Question 25 from the first paper of 2010 and the near-identical Question 5 from the second paper of 2004 both asked the candidate to "critically evaluate the use of albumin" in critically ill patients.  Judging by the college answers, these questions were not after an indepth dissction of outcomes literature.  Rather, they may be better worded as "how many legitimate uses of albumin can you think of?"

As noted in the discussion of Question 25, the best single reference for this sort of question is a review article by Luciano Gattinoni and Pietro Caironi, "The clinical use of albumin: the point of view of a specialist in intensive care" (2009). Unless otherwise referenced, this article is the main source of the information summarised below. Another good resource is the 2014 article by J.L Vincent for Critical Care. Additional local sources of information include the following articles:

• Albumin (20% concentrated human albumin)
• Response to 100ml of concentrated (20%) human albumin
• Metabolic fate of transfused albumin
• Colloids vs. crystalloids as resuscitation fluids
• Fluid resuscitation for septic shock

The routine maintenance of normal albumin levels in critical illness

As with any "critically evaluate" style question, a discussion of albumin should begin with the exploration of those properties which render it desirable as a pharmacological agent. What makes it so attractive?

Rationale for the routine replacement of albumin

• Oncotic effects: albumin plays a critical role in determining plasma oncotic pressure (it contributes 80% of it),  which makes it an attractive agent as a means of manipulating extracellular fluid movement between the intravascular and extravascular compartments.
• Scangenger effects: albumin is able to neutralise toxic compounds such as oxygen radicals and nitrite peroxides.
• Anti-NOS effect: albumin may be able to neutralise the vasodilating effect of nitric oxide.
• Extravascular buffer effects: Outside of the circulating blood, wherever there is no haemoglobin, albumin plays a major role as a buffer (in fact it is the only protein with buffer action in the extravascular space).
• Association with mortality: low albumin levels have been associated with mortality in the critically ill population, in a variety of clinical settings. JL Vincent (2003) found that every 10g/L decrease in serum albumin resulted in a 137% increase in the risk of death, 89% increase in morbidity and a 71% increase in the length of hospital stay.
• Dubois et al (2006) actively maintained albumin levels at over 31g/L in an RCT among 100 critically ill patients, and found that fluid balances were more neutral and SOFA scores were lower (mainly respiratory, cardiovascular and CNS scores).

Arguments against the routine replacement of albumin

• It is a natural  product, and therefore may cause allergic reactions.
• Its healthful effects are clearly non-essential, as is demonstrated by the excellent health of people in whom it is totally congenitally absent.
• The normal response to critical illness is to reduce the synthesis of albumin (i.e. the liver decreases its normal production of 10-12g per day, and this is an expected and beneficial protein conserving response). Ergo, the hypoalbuminaemia of critical illness is not to be viewed as a pathological state. Rather, albumin is a "negative acute phase reactant", and its association with mortality can be viewed in the same way as the association of CRP with mortality (a more severe illness results in a higher CRP and a lower albumin). Association is not causation.
• The normal effect of albumin on oncotic pressure will be completely frustrated by the leakiness of the capillary endothelium in critical illness, which destroys the particle size barrier and therefore allows albumin concentrations to equlibrate between the intravascular and extravascular space
• The immobility of the critically ill patient impairs the normal movement of lymphatic fluid, and therefore the deproteination of the extravascular fluid is delayed. In contrast, circulation will (hopefully) continue, and degradation of circulating albumin by the reticuloendothelial system will not be impaired. Albumin given to critically ill patients will therefore end up being rapidly cleared from the circulating blood, and very slowly cleared from the oedema fluid. Albumin may therefore contribute to the persistence of tissue oedema in intensive care patients.
• That Dubois study (2006) was small, conducted in a single centre, and did not find any statistically significant difference in mortality (nor was it powered to do so). The 100 patients were very heterogeneous. The study was not blinded to anybody, and the results have not been replicated in any of the subsequent larger trials.

Albumin as a resuscitation fluid

Albumin as a colloid resuscitation fluid for non-specific shock

This enjoys a ...robust discussion... in the chapter which deals directly with the colloid vs crystalloid debate. In short, if one looks for mortality benefits, there will be none. Highly criticised post-hoc analysis of subgroups has revealed advantages (and disadvantages) in specific populations, but meta-analysis has not revealed any sort of miraculous resuscitative properties to albumin. Even when compared to the (hideously toxic) normal saline, 4% albumin was found to be equivalent in terms of mortality (SAFE study). Much has been made of the findings of the SAFE study. The most recent ALBIOS study has supported the notion that albumin and saline are quivalent as resuscitation fluids. Furthermore, the authors found that the patients enrolled in early stages of sepsis did not demonstrate an early benefit, and that patients treated with albumin for longer tended to benefit more. This suggests that the benefit of albumin is derived not from a purely oncotic effect, but rather due to its ancillary functions as a nitric oxide modulator, antioxidant and anti-immunosuppressive. This is supported by the last salvo fired by Marik, who suggested that the contribution of albumin infusion to maintaining the integrity of the vascular endothelial glycocalyx is enough to support its role as "a reasonable intervention" in sepsis.

In brief, decades of debate can be glibly summarised as follows:

• Albumin is equvalent to saline in terms of mortality effects
• 4% albumin does not decrease fluid requirements very much
  (giving 1000ml of 4% albumin is haemodynamically equvalent to giving 1400ml saline)
• 20% albumin may acheieve haemodynamic goals sooner, but for some reason this does not influence the mortality data (ALBIOS trial)
• The haemodynamic benefit of albumin is statistically significant - but not clinically significant - because the inter-group difference in the ALBIOS trial was 1mmHg of MAP.

Albumin for resuscitation of septic shock

• Slightly superior to saline in terms of mortality (on post-hoc subgroup analysis of the SAFE study)
• Equivalent to saline in terms of mortality (ALBIOS trial)
• Improves mortality of septic shock patients once hemodynamic stability has been achieved (also the ALBIOS trial).
• According to a recent meta-analysis, the results of the available studies support safety, but suggest that albumin is "not robustly effective at reducing all-cause mortality".

Albumin should probably be avoided in traumatic brain injury

• Again from the SAFE study, in the same way as a post-hoc subgroup analysis revealed some benefit from albumin in sepsis, so did a similar subgroup analysis reveal some evidence of harm in patients with traumatic brain injury.
• Specifically, there were 406 patients with brain injury in the n=7000 study, and this group underwent post-hoc analysis (of them, 70% were "severe TBI" with a GCS 3-8)
• After two years, the mortality of those treated with 4% albumin appeared higher than the mortality of those treated with saline. The difference was actually quite significant: 33.2 vs 20.4%. When only the "severe TBI" group were considered, the mortality difference was 41.8% vs 22.2%. Though the mortality statistics were complied at 24-month follow up, the majority of the deaths in this subgroup occurred in the first week.
• This mortality difference had been explored with the SAFE data reexamined by Cooper et al (2013). It seems the increased mortality was due to an increase in intracranial pressure, and the increased use of toxic drugs to manage it (sedatives, vasopressors, cooling, etc). Increased cerebral oedema is thought to be the mechanism at fault for this.

Use of albumin in advanced liver disease

The recent AISF statement paper (2016)  is an excellent resource for this material: it is in effect a literature review of all the available evidence for the use of albumin in liver disease patients.

Albumin for spontaneous bacterial peritonitis

• Albumin - if administered early in SBP - reduces the incidence of renal impairment and death (Sort et al, 2011).
• A more recent meta-analysis by Salerno et al (2013) included 288 patients from 4 trials, and confirmed a significant benefit. Risk of renal failure in SBP was 30.6% in the non-albumin group, compared to 8.3% in the albumin group. Mortality was more than halved, from 35.4% to 16.0%.
• The recent AISF statement paper (2016)  recommends a dose of 1.5g/kg at diagnosis, and 1g/kg at day 3.
• Interestingly, albumin has little effect on renal failure and mortality if the infection is not SBP; the cirrhosis patients in Thierry Thevenot's 2015 trial did not benefit very much. The onset of renal failure was delayed, but it developed anyway, and the 3-month mortality was unchanged.

Albumin for volume replacement in paracentesis

• Albumin prevents paracentesis-associated circulatory dysfunction (Bernardi et al, 2012). The recent AISF statement paper (2016) strongly recommends in favour of its use in this setting.
• It seems that replacing 8-10g of albumin for every 1L of drained ascites is the best ratio for balancing cost with effectiveness (Johnson et al, 2015), although the Italians report favourable results with as little as 4g/L.
• Post-paracentesis circulatory dysfunction occurs in about 70% of patients without albumin. The use of albumin reduces this risk by 66%.
• Will any colloid do?  In a 2013 review by Kwok et al, there was no specific evidence to help decide whether albumin had any sort of advantage over other colloid volume expanders in this setting.

Albumin as an adjunct in Type 1 hepatorenal syndrome

• When combined with terlipressin, albumin infusion is an effective strategy to prolong survivali n hepatorenal syndrome while waiting for liver transplantation.
• Albumin alone is not as effective.
• The recent AISF statement paper (2016) recommends the combined use of terlipressin and albumin, quoting recovery rates of 40-50%.
• The recommended dose is 1g/kg on day 1, and then 20-40g/day until terlipressin is ceased.

Albumin to treat the hyponatremia of liver disease

• By giving albumin, one expects to improve the "apparent" hypovolemia and thereby decrease the hypersecretion of vasopressin. This should drive the sodium down.
• The results of a small RCT published only as abstract (Jalan et al, 2007) suggest that this works.
• In absence of good quality trials, the AISF could not recommend this practice with a straight face.

Albumin for extracorporeal detoxification in liver failure

• 20% albumin is used as a molecular adsorbent in the Molecular Adsorpent Recirculation System (MARS) which can be used as a bridge to transplant.
• Small trials (eg. Sponholz et al, 2015) have demonstrated efficacy of this system in influencing "paraclinical parameters" such as bile acids, endotoxin, bilirubin, creatinine and urea. Patient-centered outcomes have been mixed.
• A recent meta-analysis (Shen et al, 2016) found some evidence of a short term mortality benefit (30%) with artificial liver support systems in a broader sense.  It is hard to tell what role albumin plays in this effect.

Use of albumin to manipulate body water

Albumin as an adjunct to frusemide in ARDS

• Albumin and frusemide together improve oxygenation in hypoproteinaemic ARDS patients
• There is no mortality benefit, and robust evidence is lacking.
• Two trials (Martin et al, 2002, and Martin et al, 2005) had some success, but excluded patients with haemodynamic instability, suggesting that it is effective mainly in those whose capillary integroity has been largely restored.
• Gattinoni et al (2014) confess that they continue to recommend albumin for ARDS, mainly because it is "a safer choice" among all the colloids.

Albumin in the resuscitation of burns patients

• Some centres use albumin (4%) preferentially as the maintenance fluid in burns patients. For instance, the Alfred protocol calls for albumin infusion (4%, at 100ml/hr) to keep the serum albumin level around 25g/L for the first 4-5 days after a burn.
• Some formulae used to estimate fluid goals seem to specify albumin. The modified Parkland and Shriner's formula calls for 5% albumin as the specific resuscitation fluid in the acute burns setting.
• The rationale for this is that burns patients lose both protein and fluid, and that normal processes of protein synthesis are going to be ineffective at replacing the lost protein. Ergo, protein replacement (nutritionally and pareneterally) is being recommended.
• . In the spirit of controlling resuscitation volume, one could think that 20% concentrated albumin would be the better choice.
• There is little evidence to support this practice, other than to suggest that it is probably safe (i.e. the SAFE study did not show any increased mortality in association with albumin).
• Melinyshyn et al (2013) were unable to find any difference in outcomes or organ dysfunction scores between retrospectively audited groups of patients whose albumin levels were 20g/L and 25g/L in the first 30 days after the burn.
• Cooper et al (ALBUR, 2006) could not find any improvement in organ system dysfunction scores when 5% albumin was added to Ringer's lactate in the first 14 days of resuscitation.
• Mohammadi et al (2010) in their RCT of 141 burns patients did not find any difference in mortality, ICU stay or healing time

Albumin to aid water elimination in oedematous ICU patients

• This practice is based on physiological principles, using albumin as an oncotic agent to attract water into the intravascular space to improve diuresis.
• Proponents of this strategy admit that their recommendations "appear in open contrast with what is called “evidence-based medicine” (Caironi et al, 2009)
• The FADE trial (Oczkowski et al, 2018) aimed to assess this issue by looking at ventilator-free days and mortality. It was not so much a clinical trial, rather a feasibility study. Unfortunately, "the current study design did not demonstrate feasibility", and in any case clinical outcomes were similar between groups.

Experimental and esoteric applications of albumin

• Protection of the endothelial glycocalyx: albumin comprises up to 70% of the glycocalyx, and is rapidly adsorbed onto its surface. it is thought to have some sort of protective properties, but so far this has only been demonstrated reliably in disembodied organs. Becker et al (2010) suggested that of all the currently suggested therapies targeting the glycocalyx, it might be the easiest to implement ("to maintain mechanical and chemical stability of the endothelial surface layer ").
• Prevention and management of fat embolism and fat overload syndrome: in the circulation free fatty acids are normally bound to albumin, with only 1% in their unbound state (Habashi et al, 2006). Albumin may have some sort of a scavenging role in fat embolism and in fat overload syndrome of parenteral nutrition, particularly where it comes to oleic acid (which has lung-damaging properties - see the chapter on the properties and contents of parenteral nutrition).