The commonest causes of absolute hypovolemia are persistent diarrhea, vomiting, and massive bleeding, either GI or as a result of trauma. The reduction in ECF volume is isotonic inasmuch as there is a proportionate loss of water and plasma. The consequent fall in systemic BP leads to compensatory tachycardia and vasoconstriction, and the ensuing altered transcapillary Starling hydraulic forces enable a shift of fluid from the interstitial to intravascular compartment. In addition, neural and hormonal responses to hypovolemia result in renal Na ⁺ and water retention, with the aim of restoring intravascular volume and hemodynamic stability.

Similar compensatory mechanisms become activated after fluid losses from the skin, GI system, and respiratory system. Because of the large surface area of the skin, large amounts of fluid can be lost from this tissue because of burns or excessive perspiration. Severe burns cause the loss of large volumes of plasma and interstitial fluid and can lead rapidly to profound hypovolemia. Without medical intervention, hemoconcentration and hypoalbuminemia supervene. As occurs after massive bleeding, the fluid loss is isotonic, so plasma Na ⁺ concentration and osmolality remain normal. In contrast, excessive sweating, induced by exertion in a hot environment, leads to hypotonic fluid loss because of the relatively low Na ⁺ concentration (20 to 50 mmol/L) in sweat. The resulting hypovolemia may therefore be accompanied by hypernatremia and hyperosmolality and the type of fluid replacement must be tailored accordingly (see Chapter 14).

In addition to oral intake, the GI tract is characterized by the entry of approximately 7 L of isotonic fluid, the overwhelming majority of which is reabsorbed in the large intestine. Hence in normal conditions, fecal fluid loss is minimal. However, in the presence of pathologic conditions, such as vomiting, diarrhea, and colostomy and ileostomy secretions, especially those caused by infection, considerable or even massive fluid loss may occur. The ionic composition, osmolality, and pH of secretions vary according to the part of the GI tract involved; therefore the resulting hypovolemia is associated with a large spectrum of electrolyte and acid-base abnormalities (see Chapter 15 , Chapter 16 for further discussion).

In contrast to the massive losses that can occur from the skin and GI system, fluid loss from the respiratory tract—as occurs in febrile states and in patients who receive mechanical ventilation with inadequate humidification—is usually modest and hypovolemia ensues only in the presence of accompanying causes. Finally, a special situation in which hypovolemia can occur is after excessive ultrafiltration in dialysis patients (see Chapter 62 , Chapter 63 ).

Even when glomerular filtration rate (GFR) is markedly impaired, the amount of filtered Na ⁺ far exceeds the dietary intake and all but ∼1% of the filtered load is reabsorbed. However, if one or more of the tubular reabsorptive mechanisms are impaired, serious Na ⁺ deficit and absolute volume depletion can occur. The causes of absolute renal Na ⁺ losses include pharmacologic agents and renal structural, endocrine, and systemic disorders (see Table 13.3 ). All diuretics used to treat hypervolemic states may induce hypovolemia if administered in excess or inappropriately (see Chapter 49 ). Particularly, the powerful loop diuretics, furosemide, bumetanide, and torsemide, are often given in combination with diuretics acting on other tubular segments (e.g., thiazides, aldosterone antagonists, distal ENaC blockers, and carbonic anhydrase inhibitors). Patients receiving these combinations need to be carefully monitored and fluid balance should be scrupulously adjusted to prevent hypovolemia. Patients commonly at risk are those with HF or underlying hypertension who develop intercurrent infections.

In patients with hypertension, diuretic treatment appreciably increases the risk of volume depletion. Osmotic diuretics, endogenous or exogenous, may also reduce tubular Na ⁺ reabsorption. Endogenous agents include urea, the principal molecule involved in the polyuric recovery phase of acute kidney injury and postobstructive diuresis, and glucose in hyperglycemia. In patients with increased intracranial pressure, exogenous agents, such as mannitol or glycerol, may be used to induce translocation of fluid from the ICF to the ECF compartment and decrease brain swelling. The resulting polyuria may be associated with electrolyte and acid-base disturbances, the nature of which depends on the complex interplay between fluid intake and intercompartmental fluid shifts.

Na ⁺ reabsorption may also be disrupted in inherited and acquired tubular disorders. Inherited disorders of the proximal tubules (e.g., Fanconi syndrome) and distal tubules (e.g., Bartter and Gitelman syndromes) may lead to salt-wasting states in association with other electrolyte or acid-base disturbances. Acquired disorders of Na ⁺ reabsorption may be acute, as in nonoliguric acute kidney injury, the period immediately after renal transplantation, the polyuric recovery phase of acute kidney injury and postobstructive diuresis (please see, respectively, Chapter 27 , Chapter 28 , Chapter 39 , Chapter 69 ), or they may be chronic as a result of tubulointerstitial diseases with a propensity for salt wasting.

In addition to intrinsic tubular disorders, endocrine and other systemic disturbances may lead to impaired Na ⁺ reabsorption. The principal endocrine causes are mineralocorticoid deficiency and resistance states. A controversial cause is the systemic disturbance known as cerebral salt wasting (CSW). In this condition, salt wasting is thought to occur in response to an as yet unidentified factor released in the setting of acute head injury or intracranial hemorrhage. ^, CSW is usually diagnosed because of concomitant hyponatremia and signs of volume depletion in contrast to the normovolemia characteristic of the syndrome of inappropriate antidiuresis. Though one research group has proposed the designation of renal salt wasting and also accumulated single-center evidence that a signal peptide cleaved form of haptoglobin-related peptide drives natriuresis, the entity remains an enigmatic and not universally accepted clinical entity.

An underappreciated, but not uncommon, clinical setting for renal Na ⁺ loss is after the administration of large volumes of intravenous saline to patients over several days after surgery or after trauma. In this situation, tubular reabsorption of Na ⁺ is downregulated. If intravenous fluids are stopped before full reabsorptive capacity is restored, volume depletion may ensue. The phenomenon can be minimized by graded reduction in the infusion rate, which allows Na ⁺ reabsorptive pathways to be restored gradually.

In the context of volume depletion, diabetes insipidus should be mentioned. However, because this results from a deficiency of or tubular resistance to arginine vasopressin (AVP), water loss is the main consequence and the impact on ECF volume is only minor. AVP-related disorders are considered in Chapter 14.

As outlined previously, the principal causes of relative hypovolemia are edematous states, vasodilation, and third-space loss (see Table 13.3 ). Vasodilation may be physiologic, as in normal pregnancy, or induced by drugs (hypotensive agents, such as hydralazine or minoxidil, that cause arteriolar vasodilation), or it may occur in sepsis during the phase of peripheral vasodilation and consequent low SVR.

Edematous states in which the EABV and, hence, tissue perfusion are reduced include HF, decompensated cirrhosis with ascites, and nephrotic syndrome. In severe HF, low cardiac output and resulting low systemic BP lead to a fall in renal perfusion pressure (RPP). As in absolute hypovolemia, the kidneys respond by retaining Na ⁺ . Because the increased venous return cannot raise the cardiac output, a vicious cycle is created in which edema is further exacerbated and the persistently reduced cardiac output leads to further Na ⁺ retention. In decompensated cirrhosis, splanchnic venous pooling leads to decreased venous return, a consequent fall in cardiac output, and compensatory renal Na ⁺ retention. The pathophysiology of edematous states is discussed later (see section “Hypervolemia” later). Third-space loss occurs when fluid is sequestered into compartments not normally perfused with fluids, as in states of GI obstruction, after trauma, burns, or in pancreatitis, peritonitis, or malignant ascites. The result is that, even though total body Na ⁺ is markedly increased, the EABV is severely reduced.

Approximately 10% of patients with nephrotic syndrome—especially children with minimal change disease, but also any patient with a serum albumin level lower than 2 g/ dL—manifest clinical signs of hypovolemia. The low plasma oncotic pressure is conducive to movement of fluid from the ECF compartment to the interstitial space, thereby leading to reduced EABV.

Hemoglobin may decrease if significant bleeding has occurred or is ongoing, but the change, which is caused by hemodilution owing to fluid translocation from the interstitial to intravascular compartment, may take up to 24 hours. Therefore stable hemoglobin does not rule out significant bleeding. Moreover, the adaptive response of hemodilution may moderate the severity of hemodynamic compromise and resulting physical signs. In hypovolemic situations that do not arise from bleeding, hemoconcentration is often, but not universally, seen, inasmuch as underlying anemia of chronic disease may mask the differential loss of plasma.

Hemoconcentration may also be manifested as a rise in plasma albumin concentration if albumin-free fluid is lost from the skin, GI tract, or kidneys. On the other hand, when albumin is lost, either in parallel with other extracellular fluids (as in proteinuria, hepatic disease, protein-losing enteropathy, or catabolic states) or in protein-rich fluid (third-space sequestration, burns), significant hypoalbuminemia is observed.

This may be low, normal, or high, depending on the solute composition of the fluid lost and the replacement solution administered by the patient or physician. For example, the hypovolemic stimulus for AVP release may lead to preferential water retention and hyponatremia, especially if hypotonic replacement fluid is used. In contrast, the fluid content of diarrhea may be hypotonic or hypertonic, resulting in hypernatremia or hyponatremia, respectively. The plasma Na ⁺ concentration reflects the tonicity of plasma and provides no direct information about volume status, which is a clinical diagnosis.

These can also change in hypovolemic conditions. After vomiting and after some forms of diarrhea, loss of K ⁺ and Cl ⁻ may lead to metabolic alkalosis. More often, the principal anion lost in diarrhea is bicarbonate, which leads to hyperchloremic (nonanion gap) acidosis. When diuretics or Bartter and Gitelman syndromes (the inherited tubulopathies; see Chapter 44 ) are the cause of hypovolemia, hypokalemic alkalosis is again typically seen. On the other hand, urinary Na ⁺ loss that occurs in adrenal insufficiency or due to aldosterone hyporesponsiveness is accompanied by a tendency for hyperkalemia and metabolic acidosis. Finally, when hypovolemia is sufficiently severe to impair tissue perfusion, high anion gap acidosis caused by lactic acid accumulation may be observed.

These frequently rise in hypovolemic states and reflect impaired renal perfusion. If tubular integrity is preserved, then the rise in urea levels is typically disproportionate to that of creatinine, so-called prerenal azotemia (see Chapter 27 , Chapter 28 ). This results mainly from AVP-enhanced urea reabsorption in the medullary collecting duct (MCD), but also from augmented proximal tubular reabsorption due to increased filtration fraction. In critically ill patients, an increased urea generation rate (from exogenous or endogenous protein catabolism) or low creatinine generation rate due to muscle wasting may lead to an erroneous diagnosis of prerenal azotemia. In the presence of severe hypovolemia, acute kidney injury may ensue, leading to loss of the differential rise in urea level. Proportional rises in urea and creatinine are also observed when hypovolemia occurs against a background of underlying renal functional impairment, as in chronic kidney disease, stages 3 to 5.

In hypovolemia due to extrarenal fluid losses, the intact kidney will respond to hypoperfusion by enhanced tubular reabsorption of Na ⁺ and water. The ensuing oliguria will be characterized by urine-specific gravity >1.020, Na ⁺ concentration <10 mmol/L, and osmolality >400 mOsm/kg. When urine Na ⁺ concentration is 20 to 40 mmol/L, the finding of a fractional excretion of Na ⁺ of <1%, in the presence of oliguria, may be helpful. However, in a patient with intrinsic tubular disease or injury or on previous diuretic therapy, especially with loop diuretics, these indices may merely reflect U _Na losses. In that case, fractional excretion of urea of <30% to 35% may help in the diagnosis of hypovolemia, although the specificity of this test is rather low. ^,

Conversely, the presence of selective renal or glomerular ischemia (e.g., because of bilateral renal artery stenosis or acute glomerular injury) will be misinterpreted as poor renal perfusion and is associated with renal Na ⁺ retention (low U _Na ).

The treatment goals of hypovolemia are to restore normal hemodynamic status and tissue perfusion. These goals are achieved by reversal of the clinical symptoms and signs, described previously. Treatment can be divided into three stages: 1. initial replacement of the immediate fluid deficit; 2. maintenance of the restored ECF volume in the presence of ongoing losses; and 3. treatment of the underlying cause, whenever possible. The main strategies to be addressed by the clinician are the route, volume, rate of administration, and composition of the replacement and maintenance fluids. These are liable to change according to the patient’s response.

In general, when hypovolemia is associated with a significant hemodynamic disturbance, intravenous rehydration is required. ^, (The use of oral electrolyte solutions in the management of infants and children is discussed in Chapter 71 ). The volume of fluid and rate of administration should be determined on the basis of the urgency of the threat to circulatory integrity, adequacy of the clinical response, and underlying cardiac function. Older patients are especially vulnerable to aggressive fluid challenge, and careful monitoring is required, particularly to prevent acute left ventricular failure and pulmonary edema from overzealous correction.

Sometimes the clinical signs do not point unequivocally to the diagnosis of hypovolemia, even though the history is strongly suggestive. Several tests and indices have been developed and validated. The use of dynamic measures is encouraged to help guide treatment. The gold standard is the invasive measure of cardiac output (or index) by thermodilution and its response to fluid challenge. Responsiveness to volume could also be predicted using passive leg raising, end-expiratory occlusion, or minifluid challenge tests (comprehensively reviewed by Monnet and colleagues). Invasive monitoring of central venous and pulmonary venous pressures has not been shown to improve outcomes in this situation; monitoring preload by stroke volume variation may improve outcomes, at least after major abdominal surgery. Additionally, expanding the knowledge and practical application of noninvasive echocardiography and point-of-care ultrasound (POCUS, Chapter 26 ) could reliably assist clinicians in achieving the goals of volume management. ^{, ,} Inferior vena cava (IVC) collapsibility index (CI) has been validated in both ventilated and nonventilated patients. An IVC maximum diameter <2.1 cm and IVC CI >50% are considered inconsistent with hypervolemia, while IVC CI <20% (with moderate to large IVC diameter) is inconsistent with intravascular volume depletion. Furthermore, combination of IVC ultrasound and lung ultrasound provides a more accurate assessment of intravascular and extravascular volume excess.

However, in case of doubt, a diagnostic fluid challenge should be performed. When a flat response to fluid challenge in any approved measure is achieved, further bolus volume therapy is futile and other therapies should be considered for shock management, such as vasopressors and correction of acidemia, among others. ^,

In resource-limited countries or when dynamic measures are not available, good clinical judgment is essential for successful management because initial fluid deficits are difficult to calculate. Therefore patients with life-threatening circulatory collapse and hypovolemic shock require rapid intravenous replacement through the cannula with the widest bore possible. Replacement should continue until BP and tissue perfusion are restored. In the second stage, the rate of fluid replacement should be reduced to maintain BP and tissue perfusion. In older patients and those with underlying cardiac dysfunction, the risk of overrapid correction and precipitating pulmonary edema is heightened; therefore slower treatment is preferable to allow gradual filling of the ECF volume rather than causing pulmonary edema and the threat of mechanical ventilation associated with adverse outcomes.

Of note, evaluation and treatment of hypovolemia, electrolyte, and acid-base abnormalities associated with hyperglycemic crises (diabetic ketoacidosis and hyperglycemic hyperosmolar state) in adults with diabetes have been updated. In patients without renal or cardiac compromise, the 2024 consensus report recommends starting the administration of isotonic saline or balanced solutions at an initial rate of 500 to 1000 mL/hour during the first 2 to 4 hours and correction of estimated deficits within the first 24 to 48 hours. Again, caution should be used in older patients with additional serious comorbidities and in pregnant individuals.

The composition of replacement fluid may also affect outcomes. The two main categories of replacement solutions are crystalloid and colloid solutions. Crystalloid solutions are based largely on NaCl of varying tonicity or dextrose. Isotonic (0.9%) saline, containing 154 mmol of Na ⁺ /L, is the mainstay of volume replacement therapy being confined to the ECF compartment in the absence of deviations in Na ⁺ concentration. One L of isotonic saline increases plasma volume by approximately 300 mL; the rest is distributed to the interstitial compartment. In contrast, 1 L of 5% dextrose in water (D ₅ W), which is also isosmotic (277 mOsm/L), is eventually distributed throughout all the body fluid compartments so that only 10% to 15% (100–150 mL) remains in the ECF. Therefore D ₅ W should not be used for volume replacement.

Administration of 1 L of 0.45% saline (77 mmol of Na ⁺ /L) in D ₅ W is equivalent to giving 500 mL of isotonic saline and the same volume of solute-free water. The distribution of the solute-free compartment throughout all the fluid compartments would result in plasma dilution and reduction in the plasma Na ⁺ . Therefore this solution should be reserved for the management of hypernatremic hypovolemia. Even in that situation, it must be remembered that volume replacement is less efficient than with isotonic saline and, early in the treatment course, may cause plasma tonicity to fall too rapidly.

When hypovolemia is accompanied by severe metabolic acidosis (pH <7.10; plasma HCO ₃ ⁻ <10 mmol/L), bicarbonate supplementation may be indicated. (For a discussion of bicarbonate balance, see Chapter 15 .) Because this anion is manufactured as 8.4% sodium bicarbonate (1000 mmol/L) for use in cardiac resuscitation, appropriate dilution is required for the treatment of acidosis associated with hypovolemia. Two convenient methods are suggested. Either 75 mL (75 mmol) of 8.4% NaHCO ₃ ⁻ can be added to 1 L of.45% saline, or 150 mL of concentrated bicarbonate can be added to 1 L of D ₅ W. Although the latter is hypertonic in the short term, it is unlikely to be harmful.

In the presence of accompanying hypokalemia, especially if metabolic alkalosis is also present, volume replacement solutions should be supplemented with K ⁺ . Commercially available 1-L solutions of isotonic saline supplemented with 10 or 20 mmol of KCl make this option safe and convenient. (For details, see Chapter 16 .) On the other hand, newer, commercially available crystalloid solutions containing lactate (converted by the liver to bicarbonate) and low concentrations of KCl may offer advantages over isotonic saline. In a large prospective observational study performed in the intensive care unit setting, two periods were compared; in the control period, all patients received isotonic saline as fluid replacement, whereas during the intervention period, Hartmann solution (lactate-containing), Plasma-Lyte 148 (a balanced multielectrolyte solution, BMES), or chloride-poor 20% albumin solution was administered. The chloride-poor solutions were associated with a significantly lower risk of subsequent acute kidney injury, even after adjustment for covariates. Clearly, these provocative results indicated the need for well-controlled studies comparing BMES with chloride-rich solutions for fluid resuscitation. ^, Indeed, two recent double-blind, randomized controlled trials (RCT) addressed this therapeutic dilemma in critically ill patients. ^, The PLUS study, an investigator-initiated, double-blind RCT, was conducted in 53 intensive care units (ICUs) in Australia and New Zealand. The number of patients screened was 16,828, and about 30% were randomized (2515 and 2522 patients were assigned to receive BMES or saline, respectively). Compared with saline, the study showed that Plasma-Lyte 148 did not reduce 90-day all-cause mortality or risk of AKI and conferred no advantage in use of kidney replacement therapy (KRT) within 90 days, days alive and free of mechanical ventilation, or all-cause mortality at 28 days. These results were in line with the outcomes of the BaSICS trial, which was conducted in 75 ICUs in Brazil. It was designed as a 2 × 2 RCT, assigning 10,520 critically ill patients into BMES (Plasma-Lyte 148) or 0.9% NaCl and a second randomization according to fluid rate delivery (333 mL/hour vs. 999 mL/hour). The 5290 patients were randomized to receive 0.9% NaCl; 2641 received the rapid infusion rate, whereas 5230 received BMES—2601 at rapid infusion rate. Results showed that there was no significant interaction between the two interventions, and neither demonstrated any significant difference between groups in primary outcome (90-day mortality) or secondary outcomes. However, in subgroup analysis, statistically significant interaction among presence of traumatic brain injury, fluid type, and 90-day mortality was evident (31.3% for the BMES group vs. 21.1% for the saline solution group [HR, 1.48; 95% CI, 1.03–2.12]). Subsequently, two meta-analyses, based on data from six trials, suggested a high probability that the use of BMES in the ICU reduces in-hospital mortality, although the certainty of the evidence was moderate and the absolute risk reduction was small. Yet using BMES in the setting of traumatic brain injury was associated with increased in-hospital mortality. ^,

Colloid solutions include plasma, albumin, and high-molecular-weight carbohydrate molecules, such as hydroxyethyl starch and dextrans, at concentrations that exert colloid oncotic pressure (COP) equal to or greater than that of plasma. Because the transcapillary barrier is impermeable to these large molecules, in theory, they expand the intravascular compartment more rapidly and efficiently than crystalloid solutions. Colloid solutions may be useful in the management of burns and severe trauma when plasma protein losses are substantial and rapid plasma expansion with relatively small volumes is efficacious. However, when capillary permeability is increased, as in states of multiorgan failure or the systemic inflammatory response syndrome, colloid administration is ineffective. Moreover, RCTs, in which crystalloid solutions were compared with colloid solutions, have shown no survival benefit and even harm with some colloid solutions, particularly hydroxyethyl starch. Therefore the much cheaper and more readily available crystalloid solutions should remain the mainstay of therapy.

Renal Na ⁺ (and water retention) in edematous disorders occurs due to reduced EABV, despite an increase in total blood and ECF volumes and normal intrinsic renal function. If the underlying stimulus for hypervolemia is removed, as dramatically seen after heart or liver transplantation, Na ⁺ excretion is restored to normal. Conversely, when kidneys from patients with end-stage liver disease are transplanted into patients with normal liver function, Na ⁺ retention no longer occurs.

Because 85% of blood circulates in the venous compartment, expansion of that compartment leads to overall ECF volume excess that could occur concurrently with arterial underfilling. The latter could result from low cardiac output, peripheral arterial vasodilation, or a combination of the two. In turn, low cardiac output could result from true ECF volume depletion (see earlier discussion), cardiac failure, or decreased π _c , with or without increased capillary permeability. All these stimuli would cause activation of ventricular and arterial sensors. Similarly, conditions such as high-output cardiac failure, sepsis, cirrhosis, and normal pregnancy lead to peripheral arterial vasodilation and activation of arterial baroceptors. Activation of these afferent mechanisms would then induce the neurohumoral mechanisms that result in renal Na ⁺ and water retention (see Fig. 13.2 ).

Although the mechanisms leading to Na ⁺ retention in HF and cirrhosis are similar, specific differences between the two conditions have been observed and are discussed separately in the following sections.

Both the cardiopulmonary and baroceptor reflexes are blunted in HF, so they cannot exert an adequate tonic inhibitory effect on sympathetic outflow. The resulting sympathetic nervous system (SNS) activation triggers, among others, renal Na ⁺ retention. ^, A variety of models of HF have shown marked attenuation of atrial receptor firing and loss of nerve-ending arborization in HF. Similarly, altering central cardiac filling pressures in response to postural stimuli (e.g., head-up tilt and LBNP) in HF patients, in contrast to normal subjects, usually do not demonstrate significant alterations in limb blood flow, circulating catecholamines, AVP, or renin activity. ^, This diminished reflex responsiveness is proportionate to the severity of ventricular dysfunction.

Arterial baroceptor reflex impairment has been observed in HF. High baseline values of muscle sympathetic activity were found in patients with HF who failed to respond to activation and deactivation of arterial baroreceptors by infusion of phenylephrine and Na ⁺ nitroprusside, respectively. Carotid and aortic baroreceptor function were also depressed in experimental models of HF. These changes were associated with upward resetting of receptor threshold and a reduced range of pressures over which the receptors functioned.

Multiple abnormalities have been described in cardiopulmonary and arterial baroreceptor control of renal sympathetic nerve activity (RSNA) in HF. Thus rats with coronary ligation displayed an increased basal level of efferent RSNA that failed to decrease normally during volume expansion. ^, Similarly, in sinoaortic denervated dogs with pacing-induced HF, or following left atrial baroreceptor stimulation, the cardiopulmonary baroreflex control of efferent RSNA was markedly attenuated.

The abnormal regulation of efferent RSNA was caused by impaired function of aortic and cardiopulmonary baroreflexes; the latter defect was functionally more important. Mechanisms implicated in the pathogenesis of these abnormal baroreflexes include loss of compliance in the dilated hearts, gross changes in receptor structure, and augmented Na ⁺ -K ⁺ -ATPase activity in the baroreceptor membranes. Increased activity of angiotensin II (Ang II) through the AT ₁ receptor also contributes to depressed baroreflex sensitivity. Thus renin-angiotensin-aldosterone system (RAAS) inhibition in rats or rabbits with HF significantly improved arterial baroreflex control of RSNA or heart rate, respectively. ^, This effect of Ang II could also be blocked by central α ₁ -adrenoreceptor stimulation.

Newer studies have indicated that Ang II in the paraventricular nucleus potentiates—and AT ₁ receptor antisense mRNA normalizes—the enhanced cardiac sympathetic afferent reflex in rats with chronic HF. AT ₁ receptors in the nucleus tractus solitarii are thought to mediate the interaction between the baroreflex and cardiac sympathetic afferent reflex. Consistent with this notion, Ang II generation is enhanced and its degradation is reduced in central sympathoregulatory neurons, as shown by upregulation of angiotensin-converting enzyme (ACE)1 and downregulation of ACE2. AT ₂ receptors in the rostral ventrolateral medulla inhibited sympathetic outflow, an effect mediated at least partly by an arachidonic acid metabolic pathway. These studies indicated that a downregulation in the AT ₂ receptor was a contributory factor in the sympathetic neural excitation in HF.

Together, these data provide evidence of the role of high endogenous levels of Ang II, acting through the AT ₁ receptor in concert with downregulation of the AT ₂ receptor, in the impaired baroreflex sensitivity observed in HF, both in the afferent limb of the reflex arch and at more central sites. The central effect may be mediated through a central α ₁ -adrenoreceptor. The blunted cardiopulmonary and arterial baroreceptor sensitivity in HF may also lead to an increase in AVP release and renin secretion.

The disturbances in sensing mechanisms that initiate and maintain renal Na ⁺ retention in HF are summarized in Fig. 13.2A . As indicated, a decrease in cardiac output or a diversion of systemic blood flow diminishes the blood flow to the critical sites of the arterial circuit with pressure- and flow-sensing capabilities. The responses to diminished blood flow culminate in renal Na ⁺ retention, mediated by the effector mechanisms. An increase in systemic venous pressure promotes the transudation of fluid from the intravascular to interstitial compartment by increasing the peripheral transcapillary ΔP. These processes augment the perceived loss of volume and flow in the arterial circuit. In addition, distortion of the pressure-volume relationships as a result of chronic dilation in the cardiac atria attenuates the normal natriuretic response to central venous congestion. This attenuation is manifested predominantly as a diminished neural suppressive response to atrial stretch, which results in increased sympathetic nerve activity and augmented release of renin and AVP.

The adaptive changes in the efferent limb of the volume control system in HF are generally similar to those seen in states of true Na ⁺ depletion. These include adjustments in glomerular hemodynamics and tubular transport, brought about by alterations in the neural, humoral, and paracrine systems. However, in contrast to true Na ⁺ depletion, HF is also associated with activation of vasodilatory/natriuretic agents, which tend to oppose the effects of the vasoconstrictor/antinatriuretic systems. The final effect on urinary Na ⁺ excretion is determined by the dynamic balance among these antagonistic effector systems.

HF is characterized by not only increased renal vascular resistance and reduced GFR but also an even greater reduction in renal plasma flow (RPF), so the filtration fraction is increased. As shown in rat models of HF, these changes seem to result from diminished Kf and elevated afferent and efferent arteriolar resistances. The rise in filtration fraction is probably caused by a disproportionate increase in efferent arteriolar resistance.

In Fig. 13.3 , a comparison of the glomerular capillary hemodynamic profile in the normal versus the HF state is illustrated on the left graph of each panel. First, ΔP declines along the length of the glomerular capillary in normal and HF states, but much more so in HF because of the increased efferent arteriolar resistance. Second, Δπ increases over the length of the glomerular capillary in both states as fluid is filtered into the Bowman space, but again to a greater extent in HF because of the increased filtration fraction. As outlined later (see “Renin-Angiotensin-Aldosterone System” later), the preferential increase in efferent arteriolar resistance is mediated principally by Ang II and is critical for the preservation of GFR in the presence of reduced RPF. Because of the intense efferent arteriolar vasoconstriction, further compensation is not possible if RPP falls as a result of systemic hypotension, causing a sharp decline in GFR. This phenomenon is dramatically illustrated by HF patients whose Ang II drive is removed by RAAS inhibitors, particularly those with preexisting renal failure, massive diuretic treatment, and limited cardiac reserve. In these patients, BP may fall below the level necessary to maintain renal perfusion.

Peritubular control of proximal tubule fluid reabsorption.

Fluid reabsorption in the normal state (left) and in patients with heart failure (right) is shown. Increased postglomerular arteriolar resistance in heart failure is depicted as narrowing. The thickness and font size of the block arrows depict relative magnitude of effect. The increase in filtration fraction (FF) in heart failure causes Δπ to rise. The increase in renal vascular resistance in heart failure is believed to reduce ΔP. Both the increase in Δπ and fall in ΔP enhance peritubular capillary uptake of proximal reabsorbate and thus increase absolute Na ⁺ reabsorption by the proximal tubule. Numbers and red block arrows depict renal plasma flow [mL/min] in preglomerular and postglomerular capillaries; ΔP and Δπ are the transcapillary hydraulic and oncotic pressure differences across the peritubular capillary, respectively; yellow block arrows indicate transtubular transport; purple block arrows represent the effect of peritubular capillary Starling forces on uptake of proximal reabsorbate.

Modified from Humes HD, Gottlieb M, Brenner BM. The Kidney in Congestive Heart Failure: Contemporary Issues in Nephrology. Vol 1. New York; Churchill Livingstone; 1978:51–72.

A direct consequence of the glomerular hemodynamic alterations and augmented single-nephron filtration fraction is an increase in the fractional reabsorption of filtered Na ⁺ in the proximal tubule. In Fig. 13.3 , the peritubular capillary hemodynamic profile of the normal state is compared with that of HF on the right graph of each panel. In HF, in comparison with the normal state, the average value of Δπ along the peritubular capillary is increased and that of ΔP is decreased. These values favor fluid movement into the capillary and may also help reduce paracellular backleak of fluid into the tubule, promoting overall net reabsorption.

The peritubular control of proximal fluid reabsorption in normal and HF states is illustrated schematically in Fig. 13.3 . A critical mediator of the enhanced tubular reabsorption of Na ⁺ is Ang II, which, by increasing efferent arteriolar resistance, increases the filtration fraction and augments proximal epithelial transporter activities directly, thereby amplifying the overall increase in proximal Na ⁺ reabsorption. This is clearly illustrated by the favorable effects of RAAS blockers in HF to modulate single-nephron filtration fraction and normalize proximal peritubular capillary Starling forces and Na ⁺ reabsorption.

Enhanced reabsorption of Na ⁺ in HF has been shown in the loop of Henle, probably due to altered renal hemodynamics, as in the proximal tubule. In the distal tubule and collecting duct, elevated Ang II and aldosterone levels, respectively, enhance activities of the NaCl cotransporter and ENaC.

The primary vasoconstrictor/antinatriuretic (and antidiuretic) systems mediating Na ⁺ and water retention in HF include the RAAS, SNS, AVP, and endothelins (ETs). The antagonistic vasodilator/natriuretic substances include nitric oxide (NO), prostaglandins (PGs), adrenomedullin (AM), urotensin II (UT II), and neuropeptide Y (NPY). The development of positive Na ⁺ balance and edema in HF occurs at the point when the vasoconstrictor/antinatriuretic forces predominate ( Fig. 13.4 ). The dominant activity of Na ⁺ -retaining systems in HF is clinically important, as it is associated with globally impaired renal function, a strong predictor of mortality ; moreover, reversal of neurohumoral impairment is associated with improved outcomes.

Efferent limb of extracellular fluid volume control in heart failure.

Volume homeostasis in heart failure is determined by the balance between natriuretic and antinatriuretic forces. In decompensated heart failure, enhanced activities of the Na ⁺ -retaining systems overwhelm the effects of the vasodilatory/natriuretic systems, which leads to a net reduction in Na ⁺ excretion and an increase in ECF volume. ANP, Atrial natriuretic peptide.

Modified from Winaver J, Hoffman A, Abassi Z, et al. Does the heart’s hormone, ANP, help in congestive heart failure? News Physiol Sci . 1995;10:247–253.

Plasma levels of ANP and NH ₂ -terminal ANP are frequently elevated in HF and are correlated positively with the severity of cardiac failure, elevated atrial pressure, and left ventricular dysfunction. Hence circulating ANP level was proposed as a diagnostic marker of cardiac dysfunction and as a predictor of survival in HF. However, in this context, ANP has since been superseded by BNP. There is also evidence that mid-regional (MR) proANP may perform similarly to BNP as a biomarker of ADHF (see “Brain Natriuretic Peptide” later).

The high plasma ANP levels are attributed to increased production rather than to decreased clearance. Although volume-induced atrial stretch is the main source for the elevated circulating ANP levels in HF, enhanced synthesis and release of the hormone by ventricular tissue in response to Ang II and ET also contribute to the elevated levels. Despite the high levels, patients and experimental animals with HF retain salt and water because renal responsiveness to NPs is attenuated. However, infusion of ANP to patients with HF does lead to hemodynamic improvement and inhibition of activated neurohumoral systems. These data are in line with findings that ANP is a weak counterregulator of the vasoconstriction mediated by the SNS, RAAS, and AVP. However, despite the blunted renal response to ANP in HF, elimination of ANP production by atrial appendectomy in dogs with HF aggravated the activation of the vasoconstrictive factors and resulted in marked Na ⁺ and water retention. These data suggest that ANP plays a critical role in suppressing Na ⁺ -retaining systems and as an important adaptive or compensatory mechanism aimed at reducing pulmonary vascular resistance and hypervolemia.

Plasma levels of BNP and N-terminal (NT)–proBNP are elevated in severe HF in proportion to the degree of myocardial systolic and diastolic dysfunction and the New York Heart Association (NYHA) classification. ^, The extreme elevation of plasma BNP in severe HF stems mainly from increased synthesis by the hypertrophied ventricular tissue, but the atria also contribute. ^,

Although echocardiography remains the gold standard for the evaluation of left ventricular dysfunction, plasma levels of BNP and NT-proBNP are reliable markers and, in fact, superior to ANP and NT-proANP for the diagnosis and prognosis of HF. NT-proBNP, in particular, has high sensitivity, specificity, and negative predictive value in patients with an ejection fraction <35%. Similar high predictive values are found in patients with concomitant left ventricular hypertrophy, either in the absence of or after myocardial infarction. The added presence of renal dysfunction appears to enhance these predictive values, ^, and graded increases in mortality throughout each quartile of BNP levels have been shown in several clinical trials. In addition, elevated plasma BNP (or NT-proBNP) levels and LVEF lower than 40% are complementary independent predictors of death, HF, and new myocardial infarction at 3 years after a first infarction. Moreover, risk stratification with the combination of LVEF lower than 40% and high levels of NT-proBNP is substantially better than that provided by either alone. However, even though BNP levels tend to be lower in patients with preserved LVEF than in HF patients with reduced LVEF, the prognosis in patients with preserved LVEF is as poor as in those with reduced LVEF for a given BNP level.

In asymptomatic patients with preserved LVEF, elevated BNP levels are correlated with diastolic abnormalities in Doppler studies. Conversely, a reduction in BNP levels with treatment is associated with a reduction in left ventricular filling pressures, lower readmission rates, and a better prognosis; thus monitoring of BNP levels may provide valuable information regarding treatment efficacy and expected patient outcomes.

Another diagnostic role for BNP is in the distinction of dyspnea caused by HF from that caused by noncardiac diseases. Using specific cut points, NT-proBNP levels are highly sensitive and specific for the diagnosis of acute HF. Levels lower than 300 pg/mL rule out acute HF, with a negative predictive value of 99%. An increased level of NT-proBNP is the strongest independent predictor of a final diagnosis of acute HF. NT-proBNP testing alone was superior to clinical judgment, the National Health and Nutrition Examination score, and Framingham clinical parameters alone for diagnosing acute HF; NT-proBNP plus clinical judgment was superior to NT-proBNP or clinical judgment alone.

There is also evidence, albeit of rather low quality, that circulating BNP and NT-proBNP levels can be useful as a guide to therapeutic efficacy of drugs typically prescribed in HF including RAAS inhibitors, diuretics, digitalis, and β-blockers. ^, Also, BNP, but not NT-proBNP, levels at 24 and 48 hours after admission for acute decompensated HF (ADHF) predicted both 30-day and 1-year mortality. Predischarge levels of both peptides were predictive of 30-day and 1-year mortality but not 1-year readmission due to HF. In contrast, BNP levels were not helpful in reducing length of hospital stay and costs.

Together, these findings suggest that a simple and rapid determination of plasma levels of BNP or NT-proBNP in HF patients, together with clinical and echocardiographic measures, can be used to assess cardiac dysfunction, serve as a diagnostic and prognostic marker, and possibly assist in titrating relevant therapy. However, it should be emphasized that plasma NP levels are affected by age, salt intake, gender, obesity, hemodynamic status, and renal function leading to considerable overlap among diagnostic groups. ^, Measurement of a panel may enhance the ability of biomarkers to distinguish between cardiac and noncardiac causes of dyspnea.

Like those of ANP and BNP, plasma C-type NP (CNP) levels are increased in HF and are directly correlated with NYHA classification; levels of BNP, ET-1, and AM; pulmonary capillary wedge pressure; ejection fraction; and left ventricular end-diastolic diameter. CNP is synthesized mainly in the kidney, but it is also processed by the myocardium. Overexpression of CNP in the myocardium during HF may be involved in counteracting cardiac remodeling. On the other hand, renal CNP secretion is blunted in HF. In contrast to the diminished physiologic responses to ANP and BNP in animals with HF, CNP elicited twice as much soluble guanylate cyclase (sGC) activity as ANP due to dramatic reductions in NP receptor-A (NPR-A) but not NPR-B activity. These findings imply a significant role for NPR-B–mediated NP activity in HF and may explain the modest effects of the NPR-A–selective nesiritide (BNP) treatment in HF.

Overall, current evidence points to a role of CNP in either the peripheral vascular compensatory response in HF or mitigation of the cardiac remodeling characteristic of HF. Elaboration of the exact role of CNP in HF appears crucial for the design of more effective NP analogs than those currently available for the management of HF.

The initial trigger for splanchnic arterial vasodilation is hepatic tissue damage itself, which leads to venous outflow obstruction, reduced portal venous, blood flow and increased hepatic arterial blood flow. Moreover, the lower the portal venous flow, the higher the hepatic arterial flow ( Fig. 13.5A ). These changes lead to increased intrahepatic vascular resistance and sinusoidal pressure. Increased hepatic resistance to portal flow causes the gradual development of portal hypertension, collateral vein formation, and shunting of blood to the systemic circulation. As portal hypertension develops, local production of vasodilators—mainly NO, but also carbon monoxide, glucagon, prostacyclin, AM, and endogenous opiates—increases, leading to splanchnic vasodilation. Other contributing factors to splanchnic vasodilation include intestinal bacterial translocation, proinflammatory cytokines, and mesenteric angiogenesis. ^,

Characteristics of hepatic blood flow.

(A) Hepatic circulation. I, The normal liver receives two-thirds of its blood flow from the portal vein (PV) and the remaining third from the hepatic artery (HA). II, Both the portal venules and hepatic arterioles drain into hepatic sinusoids, but the exact arrangement that allows forward flow of the mixed venous and arterial blood remains unclear. III, Cirrhosis increases intrahepatic vascular resistance and sinusoidal pressure. In addition, PV flow is markedly decreased, and HA flow is unchanged or increased. (B) Hepatic vascular hemodynamics and sodium balance. I, Cirrhosis or restriction of HV flow increases intrahepatic vascular resistance and sinusoidal pressure, markedly decreasing PV flow and increasing HA flow. Changes in the physical forces or in the composition of the hepatic blood trigger Na ⁺ retention and edema formation. II, Insertion of a side-to-side portocaval shunt decreases sinusoidal pressure and maintains mixing of PV and HA blood, irrigating the liver. Under these conditions and despite cirrhosis, there is no Na ⁺ retention. III, Insertion of an end-to-side portocaval shunt only partially decreases the elevated sinusoidal pressure and prevents mixing of PV and HA blood supplies, inasmuch as the PV blood is diverted to the inferior vena cava (IVC). Under these conditions and, despite normalization of PV pressure, Na ⁺ retention continues unabated.

Modified from Oliver JA, Verna EC. Afferent mechanisms of sodium retention in cirrhosis and HRS. Kidney Int . 2010;77:669–680.

The decreases in SVR associated with low arterial BP and high cardiac output account for the well-known clinical manifestations of the hyperdynamic circulation commonly seen in patients with cirrhosis. These include warm extremities, cutaneous vascular spiders, wide pulse pressure, capillary pulsations in the nail bed, and pulmonary vasodilation, associated with the hepatopulmonary syndrome.

The NO system is integrally involved in the pathogenesis of the hyperdynamic circulation and Na ⁺ and water retention in cirrhosis, as well as in hepatic encephalopathy, hepatopulmonary syndrome, and cirrhotic cardiomyopathy. NO is produced in excess by the vasculature of different animal models of portal hypertension and in cirrhotic patients. In animal models, the increased production of NO can be detected at the onset of Na ⁺ retention and before the appearance of ascites and NO has been implicated in the impaired vascular responsiveness to vasoconstrictors. Moreover, removal of the vascular endothelial layer abolishes the difference in vascular reactivity between cirrhotic and control vessels.

Inhibition of NOS has beneficial effects in experimental models of cirrhosis and in cirrhotic patients. By reducing the high NO production to control levels, the hyperdynamic circulation in cirrhotic rats with ascites was corrected and accompanied by a marked increase in Na ⁺ and water excretion and regression of ascites. Concomitant decreases in PRA, aldosterone, and vasopressin concentrations were also observed. ^, In cirrhotic patients, the vascular hyporesponsiveness of the forearm circulation to NE could be reversed by NOS inhibition. Inhibition of NO production also corrected the hypotension and hyperdynamic circulation, led to improved renal function and Na ⁺ excretion, and caused a decrease in plasma NE levels in these patients. However, in patients with established ascites, NOS inhibition did not improve renal function.

The main enzymatic isoform responsible for the increased systemic vascular NO generation in cirrhosis appears to be eNOS in the systemic and splanchnic circulations. Upregulation of eNOS appears, at least in part, to be caused by increased shear stress as a result of portal venous hypertension with increased splanchnic blood flow. Increased NO release, as well as eNOS upregulation, in the superior mesenteric arteries was found to precede the development of the hyperdynamic splanchnic circulation. In accord with this concept, upregulation of hepatic eNOS (or neuronal NO synthase, nNOS) expression in rats with experimental cirrhosis was associated with a decrease in portal hypertension. However, mice with targeted deletion of eNOS alone or combined deletions of eNOS and inducible NO synthase (iNOS) can still develop hyperdynamic circulation in association with portal hypertension. This suggests that activation of other vasodilatory agents such as PGI ₂ , endothelium-derived hyperpolarizing factor, carbon monoxide, and AM may participate in the pathogenesis of hyperdynamic circulation in experimental cirrhosis.

In addition to eNOS, other isoforms may be involved in the generation of the hyperdynamic circulation and fluid retention in experimental cirrhosis. Increased expression of nNOS in mesenteric nerves may compensate partially for eNOS deficiency in eNOS knockout mice and reduce intrahepatic venous resistance and portal hypertension. Also, splanchnic vasodilation is modestly promoted, possibly by modulating neurogenic NE release. In contrast, the role of iNOS remains controversial; some researchers have shown increased iNOS in the superior mesenteric arteries of animals with experimental biliary cirrhosis but not in other forms of experimental cirrhosis. Specific iNOS inhibition led to peripheral vasoconstriction but had no effect on portal hypertension. iNOS is primarily regulated at the transcription level by many proinflammatory factors, principally nuclear factor-kappaB (NF-κB), which could be induced by endotoxin, from translocated intestinal bacteria. Interestingly, there is also an interaction between eNOS and iNOS in the vasculature in cirrhosis. Overexpression of eNOS in large arteries results in systemic hypotension and increased blood flow. These effects can be abrogated by activated iNOS in the small splanchnic vessels. Thus overall, available data indicate a predominant role for eNOS deficiency, with possible modulation by both nNOS and iNOS.

In marked contrast to the increased NO generation in the splanchnic and systemic circulation, NO production and endothelial function in the intrahepatic microcirculation are impaired in cirrhotic rats. The resulting paradoxical increase in intrahepatic vascular resistance is likely to result from contraction of myofibroblasts and stellate cells and mechanical distortion of the vasculature by fibrosis. A second mechanism for the increased intrahepatic vascular resistance could be that the locally decreased NO production shifts the balance in favor of local vasoconstrictors (ET, leukotrienes, thromboxane A2, Ang II). The increased vascular resistance may also play a role in the pathogenesis of intrahepatic thrombosis and collagen synthesis in cirrhosis.

Several cellular mechanisms have been implicated in the upregulation of splanchnic eNOS and downregulation of intrahepatic eNOS. Elevation in shear stress as a result of the hyperdynamic circulation and portal hypertension has already been mentioned and is generally consistent with this well-documented mechanism for upregulating eNOS gene transcription. However, eNOS activity is regulated not only transcriptionally but also posttranscriptionally, by tetrahydrobiopterin (THB ₄ ) and direct phosphorylation of eNOS protein. Furthermore, circulating endotoxins may increase the enzymatic production of THB ₄ , thereby enhancing mesenteric vascular eNOS activity.

Potential contributors to intrahepatic eNOS downregulation include interactions with caveolin, calmodulin, heat shock protein 90, eNOS trafficking inducer, disorders of GC activity, and increased levels of the NO inhibitor, ADMA. In fact, ADMA levels correlate with the severity of portal hypertension during hepatic inflammation and levels are higher in patients with decompensated than compensated cirrhosis. Raised ADMA levels have been linked to reduced activity of dimethylarginine dimethylaminohydrolases (DDAHs) that metabolize ADMA to citrulline. Similarly, patients with alcoholic cirrhosis and superimposed alcoholic hepatitis have higher plasma and tissue levels of ADMA, higher portal venous pressures, and decreased DDAH expression. However, attempts at pharmacologic or genetic upregulation of DDAH to reduce ADMA levels and increase NO in experimental cirrhosis have not translated into improved management of decompensated portal hypertension.

In the final analysis, the relative importance of the various mechanisms involved in the reduced intrahepatic and increased splanchnic and systemic NOS activity in cirrhosis remains to be determined.

Endogenous cannabinoids are lipid-signaling molecules that mimic the activity of Δ9-tetrahydrocannabinol, the main psychotropic constituent of marijuana. N -arachidonoylethanolamide (anandamide) and 2-arachidonoylglycerol are the most widely studied endocannabinoids that bind the two specific receptors, CB ₁ and CB ₂ . Anandamide also interacts with the vanilloid receptor.

In animal models of cirrhosis, both CB ₁ and CB ₂ receptors and endocannabinoid production are greatly upregulated and anandamide caused a dose-dependent increase in intrahepatic vascular resistance, especially in the isolated perfused cirrhotic liver. The effect appeared to be mediated by CB ₁ receptor–enhanced production of COX-derived vasoconstrictive eicosanoids. Also, the CB ₁ receptor antagonist, rimonabant, in cirrhotic rats reversed arterial hypotension, increased hepatic vascular resistance, decreased mesenteric arterial blood flow and portal venous pressure, and prevented ascites formation. The reduction in splanchnic blood flow was enhanced by the vanilloid receptor antagonist, capsazepine. These findings indicate that the transient receptor potential vanilloid type 1 protein and the CB ₁ receptor have a dual role in the splanchnic vasodilation characteristic of cirrhosis.

Endotoxin was found to be a major stimulus for endocannabinoid generation in monocytes and platelets of cirrhotic animals. This pathway could operate in patients with advanced cirrhosis in whom elevated circulating endotoxin levels are frequently found. Endocannabinoid production could then trigger splanchnic and peripheral vasodilation, arterial hypotension, and intrahepatic vasoconstriction through activation of CB ₁ receptors in the vascular wall and perivascular nerves. The potentially favorable effect of CB ₁ receptor blockade on Na ⁺ excretion opens the possibility of pharmacologic modification of human HRS.

In summary, afferent sensing of volume in cirrhosis is characterized by increased intrahepatic vascular resistance and sinusoidal pressure, decreased portal venous blood flow, and increased hepatic arterial flow. Either changes in intrahepatic physical forces or in the composition of the mixed intrahepatic blood could initiate abnormal Na ⁺ retention and edema formation (see Fig. 13.5B ). Side-to-side portocaval shunt (currently performed by transjugular intrahepatic portovenous shunt insertion) prevents (if inserted before induction of cirrhosis) or corrects (if inserted after induction of cirrhosis) renal Na ⁺ retention. This outcome could result from decreases in sinusoidal pressure or maintenance of the mixing of portal venous and hepatic arterial blood perfusing the liver. In contrast, it diverts blood to the IVC but only partially decreases sinusoidal pressure and prevents mixing of portal venous and arterial hepatic blood supplies. Although portal venous pressure is normalized, Na ⁺ retention continues unabated (see Fig. 13.5B ). Consequently, end-to-side shunting is no longer used clinically.

Available data are most consistent with the view that the putative EABV sensor(s) in the hepatic circulation are pathologically activated in cirrhosis, failing to respond to the expanded ECF volume. The sensing mechanisms likely respond specifically to elevated hepatic venous pressure with increased hepatic afferent nerve activity. The relays for these impulses consist of two autonomic nerve plexuses, surrounding the hepatic artery and portal vein, respectively. These neural networks connect hepatic venous congestion to enhanced renal and cardiopulmonary sympathetic activity.

Occlusion of the IVC at the diaphragm was associated with increases in hepatic, portal, and renal venous pressures and resulted in markedly increased hepatic afferent nerve traffic and renal and cardiopulmonary sympathetic efferent nerve activity. Section of the anterior hepatic nerves eliminated the reflex increase in renal efferent nerve activity and hepatic denervation in dogs with IVC constriction increased urinary Na ⁺ excretion. This effect of hepatic denervation was shown to be mediated by the adenosine A1 receptor in cirrhotic rats.

Apart from the adenosine-mediated hepatorenal reflex, other currently undefined humoral pathways could provide an anatomic or physiologic basis for the primary effects of alterations in intrahepatic hemodynamics on renal function. Despite the wealth of information on hepatic volume sensing, the molecular identity, cellular location of the sensor, and what is sensed remain elusive.

Several mechanisms have been proposed to account for the development of relative hypovolemia. The first is disruption in normal Starling relationships governing fluid movement in the hepatic sinusoids. Unlike other capillaries, these are highly permeable to plasma proteins. As a result, partitioning of ECF between the intravascular (intrasinusoidal) and interstitial (space of Disse) and lymphatic compartments of the liver is determined predominantly by the ΔP along the length of the hepatic sinusoids. Obstruction of hepatic venous outflow promotes enhanced efflux of a protein-rich filtrate into the space of Disse and results in augmented hepatic lymph formation. ^, In parallel, vastly increased hepatic lymph formation is accompanied by increased flow through the thoracic duct. When the rate of enhanced hepatic lymph formation exceeds the capacity for return to the intravascular compartment via the thoracic duct, hepatic lymph accumulates as ascites, and the intravascular compartment is further compromised. As liver disease progresses, fibrosis around Kupffer cells lining the sinusoids renders the sinusoids less permeable to serum proteins. Under such circumstances, termed “capillarization of sinusoids,” a decrease in oncotic pressure also promotes transudation of ECF within the hepatic lymph space, as in other vascular beds.

Another consequence of intrahepatic hypertension is transmission of elevated intrasinusoidal pressures to the portal vein. This leads to expansion of the splanchnic venous system, collateral vein formation, and portosystemic shunting, resulting in increased vascular capacitance and diversion of blood flow from the arterial circuit. Both splanchnic and systemic vasodilation occur, and this has been attributed to refractoriness to the vasoconstrictive effects of hormones such as Ang II and catecholamines, although the mechanism remains unknown. Along with diminished hepatic reticuloendothelial cell function, portosystemic shunting allows various products of intestinal metabolism and absorption to bypass the liver and escape hepatic elimination. Among these products, endotoxins are thought to contribute to perturbations in renal function in cirrhosis, either due to intestinal bacterial translocation, stimulating the release of proinflammatory cytokines (e.g., tumor necrosis factor-α [TNF-α] and interleukin-6), secondary to the hemodynamic consequences of endotoxemia, or through direct renal effects.

Levels of conjugated bilirubin and bile acids may become elevated as a result of intrahepatic cholestasis or extrahepatic biliary obstruction. Bile acids directly decrease proximal tubular reabsorption of Na ⁺ , tending to promote natriuresis. This diuretic effect might contribute to the underfilling state in advanced cirrhosis. ^,

Hypoalbuminemia in advanced cirrhosis, either as a result of decreased synthesis by the liver or dilution caused by ECF volume expansion, could also contribute to the development of hypovolemia by diminishing COP in the systemic capillaries and hepatic sinusoids. In addition, tense ascites might reduce venous return (preload) to the heart, leading to reduced cardiac output and diminished arterial BP.

Other factors that may also adversely affect cardiac performance include diminished β-adrenergic receptor signal transduction, cardiomyocyte cellular plasma membrane dysfunction, and increased activity or levels of cardiodepressants, such as cytokines, endocannabinoids, and NO. Although the cardiac dysfunction, termed “cirrhotic cardiomyopathy,” usually is clinically mild or silent, overt HF can be precipitated by stresses such as liver transplantation or transjugular intrahepatic portosystemic shunt (TIPS) insertion. Finally, intravascular volume depletion in cirrhotic patients may be aggravated by vomiting, occult variceal bleeding, and excessive diuretic use, leading to cardiovascular collapse.

Table 13.5 summarizes the various causative factors contributing to underfilling of the circulation in patients with advanced liver disease. In summary, the early stages of compensated cirrhosis are characterized by increased plasma volume, which frequently antedates ascites formation. However, as cirrhosis progresses, EABV decreases, leading to increased neurohumoral activity (RAAS, SNS, and AVP) and severe Na ⁺ and water retention.

The efferent limb of volume regulation in cirrhosis is similar to that in HF, consisting of adjustments in glomerular hemodynamics and tubular transport mediated by vasoconstrictor/antinatriuretic forces (RAAS, SNS, AVP, and ET) and counterbalanced by vasodilator/natriuretic systems (NPs and PG). Tilting the balance toward Na ⁺ retaining forces leads to renal Na ⁺ and water retention, as in HF.

In recent years, measurements of BNP and NT-proBNP have largely superseded ANP as a biomarker of cirrhosis and portal hypertension. Nevertheless, the role of NPs in the pathogenesis of HRS was largely elucidated through studies on ANP, and these are summarized here. Plasma ANP is elevated in cirrhosis at all stages, irrespective of the EABV. ^, In the preascitic stage, increased plasma ANP may be important for the maintenance of Na ⁺ homeostasis, but with progression of the disease, resistance to the natriuretic action of the peptide develops. ^, The high levels of ANP mostly reflect increased cardiac release rather than impaired clearance. The stimulus for increased cardiac ANP synthesis and release in early cirrhosis is likely increased left atrial size caused by overfilling of the circulation, secondary to intrahepatic hypertension–related renal Na ⁺ retention.

In addition to elevated ANP, preascitic patients also had significantly elevated left and right pulmonary volumes, despite normal BP, PRA, aldosterone, and NE levels. High Na ⁺ intake in these patients resulted in weight gain and positive Na ⁺ balance for 3 weeks, followed by a return to normal Na ⁺ balance, thereby preventing fluid retention and the development of ascites. The factors responsible for maintaining relatively high levels of ANP during the later stages of cirrhosis, in association with arterial underfilling, may be related to a futile cycle of mutual interactions between vasoconstrictor/Na ⁺ -retaining and vasodilatory/natriuretic forces. The fact that ANP levels do not increase further as patients proceed from early to late decompensated stages of cirrhosis would be consistent with this explanation. Furthermore, infusion of Ang II mimicked the nonresponder state by causing patients with cirrhosis, who still responded to ANP, to become unresponsive. This Ang II effect occurred at proximal (decreased distal delivery of Na ⁺ ) and distal nephron sites to abrogate ANP-induced natriuresis and was reversible. The importance of distal solute delivery was confirmed using mannitol, which also resulted in an improved natriuretic response to ANP in responders but not nonresponders. ^, ANP resistance was also ameliorated by endopeptidase inhibitors, renal sympathetic denervation, peritoneovenous shunting, and orthotopic liver transplantation.

To summarize, ANP resistance is best explained by an effect of decreased delivery of Na ⁺ to ANP-responsive distal nephron sites (glomerulotubular imbalance caused by abnormal systemic hemodynamics and activation of the RAAS) combined with an overriding effect of more powerful antinatriuretic factors to overcome the natriuretic action of ANP in the medullary collecting tubule. The latter effect could result from decreased delivery, as well as permissive paracrine/autocrine cofactors, such as PGs and kinins.

BNP levels are also elevated in cirrhosis with ascites and, like ANP, its natriuretic effect is blunted in these patients. Plasma BNP levels may be correlated with cardiac dysfunction and severity of disease and may be of prognostic value in the progression of cirrhosis. ^, Plasma CNP levels in cirrhotic preascitic patients, although normal, were directly correlated with natriuresis and urine volume and inversely correlated with arterial compliance but not SVR. These data suggested that compensatory downregulation of CNP occurs in cirrhosis when vasodilation persists and that regulation of large and small arteries by CNP may differ.

In contrast with the preascitic stage, patients with more advanced disease and impaired renal function had lower plasma and higher urinary CNP levels than those with intact renal function. Moreover, urinary CNP was correlated inversely with urinary Na ⁺ excretion. In patients with refractory ascites or HRS treated with terlipressin infusion or TIPS (see “Specific Treatments Based on the Pathophysiology of Sodium Retention in Cirrhosis” later), urinary CNP declined and urinary Na ⁺ excretion increased 1 week later. ^, Thus CNP may have a significant role in renal Na ⁺ handling in cirrhosis.

Finally, Dendroaspis NP levels were found to be increased in cirrhotic patients with ascites but not in those without, and levels were correlated with disease severity. The significance of these findings remains unknown.