Colloids versus crystalloids for fluid resuscitation in critically ill patients

  • Review
  • Intervention

Authors


Abstract

Background

Colloid solutions are widely used in fluid resuscitation of critically ill patients. There are several choices of colloid and there is ongoing debate about the relative effectiveness of colloids compared to crystalloid fluids.

Objectives

To assess the effects on mortality of colloids compared to crystalloids for fluid resuscitation in critically ill patients.

Search strategy

We searched the Injuries Group specialised register, Cochrane Controlled Trials Register, MEDLINE, EMBASE and BIDS Index to Scientific and Technical Proceedings and checked the reference lists of trials and review articles.

Selection criteria

All randomised and quasi-random trials of colloids compared to crystalloids, in patients requiring volume replacement. Cross-over trials and trials in pregnant women and neonates were excluded.

Data collection and analysis

Two reviewers independently extracted data and rated quality of allocation concealment. Trials with a 'double-intervention' such as those, which compared colloid in hypertonic crystalloid to isotonic crystalloid, were analysed separately. The analysis was stratified according to colloid type and quality of allocation concealment.

Main results

Colloids compared to crystalloids:

Albumin or plasma protein fraction: Eighteen trials reported data on mortality, including a total of 641 patients. The pooled relative risk from these trials was 1.52 (95% confidence interval 1.08 to 2.13). The risk of death in the albumin treated group was 6% higher than in the crystalloid treated group (1% to 11%). When the trial with poor quality allocation concealment was excluded the pooled relative risk was 1.34 (0.95 to 1.89).

Hydroxyethylstarch: Seven trials compared hydroxyethylstarch with crystalloids including a total of 197 randomised participants. The pooled relative risk was 1.16 (0.68 to 1.96).

Modified gelatin: Four trials compared modified gelatin with crystalloid including a total of 95 randomised participants. The pooled relative risk was 0.50 (0.08 to 3.03).

Dextran: Eight trials compared dextran with a crystalloid including a total of 668 randomised participants. The pooled relative risk was 1.24 (0.94 to 1.65).

Colloids in hypertonic crystalloid compared to isotonic crystalloid:

Eight trials compared dextran in hypertonic crystalloid with isotonic crystalloid, including 1283 randomised participants. The pooled relative risk was 0.88 (0.74 to 1.05).

Reviewers' conclusions

There is no evidence from randomised controlled trials that resuscitation with colloids reduces the risk of death compared to crystalloids in patients with trauma, burns and following surgery. As colloids are not associated with an improvement in survival, and as they are more expensive than crystalloids, it is hard to see how their continued use in these patient types can be justified outside the context of randomised controlled trials.

Plain language summary

Synopsis

No evidence that colloids are more effective than crystalloids for increasing blood volume and reducing mortality in people who are critically ill or injured

Trauma, burns or surgery can cause people to lose large amounts of blood. Fluid replacement, giving fluids intravenously (into a vein) to replace lost blood, is used to try to maintain blood pressure and reduce the risk of dying. Blood products, non-blood products or combinations are used, including colloid or crystalloid solutions. Colloids are increasingly used but they are more expensive than crystalloids. The review of trials found no evidence that colloids reduce the risk of dying compared with crystalloids.

Background

Fluid resuscitation for hypovolaemia is a mainstay of the medical management of critically ill patients, whether as a result of trauma, burns, major surgery or sepsis. Although recent studies (Bickell 1994) have suggested that the timing of volume replacement deserves careful consideration, when it comes to selecting the resuscitation fluid clinicians are faced with a range of options. At one level the choice is between a colloid or crystalloid solution. Colloids are widely used, having been recommended in a number of resuscitation guidelines and intensive care management algorithms (Vermeulen 1995; Armstrong 1994). The US Hospital Consortium Guidelines recommend that colloids are used in haemorrhagic shock prior to the availability of blood products, and in non-haemorrhagic shock following an initial crystalloid infusion. A 1995 survey of US academic health centres, however, found that the use of colloids far exceeded even the Hospital Consortium recommendations (Yim 1995). Surveys of burn care in the US (Fakhry 1995) and in Australia (Victorian Drug.1991) found that the use of colloids for resuscitation varied without a set pattern. The choice of fluid has considerable cost implications. Volume replacement with colloids is considerably more expensive than with crystalloids. Clinical studies have shown that colloids and crystalloids have different effects on a range of important physiological parameters. Because of these differences, all-cause mortality is arguably the most clinically relevant outcome measure in randomised trials comparing the two fluid types. Although there have been previous meta-analyses of mortality in randomised trials comparing colloids and crystalloids (Velanovich 1989, Bisonni 1991), neither of these satisfy the criteria that have been proposed for scientific overviews (Oxman 1994), and they predate most of the trials that have been conducted using synthetic colloids, and hypertonic crystalloid solutions. The purpose of this review was to identify and synthesize all available unconfounded evidence of the effect on mortality in critically ill patients of colloids compared to crystalloids for volume replacement.

Objectives

To determine the effects on mortality of using colloids compared to crystalloids during fluid resuscitation in critically ill patients.

Criteria for considering studies for this review

Types of studies

Controlled trials in which participants were randomised to treatment groups (colloid or control) on the basis of random or quasi-random allocation. As the comparison between fluid type was in terms of effects on mortality, randomised cross-over trials were excluded.

Types of participants

Critically ill patients (excluding neonates) who required volume replacement. Types of participants included were those who were critically ill as a result of trauma, burns, were undergoing surgery, or had other critical conditions such as complications of sepsis.

Preoperative elective surgical patients were excluded.

Types of intervention

The colloids considered were Dextran 70, hydroxyethyl starches, modified gelatins, albumin or plasma protein fraction.

There is overlap between albumin given for volume replacement and albumin given as a nutritional supplement, and many patients with a critical illness have low serum albumin. Where the trial was of total parenteral nutrition with or without albumin, it was excluded. We included trials where the albumin was given as part of volume replacement guided by colloid osmotic pressure or albumin levels.

The control group received crystalloid (isotonic or hypertonic) for fluid replacement. Trials where both groups received blood were included.

Trials of fluids used for other purposes were excluded. For example, trials of pre-loading in preparation for elective surgery, and trials in patients undergoing fluid loading before cardiopulmonary bypass, were excluded.

Types of outcome measures

The principal outcome measure was mortality from all causes assessed at the end of the follow up period scheduled for each trial.

Search strategy for identification of studies

See: Unavailable search strategy

MEDLINE, latest search, March 1999

#1 colloid*
#2 albumin*
#3 ppf
#4 dextran
#5 gelatin*
#6 gentran*
#7 haemaccel*
#8 hemaccel*
#9 pentastarch
#10 pentaspan
#11 hetastarch
#12 #1 or #2 or #3 or #4 or #5 or #6 or #7 or #8 or #9 or #10 or #11
#13 crystalloid*
#14 ringer*
#15 hartman*
#16 sodium*
#17 potassium*
#18 salin*
#19 #13 or #14 or #15 or #16 or #17 or #18
#20 fluid
#21 therapy
#22 fluid near2 therapy
#23 volume
#24 restor*
#25 volume near2 restor*
#26 fluid
#27 resuscit*
#28 restor*
#29 therap*
#30 fluid near2 (resuscit* or restor* or therap*)
#31 plasma
#32 substit*
#33 fluid
#34 volume
#35 substit*
#36 replace*
#37 (plasma near2 substit*) or ((fluid or volume) near2 (substit* or replace*))
#38 #22 or #25 or #30 or #37
#39 #12 and #19 and #38
#40 controlled
#41 clinical
#42 trial*
#43 controlled clinical trial*
#44 randomi*
#45 controlled
#46 trial*
#47 randomi* controlled trial*
#48 explode research design/ all subheadings
#49 double
#50 blind
#51 double blind
#52 meta
#53 analysis
#54 metaanalysis
#55 meta analysis or metaanalysis
#56 clinical trial in pt
#57 #43 or #47 or #48 or #51 or #55 or #56
#58 #39 and #57

The reference lists of all identified trials and review articles were checked, and we contacted the trialists, to ask if any studies had been missed.

Searches were performed in November 1998, using variations on the above strategy, covering the following databases and time periods:

EMBASE 1998
PubMed 8/1998 - 12/1998
Science Citation Index 1990-1998
Cochrane Controlled Trials Register all years

In addition searching done in November 1999, of Medline and The Cochrane Controlled Trials Register, for the update of the Injuries Group review on Human albumin, found new papers which were relevant for this review.

Full details of the search strategies used can be obtained from the Injuries Group Trials Search Co-ordinator.

Methods of the review

For the first version of this review, searching and decisions about the inclusion of studies were by Schierhout and Roberts. The review has been updated in the light of post-publication comments in the British Medical Journal, and the search has been updated.

For the updated review, Alderson and Roberts examined trials for inclusion or exclusion, reaching agreement by discussion. Alderson and Bunn rechecked all the extracted data from the original review and any new studies.

Allocation concealment was scored as described by Schulz (Schulz 1995). In particular, the presence of solutions in identical containers was only taken to mean adequate concealment if the fluid containers were used sequentially.

Information on blinding and loss to follow up was collected but not scored.

As a result of comments on the previous version of this review, trials were stratified by type of fluid rather than type of original injury.

Relative risks and 95% confidence intervals were calculated for each study using a fixed effects model. Each comparison was then inspected visually for evidence of heterogeneity and a chi-squared test performed. If there was no evidence of heterogeneity (visually or with a p value < 0.1) the trials were pooled within each type of fluid, but not combined between type of fluid.

Trials with allocation concealment judged as inadequate were then excluded and the calculations repeated.

Description of studies

There were 48 trials meeting the inclusion criteria for study design, participants and interventions. We were able to obtain data on deaths for 38 of these. Details of the remaining ten trials are also reported in the Table of Included Studies for completeness.

Reasons for exclusion of trials were the use of a cross-over design, testing a resuscitation algorithm, giving the control group oral fluids, the intervention being directed to the maintenance of serum albumin levels, for haemodilution, for fluid loading and for the reduction of intracranial pressure (see Table of Excluded Studies).

Of the 38 randomised controlled trials with data on deaths, the quality of allocation concealment was adequate in four trials and unclear in most of the others.

There were 36 comparisons of colloids and crystalloids (add-on colloid), nine comparisons of colloid in hypertonic crystalloid with isotonic crystalloid, and three comparisons of colloid with hypertonic crystalloid (see Table of Included Studies).

Methodological quality

In general, the design of studies was not well reported. This is reflected in the number of unclear scores given for allocation concealment. We also collected information on blinding and loss to follow up. Blinding was not well reported and loss to follow up was generally small. The characteristics for each trial are listed in the Table of Included Studies.

Results

Colloids compared to crystalloids:

Albumin or plasma protein fraction: Eighteen trials reported data on mortality, including a total of 641 patients. The pooled relative risk from these trials was 1.52 (95% confidence interval 1.08 to 2.13). The risk of death in the albumin treated group was 6% higher than in the crystalloid treated group (1% to 11%). When the trial with poor quality allocation concealment was excluded the pooled relative risk was 1.34 (0.95 to 1.89).

Hydroxyethylstarch: Seven trials compared hydroxyethylstarch with crystalloids including a total of 197 randomised participants. The pooled relative risk was 1.16 (0.68 to 1.96).

Modified gelatin: Four trials compared modified gelatin with crystalloid including a total of 95 randomised participants. The pooled relative risk was 0.50 (0.08 to 3.03).

Dextran: Eight trials compared dextran with a crystalloid including a total of 668 randomised participants. The pooled relative risk was 1.24 (0.94 to 1.65).

Colloids in hypertonic crystalloid compared to isotonic crystalloid:

One trial compared albumin and hypertonic saline with isotonic crystalloid. Its relative risk of death was 0.50 (0.06 to 4.33).

Eight trials compared dextran in hypertonic crystalloid with isotonic crystalloid, including 1283 randomised participants. The pooled relative risk was 0.88 (0.74 to 1.05).

Colloids in isotonic crystalloid compared to hypertonic crystalloid:

Three trials compared colloids in isotonic crystalloid with hypertonic crystalloid. In two of these, where the colloid was either gelatin or starch, there were no deaths in either group. In the remaining trial, with 38 participants, there was a relative risk of death of 7 (0.39 to 126.93) for use of colloid, based on three deaths in the treatment group and none in the control group.

Discussion

This systematic review synthesises the evidence from randomised controlled trials comparing colloid and crystalloid fluid resuscitation across a wide variety of clinical conditions. The review has been updated and extensively revised to take into account the comments made since it was first published. In particular, several commentators pointed out that it is inappropriate to combine effect estimates from studies of different colloids. For example, it was argued that large molecular weight colloids such as hydroxyethyl starch may be better retained in the vascular compartment than albumin and gelatins, and would therefore be more likely to show a favourable effect on mortality (Gosling 1998). In response to these concerns, the review has been stratified by type of colloid. However, the pooled relative risks fail to show a mortality benefit for resuscitation with any type of colloid.

There was a trend towards a favourable effect on mortality for colloids in hypertonic crystalloid compared to isotonic crystalloids. Nevertheless the results are compatible with the play of chance.

Common to all meta-analyses, this systematic review may have included studies whose interventions and patient characteristics are sufficiently incomparable that the calculation of a summary effect measure may be questioned. The resuscitation regimen differed between trials. Some trials randomised participants to an initial quantity of colloid or crystalloid, and then proceeded with some form of standard resuscitation for all participants. Other trials resuscitated with the allocated fluid to pre-determined end-points, either resuscitation end-points, or in the case of trauma, until corrective surgery. In addition, the type of colloid or crystalloid, the concentration, and the protocol to determine the quantity of fluid varied. Despite these differences, all participants were in need of volume replacement, and we believe that this variation in the intervention would have an impact on the size of the effect, rather than on its direction.

A systematic review recently published in a paper journal addressed a similar question to this review (Choi 1999). As would be expected there are differences in the studies included in the two reviews. The results were not presented stratified by fluid type, so direct comparison of the results of these reviews is difficult. Nevertheless, the overall result of the Choi review was that there was no clear benefit of colloids over crystalloids.

Reviewers' conclusions

Implications for practice

There is no evidence from randomised controlled trials that resuscitation with colloids reduces the risk of death in patients with trauma, burns and following surgery. As colloids are not associated with an improvement in survival, and further, colloids are considerably more expensive than crystalloids, it is hard to see how their continued use outside the context of randomised controlled trials in subsets of patients of particular concern, can be justified.

Implications for research

Future trials may need to concentrate on specific subgroups of patients to identify people who may benefit from colloids rather than crystalloids.

Potential conflict of interest

None known.

Acknowledgements

We would like to acknowledge the Intensive Care National Audit and Research Network in London, for assistance with identification of trials for this review, and help from Reinhard Wentz in designing the search strategies for the updated review.

Characteristics of included studies

Study Boldt 1986
MethodsRandomised controlled trial, using sealed opaque envelopes.
Information on allocation concealment was obtained on contact with the authors.
Blinding and loss to follow up not mentioned.
Participants55 patients undergoing elective aorto-coronary bypass surgery.
Exclusion criteria were ejection fraction < 50% and LVEDP > 15 mmHg.
Interventions1) 300ml 20% human albumin solution (n=15)
2) 500ml 3% hydroxyethylstarch (n=13)
3) 500ml 3.5% gelatine (n= 14)
4) no colloid (n=13)
OutcomesHaemodynamic variables were measured.
Deaths not reported.
NotesFollow up until discharge from intensive care.
Allocation concealmentB
Study Boldt 1993
MethodsRandomised controlled trial.
Allocation concealment by sealed opaque envelopes (information from author).
Blinding and loss to follow up not mentioned.
Participants75 men undergoing elective aortocoronary bypass grafting, who had a pulmonary capillary wedge pressure of less than 5 mm Hg after induction of anaesthesia.
Interventions1) 5% albumin (n=15)
2) 6% HES, mean molecular weight 450,000 (n=15)
3) 6% HES, mean molecular weight 200,000 (n=15)
4) 3.5% gelatin (n=15)
5) No colloid (n=15)
Fluid used through operation and on intensive care post-op
OutcomesDeaths not reported, author confirmed there were no deaths.
NotesFollow up to 1 day.
Allocation concealmentB
Study Boutros 1979
MethodsRandomised controlled trial ("randomly divided").
Method of allocation concealment not described.
Blinding not mentioned.
No loss to follow up.
Participants24 people undergoing major operative procedures on the abdominal aorta.
Interventions1) Albumin in 5% dextrose. (n=7)
2) 5% dextrose and Ringer's lactate. (n=8)
3) 5% dextrose in 0.45% saline. (n=9)
Allocated fluids were used on admission to ICU, following surgery, guided by PAWP. Whole blood also given if clinically needed.
OutcomesDeaths reported.
NotesFollow up to discharge from hospital.
Allocation concealmentB
Study Bowser-Wallace 1986
MethodsQuasi-randomised controlled trial (allocation by alternation).
Blinding not mentioned.
No loss to follow up.
ParticipantsAdmitted for burns of 30% or more.
Age range 5 months to 21 years.
Excluded if already given more than half calculated daily requirement before reaching hospital.
Interventions1) 2ml/kg/%burn Ringer's lactate over 24 hrs, then 0.5ml plasmanate/kg/%burn over 24 hrs plus 5% dextrose. n=19
2) 2ml/kg/%burn hypertonic lactated saline over 24 hrs, then 0.6ml/kg/%burn hypertonic lactated saline over 24 hrs plus oral Haldane's solution. (n=19)
IV fluids stopped at 48 hrs.
OutcomesDeaths reported.
Fluid and electrolytes given, weight, haematocrit.
NotesFollow up to 5 days.
Allocation concealmentC
Study Chavez-Negrete 1991
MethodsRandomised controlled trial (allocation by "random numbers").
Blinding not mentioned.
No loss to follow up.
ParticipantsAdults admitted to an emergency room with acute gastrointestinal haemorrhage, systolic blood pressure 90 mmHg or less for up to 1 hr and normal electrocardiograph.
Excluded if pregnant or had renal, cardiac or neurological disease.
Interventions1) Initial infusion of 250ml 7.5% saline/6% Dextran 60 given IV (16 patients) or intraosseous (10 patients).
2) Initial IV infusion of 250ml Ringer's lactate. (n=23)
Resuscitation continued with red cells, 0.9% saline and Dextran 40 according to clinical judgement.
OutcomesDeath.
Haemodynamic variables.
NotesFollow up to 24 hours.
Allocation concealmentB
Study Dawidson 1991
MethodsRandomised controlled trial (allocation by drawing a card from a deck).
Blinding not mentioned.
No loss to follow up.
ParticipantsAdults undergoing elective abdominal aortic surgery.
No exclusions mentioned.
Interventions1) IV 3% Dextran 70 in Ringer's lactate. (n=10)
2) IV Ringer's lactate. (n=10)
Fluid used during and for 24 hrs after operation, guided by haemodynamic variables.
OutcomesDeath.
Volume transfused, weight change, haemodynamic variables.
NotesFollow up to discharge from hospital.
Allocation concealmentC
Study Eleftheriadis 1995
MethodsPatients "randomizedly distributed".
Blinding not mentioned.
Unable to assess loss to follow up.
ParticipantsParticipants were undergoing coronary artery bypass surgery.
Interventions1) 6% hydroxyethylstarch
2) 3.5% gelatine
3) Ringer's lactate
Allocated fluid was used in the post-operative period only guided by mean arterial pressure.
OutcomesDeaths were not reported.
Haemodynamic variables.
NotesFollow up period unspecified.
Allocation concealmentB
Study Ernest 1999
MethodsRandomised controlled trial, allocation concealment not described.
No blinding.
No loss to follow up mentioned.
ParticipantsPatients with a clinical diagnosis of sepsis.
Interventions1) 5% albumin (n=9)
2) 0.9% saline (n=9)
Volume of infusion guided by PAWP
OutcomesHaemodynamic variables and volume measurements.
Deaths not reported.
NotesFollow up to immediately after infusion.
Allocation concealmentB
Study Evans 1996
MethodsQuasi-randomised trial (allocation by day of the week).
Blinding not mentioned.
No loss to follow up.
ParticipantsAged 16 or more, admitted with trauma to an emergency centre within 2 hours after injury, only crystalloid as a pre-hospital infusion.
Excluded if had underlying illness likely to affect clotting.
Interventions1) IV Haemaccel. (n=11)
2) IV Ringer's lactate. (n=14)
Fluid was used until vital signs were stable.
OutcomesDeaths from author.
Clotting variables.
NotesFollow up period unspecified.
Allocation concealmentC
Study Gallagher 1985
MethodsRandomised controlled trial. Method of allocation concealment not described. Author contacted - allocation concealment by computerised system - patient details were entered before treatment assignment was revealed.
Blinding not mentioned.
No loss to follow up.
ParticipantsPatients after coronary artery bypass graft surgery.
Exclusions: patients with significant left main coronary artery stenosis, poor left ventricular function or poor pulmonary function.
Interventions1) IV 5% albumin (n=5)
2) IV 6% hydroxyethylstarch (n=5)
3) IV Ringer's lactate (n=5)
Fluid used from admission to intensive care post op, guided by PAWP. RBC given if needed.
Five patients received 5% albumin. Five patients received lactated Ringers.
OutcomesDeaths were not reported. Author contacted and confirmed that there were no deaths in any group.
Haemodynamic data.
NotesFollow up to 1 day.
Allocation concealmentA
Study Goodwin 1983
MethodsRandomised controlled trial - assigned by 'random numbers table'.
Method of allocation concealment unclear.
Blinding not mentioned.
No loss to follow up.
Participants79 previously healthy young adults admitted with burns.
No exclusion criteria reported.
Interventions1) 2.5% albumin in Ringer's lactate (n=40)
2) Ringer's lactate (n=39)
Fluids on day 1 guided by haemodynamic variable. On day 2, given at 0.3-0.5ml/kg/%burn, then 5% dextrose.
OutcomesDeaths reported.
Lung water in some.
Infections.
NotesFollow up to discharge from hospital.
Allocation concealmentB
Study Grundmann 1982
MethodsRandomised controlled trial.
Method of allocation concealment unclear.
Blinding not mentioned.
No loss to follow up.
Participants20 people undergoing partial gastrectomy.
The average age was 50 years (range 19-84).
No exclusion criteria reported.
Interventions1) Colloid group received human albumin solution. (n=14)
2) Details of crystalloid were not reported. (n=6)
Allocated fluid was continued for 4 days after operation.
OutcomesDeaths reported.
Volumes of fluid given.
Haemodynamic variables.
NotesFollow up to discharge from hospital.
Allocation concealmentB
Study Hall 1978
MethodsQuasi-randomised controlled trial (participants were stratified by age, extent of burn and aetiology, and then allocated by alternation).
Blinding not mentioned.
No loss to follow up.
ParticipantsBurns covering more than 10% of the body surface (for children), and more than 15% of the body surface (for adults).
No exclusions mentioned.
Interventions1) 120ml/%burn IV 6% Dextran 70 in 0.9% saline over 48 hrs plus oral water or IV 5% dextrose for 'metabolic requirements'. (n=86)
2) 4ml/kg/%burn IV Ringer's lactate over 24 hrs, then 10% of initial body weight of fluid over 24 hrs plus oral water. (n=86)
OutcomesDeath.
Fluid given, haemodynamic variables.
NotesFollow up to discharge from hospital.
Allocation concealmentC
Study Hartmann 1993
MethodsRandomised controlled trial (method of allocation unclear).
Blinding not mentioned.
No loss to follow up.
ParticipantsAdults undergoing major abdominal surgery.
Exclusions: cardiorespiratory dysfunction, uraemia, diabetes, taking steroids, anticoagulants or diuretics.
Interventions1) IV Dextran 70 in saline (concentration not given) with 2.5% dextrose. (n=15)
2) IV saline (concentration not given) with 2.5% dextrose. (n=14)
Both groups given red cells, plasma, Dextran 70 and crystalloids during the operation as decided by the clinician. Post operative fluids according to the trial group guided by tissue oxygen tension to the end of resuscitation.
OutcomesDeath not reported.
Fluid given, haemodynamic variables.
NotesFollow up to 7 days.
Allocation concealmentB
Study Jelenko 1978
MethodsRandomised controlled trial, method of allocation concealment unclear.
Blinding not mentioned.
No loss to follow up.
Participants19 people with burns covering more than 20% of body surface.
Interventions1) 12.5% albumin in hypertonic saline (240MeQ Na+, 120 MeQ Chloride, 120 MeQ lactate). (n=7)
2) Hypertonic saline (240MeQ Na+, 120 MeQ Chloride, 120 MeQ lactate). (n=5)
3) Ringer's lactate (n=7)
Allocated fluid was used, guided by haemodynamic variables, to the end of resuscitation.
OutcomesDeaths reported.
Haemodynamic variables.
NotesFollow up to end of resuscitation.
Allocation concealmentB
Study Karanko 1987
MethodsRandomised controlled trial. Description of allocation procedure unclear.
Blinding not mentioned.
No loss to follow up.
Participants32 adult men scheduled for coronary artery bypass surgery.
Exclusions: left ventricular ejection fraction under 40%, abnormal lung function.
Interventions1) Colloid group received 6% dextran 70. (n=14)
2) Ringer's lactate. (n=18)
Allocated fluid was used to the end of resuscitation.
OutcomesDeaths reported.
Haemodynamic variables.
Lung water.
NotesFollow up 2 weeks.
Allocation concealmentB
Study Ley 1990
MethodsRandomised controlled trial.
Method of allocation concealment unclear.
Assessment of chest x-ray blinded.
No loss to follow up.
Participants21 people undergoing coronary artery bypass grafting or valve surgery.
Interventions1) 6% hetastarch up to 1.5L then 5% plasma protein fraction. (n=11)
2) 0.9% saline (n=10)
Allocated fluid was used for post-operative fluid resuscitation.
OutcomesDeaths were not reported.
Pulmonary and peripheral oedema.
Haemodynamic variables.
NotesFollow up to discharge.
Allocation concealmentB
Study Lowe 1977
MethodsRandomised controlled trial, allocation by sealed envelopes.
Blinding not mentioned.
No loss to follow up.
ParticipantsParticipants with serious trauma.
Interventions1) 25% albumin in Ringers lactate.
2) Ringers lactate.
Allocated fluid was used throughout the pre- and intra-operative period.
OutcomesDeaths reported.
NotesFollow up to 5 days post-operatively. Data on the 30 participants with chest injuries who were left out of the Lowe 1977 report, but included in Moss 1981, have been included in the meta-analysis.
Allocation concealmentB
Study Lucas 1978
MethodsRandomised controlled trial.
Randomisation was based on the last digit of each patient's case number.
Participants52 seriously injured patients.
Interventions1) Standard resuscitation regimen ('balanced electrolyte', blood, fresh frozen plasma) plus salt poor albumin, maximum 150g during surgery and 150g per day for the next 5 days. (n=27)
2) Standard resuscitation regimen as above. (n=25)
OutcomesDeaths reported in some patients.
NotesIn the final report of 94 randomised patients deaths were not reported. However, in this preliminary report of 52 injured patients deaths were reported.
Allocation concealmentC
Study Mattox 1991
MethodsQuasi-randomised, allocation by alternation.
Double-blind.
2 patients excluded from the analysis as code of fluid lost.
ParticipantsParticipants were pre-hospital trauma victims attended to by emergency personnel within an hour of injury, who had systolic blood pressure of 90m mmHg or less and were 16 years or older. 72% of participants had sustained penetrating trauma.
Interventions1) 250 mL Dextran-70 in 7.5% NaCl. (n=211)
2) 250 mL Ringers lactate, saline or plasmalyte. (n=211)
Allocated fluid was for initial pre-hospital resuscitation only.
OutcomesDeaths reported.
NotesFollow up to hospital discharge or transfer.
Allocation concealmentC
Study Mazher 1998
MethodsPatients 'randomized'.
Blinding of care givers by use of pharmacy prepared solutions.
No loss to follow up.
ParticipantsPatients undergoing elective coronary artery surgery.
Exclusions: age over 75, ejection fraction under 35%, creatinine over 135umol/L, ACE inhibitors.
Interventions1) 5mL/kg Polygeline. (n=10)
2) 5mL/kg 7.2% saline. (n=10)
Allocated fluid given post op over one hour. All patients subsequently receive Polygeline and red blood cells.
OutcomesHaemodynamic variables.
Death.
NotesFollow up to discharge from intensive care.
Allocation concealmentB
Study McNulty 1993
MethodsRandomised controlled trial. Method of allocation concealment not described.
Blinding not mentioned.
No loss to follow up.
ParticipantsPatients following elective cardiopulmonary bypass.
Interventions1) 5% albumin and cell-saved blood. (n=14)
2) Plasmalyte and cell-saved blood. (n=14)
Allocated fluid used as part of fluid volume replacement.
OutcomesDeaths not reported.
Study was designed to look at the effect of protein infusion on the accuracy of a haematocrit measuring device.
NotesLength of follow up unspecified.
Allocation concealmentB
Study Metildi 1984
MethodsRandomised controlled trial.
Blinding not mentioned.
No loss to follow up.
ParticipantsParticipants were admissions to an intensive care and a trauma unit with adult respiratory distress syndrome and established pulmonary failure. Included both trauma and non-trauma patients.
Interventions1) 5% salt poor albumin. (n=20)
2) Ringers lactate. (n=26)
Allocated fluid was used throughout resuscitation, and if an operation was required the allocated fluid was used for volume replacement before and during the operation.
OutcomesDeaths reported.
Haemodynamic variables.
NotesFollow up to discharge.
Allocation concealmentB
Study Modig 1983
MethodsQuasi-randomised controlled trial, allocation by admission date.
Blinding not mentioned.
No loss to follow up.
ParticipantsParticipants were trauma admissions to an emergency department with a systolic blood pressure of less than 70mmHg. Age range was 20-58 years.
Interventions1) Dextran-70 in Ringer's lactate. (n=12)
2) Ringer's lactate. (n=11)
Allocated fluids were given as the initial resuscitation fluid on admission to the emergency department, and continued as needed until after the 6th day when major reconstructive surgery was undertaken.
OutcomesDeaths reported.
Development of respiratory distress syndrome.
NotesFollow up to definitive reconstructive surgery.
Allocation concealmentC
Study Nagy 1993
MethodsRandomised controlled trial, contact with author showed it was an open label study.
Blinding not mentioned.
No loss to follow up.
ParticipantsParticipants were adult admissions to a trauma unit, with measurable systolic blood pressure less than 90 mmHg.
Interventions1) Pentastarch in 0.9% NaCl. (n=21)
2) Ringer's lactate. (n=20)
Allocated fluid was used throughout resuscitation with the exception that colloid patients recieved a maximum 4L of pentastarch, after which Ringers lactate was given.
OutcomesDeaths were not reported.
Haemodynamic variables.
NotesFollow up to discharge.
Allocation concealmentC
Study Nielsen 1985
MethodsRandomized controlled trial.
Method of allocation concealment not described.
Blinding not mentioned.
No loss to follow up.
Participants26 patients admitted for reconstructive surgery of the abdominal aorta.
Interventions1) Whole blood, crystalloid plus 80g albumin on the day of the operation, and 20g per day for the next 3 days. Albumin given as 100mL 20% human albumin solution. (n=13)
2) Whole blood and crystalloid, type not specified. (n=13)
OutcomesDeaths not reported.
Author when contacted confirmed that there were no deaths in either group.
NotesLength of follow up 4 days.
Allocation concealmentB
Study Pockaj 1994
MethodsRandomised controlled trial, allocation concealment unclear.
Blinding not mentioned.
Loss to follow up 18/54 in colloid group, 13/53 in saline group.
ParticipantsParticipants required fluid resuscitation as a result of vascular leak syndrome associated with Interleukin-2 therapy for metastatic cancer.
Interventions1) 250 mL boluses of 5% albumin in saline. (n=36 reported)
2) 250 mL boluses of 0.9% normal saline. (n=40 reported)
Boluses guided by haemodynamic variables. Both groups also received 0.45% saline with 10mmol/L KCl.
OutcomesDeaths.
Toxic effects of chemotherapy.
Haemodynamic variables.
Notes 
Allocation concealmentB
Study Prien 1990
MethodsRandomised controlled trial.
Blinding not mentioned.
No loss to follow up.
ParticipantsParticipants were undergoing modified Whipple's operation.
Interventions1) 10% hydroxyethyl starch in 0.9% saline plus plasma protein fraction if requirements > 20mL/kg. (n=6)
2) 20% human albumin solution. (n=6)
3) Ringer's lactate.
Allocated fluid was administered intra-operatively only.
OutcomesDeaths.
Intestinal oedema formation.
NotesFollow up period was unspecifed.
Allocation concealmentB
Study Rackow 1983
MethodsRandomised controlled trial, allocation concealment unclear.
Blinding not mentioned.
No loss to follow up.
ParticipantsParticipants were aged 54 to 97, and had any one of the following pre-determined indicators of shock: systolic blood pressure of 90 mmHg or less, a cardiac index of less than 2.2 L./min.m2, a serum arterial lactate greater than 18mg/dl and WP less than 15mmHg.
Interventions1) 6% hydroxoethyl starch. (n=9)
2) 5% albumin. (n=9)
3) 0.9% saline. (n=8)
Allocated fluid was given as needed until the end of resuscitation.
OutcomesDeaths reported.
Fluid balance.
NotesFollow up to discharge from hospital.
Allocation concealmentB
Study Rocha e Silva 1994
MethodsRandomised controlled trial.
ParticipantsParticipants were admissions to the emergency room, with a systolic blood pressure of 90 mmHg or less and were 16 years of age or older.
InterventionsColloid group received 6% dextran-70 in 7.5% NaCl; crystalloid group received Ringers lactate. Allocated fluid was used for the first intravenous infusion only.
OutcomesDeath was the main outcome measure, but the data are unpublished.
NotesFollow up to 30 days. By April 1994, 125 patients had been entered into the study.
Allocation concealmentB
Study Shah 1977
MethodsRandomised controlled trial. Allocation by sealed envelope.
Blinding not mentioned.
No loss to follow up.
ParticipantsPatients with severe, multiple trauma and a systolic blood pressure of less than 90mmHg. All patients were adults and both sexes were included.
Interventions1) 5% salt-poor albumin in Ringers lactate. (n=9)
2) Ringer's lactate. (n=11)
Volume infused guided by physiological parameters.
OutcomesDeath reported.
Haemodynamic variables.
NotesLength of follow up not stated.
Allocation concealmentB
Study Shires 1983
MethodsPatients 'assigned randomly'.
Blinding not mentioned.
No loss to follow up.
ParticipantsPeople undergoing aortic reconstruction surgery.
No exclusion criteria mentioned.
Interventions1) Plasmanate. (n=9)
2) Ringer's lactate. (n=9)
Allocated fluid used guided by haemodynamic variables until the first postoperative morning. All patients then received 0.45% saline.
OutcomesLung water.
Haemodynamic variables.
Death.
NotesFollow up to two days post op.
Allocation concealmentB
Study Sirieix 1999
MethodsPatients 'randomly assigned'. Blinding not described.
Two patients excluded after randomisation due to arrhythmias on giving the fluid (both in hypertonic saline group).
ParticipantsPatients undergoing mitral valve repair.
Exclusions: LVEF<0.4, systolic PAP>50mmHg, coagulation disorders, creatinine>150mmoL/L, electrolyte imbalance, diabetes, previous atrial fibrillation lasting > 1 year.
Interventions1) 250mL 7.2% hypertonic saline, 6%HES (n=8)
2) 250mL 7.2% hypertonic saline (n=10)
3) 250mL 6% HES (n=8)
Fluid given over 15mins, I hour after admission to post op intensive care
OutcomesHaemodynamic variables.
Deaths reported.
Side effects (2 had severe hypotension in group 2 and 1 in group 1; arrhythmias in 1 patient in group 1, 3 in group 2 and 1 in group 3)
NotesFollow up to discharge from hospital (all within 10 days)
Allocation concealmentB
Study Skillman 1975
MethodsRandomised controlled trial, allocation concealment unclear.
Blinding not mentioned.
No loss to follow up.
ParticipantsParticipants were undergoing elective abdominal reconstructive surgery.
Interventions1) 25% salt-poor albumin 1g/kg and 5% albumin 1L. (n=7)
2) Ringers lactate.
Allocated fluid was given intra-operatively. All patients received crystalloids only for pre-loading before surgery.
OutcomesDeaths were not reported.
Notes 
Allocation concealmentB
Study Tollofsrud 1995
MethodsRandomised controlled trial, allocation by sealed envelopes.
Blinding not mentioned.
No loss to follow up.
ParticipantsParticipants were adult patients in need of volume replacement during and after coronary artery bypass surgery.
Interventions1) Haemaccel (n=10)
2) Dextran 70 (n=10)
3) Albumin 40 (n=10)
4) Ringer's lactate (n=10).
Allocated fluid was used throughout resuscitation.
OutcomesDeaths reported.
Fluid balance.
NotesFollow up to 48 hours
Allocation concealmentB
Study Tollofsrud 1998
MethodsRandomised controlled trial, allocation by sealed envelope. Described as double blind, no loss to follow up mentioned.
ParticipantsPatients with three vessel coronary artery disease undergoing elective coronary artery surgery. Exclusions: LVEF<0.4, ventricular aneurysm, significant arrhythmia, diabetes, renal failure, lung disease.
Interventions1) 4mL/kg of 75mg/mL hypertonic saline in dextran 70 60mg/mL over 30 mins (n=10)
2) Same volume and rate of isotonic saline (n=10)
Fluid given just after surgery while still in operating theatre. Ringer's lactate for additional fluid.
OutcomesFluid balance.
Haemodynamic variables.
Deaths not reported.
NotesFollow up to 48 hours.
Allocation concealmentB
Study Vassar 1990
MethodsRandomised controlled trial, allocation concealment unclear.
Double blind study (solutions prepared in identical containers).
No loss to follow up.
ParticipantsParticipants were emergency department admissions with trauma and a systolic blood pressure below 80mmHg and were 18 years or older.
Pregnant women and people with preexisting cardiac, hepatic or renal disease were excluded.
Interventions1) 6% dextran 70 in 7.5% saline. (n=23)
2) Ringers lactate. (n=24)
Allocated fluids were given as the initial resuscitation in the emergency department. Additional isotonic crystalloids (Ringer's lactate) were given as needed.
OutcomesDeaths reported.
Haemodynamic variables.
NotesFollow up to hospital discharge.
Allocation concealmentB
Study Vassar 1991
MethodsRandomised controlled trial, allocation by randomised sequence of coded containers.
Double blind study.
No loss to follow up.
ParticipantsParticipants were pre-hospital trauma cases undergoing helicoptor transport to an emergency centre, with a systolic blood pressure of 100mmHg or less and were 18 years or older.
Exclusions: preexisting cardiac renal, hepatic or eurological disease. Peripheral oedema.
Interventions1) 4.2% dextran 70 in 7.5% saline or 6% dextran 70 in 7.5% saline. (n=83)
2) Ringers lactate. (n=83)
Fluids were given as the initial resuscitation fluid in the pre-hospital setting. Supplemental isotonic fluids were given at the discretion of the flight nurses.
OutcomesDeaths reported.
Haemodynamic variables
NotesFollow up to discharge. Allocation was to 4.2% dextran-70; to 6% dextran-70; or to crystalloid; for the calculation of the summary effect measure, the two dextran groups are combined.
Allocation concealmentA
Study Vassar 1993a
MethodsRandomised controlled blind trial, allocation concealed by random sequence of identical containers.
Double blind study.
36 people excluded post randomisation as deemed not to have met eligibility criteria.
No loss to follow up.
ParticipantsParticipants, who were undergoing ambulance transport to an emergency centre, had systolic blood pressure 90 mmHg or less, and were 18 years or older.
Exclusions: asystolic, undergoing CPR, lack sinus complex on ECG, more than 2 hours after trauma, pregnant, preexisting seizures, bleeding disorder, hepatic, cardiac or renal disease.
Interventions1) 6% dextran 70 in 7.5% saline. (n=89)
2) 7.5% saline. (n=85)
3) 0.9% saline (n=84)
Participants received 250mL of the allocated fluid in the pre-hospital setting. Additional isotonic crystalloids were given as needed.
OutcomesDeaths reported.
Haemodynamic variables.
Trauma scores.
NotesFollow up was to discharge from hospital.
Allocation concealmentA
Study Vassar 1993b
MethodsRandomised controlled trial, allocation concealed by sequential use of coded identical containers.
Double blind study.
39/233 patients excluded as deemed not to meet eligibility criteria, unclear from which groups.
ParticipantsParticipants were pre-hospital trauma cases undergoing helicoptor transport to an emergency centre, had a systolic blood pressure of 100mmHg or less and were 18 years or older.
Exclusions: asystolic, undergoing CPR, lack sinus complex on ECG, more than 2 hours after trauma, pregnant, preexisting seizures, bleeding disorder, hepatic, cardiac or renal disease.
Interventions1) 12% dextran70 in 7.5% saline. (n=49)
2) 6% dextran 70 in 7.5% saline. (n=50)
3) 7.5% saline. (n=50)
4) Ringer's lactate. (n=45)
Participants received 250mL of the allocated fluid in the pre-hospital setting. Additional isotonic crystalloids were given as needed.
OutcomesDeaths reported.
Haemodynamic variables.
Trauma scores and neurological outcome scores.
NotesFollow up to hospital discharge.
Allocation concealmentA
Study Virgilio 1979
MethodsAllocation 'by random number'.
Blinding not mentioned.
No loss to follow up.
ParticipantsParticipants were undergoing abdominal aortic surgery.
Interventions1) 5% albumin. (n=15)
2) Ringer's lactate. (n=14)
Allocated fluid was used during operation for maintenence of pre-defined physiological parameters, and the resuscitation was continued with the allocated fluid until the day following the operation. This was followed by 5% dextrose in half-normal saline, with potassium chloride as needed.
OutcomesDeaths reported.
NotesFollow up two and a half weeks
Allocation concealmentB
Study Wahba 1996
MethodsPatients 'randomly allocated'.
Blinding not mentioned.
Two patients excluded as they required reoperation for bleeding.
Participants22 adult patients in need of volume replacement following coronary artery bypass surgery.
Exclusions: abnormal left ventricular function, platelet active medication or heparin.
Interventions1) Haemaccell. (n=10)
2) Ringer's lactate. (n=10)
Allocated fluid was used from the time of admission to intensive care following operation, to the end of resuscitation.
OutcomesDeaths reported.
Pulmonary oedema.
NotesFollow up to discharge.
Allocation concealmentB
Study Woittiez 1997
MethodsRandomised controlled trial, allocation concealment by sealed opaque envelopes.
No information on blinding or loss to follow up.
Participants60 patients who had developed hypoalbuminaemia (<20g/l) after major surgery.
2 patients died after randomisation and before treatment started. They were excluded from the analysis.
Interventions1) saline (500ml/24 hr) (n=16)
2) albumin 20% (300 ml/24h) (n=15)
3) HES 10% (500ml/24h) for 3 days (n=27)
Aim was to restore colloid osmotic pressure.
OutcomesChanges in fluid balance, serum albumin, COP and clinical signs of oedema were followed daily.
Death rates supplied by the author.
NotesLength of follow up unspecified.
Allocation concealmentB
Study Younes 1992
MethodsRandomised 'in a double blind fashion'.
Blinding by use of similar bottles.
No loss to follow up.
ParticipantsParticipants were emergency department admissions, who had a systolic blood pressure of less than 80mm Hg and were 19 years and older.
Exclusions: pregnant, preexisting cardiac or metabolic disease.
Interventions1) 6% dextran 70 in 7.5% saline. (n=35)
2) 7.5% saline. (n=35)
3) 0.9% saline. (n=35)
Allocated fluid was for initial bolus of 250mL, followed by isotonic crystalloids as needed.
OutcomesDeaths reported.
Fluid balance.
NotesFollow up to discharge from hospital.
Allocation concealmentB
Study Younes 1994
MethodsTrial conducted in a 'double blind randomised fashion'.
Blinding by use of coded, identical containers.
ParticipantsParticipants were trauma admissions to the emergency room requiring treatment for haemorrhagic hypovolaemia; all were over 15 years old.
Exclusions: pregnant, cardiac or renal failure, cardiac arrest on arrival.
Interventions1) 6% dextran 70 in 7.5% saline. (n=101)
2) 0.9% saline. (n=111)
Allocated fluid was for the first intravenous infusion only.
OutcomesDeaths reported.
Complications.
NotesFollow up period was 30 days.
Allocation concealmentB
Study Younes 1998
MethodsRandomised controlled trial, allocation by sealed envelope. Blinding not mentioned, no apparent loss to follow p.
ParticipantsTrauma patients with systolic blood pressure <90mmHg admitted to the emergency room, with no previous treatment.
Interventions1) 10% pentastarch (n=12)
2) 0.9% saline (n=11)
Fluid given in 250mL boluses until systolic blood pressure>100mmHg
OutcomesDeaths reported.
No complications reported in either group.
NotesFollow up to 24 hours.
Allocation concealmentB
Study Zetterstrom 1981a
MethodsThe patients were randomly divided into two groups.
Allocation concealment was by sealed opaque envelopes (information supplied by author).
Blinding not mentioned.
No loss to follow up.
ParticipantsAdult patients undergoing elective major abdominal surgery.
Interventions1) Standard volume replacement regimen (1L Dextran 70 then up to 4 units of RBC with electrolyte, then whole blood or RBC with plasma; post op patients were given crystalloids and whole blood) plus 20% human albumin solution 100ml at end of operation, 200-300ml on same day, then 200ml on first post op day, then 100ml for next 3 days. (n=15)
2) Standard volume replacement regimen as above. (n=15)
OutcomesDeaths reported.
Haemodynamic variables.
NotesLength of follow up unspecified.
Allocation concealmentB
Study Zetterstrom 1981b
MethodsThe patients were randomly divided into two groups.
Allocation concealment was by sealed opaque envelopes (information supplied by author).
Blinding not mentioned.
No loss to follow up.
Participants18 patients who had undergone elective abdominal aortic surgery.
No exclusions mentioned.
Interventions1) 5% human albumin solution. (n=9)
2) Ringer's lactate solution. (n=9)
Administration guided by pulmonary arterial occlusion pressure.
OutcomesDeaths reported.
Haemodynamic variables.
NotesFollow up to discharge from hospital.
Allocation concealmentB

COP = colloid osmotic pressure
HES = hydroxyethylstarch
LVEDP = left ventricular end diastolic pressure
LVEF = left ventricular ejection fraction
RBC = red blood cells
PAWP = pulmonary artery wedge pressure
PAP = pulmonary artery pressure
WP = wedge pressure

Characteristics of excluded studies

StudyReason for exclusion
Artru 1989Intervention to control intracranial pressure not directed at fluid resuscitation.
Bocanegra 1966This study contained two quasi-randomised comparisons of colloid with glucose and plasma/saline with saline. In both studies, the control solution was only given IV if the patient was in coma or shock. It was therefore not a reasonable comparison of colloid and crystalloid.
Boldt 1996All groups received some colloid.
Bothner 1998Participants were having minor elective surgery, therefore not considered to be critically ill.
Breheme 1993Intervention directed at haemodilution, not at volume replacement.
Golub 1994Albumin given solely as a nutritional supplement.
Goslinga 1992Intervention directed at haemodilution, not volume replacement.
Greenhalgh 1995Intervention directed at the maintenance of serum albumin levels, not for volume replacement.
Hauser 1980Cross-over trial.
Lagonidis 1995Intervention was pre-loading for coronary artery bypass sugery.
Marhofer 1999Trial of fluid for preloading before spinal anaesthesia.
Nilsson 1980Albumin given as a nutritional supplement.
Steinberg 1989Cross-over trial.
Woods 1993This quasi-randomised trial looked at albumin supplementation in post operative patients, with the aim of maintaining the serum albumin. Since the main aim of giving albumin was not to replace volume, the study was excluded.

Graphs

Comparison 01. colloid vs crystalloid (add-on colloid)
Outcome titleNo. of studiesNo. of participantsStatistical methodEffect size
01 deaths  Relative Risk (Fixed) 95% CISubtotals only
Figure 1.

Figure 1.

    01 deaths

    Comparison 02. colloid and hypertonic crystalloid vs isotonic crystalloid
    Outcome titleNo. of studiesNo. of participantsStatistical methodEffect size
    01 deaths  Relative Risk (Fixed) 95% CISubtotals only
    Figure 2.

    Figure 2.

      01 deaths

      Comparison 03. colloid vs hypertonic crystalloid
      Outcome titleNo. of studiesNo. of participantsStatistical methodEffect size
      01 deaths  Relative Risk (Fixed) 95% CISubtotals only
      Figure 3.

      Figure 3.

        01 deaths

        Sources of support

        External sources of support

        • NHS R&D Programme: Mother and Child Health UK

        Internal sources of support

        • Institute of Child Health, University of London UK

        • UK Cochrane Centre, NHS R&D Programme UK

        Ancillary