Type-1 hepatorenal syndrome (HRS) is a severe and frequently lethal complication of cirrhosis. It is characterized by a rapid deterioration of kidney function related to derangement of systemic hemodynamic with vasodilatation in the splanchnic bed, low blood pressure, insufficient cardiac output, and release of vasoconstrictors with renal ischemia.
Patients at risk of developing HRS are those with decompensated cirrhosis and patients with compensated cirrhosis who develop a severe clinical deterioration, the so-called acute-on-chronic-liver-failure. In many cases a “trigger” can be identified such as gastrointestinal bleeding, alcohol, hepatotoxic drugs, or bacterial infections.
Until recently, when a patient with cirrhosis developed renal dysfunction concomitant to a bacterial infection, the diagnostic work-up and treatment for HRS were postponed until antibiotic treatment was completed and only the renal failure persisted, despite resolution of infection, was called HRS. This strategy was based on the opinion that the pathogenesis of renal failure precipitated by bacterial infections was different from that of HRS precipitated by other triggers. This practice, however, delayed specific HRS treatment with vasopressin analogs. Accordingly, the International Ascites Club removed the recommendation and determined that the diagnosis (and treatment) of HRS could be made before the precipitating infection had resolved. We now know that bacterial infections are a frequent trigger of HRS, probably through a cytokine-mediated impairment of circulatory function, worsening effective hypovolemia. Accordingly, the administration of human albumin to patients with spontaneous bacterial peritonitis (SBP) was shown to decrease mortality from type-1 HRS. This result was not replicated in a trial performed in patients with cirrhosis with bacterial infections other than SBP, making it uncertain whether the recommendation to infuse albumin should be extended to all patients with bacterial infection.
Therefore, a better understanding of the natural history of HRS associated with bacterial infections can lead to an improved therapeutic approach. Moreover, we need to ascertain whether the antibiotics empirically used to treat infections in patients with cirrhosis with normal renal function should be the same as that used in patients with infections associated with HRS.
Finally, as the survival of patients with cirrhosis after infection-related HRS can be insufficient for liver transplantation, the new therapeutic strategies should ensure longer survivals.
In this issue, Barreto et al. studied a cohort of 70 patients with cirrhosis with HRS associated with bacterial infections. Their study's main objectives were to evaluate the renal response to antibiotic therapy, short-term survival, and predictors of treatment failure and survival.
The use of routinely recommended antibiotic therapy failed to resolve renal dysfunction in two-thirds of patients, suggesting that the mechanism of renal damage in such patients is more complex than that causing HRS in the absence of bacterial infections (Fig. 1), or that the efficacy of antibiotics is impaired by renal dysfunction. Unfortunately, the investigators did not report on whether doses and types of antibiotics were adjusted according to the loss of renal filtration capacity.
The independent predictors of irreversible HRS were older age, Child-Pugh and Model for Endstage Liver Disease (MELD) scores, high bilirubin levels, and nosocomial infection. This confirms that the syndrome's severity is predominantly due to the severity of the bacterial infection and to liver dysfunction, rather than to renal impairment. This suggests that the best results of treatment with terlipressin and albumin could be obtained by starting therapy before bilirubin and MELD scores reach high levels.
The most important finding of Barreto et al.'s study was the high prevalence of nosocomial infections associated with irreversible renal failure. This could be explained by a high prevalence of infections due to multidrug-resistant bacteria characteristic of nosocomial infections. If this hypothesis is confirmed, the treatment of such patients could be improved by: (1) first-line (empiric) therapy with very broad-spectrum antibiotics; (2) collecting samples of urine, blood, and pathological fluids before starting antibiotic therapy to increase the possibility of isolating an infecting organism; (3) promptly replacing ineffective antibiotics with those identified as effective by cultures; (4) using a de-escalation policy (replacing empiric broad-spectrum antibiotics with antibiotics with a narrower spectrum of activity once culture results and antibiotic susceptibility are established).
Barreto et al.'s article also gave information regarding patient survival. Unfortunately, this information is outdated, as only a few patients were treated with vasoconstrictors.
Nevertheless, even considering the effectiveness of the current standard of treatment, the association of renal failure with bacterial infection in patients with cirrhosis remains one of the most serious complications occurring in patients with portal hypertension.
High serum bilirubin levels, failure of antibiotic therapy, and septic shock were predictors of death. This confirms that, in addition to the severity of liver failure, bacterial infection is an important determinant of the high mortality rate in patients with type-1 HRS. This justifies that patients with HRS and the highest risk of death (for example, patients needing mechanical ventilation or intubation, those needing blood pressure support, or patients with multiorgan failure) should be admitted to an intensive care unit.
Given the low efficacy of current treatments, it is important to develop a good strategy to prevent HRS. This strategy could include: (1) antibiotic prophylaxis for patients who had an episode of SBP; patients with low protein in ascites associated with renal dysfunction or severe liver failure,[9, 10] or patients with variceal bleeding; (2) reducing the risk of nosocomial infections by enforcing infection control programs including avoidance of urinary catheterization and to limit the use of intravenous lines; and (3) albumin administration to patients with a new diagnosis of SBP. The effectiveness of such measures has already been validated.
In conclusion, the findings of Barreto et al. are a first important step toward improving the outcome of patients with cirrhosis with type-1 HRS combined with bacterial infection. The study provides support to the recommendation that in patients with cirrhosis who develop renal failure mimicking type-1 HRS, bacterial infections, particularly SBP, should be rapidly investigated. If an infection is uncovered, great effort should be spent in resolving it as quickly as possible and restoring renal function. Blood and urine samples should be collected for cultures and empiric antibiotic therapy must be based on the local microbiological flora and based on the setting of infection acquisition (community acquired, nosocomial, or healthcare-associated infections). Early treatment with terlipressin and albumin in patients without contraindications is also highly recommended. Finally, a correct prophylactic strategy to prevent bacterial infections in patients with advanced cirrhosis is recommended in order to decrease new episodes of HRS.
Francesco Salerno, M.D. and Valentina Monti
Policlinico IRCCS San Donato Medicina Interna ed Epatologia Università di Milano, Italy