Albumin as a drug: its biological effects beyond volume expansion

Abstract

• •Albumin is the most abundant and perhaps most important protein in human blood. Research has identified many of albumin’s possible roles in modulating acid-base balance, modifying inflammation, maintaining vascular endothelial integrity, and binding endogenous and exogenous compounds.
• •Albumin plays a key role in the homeostasis of vascular endothelium, offering protection from inflammation and damage to the glycocalyx.
• •Albumin binds a diverse range of compounds. It transports, delivers and clears drugs, plus it helps with uptake, storage and disposal of potentially harmful biological products.
• •The biological effects of albumin in critical illness are incompletely understood, but may enhance its clinical role beyond use as an intravenous fluid. In this article, we summarise the evidence surrounding albumin’s biological and physiological effects beyond its use for plasma volume expansion, and explore potential mechanistic effects of albumin as a disease modifier in patients with critical illness.

Albumin is widely used in critical care and other conditions to expand plasma volume, replace losses, and restore serum albumin levels toward normal in hypoalbuminaemic states.1, 2, 3 This use is driven by the concept that albumin’s oncotic pressure effects have primacy over any other biological, physiological or clinical effects.¹ However, as a naturally occurring protein, albumin has other important biological properties beyond its oncotic and volume expansion effects. These effects may enhance its clinical role beyond use as an intravenous fluid.⁴ Such effects should be understood by clinicians to fully appreciate the potential effects of human albumin solution therapy.

Albumin physiology

Human albumin constitutes some 50% of plasma protein present in normal healthy individuals. It is a 66-kilodalton protein, which is small compared with other plasma proteins.⁵ Albumin is highly water soluble, has an elliptical shape and has low intrinsic viscosity. It is synthesised in polysomes bound to the endoplasmic reticulum of hepatocytes. Albumin is not stored in the liver and there is no reserve for release on demand.⁶ However, under basal physiological circumstances only 20–30% of hepatocytes produce albumin; so with factors such as change in interstitial colloid oncotic pressure, synthesis can be increased if required by a factor of 200–300%.³

Despite being the main protein in plasma, albumin is predominantly an extravascular protein. Specifically, its serum concentration is about 40 g/L, equating to a total intravascular mass of about 120 grams. In contrast, its interstitial concentration is lower (about 14 g/L) and varies between different anatomical regions. However, its estimated total extravascular mass is about 160 grams. A proportion of this albumin can be easily mobilised from loose interstitial tissue, while some is tightly bound, particularly in the skin. There appears to be a circular flow of albumin from the intravascular to extravascular space, followed by return to the circulation via lymphatic vessels. Such movement has been objectively measured and represents a circulation halflife of about 16–18 hours.⁵

Evidence also suggests that albumin has effects independent of its oncotic effect. These include protection of the glycocalyx,⁷ improvement in endothelial integrity,⁸ inhibition of endothelial apoptosis,⁹ modulation of nitric oxide pathways,10, 11 anti-inflammatory effects,12, 13 antioxidant effects,14, 15, 16 modulation of acid-base status,¹⁷ binding of drugs and other plasma substances,18, 19 activation of intracellular processes,20, 21 regulation of electrolyte shifts, and changes in intracellular volume.22, 23, 24 In this review, we examine these biological effects of albumin that go beyond its intravascular volume expansion properties.

Protection of the glycocalyx

The glycocalyx is a complex gel that lies between flowing blood and the endothelial cell wall, and interacts with both plasma proteins and lipids.²⁵ It contributes to the epithelial barrier, and its degradation is sufficient to increase lung permeability.²⁶ Similarly, the endothelial glycocalyx is known to contribute to critical vascular functions, such as endothelial barrier integrity, transduction of shear stress, regulation of vascular tone, and inhibition of leucocyte adhesion.27, 28 Damage to the glycocalyx structure causes increased vascular permeability, leading to tissue oedema, uncovering adhesion molecules for blood leucocyte and platelet activation, and initiating local inflammation plus proaggregatory and procoagulatory conditions.29, 30

Albumin attenuates glycocalyceal shedding during cold ischaemia and reperfusion, reducing interstitial oedema and intracoronary adhesion of leucocytes.⁷ This makes it an important and beneficial ingredient of a preservation solution used during heart transplantation.⁷ In an isolated perfused organ model, albumin, in contrast to artificial colloids, leads to a significant increase in coronary flow through its interaction with the glycocalyx.³¹ Furthermore, human albumin, through its interaction with glycocalyceal hyaluronan,⁸ improves endothelial integrity, alleviating postischemic intracoronary retention of granulocytes. This effect is confirmed by a morphologically intact endothelial glycocalyx and a markedly decreased coronary venous release of glycocalyx constituents.³² This protective effect of albumin might be mediated by a protein-bound substance that inhibits matrix metalloproteinase cleavage of the endothelial glycocalyx, such as the lipid mediator sphingosine 1-phosphate.⁴

The role of the glycocalyx in fluid shift across the endothelium, as a determinant of hydraulic conductivity and in the revised Starling model, is well recognised.³³ A low- protein environment has long been known to cause a rapid breakdown, or shedding, of the endothelial glycocalyx.³⁴ In preclinical studies, this phenomenon is independent of the effect on osmotic pressure, and albumin is more effective than semisynthetic colloids at preserving and restoring the endothelial glycocalyx, reducing vascular permeability, and reducing platelet and leucocyte adhesion.35, 36 However the ability for albumin to protect glycocalyx integrity in clinical studies has yet to be examined.

Protection of endothelium

Albumin inhibits endothelial apoptosis9, 37 via a G-coupled protein PI3K (phosphoinositide 3-kinases)-dependent mechanism.³⁸ It modulates arachidonic acid release and membrane fluidity, and protects against ischaemia and reperfusion injury.21, 39 Albumin inhibits apoptosis triggered by oxidative stress by scavenging reactive oxygen species in cultured macrophages, neutrophils, lymphocytes, and kidney tubular epithelial cells.14, 15, 16 This protective activity depends on radical scavenging, primarily through albumin’s cysteine residue at position 34 (Cys34) and its binding of lysophosphatidic acid, an abundant serum lipid.14, 15, 16, 40

Albumin prevents flow-induced vasoconstriction⁴¹ and suppresses angiotensin-converting enzyme activity,⁴² and its administration increases angiopoietin-1 to angiopoietin-2 ratio⁴³ — an effect which has been associated with endothelial integrity.⁴⁴ It significantly inhibits tumour necrosis factor-α (TNF-α)-induced adhesion of THP-1 (human leukaemia monocytic cell line) cells and dose-dependently inhibits TNF-α-induced mRNA and protein expression of vascular cell adhesion molecule-1.²⁰ Moreover, it inhibits activation and nuclear translocation of nuclear factor kappa-light-chain-enhancer of activated B cells (a transcription factor) in a dose-dependent manner.

Since excessive endothelial apoptosis may contribute to vascular disease,⁴⁵ it is possible that albumin’s protective effect on the vascular endothelium is due to its antiapoptotic activity.9, 37 Albumin may directly influence vascular integrity, by binding to the interstitial matrix and subendothelium and altering the permeability of these layers to large molecules and solutes.²²

In patients with congenital analbuminaemia — a rare, inherited, autosomal recessive disorder with about 90 patients known worldwide — there is low or complete absence of albumin. However, due to a compensatory increase in other plasma proteins, these patients have mild oedema, reduced blood pressure and chronic fatigue.⁴⁶ However, even very low concentrations of albumin, as seen in these patients, help maintain endothelial barrier function.⁸

Studies have demonstrated a protective effect of 4% human serum albumin treatment on endothelial dysfunction which is achieved by inhibiting inflammatory and oxidative stress pathways induced by endotoxins in animal models.⁴⁷ Multiple mechanisms have been proposed to explain these effects (Figure 1), but there is a lack of data to support any such improvement in endothelial function in critically ill humans — an aspect of human albumin biology that requires investigation.

Figure 1.

Mechanisms by which albumin protects endothelial cells

Nuclear factor Kb = nuclear factor kappa-light-chain-enhancer of activated B cells; PI3K = phosphoinositide 3-kinases; TNF-α = tumour necrosis factor-α; VCAM-1 = vascular cell adhesion molecule-1.

Nitric oxide modulation

Human serum albumin plays a key role in storing, scavenging and transporting nitric oxide in humans.10, 11 Hypoalbuminaemia is viewed as a causal factor in the pathogenesis of endothelial dysfunction and cardiovascular disease in patients with proteinuric chronic kidney disease; low albumin levels have been shown to increase endothelial nitric oxide production and decrease vascular sensitivity to nitric oxide.⁴⁸

Albumin is the major extracellular source of reduced sulphydryl groups, which are avid scavengers of reactive oxygen and nitrogen species, especially superoxide hydroxyl and peroxynitrite radicals. Albumin can also limit the production of these reactive species through its binding of free copper, an ion known to be particularly important in accelerating production of free radicals.⁵ Similarly, the anticoagulant effect of albumin is possibly due to its capacity to bind nitric oxide to form S-nitrosothiols, thereby inhibiting the rapid inactivation of nitric oxide and allowing prolongation of its inhibitory effects on platelets.⁴⁹

Influence on inflammation

When inflammation-related reactive oxygen species are absorbed, stored and transported by albumin, albumin’s Cys34 becomes oxidised. This results in a structural change in the albumin molecule — from its reduced form mercaptoalbumin (HMA) to its oxidised form nonmercaptoalbumin (HNA). In a small study of dialysisdependent patients receiving oral calcitriol for secondary hyperparathyroidism due to chronic kidney disease, a correlation between HNA to HMA ratios and serum levels of C-reactive protein and interleukin (IL)-6 was found.¹² In another observational study, which included 87 patients admitted to a Brazilian intensive care unit (ICU), levels of serum albumin were inversely associated with inflammatory cytokines including IL-6, IL-7, IL-8, IL-10, TNF-α and interferon-α.¹³ These data support the hypothesis that albumin may be a significant component of the proteindependent response to oxidative stress during inflammation. Thus, hypoalbuminaemia may result in an overall reduction in plasma antioxidant capacity.

In healthy individuals, serum albumin is not glycated. Glycation of albumin has been associated with inflammation and chronic diseases such as diabetes, kidney disease and eye disease.⁵⁰ Levels of glycated albumin have also been correlated with progression of these diseases, their complications, and even mortality.⁵¹ In patients with diabetes, levels of glycated haemoglobin and glycated albumin closely correlate. However, as albumin has a shorter half-life than haemoglobin, changes in glycated albumin reflect short term glycaemic control (2–3 weeks) as opposed to glycated haemoglobin, which reflects long term glycaemic control (2–3 months). On the other hand, while glycated albumin may be the result of hyperglycaemia, there is also evidence that it may, itself, further exacerbate insulin resistance.⁵²

Influence on acid–base status

Acute critical illness is often characterised by acidosis. Albumin is a weak acid comprising the majority of total extracellular acid in human plasma. According to Stewart’s model of acid-base physiology, such acids are an independent variable controlling acid–base status.¹⁷ During episodes of acute inflammation, albumin behaves as a negative acute phase reaction protein. The subsequent hypoalbuminaemia and reduction in total extracellular acid contributes to metabolic alkalosis and attenuates any acute illness-associated metabolic acidosis. This phenomenon is often seen in patients with chronic critical illness long after their initial acidifying insult (eg, lactic acidosis in septic shock) has resolved. In theory, the intravenous administration of albumin should result in an increase in serum albumin concentration and thus total extracellular acid, causing acidification of blood. However, albumin for infusion is presented as sodium chloride solution at 4-5% (iso- oncotic) concentration or chloride-poor solution at 20–25% (hyperoncotic) concentration. Thus 4-5% albumin solutions can be expected have different acid-base effects compared with 20–25% albumin solutions. In this regard, a secondary analysis of a sepsis trial in African children revealed that fluid resuscitation with 4% albumin, compared with no bolus fluid therapy, resulted in hyperchloraemic acidosis.⁵³ In contrast, infusion of hyperoncotic albumin, with its low chloride content, results in a fall in serum chloride and no overall change in blood pH or base excess.⁵⁴ Studies directly comparing 4-5% and 20% albumin solutions also confirm that the use of 4-5% preparations results in increased serum chloride levels and a tendency for metabolic acidosis.55, 56 Thus the concentration of albumin solutions has a differential effect on acid-base status. Finally, there are strong relationships with inflammation and acid-base status — albumin appears to have independent influence over both. However, the relationship between these acid–base variables and inflammation is not entirely understood or explained by illness severity alone.¹³

Binding of drugs and other ligands

Albumin has a strong negative charge, enabling it to reversibly bind other charged particles. In particular, it has strong binding capacity for water, calcium, sodium and trace elements. Albumin also plays a vital role in the storage and transport of endogenous compounds including fatty acids, bilirubin and hormones, plus a range of (exogenous) drugs. Albumin’s physical structure is based around three homologous domains (I, II and III) with each containing two subdomains (A and B). This modular structure is held together by disulphide bonds and provides multiple sites for the binding of ligands (Figure 2). Sudlow and colleagues first described the most important drug-binding sites in 1975, after identifying two specific locations in subdomains IIA and IIIA based on patterns of fluorescent probe displacement by competitive drugs.18, 19 Albumin also possesses esterase capability, allowing it to hydrolyse ester bonds of bound drugs and subsequently modify their pharmacological effects. Non-ester drugs tend to bind albumin at Sudlow site I, whereas drugs with an ester bond have a greater affinity at Sudlow site II.⁵⁸

Figure 2.

Physical structure of human albumin, including its domains (I-III), subdomains (A and B) and Sudlow binding sites (I and II)*

* Reproduced from Lakshmi et al,⁵⁷ under the Creative Commons Attribution 3.0 Unported license (available at https://creativecommons.org/licenses/by/3.0)

The combination of albumin’s binding capacity, esterase activity and abundance in human blood is responsible for its influence on the clinical efficacy and safety of several drugs. For example, drugs such as warfarin and phenytoin compete for the same albumin binding sites, meaning they will displace each other, and this renders their subsequent pharmacodynamic effects unpredictable and potentially harmful. If such drugs are used in critical illness, they should be titrated carefully according to established therapeutic drug monitoring procedures. In addition, the use of common drugs (such as non-steroidal anti-inflammatories and oral hypoglycaemic agents) alongside fluoroquinolone antibiotics results in an increased free fraction of antibiotics,⁵⁹ which may result in increased risk of important clinical adverse effects such as peripheral neuropathy.⁶⁰ A lot of the data concerning drug-binding interactions with albumin come from in vitro experiments, as it is difficult to replicate such measurements in vivo, especially in critical illness. However, there have been observations of the potential increased free drug fraction in hypoalbuminaemic states, such as faster onset of sedation in patients given midazolam.⁶¹

Many other factors influence drug binding to serum albumin, including acid-base status, inflammatory states (acute and chronic), sex, age and pathological condition. In the context of critical illness, however, it is unlikely that any one factor has a large enough effect on drug binding to cause any significant clinical impact.⁶²

Significance of hypoalbuminaemia

Hypoalbuminaemia is a common feature of chronic kidney diseases and chronic liver disease and, when combined with episodes of systemic inflammation, has been associated with increased mortality.⁶³,⁶⁴ Recently, two trials have found that albumin treatment reduced systemic inflammation and cardiocirculatory dysfunction in patients with decompensated cirrhosis.⁶⁵ Although hypoalbuminaemia is commonly seen in patients with acute conditions such as burns and sepsis, this is more likely to be because of increased transcapillary escape and the dilutional effects of exogenous fluid administration rather than increased consumption or clearance. In this regard, in vivo kinetics studies using iodine-labelled albumin have demonstrated increased transcapillary escape during sepsis and after major surgery.66, 67 While malnutrition is characterised by hypoalbuminaemia, observational data in burns patients associates hypoalbuminaemia with increasing inflammation as measured by C-reactive protein rather than the provision of adequate nutrition.⁶⁸ In fact, acute systemic inflammation actually stimulates albumin production to increase the body’s functional synthetic rate,⁶⁹ but such increased production is unable to keep up with transudation, catabolism or both. Clinical measurements of serum albumin concentration, therefore, do not reflect the complexity of its bodily kinetics during acute inflammatory states or the variety of aetiological factors at play.⁷⁰

Hypoalbuminaemia commonly seen in critical illness can result in an increased free fraction of drugs that bind albumin. This may increase a drug’s volume of distribution and reduce its efficacy. In addition, some drugs rely on albumin for more than binding and transport. Frusemide, for example, binds to albumin, which appears to facilitate its secretion into the tubular lumen. Clinical trials have shown enhanced diuresis and improved oxygenation when frusemide is given in combination with albumin in both chronic and acute disease states.71, 72 Other physiological factors, such as temperature and pH, have also been shown to alter the structure of binding sites and their subsequent ability to bind drugs.13, 73

Current clinical use of albumin

Albumin is clinically available as iso-oncotic albumin (4% or 5%) or hyperoncotic albumin (20% or 25%). Hyperoncotic albumin contains minimal sodium and chloride content and is available in smaller volumes. Current albumin solutions have excellent safety records, and substantial improvement in purity and tolerability has been achieved in the past 20 years.⁷⁴

Albumin is commonly used for fluid resuscitation in critically ill patients. The Saline versus Albumin Fluid Evaluation (SAFE) study, which randomly assigned 6997 critically ill patients to 4% albumin or isotonic sodium chloride solution for fluid resuscitation during the 28 days after ICU admission, found no differences in 28-day all-cause mortality, ICU or hospital length of stay, or duration of mechanical ventilation or renal replacement therapy. However, in the subgroup of 1218 patients with severe sepsis, albumin use was associated with a trend toward reduced mortality.⁷⁵ More recently in the Albumin Italian Outcome Sepsis (ALBIOS) trial, patients with severe sepsis were randomly assigned to receive daily supplementation with 20% albumin or crystalloid fluid alone, but mortality did not differ significantly between groups.⁷⁶ However, in the subgroup of 1121 patients with septic shock at enrolment, there was a lower risk of death for those receiving 20% albumin even after adjustment for baseline covariates, supporting the rationale for use of albumin as a drug in these patients. Large clinical trials examining the use of albumin in septic patients are currently underway (ClinicalTrials.gov identifiers NCT01337934 and NCT03869385), and the results will further inform albumin’s role in patients with sepsis.

Albumin replacement is a common practice in patients with liver failure. Long term albumin administration prolongs overall survival and might act as a disease-modifying treatment in patients with decompensated chronic liver disease⁷⁷ and hepatorenal syndrome.⁷⁸ Similarly, albumin replacement is commonly used in large volume paracenteses, to prevent effective hypovolaemia, where albumin (6–8 grams per litre of fluid removed) is administered.⁷⁹ Finally, albumin is used as a replacement fluid for therapeutic plasma exchanges.⁸⁰

Albumin solutions, however, can also be harmful. In a substudy of a large double-blind randomised controlled trial involving critically ill patients with traumatic brain injury, fluid resuscitation with 4% hypotonic albumin was associated with higher mortality rates than resuscitation with saline.⁸¹ However, this injurious effect did not appear secondary to the albumin molecule itself but rather to the low tonicity of the solution containing it. In particular, when compared with saline or an isotonic albumin solution, hypotonic albumin solution increased intracranial pressure in normal animals, while the other two solutions did not.82, 83

Recently, the use of small-volume resuscitation with hyperoncotic albumin has been investigated⁸⁴ and shown to be safe and effective in ICU patients.⁵⁵ More specifically, in patients after cardiac surgery, it results in less positive cumulative fluid balance and offers several haemodynamic and potential ICU treatment advantages such as reduced risk of postoperative acute kidney injury and requirement of vasopressors.85, 86 A phase 2 randomised controlled trial of 20% albumin resuscitation in cardiac surgery patients is about to start in Australia (Australian New Zealand Clinical Trials Registry registration number ACTRN12620000137998). Despite some early concerns that the plasma volume-expanding effect of albumin relative to that of crystalloids might be decreased under conditions characterised by increased permeability, such as cardiac surgery, this has not been shown to be correct under experimental conditions.⁸⁷ Hyperoncotic albumin causes long-lasting plasma volume expansion of similar magnitude in postoperative patients and volunteers23, 24, 88 and its bolus administration is effective and safe in healthy subjects when compared with other commonly available crystalloids and colloidal solution.⁴³

Finally, there are substantial financial implications with the use of albumin-based solutions. Even though it is available free of cost in Australia, there is societal cost of requiring blood donors, and the cost in countries outside Australia is quite notable when compared with crystalloid solutions.⁸⁹

Conclusion

Albumin may be considered the most important protein in human blood. It plays a key role in the homeostasis of vascular endothelium, offering protection from inflammation and damage to the glycocalyx. Albumin’s ability to bind a diverse range of compounds enables the transport, delivery and clearance of drugs, as well as the uptake, storage and eventual disposal of potentially harmful biological products. Its interplay with acid–base balance and inflammatory processes is only partially understood, with data on this role mainly coming from basic science experiments and clinical trials of patients with chronic disease. Pilot studies suggest a beneficial effect in cardiac surgery patients. Much evidence from randomised controlled trials suggests that albumin improves outcomes in conditions related to chronic liver disease. Similarly, and more relevant to critically ill patients, data from the ALBIOS and SAFE studies suggest beneficial effects on survival in patients with septic shock and/or severe sepsis. As a consequence, several randomised controlled studies are currently underway or being planned to further test the effect of hyperoncotic albumin solutions in septic shock and cardiac surgery patients. The role of albumin therapy in critical illness is yet to be fully explored.

Competing interests

None declared.