The colloid:crystalloid confrontation

Journals Reviewed: Selected recent articles from Critical Care Medicine, British Medical Bulletin
Abstracted by: Dr SG Smith; MB BCh (Registrar, University of the Witwatersrand)

Summary of abstracts

Is there any hope in finding an answer to the crystalloid:colloid debate? Here is a concise summary of the controversy, with pointers to recent fairly substantial reviews. (You may also wish to briefly browse our editorial comment)

Crystalloid versus Colloid - the controversy rages

The debate over whether to use crystalloids or colloids (and which specific fluid in each group to use) continues to be a controversial point in trauma, surgical and acutely ill patients.

The shift towards fluid as opposed to blood products is based on the risks of blood product infusion and the fact that acute anaemias are tolerated better than tissue perfusion deficits.

The arguments for the use of one type of fluid over another are based on, amongst others: cost, risk of anaphylactoid reactions, effects on coagulation and post- operative organ function. The major point of contention however is based on the propensity of either fluid to cause pulmonary oedema.

What both groups do agree on is that hypovolaemia leads to cardiovascular decompensation, decreased cellular perfusion and oxygen delivery (DO2), increasing oxygen debt, lactic acidosis and ultimately, cell death. Fluid provision will decrease the amount of irreversibly damaged cells BUT this does not necessarily mean an improvement in DO2. Agreement is also reached that fluids should be warmed to prevent the further complications of hypothermia.

Some factors responsible for non-resolution of the crystalloid –colloid controversy are based on none of the studies having:

  • clearly defined criteria when to resuscitate
  • clearly defined physiological end-points of resuscitation
  • clear criteria to define pulmonary oedema
  • different responses and end-points in the various subgroups of patients i.e. trauma vs. critically-ill vs. the medical-surgical patient
  • randomising difficulty in the resuscitation setting
  • treatment crossover
  • effect of co-interventions

The number of studies done in this area and the failure to reach consensus indicates the poor nature of these studies.

1. Colloid vs Crystalloids in Fluid Resuscitation: A systematic Review

This review of all randomised clinical trials from 1966 to1996 looks at outcomes of pulmonary oedema, mortality and length of hospital stay in all those studies looking at crystalloid vs. colloid resuscitation. A total of 16 studies met the inclusion criteria and the overall results indicated no difference in any of the observed outcomes in the 2 clinical groups. There was however a statistically significant improvement in mortality in the crystalloid -treated subset of trauma patients.

There are numerous problems with this study review. Is mortality an appropriate outcome to be looking at in these trials? The methodology of the reviewed studies were rated on 8 variables from 0-2; i.e. A total score of 16. The range was 4-12, indicating poor study methodology. The total number of patients from the pooled data was insufficient and this affects the power of the study.

Crit.Care Med. 1999 27 (1) 200-210
Article type: Review
Author: Choi et al, McMaster Univ., Hamilton, Ontario, Canada

2. Distribution of Normal Saline and 5% Albumin infusions in septic patients

One of the major points of contention with type of fluid is the development of pulmonary oedema as a morbid complication. A study looking at distribution of colloid vs. crystalloid may give insight into the distribution of each fluid when infused.

This prospective randomised trial looks at distribution of NS and 5% albumin in 18 critically ill patients. Plasma and extracellular fluid volume were measured using I-131 albumin and S-35 sodium sulphate radio- isotope dilution techniques. Then interstitial fluid volume ISFV = ECFV-PV. From plasma counts, extrapolated zero time plasma counts for each isotope was calculated using linear regression. It was found that NS distributed in a ratio of 1:3 through PV and ISFV in the same volume as was infused. Albumin increased ECFV by twice the volume infused PV and ISFV increasing in equal proportions. The likely explanation is that leaky capillaries allow colloid to leak into the ISFV, where the colloid then draws in water from the intracellular compartment.

Problems with the study include:

  • A very small sample number;
  • End-points of resuscitation were determined by the patient's clinician, not necessarily the same in each case
  • In the treatment groups, there was no data pointing to systemic perfusion defects and PV was in the normal range prior to infusion. i.e. was transfusion indicated?

However, this study still gives useful insights into the distribution of infused fluids, and the potential for development of interstitial and pulmonary oedema.

Although not a primary objective of the study, it was found that none of the fluid infused increased DO2, but this may be due to the study lacking statistical power due to the small number of study patients.

Crit.Care Med 1999 27 (1) 46-50
Article type: Clinical research
Author: Ernest et al; St Paul's Hospital and University of British Columbia, Vancouver, B.C., Canada

3. Fluid Replacement

This review highlights the fact that in the initial stages of trauma, it is not the type of fluid, but rather an amount appropriate to restore tissue perfusion that is important. As well as discussing the effects and side effects of the various crystalloids and colloids, other alternatives such as synthetic haemoglobin solutions are considered.

Due to the difficulty in assessing tissue perfusion, it is suggested that failure to supranormalise VO2I and DO2I is a predictor of the development of multi-organ failure, and that these should be considered goals of therapy. The other suggestion made is that in the later stages of critical illness, once `capillary leakiness` has occurred, the low molecular weight (MW) colloids e.g. Gelatins have no role and high MW colloids should be used in this setting. The review also looks at prehospital fluid management and the principles of permissive hypotention.

It is the lack of any proper studies in this field, which has led to the perpetuation of, and non-resolution of this specific controversy. At this time the, ideal fluid to use in resuscitation is most likely a combination of crystalloid and colloid, but, until an adequate study is produced, this is based not so much on evidence as on a middle road between the two sets of protagonists.

British Medical Bulletin 1999; 55 (4) 821-843
Article type: Review
Author: Jerry Nolan, Dept. Anaesthesia, Royal United Hospital, Bath, UK

Editorial pointers

Enthusiastic antagonists of colloid therapy for volume replacement have probably already perused the following two articles:

Human albumin administration in critically ill patients: systematic review of randomised controlled trials.
BMJ 1998 Jul 25;317 (7153):235-40
(Cochrane Injuries Group Albumin Reviewers)

Fluid resuscitation with colloid or crystalloid solutions in critically ill patients: a systematic review of randomised trials. BMJ 1998 Mar 28;316 (7136):961-4

The former concluded that "These data suggest that for every 17 critically ill patients treated with albumin there is one additional death", and the latter that "Resuscitation with colloids was associated with an increased absolute risk of mortality of 4% (95% confidence interval 0% to 8%), or four extra deaths for every 100 patients resuscitated"

Needless to say, as in any religious argument, these conclusions have been greeted with screams of protest or enthusiasm by the various sects.

Come on chaps. Let's be honest. Most of the studies haven't been that hot. There is also no convincing reason to lump 'all colloids' or 'all crystalloids' into a single group - for example, it can be argued fairly convincingly that anyone who performs large-volume resuscitation with normal saline instead of Ringer's lactate (RL) is indescribably wicked. Is it then really fair to compare (say) RL resuscitation with a mixture of what is effectively normal saline and added colloid? In addition, a colloid mixture with an average molecular weight of 70 kDa is hardly comparable with one of 200 kDa, especially in the face of 'leaky' capillaries!

Before we argue, we need to define our resuscitative goals, desirable properties of resuscitative fluids apart from whether they contain colloid or not, and then create large prospective randomised blinded trials where resuscitative end-points are met with either crystalloids, or similar solutions containing added colloid of an appropriate molecular weight. At present honest intensivists and trauma specialists are really not sure whether they should enthusiastically embrace colloids or expel them from their units (although we seem to be tending towards the latter)!


Old scopes logos