The 12th leading cause of death in the United States in 2006 was cirrhosis and chronic liver failure.1 An association between decompensated cirrhosis and renal failure was first described by Flint2 in 1863 and is referred to clinically as hepatorenal syndrome (HRS). HRS is a common complication of patients with end-stage cirrhosis with portal hypertension (HTN) and ascites. It is characterized by intense renal vasoconstriction in the presence of splanchnic and systemic vasodilation. The annual incidence for developing HRS is estimated at 8% for patients with both cirrhosis and ascites with a 40% probability of developing HRS within 5 years.3
Critical care nurses occasionally confront patient conditions that are not common. One such condition is HRS. Ongoing education about this complex syndrome may prevent complications and death. This article reviews the pathophysiology of chronic liver failure, portal HTN, and ascites as well as the possible mechanisms that predispose patients with end-stage cirrhosis to develop HRS. Diagnostic clues for differential diagnosis, current treatment, and nursing strategies to effectively care for the patient with HRS will be addressed. An understanding of the pathophysiologic and clinical manifestations of HRS by the health care team is essential for prevention, early detection, and management of this complex syndrome.
Pathophysiology
The presence of both liver failure and ascites is requisite to the development of HRS. Chronic liver failure is a silent disease whereby the individual remains asymptomatic until 80% to 90% of the liver parenchymal cells are destroyed. Major causes of liver failure include alcohol abuse (60%-70%), chronic hepatitis C (10%), obesity with nonalcoholic fatty liver disease (10%), and biliary obstruction (5%-10%).4 Cirrhosis results from chronic insult to the liver from these causes. Injury to the parenchyma in response to toxins and inflammation over time produces a progressive, irreversible, diffuse, fibrosing, and nodular condition known as cirrhosis that disrupts the normal liver architecture and liver function. Decompensated liver failure may result in complications such as ascites, spontaneous bacterial peritonitis (SBP), hepatic encephalopathy, variceal bleeding, and portal HTN. Clinical, laboratory, and radiographic data are suggestive of liver disease; however, a definitive diagnosis is made by liver biopsy. Skelly and colleagues5 reported a strong correlation between liver function tests that are 2 times normal values for 6 months and an abnormal liver biopsy. Abdominal ultrasonography provides a noninvasive, cost-effective, and reliable diagnostic procedure to identify cirrhosis as well as ascites.
Ascites
The presence of ascites in patients with cirrhosis is requisite to developing HRS. An understanding of how ascites develops provides a hint to the circulatory dysfunction that contributes to HRS. Patients with cirrhosis develop portal HTN from increased intrahepatic resistance. The rise in hepatic resistance is attributed to processes that produce obstruction and portal fibrosis, for example, nodular regeneration, contraction of perisinusoidal liver cells, and collagen deposits in the Disse space. Impaired antioxidant function may be responsible for the progressive, irreversible damage to the liver that contributes to the loss of normal architecture and parenchyma.6 An increased release of vasoconstrictive substances, endothelin-1 and cyclooxygenase-derived (COX) prostaglandins, from activated hepatic satellite cells occurs that further contributes to resistance and portal HTN.7 Nitric oxide (NO) is the major vasodilator substance released in response to portal HTN. An imbalance exists between the hyporesponsive and underproduced vasodilator and the hyperresponsive and overproduced vasoconstrictor substances that perpetuate portal HTN, leading to the development of ascites.
Portal HTN promotes the development of ascites, the accumulation of fluid and albumin in the peritoneal cavity (Figure 1). Increased hydrostatic pressure and disruption of epithelial junctions from metabolites allow fluid and colloidal osmotic particles to move into the interstitial space. Likewise, the liver is responsible for manufacturing colloidal substrates (eg, albumin, fibrinogen). Thus, the consequent low colloidal osmotic pressure in the capillary from albumin loss and reduced production results in a failure to pull fluid back into the capillary on the venous side for recirculation. The loss of fluid from the circulating blood volume creates a relative hypovolemia and triggers compensatory mechanisms to maintain cardiac output.
The debate continues whether HRS results from the pathophysiologic changes associated with advanced cirrhosis alone or whether a second hit occurs whereby a triggering event causes intravascular volume depletion. The most likely triggers are sepsis, SBP, acute alcoholic hepatitis, variceal hemorrhage, overdiuresis, or large-volume paracentesis. This second-hit hypothesis focuses the attention of health care professionals on surveillance and the implementation of preventative strategies to limit the development of HRS in this vulnerable population.
Circulatory Changes
Three physiologic processes contribute to circulatory changes in the renal and splanchnic vascular beds placing the patient with end-stage liver disease at risk for developing HRS. These processes include hypovolemia, systemic vasodilation, and a hyperdynamic circulation.8 The body's first response to counteract the effects of reduced perfusion to the regional beds resulting from hypovolemia is to stimulate the release of local and peripheral-acting vasodilator substances.
Nitric oxide, a well-studied signaling molecule in the body, is an endogenous vasodilating product of endothelial cells, neutrophils, and monocytes. It plays a direct role in both splanchnic and peripheral vasodilation. Likewise, NO indirectly promotes vasodilation by decreasing the sensitivity of peripheral receptors to vasoconstrictive substances such as catecholamines and angiotensin II.9 Increased levels of both NO and factors responsible for increased NO synthesis have been reported in patients with cirrhosis and have been linked to the development of a hyperdynamic circulation characterized by low peripheral and systemic vascular resistance (SVR) leading to increased cardiac output.
Similarly, prostaglandins, vasoactive intestinal peptide, glucagon, and calcitonin gene-related peptide have vasoactive properties (Table 1). These mediators acting alone or in combination may be responsible for the further decline in vascular reactivity by a decreased response to vasoconstrictor substances, allowing for the opening of preexisting or new anatomical shunts in the splanchnic beds.8
The totality of vasodilation of the splanchnic and renal beds, and ultimately peripheral circulatory beds, produces a decline in total peripheral vascular resistance and SVR. Reduced circulatory resistance results in underfilling of the systemic circulation. Thus, because of increased capacitance in the vasodilated systemic circulation, blood volume becomes inadequate to fill the vessels, and a relative state of hypovolemia develops. To counteract gross vasodilation from high levels of NO and other vasoactive substances, high plasma renin-angiotensin-aldosterone system (RAAS) activity and arginine vasopressin (AVP) release occur to increase both vascular resistance and volume by promoting salt and water retention.10
The proposed mechanisms to compensate for underfilling and circulatory dysfunction have been described by 2 theories, vasodilator theory11 and overflow theory.7,12 These theories provide insight into how key compensatory processes are initiated to deal with circulatory underfilling that ultimately contributes to dysfunction. Although these theories provide differing views on the onset of HRS, they overlap when describing the resultant compensatory feedback mechanisms. When these mechanisms fail to keep the dysfunctional processes associated with end-stage liver disease in check, the classic HRS findings of intense renal vasoconstriction in the face of profound splanchnic and peripheral vasodilation occur.
Compensation Mechanisms
Neurohormonal responses to fill the circulation caused by systemic and splanchnic vasodilation and fluid loss from the circulation by ascites involve activation of the sympathetic nervous system (SNS), RAAS, and renal autocoid system as well as secretion of AVP. All responses are aimed at restoring volume by vasoconstriction and promoting renal sodium and/or water retention. These responses become central to the evolving circulatory dysfunction that ends in HRS (Figure 2).
Both SNS and RAAS activities are significantly increased in patients who develop HRS.13 High-pressure arterial baroreceptors located in the carotid body and aortic arch respond to the low circulating blood volume by triggering the SNS (norepinephrine; epinephrine) and the RAAS (angiotensin II). Vasoactive mediators in the renal beds become overwhelmed by intense SNS and RAAS stimulation, resulting in preferential vasoconstriction of the outer cortex of the kidneys leaving it relatively hypoxic. Hypoxia exhausts membrane arachidonic acids and other essential molecules that generally lead to an activation of the autocoid system and the production of critical vasodilatory prostaglandins. Furthermore, the presence of the natural endothelial nitric oxide synthase (eNOS) inhibitor, asymmetric dimethylarginine, antagonizes NO-promoting vasodilation, thus producing profound renal vasoconstriction.10 Both SNS activation of the kidneys and the loss of vasodilatory effect produce intense renal vasoconstriction, resulting in a steep decline in renal blood flow and glomerular filtration rate (GFR).
To correct the relative hypovolemia state, aldosterone, a hormone product of RAAS activation, promotes the conservation of sodium and free water. Arginine vasopressin is also released from the posterior pituitary gland in response to increased plasma osmolality, decreased blood pressure (BP) and volume, and conditions sensed by stretch receptors in the heart and large arteries. A major function of AVP is water conservation. Arginine vasopressin increases the reabsorption of water in the distal convoluted and collecting tubules by inserting water channels (aquaporin-2 channels) into the apical membranes of the epithelial cells that line these tubules.14 In the patient with end-stage liver disease, there is hypersecretion of AVP and impaired metabolic clearance; both contribute to the impaired ability of the kidneys to excrete a water load. The increased half-life of AVP in patients with advanced cirrhosis results in the increased expression of vasopressin-regulating water channels in the cell membrane. The number of water channels directly correlates with the amount of ascites.15 As AVP accumulates, it becomes a potent vasoconstrictor via stimulation of the V1 receptors, inhibits sympathetic efferent stimuli, and potentiates baroreflexes. In turn, renal vasoconstriction triggers the renal autocoid system to produce COX-derived prostaglandin hormones to promote venodilation. With decompensation, renal vasoconstriction predominates.
One final group of hormones that influence water equilibrium involves natriuretic factors. One or more natriuretic factors (atrial natriuretic peptide, brain natriuretic peptide, C-type natriuretic factor) isolated from the systemic and regional circulation of persons with decompensated cirrhosis may be defective. All 3 natriuretic peptides are capable of promoting local sodium and water transport, thus altering renal hemodynamics in patients with cirrhosis.7,16 The neuroendocrine battle to expand volume eventually disrupts the delicate balance between vasoconstrictor and vasodilator substances, ending in the classic HRS findings of renal vasoconstriction.
Splanchnic beds do not encounter the same vasoconstrictive fate as the renal beds because of their enhanced production of local vasodilators. Circulating splanchnic vasodilators exert their effect on peripheral circulation, which becomes a driving force behind the development of a hyperdynamic circulation similar to what is seen with sepsis.
In summary, neurohormonal and SNS responses to correct underfilling in the vascular space are initially effective in increasing the circulating blood volume by improving the volume to vessel size relationship. However, when unchecked, the outcome is a hyperdynamic circulation with increased cardiac output and hypotension. The degree of hyperdynamic circulation present directly correlates with the extent of liver decompensation, placing the patient with advanced cirrhosis and ascites at risk for developing HRS and organ failure, for example, renal, heart, brain, and adrenal glands.17,18 The tentative balance between severely reduced SVR, profound SNS and RAAS activation, and release of vasoconstrictive and vasodilating substances can easily be disrupted by a second hit or insult to an already precarious circulatory system, for example, sepsis, overdiuresis, SBP, or large-volume paracentesis. A second-hit disturbance can result in the classic findings in HRS-profound renal hypoperfusion and splanchnic and peripheral vasodilation.13
Diagnostic and Prognostic Criteria for HRS
The International Ascites Club published a consensus statement to standardize the nomenclature and diagnostic criteria of syndromes associated with ascites and cirrhosis.19 Until recently, the diagnostic criteria for HRS were characterized according to major and minor criteria. Clinical findings listed under major criteria were essential for diagnosing HRS, whereas clinical findings listed under minor criteria were based on functional evidence of low GFR and sodium retention. Although minor criteria frequently accompanied HRS, their presence was not required for the diagnosis. The updated version (Table 2) does not subdivide diagnostic criteria as major or minor but rather reflects a blending of the most prognostic indices.
Hepatorenal syndrome is further delineated as type I (acute) or type II (chronic). Criteria separating type I HRS from type II HRS are reviewed in Table 3. More recently, type III and IV categorization have been proposed.20 Type III categorization defines patients with coexistent intrinsic renal dysfunction and advanced liver disease, whereas type IV includes patients with acute liver failure. Defining causation more specifically is designed to facilitate managing precipitating events as well as to tailor treatment strategies. Our discussion will be limited to the accepted type I and II categorizations of HRS.
Type I is characterized by the rapid progression of severe renal failure and failure of the compensatory mechanisms to maintain renal perfusion. Spontaneous recovery from type I HRS is rare. In 70% to 100% of patients with type I HRS, a precipitating event is identifiable, which is in line with the second-hit hypotheses.21 Likewise, it is possible to have more than 1 precipitating event occur for a single patient.22 Hospital survival rate has been reported as high as 40%23 to a low of 10% to 20%19,24 with a median survival time of 2 weeks.
Type II HRS is less aggressive, taking many weeks or months to develop and has been linked to the progression of cirrhosis with no apparent precipitating cause. In contrast to type I, the reduction in GFR and elevation of creatinine concentration in type II HRS are less severe, allowing for outpatient management.20 The eventual progression to acute renal failure (ARF) and conversion to type I should be anticipated, but recent evidence suggests that progression from type II to type I is uncommon.25 For patients with type II who develop type I HRS, however, a second hit or triggering event is responsible.20 Type II HRS has no consistent relationship to jaundice and has a median survival of 6 months, sufficient time for organ transplantation therapies if the patient meets transplant criteria.
The Child-Pugh Classification Scale and more recently the updated Child-Turcotte-Pugh Scoring System26,27 (Table 4) have been used with variable success to predict outcome related to liver failure and thus have been extended for use with the HRS patient. The Child-Pugh scale divides patients into classes by severity of liver function impairment based on the presence/ absence and/or degree of encephalopathy, ascites, bilirubin, albumin, and prothrombin time (international normalized ratio). Calculators for these classification systems are available online to facilitate scoring.28 A noted omission in this scale has been the lack of inclusion of renal failure, which may be a key contributing factor to outcome in patients with severe liver failure.
To enhance prognostic capability, the Model for End-Stage Liver Disease (MELD) score, which reflects kidney dysfunction and includes creatinine concentration, was added to the assessment.25 MELD calculators are readily available online16 and may be a useful adjunct for tracking changes over time and predicting patients with cirrhosis and ascites at risk for HRS.20,29 For patients with a MELD score approaching 18, Fernandez and colleagues36 reported that approximately 40% of patients developed HRS within 1 year. As a predictor of 3-month survival, a high score (40 or more out of 50) correlated with a less than 20% survival rate.30 The MELD scoring system, however, was found to underrate mortality when type I HRS was present. One possible explanation may be that creatinine values are underestimated in patients with hyperbilirubinemia, thus limiting the MELD's usefulness as a prognostic score in this cohort. The Child-Pugh score and response to treatment together are believed to be better predictors of survival for HRS patients.25,31 The Child-Turcotte-Pugh and MELD scoring systems are presented in Tables 4 and 5, respectively.
HRS Clinical Presentation
The clinical onset of HRS (type II) is usually insidious, progressing slowly over weeks or months with the classic prerenal features of low urine output in the absence of diuretics, generally benign urine sediment, a very low rate of urinary sodium excretion, dilutional serum hyponatremia, moderate hyperkalemia, and a rise in plasma creatinine concentration with a subsequent decline in creatinine clearance. Although serum creatinine is the standard clinical marker to evaluate renal function and GFR, it is not the ideal marker in patients with advanced liver failure because these patients have a lower production of creatinine clearance related to reduced muscle mass, lack of capacity for converting creatine to creatinine, and low protein intake.32 Hence, it is common for creatinine values in HRS patients to fall into reference range, for example, 1.0-1.3 mg/dL. Values within this normative range may be associated with a GFR between 20 and 60 mL/min (normal >100 mL/min). Thus, for patients with decompensated cirrhosis, the practitioner should expect to find a lower serum creatinine level than anticipated for any given GFR.
Unlike patients with moderate liver disease, most HRS patients have signs of advanced liver failure and portal HTN, particularly jaundice, coagulopathy, malnutrition, and hepatic encephalopathy. No consistent relationship has been noted between jaundice or the degree of hepatic encephalopathy and the development of HRS.33
Differential Diagnosis of HRS
Renal Indicators
The diagnosis of HRS is one of exclusion because various other conditions can cause ARF in the patient with advanced cirrhosis. Close attention during the history and physical examination is necessary to correctly identify causation. Hepatorenal syndrome is a prerenal disease, characterized by normal renal histology and initially preserved reabsorption capacity of tubular cells with the capacity to concentrate urine. Conditions that similarly produce a picture of prerenal ARF include severe dehydration from gastrointestinal losses (vomiting, diarrhea, nasogastric tube drainage, gastrointestinal or esophageal bleeding), renal losses from over-diuresis, septic shock, or SBP.20 If ARF is secondary to volume depletion, the infusion of a fluid challenge (1.5 L) should improve renal output, whereas there will be no improvement if HRS is present. Also, urinalysis results may be misleading because granular and epithelial cell casts are seen in marked hyperbilirubinemia.18,33
It becomes more difficult to separate the diagnosis of HRS from renal etiology if the patient develops acute tubular necrosis (ATN). Acute tubular necrosis of intrarenal origin can occur following the administration of nephrotoxic drugs or dyes or from intrinsic renal parenchymal disease, hypoxic/ischemic events, or an infectious cause. ATN presents as a rapid progressive rise in plasma creatinine concentration. Again, caution is needed when interpreting a blunted rise in creatinine concentration and the adequacy of GFR in the patient with chronic liver disease. Because of the variability in renal findings in patients with HRS, renal indices were not used initially as major diagnostic criteria,19 but some cellular changes to reflect renal function have been added.34 True HRS is characterized by the absence of factors causing volume depletion, for example, dehydration or sepsis and intrarenal conditions such as ATN, nephrotoxic exposure, or ischemia.
Electrolyte Status
Evaluating the electrolyte and acid-base balance of the patient with decompensated cirrhosis may also assist with the identification of HRS (Table 6). Hyponatremia is universal to HRS and likely results from the retention of excess water from high circulating levels of AVP. Thus, in the patient with cirrhosis and renal failure, the finding of a normal serum sodium level should prompt the health care practitioner to look for a different cause of ARF. Hyperkalemia is associated with renal failure. While the HRS patient's potassium level is elevated, it is usually in the moderate range and is not useful for confirming the diagnosis of HRS. Finally, the presence of severe metabolic acidosis is uncommon in HRS unless it is precipitated by severe infection.33
Other Indicators
Other distinguishing assessment and clinical findings to evaluate HRS are the presence or absence of pulmonary edema, infection, or ascites. Pulmonary edema, a common finding in patients with severe ARF, is rare in the HRS patient because of an already-overdilated systemic vasculature (reduced afterload results in decreased preload and relative hypovolemia). For the HRS patient who has not been aggressively treated with volume and/or plasma expanders, filling the circulatory system with the fluid bolus will not cause adverse pulmonary effects.35
Infectious Sources
Susceptibility to infection is increased for persons with ARF and advanced liver disease and with sepsis, a common complication. SBP leads to renal impairment in approximately 10%-30% of patients with cirrhosis.36 An estimated 28% of patients with SBP will develop HRS, despite appropriate treatment with nonnephrotoxic drugs.37 The 2 most common causative organisms are gram-negative bacteria- Escherichia coli or Klebsiella.38 In addition to SBP, cellulitis, urinary tract infection, sepsis, and pneumonia may precipitate HRS. The health care team must be vigilant in monitoring for infection because delayed treatment results in the release of vasorelaxant proinflammatory endotoxins and cytokines that worsen both the underlying hemodynamic circulatory condition and renal function. Septic shock may be more difficult to rule out in the patient with HRS because of the subclinical presentation of symptoms of bacterial infection in patients with cirrhosis.
Because arterial hypotension due to sepsis can mimic reductions in BP during the early stages of advanced liver disease, BP is an unreliable indicator of HRS. Hemodynamic changes from systemic vascular vasodilation in the patient with advanced cirrhosis result in the "normal" BP being low, but stable and mean arterial pressure (MAP) 60 to 65 mm Hg is reduced from a premorbid baseline of perhaps 75 to 80 mm Hg.39 This drop in MAP is more significant if the patient has a preexisting history of HTN. The presence of hemodynamic instability and a severe reduction in BP suggest an infectious complication. Conversely, HTN more likely suggests the presence of glomerulonephritis. Finally, the absence of ascites in a patient with cirrhosis and renal failure argues against the diagnosis of HRS.
Treatment
A 95% 30-day mortality rate has been reported for patients with type I HRS.3,8 Spontaneous recovery from HRS is rare, and a quick, accurate, differential diagnosis is essential. Two courses of treatment have been proposed for HRS. The first centers on fluid and pharmacologic management to reduce circulatory dysregulation and the second focuses on correcting liver function via surgical shunts or transplantation.
Fluid Management
The treatment goal for fluid balance is to achieve diuresis without further loss of sodium, while maintaining adequate circulating volume and hemodynamic stability. In conjunction with optimizing fluid status to fill excess vessel capacitance and achieve hemodynamic stability, it is necessary to vasoconstrict the splanchnic beds to enhance the circulating blood volume without overconstriction of the renal or systemic vasculature. The BP should be optimized by maintaining the MAP at 10 to 15 mm Hg above BP measured on presentation or until urine output increases (0.5 mL/kg/h). Intravenous fluid replacement solutions vary on the basis of the patient serum sodium level with a focus on preventing additional sodium retention. Isotonic sodium chloride is preferred to hypertonic solutions that may cause a rapid return of interstitial fluid to the vascular space leading to pulmonary edema. More recently, the recommendation for fluids is away from sodium chloride because it can escape the vascular compartment in patients with severe hypoalbuminemia (<2.0 g/dL). Albumin is recommended for volume expansion in patients with hypoalbuminemia and should be administered at 1 g/kg body weight not to exceed 100 g.20,40 Central venous pressure (CVP) trends can be used to guide fluid replacement. One recommendation is to maintain the CVP between 8 and 12 cm H2O or, where pulmonary artery monitoring is available, a pulmonary artery wedge pressure between 14 and 19 mm Hg.41
Performing paracentesis is recommended if the decline in renal function is linked to a reduction in renal venous pressure from ascites compression. Only for large-volume paracentesis (>5 L) is albumin replacement recommended to maintain intravascular volume and hemodynamic stability (albumin replacement 8-10 g/L removed).42 Improvement of renal function postparacentesis is usually transient because of the unrelenting pathophysiologic mechanisms responsible for ascites. Initially, diuretics should be stopped and their use reevaluated for the HRS patient. If deemed relevant to the treatment plan, diuretics should be used with caution to avoid further reduction of the circulating blood volume. Should overdiuresis occur, it is usually reversible by discontinuing diuretic administration.20
Hemodialysis is of limited use in the patient with HRS because of hemodynamic instability and the potentially dangerous outcome of severe hypotension. If the MAP is sufficient, continuous arteriovenous or venous-to-venous hemofiltration may be preferred because of reduced hemodynamic swings. Situations in which hemodialysis may be needed in HRS treatment include the development of acidosis, hyperkalemia, uremic symptoms, or volume overload.20 Peritoneal dialysis has been rated minimally effective and compounds the risk of hemorrhage and infection (SBP) in an already-compromised system.
Pharmacologic Management
The approach to pharmacologic management of HRS is based on underlying pathophysiologic mechanisms. The treatment goal is to restore systemic and splanchnic vasoconstriction, while promoting renal vasodilation, balanced sodium, and euvolemia. Combinations of peripheral vasoconstrictors and plasma volume expanders have been most effective in reversing the extreme vasodilation and decreased effective blood volume characteristic of HRS. Administering a vasoconstrictor agent such as norepinephrine or neosynephrine in combination with albumin improved renal function and survival in comparison to vasoconstrictor therapy alone.43
Because of the lack of availability of Food and Drug Administration-approved drug therapies for treating HRS in the United States, combination therapy with albumin and a vasoconstrictor drug offers the best chance to reverse or delay the lethal progression of HRS. Screening for a history of coronary artery disease, cardiomyopathies, cardiac or respiratory failure, HTN, vascular disease, and asthma prior to initiation of vasoconstrictor therapy should be done and a decision to use this approach be based on a risk-to-benefit analysis. Recommendations for the administration of norepinephrine, albumin, and the European-available AVP analogue terlipressin are presented in Table 7. Continued therapy is recommended until HRS is reversed or for a maximum of 14 days.20 Of note, small sample sizes have been reported in study trials for combination therapy (vasoconstrictor therapy plus albumin). The success of combination therapy for type I HRS patients was defined by extended survival time until liver transplantation. Combination therapy can result in a reduced creatinine level and a change in the MELD score, which is heavily weighted on the creatinine concentration. The resultant reduced MELD score potentially changes the patient's United Network for Organ Sharing (UNOS) categorization and time to transplant. Advocacy by health care workers is needed to reflect the continued medical urgency for assigning a higher MELD score for the HRS patient. Thus, combination therapy is particularly beneficial for type I HRS patients who often do not survive long enough for transplantation.
The administration of angiotensin-converting enzyme inhibitors to limit the production of angiotensin II, which causes renal vasoconstriction and ischemia, is not recommended to treat HRS patients because of the potential adverse effect of systemic hypotension (at high doses), reduced GFR caused by preferential efferent arteriolar dilation, and a decrease in intraglomerular pressure.7 Other contraindications for using vasoconstrictor therapy in HRS patients include the presence of coronary artery disease, cardiomyopathies, cardiac arrhythmias, arterial HTN, cerebrovascular disease, peripheral vascular disease, bronchospasm, terminal liver disease, or age greater than 70 years.20
Additional pharmacologic treatments of HRS revolve around antagonizing the endogenous effects of renal vasoconstrictors. The use of agonists to promote renal dilation singly or in conjunction with volume expanders, primarily albumin, has been studied. European drug trials report using AVP analogues (eg, terlipressin, unavailable in the United States) as preferential vasoconstrictors of the splanchnic vasculature with an added effect of reducing both renin and angiotensin levels, thereby constricting the splanchnic circulation and increasing renal perfusion. Terlipressin was found to increase GFR and BP, while reducing creatinine and neurohormonal levels in 77% of cases (n = 42). When terlipressin therapy was related to the Child-Pugh score, patients with a score of 13 or less had improved renal function, whereas patients with higher Child-Pugh scores related to more severe cirrhosis did not have improved renal function.21
Other trials using AVP analogues alone (ornipressin, terlipressin) were minimally effective in reversing HRS, despite increasing urine output and creatinine clearance,44 but, when used together, the 2 AVP analogues improved renal function and 30-day survival.45 The use of AVP analogues has been associated with serious ischemic adverse effects in up to 10% of HRS patients, and AVP analogues should be given with extreme caution. When AVP analogues were given in conjunction with adequate volume expansion using albumin, a significant reduction in serum creatinine level was noted.41,46 Response to fluid and pharmacologic therapies provided a temporary respite from HRS, but without transplantation, 80% of patients died within 3 months.21
Renal vasodilators have also been studied in HRS. Oral and intrarenal administration of the prostaglandin analogue misoprostol was performed to correct renal cortical ischemia.47 Of studies using prostaglandins, no change in GFR or sodium excretion occurred.24,48 Low-dose infusion of dopamine produced increased renal cortical blood flow with no increase in GFR or urinary output in patients with ascites or HRS.49
In summary, vasoactive drugs used to treat HRS are shown to be variably effective. Drug therapies in combination with adequate volume replacement play a role in the short-term management of the HRS patient by serving as a bridge to transplantation.
Surgical Management
Surgical treatment of decompensated liver failure is key to preventing and managing HRS. However, early diagnosis is essential for the patient to be able to receive a liver transplant. Of patients awaiting liver transplantation, between 10% and 48% will develop HRS.3,50 Surgery offers hope for reversal of renal failure by improving hepatic function and, thus, partial resolution of the primary disease. Available surgical options include shunts for portal decompression, liver transplantation, or combined liver-kidney transplantation.
Shunts
The transjugular intrahepatic portosystemic shunt (TIPS) is used to manage refractory ascites of cirrhosis by noninvasively reducing portal HTN. A tract is established across the liver tissue that is stented to ultimately connect the portal and hepatic veins within the hepatic parenchyma. A TIPS procedure is performed by an interventional radiologist. The reduction in pressure produced by TIPS has been associated with improved renal function via improved flow and the subsequent reduction in SNS and RAAS activities and neurohormonal factors.51 TIPS has improved survival rates and allowed some patients to stop hemodialysis.52 However, TIPS is not useful for the majority of patients with type I HRS, because it may further reduce portal perfusion and lead to worsening liver function and GFR.25
Transplantation
Pharmacologic and surgical therapies described thus far are palliative and serve as a bridge to transplantation. Use of these therapies must be justified by investigating whether the patient is a liver transplant candidate. Orthotopic liver transplantation is the only treatment modality for permanent reversal of HRS.27 Varied survival rates for HRS patients have been reported posttransplant with the best being a 4-year survival of approximately 60%.53 This study reported that pretransplantation, neither the duration of HRS nor the use of hemodialysis was adversely associated with outcome. Poorer outcome was associated with liver disease from chronic alcoholism and the need for posttransplant dialysis.54 Because prolonged transplant waiting time is associated with progressive renal damage and renal failure adversely affects outcome post-liver transplantation, the current thought is to perform a combined liver and kidney transplant. In HRS patients receiving dialysis for more than 8 weeks, combined liver and kidney transplant improved both survival and use of hospital resources, whereas it offered no survival advantage over liver transplant alone for patients receiving dialysis less than 8 weeks.55
The benefits of transplantation must always be weighed against the availability and need for organs by other candidates. Ethical considerations relevant to patient characteristics and history as well as the potential for survival should be reviewed by the ethics committees in conjunction with the family.
Nursing Care
Monitoring the HRS patient is a nursing challenge. Prevention of complications due to alterations in hemodynamic status may be critical to preventing HRS. Attention to history findings such as recent exposure to nephrotoxic drugs and angiographic studies involving dye contrast, paracentesis, gastrointestinal or esophageal bleeding, infection, and nutritional status is important because these findings represent potential triggers for developing HRS (second-hit hypothesis).
In addition to the diagnostic findings associated with HRS (Table 2), surveillance of other laboratory and diagnostic findings is necessary (Table 6). When monitoring the patient with end-stage liver failure and ascites for HRS, it is important to recall that a prerenal pattern of renal failure is expected. The serum creatinine concentration may be underestimated in this cohort of patients. Liver enzymes, aspartate aminotransferase and alanine aminotransferase, may also be normal or low in advanced liver disease. Early signs of infection may be subclinical. Because total SVR is severely reduced despite profound activation of SNS and vasoconstrictors from the RAAS, BP is usually low but stable (MAP [equivalent to] 70 mm Hg). In the absence of a pulmonary artery catheter, the diastolic BP provides an indicator of SVR and vessel size; for example, low diastolic BP indicates vasodilation and low SVR. A relationship between the degree of arterial hypotension in cirrhosis and the severity of hepatic dysfunction, decompensation, and survival has been shown.39 The presence of MAP instability points toward infectious complications. See the "Case Study" section for an example of diagnosis and ongoing assessment of an HRS patient.
Medication Monitoring
In the presence of reduced renal function, all drugs should be started at the adjusted renal dosage range and titrated to therapeutic effect while continuously monitoring for toxicity. To prevent ATN, the nurse should exercise caution in administering drugs with nephrotoxic and hepatotoxic adverse effects, particularly aminoglycoside antibiotics and radiocontrast agents.56 If radiocontrast agents are required in the course of treatment, oral acetylcysteine or intravenous fenoldopam in conjunction with saline hydration can be administered prior to the test to protect the kidneys from nephrotoxic effects.57
Other categories of drugs that pose a risk for renal dysfunction are antivirals and nonsteroidal anti-inflammatory drugs because of their effects on the COX-derived prostaglandin synthesis pathway. Marked elevations in prostaglandins to preserve vasodilation have been reported in the patient with decompensated cirrhosis.58 The main COX isoforms are nonsteroidal anti-inflammatory drugs. COX-1 (aspirin, ibuprofen, naproxen) and COX-2 (celecoxib) medications should be avoided to prevent interference with prostaglandin synthesis and avoid profound renal hypoperfusion and hypoxia59 (Figure 3). Discontinuing COX inhibitors usually is sufficient for restarting prostaglandin synthesis.
From a hemodynamic standpoint, other medications should be evaluated for their effects on the HRS patient. Diuretics should be discontinued, if HRS is suspected, to avoid hypovolemia. The decision to restart diuretics should be guided by the CVP reading (goal: 8-12 cm H2O) or, where pulmonary artery monitoring is available, a pulmonary artery wedge pressure between 14 and 19 mm Hg (goal).41 If diuretics are restarted, the dosage should be reduced and BP and overall fluid status carefully monitored. Angiotensin-converting enzyme inhibitors, nitrates, and other vasodilators should also be withheld to prevent further dilation of an already overdilated compartment. The hemodynamic goal is to maintain the MAP near 70 mm Hg and to achieve a euvolemic state.
Infection Prevention
The increased risk of infection dictates strict adherence to universal precautions and careful monitoring of catheter sites and wounds. Prophylaxis and treatment of bacterial infection should include the use of nonnephrotoxic antibiotics. Prescribed treatments should be promptly initiated to avoid worsening of hyperdynamic circulatory status and renal function related to bacterial endotoxin and inflammatory cytokine production. After gastrointestinal hemorrhage, short-term antibiotic prophylaxis is recommended.60 A randomized study of patients with cirrhosis treated with cefotaxime plus albumin at the diagnostic point of SBP found a decreased incidence of type I HRS and a decreased hospital mortality rate.61 For a polymorphonuclear leukocyte count greater than 250 cells per mm3 in ascites fluid, antibiotic prophylaxis (cefotaxime) is recommended to prevent SBP. Post-SBP survival, a regimen of long-term prophylaxis is suggested.60 Continued monitoring for treatment effectiveness is required as cefotaxime-resistant gram-negative organisms continue to increase (from 7% to 28%).62
Fluid and Nutrition
Equally important is attention to nutritional and fluid needs. Up to 60% of patients with decompensated liver disease experience protein-calorie malnutrition and micronutrient deficiencies.63 Despite patients with advanced liver disease being protein malnourished, a diet high in protein results in hepatic encephalopathy and worsening of existing metabolic abnormalities. Likewise, a diet too low in protein can produce a paradoxical increase in plasma amino acids that mimics a high-protein diet because of increased gluconeogenesis and muscle catabolism. The current recommendation is to maintain protein intake at normal levels unless severe malnutrition is present. The use of enteral nutrition as a choice of nutritional support has been questioned for the patient with end-stage liver disease because of susceptibility to bleeding from nasogastric tube irritation to esophageal and gastric varices. The presence of coagulopathy, abdominal venous collaterals, or ascites also requires careful consideration when making the decision to tube-feed.64 The altered level of consciousness resulting from rising ammonia levels with hepatic encephalopathy makes oral feeding unrealistic.
Nutritional supplements for patients experiencing liver disease frequently include branch chain amino acids, low aromatic and ammonia-forming amino acids, high calorie-to-nitrogen ratio, low total fat, and multivitamins with or without trace minerals and key micronutrients. Supplementation of vitamins and minerals, particularly zinc, is necessary, because of renal losses, gastrointestinal malabsorption, and anorexia.65 Close collaboration with a dietitian is vital for assessing and monitoring the nutritional needs of this complex patient condition. While complete nutrition may not be tolerated in patients with tense ascites, some enteral feeding may be possible to preserve the intestinal villi and prevent bacterial translocation. Precautions to take when enteral feeding include elevating the head of the bed 30[degrees] to minimize aspiration, limiting the formula hang time, and changing the bag daily to avoid sepsis and pneumonia. Limiting the use of gastric acid suppression therapies to prevent bacterial overgrowth may be useful.66
A fluid-restricted diet is suggested if serum sodium levels are 120 to 125 mEq/L or less. Sodium restriction at or below 2000 mg/d and diuretics (as necessary), particularly oral spironolactone and furosemide, are first-line management strategies.60 Spironolactone, the aldosterone antagonist, is given in conjunction with furosemide to limit activation of aldosterone caused by the loss of sodium from this loop diuretic.67 Ensuring adequate nutritional and fluid status by monitoring calories, urinary output, weight, BP, and evaluation with replacement of electrolytes may reduce the risk for HRS.
Summary
Because patients with advanced liver disease are at increased risk for developing renal insufficiency from other causes, the HRS diagnosis is often one of exclusion. When diagnoses are made by exclusion, continuous reassessment and surveillance are necessary to avoid delay in identifying and treating the underlying condition. An understanding of the pathophysiology of end-stage liver disease and neurohormonal alterations that underlie HRS may prompt health care professionals to suspect HRS and lead to prompt interventions to improve patient outcome.
Case Study
J.T., a 55-year-old man with shortness of breath, fatigue, weakness, and anorexia, was admitted through the emergency department to the critical care unit. On physical examination, he had slow, slurred speech and difficulty following simple commands. He demonstrated asterixis and ataxia. He was oriented to person. The history was obtained from J.T.'s wife, who reported that he had a 29-year history of alcohol abuse, with no alcohol consumption for the past 5 years. She reported that J.T. had paracentesis 2 weeks ago, and 5 L of yellow fluid was removed. Following the fluid removal, J.T. had "low BP" and received albumin and fluid replacement. She stated that diuretics were resumed, but J.T.'s abdomen was nearly as large as it was 2 weeks ago. The wife stated that they had been considering the TIPS procedure, which J.T.'s doctor had recommended performing 3 months previously to manage J.T.'s refractory ascites. His current medication list was as follows: neomycin 500 mg 4 times/d; lactulose 2 tablespoons/d (20 g); spironolactone 200 mg/d; furosemide 100 mg/d; zinc 600 mg/d; and sodium benzoate 5 g/d. He had been following a diet containing 80 g protein/d.
Vital signs were as follows: temperature 36.7[degrees]C, heart rate (HR) 110 beats per minute, respiratory rate 36/min, BP 96/80 mm Hg, and MAP 86 mm Hg. Serum laboratory values are reported in Table 8. J.T.'s electrocardiogram revealed sinus tachycardia with PACs and tall, tent-shaped T waves. J.T. was malnourished and weighed 54 kg. He was classified to have a grade II encephalopathy on the basis of his drowsiness, disorientation, and poor short-term memory. Two weeks prior, his liver ultrasonography indicated a small nodular liver and ascites. A Doppler flow study revealed a significant decline in portal circulation. The diagnosis at the time of his admission to the intensive care unit showed respiratory distress, altered level of consciousness, chronic liver failure with tense ascites, and rule out HRS.
Paracentesis was performed with removal of 6 L of yellow fluid. J.T.'s vital signs postparacentesis were as follows: temperature 36.9[degrees]C, HR 88 beats per minute, respiratory rate 20/min, BP 80/64 mm Hg, and MAP 69 mm Hg. Given his previous episode of hypotension and hypovolemia, a pulmonary artery catheter was inserted through the internal jugular vein, and a urinary catheter was inserted to monitor his fluid status. Hemodynamic values are reported in Table 9. J.T. had no history of esophageal varices bleeding. An orogastric tube was placed for nutritional support, and home medications were resumed with the exception of furosemide. Hepatic formulation tube feeding was started at 30 mL/h with an order to increase the rate 10 mL/h every 4 hours for residual volume less than 200 mL, with a target goal of 60 mL/h (60 g of protein/24 h).
J.T. was given 50 g of albumin intravenously postparacentesis (recommended: 8-10 g/L fluid removed). Preliminary analysis of the ascites fluid revealed a polymorphonuclear leukocyte count of less than 150 cells per mm3. Antibiotic therapy was not initiated on the basis of a polymorphonuclear leukocyte count of less than 250 cells per mm3.
Twelve Hours After Admission
During the evening, J.T. became more agitated. His BP was 80/52 mm Hg with a MAP of 61 mm Hg. His urine output had been approximately 15 mL for each of the past 3 hours (urine output < 0.5 mL/kg; 54-kg body weight; anticipated output 27 mL/h). A norepinephrine drip was started at 0.1 mcg/kg/min with a target goal to increase MAP by 10 mm Hg. Four hours later, J.T.'s MAP was 68 mm Hg, and the norepinephrine dose was increased by 0.05 mcg/kg/min. Orders were written to titrate the norepinephrine drip by 0.05 mcg/kg/min every 4 hours up to a maximum dose of 0.7 mcg/kg/min to maintain MAP between 70 and 75 mm Hg. J.T. was continuously monitored for HTN, dysrhythmias, and peripheral vascular complications. Urinary output remained 15 to 20 mL/h.
Day 1
J.T.'s type I HRS was diagnosed after the large-volume paracentesis (second-hit model). J.T. met the criteria for HRS as evidenced by a continued rise in creatinine level, low urine output, absence of shock, normal renal ultrasonography, and absence of microhematuria and proteinuria. His liver and renal status continued to decline as evidenced by trend changes in bilirubin, aspartate aminotransferase, alanine aminotransferase, international normalized ratio, and creatinine concentration (Table 8).
Continuous venovenous hemodialysis was initiated to reduce uremia and volume overload. The MELD score was computed using an online calculator on the basis of J.T.'s serum laboratory values. When using this calculator, laboratory values less than 1.0 are set to 1.0 for MELD score calculation.
On the basis of the prognostic indicators, MELD score, and the Child-Turcotte-Pugh Scoring System, the physician explained to J.T. and his wife that a liver transplant was the only treatment for reversing his HRS. A MELD score more than 15 classified J.T. as having liver status 1A or 1B on the UNOS. J.T.'s wife asked the nurse why the doctor had changed his mind about performing the TIPS procedure. The nurse explained that the TIPS procedure was contraindicated because of J.T.'s HRS and advanced liver disease (eg, severe jaundice and hepatic encephalopathy) and might worsen perfusion to the liver. J.T. was added to the UNOS patient waiting list on the basis of his mortality risk score (MELD). Continued dialysis reduced J.T.'s MELD score, which was heavily weighted on creatinine. The health care team continued reassessment and recertification because the MELD could be used no longer than 48 hours to maintain transplant status.
Days 2 to 3
The norepinephrine drip was continuing to maintain MAP greater than 70 mm Hg in order to support liver and renal perfusion. Albumin 25 g was ordered every 8 hours to keep the CVP between 10 and 15 cm H2O. Continuous venovenous hemodialysis was continued to reduce creatinine level. A reduction of approximately 30% in creatinine level had occurred since the initiation of therapy.
During the evening of day 3, J.T. developed an acute drop in MAP (<60 mm Hg) and became febrile (38.3[degrees]C). Despite titration of the norepinephrine drip and albumin and fluid replacement, his BP remained labile. His wife was informed that J.T. would be removed from the UNOS patient waiting list and that J.T. was ineligible for a liver transplant unless his condition could be stabilized. After discussing possible benefits of continued therapy with the health care team, she agreed to make J.T. "Do Not Resuscitate, Comfort Care Arrest" status.
Acknowledgment
This work was supported by an NIH:NINR Mentored Scientist Award (KO1 NR009787-01) to the first author.
REFERENCES