Critical Care, Critical Care Alert

Critical Care Alert: Crystalloids vs Colloids for Fluid Resuscitation in the Intensive Care Unit: A Systematic Review and Meta-analysis

Critical Care Alert

Martin GS, Bassett P. Crystalloids vs colloids for fluid resuscitation in the Intensive Care Unit: A systematic review and meta-analysis. Journal of Critical Care. 2019;50:144-154.

Surviving Sepsis Guidelines recommend crystalloids for initial fluid resuscitation then switching to albumin when crystalloids are insufficient. There is debate regarding using colloid as an initial resuscitative fluid over crystalloid. The authors of this article aimed to compare crystalloid and colloid in their sufficiency for volume replacement.

Current international guidelines follow recommendations from the Surviving Sepsis Campaign with regards to fluid resuscitation. The current avoidance of colloid as an initial resuscitative fluid is largely the result of limited efficacy and safety data. While some data do suggest that specific colloids should be avoided in certain patient populations, there is no one size fits all with regards to fluid resuscitation. The evidence provided by the authors in this article suggests that the Surviving Sepsis guideline of 30 mL/kg crystalloid should not be followed in all cases of sepsis.

Study Design

  • Systematic review and meta-analysis of randomized control trials
  • Sub-analysis, statistical analysis, data extraction and bias assessment
  • Inclusion Criteria
    • Published randomized control clinical trials
    • Human subjects
    • Adult subjects > 18 years of age
    • Critically ill patients
    • A requirement for acute volume resuscitation
    • Allocation to resuscitation with crystalloid compared with non-crystalloid active comparator fluid therapy
    • Full, original articles
  • Exclusion Criteria
    • Observational studies, quasi-randomized, cross-over, or cluster randomized trials
    • Non-clinical reports
    • Studies focused on pediatric or neonatal subjects
    • Studies in healthy volunteer
    • Studies on elective surgical procedures
    • No other relevant or active comparator fluid
    • Articles not reporting original data
    • Conference abstracts or proceedings
    • Papers that have been retracted
    • Guidelines, meta-analysis, systematic review, literature review, editorial, commentary, case report or series, conference abstract or proceeding
    • Studies with Boldt as an author
    • Full text not available in English
  • 55 total studies included, the earliest dating to 1977
  • All studies included a crystalloid arm and at least one colloid comparator

Outcome Measures

  • Primary outcome
    • Sufficiency of crystalloids
    • Volume efficacy
    • Patient-centered outcomes
    • Hemodynamic outcome
  • Secondary outcomes
    • Tissue perfusion (UOP, lactate)
    • Colloid osmotic pressure
    • Creatinine
    • Oxygen saturation
    • pH
    • Serum albumin
    • Hematocrit/hemoglobin
    • Coagulation
    • Adverse events


Efficacy of crystalloids vs colloids

  • Crystalloids sufficient in achieving target thresholds in 11/14 (79%) of studies; target MAP achieved within target range but unable to reach target CVP in the remaining 3 studies
  • Colloids sufficient at raising CVP to target threshold (8-12 mm Hg) in 14/14 (100%) of studies
  • In all instances, maximum CVP significantly lower in the crystalloid group compared to the colloid group (-2.0 mm Hg vs -4.3 mm Hg, p < 0.001)
  • Maximum MAP lower in crystalloid group compared to albumin (-3.5 mm Hg, p = 0.03) or gelatin (-9.2 mm Hg, p = 0.02)

Fluid Volume

  • Volume administered with crystalloids significantly higher than with Hydroxyethyl Starch (HES) (+1775 mL, p = 0.006)


  • Large and relevant sample size (55 studies, ~27000 critically ill patients)
  • Meta-analysis
  • Majority of RCTs included were in fields of sepsis, shock, trauma, and critical care (thus relevant to ED critical care)


  • Broad selection criteria (no restrictions on publication date or ICU setting)
  • Heterogeneity between selected studies
  • Questionable validity to CVP, MAP, and CI as measure of fluid resuscitation

This article gathered data from high-quality RCTs in critical care settings and showed that crystalloids were statistically less effective than colloids in reaching resuscitative hemodynamic endpoints (CVP, MAP, and CI). However, the overall lack of a significant difference in mortality between crystalloid and albumin, gelatin, or dextran suggests that these endpoints may not be clinically significant. HES was the only colloid with an associated increased mortality in comparison to crystalloid, suggesting that other colloids (i.e. albumin) are more appropriate options in fluid resuscitation.

ED Take Away Points
As the authors of this study suggest, there is no “one size fits all” resuscitative fluid. The statistically significant difference in fluid volumes for crystalloid and colloid suggests that clinicians should consider colloid as the resuscitative fluid of choice in patients with tenuous fluid states. This article is another step in the direction toward proving that colloids are, at least, not harmful in comparison to crystalloids. However, more quality prospective studies need to be performed before this data can be extrapolated toward standard use in clinical practice.

Related Articles

Critical Care ECMO Series: Introduction to ECMO

Extracorporeal membrane oxygenation (ECMO) allows for temporary life support in cardiopulmonary failure refractory to conventional medical treatment. Given its capabilities and increasing presence in

Critical Care Alert: Point-of-Care Ultrasound in Out-of-Hospital and in-ED Cardiac Arrests

Critical Care Alert: Point-of-Care Ultrasound in Out-of-Hospital and in-ED Cardiac Arrests A 57-year-old male with unknown past medical history becomes unresponsive as witnessed by bystanders. He is