Albumin is predominantly an extravascular protein, and its serum concentration is -40 g/L, suggesting a total intravascular mass of -120 g [5]. The interstitial concentration is lower (14 g/L) and varies among anatomic regions. However, the total extravascular mass is -160 g [5]. Some is easily mobilized from loose interstitial tissues, whereas the remainder is tightly bound. Albumin appears to circulate from the intravascular to extravascular spaces, and the transcapillary escape rate is determined by the capillary and interstitial free albumin concentrations, microvascular permeability to albumin, movement of solvents and solutes, and transcapillary electrical charge. In patients with hypoalbuminemia (especially when associated with inflammation or sepsis), whose capillaries are known to be hyper-permeable, leakage of albumin into the interstitial space draws water and produces edema [6].

Albumin has a strong negative charge, but binds weakly and reversibly to both cations and anions. Therefore, it functions as a transport molecule for a large number of metabolites, including fatty acids, ions, thyroxine, bilirubin, and amino acids (Table 1). Albumin also binds covalently and irreversibly with d-glucose and d-galactose. The glycosylation of albumin, which is to a certain extent age-dependent, affects its charge and, therefore, may influence capillary permeability characteristics [5].

Albumin is the major extracellular source of thiols. These sulphydryl groups are scavengers of reactive oxygen and nitrogen species. Albumin can also limit the production of reactive oxidative species by binding free copper, an ion known to be particularly important in accelerating the production of free radicals. In sepsis, the administration of human albumin led to significantly increased levels of total plasma thiol levels. Unlike albumin concentrations, however, which fell significantly between 5 minutes and 4 hours following administration, thiol remained significantly elevated for up to 18 hours following albumin administration [7]. These results suggest that the increase in plasma protein thiols associated with albumin administration is sustained long-term compared with plasma albumin levels, which is indicative of a beneficial albumin-mediated thiol exchange in the plasma of these septic patients. Also, albumin may in this way influence redox balance, which has several important implications for other indices of critical illness, including capillary permeability, cell signaling processes, and drug metabolism and transport [4].

Albumin's ability to reduce injury to the endothelium caused by reactive oxygen and nitrogen species means that it may stabilize the endothelium and help to maintain capillary permeability. Albumin also interferes with neutrophil adhesion to the capillary endothelium [8], thereby reducing inflammation and aiding the maintenance of endothelial integrity.

Drugs with which albumin interacts are clinically significant owing to their highly protein-bound state and low margins of safety, and include warfarin, phenytoin, non-steroidal anti-inflammatory drugs, digoxin, midazolam, thiopental, and several antibiotics. The distribution volume of drugs bound to albumin may increase in hypoalbuminemia, reducing their efficacy [5]. Administration of mixtures of loop diuretics with albumin has therefore been advocated, although this has been shown to be ineffective in cirrhotic patients with ascites.

The standard treatment for cirrhotic ascites is sodium restriction and diuretic therapy. One randomized, controlled trial assessed the effects of albumin plus standard diuretic therapy in cirrhotic patients with ascites; weekly infusions of 25 g albumin produced a significantly better diuretic response, shorter hospital stays, and a lower likelihood of readmission to hospital than treatment with standard therapy [17]. Suppression of the activity of anti-natriuretic systems, particularly the renin-angiotensin-aldosterone system, probably accounts for an increase in the natriuretic response to diuretics with repeated albumin infusions [18]. Survival, however, was not affected by the addition of albumin [6]. Moreover, compared with the simple performance of paracentesis in a day-care unit, the logistic problems of intravenous albumin administration on a weekly basis, and its lack of cost-effectiveness, render this indication unjust and impractical in clinical practice [19]. Infusions of albumin plus diuretic therapy, therefore, cannot be recommended as the standard of care for these patients [6].

To date, two different situations that may further impair circulatory function in cirrhotic patients with ascites have been identified: large-volume paracentesis and spontaneous bacterial peritonitis (SBP).

The administration of albumin to patients with cirrhosis and ascites causes an increase in total blood volume, followed by a moderate reduction, but not normalization, of the activity of vasoconstrictor and anti-natriuretic systems. These circulatory changes are associated with favorable effects on renal function. However, these renal effects are modest and limited only to patients with normal or slightly impaired renal function, whereas patients with severe renal dysfunction show no beneficial response [29-32]. Albumin infusion alone fails to consistently improve circulatory and renal functions because albumin cannot increase effective arterial blood volume efficiently due to the extreme splanchnic vasodilatation present in these cirrhotic patients [1].

Hepatorenal syndrome (HRS) is characterized by very low arterial pressure and total systemic vascular resistance, marked over-activity of vasoconstrictor factors, and marked arterial vasoconstriction in the kidney and other vascular territories (muscle, skin, and brain) [33]. For many years, HRS was considered a terminal irreversible event in patients with decompensated cirrhosis. However, it was demonstrated that patients with type 1 HRS may be effectively treated with a combination of vasoconstrictor and plasma volume expansion [34]. The use of albumin appears to increase the efficacy of vasoconstrictor drugs. Two studies have shown that the treatment of cirrhotic patients with HRS for several days or weeks with a combination of vasoconstrictors and plasma volume expansion with albumin results in a marked improvement in circulatory and renal functions in most cases, with normalization of plasma levels of vasoconstrictor factors and serum creatinine [35,36]. However, the need for a plasma expander agent as a co-therapy remains unclear. The administration of vasoconstrictors with albumin has been shown to reverse type 1 HRS and normalize renal function in 60-70% of treated patients. However, these studies included only small numbers of patients, some of whom were not randomized, and the impact on long-term (> 1 month) survival has not been shown. Available data on the treatment of type 2 HRS are much scarcer than for type 1 HRS [19]. Whether albumin is required to achieve the beneficial effect of vasoconstrictor therapy in HRS is not known. However, albumin likely improves the therapeutic efficacy of vasoconstrictors, as the improvement in circulatory and renal functions is more marked in patients treated with terlipressin and albumin than that in patients treated with terlipressin alone. Given that albumin has volume-expanding, ligand-binding, and antioxidant properties, it seems prudent to use albumin infusions in the treatment of HRS unless there is evidence that albumin actually does some harm. Currently, the AASLD recommends that albumin infusion plus administration of vasoactive drugs such as octreotide and midodrine should be considered in the treatment of type 1 HRS (level II-1).

The recent development of the molecular absorbent recirculating system (MARS), an albumin dialysis that removes albumin-bound and water-soluble substances in patients with acute and chronic liver failure, is a clinical application for albumin, based on its capacity to remove water-insoluble substances [37]. The MARS system has been shown to be very effective in the treatment of hepatic encephalopathy [38] and intractable pruritus. Furthermore, it markedly improves circulatory and renal functions in patients with cirrhosis and ascites [39,40]. Several randomized controlled trials have been performed to assess the use of MARS in patients with acute and chronic liver failure. Despite encouraging results, the role of MARS in the management of patients with end-stage cirrhosis remains unsettled because all reports published to date were based on small studies. Furthermore, MARS is an expensive treatment that requires skilled personnel. The use of MARS in patients with acute or chronic liver failure showed an improvement in hepatic encephalopathy, but not in systemic hemodynamic parameters or renal function [41]. Future studies will need to define the indications for MARS treatment, patient selection, and predictors for response.

Hyponatremia is common complication in patients with advanced cirrhosis. Hyponatremia is usually the result of vasopressin overactivity in response to a reduction in the effective arterial blood volume. As albumin is capable of refilling the effective arterial blood volume, albumin infusions have been used in the treatment of this complication [42]. To date, however, no study has compared the efficacy of albumin infusions with those of other volume expanders in the treatment of hyponatremia. Diuretic-induced electrolyte abnormalities can also lead to the development of hepatic encephalopathy in patients with cirrhosis. In a study conducted in cirrhotic patients with diuretic-induced hepatic encephalopathy, the infusion of albumin was comparably effective to the infusion of a colloid solution. These infusions caused similar reductions in plasma ammonia concentrations and increases in urinary ammonia excretion. The improvement of hepatic encephalopathy grade, although initially observed in both groups of patients, was only sustained at 72 hours in the group who received albumin infusions [43].