Acute kidney injury (AKI) is frequent in patients with cirrhosis. AKI and hyponatraemia are major determinants of the poor prognosis in advanced cirrhosis. The hepatorenal syndrome (HRS) denotes a functional and potential reversible impairment of renal function. Type 1 HRS, a special type of AKI, is a rapidly progressive AKI, whereas the renal function in type 2 HRS decreases more slowly. HRS is precipitated by factors such as sepsis that aggravate the effective hypovolaemia in decompensated cirrhosis, by lowering arterial pressure and cardiac output and enhanced sympathetic nervous activity. Therefore, attempts to prevent and treat HRS should seek to improve liver function and to ameliorate arterial hypotension, central hypovolaemia and cardiac output, and to reduce renal vasoconstriction. Ample treatment of HRS is important to prevent further progression and death, but as medical treatment only modestly improves long-term survival, these patients should always be considered for liver transplantation. Hyponatraemia, defined as serum sodium <130 mmol/L, is common in patients with decompensated cirrhosis. From a pathophysiological point of view, hyponatraemia is related to an impairment of renal solute-free water excretion most likely caused by an increased vasopressin secretion. Patients with cirrhosis mainly develop hypervolaemic hyponatraemia. Current evidence does not support routine use of vaptans in the management of hyponatraemia in cirrhosis.
In patients with chronic liver disease, the outcome and the course of the disease are largely determined by the development of complications . They include bleeding from oesophageal varices and dysfunction of multiple organ systems such as the brain, the lungs, the heart and the kidneys [2, 3]. Chronic kidney disease (CKD) is seen in about 1% , whereas acute kidney injury (AKI) often develops in decompensated patients with ascites and predicts short-term mortality [5-7]. Recently, there has been considerable focus on acute-on-chronic liver failure as a distinct syndrome in patients with acute decompensation . The hepatorenal syndrome (HRS) is a special type of acute or subacute kidney failure that occurs in advanced cirrhosis characterised by functional impairment of the kidneys owing to vasoconstriction of the renal arteries with preserved tubular function and near-normal renal histology . Clinically, HRS is characterised as a prerenal failure not responding to volume expansion . HRS should be considered a diagnosis of exclusion of other causes of AKI in patients with cirrhosis . Recent data confirm that renal failure is associated with a high mortality and that more than 50% of the patients die within 1 month. Further progression of renal failure is associated with a seven-fold increased risk of death within 1 year [6, 12]. In addition, in cirrhotic patients with imminent renal failure, hyponatraemia is a common concomitant finding [13, 14].
This study aims to review recent advances in our understanding of the pathophysiology of acute and chronic kidney injury behind modern treatment and preventive strategies of HRS in cirrhosis and ascites.
Cirrhosis: a multiple organ syndrome
A preferential splanchnic arteriolar vasodilatation seems to be a key factor for the development of complications in cirrhosis [2, 15, 16]. It may be brought about by a combination of overproduction of circulating vasodilators, primarily of intestinal or systemic origin. The arterial vasodilation leads to reduction in the systemic vascular resistance, central arterial underfilling with effective hypovolaemia, activation of vasoconstrictor systems, such as the sympathetic nervous system (SNS), the renin-angiotensin-aldosterone system (RAAS), vasopressin and endothelins (ETs) [2, 17, 18]. The haemodynamic consequences are development of a hyperdynamic circulation with increased cardiac output and heart rate and low systemic vascular resistance and arterial blood pressure [2, 18]. Among potential vasodilators, research has focused especially on substances like nitric oxide (NO) and adrenomedullin, but natriuretic peptides, cytokines, hydrogen sulphide, ETs and endocannabinoids have also been implicated [16, 19-21]. The result is a multiorgan failure with disturbed function of the organs such as the lungs (the hepatopulmonary syndrome), the heart (cirrhotic cardiomyopathy) and the kidneys (HRS) [22-24]. For example, in the kidneys, vasoconstriction prevails and plays a pivotal role along with the development of hepatic decompensation. Liver dysfunction, central hypovolaemia, arterial hypotension and neurohumoral activation of particularly the RAAS and SNS with renal vasoconstriction are of major importance for the development of HRS [9, 20, 25, 26].
Renal dysfunction in cirrhosis
Acute kidney injury in cirrhosis
Acute kidney injury (AKI) is estimated to occur in approximately 20% of hospitalised patients with cirrhosis . Several attempts have been made to achieve agreement on a more precise definition of AKI. About 10 years ago, the Acute Dialysis Quality Initiative (ADQI) group developed the RIFLE criteria for AKI. This classification encompasses Renal risk, Injury, Failure, Loss of kidney function and End-stage renal disease based on changes in serum creatinine and urine volume [27, 28] (see Table 1). AKI has been defined as an increase in serum creatinine ≥133 μmol/L (≥1.5 mg/dl) or an increase >50%, but initiatives from the Acute Kidney Injury Network (AKIN) and others have defined AKI as an absolute increase in serum creatinine >26.4 μmol/L (≥0.3 mg/dl) (or a 50% increase over 48 h) . It is, however, still discussed whether the introduction of the new criteria improves the traditional diagnostic criteria in terms of a better prediction of mortality . Thus, Fagundes et al. were able better to stratify their patients by combining the AKIN criteria with classical criteria in terms of maximum serum creatinine, and three risk groups were identified: (i) patients with AKI stage 1 with peak creatinine ≤1.5 mg/dl; (ii) patients with stage 1 with peak creatinine >1.5 mg/dl; and (iii) patients with stages 2–3 (survival 84%, 68% and 36% respectively; P < 0.001) . Piano et al. reported somewhat similar results with the traditional diagnostic criteria being more accurate than the AKIN criteria in the prediction of in-hospital mortality . It therefore appears that serum creatinine values should be taken into account in the model of AKIN and both studies suggest that AKI-1 with a serum creatinine <1.5 mg/dl is less harmful and it is the progression of renal impairment with serum creatinine >1.5 mg/dl that determines a poor prognosis . Differential diagnosis of AKI in cirrhosis is difficult as it ranges from prerenal AKI (45%), intrarenal AKI including acute tubular necrosis and glomerulonephritis (32%), HRS (23%) and seldom post-renal AKI (<1%) . When applying the creatinine criteria for AKI in Table 1, it may identify a considerable number of patients with AKI, with normal serum creatinine, but reduced glomerular filtration rate (GFR). Similarly, the value of measuring urine production may not be ideal as many patients with decompensated cirrhosis and refractory ascites have a low urine volume <0.5 ml/kg/h without AKI. Differentiation between HRS and other types of AKI should be based on exclusion of dehydration, hypotension/shock, intolerance to dose of diuretics, bacterial infections, biochemical and renal ultrasonographic examinations, the presence of proteinuria and knowledge of precipitating factors . In a new study, Belcher et al. applied the AKIN criteria among a group of 192 hospitalised patients with evidence of AKI . The authors observed that 44% of the patients progressed to a higher AKIN stage after initially fulfilling the AKI criteria. Moreover, complications such as ascites, infections and mortality increased with increased severity of AKI .
Table 1. The Acute Kidney Injury Network (AKIN) criteria for the definition and classification of acute kidney injury (AKI) (modified RIFLE) (Mehta et al. Crit Care 2007) and the K/DOQI guidelines definition and stages of chronic kidney disease (CKD) (Am J Kidney Dis 2002)
Serum creatinine criteria
Urine output criteria
Increase >26.4 μmol/L or increase to >150–200% from baseline
>0.5 ml/kg/h for >6 h
Increase to 200–299% from baseline
>0.5 ml/kg/h for >12 h
Increase to >300% from baseline or >354 μmol/L after a rise of >44 μmol/L or treatment with renal replacement therapy
>0.3 ml/kg/h for >24 h or anuria for 12 h
GFR (ml/min/1.73 m2)
Kidney damage with normal or increased GFR
Kidney damage with mildly decreased GFR
Moderately decreased GFR
Severely decreased GFR
<15 or dialysis
Chronic kidney disease in cirrhosis
Patients with cirrhosis and chronic renal impairment may erroneously be classified as having type 2 HRS when the serum creatinine exceeds 133 μmol/L. On the other hand, they may not meet the criteria from the guidelines by the Kidney Disease Outcomes Quality Initiatives (K/DOQI) Workgroup , see Table 1. According to the criteria as stated in these guidelines, chronic kidney disease (CKD) is defined by a GFR <60 ml/min/1.73 m2 over 3 months. It is a challenge to distinguish patients with type 2 HRS from CKD or . As a diagnostic criteria, Wong et al. have recently proposed CKD in patients with cirrhosis to be defined as GFR<60 ml/min over 3 months calculated from Modification of Diet in Renal Disease (MDRD) .
Acute-on-chronic kidney disease in cirrhosis
Acute-on-chronic liver failure is an increasingly recognised entity that denotes an acute deterioration of liver function in patients with cirrhosis . Alike, AKI may also develop in patients with HRS-2. The distinction between the various combinations of hepatorenal disorders can have important clinical and therapeutic implications. An empirical definition of this entity includes a rise in serum creatinine ≥50% from baseline or a rise of serum creatinine by ≥26.4 μmol/L (≥0.3 mg/dl) in <48 h in a patient with cirrhosis with a GFR<60 ml/min .
The hepatorenal syndrome
Approximately 20% of the cirrhotic patients with ascites who are resistant to diuretics progress to HRS . The new definition of HRS denotes a functional prerenal failure that is unresponsive to volume expansion in patients with chronic liver disease and ascites without significant morphological changes in renal histology, and with a largely normal tubular function [9, 38]. The prognosis of patients with a full-blown HRS is poor ranging from days to weeks, and liver transplantation is the only radical treatment for the HRS that changes long-term prognosis [12, 39]. The new definition and diagnostic criteria of the HRS are shown in Table 2. Two types of HRS have been defined depending on the rapidness and the extent of the renal failure [10, 36, 38]. Type-1 HRS is an acute form with a rapid decrease in renal function and renal failure as an independent predictive factor; type-2 HRS is a chronic form with a more stable renal dysfunction, see Table 3 [36, 40].
Table 2. New diagnostic criteria for the hepatorenal syndrome (HRS) from the International Ascites Club
No improvement of serum creatinine (decrease to a level of <133 μmol/L) after at least 2 days with diuretic withdrawal and volume expansion with albumin. 1 g/kg of bodyweight per day up to a maximum of 100 g/day
Absence of shock
No current treatment with nephrotoxic drugs
Absence of parenchymal kidney disease as indicated by proteinuria >500 mg/day, or microhaematuria, (>50 red blood cells per high power field) and/or a normal renal ultrasonography
Table 3. Type I and type II hepatorenal syndromes (HRS)
Type I HRS
Type II HRS
Serum creatinine double to >222 μmol/L or creatinine clearance<20 ml/min in <2 weeks
Serum creatinine >133 μmol/L or creatinine clearance <40 ml/min with a slow decline
Prognosis: >50% die within 1 month without transplantation
Prognosis: Median survival time is approximately 6 months
Often precipitating events
Most frequent cause of refractory ascites
The major elements in the development of HRS are the diseased liver, the circulatory dysfunction with vasodilatation and lowering of the arterial blood pressure, the abnormal systemic neuro-humoral regulation with activation of the SNS, which alters renal autoregulation, a cardiac dysfunction owing to cirrhotic cardiomyopathy with a preterminal decline in cardiac output [21, 25, 41].
Low systemic vascular resistance, central hypovolaemia, reduced baroreflex sensitivity and abnormal renal autoregulation play a pivotal role in the circulatory dysfunction [25, 42, 43]. Figure 1 summarises the pathophysiological mechanisms. A normal blood pressure is essential for an adequate renal perfusion. At pressure levels below 70 mmHg, the renal autoregulation is abrogated and the renal blood flow (RBF) is directly related to the renal perfusion pressure [44, 45]. In patients with increased sympathetic nervous activity, the autoregulation curve may be shifted towards the right side . Because of this, even minor reductions in arterial blood pressure may be harmful to renal perfusion and function and in these patients, low arterial pressure relates to survival [20, 46]. Figure 2 illustrates how the RBF decreases with the advancement of the clinical stage of liver dysfunction .
Cirrhotic cardiomyopathy has been described as a condition with impaired contractile responsiveness to stress and altered diastolic relaxation . With the progression of the disease, the reduction in the systemic vascular resistance becomes so severe that the hyperdynamic cirrhotic heart is unable further to increase the high cardiac output, which leads to an underfilling of the central vascular bed and effective central hypovolaemia [9, 42, 48]. There is now evidence from several studies of a relationship between the terminal decline in cardiac output and the progression of the disease, development of HRS and survival [49, 50]. We have therefore recently hypothesised a cardiorenal interaction in patients with advanced cirrhosis and renal dysfunction that refers to a condition where cardiac dysfunction in cirrhosis is a major determinant of the course of patients who develop HRS . In addition, a relative adrenal insufficiency has been reported in cirrhosis as part of a hepato-adrenal syndrome with inhibition ACTH and CRH owing to high levels of pro-inflammatory cytokines [37, 52]. As patients with adrenal insufficiency may exhibit similar characteristics in terms of cardiac dysfunction, we recently hypothesised that adrenal insufficiency may contribute to cardiac dysfunction and to precipitate HRS . The relationship between cardiac dysfunction and development of HRS should, therefore, be the focus for treatment strategies that seek to improve cardiac function . It has been discussed whether beta-blockers should be discontinued in patients with end-stage liver disease because of its negative chronotropic and inotropic effects . A reduced cardiac output may lead to a decrease in renal perfusion, azotaemia and increased risk of HRS . At least beta-blockers in these patients should be used in low doses and probably discontinued in patients with low blood pressure [56, 57].
Hyponatraemia is often seen in patients with HRS and is associated with the severity and prognosis [58, 59]. In particular, hyponatraemia is prevalent in decompensated patients complicated with HRS, SBP and hepatic encephalopathy [13, 60]. In patients with cirrhosis and refractory ascites, serum sodium included in the MELDNa score is a better predictor for the outcome than the model of end-stage liver disease (MELD) score alone [61, 62]. In patients, who are candidates for liver transplantation, hyponatraemia is related to increased mortality prior to and after liver transplantation [13, 63, 64]. Primarily hyponatraemia below 130 mmol/L should be the target for treatment [13, 65].
Infections and bacterial translocation
It is well known that cirrhotic patients are susceptible to bacterial infections and in particular spontaneous bacterial peritonitis (SBP) [66, 67]. In particular, SBP is a main cause for the development of HRS and about 33% of patients with SBP develop HRS [66, 68]. Infections and sepsis are serious complications in cirrhosis as it increases mortality approximately four times and one-third of patients die within a month of infection [69, 70]. Bacterial translocation from the gut plays a significant role for spontaneous infections and the circulatory dysfunction characterised by aggravation of vasodilatation elicited by an inflammatory response with the production of pro-inflammatory cytokines such as TNF-α and IL-6 as a ‘cytokine storm’ [71, 72]. SBP is frequent in patients with cirrhosis and is an important risk factor for the development of circulatory dysfunction and HRS [38, 73, 74]. Recently, Barreto et al. reported that in 70 patients with HRS and bacterial infections,  renal dysfunction was not reverted in two-thirds of the HRS type-1 patients with infections despite antibiotic treatment. This very high prevalence of nosocomial infections associated with irreversible renal failure and points to complex mechanisms of infections in developing HRS .
All events that affect the vascular homeostasis and volume regulation in decompensated cirrhosis may precipitate development of HRS. Such precipitating factor includes, apart from infections and sepsis, bleeding from oesophageal varices, large volume paracentesis with inadequate albumin substitution, severe sodium retention and hyponatraemia, which further reduce arterial blood pressure and cardiac output, thereby precipitating development of HRS [38, 49, 50, 77]. New data suggest that a relative adrenal insufficiency that is frequently seen in decompensated cirrhotic patients may contribute to circulatory and renal dysfunction and a higher risk of severe sepsis and type-1 HRS . Recent data from a large randomised study of 110 patients with other infections than SBP suggest that the addition of albumin to antibiotic treatment has beneficial effects on renal and circulatory function and may reduce mortality and risk of type-1 HRS . These findings have been confirmed in a recent meta-analysis of four randomised trials where albumin prevented renal impairment and reduced mortality among patients with SBP . Combination of albumin with the vasoconstrictor terlipressin may further improve arterial pressure and central hypovolaemia .
Treatment of renal failure in cirrhosis
Diagnosis of acute kidney injury in cirrhosis
Diagnosis of HRS and other types of AKI may not be easy in cirrhosis. The International Club of Ascites has published the diagnostic criteria as presented in Table 2 . Differentiation between HRS and other types of AKI should be based on knowledge of precipitating factors , see Fig. 1. One among several clinical challenges is the differentiation between type 1 HRS and acute tubular necrosis (ATIN). A medical history of precipitating events, alcoholic hepatitis, arterial hypotension, nephrotoxic drugs may here be of value. A proposal for a diagnostic algorithm is shown in Fig. 3.
Methods and biomarkers of kidney function
The most widely used biomarker of renal failure is serum creatinine, although it is inaccurate and may first rise when the GFR is decreased by 50% . Moreover, cirrhotic patients are often muscle wasted with reduced production of creatinine, which further limits its use . For the same reasons are the use of creatinine clearance including the MDRD criteria that requires 24-h urine collection not a reliable estimate of GFR, which is over-estimated by these methods . The renal clearance of 51Cr-EDTA is accurately assessed in compensated patients but overestimates the true renal function in patients with ascites because of spill-over into the ascitic fluid. To overcome this, it is necessary to determine the true clearance, which is the total plasma clearance with urine sampling considered as the gold standard .
During the last 10 years, a number of renal biomarkers have been discovered such as kidney injury molecule-1 (KIM-1), liver-type fatty acid-binding protein (L-FABP), interleukine-18 (IL-18), cystatin C, neutrophil-gelatinase-associated lipocalin (NGAL) and others . One of the most promising biomarkers is NGAL, a 25-kDa protein that is up-regulated after kidney injury [87, 88]. NGAL primarily reflects distal tubular injury . Plasma levels relate to severity of liver dysfunction and are particularly elevated in patients with infections [90-92]. In a recent study, patients with HRS had significantly higher urinary NGAL values than patients without renal failure . In that study, NGAL showed potential both to detect AKI and to diagnose HRS and NGAL also proved to be a better prognostic factor than the MELD score . Cystatin C has also shown potentials to be a marker of renal dysfunction in cirrhosis and a strong predictor of survival [94, 95]. Data from a recent study have shown that cystatin C together with the GFR is an accurate predictor of the development of AKI . Few studies on KIM-1 are available in cirrhosis, but its value in cirrhosis seems to be disappointing [97, 98]. Future studies will need to validate the predictive and diagnostic power of these new biomarkers compared to standard methods of assessment of renal function.
Treatment of the hepatorenal syndrome
The hypothetically ideal drug would be a substance that improves liver function, reduces portal pressure, exerts arterial volume expansion, systemic splanchnic vasoconstriction and renal vasodilatation. Such a drug will probably never be developed, but the specific pathogenic mechanisms are its important targets for potential pharmacological treatment combinations.
A prophylactic and surveying approach is essential in the prevention of HRS to avoid precipitating factors to occur and when possible initiate early treatment of imminent infections. Treatment of HRS includes general supportive care to protect respiratory and circulatory function and liver function. Diuretic treatment should be considered discontinued as soon as the diagnosis of HRS is suspected . In some cases, however, furosemide may be used to maintain an adequate diuresis and to avoid over-hydration or when overlapping acute tubular necrosis is suspected. Spironolactone is contraindicated in patients with HRS because of the very high risk of hyperkalaemia . Paracentesis should be considered in decompensated patients as it may improve renal perfusion by reducing the renal venous pressure . However, a post-paracentesis circulatory failure would have a negative effect on renal perfusion pressure because of a reduced arterial blood pressure, and simultaneous infusion of albumin is therefore mandatory also in these patients [100, 101]. Treatment should then be directed to support cardiac function and to treat bacterial infections.
Specific treatment of the HRS includes administration of vasoconstrictors such as terlipressin in combination with infusion of human albumin, also with revised HRS criteria including infection as a precipitating event [81, 102, 103]. A recent meta-analysis has shown that this treatment improves kidney function and reverses HRS type 1 [47, 104]. Terlipressin does not seem to have a sustained effect in patients with HRS type 2. Terlipressin should be given as an initial dose of 1 mg four to six times daily and increased to a maximal dose of 2 mg, six times per day depending of the effect. Human albumin is supplemented in a dose of 40 g per day, equal to 200 ml 20% human albumin . The treatment is continued until normalisation of serum creatinine (below 133 μmol/L) is achieved. After cessation of treatment, relapse of HRS is rather seldom and patients may often respond adequately on a repeated treatment attempt. Approximately 30% of the patients develop side effects to terlipressin, most frequently abdominal pain and diarrhoea (16%) and cardiac and skin manifestations [68, 105, 106]. Appearance of serious side effects, which is seen in 5–7% of the patients, should result in immediate termination of the treatment . Continuous infusion of terlipressin may reduce the risk of serious side effects as shown by Gerbes et al. . However, in the available clinical trials, no deaths because of adverse events have been reported. As renal replacement therapy, extracorporal albumin dialysis has been validated in one randomised controlled trial , but the results have been disappointing and this treatment is no longer recommended [37, 109].
Liver transplantation is the ultimate treatment of HRS types 1 and 2 and should be considered in all patients [38, 110]. Peri-operatively, there may be a further deterioration of renal function, but within 1–2 months, GFR and RBF improve and haemodynamics and neurohumoral changes normalise and most patients with pretransplant kidney dysfunction do not experience progression to advanced kidney disease after liver transplantation . Survival after liver transplantation in patients with HRS type 1 is, however, somewhat lower (65%) than the general survival in cirrhotic patients . The presence of HRS at the time of transplantation is one of the strongest predictors for a poor prognosis after liver transplantation . Patients with HRS who do not respond to terlipressin should likewise be considered for liver transplantation as the renal function usually improves after liver transplantation . Selected patients with HRS and a prolonged need of dialysis (more than 12 weeks prior to liver transplantation) should be considered for a combined liver/kidney transplantation . Nutrition has not been specifically investigated in HRS, but HRS is associated with severe stress metabolism and patients should be offered sufficient nutrition from day one. It is recommended to ensure an energy intake of 35–40 kcal/kg body weight per day (147–168 kJ/kg body weigh per day) and a protein intake of 1.2–1.5 g/kg body weight per day. If possible, nutrition should be oral or subsidiary enteral or parenteral. Isotonic glucose should only be used until other nutritional support is established because it is nutritionally insufficient and may cause hyponatraemia [113, 114].
HRS and hyponatremia are still severe conditions associated with poor prognoses despite progress in treatment options. However, our knowledge on the pathophysiology behind these severe complications has improved considerably and there are indications that novel medical treatments with the combination of different pharmacological principles may have a positive impact on survival. The future approach will probably be to attack different aspects in the pathophysiological process. A multitarget strategy should seek efficiently to counteract the arterial vasodilatation, central hypovolaemia and arterial hypotension by administration of potent vasoconstrictors such as terlipressin combined with human serum albumin. Development of long-acting systemic vasoconstrictors should be encouraged. All patients with HRS including those who respond to terlipressin and albumin should be properly prioritised on the waiting list for liver transplantation. Treatment with vaptans has shown disappointing results in the treatment of dilutional hyponatraemia and an effect on clinical outcomes has not been proven . In particular, focus should be directed towards management of bacterial infections and circulatory and cardiac support.
Professor Søren Møller received a grant from the NovoNordisk Foundation.
Conflict of interest: The authors do not have any disclosures to report.