Literature Summary

Major Literature

1. Finfer. A comparison of albumin and saline for fluid resuscitation in the intensive care unit. N Engl J Med 2004;350(22):2247-56
  • In a  multicenter, randomized, double-blind trial comparing 0.9% saline or 4% albumin for fluid resuscitation in 6997 critically ill patients in the ICU, there was no difference in mortality (729 v 726, RR 0.99; 95 CI 0.91 to 1.09; P=0.87), new single-organ and multiple-organ failure (P=0.85), mean (SD) numbers of ICU days (6.2±6.2 v 6.5±6.6, P=0.44), hospital days (15.6±9.6 v 15.3±9.6; P=0.30), days of mechanical ventilation (4.3±5.7 v 4.5±6.1; P=0.74), or days of renal-replacement therapy (0.4±2.0 v 0.5±2.3) respectively
  • In a blinded randomized controlled trial comparing 6% hydroxyethyl starch 130/0.42 (Voluven) with 0.9% saline for fluid resusciation in 7000 critically ill patients, this colloid therapy was associated with a 21% increased risk of the requirement for renal replacement therapy ( HES RRT requirement 7.0% versus saline 5.8%; relative risk 1.21; 95% CI 1.00 to 1.45; P=0.04 and no mortality benefit (HES mortality 18.0% versus  saline mortality 17.0%; relative risk in the HES group, 1.06; 95% CI 0.96 to 1.18; P=0.26). Starch therapy was also associated with increased rates of hepatic failure, rash and pruritus.
3. Maitland. Mortality after Fluid Bolus in African Children with Severe Infection (FEAST Trial). N Engl J Med 2011;364:2483-2495   (Paediatric Study)
  • Maitland et al performed a stratified (severe hypotension or not), multicenter, randomized control trial, in a resource-limited setting in sub-Saharan Africa, comparing a fluid bolus (20 to 40 ml of 5% albumin or 0.9% saline) with no fluid bolus at admission to hospital in 3,141 children with febrile illness and impaired perfusion, and found fluid bolus therapy was associated with a higher mortality at 48 hours (albumin  10.6%, saline 10.5%, no bolus 7.3%; relative risk bolus therapy versus no bolus 1.45, 95% CI 1.13 to 1.86, P=0.003), and 28 days (12.2%, 12.0% & 8.7%, respectively; RR bolus therapy versus no bolus p=0.004), with similar incidences of pulmonary oedema, increased intracranial pressure (2.6%, 2.2% versus 1.7% P=0.17), and neurological sequela in the three groups (P=0.92).

Albumin vs. Crystalloid:

Current Guidelines

2012 - Adult Surviving Sepsis Campaign
  • "Crystalloids as the initial fluid of choice in the resuscitation of severe sepsis and septic shock"
  • "Use of albumin in the fluid resuscitation of severe sepsis and septic shock when patients require substantial amounts of crystalloids"
    • "Substantial amounts" is undefined in the guideline
2012 - Pediatric Surviving Sepsis Campaign
  • "...initial resuscitation of hypovolemic shock begins with infusion of isotonic crystalloids or albumin..."
So the guidelines are unclear when to use one over the other...

Recent Meta-analyses

2011 - The Role of Albumin as a Resuscitation Fluid for Pts with Sepsis (Critical Care Med)
  • Methods
    • Meta-analysis of 17 RCTs (total of 1977 pts, 14 adult and 3 pediatric trials) comparing albumin with other fluids (NS, LR, starches, gelatin) for resuscitation in sepsis
    • Included trials available up to 2010
  • Results
    • Albumin reduced mortality in all age groups when...
      • Compared to all types of fluids (OR 0.82, p = 0.05)
      • Compared to crystalloids only (OR 0.78, p = 0.04)
    • Concentrated albumin (20% or greater albumin) increased mortality in all age groups when...
      • Compared to all types of fluids (OR 1.08, p = 0.73)
    • Diluted albumin (4-5% albumin) reduced mortality in all age groups when...
      • Compared to all types of fluids (OR 0.76, p = 0.02)
    • Albumin reduced mortality in pediatric population when...
      • Compared to all types of fluids (OR 0.29, p = 0.008)
    • Albumin reduced mortality in adult population when...
      • Compared to all types of fluids (OR 0.87, p = 0.18)
  • Bottom line
    • Trend towards reduced mortality in children and adults when using albumin
    • Do not use boluses of concentrated albumin (25% albumin at U of M)
    • Peds population seems to like albumin, but only based on 3 small trials
2013 - Colloids vs. Crystalloids for Fluid Resuscitation in Critically Ill Pts (Cochrane Review)
  • Methods
    • Meta-analysis of 24 RCTs (total of 9920 pts who were ill due to trauma, burns, surgery, sepsis, excluding neonates and pregnant women) comparing colloids (albumin, plasma) to crystalloids (isotonic, hypertonic)
    • Included trials available up to 2012
  • Results
    • No difference in mortality (OR 1.01, CI 0.93-1.10)
  • Bottom line
    • No mortality benefit for resuscitation with albumin/plasma

Recent RCT Not Included in the Above Meta-analyses

2013 - Effects of FLuid Resuscitation with Colloids vs. Crystalloids in Hypovolemic Shock (JAMA)
  • Methods
    • European study performed at 57 ICUs
    • Adult pts with hypovolemic shock from sepsis, trauma, or other causes randomized to cyrstalloids (isotonic or hypotonic saline) vs. colloids (hypo-oncotic [gelatins, 4-5% albumin] or hyper-oncotic [dextrans, hydroxyethyl starches, 20-25% albumin])
    • All MIVF were isotonic saline
    • If pts became hypoalbuminemic (albumin <2.0), then they could receive albumin regardless of treatment group
  • Results
    • 1414 pts in colloid group, 1443 pts in crystalloid group
      • No difference in baseline characteristics
      • 16% of pts in crystalloid group received albumin supp
      • No difference in total amount of blood products transfused
      • No difference in proportion of pts who received renal replacement therapy (RRT)
    • 28 day mortality (colloid vs. crystalloid) [NS = non-significant]
      • In all group:  25.4% vs. 27% (NS)
      • Just for sepsis:  27.8% vs. 29.0% (NS)
      • Just for albumin vs. NS:  30.0% vs. 26.6% (NS)
      • Just for albumin vs. NS in sepsis:  27% vs. 28.2% (NS)
    • 90 day mortality (colloid vs. crystalloid)
      • In all groups:  30.7% vs. 34.2% (p = 0.03)
      • Just for sepsis:  32.6% vs. 36.7% (NS)
      • Just for albumin vs. NS:  35.0% vs. 33.4% (NS)
      • Just for albumin vs. NS in sepsis:  37.3% vs. 35.4% (NS)
    • Colloid group had more days free of...
      • Mechanical ventilation
      • Vasopressor
    • No difference between groups in days free of...
      • RRT
      • Low SOFA score
      • ICU or hospital stay
  • Bottom line
    • No short-term mortality difference in using colloids over crystalloids in septic and other hypovolemic pts
    • Colloids associated with better long-term mortality, but no difference when comparing albumin to NS or when comparing colloids to crystalloids in septic pts
    • Colloids seem to reduce days on mechanical ventilation or vasopressor but does not influence the trend of overall organ failure score or length of hospital stay
Myburgh, NEJM 2013
  • Ideal resuscitation fluid (does not currently exist)
    • Increases intravascular volume
    • Chemical composition close to ECF
    • Does not accumulate in tissue
    • No associated metabolic/systemic effects
    • Cheap!
  • Colloid (albumin - Osmol 250, Na 148, Cl 128; semi-synthetic - varies)
    • Increases oncotic pressure
    • Remains in the intravascular space
    • Requires less volume compared to crystalloids to maintain intravascular volume
    • Expensive
  • Crystalloid (NS- Osmol 308, Na 154, Cl 154; LR - Osmol 281, Na 131, Cl 111, K/Ca/lactate)
    • Affects tonicity of fluid
    • Cheap
    • Can lead to interstitial edema
    • Can lead to hyperchloremic metabolic acidosis --> possible immune and renal dysfunction
  • Evidence of Benefit/Harm of Various Fluids
    • Colloids vs. Crystalloids
      • 1998:  meta-analysis on albumin vs. crystalloids in hypovolemic, burn, or hypoalbuminemic pts
        • Albumin associated with increase in rate of death
      • 2004:  RCT on 4% albumin vs. saline in ICU pts (SAFE study mentioned by Dr. Annich)
        • No difference in death or organ failure at 28 days
      • 2007:  Subanalysis of 2004 study
        • Increased rate of death in TBI pts with albumin (likely elevates ICP)
        • Decreased risk of death in severely septic pts with albumin
        • No difference in death in hypoalbuminemic pts with albumin vs. saline
        • No difference in MAP or HR endpoint with albumin vs. saline but required less volume of albumin for resuscitation
      • 2011:  RCT on boluses of albumin or saline vs. no boluses of albumin or saline in children
        • Bolus albumin vs. saline had equal mortality rate at 48 hours
        • "Bolus albumin or saline" had higher mortality at 48 hours compared to "no bolus albumin or saline"
          • Cause of death was CV collapse (potential interaction between bolus'ing and compensatory neurohormonal response?)
    • Semi-synthetic Colloids vs. Crystalloids
      • Since we don't use this, I'll keep this section short
        • Increased rate of death and CRRT with semi-synthetic
        • No difference in hemodynamic resuscitation end points
    • Hypertonic Saline vs. Normal Saline
      • 2004:  RCT on "small volume" resuscitation with hypertonic saline vs. normal saline in TBI pts --> no short/long-term benefits
    • Chloride Restrictive Fluid vs. Chloride Rich Fluid
      • 2012:  RCT of Cl low fluid (LR, chloride poor albumin, plasmalyte) vs. Cl high fluid (NS, albumin, semi-synthetic) in ICU --> decreased incidence of AKI and CRRT
  • In summary
    • NS for hypovolemic and alkalotic pts
    • Consider albumin for early resuscitation of severely septic pts
    • Do not use albumin for TBI pts
    • Semi-synthetic fluids = FAIL
    • Perhaps we should be using LR rather than NS...can it be that surgeons were actually correct all this time?
      • This idea was further strengthened by the most recent study looking at the choice of crystalloid fluid published in CCM this year