Acute renal failure is characterized by the association

Transcription

1 PERSPECTIVES IN CLINICAL HEPATOLOGY Acute Renal Failure in Patients With Cirrhosis: Perspectives in the Age of MELD Richard Moreau and Didier Lebrec In patients with cirrhosis, acute renal failure is mainly due to prerenal failure (caused by renal hypoperfusion) and tubular necrosis. The main causes of prerenal failure are true hypovolemia (induced by hemorrhage or gastrointestinal or renal fluid losses), sepsis, or type 1 hepatorenal syndrome (HRS). The frequency of prerenal failure due to the administration of nonsteroidal anti-inflammatory drugs or intravascular radiocontrast agents is unknown. Prerenal failure is rapidly reversible after restoration of renal blood flow. Treatment is directed to the cause of hypoperfusion, and fluid replacement is used to treat most cases of non-hrs prerenal failure. In patients with type 1 HRS with very low short-term survival rate, liver transplantation is the ideal treatment. Systemic vasoconstrictor therapy (with terlipressin, noradrenaline, or midodrine [combined with octreotide]) may improve renal function in patients with type 1 HRS waiting for liver transplantation. MARS (for molecular adsorbent recirculating system) and the transjugular intrahepatic portosystemic shunt may also improve renal function in these patients. In patients with cirrhosis, acute tubular necrosis is mainly due to an ischemic insult to the renal tubules. The most common condition leading to ischemic acute tubular necrosis is severe and sustained prerenal failure. Little is known about the natural course and treatment (i.e., renal replacement therapy) of cirrhosis-associated acute tubular necrosis. (HEPATOLOGY 2003;37: ) Abbreviations: HRS, hepatorenal syndrome; GFR, glomerular filtration rate; NSAID, nonsteroidal anti-inflammatory drug; SBP, spontaneous bacterial peritonitis; COX, cyclooxygenase; MARS, molecular adsorbent recirculating system; TNF-, tumor necrosis factor. From the Laboratoire d Hémodynamique Splanchnique et de Biologie Vasculaire, INSERM U-481, and Service d Hépatologie, Hôpital Beaujon, Clichy, France. Received July 15, 2002; accepted December 6, Address reprint requests to: Richard Moreau, M.D., INSERM U-481 and Service d Hépatologie, Hôpital Beaujon, Clichy, France. rmoreau@ bichat.inserm.fr; fax: (33) Copyright 2003 by the American Association for the Study of Liver Diseases /03/ $35.00/0 doi: /jhep Acute renal failure is characterized by the association of a rapid decline in glomerular filtration (over a period of hours to days), perturbation of extracellular fluid volume, electrolyte and acid-base homeostasis, and retention of nitrogenous waste from protein catabolism. 1-3 Acute renal failure is thought to be common in patients with cirrhosis, 4 but its exact incidence is unknown. Patients with cirrhosis are predisposed to acute renal failure following complications (e.g., portal hypertensive bleeding) 5 or the administration of drugs (e.g., aminoglycoside antibiotics). 6 Moreover, patients with cirrhosis may develop a specific acute renal failure called type 1 hepatorenal syndrome (HRS) Finally, acute renal failure has a prognostic importance; for example, in patients admitted to the hospital for acute upper gastrointestinal hemorrhage, the occurrence of acute renal failure during hospitalization is an independent predictive factor of death. 5 Definitions In patients with cirrhosis, acute renal failure frequently develops during hospitalization. In patients without renal impairment at admission, acute renal failure is diagnosed when the serum creatinine level increases by more than 50% of the baseline value, to above 133 mol/l (1.5 mg/dl). 11,12 In patients with preexisting renal impairment, acute renal failure is diagnosed when serum creatinine increases by more than 50% above baseline. 11 Acute renal failure resulting from renal hypoperfusion without cellular injury (i.e., without glomerular or tubular lesions) is called prerenal failure, or prerenal azotemia. 3 Renal dysfunction due to obstruction of the urinary outflow tract is termed postrenal failure, or postrenal azotemia. 3 Acute renal failure due to a primary intrarenal cause is called intrinsic renal failure, or renal azotemia. 3 Intrinsic renal failure may be caused by either tubular necrosis (due to an ischemic or toxic insult to the tubule), glomerulonephritis (due to a primary decrease in the filtering capacity of the glomerulus), or interstitial nephritis 233

2 234 MOREAU AND LEBREC HEPATOLOGY, February 2003 (due to a tubulointerstitial process with inflammation and edema). 3 HRS is a prerenal failure. 3 There are two forms of HRS: type 1 is the acute form, whereas type 2 is a moderate and chronic form. 4 Type 2 HRS is not considered in this review. Prerenal failure is a preischemic state; it may lead to ischemic tubular necrosis when the reduction in blood flow is sufficient to result in the death of tubular cells. 3 A recent multicenter retrospective study investigated 355 patients with cirrhosis and acute renal failure. 13 Two hundred six (58.0%) patients had prerenal failure, including 71 who had type 1 HRS. One hundred forty-eight (41.7%) patients had acute tubular necrosis; ischemic tubular injury occurred in 139 cases. Only 1 (0.3%) patient had postrenal failure. In this study, there was no case of acute renal failure due to acute glomerulonephritis confirming that this is an uncommon cause of acute renal failure in cirrhosis (see below). Together, these findings indicate that renal hypoperfusion can explain more than 90% of cases of cirrhosis-associated acute renal failure. Fig. 1. Proposed mechanisms explaining the decrease in renal blood flow and the preservation of glomerular filtration rate in nonazotemic patients with cirrhosis and ascites. The decrease in renal blood flow is a result of the overactivity of endogenous vasoconstrictor systems (the sympathetic nervous system and the renin-angiotensin system). (A) This neurohumoral overactivity is induced by decreases in effective arterial blood volume (and arterial pressure) resulting from marked splanchnic and systemic vasodilation. (B) At this stage, the glomerular filtration rate does not markedly decrease because compensatory vasodilator mechanisms limit the constrictor action of neurohumoral vasocontrictors in afferent (preglomerular) arterioles. An important compensatory mechanism is the biosynthesis of cyclooxygenase-derived vasodilator prostaglandins in renal arterial walls. The decrease in glomerular filtration may be prevented by another mechanism, i.e., angiotensin II-induced efferent (postglomerular) arteriolar vasoconstriction which maintains glomerular capillary pressure, the driving force for glomerular filtration. Expected Causes of Acute Renal Failure in Cirrhosis Prerenal Failure Before commenting on the expected causes of cirrhosis-associated prerenal failure, two important points should be noted. First, in cirrhosis, prerenal azotemia usually develops in patients with ascites. These patients already have significant circulatory alterations such as low arterial pressure, renal vasoconstriction, and decreased renal blood flow (Fig. 1A); but they have no or only slight decreases in the glomerular filtration rate (GFR) due to compensatory mechanisms (Fig. 1B). 14 Second, in cirrhosis, the causes of prerenal failure further decrease renal blood flow and the GFR (Fig. 2). Prerenal failure may be rapidly reversible if the underlying cause is corrected. 1-3 True hypovolemia (induced by hemorrhage or gastrointestinal or renal fluid losses) associated with shock or arterial hypotension is one cause of prerenal azotemia. 1-3 There is a short delay between hemorrhage and the resulting renal dysfunction. Acute upper gastrointestinal hemorrhage is a common complication of cirrhosis. Randomized and nonrandomized studies investigating acute upper gastrointestinal hemorrhage in patients with cirrhosis show that 10% to 20% of patients have hypovolemic shock at inclusion. Moreover, when treating acute bleeding in cirrhotic patients, blood volume restitution should be performed cautiously and conservatively. 15 For all these reasons, prerenal failure is expected to occur in patients with cirrhosis and gastrointestinal bleeding. However, the incidence of hemorrhageinduced renal failure is unknown because the immediate impact of shock or arterial hypotension on renal function was not an end point in most studies. A retrospective study has shown that 5% of patients with cirrhosis hospitalized for acute upper gastrointestinal hemorrhage have early renal failure that lasts less than 7 days after index bleeding. 5 Finally, it should be emphasized that patients admitted for hemorrhage may develop prerenal failure due to other causes. Indeed, acute gastrointestinal hemorrhage predisposes patients with cirrhosis to bacterial

3 HEPATOLOGY, Vol. 37, No. 2, 2003 MOREAU AND LEBREC 235 Fig. 2. Expected causes of prerenal failure and acute ischemic tubular necrosis in patients with cirrhosis. infections, 16 which may induce prerenal azotemia or acute tubular necrosis (see below). On the other hand, a significant proportion of patients with cirrhosis admitted for acute upper gastrointestinal hemorrhage have received nonsteroidal anti-inflammatory drugs (NSAIDs) in the week preceding bleeding 17 and NSAIDs may cause prerenal failure (see below). True hypovolemia and subsequent renal failure may be a result of vomiting or diarrhea. 4 Diuretic treatment to mobilize tense or large ascites may induce prerenal failure. 4 Glycosuria is a cause of renal fluid loss, which may lead to renal dysfunction. Since diabetes is often associated with cirrhosis, glycosuria-induced prerenal failure may occur in patients with chronic liver disease. Severe sepsis is defined as the association of sepsis (i.e., known or suspected infection plus at least two signs of the systemic inflammatory response syndrome) and acute organ dysfunction (Table 1). 18 In the general population, patients with severe sepsis frequently develop prerenal azotemia as a result of septic shock. 18 However, certain patients with severe sepsis without shock develop prerenal failure. 18 Renal vasoconstriction plays a role in sepsisinduced prerenal azotemia. 2,3 Since decreased intravascular volume is a mechanism of sepsis-induced renal vasoconstriction, prerenal failure may be sensitive to fluid replacement. 3 Patients with cirrhosis are susceptible to bacterial infections, in particular spontaneous bacterial peritonitis (SBP). 19 Septic shock and subsequent renal impairment occurs in 10% of patients with SBP. 19 At the onset of SBP, 20% to 40% of patients have renal failure without shock. 11,20 However, studies on SBP provide little information on the cause of this renal dysfunction. In particular, the following questions remain unanswered. Do patients have prerenal azotemia as suggested? If so, how many patients have acute renal failure due to sepsis or other causes? And how many patients have preexisting chronic renal impairment that is not due to sepsis (e.g., type 2 HRS)? Five percent of patients hospitalized for acute upper gastrointestinal hemorrhage 5 and 30% of those admitted for SBP develop type 1 HRS during hospitalization. 11 Type 1 HRS also occurs in 10% of patients with ascites treated by total paracentesis 21 and in 25% of patients with severe acute alcoholic hepatitis. 22 Marked renal vasoconstriction explains renal failure in patients with HRS. 2 However, the mechanisms of renal vasoconstriction have not yet been fully elucidated In type 1 HRS, there is a rapid increase in serum creatinine. 4 At diagnosis of HRS, there is no recent history of circulatory shock, increased intestinal or renal fluid loss, or nephrotoxic drug administration. 4 In addition, there is no evidence for ongoing bleeding or sepsis. 4 Finally, renal failure does not improve following administration of plasma expanders and discontinuation of diuretics. 1,4 Administration of nonselective NSAIDs (i.e., drugs that can inhibit the two cyclooxygenase [COX] isoforms, COX-1 and COX-2) is not recommended in patients with cirrhosis. 23 However, studies have shown that these patients may receive nonselective NSAIDs. 17 Since COXderived vasodilator prostaglandins protect renal perfusion in patients with cirrhosis and ascites, these patients may develop marked renal hypoperfusion and subsequent prerenal failure following COX inhibition induced by nonselective NSAIDs (reviewed in Clive and Stoff 23 ). However, the incidence of nonselective NSAID-induced prerenal failure in patients with cirrhosis is unknown. Newly developed selective COX-2 inhibitors (celecoxib and rofecoxib) are thought to be safer for renal function than nonselective NSAIDs. 24 However, administration of selective NSAIDs has been shown to induce a decrease in GFR in salt-depleted elderly subjects, i.e., in subjects with decreased intravascular volume. 24 In addi- Table 1. Definitions in Sepsis Sepsis Known or suspected (e.g., polymorphonuclear cell count 250 mm 3 in the ascitic fluid) infection Presence of at least 2 signs of the systemic inflammatory response syndrome (referred to as SIRS): Core temperature 38 C or 36 C Heart rate 90 beats/min Respiratory rate 20 breaths/min or PaCO 2 32 mm Hg White cell count 12,000/mm 3 or 4,000/mm 3 or a differential count showing 10% immature neutrophils Severe sepsis Presence of sepsis plus at least one organ dysfunction (e.g., kidney dysfunction) Abbreviation: PaCO 2, partial pressure of arterial carbon dioxide. Adapted from reference 18.

4 236 MOREAU AND LEBREC HEPATOLOGY, February 2003 tion, preliminary results have shown that short-term celecoxib administration induces a marked decrease in GFR in certain patients with cirrhosis and ascites. 25 Together, these findings suggest that COX-2 inhibitors, like nonselective NSAIDs, may induce prerenal failure in patients with cirrhosis and ascites. The intravascular administration of iodinated agents may cause prerenal failure by inducing renal vasoconstriction. 26 There are 3 important risk factors of radiocontrastinduced acute renal failure: chronic renal failure, diabetes, and decreased effective arterial blood volume. 26 Nonazotemic patients with cirrhosis and ascites have decreased effective arterial blood volume. 4 They may also have chronic renal impairment (e.g., those with type 2 HRS) or diabetes. Since patients with cirrhosis often receive radiocontrast agents, they may develop prerenal failure due to these agents. 26 However, preliminary results show that radiocontrast agents did not impair renal plasma flow and GFR in patients with cirrhosis. 27 It should be noted that these results were obtained in a small series of patients, one third of whom did not have ascites. 27 More information is needed on the effects of radiocontrast agents on renal function in patients with cirrhosis. Drugs (such as nitrovasodilators or substances inhibiting angiotensin II action) that are used in the treatment of portal hypertension may induce renal dysfunction in patients with cirrhosis and ascites. It has been shown that acute administration of the nitrovasodilator 5-mononitrate induced decreases in effective arterial blood volume, arterial pressure, renal plasma flow, and GFR. 28 However, other studies did not show any increase in the incidence of prerenal failure in patients receiving long-term administration of 5-mononitrate isosorbide. 29 Inhibition of angiotensin II action by inhibiting the angiotensinconverting enzyme with captopril or blocking type 1 angiotensin receptors with losartan or irbesartan may induce marked arterial hypotension and vasodilation of postglomerular arterioles, causing prerenal failure in patients with cirrhosis and ascites Fig. 3. Proposed mechanisms for the decrease in the glomerular filtration rate in patients with ischemic or toxic acute tubular necrosis. The glomerular filtration rate falls because of impaired renal blood flow and glomerular capillary pressure, disrupted integrity of the tubule epithelium with backleak of glomerular filtrate, and obstructed urine flow due to intratubular formation of casts composed of detached epithelial cells and cellular debris. Intrinsic Renal Failure Acute Tubular Necrosis. Injury to the renal tubules may be ischemic or toxic in origin. The mechanisms of renal failure are very similar in ischemic and toxic tubular necrosis (Fig. 3). The course of acute tubular necrosis can be divided into the initiation (hours to day), maintenance (1-2 weeks), and recovery phases. 2 All causes of prerenal azotemia may lead to ischemic tubular injury. 3 Administration of aminoglycoside antibiotics is thought to be the most common cause of toxic tubular necrosis in patients with cirrhosis. 4 The incidence of radiocontrast agent-induced tubular necrosis is unknown in patients with cirrhosis. Ischemia and toxins may combine to cause renal failure in severely ill patients. 4 Acute Glomerulonephritis. Some cases of infectious acute glomerulonephritis have been reported in patients with cirrhosis. 33 The site of infection was the oropharynx, the skin, or the endocardium. 33 In patients with cirrhosis due to hepatitis C virus, cryoglobulinemic membranoproliferative glomerulonephritis is a rare cause of acute renal failure. 34 Macroscopic hematuria associated with acute renal failure may occur in certain patients with alcohol-induced cirrhosis and IgA nephropathy. 35 In these patients, acute renal failure is due to tubular necrosis and not caused by glomerular lesions. 35 Tubular necrosis may be caused by a direct toxic insult to the tubules by red blood cells that are present in the tubular lumen. 3 There are other causes of acute renal failure that are not associated with cirrhosis (Table 2). Predictive Factors and Prognosis of Acute Renal Failure in Cirrhosis In patients with cirrhosis admitted for acute upper gastrointestinal hemorrhage, the presence of severe cirrhosis at admission is an independent predictive factor of the development of acute renal failure. 5 Similarly, the severity of cirrhosis at the time of diagnosis of SBP is an independent predictive factor of the occurrence of type 1 HRS. 11 In addition, increased serum creatinine at the onset of SBP (see above) is an independent predictive factor of the development of type 1 HRS. 11 In patients with severe sepsis and prerenal failure, delayed fluid replacement and the early use of mechanical

5 HEPATOLOGY, Vol. 37, No. 2, 2003 MOREAU AND LEBREC 237 Table 2. Causes of Acute Renal Failure Not Associated With Cirrhosis Prerenal causes Intravascular volume depletion and hypotension Gastrointestinal fluid loss (nasogastric suction) or pooling of fluid (pancreatitis, bowel disease) Trauma, surgery, burns Decreased effective intravascular volume Congestive heart failure or other causes of myocardial failure Nephrotic syndrome Anaphylaxis Anesthetic agents Renal artery or renal vein occlusion by thrombosis; atheroembolism Intrinsic causes Tubular necrosis Ischemic (as a consequence of above-mentioned prerenal events) Toxic due to cyclosporin, tacrolimus, amphotericin B, methotrexate, foscarnet, pentamidine, organic solvents (carbon tetrachloride, ethylene glycol), heavy metals (mercury, cisplatin), heme pigments (rhabdomyolysis), myeloma light-chain Interstitial nephritis related to drugs (penicillins, cephalosporins, sulfonamides, rifampicin, ciprofloxacin, selective and nonselective NSAIDs, thiazide diuretics, furosemide, cimetidine, phenytoin, allopurinol), infection, cancer, or sarcoidosis Small-vessel vasculitis or acute glomerulonephritis due to connective-tissue disorders, scleroderma, microscopic polyarteritis, rapidly progressive glomerulonephritis, or other disorders Postrenal causes Upper urinary tract obstruction: ureteral obstruction of one or both kidneys Lower urinary tract obstruction: bladder-outlet obstruction NOTE. Expected causes of acute renal failure in patients with cirrhosis are shown in Fig. 2. ventilation (an indicator of severe sepsis) are 2 independent predictive factors of the occurrence of acute tubular necrosis. 36 Patients with nonoliguric (urinary output 400 ml/d) acute renal failure have a better prognosis than patients with oliguric (urinary output 400 ml/d) renal failure. 3 Despite advances in intensive care, the in-hospital mortality rate remains elevated in patients with type 1 HRS (75%), mainly due to sepsis-induced multiorgan failure. 13 The mortality rate in patients with cirrhosis admitted for septic shock associated with the acute respiratory distress syndrome is 80% to 100%. 37 A study in patients who received a liver transplant for end-stage liver disease has shown that the presence of pretransplantation renal failure was an independent predictive factor of short- and long-term postoperative mortality. 38 This study, however, did not identify the causes of pretransplantation renal failure. 38 Nevertheless, other studies have shown that, among patients who received a liver transplant, the probability of postoperative death was higher in patients with pretransplantation HRS than in those without (reviewed in Bataller et al. 7 ). The impact of pretransplantation acute tubular necrosis on posttransplantation outcome has not yet been specifically studied. Diagnostic Evaluation Since acute renal failure mainly occurs in patients with decompensated cirrhosis, diagnosis of chronic liver disease is generally easy. In contrast, the diagnosis of the cause of renal failure may not be as easy. Clinical Evaluation A history of exposure to nephrotoxic medications, a recent history of angiography, and physical findings of gastrointestinal hemorrhage or suggesting sepsis (Table 3) all provide important diagnostic information. Anuria suggests postrenal azotemia. 3 The existence of arterial hypertension, which is an unexpected finding in cirrhosis, suggests glomerulonephritis. 39 Macroscopic hematuria should suggest IgA nephropathy as a cause. 39 Purpura, arthralgia, weakness, Raynaud s syndrome, or leg ulcers suggest cryoglobulinemia. 3 Urine Evaluation Diagnostic information may be obtained from the urinanalysis. 1-3 For example, pigmented granular casts are typical of ischemic and toxic acute renal failure and red cell casts of glomerulonephritis. 3 Patients with renal azotemia due to acute or subacute glomerulonephritis have significant proteinuria (around 3 g/d). 2,3,39 In contrast, proteinuria is absent or moderate in other causes of acute renal failure. 4 Urine indices (urine osmolality, urinary sodium concentration, and fractional excretion of sodium) may help distinguish prerenal failure (including type 1 HRS) from tubular necrosis (Table 4). 2,3 The tubular ability to reabsorb sodium and to concentrate urine is preserved in prerenal azotemia and impaired in tubular necrosis. 2,3 Thus, patients with prerenal failure have low urinary sodium concentrations (below 20 mmol/l) and elevated urine osmolality (higher than 500 mosm/kg). 2,3 In contrast, patients with tubular necrosis have high urinary sodium concentrations (above 40 mmol/l) and urine osmolality below 350 mosm/kg. 2,3 However, the urinary sodium concentration may be low early in the course of certain Table 3. Signs That Suggest the Presence of Sepsis in Patients With Acute Renal Failure Traditional criteria of sepsis: i.e., known or suspected infection plus signs of SIRS Increased serum concentrations of C-reactive protein or procalcitonin Increased cardiac output, low systemic vascular resistance, low oxygen extraction ratio Increased insulin requirements Altered tissue perfusion: e.g., altered skin perfusion Organ dysfunction: e.g., coagulopathy Abbreviation: SIRS, systemic inflammatory response syndrome. See also Table 2.

6 238 MOREAU AND LEBREC HEPATOLOGY, February 2003 Table 4. Examples of Urine Findings According to Causes of Acute Renal Failure Condition Urine Osmolality (mosm/ kg) Sodium Concentration (mmol/l) Fractional Excretion of Sodium (%) Prerenal failure Intrinsic renal failure Tubular necrosis* Acute interstitial nephritis Acute glomerulonephritis Postrenal failure *In nonoliguric acute renal failure due to sepsis or radiocontrast agents, the initial fractional excretion of sodium can be 1%. When tubulointerstitial abnormalities are associated with glomerular lesions, urine osmolality is 350 mosm/kg and the fractional excretion of sodium is 1%. The fractional excretion of sodium can be 1% early in the course of obstruction (before tubular injury occurred). Adapted from references 1-3. processes that lead to tubular necrosis, such as sepsis, exposure to radiocontrast agents, or obstruction. 3 In addition, some cases of HRS with elevated urinary sodium concentrations have been reported. 4 These findings indicate that in clinical pratice, it is difficult to differentiate type 1 HRS from other causes of acute renal failure. The International Ascites Club has suggested that 5 major criteria be present to confirm the diagnosis of HRS (Table 5), 4 and one of the most important is the lack of improved renal response following optimization of intravascular volume in patients with type 1 HRS. 4 Indeed, although there is an increase in GFR following fluid replacement in most causes of prerenal failure, GFR continues to rise despite fluid challenge in patients with type 1 HRS. 1 On the other hand, because patients with acute tubular necrosis do not respond to fluid replacement, a lack of renal response to fluid challenge does not differentiate HRS from tubular necrosis. 1-4 This is a limitation of the International Ascites Club criteria. It should be emphasized that patients with type 1 HRS have improved renal function following administration of vasoconstrictors such as terlipressin 13 or noradrenaline, 40 whereas vasoconstrictor therapy does not improve renal function in patients with acute tubular necrosis. 3 Thus, the renal response to vasoconstrictor agents might be used to differentiate HRS from tubular necrosis. Finally, renal biopsy helps in the diagnosis of acute renal failure (see below). It has been suggested that the urinary markers of tubular damage such as 2 -microglobulin may help in the diagnosis of tubular necrosis due to aminoglycoside antibiotics. 6 However, the increase in urinary 2 -microglobulin is not specific for aminoglycoside-induced nephrotoxicity. Blood Tests The presence of antibodies to hepatitis C virus and hepatitis C virus RNA, high concentrations of cryoglobulins, positive rheumatoid factor assays, and low concentrations of complement suggest hepatitis C virus associated cryoglobulinemic glomerulonephritis. 39 Evaluation of Obstruction Early diagnosis of obstruction is important because most cases can be treated and a delayed therapy can result in irreversible renal injury. 3 Bladder catheterization should be performed initially if bladder neck obstruction is suspected. 3 Renal ultrasonography is the test of choice to exclude the diagnosis of obstruction at the ureters or above (specificity: 93%). However, sensitivity for hydronephrosis is only 80% to 85%. A nondilated collecting system does not necessarily exclude obstruction, e.g., in patients with hypovolemia. 3 Renal Biopsy Renal biopsy is generally not necessary in the diagnosis of acute renal failure. 3 However, when in-depth clinical assessment and laboratory and radiologic investigations have ruled out prerenal and postrenal azotemia and suggest a diagnosis of an intrinsic renal disorder other than acute tubular necrosis, a kidney biopsy may determine the diagnosis and treatment. 3,41 Percutaneous renal biopsy may be contraindicated in patients with cirrhosis and coagulation disorders. Transjugular renal biopsy has been shown to be a safe procedure in these patients. 41 In addition to kidney biopsy, a transjugular liver biopsy can confirm the diagnosis of cirrhosis. 41 Management of Acute Renal Failure The management of acute renal failure due to true hypovolemia, and severe sepsis is described below, because this is frequent in patients with cirrhosis. In most cases, patients should be managed in an intensive care unit. Table 5. International Ascites Club Major Criteria for the Diagnosis of Hepatorenal Syndrome Severe cirrhosis Serum creatinine above 133 mol/l (1.5 mg/dl) Absence of shock, ongoing bacterial infection, or recent treatment with nephrotoxic drugs Absence of gastrointestinal fluid loss No improvement in serum creatinine despite optimization of intravascular blood volume (by using fluid replacement and stopping diuretics) Proteinuria below 0.5 mg/d No ultrasonographic evidence of renal obstructive disease Adapted from Arroyo et al. 4

7 HEPATOLOGY, Vol. 37, No. 2, 2003 MOREAU AND LEBREC 239 Prerenal Failure Prerenal azotemia is rapidly reversible after restoration of renal blood flow. Treatment is directed at the cause of hypoperfusion. 3 True Hypovolemia. In patients with acute upper gastrointestinal hemorrhage, transfusions of packed red blood cells maintain the hematocrit between 25% and 30%, and plasma expanders maintain hemodynamic stability. 15 In patients with intestinal or renal fluid losses, crystalloids (and sometimes colloids) are administered, and any diuretic therapy should be stopped. 2 Severe Sepsis. Our knowledge in this setting is based on studies performed in noncirrhotic patients. In patients with severe sepsis, correction of volume depletion is crucial for the prevention of acute tubular necrosis. 3,36 However, the following questions remain unanswered: Should a crystalloid or colloid solution be used? Which colloid solution should be administered (i.e., either albumin, gelatin, dextran, or hydroxyethylstarch)? Nevertheless, 6% hydroxyethylstarch should not be used. Indeed, a multicenter randomized study in 129 patients with severe sepsis has shown that 6% hydroxyethylstarch significantly increased the incidence of in-hospital acute renal failure. 36 In patients with septic shock, hypotension may persist despite adequate fluid replacement leading to the use of catecholamines. Because renal vasoconstriction plays a role in sepsisinduced acute renal failure, it has been suggested that low-dose dopamine ( 3 g/kg/min, a dose known to induce renal vasodilation) may prevent acute tubular necrosis in patients with severe sepsis. However, a multicenter, randomized, placebo-controlled study performed in 324 patients with at least 2 systemic inflammatory response syndrome criteria and early renal dysfunction has shown that low-dose dopamine had no effect on renal function or the need for dialysis. 42 Thus, low-dose dopamine should not be used to prevent acute tubular necrosis in patients with severe sepsis. Because delayed treatment of circulatory disorders is a significant risk factor for acute tubular necrosis in patients with severe sepsis, 36 early hemodynamic therapy may be beneficial. A randomized study in 263 patients with severe sepsis compared early goal-directed therapy with standard therapy. 43 Patients assigned to early goal-directed therapy received, in a sequential manner, fluid replacement, vasoactive drugs, red-cell transfusions, and inotropic drugs to achieve target levels of central venous pressure (8-12 mm Hg), mean arterial pressure (65-90 mm Hg), and central venous oxygen saturation (at least 70%). Patients who did not respond to these treatments underwent sedation and mechanical ventilation. Patients assigned to early goal-directed therapy had significantly less severe organ dysfunction than those assigned to standard therapy. Moreover, in-hospital mortality was significantly lower in patients assigned to early goal-directed therapy than in the other patients. 43 These findings suggest that, in patients with severe sepsis, early goal-directed therapy may be used both to prevent renal dysfunction and to improve survival. Because the degree of severity of sepsis at admission to intensive care unit is another significant risk factor for acute tubular necrosis, 36 any intervention that decreases the severity of illness may be useful. A randomized, placebo-controlled study in 1,690 patients with severe sepsis has shown that administration of dotrecogin alfa activated (a recombinant human activated protein C with antiinflammatory and anticoagulant actions) significantly decreased the number of dysfunctional organs or systems and mortality. 44 However, dotrecogin alfa activated may be associated with an increased risk of bleeding. 44 Since preliminary results have shown that patients with cirrhosis and severe sepsis have a very marked acquired deficiency in protein C activity, 45 these patients may benefit from dotrecogin alfa activated administration. Type 1 HRS. The treatment of HRS has been reviewed elsewhere Although the best treatment for HRS is liver transplantation, patients with HRS who are transplanted have more complications and a higher inhospital mortality rate than those without HRS (see above and Bataller et al. 7 ). Therefore, in patients with HRS, it might be better to treat renal function abnormalities before liver transplantation. It has been suggested that systemic vasoconstrictor therapy may improve renal function in patients with HRS by increasing effective arterial blood volume. 10 A nonrandomized retrospective study in a large series of patients with HRS suggests that the vasopressin analogue, terlipressin, is an effective treatment of renal failure. 13 Other nonrandomized studies suggest that vasoconstrictor therapy with noradrenaline 40 or midodrine (combined with octreotide) 46 may improve renal function in these patients. On the other hand, a randomized study in a small series of patients, comparing the molecular adsorbent recirculating system (MARS) (eventually combined with intermittent venovenous hemofiltration) and intermittent venovenous hemofiltration alone, suggests that MARS may improve survival in HRS. 47 Finally, nonrandomized studies suggest that the transjugular intrahepatic portosystemic shunt (a nonsurgical method of portal decompression) may improve renal function in patients with HRS (reviewed by Moreau 10 ). Acute Tubular Necrosis Strategies to prevent or treat acute tubular necrosis have been evaluated mainly in the general population.

8 240 MOREAU AND LEBREC HEPATOLOGY, February 2003 Table 6. General Care During the Course of Prerenal Failure and Acute Tubular Necrosis in Patients With Cirrhosis Discontinue nephrotoxins (e.g., aminoglycosides, NSAIDs) Avoid radiocontrast agents Treat hyperkalemia and metabolic acidosis Provide sufficient nutritional support Prevent and treat portal hypertensive bleeding Prevent and treat bacterial infections Prevention. General care to prevent acute tubular necrosis in patients with prerenal azotemia is shown in Table 6. A randomized study performed in 1,548 critically ill surgical patients compared intensive insulin therapy (maintaining blood glucose between 4.4 and 6.1 mmol/l) with conventional treatment, in which insulin was only given when the blood glucose level exceeded 11.9 mmol/l. 48 Intensive insulin therapy significantly reduced the incidence of acute renal failure and was also associated with a significant decrease in mortality due to a reduction in the incidence of multiorgan failure due to sepsis. 48 Since this study was performed in critically ill surgical patients, other studies should evaluate the effect of intensive insulin therapy in critically ill medical patients, especially those with cirrhosis. General Principles of Treatment. General care to prevent further kidney injury is shown in Table 6. Since renal vasoconstriction is involved in the mechanism of acute tubular necrosis (Fig. 3), the renal vasodilator action of atrial natriuretic peptide may be beneficial. 3 However, 2 large randomized studies in more than 700 critically ill patients with acute renal failure and a clinical diagnosis of acute tubular necrosis did not show any effect of a 24- hour infusion of atrial natriuretic peptide on dialysis-free survival. 49,50 Since tubular obstruction plays a role in the decrease in GFR in patients with acute tubular necrosis (Fig. 3), loop diuretics may induce diuresis, washing out obstructing cellular debris and casts. 3 A randomized study in 92 patients with oliguric acute tubular necrosis treated with low-dose dopamine has shown that diuretics induced diuresis but did not change main outcome variables (renal recovery, need for dialysis, and survival). 51 Other treatments have been tested in patients with acute tubular necrosis: low-dose dopamine, mannitol, calcium channel blockers (reviewed in Brady and Singer 2 and Thadhani et al. 3 ). No beneficial effect has been shown with these substances. Renal Replacement Therapy. In the general population, renal replacement therapy is indicated in patients with symptoms or signs of uremia (such as pericardial rub or effusion, hiccups, or changes in mental status) and the management of volume overload, hyperkalemia, or acidosis that are refractory to usual treatment. 2,3 In addition, certain physicians start dialysis if blood urea nitrogen is above 35 mmol/l, even in the absence of clinical signs of uremia. 2 However, studies evaluating the efficacy of prophylactic dialysis provide conflicting results. 2 There are numerous renal replacement therapies including conventional (alternate-day) intermittent hemodialysis, daily intermittent dialysis, peritoneal dialysis, and various forms of continuous renal replacement therapies. 2,3 A randomized study in 70 patients with severe infection-associated acute renal failure has shown that the mortality rate was significantly lower with continuous renal replacement therapy than with peritoneal dialysis. 52 The hypotension that occurs during treatment with alternate-day intermittent hemodialysis may prolong tubular injury, which may increase morbidity and mortality. 2,3 Since continuous renal replacement therapy may offer the benefit of slow and controlled ultrafiltration with a decrease in the frequency and duration of hypotensive episodes, this technique is often used in critically ill patients with acute renal failure. 2,3 In addition, in patients with severe sepsis, continuous renal replacement therapy may remove injurious circulating proinflammatory cytokines. 3,53 However, there is no clear advantage between continuous and intermittent hemodialysis for hemodynamic instability and renal recovery (reviewed in Vanholder et al. 53 ). Moreover, analysis of nonrandomized studies (reviewed in Vanholder et al. 53 ) and the results of one randomized study 54 indicate that in-hospital mortality did not significantly differ between continuous and intermittent hemodialysis. Finally, it should be kept in mind that continuous renal replacement therapy has disadvantages (e.g., immobilization in bed, systemic anticoagulation, and cost). 2,3,53 It has been suggested that the use of bioincompatible dialysis membranes may activate complement and neutrophils. 3 Kidney infiltration by activated neutrophils may prolong renal damage and delay recovery of renal failure. 3 It has been shown that the rate of renal recovery and survival were significantly higher in patients treated with biocompatible membranes than in those treated with bioincompatible membranes. However, these results were not confirmed in other studies (reviewed in Vanholder and Lameire 55 ). Finally, the dose of hemodialysis has been studied. A randomized study performed in 425 critically ill patients with acute renal failure treated with venovenous hemofiltration has shown that an increase in the rate of ultrafiltration was associated with improved survival. 56 A prospective nonrandomized study performed in 160 critically ill patients with acute renal failure showed significantly higher survival in patients treated with daily intermittent hemodialysis than in those treated with alternate-day intermittent hemodialysis. 57 To date, continuous renal replacement therapy with elevated ultrafil-

9 HEPATOLOGY, Vol. 37, No. 2, 2003 MOREAU AND LEBREC 241 tration volumes has not yet been compared with daily intermittent dialysis. These findings indicate that the optimal treatment of critically ill patients with acute renal failure remains to be determined. Theoretically, in patients with cirrhosis and acute tubular necrosis, the indications for renal replacement therapy should be similar to those for the general population (see above). However, since a significant number of patients with cirrhosis and acute tubular necrosis are in very poor condition and have a high short-term probabilty of death, these patients may not benefit from renal replacement therapy. In fact, data are insufficient to help make the decision to use renal replacement therapy in patients with cirrhosis and acute tubular necrosis. If a patient with cirrhosis is an appropriate candidate for renal replacement therapy, one question must be answered: What type of therapy? Although there are no specific studies of peritoneal dialysis in patients with cirrhosis, it is not reasonable to use peritoneal dialysis because of the risk of SBP. In critically ill patients with cirrhosis and acute tubular necrosis, continuous renal replacement therapy may be safer and more effective (e.g., for controlling intracranial hypertension) than alternate-day intermittent dialysis. 58 However, nonrandomized studies have shown that, among patients who are candidates for liver transplantation, mortality was higher in patients treated with continuous renal replacement therapy than in those treated with intermittent dialysis (reviewed in Davis et al. 58 ). In patients with cirrhosis, like in the general population, the optimal treatment of critically ill patients with acute renal failure remains to be determined. Prevention of Acute Renal Failure SBP is a precipitating factor of HRS. 11,21 SBP stimulates the production of the proinflammatory cytokine tumor necrosis factor (TNF- ), which is known to induce vasorelaxant mechanisms in arteries. 20 Thus, the SBPinduced, TNF- mediated activation of vasodilator mechanisms may enhance preexisting systemic vasodilation (and arterial hypovolemia) and precipitate HRS. It has been hypothesized that in patients with SBP, simultaneous intravenous administration of albumin and antibiotics could prevent the sepsis-induced decrease in effective arterial blood volume and resulting HRS. One hundred twenty-six patients with SBP were randomly assigned to receive either cefotaxime alone or cefotaxime plus intravenous albumin. 11 Albumin was given at a dose of 1.5 g/kg on diagnosis of SBP followed by 1 g/kg on day 3. The proportion of patients who developed HRS was significantly lower in the cefotaxime-plus-albumin group than in the cefotaxime group. In addition, the in-hospital mortality rate was significantly lower in the cefotaximeplus-albumin group than in the cefotaxime group. Finally, at 3 months, the mortality rate was still significantly lower in the cefotaxime-plus-albumin group than in the cefotaxime group (22% vs. 40%, respectively). 11 HRS is the main cause of in-hospital death in patients with cirrhosis and severe acute alcohol-induced hepatitis. 22 Severe acute alcohol-induced hepatitis may precipitate HRS by activating the above-mentioned TNF- mediated stimulation of vasorelaxant mechanisms. A randomized, placebo-controlled study investigated the effects of pentoxifylline, a TNF- synthesis inhibitor, in patients with severe alcohol-induced hepatitis. 22 In-hospital mortality was significantly lower in the pentoxifylline group than in the placebo group. 22 Prevention of radiocontrast-agent induced renal failure has been studied in noncirrhotic patients at risk of radiocontrast-agent induced acute renal failure (i.e., mainly patients with chronic renal failure or with diabetes). 26 Since renal vasoconstriction plays a critical role in the development of this renal failure, randomized studies have evaluated the effects of vasodilators. However, the studies have shown that calcium-channel antagonists, captopril, adenosine antagonists, dopamine, prostaglandin E 1, atrial natriuretic peptide, and blockade of endothelin receptors were not effective in preventing radiocontrast-agent induced nephropathy (reviewed in Safirstein et al. 59 ). On the other hand, saline alone has been shown to provide better protection against radiocontrast-induced nephropathy than saline plus mannitol or furosemide. 26,59 Finally, a randomized study in 83 clinically stable patients with a mean serum creatinine level of 220 mol/l has shown that the addition of the oral antioxidant acetylcysteine to intravenous saline significantly reduced the frequency of contrast-induced renal failure (2% in the saline-plus-acetylcysteine group vs. 21% in saline-alone group). 60 In this study acetylcysteine was administered orally at a dose of 600 mg twice daily on the day before and on the day of administration of the contrast. 60 There are no studies on the prevention of radiocontrast-agent induced nephrotoxicity in clinically stable patients with cirrhosis. However, the use of intravenous saline and acetylcysteine in patients with cirrhosis and chronic renal failure or diabetes seems to be indicated. Conclusions In patients with cirrhosis, acute renal failure is mainly due to prerenal failure and tubular necrosis (due to ischemic insult to the renal tubules). However, except for type 1 HRS, little is known about the incidence, natural course, and treatment of the different causes of cirrhosisassociated acute renal failure. Studies are needed, in

10 242 MOREAU AND LEBREC HEPATOLOGY, February 2003 particular for severe sepsis, radiocontrast-induced nephrotoxicity, and renal replacement therapies. Acknowledgment: The authors thank A. Truskolaski for his expert technical assistance. References 1. Abuelo JG. Diagnosing vascular causes of renal failure. Ann Intern 1995; 123: Brady HR, Singer GG. Acute renal failure. Lancet 1998;346: Thadhani R, Pascual M, Bonventre JV. Acute renal failure. N Engl J Med 1996;334: Arroyo V, Ginès P, Gerbes AL, Dudley FJ, Gentilini P, Laffi G, Reynolds TB, et al. Definition and diagnostic criteria of refractory ascites and hepatorenal syndrome in cirrhosis. HEPATOLOGY 1996;23: Cardenas A, Ginès P, Uriz J, Bessa X, Salmeron JM, Mas A, Ortega R, et al. Renal failure after upper gastrointestinal bleeding in cirrhosis: incidence, clinical course, predictive factors, and short-term prognosis. HEPATOLOGY 2001;34: Cabrera J, Arroyo V, Ballesta AM, Rimola A, Gual J, Elena M, Rodés J. Aminoglycoside nephrotoxicity in cirrhosis. Value of urinary 2 -microglobulin to discriminate functional renal failure from acute tubular damage. Gastroenterology 1982;82: Bataller R, Ginès P, Guevara M, Arroyo V. Hepatorenal syndrome. Semin Liv Dis 1997;17: Moore K. The hepatorenal syndrome. Clin Sci 1997;92: Wong F, Blendis L. New challenge of hepatorenal syndrome: prevention and treatment. HEPATOLOGY 2001;34: Moreau R. Hepatorenal syndrome in patients with cirrhosis. J Gastroenterol Hepatol 2002;17: Sort P, Navasa M, Arroyo V, Aldeguer X, Planas R, Ruiz Del Arbol L, Castells L, et al. Effect of intravenous albumin on renal impairment and mortality in patients with cirrhosis and spontaneous bacterial peritonitis. N Engl J Med 1999;341: Moreau R, Asselah T, Condat B, De Kerguenec C, Pessione F, Bernard B, Poynard T, et al. Comparison of the effect of terlipressin and albumin on arterial blood volume in patients with cirrhosis and tense ascites treated by paracentesis: a randomised pilot study. Gut 2002;50: Moreau R, Durand F, Poynard T, Duhamel C, Cervoni J-P, Ichaï P, Abergel A, et al. Terlipressin in patients with cirrhosis and type 1 hepatorenal syndrome: a retrospective multicenter study. Gastroenterology 2002; 122: Moreau R, Gaudin C, Hadengue A, Braillon A, Roulot D, Bacq Y, Lebrec D. Renal hemodynamic in patients with cirrhosis: relationship with ascites and liver failure. Nephron 1993,65: Bosch J, Laine LA, Burroughs AK, Marcon N, Nevens F, Silvain C, Vinel JP. Baveno III consensus statements: treatment of the acute bleeding episode. In: de Franchis R, ed. Portal hypertension III. Proceedings of the third Baveno international consensus workshop on definitions, methodology and therapeutic strategies. London, Edinburgh, Malden, Carlton, Paris: Blackwell Science, 2001; Bernard B, Cadranel JF, Valla D, Escolano S, Jarlier V, Opolon P. Prognostic significance of bacterial infection in bleeding cirrhotic patients: a prospective study. Gastroenterology 1995;108: De Lédinghen V, Heresbach D, Fourdan O, Bernard P, Liebaert-Bories MP, Nousbaum JB, Gourlaouen A, et al. Anti-inflammatory drugs and varceal bleeding: a case-control study. Gut 1999;44: American College of Chest Physicians-Society of Critical Care Medicine Consensus Conference. Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. Crit Care Med 1992;20: Fernandez J, Navasa M, Gomez J, Vila J, Arroyo V, Rodés J. Bacterial infections in cirrhosis: epidemiological changes with invasive procedures and norfloxacin prophylaxis. HEPATOLOGY 2002;35: Navasa M, Follo A, Filella X, Jiménez W, Francitorra A, Planas R, Rimola A, et al. Tumor necrosis factor and interleukin-6 in spontaneous bacterial peritonitis in cirrhosis: relationship with the development of renal impairment and mortality. HEPATOLOGY 1998;27: Ginès A, Escorsell A, Ginès P, Salo J, Jimenez W, Inglada L, Navasa M, et al. Incidence, predictive factors, and prognosis of the hepatorenal syndrome in cirrhosis with ascites. Gastroenterology 1993;105: Akriviadis E, Botla R, Briggs W, Han S, Reynolds T, Shakil O. Pentoxifylline improves short-term survival in severe acute alcoholic hepatitis: a double-blind, placebo-controlled trial. Gastroenterology 2000;119: Clive DM, Stoff JS. Renal syndromes associated with nonsteroidal antiinflammatory drugs. N Engl J Med 1984;310: FitzGerald GA, Patrono C. The COXIBS, selective inhibitors of cyclooxygenase-2. N Engl J Med 2001;345: Guevara M, Abecasis R, Jimenez W, Rios H, Terg R. Effect of celecoxib on renal function in cirrhotic patients with ascites. A pilot study [Abstract]. J Hepatol 2002;36(Suppl 1): Barrett BJ, Parfrey PS. Prevention of nephrotoxicity induced by radiocontrast agents. N Engl J Med 1994;331: Guevara M, Fernandez-Esparach G, Ginès P, Calahorra B, Ozdogan O, Testuccia T, Montaya X, et al. Effect of radiocontrast agents on renal function in patients with cirrhosis and ascites [Abstract]. J Hepatol 2002; 36(Suppl 1): Salmeron J.M, Ruiz Del Arbol L, Ginès A, Garcia-Pagan JC, Ginès, P, Feu F, Claria J, et al. Renal effects of acute isosorbide-5-mononitrate adminitration in cirrhosis. HEPATOLOGY 1993;17: Garcia-Pagan JC, Villanueva C, Vila MC, Albillos A, Genescà J, Ruiz Del Arbol L, Planas R, et al. Isosorbide mononitrate in the prevention of first variceal bleed in patients who cannot receive -blockers. Gastroenterology 2001;121: Gentilini P, Romanelli RG, La Villa G, Maggiore Q, Pesciullesi E, Capelli G, Raggi VC, et al. Effects of low-dose captopril on renal hemodynamics and function in patients with cirrhosis of the liver. Gastroenterology 1993; 104: Gonzalez-Abraldes J, Albillos A, Banarez R, Ruiz Del Arbol L, Moitinho E, Rodriguez C, Gonzalez M, et al. Randomized comparison of long-term losartan versus propranolol in lowering portal pressure in cirrhosis. Gastroenterology 2001;121: Schepke M, Werner E, Biecker E, Schiedermaier P, Heller J, Neef M, Stoffel-Wagner B, et al. Hemodynamic effects of the angiotensin II receptor antagonist irbesartan in patients with cirrhosis and portal hypertension. Gastroenterology 2001;121: Montseny JJ, Meyrier A, Kleinknecht D, Callard P. The current spectrum of infectious glomerulonephritis. Experience with 76 patients and review of the literature. Medicine 1995;74: D Amico G. Renal involvement in hepatitis C infection: cryoglobulinemic glomerulonephritis. Kidney Int 1998;54: Cadranel JF, Soltani N, Guettier C, Fleury D, Poux JM, Hillion D, Vanhille P, et al. Macroscopic hematuria with glomerular bleeding: a peculiar form of acute renal failure in cirrhotic patients [Abstract]. J Hepatol 2002; 36(Suppl 1): Schortgen F, Lacherade JC, Bruneel F, Cattaneo I, Hemery F, Lemaire F, Brochard L. Effect of hydroxyethylstarch and gelatin on renal function in severe sepsis: a multicentre randomised study. Lancet 2001;357: Moreau R, Hadengue A, Soupison T, Kirstetter P, Mamzer MF, Assous M, Sicot C. Clinical, hemodynamic and metabolic characteristics and ICU outcome of septic shock in patients with cirrhosis. Crit Care Med 1992; 20: Nair S, Verma S, Thuluvath PJ. Pretransplant renal function predicts survival in patients undergoing othotopic liver transplantation. HEPATOL- OGY 2002;35: Couser WG. Glomerulonephritis. Lancet 1999;353: Duvoux C, Zanditenas D, Hezode C, Chauvat A, Monin JL, Roudot- Thoraval F, Mallat A, Dhumeaux D. Effects of noradrenalin and albumin in patients with type I hepatorenal syndrome: a pilot study. HEPATOLOGY 2002;36: