Starling curve, so a more complete description of ventricular function is helpful.

The ventricular function shown by the Starling relation is based on the mechanical properties of the relaxed (diastolic) or contracting (systolic) V-P relations of the ventricle. This section reviews the factors that influence the diastolic and systolic mechanics in health and in critical illness and relates these mechanics to the corresponding Starling function curves of the heart. In Fig. 20-4B, LV end-diastolic pressure (LVEDP) and LV end-systolic pressure (LVESP) are plotted against the corresponding volumes (LVEDV and LV end-systolic volume [LVESV]). The continuous end-diastolic V-P curve is marked with a dot at an LVEDP of 10 mm Hg, where the normal LVEDV is 120 mL. When the ventricle contracts, pressure increases at the same volume until the aortic valve is opened, and blood is ejected until the valve closes at an LVESV of 45 mL. The SV (LVEDV − LVESV) is 75 mL, as plotted on the continuous Starling curve shown in Fig. 20-4A. When hypovolemia decreases the LVEDV (1 in Fig. 20-4B), LVEDP and SV decrease along the Starling curve above; if volume expansion increases the LVEDV to position 2, SV and LVEDP increase along the Starling curve. Intracardiac pressures such as Pla are measured with respect to atmospheric pressure, so they do not represent true transmural, or filling, pressures of the heart chamber when the pressure on the outside of the heart is not atmospheric.4,5 Pericardial pressure is most often equal to Ppl, which is subatmospheric during spontaneous breathing (−3 to −10 mm Hg) and can become very negative in airflow obstruction or very positive with mechanical ventilation and positive end-expiratory pressure (PEEP). For convenience, the following discussion refers to the intravascular pressures as transmural, or filling, pressures, and any cause for altered pericardial or pleural pressure is noted.

The Diastolic V-P Curve and Ventricular Filling Disorders

Figure 20-4B plots LVEDV against LVEDP. As ventricular volume increases from zero, the transmural pressure of the ventricle does not exceed zero until about 50 mL (the unstressed volume) is added. Then LVEDP increases in a curvilinear manner with ventricular volume (the stressed volume) first as a large change in volume for a small change in pressure and then as a small change in volume for a large change in pressure. If the pericardium is removed, these V-P characteristics are more linear such that the large change in LVEDP at higher values of LVEDV is no longer evident. Thus the pericardium acts like a membrane with a large unstressed volume loosely surrounding the heart up to a given ventricular volume, but at greater LVEDV the pericardium becomes very stiff. At higher heart volumes, most of the pressure across the heart is across the pericardium, accounting for the very steep rise in the diastolic V-P relation. In the presence of pericardial effusion, the volume at which the pericardium becomes a limiting membrane is reduced by the volume of the effusion. When the effusion is large enough, reduced end-diastolic volumes are associated with quite large end-DPs (see Chap. 28). In turn, pericardial pressure decreased VR by increasing Pra, thus keeping end-diastolic volume and Q̇T abnormally low. Tension pneumothorax, massive pleural effusions, high levels of PEEP, and greatly increased abdominal pressures can increase pressure outside the heart (Ppl) and thus reduce LVEDV and SV despite high values of LVEDP (Table 20-1). Intercurrent LV hypertrophy or infiltrative diseases (amyloidosis) occasionally stiffen the relaxed ventricle such that high filling pressures are needed to maintain an adequate SV, and inadequate filling time or poorly coordinated atrial contraction also impairs ventricular filling.6

A right-to-left shift of the interventricular septum also can restrict diastolic filling. Presumably, the distention of the right ventricle causes the interventricular septum to bulge from right to left, thereby reducing the unstressed volume and compliance of the left ventricle.2,7 This effect of ventricular interdependence is much less marked when the pericardium is removed, perhaps because the limiting membrane of the pericardium restricts freedom of motion of the left ventricle, making it more vulnerable to displacements of the septum. Accordingly, conditions in which the right ventricle is abnormally loaded (e.g., acute pulmonary embolism or acute-on-chronic respiratory failure due to obstructive or restrictive lung disease) may impede the emptying of the right ventricle, causing it to work at a higher end-diastolic volume. Then LV filling pressures will be higher than expected for the end-diastolic volume. This provides one possible explanation for why PEEP is often associated with increased filling pressure to maintain a normal SV even when LVEDP is corrected to the true filling pressure by subtracting the increase in Ppl (ΔPpl) measured when PEEP is applied.4,5,8,9 Acute myocardial ischemia also displaces the diastolic V-P curve of the left ventricle up and to the left (Fig. 20-5). Conceivably, myocardial injury alters the elastic properties of the relaxed ventricle.6 Therefore, a higher ventricular filling pressure is required at each end-diastolic volume. This accounts in part for the often noted observation that patients with acute myocardial injury need values of LVEDP as high as 30 mm Hg to maintain adequate Q̇T, whereas normal patients need filling pressures below 10 mm Hg.

The End-Systolic V-P Curve and Contractility

The normal diastolic V-P relation is demonstrated by the continuous line in the lower right-hand portion of Fig. 20-5. Consider the effects when the ventricle contracts during systole without ejecting any blood, as if the aortic valve could not open. A very large pressure is generated during this isovolumic contraction from a normal LVEDV, but when the LVEDV is reduced, the pressure generated during a similar isovolumic contraction is much less, as a manifestation of the force–length characteristics of the myocardium.2,3,10,11 That is, the less the muscle is stretched, the less force it can generate, a manifestation of the length-dependent activation of actin–myosin cross-bridges. The units of force in the hollow sphere of myocardium are the units of pressure, or force per unit area. A line connecting the end-systolic V-P points is linear and extrapolates toward the origin (see the continuous end-systolic V-P line in Fig. 20-5, upper left).

Of course, the aortic valve does open in early systole when the isovolumic pressure exceeds the aortic DP; then LV volume decreases as the SV is ejected (see Fig. 20-5). The contracting ventricle shortens against the aortic afterload pressure until its volume reaches the end-systolic volume; at that lower volume, the maximum pressure that can be generated is equal to the afterload pressure, so the aortic valve closes and ejection is over. If the afterload pressure were decreased, the ventricle could eject further to a lower end-systolic volume, where the maximum generated pressure equals the reduced afterload; hence, SV would increase.

The line connecting all end-systolic V-P points is an indicator of the pumping function or contractility of the heart because this line defines the volume to which the ventricle can shorten against each afterload for a given contractile state.10–13 Agents that enhance contractility (e.g., epinephrine, calcium, dobutamine, and dopamine) shift the end-systolic V-P relation up and to the left; then the ventricle can shorten to a smaller end-systolic volume for each afterload, thereby increasing SV at a given LVEDV/LVEDP.10,11 Conversely, negative inotropic agents such as propranolol, myocardial ischemia, hypoxia, and acidemia depress the end-systolic V-P relation down and to the right, as indicated by the interrupted end-systolic V-P line shown in Fig. 20-5.12,13 Then end-systolic volume is increased for a given pressure afterload, thereby reducing the SV at a given filling pressure. Such a reduction in contractility is a common cause for the depressed Starling curve AB shown in in Fig. 20-4.

An Approach to Acute Ventricular Dysfunction

These concepts provide a framework for understanding the pathophysiology and therapy of acute myocardial infarction (see Chap. 25). Figure 20-5 depicts normal diastolic and systolic V-P relations and indicates a normal systolic ejection (continuous lines). From an LVEDV of 120 mL and an LVEDP of 10 mm Hg, the ventricle contracts isovolumically until the aortic valve opens at a DP of 80 mm Hg. Blood is then ejected as SP increases to 110 mm Hg and decreased toward an LVESP of 90 mm Hg and an LVESV of 50 mL, when the aortic valve closes to generate the dicrotic notch on the arterial pressure trace. Accordingly, SV is 70 mL at a Pla of 10 mm Hg and Q̇T is 5.6 L/min when HR is 80 beats/min. Acute myocardial infarction depresses the end-systolic V-P relation so that the end-systolic volume is increased to 90 mL at a reduced LVESP of 75 mm Hg (interrupted lines). At the same time, the end-diastolic volume is increased to 130 mL to accommodate the VR, and end-DP is increased even more than expected (LVEDP = 30 mm Hg) due to the shift up and to the left of the end-diastolic V-P relation (interrupted line). Thus SV (40 mL) and Q̇T (4.4 L/min) are reduced despite reflex tachycardia (HR = 110 beats/min) at an increased LV filling pressure, and BP is decreased (SP/DP = 90/70 mm Hg) despite the reflex increase in SVR.

Conventional therapy consists of preload reduction, inotropic agents, and afterload reduction, in addition to measures to reestablish and maintain coronary blood flow (see Chap. 25). Interventions such as morphine, furosemide, nitrates, and rotating tourniquets decrease VR by dilating venous capacitance beds to increase the unstressed volume and decrease mean systemic pressure (Pms; see below). These actions in turn decrease LVEDV and LVEDP. The reduction in end-diastolic volume tends to decrease LVEDP along the steep diastolic V-P curve so that there is a large reduction in LVEDP for a small reduction in LVEDV and SV. Further, the potential adverse effect of reduced SV is often offset by increased contractility and reduced afterload when myocardial wall stress is decreased by the reduction in LVEDV and LVEDP (see Chaps. 22 and 25). Reduced afterload and reduced myocardial O2 consumption improve ventricular pumping function by shifting the end-systolic V-P relation up and to the left, so LVESV decreases and SV increases. Further, the decreased end-DP reduces the complication of cardiogenic pulmonary edema.

Positive inotropic agents such as dopamine and dobutamine act directly on the myocardium to reduce end-systolic volume at a given end-SP, thereby increasing SV (see Chaps. 22 and 25). Dopamine also causes venoconstriction by increasing Pms and VR, so the increase in SV is often associated with increased LVEDV, whereas dobutamine tends to increase SV and decrease LVEDV.14 Afterload-reducing agents such as nitroprusside dilate peripheral arteries to decrease end-SP and afterload; in turn, end-systolic volume decreases along the depressed end-systolic V-P relation to increase SV.14 Nitroprusside and other arteriolar vasodilating agents also decrease end-DP without changing end-diastolic volume; this effect appears to enhance ventricular function viewed on the Starling relation, as discussed above.15 The decrease in LVEDP decreases pulmonary edema and may decrease myocardial oxygen demands by decreasing ventricular wall stress. To the extent that it decreases ventricular wall stress, end-systolic V-P relations may shift to the left due to enhanced contractility. In some patients with cardiogenic shock, vasodilator therapy appears to increase SV and Q̇T without decreasing or even increasing arterial BP, that is, arterial dilation appears to reduce end-systolic volume at a given end-SP as if contractility were enhanced.14,15

Noninvasive mechanical ventilation (or continuous positive airway pressure, CPAP) lowers both preload and afterload, in addition to beneficial effects on gas exchange and work of breathing.16 Q̇T is not reduced despite decreased transmural pressures, indicating improved pump function.

Other concomitant effects of critical illness cause ventricular dysfunction characterized by reduced SV at increased Pla (see Table 20-1). Arterial hypoxemia12 and acidemia13 depress the end-systolic V-P curve and increase diastolic stiffness, as shown by the interrupted curve in Fig. 20-5. Acute arterial hypertension raises the pressure afterload, so SV decreases as end-systolic volume increases along the continuous end-systolic V-P curve in Fig. 20-5. Then LVEDV increases to accommodate VR, so LVEDP increases, often more than expected, due to diastolic stiffness, in turn due to LV hypertrophy in the hypertensive patient. Accordingly, pulmonary edema is a common complication, and it responds to vasodilator therapy when BP is decreased. In some or all of these conditions, diastolic dysfunction merits special management.6 When acute or acute-on-chronic congestive heart failure is present, decreasing LVEDP and LVEDV, maintaining atrial contraction, increasing the duration of diastole, and minimizing myocardial ischemia are helpful. Each of these therapeutic measures is also helpful in managing hypoperfusion states associated with diastolic dysfunction.6

Valvular dysfunction mimics systolic and diastolic dysfunctions such that LVEDV is much increased and the forward SV is reduced (see Chap. 29). With aortic regurgitation, after a vigorous systolic ejection, aortic blood runs off forward and backward in diastole such that LVEDP increases and arterial DP decreases toward equal values at 40 mm Hg. The large LVEDV then ejects a large SV to increase SP to 120 mm Hg, causing a bounding PP of 80 mm Hg, but the aortic regurgitation reduces forward SV and Q̇T to a low value. Consider also mitral valve incompetence. During systole, a large fraction of the blood ejected from the ventricle regurgitates to the left atrium, thereby reducing forward SV and Q̇T but increasing LVEDV and LVEDP when the left atrium fills the ventricle in diastole. In this circumstance, PP and BP are decreased. In both cases, the ventricular mechanics resemble the interrupted curves shown in Fig. 20-5 and improve toward the normal continuous curves when forward flow is increased by vasodilator therapy that lowers SVR to allow more peripheral runoff and less regurgitant flow (see Chap. 29).

Ventricular pumping function generates the flow in the circulation but retains control of the VR, so Q̇T resides in the systemic vessels.1,2 The heart is best regarded as a mechanical pump having diastolic and systolic properties that determine how it accommodates the VR. This section reviews the mechanical characteristics of the systemic vessels as a basis for understanding control of Q̇T in health and in critical illness.

Mean Systemic Pressure

When the heart stops beating (see Fig. 20-1), pressure equalizes throughout the vascular system, and its new value is the Pms (10 to 15 mm Hg). This pressure is much lower than the arterial pressure and is closer to the Pra. When flow stops, blood drains from the high-pressure, low-volume arterial system into the high-volume, low-pressure venous system, which accommodates the displaced volume with little change in pressure. When the heart begins to beat again, the left heart pumps blood from the central circulation into the systemic circuit, thus increasing pressure there. At the same time, the right heart pumps blood into the lungs, thereby decreasing its pressure (Pra) with respect to Pms, so blood flows from the venous reservoir into the right atrium. Pressure on the venous side decreases slightly below Pms, whereas pressure on the arterial side increases considerably above Pms with succeeding heartbeats. This continues until a steady state is reached, when arterial pressure has increased enough to drive the whole SV of each succeeding heartbeat through the high arterial resistance into the venous reservoir. The Pms does not change between the state of no flow and the new state of steady flow because neither the vascular volume nor the compliance of the vessels has changed. What has changed is the distribution of the vascular volume from the compliant veins to the stiff arteries; this volume shift creates the pressure difference driving flow through the circuit.1,2,17,18

Pms is the driving pressure for VR to the right atrium when circulation resumes. It can be increased to increase VR by increasing the vascular volume or by decreasing the unstressed volume and compliance of the vessels.17,18 The latter two mechanisms are mediated by baroreceptor reflexes responding to hypotension by increasing venous tone and usually occur together. The unstressed volume also may be reduced by raising the legs of a supine patient or applying military antishock trousers; both methods return a portion of the unstressed vascular volume from the large veins in the legs to the stressed volume, thereby increasing Pms and VR. When the heart has an improvement in inotropic state or a reduction in afterload, blood is shifted from the central compartment to the stressed volume of the systemic circuit, thereby increasing Pms and VR;18 moreover, improved ventricular pumping function decreases Pra to increase VR further (see below).

Venous Return and Cardiac Function Curves

Before the heart was started in the discussion above, Pra was equal to the pressure throughout the vascular system, Pms. With each succeeding heartbeat, Pra decreases below Pms and VR increases. This sequence is repeated in a more controlled, steady state by replacing the heart with a pump set to keep Pra at a given value while VR is measured.17,18 Typical data are plotted in Fig. 20-6. As Pra is decreased from 12 to 0 mm Hg (indicated by the thin continuous line), VR is progressively increased with the driving pressure (Pms − Pra). The slope of the relation between VR and Pms − Pra is the resistance to VR (RVR = Δ[Pms − Pra]/ΔVR). When Pra falls below zero, VR does not increase further because flow becomes limited while entering the thorax. This occurs when the pressure in these collapsible great veins decreases below the atmospheric pressure outside the veins. Further decreases in Pra and CVP are associated with progressive collapse of the veins rather than with an increase in VR.

For a given stressed vascular volume and compliance, Pms is set and RVR is relatively constant. In the absence of pulmonary hypertension or right heart dysfunction, LV function will determine Pra and, hence, VR to the right heart, which must equal the Q̇T from the left heart. Q̇T is described by the cardiac function curve, drawn as a thick continuous line relating Pra (abscissa) to Q̇T (ordinate), in Fig. 20-6. The heart is able to eject a larger SV and Q̇T when the end-DP is greater because more distended ventricles eject to about the same end-systolic volume as less distended ventricles do. Accordingly, as Pra decreases, Q̇T decreases along the cardiac function curve. However, VR increases as Pra decreases until VR equals Q̇T at a unique value of Pra, indicated by the intersection of the cardiac function and VR curves in Fig. 20-6 (see point A in both panels).

When Q̇T is insufficient, VR can be increased in several ways. A new steady state of increased VR is achieved by increasing Pms with no change in RVR, indicated by the interrupted VR curve in the left panel of Fig. 20-6. This new VR curve intersects the same cardiac function curve at a higher value of Q̇T at point B. This method of increasing VR is associated with an increase in Pra. Due to the steep slope of the cardiac function curve in normal hearts, large increases in VR occur with only small increases in Pra. Alternatively, VR can be increased by enhanced cardiac function by increasing contractility or decreasing afterload of the heart. This is depicted as an upward shift of the cardiac function curve, as in the right panel of Fig. 20-6, such that greater Q̇T occurs at each Pra. The increase on each VR curve by this mechanism is associated with a reduction in Pra. Further, in the normal heart, only a small change in VR is possible (from point A to point B in the right panel), and greater reductions in Pra do not increase Q̇T further because VR becomes flow limited as Pra decreases to below zero. This explains why inotropic agents that enhance contractility are ineffective in hypovolemic shock.

When cardiac pumping function is depressed, as depicted by the interrupted line in Fig. 20-7, VR is decreased from point A to point B for the same value of Pms as Pra increases. The patient must then retain fluid or initiate cardiac reflexes to increase Pms toward the new value required to maintain adequate Q̇T, as in chronic congestive heart failure. This is associated with a large increase in Pra from point B to point C, which in turn causes jugular venous distention, hepatomegaly, and peripheral edema. Diuretic reduction of vascular volumes will correct these cosmetic abnormalities at the expense of decreasing Pms and VR. In contrast, inotropic and vasodilator drugs, which improve depressed cardiac function by shifting the interrupted cardiac function curve upward, increase Q̇T and decrease Pra more effectively than in patients with normal cardiac function.

Effects of Pressure Outside the Heart on Cardiac Output

In the figures cited and the preceding discussions, values of Pms and Pra were expressed relative to atmospheric pressure. However, the transmural pressure of the right atrium exceeds the Pra by the subatmospheric value (about −4 mm Hg) of the Ppl surrounding the heart. Consider the effect of opening the thorax, which raises Ppl from −4 to 0 mm Hg: VR decreases from point A to point B in Fig. 20-8 because Pra increases.19 This is indicated by the interrupted cardiac function curve shifted to the right by the increase in pressure outside the heart but parallel to the normal cardiac function curve (continuous line through point A). Normal VR can be restored (point B to point C) by increasing Pms by an amount equal to the increase in Ppl and Pra induced by thoracotomy. Then transmural Pra will be the same as at point A, and Pra will have increased from point A to point C at the same Q̇T.

This mechanism for the decrease in Q̇T with thoracotomy also partly explains the decrease in Q̇T with PEEP. The Ppl within an intact thorax increases with passive positive-pressure ventilation, thereby increasing Pra and decreasing VR.4,5,8,9,20 When 8 mm Hg of PEEP (10 cm H2O) is added to the ventilator, the end-expiratory value of Ppl increases by about half that amount, e.g., from −4 to 0 mm Hg. Accordingly, VR decreases with PEEP from point A to point B in Fig. 20-8, with no change in cardiac function or Pms. Q̇T is returned to normal by volume infusion or vascular reflexes that increase Pms by an amount equal to the increases in Ppl and Pra. Greater PEEP (20 cm H2O, as in the dotted line shown in Fig. 20-8) decreases VR further (from point A to point D) and requires greater increases in Pms to return it to normal (from point D to point E). Alternatively, Pms increases as much as Pra when PEEP is added, so the decrease in VR must be due to an increase in RVR with PEEP.20 In either event, VR can be restored on PEEP by increasing Pms.

Q̇T is much less susceptible to the deleterious effects of PEEP when Pms is high. In patients with reduced circulatory volume, vascular reflexes are already operating to maintain VR and Pms by reducing unstressed volume or vascular compliance. Such patients have little vascular reflex reserve and poorly tolerate intubation and positive-pressure ventilation without considerable intravenous infusion to increase vascular stressed volume. In contrast, well-hydrated or overhydrated patients may tolerate even large amounts of PEEP with no reduction in Q̇T because their previously inactive vascular reflexes can increase Pms in well-filled systemic vessels by the amount that Ppl increases with PEEP. These considerations allow the physician to anticipate and treat the hypotension induced by ventilator therapy; the concept should not be interpreted as an indication for maintaining high circulatory volume in critically ill patients on ventilators because this often increases lung edema and provides even more Q̇T than was already deemed sufficient. Further, pressure outside the heart can be increased by a variety of other concomitant conditions and complications of critical illness; all these actions increase pressures measured in the heart chambers and decrease heart volume and, as a consequence, are often interpreted as diastolic dysfunction (see Table 20-1).

How much is the pressure outside the heart increased by PEEP, and is there a practical approach to relating the transmural wedge pressure to SV and Q̇T? When PEEP increases end-expired lung volume, the inflated lungs push the thorax to an increased volume through greater pleural pressure, and this change in Ppl (ΔPpl) with PEEP is approximately equal to the change in pressure outside the right and left ventricles.8 During mechanical ventilation, the ratio of ΔPpl to the change in static elastic pressure across the lung and chest wall (ΔPel) for each breath is given by the ratio of respiratory system compliance (Crs) to the compliance of the chest wall (Cw); that is, ΔPpl/ΔPel = Crs/Cw (assuming no alveolar recruitment). When lung compliance (CL) is normal, CL = Cw, so ΔPpl/ΔPel = 0.5. When the lungs lose compliance in acute hypoxemic respiratory failure (AHRF), ΔPel increases because Crs decreases, but ΔPpl changes little (at constant tidal volume) because Cw is unaffected by the lung disease, and ΔPpl becomes much less than half of ΔPel. To the extent that the increase in lung volume (ΔVL) with PEEP is determined by Crs, ΔPpl/PEEP = Crs/Cw, and a decrease in Crs with AHRF would decrease Ppl for a given amount of PEEP well below the normal value of 0.5. However, ΔVL with PEEP is much greater than that predicted by Crs in AHRF because PEEP recruits many previously flooded airspaces,21,22 so ΔPpl/PEEP is as great after acute lung injury as before4 (Fig. 20-9). Accordingly, the ΔPpl with PEEP is difficult to measure and hard to predict, so many approaches have been tested to estimate the transmural pressure of heart chambers on PEEP.23 Because PEEP is used most often to decrease shunt in pulmonary edema and because accurate knowledge of transmural Pla shows that the value associated with an adequate Q̇T can differ between patients by 20 mm Hg according to the extent of LV dysfunction, a better approach is to seek the lowest pulmonary wedge pressure (Ppw) that provides adequate output on each level of PEEP. In this way, therapy to decrease Ppw and edema and maintain Q̇T is not confounded by erroneous estimates of transmural Ppw on PEEP.24,25

An Approach to Hypoperfusion States

A hypoperfusion state, or shock, is almost always signaled by systemic hypotension; commonly associated clinical features of multiple organ system hypoperfusion are tachycardia, tachypnea, prerenal oliguria (urine flow < 20 mL/h, urine Na+ > 20 mEq/L, urine K+ > 20 mEq/L, urine-specific gravity > 1.020), abnormalities of mentation and consciousness, and metabolic acidosis. The mean BP is determined by the product of Q̇T and SVR. A conceptual framework for the initial diagnosis and management of the hypotensive patient is outlined in Table 20-2. Utilization of this approach aims to categorize the patient's symptoms into one of the three common causes of shock (septic, cardiogenic, or hypovolemic) and to initiate early appropriate therapy of the presumed diagnosis (see Chap. 21). Response to the therapeutic intervention tests the accuracy of the initial diagnosis, so the hemodynamic response is reevaluated within 30 minutes. The diagnostic decision is aided by collating clinical data from the medical history, physical examination, and routine laboratory tests to answer three questions in sequence.

Septic Shock

Is BP decreased because Q̇T is decreased? If not, SVR must be reduced, a condition almost always related to sepsis or sterile endotoxemia associated with severe liver disease. As indicated in Table 20-2 (right column), a low BP is often characterized by a large PP because the SV is large and by a very low DP because each SV has a rapid peripheral runoff through dilated peripheral arterioles (see Fig. 20-3). This produces warm, pink skin with rapid return of color to the nail bed and crisp heart sounds. As in other types of shock, tachycardia is evident due in part to baroreceptor reflex response to hypotension, but the arterial vasoconstriction response to reflex sympathetic tone is blocked by relaxation of arteriolar smooth muscle induced by endothelium-derived relaxing factor (or nitric oxide). The combination of tachycardia and large PP indicates a large Q̇T that is almost always present early unless concurrent hypovolemia or myocardial dysfunction precludes the hyperdynamic circulatory state of sepsis.

Initial therapy starts with appropriate broad-spectrum antibiotics (see Chap. 46) and expands the circulating volume by intravenous infusion of fluids to treat associated hypovolemia, which is due to venodilation decreasing Pms and VR lower than needed to maintain adequate perfusion pressure of vital organs. The end point of volume infusion is obscure because Q̇T and oxygen delivery (DO2) are already increased, and although Q̇T usually increases further with intravenous infusions, BP increases little with increased Q̇T. Further, the need for an even greater Q̇T to increase DO2 is questionable because the lactic acidosis of septic shock may not be due to anaerobic metabolism.26–28 Accordingly, septic patients in whom Q̇T is maximized do not have improved survival.29,30 Conversely, pulmonary vascular pressures always increase with volume infusion, thus increasing pulmonary edema when the septic process increases the permeability of lung vessels.25,31–33 This coincidence of the acute respiratory distress syndrome (ARDS) and septic shock has created an apparent dilemma concerning fluid therapy and cardiovascular management of these conditions. My approach is to ensure resuscitation from septic shock as the first priority by ensuring a large Q̇T with a Ppw that does not exceed 15 mmHg and add dobutamine to increase Q̇T and BP as necessary.25 When early ARDS is not associated with septic shock, I seek the lowest circulating volume to provide adequate Q̇T.25

The septic myocardium does not function normally,34,35 but this dysfunction is often associated with SV values larger than 100 mL at normal values of LVEDP. Accordingly, it seems unlikely that systolic dysfunction contributes substantially to the shock, but infusion of dobutamine does increase Q̇T for a given high-normal LVEDP without increasing O2 uptake or correcting lactic acidosis in septic shock.36 Even when Q̇T and DO2 are made adequate with fluid and dobutamine infusions, the perfusion pressure for vital organs such as the brain and heart may still be too low in some septic patients. In this case, norepinephrine infusion increases BP and splanchnic blood flow37,38 without compromising renal function;39 in contrast, dopamine and epinephrine infusions cause splanchnic hypoperfusion in septic shock.37,38,40 Tachypnea and respiratory distress may be severe, so initial supportive therapy includes consideration of early intubation and mechanical ventilation and correction of hyperthermia with antipyretics, paralysis, and cooling (see Table 21-4). This prevents catastrophic respiratory muscle fatigue, respiratory acidosis, and the complications of emergent intubation and may improve tissue oxygenation by reducing O2 requirements in patients with limited DO2.41,42

Cardiogenic Shock

In contrast to septic shock, low Q̇T is signaled by low PP indicating low SV (see Fig. 20-3), signs of increased systemic vascular resistance (e.g., cold, blue, damp extremities and poor return of color to the nail bed), and a history or presentation including features suggesting a cardiogenic or hypovolemic cause of hypotension. If Q̇T is reduced in the hypotensive patient, then the heart may be too full.

A heart that is too full (see Table 20-2) is often signaled by symptoms of ischemic heart disease or arrhythmia, signs of cardiomegaly, the third and fourth sounds or gallop rhythm of heart failure, new murmurs of valvular dysfunction, increased jugular or CVP, and laboratory tests suggesting ischemia (e.g., electrocardiogram [ECG] or creatine phosphokinase determination) or ventricular dysfunction (e.g., chest x-ray suggesting cardiomegaly, a widened vascular pedicle, or cardiogenic edema or echocardiogram showing regional or global systolic dyskinesia). The most common cause of hypotension associated with a circulation that is too full on initial evaluation is cardiogenic shock due to myocardial ischemia (see Chaps. 22 and 25). Initial therapy treats this presumptive diagnosis with inotropic drug therapy (dobutamine 3 to 10 μg/kg per minute) to assist the ejecting function of the ischemic heart. Such therapy does not directly address the coronary insufficiency and may increase the myocardial O2 demand, especially if it causes tachycardia. Concurrent sublingual, dermal, or intravenous nitroglycerin ameliorates elements of coronary vasospasm to increase blood flow and reduces preload to decrease myocardial O2 consumption. Morphine also decreases pain, anxiety, and preload.43

In this situation, even a cautious volume challenge (250 mL 0.9% NaCl over 20 minutes) may be risky because ventricular function and Q̇T are decreased as often as they are increased by this intervention, and the risk of pulmonary edema is increased. When signs of pulmonary edema are present on clinical and radiologic examinations of the thorax, diuretics, morphine, and nitroglycerin often reduce preload by relaxing the capacitance veins, associated with an increase in LV systolic performance. However, about 10% of patients with myocardial ischemia present with significant hypovolemia. Accordingly, the clinical assessment of hemodynamics should be supplemented as soon as possible with other means to exclude hypovolemia (e.g., right heart catheterization, empiric volume challenge, echocardiography, or dynamic tests of the adequacy of circulating volume) so that appropriate volume infusion or reduction can be titrated. When these measures are addressed adequately but the hypoperfusion state persists, early movement toward arteriolar vasodilator therapy or a balloon-assist device is indicated to reduce LV afterload and preserve coronary perfusion pressure (see Chap. 25). These latter interventions are not relegated to the last resort but are considered early in this initial stabilization of cardiogenic shock. Similarly, early elective intubation and mechanical ventilation allow effective sedation and reduce O2 consumption,41 and PEEP improves arterial oxygenation, often without reducing VR and with improvement of pumping function in the damaged left ventricle by reducing preload and afterload.44

Hypovolemic Shock

Beyond the absence of clinical features suggesting that the heart is too full in the hypotensive patient who is presenting reduced Q̇T (see Table 20-2), hypovolemic shock is distinguished from cardiogenic shock by several positive clinical features. Often there is an obvious source of external bleeding (e.g., multiple trauma, hemoptysis, hematemesis, hematochezia, or melena); internal bleeding is often signaled by blood aspirated from the nasogastric tube or on rectal examination, by increasing abdominal girth, or by clinical and radiologic examinations of the thoracic cavity for pleural, alveolar, retroperitoneal, or periaortic blood. Each of these signals is often associated with a new reduction in the hematocrit. Nonhemorrhagic hypovolemia often presents with recognizable excess gastrointestinal fluid losses (e.g., vomiting, diarrhea, suctioning, and stomas), excess renal losses (e.g., osmotic or drug diuresis and diabetes insipidus), or third-space losses as in extensive burns. Physical examination shows dry mucous membranes with decreased tissue turgor, and routine laboratory tests often show increased serum urea nitrogen out of proportion to a relatively normal creatinine level and increased hematocrit due to hemoconcentration.

The initial management of patients with presumed hypovolemic shock necessitates early vascular access with two large-bore (14-gauge) peripheral intravenous catheters for rapid infusion of large volumes of warmed blood and fluids for hemorrhagic shock and the appropriate crystalloid solution for dehydration. Central venous access ensures adequate volume resuscitation and allows early measurement of CVP. An immediate response of increased BP and pulse volume supports the presumed diagnosis, whereas no improvement in these hemodynamic measurements necessitates emergent repair of the site of blood loss or a reevaluation of the working diagnosis. Achieving hemostasis in hemorrhagic shock is a prerequisite for adequate volume resuscitation: urgent and simultaneous pursuit of hemostasis and fluid resuscitation is encouraged.45 Vasoconstricting drugs such as norepinephrine should be used only as short-term antihypotensives to mobilize endogenous unstressed volume or enhance arteriolar vasoconstriction until the circulating volume is restored by transfusion; prolonged use of these drugs confounds the physician's assessment of the end point of volume resuscitation. Early endotracheal intubation and mechanical ventilation reduce the patient's work of breathing and allow respiratory compensation for lactic acidosis during volume resuscitation; warming the fluids and covering the patient with warm dry blankets prevent the complication of hypothermia, including cold coagulopathy and further bleeding (see Table 21-4).

Other Common Causes of Shock: A Short Differential Diagnosis

The purpose of this initial schema is to formulate a working diagnosis for the most common presentations of shock so that early and rapid therapy can be initiated. The response to the initial therapy confirms or challenges the working diagnosis. When features of the initial clinical presentation or the response of the patient to appropriate management challenges the working diagnosis, early acquisition of more objective hemodynamic data is appropriate. In the interim, other features of the clinical presentation often suggest a cause of shock that falls outside this simplistic schema, or the possibility of overlapping or concurrent causes expands. This section briefly reviews several important differential diagnostic conditions for cardiogenic shock (e.g., tamponade or acute right heart syndromes) and hypovolemic shock (e.g., anaphylactic, neurogenic, or adrenal shock; (see Table 20-2, what does not fit?).

Cardiac Tamponade

Pericardial effusion is often suggested early by the clinical setting (e.g., renal failure, malignancy, or chest pain), physical examination (e.g., elevated neck veins, systolic BP that decreases >10 mm Hg on inspiration, or distant heart sounds), or routine investigations (e.g., chest radiograph with “water bottle” heart, low voltage on the ECG, or electrical alternans). Such a constellation of clinical data requires early echocardiographic confirmation of pericardial effusion, and tamponade is signaled by right ventricular and right atrial collapse that worsens with inspiration, with a relatively small left ventricle (see Chap. 28). Tamponade requires urgent pericardiocentesis or operative drainage by pericardiostomy. While deciding on definitive treatment, one should remember that intravenous expansion of the circulating volume may produce small increases in BP, whereas reductions in circulating volume (e.g., diuretics, nitroglycerin, morphine, or intercurrent hemodialysis) are often associated with catastrophic reduction in Q̇T by reducing the venous tone and volume necessary to maintain the Pms required to drive VR back to high Pra.

Right heart catheterization typically shows a Pra increased to about 16 to 20 mm Hg and equal to pulmonary arterial DP and the arterial Pwp; Q̇T and SV are much reduced (see Chap. 28). This hemodynamic subset resembles that of cardiogenic shock (high Ppw and low SV). However, in the case of pericardial tamponade, Ppw is increased because pericardial pressure is increased, so the transmural pressure of the left ventricle approaches zero, a value consistent with the very low LVEDV accounting for the low SV. Other etiologies of hypotension associated with high cardiac pressures and small ventricular volumes include constrictive pericarditis, tension pneumothorax, massive pleural effusion, positive-pressure ventilation with high PEEP, and very high intraabdominal pressure (see Table 21-1). Up to 33% of patients presenting with cardiac tamponade have increased BP despite low Q̇T; this subset of patients has a high incidence of hypertension preceding the onset of tamponade.46

Right Ventricular Overload and Infarction

Another clinical presentation that may fall outside the simplest scheme presented in Table 20-2 is the hypotension associated with acute or acute-on-chronic pulmonary hypertension. Shock after acute pulmonary embolism is often signaled by the clinical setting including risk factors (e.g., perioperative, immobilized, thrombophilia, or prior pulmonary embolisms); symptoms of acute dyspnea, chest pain, or hemoptysis; physical examination showing a loud P2 with a widened and fixed split of the second heart sound; new hypoxemia without obvious radiologic explanation; and acute right heart strain on the ECG (see Chap. 27). Noninvasive Doppler studies of the veins in the lower extremities and helical computed tomographic angiography confirm the diagnosis. Anticoagulation or placement of a filter in the inferior vena cava reduces the incidence of subsequent emboli, and there may be some success with thrombolytic therapy (or, in some centers, surgical removal of the embolus) in patients with shock due to pulmonary embolism. Acute-on-chronic pulmonary hypertension causes shock in the setting of prior primary pulmonary hypertension, recurrent pulmonary emboli, progression of collagen vascular disease, or chronic respiratory failure (e.g., chronic obstructive pulmonary disease or pulmonary fibrosis) aggravated in part by hypoxic pulmonary vasoconstriction. In these circumstances, O2 therapy and pulmonary vasodilator therapy combine to decrease pulmonary hypertension and increase Q̇T in a small but significant proportion of patients (see Chap. 27).

Right heart catheterization shows a unique hemodynamic profile: a very high mean pulmonary artery pressure, pulmonary arterial DP considerably greater than the Pwp, and reduced Q̇T and SV. Not uncommonly, arterial Pwp is normal or increased despite a small LVEDV on echocardiographic examination, which also shows a right-to-left shift of the interventricular septum; presumably, this causes stiffening of the diastolic V-P curve of the left ventricle. A complication of pulmonary vasodilator therapy is hypotension due to systemic arterial dilation unaccompanied by increased right heart output. Such effects aggravate the hypoperfusion state, perhaps by reducing coronary blood flow to the hypertrophied, dilated right ventricle. Some evidence suggests that shock associated with pulmonary hypertension is ameliorated by α-agonist therapy (e.g., norepinephrine or phenylephrine), which acts as a predominant systemic arteriolar constrictor to increase BP sufficiently to maintain right ventricular perfusion.47,48

Right ventricular infarction causes low pulmonary artery pressures and normal LV filling pressures because the dilated, injured right ventricle is unable to maintain adequate flow to the left heart.49 Elevated neck veins and Pra tend to decrease with dobutamine infusion, perhaps because the enhanced contractility of the left ventricle improves systolic function of the mechanically interdependent right ventricle.45,46 Volume expansion often aggravates right ventricular dysfunction, and systemic vasoconstriction may preserve right ventricular perfusion.50

Anaphylactic, Neurogenic, and Adrenal Shock

Other etiologies of shock having unique clinical presentations that usually lead to early diagnosis are anaphylactic shock and neurogenic shock. Beyond identifying the etiology early through their association with triggering agents and trauma, respectively, the physician should note that the pathophysiology of each is a dilated venous bed with greatly increased unstressed volume of the circulation leading to hypovolemic shock. Accordingly, the mainstay of therapy for both conditions is adequate volume infusion; adjunctive therapy for anaphylaxis includes antihistamines, steroids, and epinephrine to antagonize the mediators released in the anaphylactic reaction (see Chap. 106), whereas a careful search for sources of blood loss and hemorrhagic shock is part of the early resuscitation of spinal shock in the traumatized patient (see Chap. 94).

Not uncommonly, the presentation of patients with nonhemorrhagic hypovolemic shock raises the concern of acute adrenal cortical insufficiency. When this possibility is not obviously excluded, it is appropriate to draw a serum cortisol level, provide adequate circulating steroids with dexamethasone, and conduct a corticotropin stimulation test to confirm or refute the diagnosis. Characteristically, hypotension and hypoperfusion in such patients will not respond to adequate vascular volume expansion until dexamethasone is administered (see Chap. 79).

Multiple Etiologies of Shock

With this differential diagnosis and management evaluation in mind, the initial approach to patients with hypoperfusion states should be completed in less time than it takes to read about it. The target is to distinguish among patients with septic shock, cardiogenic shock, and hypovolemic shock and to initiate an appropriate therapeutic challenge—antibiotics, inotropic agents, or a volume challenge—within 30 minutes of presentation. By the response, the diagnosis is confirmed or challenged, with special regard to equivocal responses to therapy or to several other diagnostic categories of shock. Sorting out the primary etiology of the hypoperfusion state often requires considerable additional data. This process is rendered more complex by concurrent etiologies contributing to the shock, for example, the patient with septic shock unable to increase Q̇T due to intercurrent myocardial dysfunction, the patient with acute myocardial infarction who is hypovolemic, or the patient with hemorrhagic shock who becomes septic. Other combinations of these major categories overlap with confounding effects of tamponade, positive-pressure ventilation, pneumothorax, and pulmonary hypertension—all to challenge ongoing diagnostic and management approaches.

Pressures, Flow, and Resistance in Pulmonary Vessels

Q̇T from the left heart is equal to VR to the right heart, so the entire Q̇T traverses the pulmonary circulation in pulsatile fashion (Fig. 20-10). The right ventricle ejects blood into the pulmonary artery, thereby increasing its pressure (Ppa) to drive flow through a branching arteriolar system into the lung parenchyma, where a network of very small alveolar septal vessels or capillaries passes between the airspaces of the lung to effect pulmonary gas exchange. These septal vessels converge into pulmonary veins that empty into the left atrium, where the pressure (Pla) is often regarded as the outflow pressure of the pulmonary circulation. When this pressure gradient across the pulmonary circulation (Ppa − Pla) is divided by the pulmonary blood flow (Q̇), the pulmonary vascular resistance is calculated (mm Hg/L per minute) and sometimes converted to metric units (dyn-s/cm5) by multiplying by 80. By this analysis, increasing blood flow from one level to another is associated with decreasing pressure across the pulmonary circulation (Ppa − Pla) along a unique pressure-flow relation given by the continuous line in Fig. 20-10B. Resistance to Q̇ may be increased by smooth muscle constriction within the pulmonary arterioles and alveolar vessels by hypoxia, by compression of the alveolar septal vessels by elevated PA, by obstruction of larger pulmonary vessels by thromboembolism, or by obliteration of many of the parallel vascular channels as they traverse the lung so that the same blood must flow through fewer channels. Such an increase in pulmonary vascular resistance would be calculated as at point A on the interrupted line in Fig. 20-10, where the pressure difference across the lung (Ppa − Pla) has increased for the same amount of Q̇. Pulmonary hypertension is a frequent abnormality in critical illness; its causes are listed in Table 22-4 and its treatment is discussed in Chaps. 22 and 26.

Figure 20-10 also depicts a common way to make these measurements with a pulmonary artery catheter (PAC) that is passed throurg systemic veins into the central circulation. When a small balloon near its tip is inflated, the balloon passes with the VR into the right atrium, right ventricle, and pulmonary artery until it wedges in a pulmonary artery branch, obstructing the flow there. Because there is no flow, the hole in the catheter tip is open to a stagnant column of blood extending through the pulmonary vessels to the left atrium. Accordingly, this Ppw approximates Pla, providing an estimate of LVEDP to evaluate ventricular function and an estimate of pulmonary microvascular pressure to help manage pulmonary edema (see below). When the balloon is deflated and flow resumes through that vessel, the pressure there is equal to pulmonary arterial pressure. Mixed venous blood drawn from the pulmonary artery provides a measure of O2 content (C −VO2); when related to the simultaneous measurement of arterial O2 content (CaO2) and Q̇T, the patient's O2 consumption (V̇O2 = Q̇T[CaO2 − C −VO2]) can be calculated and interpreted in the context of the patient's O2 transport (DO2 = Q̇T × CaO2). A sensitive thermistor at the tip of the catheter may be used to detect temperature changes after the injection of a cold saline bolus into the right atrium to allow estimation of Q̇T from the resulting thermodilution curve.

The pulmonary artery and the left atrium are surrounded by Ppl, so absolute values of Ppa and Pla change with respiration. When spontaneous active inspiration decreases Ppl, pulmonary arterial and left atrial pressures decrease, but the driving pressure of blood flow across the lung stays the same (Ppa − Pla); when positive-pressure inflation increases Ppl, Ppa and Pla increase. Accordingly, it is helpful to record pulmonary vascular measurements at end expiration when the mode of ventilation has minimally different effects; even this approach can be confounded when the patient exerts vigorous respiratory activity. When alveolar pressure (PA) exceeds Pla, the true driving pressure for pulmonary blood flow is Ppa − PA. One often overlooked adverse effect of positive-pressure ventilation with high PEEP or high tidal volume is the large increase in dead space (VD/VT) when pulmonary blood flow is interrupted by the high PA; not infrequently, alveolar ventilation can actually increase when tidal volume is reduced in these conditions, causing a paradoxical fall in PAco2. A second consequence of PA being greater than Pla is an overestimation of Ppw; this can be detected when the respiratory fluctuation in Ppa is much less than that in Ppw.51 Given these effects of respiration on measurements of Ppa and Ppw, it is not surprising that many physicians err in their interpretation of PAC data.52,53 Further, PAC use is accompanied by complications, and it can be argued that the hemodynamic data obtained can be deduced by clinical examination, are not helpful in clinical decision making, or do not improve outcome.54,55 However, physicians err in their clinical evaluations,56–57 so it seems reasonable to encourage the informed use of PAC to obtain hemodynamic data when there is clinical uncertainty and when those data will be used to titrate aspects of the patient's management.

Pulmonary Edema

Figure 20-11 shows a schematic diagram depicting the circulatory factors governing the movement of edema (Q̇E) between the pulmonary vessels and the lung interstitial tissues; the Starling equation describing lung liquid flux is written beneath the figure. The hydrostatic pressure in the microvessels of the lung (Pmv = 12 mm Hg) lies about halfway between Ppa (normally about 15 mm Hg) and LVEDP (normally about 10 mm Hg). Hydrostatic pressure in the septal interstitial space (Pis = −4 mm Hg) is subatmospheric, in part because it drains into the peribronchovascular interstitium, which has a more negative pressure, and in part because lymph vessels, valved like veins for unidirectional flow, actively remove liquid from the interstitial spaces that have intrinsic structural stability to resist collapse.58 Accordingly, there is a positive hydrostatic pressure (Pmv − Pis = 16 mm Hg) driving edema across the microvascular endothelium to the lung septal interstitium. The vascular wall presents a barrier to this bulk flow of liquid characterized by its permeability to water (Kf; mL edema/min per mm Hg); Kf includes surface area (S) and thus is heavily weighted by the characteristics of the alveolar vessels, where so much S resides.58 The microvascular membrane is also characterized by its permeability to circulating proteins, dominated by albumin and globulin. If these plasma proteins were completely reflected (σ = 1), no protein would pass from lung blood to the interstitium; in contrast, if the microvascular membrane were freely permeable (σ = 0), interstitial protein concentration (CL), as measured in lung lymph, would equal that of plasma proteins (Cp). CL/Cp is about 0.6 in the normal steady-state edema flow in most mammals; when Q̇E, as estimated from lung lymph flow (Q̇L), is progressively increased by elevating Pmv, CL/Cp decreases to a plateau value of about 0.3. This plateau value indicates the microvascular protein reflection coefficient (σ = 1 2 CL Cp = 0.7) measured in conditions of high edema flow; at lower Q̇E levels, water diffuses from the interstitium to the blood along the concentration gradient for water established by Cp > CL.58

In cardiogenic pulmonary edema, Q̇E is increased by increasing Pmv. Several factors act to keep the lungs from accumulating excess liquid: lymphatic flow increases, CL/Cp decreases, and Pis increases. The increased septal Pis drives edema through tissue planes toward the intraparenchymal peribronchovascular interstitium, where Pis is rendered even more subatmospheric (−10 mm Hg) by the outward pull of alveolar walls on the adventitia surrounding the relatively stiff bronchi and vessels.58 This adventitial pull renders Pis even more negative with each inspiration, creating a cyclic suction to move edema from the alveolar septa toward the hilum of the lung, where peribronchovascular interstitial pressures are most negative, where the tissues have the largest capacity to accommodate the edema, and where the most dense accumulation of lymphatics is arranged to clear the edema to the systemic veins. This accounts for the Kerley lines, the bronchial cuffing, and the perihilar “butterfly” distribution of interstitial cardiogenic pulmonary edema on the chest radiograph. When edema genesis continues to fill these interstitial reservoirs, Pis rises at the alveolar septa, disrupting tight junctions between alveolar type I epithelium to flood the airspaces. Histologic morphometry of edematous lungs shows that flooded alveoli have about one-eighth the volume of unflooded alveoli, indicating that a relatively small volume of alveolar edema floods eight times that volume of airspace; for example, in a patient with an end-expired lung gas volume of 4 L, 250 mL of alveolar edema fills half the airspaces (8 × 250 = 2 L), accounting for a large intrapulmonary shunt and for a large reduction in CL because only half the lung is ventilated.22

In the exudative phase of ARDS, a greater proportion of noncardiogenic edema accumulates in airspaces, so there is a much greater shunt per edema volume than in cardiogenic edema. Presumably, this different distribution of edema occurs because the lung injury that increases Kf and decreases S also damages the alveolar epithelial barrier, so increased Q̇E has access to a low-resistance pathway to a very large reservoir for edema—the airspaces of the lung.59 Often, the hydrostatic pressure driving edema from vessels to airspace is normal or reduced; as Q̇E increases at normal Pmv after an acute lung injury, CL/Cp does not decrease as in cardiogenic edema but increases slightly to a value of about 0.8, so the reflection coefficient decreases (σ = 1 2 CL/Cp = 0.2 L). Accordingly, alveolar fluid protein concentration (CA) approaches Cp in ARDS but is much lower than Cp in cardiogenic edema.56 When the vascular membrane is repaired, alveolar edema is cleared very slowly from noninjured lungs by active transport of sodium; water follows the osmotic gradient through an intact alveolar membrane, and this clearance raises CA above Cp as a clinical marker of recovery from ARDS.60

PEEP increases end-expired lung volume to decrease Pis and increase capacity in the peribronchovascular interstitium; this in turn redistributes much of the alveolar edema into this interstitial reservoir, associated with the aeration of flooded airspaces at a much larger alveolar volume to reduce shunt and to increase CL without altering the amount of edema.11,12,61 Because lung volume increases greatly when PEEP is effective in redistributing edema, Ppl must increase to push the chest wall to an equivalently higher volume (see Fig. 20-5). This raises Pra to reduce VR and BP4,5,20 unless the patient's baroreceptor reflexes, iatrogenic infusions of fluid, or vasoactive drugs maintain Pms and Q̇T.25,62 As illustrated in Fig. 20-9, this recruitment of previously flooded airspaces occurs within the large P-V hysteresis of the edematous lung, so less PEEP is required than that indicated by the inflection point of the inflation P-V curve.63

An Approach to Managing Acute Hypoxemic Respiratory Failure

As with many therapeutic interventions in critical illness, too much can cause harm, so it is helpful to define the goal of each intervention and then use the mildest intervention to achieve that goal. Ventilator management of pulmonary edema causing AHRF is summarized in Table 20-3. Because the aim of PEEP therapy is to maintain arterial saturation (>90%) of an adequate circulating hemoglobin (<12 g%) on a nontoxic fraction inspired O2 (FIO2; <0.6)—all to effect adequate DO2 without aggravating the lung injury with oxygen toxicity—it is helpful to use the least PEEP to achieve these goals so as not to reduce Q̇T. Because PEEP already increases end-expired lung volume, superimposed large tidal volumes cause pulmonary barotraumas, further reduce VR, and contribute to mortality; using the least tidal volume (e.g., 6 mL/kg ideal body weight) effecting adequate CO2 elimination at an increased rate minimizes these complications. It is remarkable how rapidly PEEP redistributes edema to reduce hypoxemia (in minutes) and how rapidly the shunt returns when PEEP is removed. Accordingly, the informed physician can implement an effective, tolerable estimate of least PEEP in less than 15 minutes in ventilated patients in whom BP and pulse oximetry are being monitored continuously. Beginning with a small tidal volume (6 mL/kg), high respiratory rate (30 breaths/min), and FIO2 of 1 in a well-sedated patient, PEEP is increased by 5 cm H2O every minute from 0 to 20 minutes. If BP does not decrease and arterial O2 saturation (SaO2) remains 95%, FIO2 is reduced to 0.8 for 5 minutes and then to 0.7 and 0.6 at 5-minute intervals. If SaO2 still exceeds 95%, PEEP is reduced by 5 cm H2O every 5 minutes until SaO2 is greater than 90% and then increased by 2 to 3 cm H2O to restore SaO2 to less than 90%. A decrease in BP as PEEP is initially increased indicates relative circulatory hypovolemia, so PEEP must be reduced again until Q̇T and BP are restored with volume infusion including packed red blood cells to achieve an adequate hematocrit or with an infusion of dobutamine titrated from 1 to 10 μg/kg per minute to maintain Q̇T at a lower circulatory volume and Ppw. Similarly, if the initial FIO2 reductions decrease SaO2 by less than 90%, PEEP should be increased in 2.5-cm H2O increments above 20 cm H2O until SaO2 is high enough to allow FIO2 reduction; at this stage, it is prudent to reduce the tidal volume further. When PEEP is effective, plans to prevent its inadvertent removal, as during routine bedside suctioning, can prevent sudden hypoxemic cardiovascular catastrophe.

Cardiovascular management of cardiogenic and noncardiogenic edema aims to reduce edema formation and accumulation without inducing inadequate Q̇T or DO2 (see Table 20-3), thereby decreasing the duration and complications of intensive care.62 Cardiogenic edema is caused by high Pmv, often related to acute or acute-on-chronic LV dysfunction that increases LVEDP. Reducing the central blood volume by venodilating agents (e.g., morphine, furosemide, or nitroglycerin) or procedures (e.g., rotating tourniquets, phlebotomy, or diuresis) reduces LVEDP and edema genesis, but excess preload reduction will adversely reduce Q̇T from a poorly functioning ventricle that often requires a higher LVEDP (16 to 30 mm Hg) than normal (8 to 12 mm Hg). Where indicated, vasoactive drugs to enhance systolic function (e.g., dobutamine, milrinone, nitroglycerin, or oxygen) or reduce afterload (e.g., enalaprilat, calcium channel blockers, or nitroprusside) and measures to correct diastolic dysfunction (e.g., prolong filling time, maintain coordinated atrial contraction, or correct myocardial hypoxia and ischemia) act to reduce the LVEDP required for adequate Q̇T and thus reduce cardiogenic edema formation by the Starling equation. Increasing πmv by colloid infusion also reduces edema formation, provided Pmv is not increased; an albumin infusion that raises πmv from 15 to 20 mm Hg and Pmv from 25 to 30 mm Hg causes more Q̇E because σ = 0.7.

Colloid infusion is even less helpful in reducing noncardiogenic edema, where σ is much reduced. In one study of oleic acid-induced noncardiogenic edema in dogs, raising πmv by 5 mm Hg had no effect on edema when Pmv was not allowed to change, but lowering Pmv by 5 mm Hg (when πmv did not change) reduced edema by 50% over the same 4 hours of treatment.57 Compelling evidence in other animal64–67 and clinical studies31–33 of acute lung injury supports the therapeutic effect of reducing pulmonary vascular pressures in reducing pulmonary edema (Fig. 20-12), and this is currently being tested prospectively in patients with acute lung injury and ARDS (ARDS Network Fluid and Catheter Treatment Trial). When noncardiogenic edema occurs at normal or low Pmv, the reduction of LVEDP by only 5 mm Hg reduces LVEDV, SV, and Q̇T by as much as a 20 mm Hg reduction in LVEDP in cardiogenic edema. Accordingly, reducing the measured Ppw in a patient with acute lung injury must avoid causing inadequate Q̇T and DO2. As in cardiogenic edema, Ppw can be reduced considerably without evidence of inadequate output and DO2, and vasoactive drugs and increased hematocrit are effective in restoring adequacy of DO2 at low Ppw when indicated. This approach constantly seeks the least Ppw associated with an adequate Q̇T and DO2 during the early stage of edema formation in ARDS. Of course, this is only symptomatic treatment; as yet there are no specific therapies for the acute lung injury that correct an increased Kf and a reduced S. The aim is to minimize the edema consequences of vascular injury and thereby shorten duration of ventilation and care in the intensive care unit.31–33