Current issues of ACP Journal Club are published in Annals of Internal Medicine


Review: Colloid and crystalloid solutions for fluid resuscitation are associated with similar mortality rates in critically ill patients

ACP J Club. 1998 Sep-Oct;129:32. doi:10.7326/ACPJC-1998-129-2-032

Source Citation

Schierhout G, Roberts I. Fluid resuscitation with colloid or crystalloid solutions in critically ill patients: a systematic review of randomised trials. BMJ. 1998 Mar 28;316:961-4.



In critically ill patients, does resuscitation with colloid or crystalloid solutions for volume replacement reduce mortality?

Data sources

Studies were identified by using the Cochrane Controlled Trials Register, MEDLINE, EMBASE/Excerpta Medica, and BIDS Index to Scientific and Technical Proceedings. 29 international journals, proceedings of several international meetings, and bibliographies of review articles and selected trials were hand searched, and authors were contacted.

Study selection

Randomized controlled trials of critically ill patients were selected if the patients required fluid resuscitation and the intervention included colloid and crystalloid fluid resuscitation protocols. Enrolled patients had sepsis, burns, trauma, or major surgery. Exclusion criteria were studies of neonates, confounded studies, crossover trials, or trials of preloading in preparation for surgery or coronary artery bypass grafting.

Data extraction

All-cause mortality was the main outcome. Data were also extracted on study quality, type of injury, treatment, length of follow-up, and other outcomes. Authors were contacted to obtain unreported data on outcomes.

Main results

48 trials were identified, and 37 met the inclusion criteria. The solutions varied across the trials—for example, by concentrations and composition. Data from 19 trials (1315 patients) showed no mortality differences between resuscitation with colloid solutions and with crystalloid solutions (relative risk [RR] 1.19, 95% CI 0.98 to 1.45, P = 0.3) (Table). Results were similar when only trials with adequate allocation concealment were pooled (RR 1.29, 0.94 to 1.77). Subgroup analysis did not show any differences based on type of injury (trauma [RR 1.30, CI 0.95 to 1.77 for 6 trials], surgery [RR 0.55, CI 0.18 to 1.65 for 7 trials], burns [RR 1.21, CI 0.88 to 1.66 for 4 trials], or other injuries [RR 1.08, CI 0.73 to 1.61 for 2 trials]).


In critically ill patients, mortality rates are similar with resuscitation using colloid or crystalloid solutions for volume replacement.

Source of funding: NHS R&D Programme: Mother and Child Health.

For correspondence: Dr. G. Schierhout, Department of Epidemiology and Public Health, Institute of Child Health, University College London Medical School, London WC1N 1EH, England, UK. FAX 44-171-813-0242.

Table. Mortality associated with fluid resuscitation using colloid or crystalloid solutions in critically ill patients (maximum follow-up 2.5 wk)*

Weighted event rate RRI (95% CI) NNT
Colloid Crystalloid
19.4% 18.7% 1.7% (-2 to 55) Not significant

*Abbreviations defined in Glossary; RRI, NNT, and CI calculated from data in article.


The meta-analysis by Schierhout and Roberts addresses the long-standing debate about colloids compared with crystalloids for fluid resuscitation. Strong features of the review include the comprehensive search strategy, evaluation of publication bias, critical appraisal of the primary studies, and exploration of heterogeneity of results using methodologic quality and patient characteristics. Different colloids (blood products, synthetic starches, dextrans, and gelatins) were compared with various crystalloids (isotonic and hypertonic solutions).

Similar to previous meta-analyses (1, 2), this study detected no difference in mortality, despite pooling data from > 1300 patients from 19 trials. These findings do not suggest that fluid choice for resuscitation is unimportant. However, it is plausible that such factors as comorbid conditions, acute illness severity, and iatrogenic complications have a greater influence on mortality than does the choice of resuscitative fluid. Selection of crystalloids or colloids is often based on physiologic rationale considering outcomes, such as oncotic pressure and perfusion. The extent to which these end points correlate directly with major morbid events or mortality is uncertain; trial results, individually and in aggregate, are inconclusive. Other issues, such as availability, cost, and habit, also determine the choice of fluid for resuscitation.

This meta-analysis provides a high-quality updated synthesis of the randomized trials of crystalloids compared with colloids in critically ill patients. If survival is truly influenced by fluid selection, however, much larger trials will be necessary to detect differences. Meanwhile, foci of current research in this field include investigation of specific populations (e.g., critically ill parturients), fluid management protocols, and resuscitation monitoring strategies with pulmonary artery catheters.

Peter Choi, MD
Deborah Cook, MDMcMaster UniversityHamilton, Ontario, Canada


1. Velanovich V. Crystalloid versus colloid fluid resuscitation: a meta-analysis of mortality. Surgery. 1989;105:65-71.

2. Wade CE, Kramer GC, Grady JJ, Fabian TC, Younes RN. Efficacy of hypertonic 7.5% saline and 6% dextran-70 in treating trauma: a meta-analysis of controlled clinical studies. Surgery. 1997;122:609-16.