Colloids

Key points

• A Cochrane review comparing colloids versus crystalloids showed little or no difference in mortality rates but failed to address whether colloids, particularly human albumin, show clinical benefit is defined patient populations or in specific clinical settings
• The use of HES has been greatly restricted across Europe and North America and Canada in response to evidence that guidelines on the safe use of HES were being ignored putting patients at risk of harm
• HES use was an independent predictor of postoperative delirium in oesphagectomy
• Human albumin has both oncotic and non-oncotic properties
• Interest in human albumin has resurged following publication of studies demonstrating negative outcomes with synthetic colloids. These studies and change in the Surviving Sepsis guidelines may have resulted in increased use of human albumin
• Gelatin solutions may increase risk of anaphylaxis and may increase mortality, renal failure and bleeding
• Cheaper and safer alternatives to gelatin are available until well-designed RTs can prove safety
• Following HES controversy, dextran, which had fallen out of use, may be re-considered as fluid therapy

The rationale for using colloids in fluid therapy is to increase the oncotic pressure in the intravascular space. Colloid molecules are too large to extravasate and remain in the blood for several hours before being distributed to other parts of the body. Due to their water-binding capacity, colloids were thought to be better than crystalloids in maintaining circulating volume in patients with acute volume depletion.

However, in recent years, studies have questioned whether this physiological principle leads to better outcomes for patients. Does the choice between colloid and crystalloid IV fluid have measurable implications in different disease conditions, e.g. trauma or surgery, sepsis, or liver failure? 

A 2018 Cochrane review of 69 studies (n=30,020) attempted to answer this question across a highly diverse range of critically ill patient populations by providing a general comparison between the two categories of IV fluids. The review found that starches, dextrans, albumin, fresh frozen plasma (FFP) (moderate-certainty evidence) and gelatins (low-certainty evidence) probably make little or no difference to mortality rates in comparison to crystalloids across a diverse population. The review suggested that starches slightly increase the need for blood transfusion and renal replacement therapy (RRT) (moderate-certainty evidence), whereas albumin and FFP are thought to make little or no difference to the need for RRT (low-certainty evidence) (Lewis et al., 2018). However, the review failed to investigate whether colloids or crystalloids confer clinical benefit in special situations or in defined patient populations.

Hydroxyethyl Starches

A range of hydroxyethyl starches (HES) have been developed differing in their mean molecular weight (MW), molar substitution and C2/C6 ratio. Hydroxyethyl starches are identified by their concentration, mean MW (in kDa) and degree of molar substitution (MS) e.g. 6% HES 130/0.4. The degree of molar substitution, i.e. 0.4, indicates that there are four hydroxyethyl residues on average per 10 glucose subunits (Westphal et al., 2009).

Pharmacokinetic studies of first- and second-generation HES, known as hetastarches (MS=0.7), hexastarches (MS=0.6) and pentastarches (MS=0.5), showed incomplete clearance within 24 hours, with following repeated infusions resulting in accumulation of HES in the circulation and tissues. Third-generation HES, known as tetrastarches (MS=0.4), were developed to enhance degradation and to minimise bodily retention (Westphal et al., 2009).

Six per cent HES solutions are iso-oncotic, whereas 10% solutions are hyperoncotic (Westphal et al., 2009).

A meta-analysis of 25 studies (16 single-arm cohort studies, 5 non-randomised controlled trials and 3 RCTs) with a total of 287 subjects estimated the tissue uptake of a range of HES compounds. Tissue uptake of low molecular weight HES (≤200 kD) was 42.3% (95% confidence interval (CI), 39.6–45.0) versus 24.6% (95% CI, 17.8–31.4) for high molecular weight HES (p<0.001). Similarly, the tissue uptake of lower substitution HES (≤0.5) was 42.4% (95% CI, 39.5–45.3) compared with 26.6% (95% CI, 19.6–33.6) for higher substitution HES (p<0.001) (Bellmann et al., 2012).

Based on the analysis of three clinical trials of HES in patients with severe sepsis, or in intensive care (Brunkhorst et al., 2008; Myburgh et al., 2012; Perner et al., 2012), the European Medicines Agency’s (EMA) Pharmacovigilance Risk Assessment Committee (PRAC) decided that HES must not be used to treat patients with sepsis or burn injuries, or critically ill patients, due to an increased risk of mortality and kidney injury (EMA, 2013).

Use of HES is contraindicated in:

• sepsis
• burns
• kidney impairment or kidney function replacement therapy
• intracranial or cerebral haemorrhage
• critically ill patients (typically admitted to intensive care)
• hyperhydrated patients, including patients with pulmonary oedema
• dehydrated patients
• severe coagulopathy
• severely impaired liver function.

(Medicines and Healthcare products Regulatory Agency [MHRA], 2014).

European and UK guidelines indicate that HES solutions may continue to be used in patients to treat hypovolaemia caused by acute blood loss, where treatment with crystalloids alone are not considered to be sufficient (EMA, 2013; MHRA, 2014).

In addition to existing restrictions, a recent study has found HES to be associated with postoperative delirium. Patients undergoing oesophagectomy are often at a high risk of developing postoperative delirium due to their advanced age, existing comorbidities and the need for Intensive Care Unit (ICU) care. However, in a retrospective study of 1,041 patients undergoing oesophagectomy, HES infusion was found to be an independent risk factor for postoperative delirium (OR 1.53; 95% CI, 1.09–2.14; p=0.0151). This important finding may change the type of fluid surgical patients are administered, in an effort to avoid postoperative delirium, as this is often associated with longer hospital stays and increased mortality rates (Jung et al., 2018).

In January 2018, the Pharmacovigilance Risk Assessment Committee (PRAC) recommended the suspension of HES marketing authorisations in response to evidence showing that HES was often administered to critically ill patients and patients with sepsis despite contraindications. The Co-ordination Group for Mutual Recognition and Decentralised Procedures – Human (CMDh) agreed with the assessment of serious risk. However, risk minimisation measures have been shown to have some effect in certain countries. Therefore, the European Commission made the EU-wide legally binding decision on 17 July 2018 that HES would continue to stay on the market, but with new measures put in place to protect patients at risk. The new measures include a controlled access programme whereby only accredited hospitals will be supplied with HES on the basis of healthcare professional training. This is along with warnings on the medicines’ packaging and SmPCs, and written contact with healthcare professionals, ensuring they are aware of the contraindications (EMA, 2018). 

Click here to see a timeline of the events and published research leading up to this decision.

The EMA’s decision to maintain the marketing authorisations of HES has been met with some criticism and is a heavily debated topic; two recent letters to editors by Bilotta et al. and Laake & Møller address the ongoing issues with HES use, sharing further evidence against its use and ask for the EMA’s decision to be reassessed (Bilotta et al., 2018; Laake & Møller, 2018). However, despite the backlash against the decision, there is also emerging evidence supporting the continued use and safety of HES. A recent 1-year follow-up study of patients who had undergone major open abdominal surgery found that those treated with balanced HES solutions had a lower World Health Organization Disability Assessment Schedule 2.0 (WHODAS) score than those treated with balanced crystalloid solutions (2.7 [0–12] vs. 7.6 [1.3–18]; p=0.015). In addition, disability-free survival was reported to be higher in the colloid group compared to the crystalloid group (79% vs. 60%, 95% CI, 2–39; p=0.024). Interestingly, the primary endpoint of this study, to measure long-term renal function by estimated glomerular filtration rates found no statistical difference between the two treatment groups. However, there are some questions over the validity of these results with the authors commenting that the study had limited power to detect differences in renal function (Joosten et al., 2018). 

Evidence surrounding the safety and efficacy of HES solutions has been historically controversial, with the high-profile case of Joachim Boldt, a German anaesthiologist and an advocate for colloids in fluid therapy. As of 2017, 96 of his papers have been retracted for fraud. In 2013, JAMA published a meta-analysis on HES in critically ill patients (Zachyranski et al. 2013). Boldt had seven studies from the 1990s that had not yet been retracted. Including them, there was no increase in mortality, but excluding them, there was a significant increase in mortality. Only the Boldt studies showed an improvement with HES; all other studies showed the opposite. It is likely that his fraudulent studies put critically ill patients at risk and caused harm.
One reason for misleading results is financial bias. Wiedermann (2018) analysed 45 anaesthesiology, critical care, and emergency medicine textbooks. Conflict of interest (COI) statements were absent from 38 of the textbooks; however, journal COI statements linked 22% of the authors to HES manufacturers. HES recommendations were often controversial and approximately a quarter of relevant textbooks were inconsistent with regulatory authority or international guideline criteria (Wiedermann, 2018).

In 2013, the FDA issued a ‘black box’ warning against the use of HES in high-risk patients and recommended that HES should not be used in critically ill patients or in those with pre-existing renal dysfunction (Nolan & Mythen, 2013; FDA, 2013). Since this warning, a public petition to ban completely the use of HES was submitted to the FDA (Regulations Gov, 2017). However, in February 2018, a joint letter from America’s Blood Centers, the American Red Cross, the American Association of Blood Banks (AABB) and the American Society for Apheresis, was sent to the FDA opposing the blanket restriction. The reasons for this opposition come from evidence supporting a good safety profile in the use of:

• HES preparations as sedimenting agents in therapeutic and donor apheresis
• HES for umbilical cord blood, bone marrow and peripheral blood hematopoietic stem cell processing, storage and transplantation (Katz et al., 2018)

A recent review of the adverse effects of HES found overwhelming, high-quality evidence for the harms of HES in critically ill patients, and no clear evidence for a benefit of HES over other fluids in any setting, including in surgical patients (Ünal & Reinhart, 2019). Find out more about this evidence in this publication digest. 

Regardless of the decisions made by governing bodies and international guideline recommendations, it is likely to take some time before practice is changed. Research suggests it can take up to 17 years for evidence to change practice (Balas & Boren, 2000). The findings from one Canadian study suggest that equipment re-design to force function behaviour may be required to implement system change successfully after messaging and education methods have been used. The study investigated the number of monthly HES-containing fluid products ordered in the 6 months prior to warnings issued in 2013, and 42 months after the warnings were issued. An 81% drop in orders occurred  in the 6 months following the warnings. However, during 2014, HES orders were persistently over 40 units per month. In order to reduce HES use, the hospital removed HES products from anaesthesia drug carts and placed them in the operating room pharmacy. After removal from drug carts, order levels were sustained at <1% of the pre-warning levels (Pysky et al., 2018). 

Albumin

Human serum albumin (HSA) is a multifunctional monomeric protein and is the main determinant (about 70%) of plasma oncotic pressure. It is the main modulator of fluid distribution between body fluid compartments (Evans, 2002; Fanali et al., 2012; Gatta et al., 2012).

The figure below shows a schematic drawing of the structure of human serum albumin at 2.5Å resolution (Sugio et al., 1999).

Structure of human serum albumin at 2.5Å resolution. Adapted from Sugio et al., 1999 and NCBI Structure, 2011.

HSA has a high ligand-binding capacity, providing a depot and carrier for many endogenous compounds, including fatty acids, cholesterol, nitric oxide (NO), amino acids, ions, thyroxine and bilirubin. HSA also interacts with exogenous compounds such as nonsteroidal anti-inflammatory drugs (NSAIDs), phenytoin, digoxin and antibiotics and thereby affects the pharmacokinetics of these and other drugs (Evans, 2002; Fanali et al., 2012; Gatta et al., 2012).

HSA is an antioxidant, providing most of the antioxidant capacity of plasma. It has a role in the scavenging of free oxygen radicals, which are involved in the pathogenesis of inflammatory diseases. This can significantly reduce reoxygenation injury, which is particularly important in sepsis (Ferrer et al., 2018). Albumin has antithrombotic and anticoagulant effects. It also inhibits platelet aggregation, which may be mediated by its binding of the anti-aggregatory nitric oxide (NO), following the formation of S-nitrosylated HSA, which stabilises NO activity (Evans, 2002; Fanali et al., 2012; Gatta et al., 2012). HSA may also act as a detoxification agent (Fanali et al., 2012; Gatta et al., 2012).

Albumin may also play a role in maintaining microvascular integrity and in modulating the inflammatory response, including neutrophil adhesion and the activity of cell signalling factors (Quinlan et al., 2005).

A summary of the oncotic and non-oncotic properties of human albumin is shown in the figure below (Caraceni et al., 2016).

Summary of the oncotic and non-oncotic functions of human albumin.
Key: Cys-34, cysteine-34; LPS, lipopolysaccharides; NO, nitric oxide; PG, prostaglandins; TNFα, tumour necrosis factor-alpha.

The 2018 systematic Cochrane review (69 studies, n=30,020) compared four types of colloid (i.e. starches, dextrans, gelatins, and albumin or FFP) versus crystalloids. The review found probably little or no difference in mortality rates between albumin or FFP versus crystalloids at the end of follow-up (RR 0.98; 95% CI, 0.92–1.06) or within 90 days (RR 0.98; 95% CI, 0.92–1.04) or 30 days (RR 0.99; 95% CI, 0.93–1.06) (moderate certainty). Albumin or FFP also appeared to make little or no difference to the need for renal replacement therapy (RRT) (RR 1.11; 95% CI, 0.96–1.27) or to the likelihood of allergic reactions (RR 0.75; 95% CI, 0.17–3.33) compared to crystalloids (Lewis et al., 2018). As might be expected, this review also confirms that, in critically ill patients, the prognosis is determined primarily by the underlying disease. This does not preclude the possibility that in special populations the choice of fluid regimen has an influence on clinical outcomes.

A retrospective study compared the use of balanced crystalloids alone versus a mixed and balanced crystalloid solution in patients receiving venoarterial extracorporeal membrane oxygenation (VA-ECMO). Patients in a single ICU between 2010 and 2017 received crystalloids alone (n=171) or crystalloids mixed with albumin (1:2 volume ratio, or 10 g/L of albumin) (n=112). Patients treated with albumin alone versus the mixed solution showed significantly improved hospital survival (38.4% vs. 25.7%, respectively; p=0.026). This significance remained after propensity score-matching (43.9% vs. 27.6%; p=0.025). A matched multivariate regression analysis showed that albumin was an independent factor for improved hospital survival (OR 3.1, 1.15–6.38). The mixed crystalloid and albumin solution was also more favourable when comparing the cumulative incidence of hospital mortality (p=0.044) (Wengenmayer et al., 2018).

Gelatin

Semisynthetic gelatin colloids include succinylated gelatin and urea-linked gelatin-polygeline formulations (Myburgh & Mythen, 2013).

Patients with severe sepsis receiving gelatin (n=87; 68%) had an increased risk of acute kidney injury (adjusted p=0.025) when compared with patients receiving crystalloids (n=141; 47%) (Bayer et al., 2011). European Society of Intensive Care Medicine (ESICM) guidelines recommend that gelatins should not be administered to organ donors (Reinhart et al., 2012). 

A systematic review and meta-analysis of the safety of gelatin for volume resuscitation identified 40 randomised controlled trials (RCTs) published between 1976 and 2010. The risk of mortality, assessed in 15 trials involving 1,766 patients (7 involving critically ill patients; 8 in trauma, emergency or elective surgery), was not statistically significantly different between gelatin and control fluids (albumin or crystalloids) (RR=1.12; 95% confidence interval [CI], 0.87–1.44). There was no significant difference in the risk of exposure to allogeneic blood products (8 RCTs; 712 patients) between gelatin and control fluids (RR=1.28; 95% CI, 0.89–1.83). Only three RCTs (n=172) reported acute kidney injury and the need for renal replacement therapy which did not differ significantly between gelatin and control groups (RR= 1.35; 95% CI, 0.58–3.14) (Thomas-Rueddel et al., 2012).

However, despite over 60 years of clinical practice, the authors concluded that the safety and efficacy of gelatin cannot be reliably assessed in at least some settings and that further research was needed to establish safety.

A systematic review which compared modified gelatin with a crystalloid solution in critically ill patients (11 trials, 506 patients) gave a pooled risk ratio (RR) for mortality of 0.91 (95% CI, 0.49–1.72) (Perel et al., 2013). 

A recent meta-analysis which compared patients receiving gelatin fluids with those receiving albumin or crystalloid solutions for hypovolaemia (16 RCTs, n=2525), reported RRs for mortality = 1.15 (95% CI, 0.96–1.38); for requiring allogeneic blood transfusion = 1.10 (95% CI, 0.86–1.41); for acute kidney injury = 1.35 (95% CI, 0.58–3.14); and for anaphylaxis = 3.01 (95% CI, 1.27–7.14) (Moeller et al., 2016). Gelatin solutions increase the risk of anaphylaxis and may be harmful by increasing mortality, renal failure, and bleeding possibly due to extravascular uptake and coagulation impairment. Until well-designed randomized controlled trials show that gelatin is safe, the authors caution against the use of gelatins because cheaper and safer fluid alternatives are available.

In a group of hypovolaemic critically ill patients (n = 115), fluid resuscitation by only 1500–1700 mL of normal saline, gelatine, HES or albumin, resulted in a small decrease in pH, regardless of the fluid used. The post-hoc analysis suggested that a progressive metabolic acidosis should not therefore be erroneously attributed to insufficient fluid resuscitation (Spoelstra-de Man et al., 2017). 

Dextran

The use of dextran solutions has largely been superseded by the use of other semi-synthetic colloid solutions (Schumacher & Klotz, 2009; Myburgh & Mythen, 2013). Dextran solutions have adverse effects on blood coagulation, cause increased risk of bleeding, and also have a high risk of provoking allergic reactions (Schumacher & Klotz, 2009). However, after the recent controversies and restrictions of HES solutions, and following the published results from the 6S-trial, some researchers have questioned if dextran could be reintroduced despite remaining concerns over the adverse effects of this solution. A study of 18 patients undergoing major gynaecological surgery assigned patients to receive either 5% albumin or 6% dextran-70 and monitored coagulation defects at different time points perioperatively. The study found that on standard plasma-based coagulation tests, the platelet count and whole blood viscoelastic clot structure was affected to a greater extent in the dextran-70 group compared to albumin. However, this difference was not seen in platelet aggregation. Further studies are needed with more advanced haemodynamic monitoring of dextran’s effects on haemostasis if dextran is to be re-considered in the future (Sigurjonsson et al., 2018).

Dextran solutions used clinically are hyperoncotic (Annane et al., 2013).

A systematic review which compared dextran with a crystalloid solution in critically ill patients (n=834; 9 trials) gave a pooled RR for mortality of 1.24 (95% CI, 0.94–1.65) (Perel et al., 2013). In contrast, a 2018 systematic review found when comparing dextran to alternative crystalloid therapies there was probably little or no difference in mortality at end of follow-up (RR 0.99, 95% CI 0.88–1.11) or within 90 days or 30 days (RR 0.99, 95% CI 0.87–1.12) (moderate certainty) (Lewis et al., 2018). Further meta-analyses have also suggested that the addition of dextran to hypertonic saline in patients with trauma or haemorrhagic shock does not confer a survival advantage (de Crescenzo et al., 2017; Wu et al., 2017).

Crystalloids

Key points

• There is the assumption that balanced crystalloids are preferable for critically ill patients, but fluid therapy should be carefully tailored to the patient considering the type amount and duration of resuscitation required
• The volume of crystalloids administered is a key consideration; high volumes appear to increase mortality and ventilator days in one study
• Saline is the most commonly used crystalloid, but there are concerns about non-desirable metabolic changes such as hyperchloraemic metabolic acidosis. Further research is needed

Crystalloids and colloids are both volume expanders that increase the circulating volume. Colloids have larger molecules, which are intended to remain in the intravascular space for longer, while crystalloids are aqueous solutions of salts and other water-soluble molecules. Saline (0.9% NaCl solution) is the most commonly used crystalloid.

Treatment with fluids is the cornerstone of management for numerous medical conditions. In this section we discuss the evidence for crystalloids.

Crystalloids are often first-line therapy for critically ill patients. This means a vast amount of literature comparing saline to compound/balanced crystalloid solutions in this population. The recent SALT-ED (Saline against Lactated Ringer’s or Plasma-Lyte in the Emergency Department) trial adds to the limited data comparing the two types of solution in non-critically ill patients. Interestingly, the number of hospital-free days didn’t differ significantly between the two treatment groups (median, 25 days in each group; adjusted odds ratio with balanced crystalloids, 0.98; 95% CI 0.92–1.04; p=0.41). However, those that received balanced crystalloids had a lower incidence of major adverse kidney events within 30 days than those who received saline solution (4.7% vs. 5.6%; adjusted odds ratio, 0.82; 95% CI, 0.70–0.95; p=0.01) (Self et al., 2018).

The volume at which crystalloids are administered is another key consideration in fluid therapy. A recent analysis of 970 trauma patients found 247 (27%) received ≥5 L of crystalloids in the first 24 hours following injury. The effects of large-volume crystalloid resuscitation were associated with increased mortality (adjusted odds ratio 2.55) and a significant increase in ventilator days (RR 2.31, 95% CI 1.81–2.96; p<0.0001). However, in-hospital complications including pneumonia and sepsis were not found to be associated with ≥5 L of crystalloid administration (Jones et al., 2018).

Saline

Saline is the most commonly used crystalloid solution globally, with use in the United States especially high. Normal saline (0.9% NaCl) is roughly isotonic with extracellular fluid (Myburgh & Mythen, 2013). A recent summary of randomised control trials found that normal saline compared to 6% HES may reduce the onset of acute kidney injury (AKI), however, no significant difference in mortality and AKI incidence was seen when comparing normal saline to 10% HES, albumin or buffered crystalloid solution. As such, it is important that intravenous fluids are prescribed on an individual basis taking into consideration the patient’s condition (Zhou et al., 2018).

Despite the popularity of normal saline, there is concern about non-desirable metabolic changes, such as hyperchloraemic metabolic acidosis. A recent literature review of the safety of normal saline revealed many studies with data highlighting these concerns, including evidence suggesting normal saline causes substantially more in vitro haemolysis in comparison to Plasma-Lyte A, a balanced solution, or similar solution,s during short-term storage (24 hours) after red blood cell washing or intraoperative salvage. Further studies are needed on the safety of saline solutions (Blumberg et al., 2018).

The use of saline was also questioned in a retrospective study of 84 patients with diabetic ketoacidosis (DKA). The study compared fluid therapy with Plasma-Lyte A to normal saline (sodium chloride 0.9%) to see if either had a beneficial effect on time to resolution of DKA or change in pH and chloride levels. The primary outcome of mean time to resolution of DKA was similar between the two groups (Plasma-Lyte A 17.74 hours vs. normal saline 18.05 hours; p=0.5080). However, patients in the Plasma-Lyte A group had a significantly greater rise in pH within the 4–6 hour and 6–12 hour periods compared to normal saline (mean difference -0.20 [95% CI -0.32 to -0.08] and -0.17 [95% CI -0.32 to -0.02] respectively). Those in the normal saline group also had a significantly higher change in chloride level over the 6–12-hour period compared to the Plasma-Lyte A group (11.40 mEq/L vs. 7.25 mEq/L, respectively). A further difference was observed when comparing the anion gap (Plasma-Lyte A 17.00 mEq/L vs. normal saline 13.58 mEq/L). These differences may suggest that Plasma-Lyte A has advantages over normal saline for fluid resuscitation in patients with DKA. However, larger randomised controlled studies are needed to further assess the differences in effectiveness between the two fluids (Oliver et al., 2018).

Compound/balanced solutions

Compound solutions comprise sodium lactate derivatives which are based on original Hartmann’s and Ringer’s solutions; with PlasmaLyte being an example of a balanced salt solution (Table 1). Neither the compounded lactate salt nor balanced salt solutions are physiological; they are hypotonic compared to extracellular fluid mainly due to their lower sodium concentration (Myburgh & Mythen, 2013).

Table 1. Composition of crystalloid solutions and human plasma.

A meta-analysis of critically ill patients (14 trials; n=956), which compared hypertonic crystalloids with near isotonic crystalloids, reported no significant differences in mortality. Pooled relative risks (RR) for death in patients given hypertonic crystalloid were 0.84 for trauma patients (95% confidence interval (CI), 0.69–1.04); 1.49 for patients with burns (95% CI, 0.56–3.95); and 0.51 for patients undergoing surgery (95% CI, 0.09–2.73) (Bunn et al., 2004).

In a sequential period study of critically ill adults of chloride-liberal (n=760) versus chloride-restrictive (n=773) IV administration, chloride-restrictive fluid was associated with a significant decrease in the incidence of acute kidney injury (AKI) and the use of renal replacement therapy (RRT). In the chloride-liberal group (mean: 694 mmoL of chloride), creatinine values increased by a mean of 22.6 μmol/L, while in the chloride-restrictive group (mean: 496 mmoL chloride), creatinine values rose by a lower amount of 14.8 μmol/L. There was a corresponding reduction in AKI and renal failure from 14% to 8.4% (p<0.001) and a reduction in the use of RRT from 10% to 6.3%. (p=0.005) (Yunos et al., 2012). It was suggested that the improved renal outcomes were attributable to the use of hyperoncotic albumin (Wiedermann & Joannidis, 2013). However, Yunos et al. (2012) considered that the effect of hyperoncotic albumin on renal function was likely to be negligible (Yunos et al., 2013).

It was suggested that the improved renal outcomes observed by Yunos and colleagues (Yunos et al., 2012) were attributable to the use of hyperoncotic albumin (Wiedermann and Joannidis, 2013). In reply, Yunos and colleagues consider that the effect of hyperoncotic albumin on renal function is likely to be negligible (Yunos et al., 2013).

A large, retrospective, cohort study showed that in patients receiving abdominal surgery, use of a calcium-free balanced crystalloid (PlasmaLyte, n=926) was associated with less postoperative morbidity than 0.9% saline (n=30,994). In-hospital mortality rates were 2.9% in the balanced crystalloid group compared with 5.6% in the saline group (p<0.001). Fewer major complications occurred in the balanced crystalloid group: rates for ≥1 complication were 23% vs. 33.7%, respectively (p<0.001). In a propensity-scoring model (3:1 matched sample), treatment with balanced crystalloid fluid was associated with fewer complications (odds ratio (OR) = 0.79; 95% CI, 0.66–0.97) (Shaw et al., 2012).

An observational study of 11,182 patients at nine tertiary and community hospitals across the US found that patients receiving crystalloid therapy within 30 minutes had the lowest mortality rates at 17.8% compared with 18.7% for those who received fluids at 31–120 minutes, and 24.5% mortality for those who received fluids after more than 120 minutes. These rates were not impacted by the severity of disease or comorbidities present (Leisman et al., 2017).

Recent data published from the Isotonic Solutions and Major Adverse Renal Events Trial (SMART) trial has added to the ongoing debate surrounding balanced solutions versus saline solutions in critically ill patients. In this pragmatic, unblinded, cluster randomised, multiple-crossover trial, patients receiving balanced crystalloids had a lower incidence of major adverse kidney events than patients receiving saline solution (14.3% vs. 15.4%: marginal OR, 0.91; 95% C,  0.84–0.99; conditional OR, 0.90; 95% CI, 0.82– 0.99; p=0.04). Those in the balanced crystalloid group also had a lower rate of in-hospital mortality than those in the saline group. However, the difference was not statistically significant (10.3% vs. 11.1%; p=0.06) (Semler et al., 2018). Although the results suggest that balanced crystalloids are the preferred fluid choice, prespecified analyses question the validity of these results. For example, much of the response was observed in patients with sepsis, who had relevant reductions in mortality rates and major kidney events, and are more prone to large-volume resuscitations. Only a small fraction of the positive differential events between the two study arms (12 out of an absolute difference of 72; total of 1,211 events with saline vs. 1,139 events with balanced crystalloids) was observed in the 84.5% of patients with a diagnosis other than sepsis. The benefits of balanced crystalloids outside of the sepsis patient population are therefore less clear and these results cannot be readily generalised to all critically ill patients (Reverter et al., 2018).

Another systematic review and network meta-analysis of 49 randomised controlled trials comparing different resuscitation fluid therapies used in critically ill patients reflected similar findings to the SMART trial as there didn’t appear to be a significant difference amongst the resuscitation fluids for mortality rates. However, balanced crystalloids were found to be most effective (80.79%), with Plasma-Lyte being the most effective of its kind (77.52%). Although this study makes the presumption that balanced crystalloids are therefore the best choice for patients who are critically ill, fluid therapy should be carefully tailored to the patient, considering the type, amount, and duration of fluid resuscitation (Liu et al., 2018).

Hypoalbuminaemia

Analysis of 112 patients with severe sepsis and septic shock associated with community-acquired bloodstream infection identified hypoalbuminaemia as the most important prognostic factor. Logistic regression analysis identified albumin as an independent risk factor for mortality (OR=0.34; 95% CI, 0.15–0.76; p=0.009 per 10 g/L increment in plasma albumin concentration) (Artero et al., 2010).

Morisaki et al. also reported low levels of albumin to be a risk factor for mortality in patients who had undergone major amputation surgery. The retrospective study found a low serum albumin concentration was a risk factor for 30-day mortality in univariate and multivariate analysis (HR, 3.87; 95% CI 1.12–16.3; p=0.03) (Morisaki et al., 2018).

The risk of low albumin preoperatively has also been shown to be independently associated with poorer postoperative outcomes, including sepsis, longer hospital stay and urinary tract infections. A retrospective analysis of 1,275 patients undergoing elective anterior lumbar interbody fusion (ALIF) surgery found 53 patients (4.15%) had preoperative hypoalbuminaemia (serum albumin <3.5 g/dL). In a univariate analysis of serious adverse postoperative complications, preoperative albumin levels were significantly associated with length of stay ≥5 days (p<.001), wound complications (p=.017), pulmonary complications (p=0.019), urinary tract infections (p<0.001), intra- and postoperative red blood cell transfusion (p=0.007), sepsis (p=0.029), and unplanned readmission (p=0.039). In a multivariate regression analysis, low preoperative albumin levels were shown to be a strong independent predictor of length of stay ≥5 days (OR = 2.56, 95% CI 1.43–4.59; p=0.002), urinary tract infection (OR = 5.93, 95% CI 2.11–16.68; p=0.001), and sepsis (OR = 5.35, 95% CI 1.13–25.42; p=0.035) (Ukogu et al., 2018).

A systematic analysis undertaken to examine the relationship between the incidence of acute kidney injury (AKI) in relation to serum albumin concentrations identified 17 studies involving a total of 3,917 patients. A lower serum albumin concentration was a significant independent predictor of AKI development, with a 10 g/L reduction in serum albumin concentration being associated with an increase in the odds of AKI by 134% (pooled OR=2.34; 95%CI, 1.74–3.14). A lower serum albumin concentration was also a significant independent risk factor for mortality among patients (n=1,172; 6 studies) who had developed AKI: for each reduction in 10 g/L serum albumin concentration the odds of death increased by 147% (pooled OR=2.47; 95% CI, 1.51–4.05) (Wiedermann et al., 2010).

A retrospective study of 582 patients with non-severe sepsis (sepsis without organ failure or shock) admitted to an emergency department, identified an initial serum albumin concentration <3.5 g/dL as an independent predictor of the development of severe sepsis or shock within 96 hours of presentation (OR=4.82; 95% CI, 2.40–9.69) (Holder et al., 2016).

Albumin

Since hypoalbuminaemia can impact negatively on sepsis treatment success, many studies have compared albumin with crystalloids to standard treatments. Three large randomised controlled trials (RCTs) compared albumin with crystalloids in patients with severe sepsis or septic shock.

An early key study compared intravascular (iv) fluid resuscitation using 4% albumin or normal saline in intensive care unit (ICU) patients. There was no significant difference in the primary outcome of the study – death from any cause at 28 days – with mortality rates of 20.9% (726/3473) in the albumin group and 21.1% (729/3460) in the saline group obtained (relative risk (RR)=0.99; 95% CI, 0.91–1.09; p=0.87). There was no significant difference (p=0.85) in the number of patients in either group who had no new organ failure or new organ failure in 1–5 organs (SAFE Study Investigators, 2004).

In patients with severe sepsis 28-day all-cause mortality rates were 30.7% (185/603) and 35.3% (217/615) respectively (RR=0.87; 95% CI, 0.74–1.02; p=0.09) (SAFE Study Investigators, 2004).

A later study of patients with severe sepsis showed there was no significant difference in 28-day all-cause mortality rates following fluid resuscitation with either albumin or saline: rates were 30.7% (185/603) and 35.3% (217/615) respectively (odds ratio (OR)=0.87; 95% CI, 0.74–1.02; p=0.09).  There were no differences between albumin and saline with regard to impairment of renal or other organ function. The number of patients treated with renal replacement therapy was similar in both groups: 18.7% (113/603) in the albumin group, and 18.2% (112/615) in the saline group (OR=1.04; 95% CI, 0.78–1.38, p=0.98). There was no difference in either the renal or total Sequential Organ Failure Assessment (Shankar–Hari) scores in the two groups (SAFE Study Investigators, 2011).

Unadjusted RRs of death for albumin versus saline were 0.87 (95% CI, 0.74–1.02) for patients with severe sepsis; and 1.05 (0.94–1.17) for patients without severe sepsis. Multivariate logistic regression analysis which adjusted for patients’ (n=919) baseline factors showed that albumin was independently associated with a decreased odds ratio for death at 28 days (OR=0.71; 95% CI, 0.52–0.97; p=0.03) (SAFE Study Investigators, 2011).

The third RCT compared 20% albumin and crystalloid solution with crystalloid solution for fluid replacement in 1,818 adult patients with severe sepsis. No significant difference between the two groups was found for the primary outcome – death from any cause at 28 days – mortality rates were 31.8% (285/895) in the albumin group; and 32.0% (288/900) in the crystalloid group (RR=1.00; 95% CI, 0.87–1.14; p=0.94). Similarly, there was no significant difference in death rates at 90 days: albumin, 41.1% (365/888); versus crystalloid, 43.6% (389/893) (RR=0.94; 95% CI, 0.85–1.05; p=0.29). Thus, although fluid replacement with albumin was efficacious it did not provide a survival advantage over crystalloids alone. The figure below shows Kaplan–Meier graphs for the probability of survival in patients receiving either fluid, showing no significant difference in survival (p=0.39) (Caironi et al., 2014).

Kaplan–Meier graphs for the probability of survival in patients receiving albumin plus crystalloids and crystalloids alone (Caironi et al., 2014).

Albumin fluid replacement compared with crystalloid was found to be safe, showing no significant difference in new organ failure(s) (p=0.99) or in median Sequential Organ Failure Assessment (SOFA) score (p=0.23) (Caironi et al., 2014).

A small prospective study also compared human serum albumin (HSA) to saline in an effort to understand if HSA has protective effects on endothelial cells in patients with septic shock, as hypothesised previously (Taverna et al., 2013). The study measured skin endothelial function in 30 patients before and 1 hour after volume expansion (with either 500 mL bolus saline [n=15] or a 100 mL bolus HSA 20% [n=15]) in the forearm area. Endothelial dysfunction is believed to be critical in the pathophysiology of sepsis and septic shock and has even been suggested as a predicter of mortality in patients with sepsis (Boisramé-Helms et al., 2013).

Prior to fluid infusion, endothelial response was not significantly different between the two groups (area under the curve [AUC] 3,295 [1,148–5,938] vs. 3082 [879-4,902], p=0.70). However, after infusion, endothelial reactivity improved twofold in the HSA group (AUC 3,082 [879–4,902] prior to infusion vs. 5857 [2,888–16,679] 1 hour after infusion, p=0.04). Saline infusion didn’t have a significant effect on endothelial function, (AUC 3,295 [1,148–5,938] prior to infusion vs. 2,388 [1,914–10,455] 1 hour after infusion; p=NS). Cardiac output and skin blood flow results weren’t significantly different between the two groups, this may suggest that the beneficial effects of albumin are independent of its oncotic properties. Further larger randomised trials are needed to validate these results, but this data presents novel insights into the benefits of albumin infusion in patients with septic shock (Hariri et al., 2018).

Network meta-analyses using Bayesian statistics (14 studies [not including Caironi et al., 2014]; n=18,916) compared different resuscitative fluids on mortality in patients with sepsis. Selected comparative results are shown in the tables below. The quality of evidence for these multiple comparisons was variable. These results suggest that the use of balanced crystalloids or albumin is associated with reduced mortality in patients with sepsis (Rochwerg et al., 2014).

Table 2. Comparative effects of resuscitation fluids on mortality in patients with sepsis: 4-node analyses.

Table 3. Comparative effects of resuscitation fluids on mortality in patients with sepsis: 6-node analyses.

A meta-analysis, which selected only 3 recent trials (the EARSS [Charpentier & Mira, 2011]; SAFE [SAFE Study Investigators, 2011]; and ALBIOS [Caironi et al., 2014] trials), reported that albumin was associated with a reduction in mortality in adults with severe sepsis (pooled RR=0.92; 95% CI, 0.84–1.00; p=0.046) (Wiedermann & Joannidis, 2014).

The findings from the ALBIOS trial have recently been reassessed due to a change in the definition of septic shock as sepsis has various clinical features and definitions. To provide a more homogeneous approach at identifying subsets of patients, a consensus conference in 1991 defined septic shock as persistent sepsis-induced hypotension despite adequate fluid resuscitation which could be further categorised with a Sequential Organ Failure Assessment (SOFA) score of 1, 3, or 4 (sepsis-1).  In 2001 a new definition (sepsis-2) expanded on the definition but retained the definition of sepsis as a systemic inflammatory response syndrome due to infection, and severe sepsis is associated with organ failure. More recently, in 2016, a definition of sepsis-3 was proposed, which is defined as both vasopressor-dependent hypotension and serum lactate >2 mmol/L after adequate fluid resuscitation. The ALBIOS trial compared the efficacy of albumin and crystalloids together versus crystalloid solutions alone in 1,818 patients with septic shock. Using the new definition of sepsis-3, the population decreased by 34%; those defined as having sepsis-3 had a higher lactate (p<0.001), a greater resuscitation fluid requirement (p=0.014), higher Simplified Acute Physiology Score II (p<0.001), a higher SOFA score (p=0.022), lower platelet counts (p=0.002), and higher 90-day mortality (51.9% vs. 43.5%; p=0.031) compared to those defined as having sepsis-2. The use of albumin was still shown to reduce mortality in both definitions of sepsis compared to crystalloid solutions alone (sepsis-2 43.7% vs. 54.9%, 12.6% relative risk reduction [p=0.04]; sepsis-3 48.7% vs. 54.9%, 11.3% relative risk reduction [p=0.22]). Although the benefits of albumin were still apparent after implementing this new criterion, this will impact future patient recruitment into randomised clinical trials (Vasques et al., 2018).A consensus statement from the European Society of Intensive Care Medicine (ESICM) on colloid volume therapy in critically ill patients, published in 2012, indicated that albumin could be used for the resuscitation of severe sepsis patients (Reinhart et al., 2012).

International Surviving Sepsis Campaign guidelines for the management of severe sepsis and septic shock, updated in 2016, suggested that albumin be used for fluid resuscitation of patients with severe sepsis and septic shock when substantial volumes of crystalloids are needed (Rhodes et al., 2017). Indeed, NICE guidelines also recommend that albumin should be considered only in patients with severe sepsis (NICE, 2013;2017).

Hydroxyethyl Starches (HES)

A series of systematic reviews and meta-analyses have highlighted that HES is associated with increased mortality and acute kidney injury (AKI) in critically ill patients (Groeneveld et al., 2011; Patel et al., 2013; Perel et al., 2013; Gattas et al., 2013; Haase et al., 2013; Mutter et al., 2013; Zarychanski et al., 2013; Serpa Neto et al., 2014; Wiedermann, 2014).

Increased mortality was reported in patients with severe sepsis or septic shock following HES therapy (Groeneveld et al., 2011; Patel et al., 2013; Serpa Neto et al., 2014), with a meta-analysis of 6 RCTs (n=3033) indicating that 90-day mortality was associated with HES exposure (relative risk [RR]=1.13; 95% confidence intervals [CI], 1.02–1.25; p=0.02) compared with crystalloid (Patel et al., 2013). The increased 90-day mortality with HES infusion was confirmed in a later meta-analysis (RR=1.14; 95% CI, 1.04–1.26; p=0.005) of 4,624 patients with sepsis, pooled from 10 trials. This meta-analysis also showed that in patients with sepsis, HES was significantly associated with an increased incidence of AKI (RR=1.24; 95% CI, 1.13–1.36; p<0.001), and increased requirement for renal replacement therapy (RR=1.36; 95% CI, 1.17–1.57; p<0.001) when compared with crystalloids (Serpa Neto et al., 2014).

Zarychanski et al., (2013) published a systematic review and meta-analysis of HES administration, examining mortality and AKI in critically ill patients. The review was an updated version of a previously published version (Zarychanski et al., 2009) which contained 7 studies by Dr Joachim Boldt. These were conducted without IRB approval and were subsequently retracted (Editors-in-Chief, 2011). Analysis of 10,290 critically ill adult patients involved in 28 trials (with the Boldt trials excluded) showed that HES was significantly associated with increased mortality (RR=1.09; 95% CI, 1.02–1.17); increased renal failure among 8,725 patients (RR=1.27; 95% CI, 1.09–1.47); and increased use of renal replacement therapy among 9,258 patients (RR=1.32; 95% CI, 1.15–1.50) (Zarychanski et al., 2013).

Alternative meta-analyses of acutely ill patients also reported significant associations of HES with increased mortality (Perel et al., 2013; Mutter et al., 2013; Anthon et al., 2017), increased use of renal replacement therapy (Gattas et al., 2013; Haase et al., 2013; Mutter et al., 2013); and incidence of kidney failure (Mutter et al., 2013).

Regulatory organisations in Europe and the United States state that HES should not be used in critically ill patients or in those with pre-existing renal dysfunction (European Medicines Agency [EMA] 2013; Medicines and Healthcare products Regulatory Agency [MHRA], 2014; Food and Drug Administration [FDA], 2013).  New measures have been put in place in Europe following increasing evidence that guidelines are not being followed leading to HES administration in critically ill patients and those with sepsis. The EMA have decided to keep HES on the market but new measures including a controlled access programme for accredited hospitals, warnings on the medicines’ packaging and SmPCs, and written contact with healthcare professionals have been put in place to protect patients at risk (EMA, 2018).

Liver Cirrhosis

Key points

• Volume replacement after large-volume paracentesis is recommended by clinical guidelines, to prevent post-paracentesis circulatory dysfunction
• Albumin performs better than other volume expanders or vasoconstrictors
• Long-term HA may have benefit in refractory ascites, but further research is needed
• A trial of TIPS in refractory ascites resulted in better 1-year survival versus LVP and albumin

The most common complication of liver cirrhosis is ascites, with other hepatic complications being hepatic encephalopathy and variceal haemorrhage (Runyon et al., 2009; Runyon, 2012). The development of ascites is associated with a poor prognosis, with 5-year survival of around 40% (Bernardi et al., 2014).

Patients with cirrhosis and ascites are at high risk for other complications of liver disease, including refractory ascites, spontaneous bacterial peritonitis (SBP), hepatorenal syndrome (HRS) and hyponatraemia (European Association for the Study of the Liver, 2010).

The figure below summarises the main pathophysiological events and complications of cirrhosis (Bernardi et al., 2014).

A study of 309 patients with cirrhosis found that 75.4% (n=233) had evidence of ascites. A bacterial infection was diagnosed in 44.6% of patients with cirrhosis and ascites (n=104), with the bacterial infection associated with renal failure in around a third of patients (33.6%; n=35) (Fasolato et al., 2007). In cirrhotic patients with spontaneous bacterial peritonitis renal failure is a risk factor for mortality (Navasa et al., 1998).

Albumin

Renal functional abnormalities and ascites formation in cirrhosis are the final consequence of circulatory dysfunction, characterised by marked splanchnic arterial vasodilatation, causing a reduction in effective arterial blood volume and homeostatic activation of vasoconstrictor and antinatriuretic mechanisms (Bernardi et al., 2014). 

Albumin infusions are very effective in preventing the deterioration in renal function associated with large-volume paracentesis (Bernardi et al., 2012), or spontaneous bacterial peritonitis. These are both conditions known to cause impairment of circulatory function in patients with cirrhosis and ascites. In this situation, albumin acts primarily as a plasma volume expander, acting to taper activated vasoconstrictor and sodium-retaining systems and improve renal perfusion.

In addition, albumin’s non-oncotic properties may also potentially target several pathophysiological mechanisms underlying decompensated cirrhosis, including albumin’s antioxidant and scavenging activities, the binding and transport of endogenous and exogenous substances, and the regulation of endothelial function and inflammatory or immune responses (Garcia-Martinez et al., 2013).

Ascites

Guidelines recommend that large-volume paracentesis is the first line therapy in patients with large ascites (tense and refractory) (European Association for the Study of the Liver, 2018; Runyon et al., 2009; Runyon, 2012). Large-volume paracentesis can lead to post-paracentesis circulatory dysfunction as a consequence of reduced effective circulating volume (Bernardi et al., 2014).

A meta-analysis of patients undergoing large-volume paracentesis (17 trials, n=1,225) showed that albumin infusion reduced morbidity and mortality compared with alternative treatments (other volume expander or vasoconstrictor). Compared with alternative treatments, albumin reduced the incidence of post-paracentesis circulatory dysfunction (OR=0.39; 95% CI, 0.27–0.55); reduced hyponatraemia (OR=0.58; 95% CI, 0.39–0.87); and reduced mortality (OR=0.64; 95% CI, 0.41–0.98) (Bernardi et al., 2012).

Data surrounding albumin infusion for treatment of refractory ascites remains limited leading to continuing debate. Di Pascoli et al. recently offered some interesting insights into the long-term use of albumin in patients with refractory ascites. In the study, patients were non-randomly assigned to receive either human albumin at a dose of 20 g twice a week in addition to standard of care (SOC), or SOC alone. Albumin appeared to have beneficial effects when administered long-term, as the cumulative incidence of 24-month mortality was significantly lower in the albumin and SOC group compared to the SOC alone group (41.6% vs. 65.5%; p=0.032). The period free of emergent hospitalisations was also significantly longer in the albumin and SOC group compared to those in the SOC alone group (p=0.008). The practical applications of this strategy could be beneficial to patients waiting for a liver transplant to increase survival during the waiting period. However, larger, multicenter, randomised control studies are needed to assess the efficacy of this treatment in this subgroup of patients with liver cirrhosis (Di Pascoli et al., 2018). Both European and American guidelines recommend an albumin infusion for large-volume paracentesis (LVP) to prevent circulatory dysfunction after LVP (European Association for the Study of the Liver, 2018; Runyon, 2012).

However, recent evidence from Austria has shown alternative reduction in portal pressure by self-expandable polytetrafluoroethylene (ePTFE)-covered transjugular intrahepatic portosystemic shunts (TIPS) to be a successful treatment option for refractory ascites with an improved 1-year survival in comparison to LVP plus albumin. The retrospective study investigated ePTFE-TIPS versus LVP + albumin for control of ascites, occurrence of hepatic encephalopathy (HE), and transplant-free survival in cirrhotic patients with refractory ascites. Control of ascites was reported in over half of the ePTFE-TIPS group (54%) without the need further paracentesis. The need for frequent LVP was also significantly lower in the ePTFE-TIPS group compared to LVP + albumin (median 49.5 [IQR: 5.07-102.60] vs. 0.67 [IQR: 0.23-2.63] months until paracentesis, log-rank p<0.001). Interestingly, ePTFE-TIPS was associated with an improved 1-year survival compared to LVP + albumin (65.6% vs. 48.4%, p=0.033), without increasing the incidence of HE. However, an association between ePTEE-TIPS and transplant-free survival wasn’t observed (Bucsics et al., 2018). 

This new evidence may impact future guideline recommendations for first line treatment of ascites.

Spontaneous bacterial peritonitis

A meta-analysis of patients with spontaneous bacterial peritonitis (4 trials; n=288), which compared with albumin infusion with no albumin infusion (3 trials) or HES (1 trial) showed improvement in renal function and reduced mortality in the albumin group. The incidence of renal impairment in control groups was 30.6% (44/144) compared with 8.3% (12/144) in albumin treated groups (OR=0.21; 95% CI, 0.11–0.42). Mortality in control groups was 35.4% (51/144), compared with 16.0% (23/144) among patients who received albumin (OR=0.34; 95% CI, 0.19–0.60) (Salerno et al., 2013).

European guidelines recommend that patients with SBP should be treated with intravenous albumin. Evidence also suggests that albumin use in SBP may prevent acute kidney injury. However, routine use of albumin is not recommended for all other bacterial infections (European Association for the Study of the Liver, 2018).

Hepatorenal syndrome

Hepatorenal syndrome (HRS) is defined as the occurrence of renal failure in a patient with advanced liver disease in the absence of an identifiable cause of renal failure (European Association for the Study of the Liver, 2010). The condition is a serious complication of end-stage liver disease, which commonly occurs in patients with advanced cirrhosis and ascites, who have pronounced circulatory dysfunction, and also in patients with acute liver failure (Salerno et al., 2007). Type 1 HRS is characterised by a rapidly progressive reduction in renal function, which is commonly triggered by a bacterial infection, mainly SBP (Salerno et al., 2007; Runyon et al., 2009; Runyon, 2012).

A systematic review of vasoconstrictor drugs for HRS (10 trials, n=376) showed that vasoconstrictor drugs (terlipressin, octreotide or noradrenaline) used alone or in combination with albumin reduced mortality compared with no intervention or albumin (RR=0.82; 95% CI, 0.70–0.96). A subgroup analysis showed that terlipressin plus albumin reduced mortality compared with albumin alone (RR=0.81; 95% CI, 0.68–0.97) (Gluud et al., 2010).

A recent systematic review and meta-analysis of 13 RCTs found terlipressin plus albumin to be more efficacious than placebo plus albumin (OR = 4.72; 95% CI 1.72–12.93; p=0.003) or midodrine plus albumin and octreotide (OR = 5.94; 95% CI 1.69–20.85; p=0.005) for HRS reversal (Nanda et al., 2018).

Both European and American guidelines recommend albumin plus a vasoactive agent (e.g. terlipressin, or midodrine plus octeotride) as first-line pharmacological therapy for type 1 HRS (Runyon et al., 2009; Runyon, 2012; Lenz et al., 2015; European Association for the Study of the Liver, 2018). 

A meta-analysis of 19 clinical studies (8 RCTs, 8 prospective studies, 3 retrospective studies; n=574) aimed to determine the impact of albumin dose on treatment outcomes in type 1 HRS. HRS reversal was calculated to be 49.5% (95% CI, 40.0–59.1; n=525) with patients having a 30-day survival of 50.6% (95% CI, 45.5–56.3; n=377). Expected percentages of patients surviving at 30 days, calculated from Cox regression analyses, were 43.2%, 51.4% and 59.0% for patients receiving 200g, 400g and 600g albumin, respectively (Salerno et al., 2015).

Guidelines

European Association for the Study of the Liver (EASL) clinical practice guidelines on the management of decompensated cirrhosis, including ascites, spontaneous bacterial peritonitis (SBP), and hepatorenal syndrome in cirrhosis are available (European Association for the Study of the Liver, 2018).

Based on these guideline the following 4-minute animation summarises current recommendations regarding the treatment of these conditions, including key information on the role of human albumin.  

Further details regarding the use of human albumin for the treatment of patients with ascites, spontaneous bacterial peritonitis, acute kidney injury and hepatorenal syndrome can be found in the Albumin section below.

More information: EASL clinical practice guidelines on the management of ascites, spontaneous bacterial peritonitis, and hepatorenal syndrome in cirrhosis.

The American Association for the Study of Liver Diseases (AASLD) has produced guidelines for the management of adult patients with ascites due to cirrhosis (Runyon et al., 2009) which were updated in 2012 (Runyon, 2012).

Revised consensus recommendations from the International Club of Ascites (ICA) for the diagnosis and management of acute kidney injury (AKI) in patients with cirrhosis were recently published. A proposed algorithm for the management of AKI in these patients is shown below (Angeli et al., 2015).

The ICA-AKI classification combines Kidney Disease Improving Global Outcomes (KDIGO) criteria and conventional criteria in patients with cirrhosis and ascites. In patients with cirrhosis and ascites with initial ICA-AKI stage 1, plasma volume expansion using crystalloids, albumin or blood is recommended for patients with clinically suspected hypovolaemia. In ICA-AKI stage 2/3 recommendations include withdrawal of diuretics and plasma volume expansion using intravenous albumin (1g/kg) for 2 days (Angeli et al., 2015).

A series of recommendations were made by the Italian Association for the Study of the Liver (AISF) and the Italian Society of Transfusion Medicine and Immunohaematology (SIMTI) regarding the appropriate use of albumin in patients with liver cirrhosis. These include mandatory indications for albumin in the prevention of post-paracentesis circulatory dysfunction; prevention of renal failure after SBP; and treatment of type I hepatorenal syndrome (in combination with vasoconstrictors) (Caraceni et al., 2016).

Cardiac Surgery

Key points

• Both pre- and postoperative hypoalbuminaemia are associated with adverse outcomes, such as increased incidence of AKI in children and an increased mortality
• Albumin is 5 times more effective than saline as plasma volume
• Albumin versus HES reduced mortality after CABG
• Discontinuation of HES in one Canadian institution led to a shorter hospital stay and reduced risk of red blood cell, plasma and platelet transfusion
• There are significant differences in the effects on blood coagulation parameters between different fluid therapies, with prolonged clot formation and reduced clot firmness with HES

Cardiac surgery including coronary-artery bypass grafting (CABG) is either performed using cardiopulmonary bypass (on-pump CABG), or without cardiopulmonary bypass (off-pump CABG) (Shroyer et al., 2009).

Appropriate fluid management in cardiac surgery patients is critical, but the choice of colloid or crystalloid – and which colloid – is still actively discussed (Schumacher & Klotz, 2009; Shaw & Raghunathan, 2013).

Hypoalbuminaemia

Both preoperative and postoperative albumin levels have an impact on outcomes in patients with cardiac surgery. The data and survival figures are discussed below, as is the role of albumin replacement in this high-risk group of patients.

Preoperative hypoalbuminaemia

Preoperative low albumin concentration (<2.5 g/dL) was independently associated with increased mortality in patients (n=5,168) undergoing cardiopulmonary bypass surgery (p≤0.0005). Analysis showed that an albumin concentration <2.5 g/dL was independently associated with increased risk of new atrial fibrillation (odds ratio [OR]=1.4; 95% confidence interval [CI], 1.2–1.7; p=0.0001); postoperative renal failure (OR=2.0; 95% CI, 1.3–3.2; p=0.002); reoperation for bleeding (OR=1.4; 95% CI, 1.0–2.1; p=0.04); prolonged ventilatory support (OR=2.5; 95% CI, 2.0–3.1; p=0.0001), ICU stay >3 days (OR=1.7; 95% CI, 1.4–2.1; p=0.0001), and prolonged total length of stay (OR=1.9; 95% CI, 1.5–2.4; p=0.0001) (Engelman et al., 1999).

Cardiac surgery patients with preoperative hypoalbuminaemia (albumin concentration <3.5 g/dL; n=337) had an increased risk of long-term mortality compared to those with pre-operative albumin concentrations ≥3.5 g/dL (n=827): hazard ratio (HR)=2.2 (95% CI, 1.4–3.6; p=0.001).

The figure below shows Kaplan–Meier survival curves after CABG surgery in the two groups (De la Cruz et al., 2011).

This study found no significant differences in 30-day mortality rates (albumin <3.5 g/dL, 2% vs. albumin ≥3.5 g/dL, 1.7%; p=0.76), nor adverse cardiac events between the two groups (De la Cruz et al., 2011).

In elderly cardiac surgical patients (≥75 years, n=92) multivariate analysis showed that preoperative hypoalbuminaemia (albumin <3.5 g/dL) was a predictor of postoperative renal dysfunction (p<0.01), and an independent predictor of increased length of stay (p<0.01) (Rich et al., 1989). 

In patients undergoing off-pump coronary artery bypass surgery with low preoperative albumin levels (<4.0 g/dL), pre-surgery administration of 20% albumin significantly reduced the risk of postoperative acute kidney injury (AKI) and increased urine output during surgery. AKI, defined by AKI Network criteria, was 13.7% in the albumin group (n=102) compared with 25.7% in the control group (n=101) (RR=0.53; 95% CI, 0.30–0.96). Median intra-operative urine output was 370 mL in the control group compared with 550 mL in the albumin group (p=0.006) (Lee et al., 2016).

A low preoperative serum albumin level has also been shown to be associated with postoperative AKI in paediatric patients undergoing congenital cardiac surgery (OR 0.506, 95% CI 0.325–0.788; p=0.003) (Lee et al., 2018).

Postoperative hypoalbuminaemia

Postoperative hypoalbuminaemia (albumin <2.3 g/dL) in patients who underwent off-pump CABG surgery (n=690) was associated independently with multiple 30-day adverse outcomes. These included renal failure (OR=7.98, 95% CI, 1.59–40.1; p=0.01); the need for postoperative intra-aortic balloon pump (OR=13.7; 95% CI, 1.53–125; p=0.02); the need for inotropes in the intensive care unit (OR=1.79; 95% CI, 1.12–2.86; p=0.02); and reoperation for bleeding (OR=4.33; 95% CI, 1.02–18.5; p=0.05) (Lee et al., 2011).

Postoperative serum albumin <1.8 g/dL was the strongest predictor of 28-day mortality in a study of 454 cardiac surgical patients. Logistic regression analysis calculated an odds ratio of 0.86 (95% CI, 0.84–0.89) for albumin as a predictor of mortality (Fritz et al., 2003).

Low serum albumin postoperatively has also been shown to be an independent risk factor for acute kidney injury (AKI) in paediatric patients undergoing cardiac surgery. In a retrospective analysis of 1,110 children undergoing cardiopulmonary bypass, Duan et al. found patients with hypoalbuminaemia (serum albumin ≤35 g/L) to have a significantly higher incidence of AKI than those with normal albumin levels (p<0.05). Patients with AKI were also observed too have significantly lower serum albumin levels 48 hours postoperatively than those without AKI (p<0.05) (Duan et al., 2018).

Albumin

A postoperative study of 40 cardiac surgical patients showed that although albumin and saline had comparable effects on oxygen delivery and changes in interstitial and extracellular fluid volume, 5% albumin was approximately five times more efficient as a plasma volume expander in comparison with normal saline. Increases in plasma volume were 52% ± 84 and 9% ± 23 respectively (<0.05) (Ernest et al., 2001).

Albumin, when compared with a non-protein colloid (hydroxyethl starch or dextran), reduced mortality after coronary artery bypass graft surgery. Mortality was significantly lower (p=0.02) in the albumin group (2.47%; n=8,084) compared to the non-protein colloid group (3.03%; n=11,494). Multivariable analysis showed that albumin use was associated with 20% lower odds of mortality compared to non-protein colloid use (OR= 0.80; 95% CI, 0.67–0.96) (Sedrakyan et al., 2003).

A comparison of the effects of 5% albumin plus Ringer's lactate, 6% HES 130/0.4 plus Ringer's lactate, and Ringer's lactate alone on blood loss in patients following cardiac surgery, showed no significant differences between the groups (p=0.085). However, significant differences between the groups were found in the quantity of blood transfused (p=0.0004); with patients in the Ringer’s lactate group (n=79) receiving fewer packed red blood cells (PRBCs) compared with those in the human serum albumin (n=76; p=0.0015) and HES (n=81; p=0.0002) groups. The proportion of patients who received PRBCs was significantly higher in the colloid groups; values were: albumin, 58%; HES, 61%; and Ringer's lactate, 34% (p=0.0013) (Skhirtladze et al., 2014).

A decreased risk of in-hospital mortality has also been reported in patients receiving 5% albumin plus crystalloids versus crystalloids alone (OR 0.5; 95% CI 0.3–0.9; p=0.02). In an observational study of patients who underwent on-pump cardiac surgery for valve and/or coronary artery procedures between 2001–2013, patients either received albumin plus crystalloids (n=1,095) or crystalloids alone (n=1,095) to assess all cause mortality rates, AKI severity, major morbidity composite, and all-cause 30-day readmissions. As well as in-hospital mortality, patients in the albumin plus crystalloids group also had lower all-cause 30-day readmission rates (OR 0.7; 98.3% CI 0.5–0.9; p<0.01). However, albumin therapy wasn’t associated with differences in overall morbidity or AKI severity in comparison to crystalloids (Kingeter et al., 2018). 

Another study compared albumin to crystalloids in cardiac surgery patients after a hospital policy change. The policy discontinued 5% albumin as the main resuscitation fluid following open cardiac surgery and crystalloids have since been used in its place (mainly normal saline and Plasma-Lyte). This retrospective study included 360 paediatric patients who had undergone cardiac surgery and received either 5% albumin (n = 208) in the 6 months prior to the policy change, or crystalloids (n = 158) in the 6-month period after the policy change. Interestingly, on day 1 postoperatively there was no association between fluid group and fluid intake (coefficient 2.84, 95% CI 5.37–11.05; p=0.497). However, those who had been administered crystalloids received significantly less fluid on day 2 (coefficient -12.8, 95% CI -22.0 to -3.65; p=0.006), day 3 (coefficient -14.9, 95% CI -24.3 to -5.57; p=0.002), and in the first 48 hours postoperatively (coefficient 10.1, 95% CI -27.9 to -1.29; p=0.032) compared to the albumin group. The length of ICU stay following surgery was also shorter in the crystalloid group compared to the albumin group (coefficient -1.29, 95% CI -2.5 to -0.08; p=0.036). The findings of the study contradict the theoretic oncotic advantage of albumin, as no measured clinical benefits were observed from administering albumin. Rather the opposite was seen, with an increase in side effects and increased health care cost (Dingankar et al., 2018).

Hydroxyethyl Starches

A meta-analysis of the effect of hydroxyethyl starches (HES) on postoperative blood loss after cardiopulmonary bypass surgery identified 18 trials with a total of 970 patients. Compared with albumin, HES (450/0.7 or 200/0.5) increased postoperative blood loss by 33.3% (95% CI, 18.2–48.3; p<0.001). The risk of reoperation for bleeding was more than doubled by HES compared to albumin (RR=2.24; 95% CI, 1.14–4.40; p=0.02). Compared with albumin, HES infusion increased transfusion of red blood cells (RBCs) by 28.4% (95% CI, 12.2–44.6; p<0.001); fresh-frozen plasma (FFP) by 30.6% (95% CI, 8.0–53.1; p=0.008); and platelets by 29.8% (95% CI, 3.4–56.2; p=0.027) (Navickis et al., 2012).

The use of artificial colloids such as HES during cardiopulmonary bypass surgery to counteract fluid extravasation remains a controversial strategy. There have been claims of beneficial effects on global fluid loading,.However, there are concerns that renal function and coagulation may be adversely affected as a result. A randomised control trial recently tested this theory by comparing HES solution (130/0.42) to Ringer’s solution as priming solutions during elective coronary artery bypass graft (CABG) surgery. Interestingly, HES was shown to be more beneficial in comparison to crystalloids as HES resulted in a lower fluid accumulation (3,374 [± 883] mL vs. 4,328 [± 1,469] mL; p=0.024). The reduction in perioperative fluid accumulation also appeared to increase cardiac index immediately after surgery (HES group 2.7 [± 0.4] L/min/m² vs. crystalloid group 2.1 [± 0.3] L/min/m²; p<0.001). Despite concerns that HES may be linked to an increase in bleeding in surgical patients, this wasn’t observed in this study. However, three patients within the HES group developed AKI postoperatively; the occurrence of AKI exclusively in the HES group does raise some questions over the safety and suitability of this solution in cardiac surgery (Svendsen et al., 2018). Direct comparison by meta-analysis of 272 patients in 4 trials who received HES 130/0.4 or HES 200/0.5 found no significant differences in postoperative blood loss (p=0 .21); the risk of reoperation for bleeding (p=0.62); or the transfusion of RBCs (p=0.24), FFP (p=0.70), or platelets (p=0.46) (Navickis et al., 2012).

Interestingly, a recent randomised study of 195 patients found HES to have no significant effect on the incidence of AKI following cardiac surgery in paediatric patients compared to a crystalloid fluid (HES group 40.8% vs. control group 30.0%; p=0.150). The study also found no difference in clinical outcomes between the two groups, such as mortality, major adverse events, intensive care unit stay, and duration of mechanical ventilation (Oh et al., 2018).

HES use remains restricted across Europe, the USA, and Canada and should not be used in critically ill patients; HES solutions are often used in perioperative settings due to conceptions that surgical patients are at a reduced risk of organ failure. However, a systematic review and meta-analysis carried out by the Centre for Medical Evidence, Decision Integrity & Clinical Impact (MEDICI) revealed that patients exposed to HES were at a greater risk of death, renal replacement therapy, and blood product transfusion compared with patients receiving alternative fluid therapies. Following the findings of this analysis, the London Health Sciences Centre (LHSC), Canada, made the decision to discontinue the use of HES for fluid replacement therapy across the entire institution. After the discontinuation, Hong et al. conducted a retrospective analysis of patients who had undergone coronary artery bypass surgery either two years prior or two years after the restriction was implemented, to compare the effects on length of hospital stay, transfusion, risk of death, AKI, and dialysis. Of the patients who had undergone cardiac surgery during this period, 1,496 had been exposed to HES and 1,261 patients had not received HES. Propensity score-matched analysis revealed that discontinuation of HES was associated with a shorter length of hospital stay (HR 1.24, 95% CI 1.14–1.35), along with a reduced risk of red blood cell, plasma, and platelet transfusions. In-hospital mortality was not associated with the discontinuation of HES solution (OR 0.74, 95% CI 0.36–1.54), nor was AKI (OR 0.84, 95% CI 0.57–1.25), or dialysis (OR 0.83, 95% CI 0.25–2.73) (Hong et al., 2018). 

Effect on blood coagulation parameters

An RCT of 45 patients following cardiac surgery found that succinylated gelatin and HES (200/0.5), but not albumin, impaired haemostasis after cardiac surgery. The coagulation time, which indicates initial fibrin formation, was prolonged immediately following HES infusion but not after albumin or gelatin infusions (p<0.05 for comparisons of HES with albumin or gelatin). Immediate reductions in maximum clot firmness were found following gelatin HES infusions but not after albumin infusion (p<0.05 for comparisons of HES or gelatin with albumin) (Niemi et al., 2006).

An RCT comparing two HES solutions (6% HES 200/0.5 or 6% HES 130/0.4), and 4% albumin in 45 patients following cardiac surgery found that HES solutions impaired blood coagulation. Both HES solutions prolonged the thromboelastometric parameters – clot formation time and reduced maximum clot firmness – when compared with albumin (p<0.05 for all comparisons) (Schramko et al., 2009).

Significant differences between plasma expanders were reported for thromboelastometric parameters on arrival in ICU in patients following cardiac surgery who had received albumin (n=76), HES 130/0.4 (n=81), or Ringer's lactate (n=79) solutions. Clot formation time was significantly different (p<0.001) in albumin (median=137s), HES (median=185s) and Ringer's lactate (median=107s) groups. Differences between groups were also significant (p<0.05 for each comparison). Maximal clot firmness values were also significantly different in the treatment groups (p<0.001): median values were 10mm for albumin; 7mm, HES; and 13mm, Ringer's lactate. Differences between groups were also significant (p<0.05 for each comparison) (Skhirtladze et al., 2014).

Complications

An RCT of 156 patients receiving off-pump CABG surgery showed an increase in bleeding complications following hydroxyethyl starch (450/0.7) infusion in comparison with albumin infusion.

The table below shows data for the primary outcome (transfusion requirements) and blood loss following surgery (Hecht-Dolnik et al., 2009).

Table 4. Primary outcome (transfusion requirements) and blood loss following surgery.

A meta-analysis of general surgical patients (n=1,230; 17 studies) who were treated with HES 130/0.40 or a colloidal or crystalloidal control solution reported no abnormal effect of the tetrastarch on renal function. No significant differences between HES 130/0.40 and a control solution were found for acute renal failure (p=0.98), incidence of renal replacement therapy (p=0.85), maximum serum creatinine values (p=0.65) or calculated creatinine clearance values (p=0.14) (Martin et al., 2013).

A meta-analysis which identified 49 studies with a total of 3,439 cardiac surgical patients, compared HES with other colloid and crystalloid solutions for blood loss and transfusion requirements. Comparisons of tetrastarch to albumin reported lower blood loss with tetrastarch (3 studies, n=185: standardised mean difference (SMD) =−0.34; 95% CI, −0.63 to −0.05; p=0.02); and reduced transfusion requirements with tetrastarch (3 studies, n=185: RR=0.70; 95% CI, 0.56–0.89). In contrast, there was a higher transfusion need for hetastarch, when compared to albumin (4 studies: RR=1.48; 95% CI, 1.04–2.10; p=0.03) (Jacob et al., 2014).

The study by Jacob and colleagues was heavily criticised by Navickis and colleagues who demonstrated by meta-analysis of three RCTs, that postoperative blood loss was lower in patients receiving albumin than those treated with HES, although the difference was not statistically significant (Navickis et al., 2015).

Effects on acute kidney injury and postoperative bleeding

A retrospective review of the medical records of 771 patients who underwent off-pump CABG surgery compared outcomes in patients treated with and without intraoperative HES. The results in matched cohorts showed no significant differences for median postoperative blood loss at 24 hours (525 mL with HES, vs. 540 mL without HES; p=0.203). However postoperative AKI was more common with patients treated with HES (10.7%) vs. without HES (3.6%; p<0.001). Renal safety therefore remains a concern in patients undergoing off-pump CABG and HES should be used with caution in these patients (Min et al., 2017).