Intervention Review

You have free access to this content

Human albumin solution for resuscitation and volume expansion in critically ill patients

  1. Ian Roberts1,*,
  2. Karen Blackhall1,
  3. Phil Alderson2,
  4. Frances Bunn3,
  5. Gillian Schierhout4

Editorial Group: Cochrane Injuries Group

Published Online: 9 NOV 2011

Assessed as up-to-date: 2 AUG 2011

DOI: 10.1002/14651858.CD001208.pub4


How to Cite

Roberts I, Blackhall K, Alderson P, Bunn F, Schierhout G. Human albumin solution for resuscitation and volume expansion in critically ill patients. Cochrane Database of Systematic Reviews 2011, Issue 11. Art. No.: CD001208. DOI: 10.1002/14651858.CD001208.pub4.

Author Information

  1. 1

    London School of Hygiene & Tropical Medicine, Cochrane Injuries Group, London, UK

  2. 2

    National Institute for Health and Clinical Excellence, Manchester, UK

  3. 3

    University of Hertfordshire, Centre for Research in Primary and Community Care, Hatfield, Hertfordshire, UK

  4. 4

    London School of Hygiene & Tropical Medicine, c/o Cochrane Injuries Group, London, UK

*Ian Roberts, Cochrane Injuries Group, London School of Hygiene & Tropical Medicine, North Courtyard, Keppel Street, London, WC1E 7HT, UK. Ian.Roberts@Lshtm.ac.uk.

Publication History

  1. Publication Status: New search for studies and content updated (no change to conclusions)
  2. Published Online: 9 NOV 2011

SEARCH

 
Characteristics of included studies [ordered by study ID]
Bland 1973

MethodsRandomised controlled trial. Therapy cards were randomised in pairs matched for weight. Method of allocation concealment not fully described.


ParticipantsNewborn infants considered at high risk for developing respiratory distress. Those with a cord serum protein level less than 4.6g/100ml and at least one of the following; birthweight less than 2500g, gestational age less than 37 weeks, arterial pH less than 7.25.


Interventions1) Intervention (n=50) received 8ml/kg 25% salt poor albumin.
2) Control group (n=50) received 8ml/kg 5% dextrose in water.


OutcomesDeaths reported within 28 days.


Notes

Bland 1976

MethodsRandomised controlled trial. Method of allocation concealment not fully described.


ParticipantsPremature infants (less than 37 weeks gestation), with hypoproteinaemia (cord serum total protein of 4.6g/100ml or less).


Interventions1) Intervention group (n=14) received 8ml/kg salt-poor albumin.
2) Comparison group (n=13) received 8ml/kg glucose in water.


OutcomesDeaths reported.


NotesLength of follow-up unspecified.

Boldt 1993

MethodsRandomised controlled trial. Method of allocation concealment not described in published report. Authors were contacted and confirmed the use of sealed opaque envelopes.


ParticipantsMen undergoing elective aortocoronary bypass grafting, who had a pulmonary capillary wedge pressure of less than 5mmHg after induction of anaesthesia.


Interventions1) Intervention (n=15): Albumin 5%.
2) Control (n=15): No additional volume.


OutcomesDeaths not reported.
Authors were contacted and confirmed that there were no deaths in the albumin nor the control group.


NotesFollow-up to 1 day.


Risk of bias

BiasAuthors' judgementSupport for judgement

Allocation concealment (selection bias)Low risk

Boutros 1979

MethodsRandomised controlled trial. Method of allocation concealment not fully described.


ParticipantsParticipants were undergoing major operative procedures on the abdominal aorta.


Interventions1) Intervention group (n=7) received albumin in 5% dextrose.
2) Control group (n=17) received 5% dextrose in lactated Ringers and 5% dextrose in 0.45 NaCl. Allocated fluids were used on admission to ICU, following surgery.


OutcomesDeaths reported.


NotesFollow-up to 48 hours after the end of the operation.

Brown 1988

MethodsRandomised controlled trial. Patients entered into the study were assigned to one of two treatment groups by a table of random numbers. Method of allocation concealment not described. Author contacted - no allocation concealment.


ParticipantsAll patients who required central TPN and had hypoalbuminaemia (serum albumin concentration below 3.0g/dl). Patients who were thermally injured, had nephrotic syndrome or required protein restriction were excluded.


Interventions1) The intervention group received central TPN plus normal serum albumin (n=33).
2) The control group (n=34) received central TPN alone.


OutcomesDeaths reported.


NotesFollow-up to discharge.


Risk of bias

BiasAuthors' judgementSupport for judgement

Allocation concealment (selection bias)High risk

Cooper 2006

MethodsRandomised controlled trial, not blinded.


ParticipantsEligible adults over 15 years old suffering from thermal injury not more than 12 hours before enrolment.


Interventions1) Intervention (n=19) received 5% albumin plus Ringer's lactate.
2) Control (n=23) received Ringer's lactate.


OutcomesMultiple Organ Dysfunction Score (MODS) and deaths reported.


Notes


Risk of bias

BiasAuthors' judgementSupport for judgement

Allocation concealment (selection bias)Low risk"Patients were allocated to study groups with stratified randomization with a computer-generated randomization list and sequentially numbered sealed, opaque envelopes. Randomization was stratified by center..." p. 81

Dubois 2006

MethodsProspective, controlled, randomized, open, single centre study.


ParticipantsAll patients admitted to the hospital ICU with serum albumin less than or equal to 30g/L/. Exclusion criteria were expected length of stay <72 hrs., life expectancy <3 months or a do not resuscitate order, albumin administration in the preceding 24 hrs. or evidence of fluid overload.


Interventions1) Intervention (n=50) received 300 mL of 20% albumin solution on day 1 and 200 mL on subsequent days when serum albumin concentration was lower than 31 g/L.
2) Control (n=50) received Ringer's Lactate.


OutcomesEffect of albumin on organ function as assessed by a delta Sequential Organ Failure Assessment score. Deaths not reported.


Notes


Risk of bias

BiasAuthors' judgementSupport for judgement

Allocation concealment (selection bias)Low risk"Eligible patients were randomized in a 1:1 ratio using sealed envelopes. When a patient was assigned to one group, he or she remained in that group whether or not he or she received the planned treatment."

Ernest 1999

MethodsRandomised controlled trial, not blinded.


Participants18 septic, critically ill patients where a fluid infusion was clinically indicated.


Interventions1) 5% albumin (n=9).
2) normal saline (n=9).


OutcomesInformation on death not collected.


NotesFollow up for about an hour after infusion.


Risk of bias

BiasAuthors' judgementSupport for judgement

Allocation concealment (selection bias)High risk

Ernest 2001

MethodsRandomised controlled trial, not blinded.


Participants40 postoperative cardiac surgical patients.


Interventions1) 5% albumin (n=23).
2) normal saline (n=17).


OutcomesInformation on death not collected.


NotesFollow up for 40 minutes after infusion. Trial conducted in 1992.


Risk of bias

BiasAuthors' judgementSupport for judgement

Allocation concealment (selection bias)High risk

Foley 1990

MethodsPatients were randomly assigned to either a treatment or non-treatment group by medical record number.


ParticipantsHypoalbuminaemic (serum albumin <25g/L) critically ill patients. Potential subjects with Child's class C cirrhosis were excluded.


Interventions1) The treatment group (n=18) received 25-50g per day of 25% albumin in addition to full nutritional support with parenteral nutrition. Albumin administration was continued daily until serum albumin levels exceeded 25 g/L after which patients received additional albumin as needed to keep the albumin level at 25 g/L or higher.
2) The non treatment group (n=22) received no exogenous concentrated albumin.


OutcomesDeaths reported.


NotesFollow up to discharge.


Risk of bias

BiasAuthors' judgementSupport for judgement

Allocation concealment (selection bias)High risk

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.


ParticipantsPatients after coronary artery bypass graft surgery.


Interventions1) Treatment group received 5% albumin (n=5).
2) The control group received lactated Ringers (n=5).


OutcomesDeaths were not reported. Author contacted and confirmed that there were no deaths in either group.


NotesFollow-up to 1 day.


Risk of bias

BiasAuthors' judgementSupport for judgement

Allocation concealment (selection bias)Low risk

Golub 1994

MethodsComputer randomisation - method of allocation concealment not described. Author contacted and confirmed that allocation concealment was by the use of sealed opaque envelopes.


ParticipantsPatients in the surgical intensive care unit of a community hospital with circulating albumin concentrations of <3.0g/dL.


Interventions1) The treatment group (n=116) received 37.5g/day of albumin until the circulating albumin concentration increased to 3.0g/dL.
2) The control group received no supplemental albumin.
Both groups received standard nutritional support.


OutcomesDeaths reported.


NotesFollow-up to discharge.


Risk of bias

BiasAuthors' judgementSupport for judgement

Allocation concealment (selection bias)Low risk

Goodwin 1983

MethodsRandomised controlled trial. Method of allocation concealment not described.


Participants79 thermally injured patients. No other inclusion criteria were reported. All of the participants were previously healthy young adults.


Interventions1) The treatment group (n=40) group received 2.5% albumin in Ringer's lactate.
2) The control group (n=39) Ringers lactate. Allocated fluid was used throughout resuscitation.


OutcomesDeaths reported.


NotesFollow-up to discharge.

Greenhalgh 1995

MethodsMethod of random allocation not described. Author contacted and confirmed the use of a randomisation scheme controlled by the pharmacy.


ParticipantsPatients aged 18 years or younger with acute burns.


Interventions1) High albumin group (n=34): Patients were supplemented with human albumin to maintain serum levels between 2.5 and 3.5g/dL. Albumin was supplied as a continuous drip of 25% human albumin at a rate of 3-10mL/hour. Supplementation was discontinued if serum albumin levels remained >2.5 g/dL without supplementation or if intravenous support was discontinued.
2) Low albumin group (n=36): Patients were not given albumin supplementation unless levels dropped <1.5 g/dL. During burn shock, patients were allowed to receive albumin if they had levels <2.0 g/dL and were receiving >4 mL/Kg/% burn fluid resuscitation.


OutcomesDeaths reported.


NotesFollow-up to discharge.


Risk of bias

BiasAuthors' judgementSupport for judgement

Allocation concealment (selection bias)Low risk

Greenough 1993

MethodsRandomised controlled trial. Allocation concealment by sealed opaque envelopes.


ParticipantsInfants between 24 and 34 weeks gestational age, who were ventilator dependent, and had a serum albumin level of less than or equal to 30g/l.


Interventions1)Intervention group (n=20) received 5ml/kg 20% salt-poor human albumin.
2) Control group (n=20) received 5ml/kg of the infant's maintenance fluids.


OutcomesDeaths were not reported. Author contacted and provided data on deaths.


NotesFollow-up to 24 hours after infusion.


Risk of bias

BiasAuthors' judgementSupport for judgement

Allocation concealment (selection bias)Low risk

Grundmann 1982

MethodsRandomised controlled trial. Method of allocation concealment not fully described.


ParticipantsParticipants were undergoing partial gastrectomy. The average age was 50 years (range 19-84).


Interventions1) Intervention group (n=14) group received human albumin.
2) Control group (n=6) details of crystalloid were not reported.
Allocated fluid was continued for 4 days after operation.


OutcomesDeaths reported.


NotesFollow-up to discharge.

Jelenko 1978

MethodsRandomised controlled trial. Method of allocation concealment not described.


ParticipantsParticipants had burns covering more than 20% of body surface.


Interventions1) Intervention group (n=7) received albumin in hypertonic saline (240MeQ Na+, 120 MeQ Chloride, 120 MeQ lactate, 3.5 torr/liter);
2) Control group (n=7) received hypertonic saline (240MeQ Na+, 120 MeQ Chloride, 120 MeQ lactate). Allocated fluid was used to the end of resuscitation.


OutcomesDeaths reported.


NotesFollow-up to 5 days.

Kanarek 1992

MethodsRandomised controlled trial. Allocation concealment by sealed opaque envelopes.


ParticipantsSick premature newborn infants whose serum albumin was less than 3g/dL.


Interventions1) Intervention group (n=12) received TPN with added albumin.
2) Control group (n=12) received no added albumin.


OutcomesDeaths reported.


NotesLength of follow-up unspecified.


Risk of bias

BiasAuthors' judgementSupport for judgement

Allocation concealment (selection bias)Low risk

Lowe 1977

MethodsRandomised controlled trial. The solutions were randomised by opening a sealed envelope containing a card denoting the appropriate fluid.


ParticipantsParticipants were undergoing emergency laparotomy for acute abdominal trauma.


Interventions1) Intervention group (n=57) received 50g albumin in 200ml in Ringers lactate;
2) Crystalloid group (n=84) received Ringer's 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.


Risk of bias

BiasAuthors' judgementSupport for judgement

Allocation concealment (selection bias)Low risk

Lucas 1978

MethodsRandomised controlled trial. Randomisation was based on the last digit of each patient's case number.


Participants52 seriously injured patients.


Interventions1) Intervention group (n=27) received supplemental salt-poor albumin totaling a maximum of 150g during operation and 150g per day over the next five days.
2) Control group (n=25) received standard resuscitation regimen but no supplemental albumin.


OutcomesDeaths reported.


NotesIn the final report of 94 randomised patients deaths were not reported. However, in this preliminary report of 52 injured patients deaths were reported.


Risk of bias

BiasAuthors' judgementSupport for judgement

Allocation concealment (selection bias)High risk

Maitland 2005

MethodsUnblinded, randomised, controlled trial.


ParticipantsKenyan children older than 2 months with symptomatic sever malarial anaemia (haemoglobin less than 5g/dl).


Interventions1) Intervention (n=23) received 4.5% albumin.
2) Intervention (n=20) received 0.9% saline.
3) Control (n=18) received only maintenance (rescue emergency intervention when required).


OutcomesChange in perfusion rates. Deaths not reported.


Notes


Risk of bias

BiasAuthors' judgementSupport for judgement

Allocation concealment (selection bias)Low risk"On admission to the HDU, children were randomly assigned to PTM with one of the three treatments using sealed card system. The sealed cards were available 24h/day; either the admitting doctor or nurse assigned the therapy at admission based on laboratory results or, in an emergency, on clinical suspicion." p. 128.

Maitland 2011

MethodsMulticentre open, randomised, controlled study


ParticipantsChildren (n=3170) aged between 60 days and twelve years of age, with severe febrile illness, randomly assigned within two strata (stratum A was children with severe febrile illness and impaired perfusion but without severe hypotension - stratum B was children with severe hypotension).


InterventionsChildren were randomly allocated to rapid volume replacement over the course of 1 hour with either:
1) 20 ml 5% Human Albumin solution per kg body weight (n=1063)
2) 20 ml 0.9% Saline solution per kg body weight (n=1063)


OutcomesMortality at 4 weeks after randomisation


NotesChildren (n=1044) assigned to no treatment were not included in the analysis.


Risk of bias

BiasAuthors' judgementSupport for judgement

Allocation concealment (selection bias)Low riskTrial numbers kept inside opaque, sealed envelopes. Opened in numerical order by clinician.

McNulty 1993

MethodsRandomised controlled trial. Method of allocation concealment not described.


Participants28 Patients following elective cardiopulmonary bypass.


Interventions1) Intervention group (n=14) received 5% albumin.
2) Control group (n=14) received isotonic crystalloid.


OutcomesDeaths not reported.


NotesLength of follow-up unspecified.

Myburgh 2007

MethodsA double blind, randomised, controlled SAFE 2004 trial. "Randomization was stratified by a diagnosis of trauma (defined as injury to the body caused by mechanical forces, excluding burns)." p. 875.


Participants460 patients with traumatic brain injuries.


Interventions1) Intervention (n=231) received 4% albumin.
2) Control (n=229) received normal saline.


OutcomesMortality rate at 28 days.


Notes


Risk of bias

BiasAuthors' judgementSupport for judgement

Allocation concealment (selection bias)Low risk"Randomization was carried out centrally with the use of a minimization algorithm, and the service was accessed on the Internet through a secure Web site." SAFE 2004 p. 2248

Nielsen 1985

MethodsRandomised controlled trial. Method of allocation concealment not described.


ParticipantsPatients admitted for reconstructive surgery of the abdominal aorta. Twenty six patients were randomised.


Interventions1) Intervention (n=13): 80 g of albumin administered in units of 100 ml 20% human serum albumin on the day of the operation and 20 g albumin daily on the following three postoperative days.
2) Control group (n=13): no supplemental albumin.


OutcomesDeaths not reported. Author when contacted confirmed that there were no deaths in either group.


NotesFollow-up 4 days.

Nilsson 1980

MethodsRandomised controlled trial. Allocation concealment by sealed opaque envelopes.


ParticipantsPatients with colorectal cancer undergoing elective surgery with resection of the tumour and primary anastomosis.


Interventions1) Intervention group (n=29) received 20-25g per day of albumin (as 5% albumin or 20% albumin) for three days, starting on the day after the operation.
2) Control group (n=30) received no albumin.


OutcomesDeaths reported.


NotesFollow up to discharge.


Risk of bias

BiasAuthors' judgementSupport for judgement

Allocation concealment (selection bias)Low risk

Oca 1999

MethodsRandomised controlled trial. Allocation concealment was by the use of sealed opaque sequentially numbered envelopes. Information obtained on contact with the author.


Participants24 neonates being treated for hypotension. Hypotension was defined as an oscilometric mean arterial blood pressure <30 mmHg for at least 30 minutes. Exclusion criteria consisted of proven sepsis, life-threatening congenital abnormalities, congenital hear disease, unresolved thoracic air leak, insulin-requiring maternal diabetes mellitus or treatment with high-frequency ventilation.


Interventions1) 5% albumin (n=11).
2) normal saline (n=13).


OutcomesMean arterial blood pressure.


NotesFollow-up to discharge.


Risk of bias

BiasAuthors' judgementSupport for judgement

Allocation concealment (selection bias)Low risk

Oca 2003

MethodsRandomised, unblinded, controlled trial.


ParticipantsNewborn infants who were <24 hours old and were admitted to the ICU to receive one of two solutions for volume expansion. 41 infants were included in the trial.


Interventions1) Intervention (n=21) received 4% albumin.
2) Control (n=20) received normal saline.


OutcomesMagnitude of change in mean arterial blood pressure.


Notes


Risk of bias

BiasAuthors' judgementSupport for judgement

Allocation concealment (selection bias)Low risk"Randomization was accomplished by random number tables via sealed, serially numbered envelopes" p. 474

Pockaj 1994

MethodsRandomised controlled trial. Method of allocation concealment not described.


ParticipantsParticipants required fluid resuscitation as a result of vascular leak syndrome associated with Interleukin-2 therapy for metastatic cancer.


Interventions1) Intervention group (n=54) received 5% albumin n 154meq/L NaCl;
2) Control group (n=53) received 0.9% normal saline with 154Meq/L NaCl.


OutcomesDeaths reported.


NotesLength of follow-up unspecified.

Prien 1990

MethodsRandomised controlled trial. Method of allocation concealment not described.


ParticipantsPatients undergoing hemipancreato-duodenectomy (Whipple's operation).


Interventions1) Intervention group (n=6) received 20% human albumin to maintain central venous pressure at the pre-operative level.
2) Control group (n=6) received Ringer's lactate.


OutcomesDeaths reported.


NotesLength of follow-up unspecified.

Quinlan 2004

MethodsProspective, randomised, placebo-controlled study.


ParticipantsPatients meeting the American European Consensus criteria for acute lung injury.


Interventions1) Intervention (n=10) received albumin (25g of a 25% solution) every 8 hrs for a total of nine doses.
2) Control (n=10) received normal saline administered in identical fashion and volume.


OutcomesDeaths noted.


Notes


Risk of bias

BiasAuthors' judgementSupport for judgement

Allocation concealment (selection bias)Unclear risk"Patients were randomized to receive 25g of human albumin every 8 hrs or a placebo (normal saline, administered in a double-blind fashion and targeted to normalization of serum total protein." p. 755 - 756.

Rackow 1983

MethodsRandomised controlled trial. Method of allocation concealment not described.


ParticipantsParticipants were above 18 years of age, and had any one of the following pre-determined indicators of shock: systolic blood pressure of 90mmHg or less, a cardiac index of less than 2.2L./min.m2, a serum arterial lactate greater than 18mg/dl and WP less than 15mmHg.


Interventions1) Intervention group ( n= 9) received 5% albumin.
2) Control group (n=8) received 0.9% NaCl. Allocated fluid was given as needed until the end of resuscitation.


OutcomesDeaths reported.


NotesFollow-up to discharge.

Rubin 1997

MethodsPatients were randomised using "a closed envelope system in the pharmacy".


ParticipantsPatients with hypoalbuminaemia (<2.5g/dL) who required TPN for at least six days, were not pregnant or under age, and did not have metastatic cancer, cirrhosis, or nephrotic syndrome.


Interventions1) Intervention group (n=16) 25g on normal serum albumin.
2) Control group (n=15) 100 mL of normal saline placebo over a 1 hour period daily.


OutcomesDeaths reported.


NotesFollow-up to discharge.


Risk of bias

BiasAuthors' judgementSupport for judgement

Allocation concealment (selection bias)Low risk

SAFE 2004

MethodsCentral randomisation accessed on the Internet through a secure website with use of a minimisation algorithm. Blinding was assured through the use of specially designed masking cartons and specially designed and manufactured administration sets. The authors report that the effectiveness of the blinding was confirmed in a formal study before the trial was initiated.


ParticipantsPatients 18 years of age or older admitted to ICU who the treating clinician judged to require fluid administration to maintain or increase intravascular volume, with this decision supported by the fulfilment of at least one objective criterion. Patients admitted after cardiac surgery, after liver transplantation, or for the treatment of burns were excluded.


Interventions1) 4% Albumin or
2) Normal saline

The allocated study treatment was used for all fluid resuscitation in the ICU until death or discharge or until 28 days after randomisation. The treating clinicians determined the amount and rate of fluid administration according to each patient's clinical status and response to treatment.


OutcomesDeaths reported.


Notes28 day mortality.


Risk of bias

BiasAuthors' judgementSupport for judgement

Allocation concealment (selection bias)Low risk"Randomization was carried out centrally with the use of a minimization algorithm and the service was accessed on the Internet through a secure Web site." p. 2248.

Shah 1977

MethodsRandomised controlled trial, allocation by sealed envelope.


ParticipantsPatients with severe, multiple trauma and a systolic blood pressure of less than 90mmHg. All patients were adults and both sexes were included.


Interventions1) Intervention group (n=9) 5% salt-poor albumin in Ringers lactate.
2) Control group (n=11) Ringer's lactate for resuscitation, volume infused guided by physiological parameters.


OutcomesDeath reported.


NotesLength of follow-up not stated.


Risk of bias

BiasAuthors' judgementSupport for judgement

Allocation concealment (selection bias)Low risk

Skillman 1975

MethodsRandomised controlled trial. Method of allocation concealment not described.


ParticipantsParticipants were undergoing elective abdominal reconstructive surgery.


Interventions1) Intervention group received 25% concentrated salt-poor albumin and 5% albumin in saline.
2) Control group received Ringer's lactate with 5% dextrose. Allocated fluid was given intra-operatively. All patients received crystalloids only for pre-loading before surgery.


OutcomesDeaths were not reported. Author could not be contacted.


NotesFollow-up to 1 day.

So 1997

MethodsRandomised controlled trial. Method of allocation concealment not described. Author contacted and confirmed that allocation concealment was by computer randomisation. Details of patient were entered before group allocation revealed.


ParticipantsPre-term infants weighing 540 to 1959g at birth, with gestational ages of 23 to 34 weeks, who developed hypotension within the first two hours of life.


Interventions1) Intervention group (n=32) were given 5% albumin at a dose of 10mg/Kg by slow intravenous infusion over 30 minutes.
2) Control group (n=31) were given 0.9%NaCl at a dose of 10mg/kg by slow intravenous infusion over 30 minutes.


OutcomesDeaths reported.


NotesFollow up to discharge.


Risk of bias

BiasAuthors' judgementSupport for judgement

Allocation concealment (selection bias)Low risk

Tollofsrud 1995

MethodsRandomised controlled trial. Allocation concealment by sealed opaque envelopes.


ParticipantsPatients undergoing elective coronary artery bypass surgery. Patients with left ventricular ejection fraction less than 40%, valvular heart disease, ventricular aneurysm, arrhythmia, diabetes mellitus, renal failure or lung disease were excluded.


Interventions1) Intervention group (n=10) received albumin 40mg/ml whenever fluid was required to stabilise haemodynamics.
2) Control group (n=10) received Ringers acetate.


OutcomesDeaths reported.


NotesFollow-up to 48 hours.


Risk of bias

BiasAuthors' judgementSupport for judgement

Allocation concealment (selection bias)Low risk

Virgilio 1979

MethodsRandomised controlled trial. Method of allocation concealment not described.


ParticipantsParticipants were undergoing abdominal aortic surgery.


Interventions1) Intervention group (n=15) received 5% albumin in Ringer's lactate
2) Control group (n=14) received Ringers lactate. Allocated fluid was used during operation for maintenance 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 2 and a half weeks.

Woittiez 1998

MethodsRandomised controlled trial. Allocation concealment by sealed opaque envelopes.


ParticipantsPost-operative intensive care patients.


Interventions1) Intervention group (n=15) received 20% albumin.
2) Control group (n=16) received 0.9% NaCl.


OutcomesUnpublished data on deaths were provided by the trialist.


NotesLength of follow-up unspecified.


Risk of bias

BiasAuthors' judgementSupport for judgement

Allocation concealment (selection bias)Low risk

Wojtysiak 1992

MethodsRandomised controlled trial. Table of random numbers was used to generate the random sequence. Method of allocation concealment not described in published report. The author was contacted and indicated that there was inadequate allocation concealment.


ParticipantsPatients between the ages of 18 and 75 years who were to receive parenteral nutrition and had a serum albumin concentration <3.0 g/dL. Patients were excluded if they had renal impairment, liver impairment or were haemodynamically unstable.


Interventions1) Intervention group (n= 15) had 25g of human albumin added to each litre of parenteral nutrition.
2) Control group (n=15) had no supplemental albumin.


OutcomesDeaths not reported in published report. Author when contacted confirmed that there were no deaths in either group.


NotesFollow-up to 5 days.


Risk of bias

BiasAuthors' judgementSupport for judgement

Allocation concealment (selection bias)High risk

Woods 1993

MethodsRandomised controlled trial. Patients with even hospital numbers were randomised to the group receiving albumin while those patients with odd hospital numbers were randomised to the group not receiving supplemental albumin.


ParticipantsPatients undergoing surgery for abdominal aortic aneurysm, aortoiliac or aortofemoral bypass.


Interventions1) Intervention group (n=37): albumin was replaced to a level greater than or equal to 3.5 g/dL.
2) Control group (n=32): received no supplemental albumin.


OutcomesDeaths reported.


NotesFollow-up to discharge.


Risk of bias

BiasAuthors' judgementSupport for judgement

Allocation concealment (selection bias)High risk

Zetterstrom 1981a

MethodsThe patients were randomly divided into two groups. The method of allocation concealment is not described. Author was contacted and confirmed the use of sealed opaque envelopes.


ParticipantsAdult patients undergoing elective major abdominal surgery.


Interventions1) Intervention group (n=15)
2) Control group (n=15)
A similar schedule of fluid therapy and blood replacement was followed in the intervention and control groups. However, the albumin group received a 20% solution of human albumin intravenously according to the following scheme:
At the end of the operation: 100ml.
Postoperatively on the day of the operation: 200-300 ml.
First day after the operation: 200 ml.
Following 3 days 100 ml each day.


OutcomesDeaths reported.


NotesLength of follow-up unspecified.


Risk of bias

BiasAuthors' judgementSupport for judgement

Allocation concealment (selection bias)Low risk

Zetterstrom 1981b

MethodsPatients were randomly divided into two groups. Method of allocation concealment was not described. Author was contacted and confirmed the use of sealed opaque envelopes.


ParticipantsPatients undergoing elective reconstruction of the abdominal aorta.


Interventions1) Intervention group (n=9)
2) Control group (n=9)
Postoperatively, the aim of fluid administration was to keep the pulmonary arterial occlusion pressure equal to the preoperative level. When lower values were recorded, the patients in the control group were given a balanced electrolyte solution of the Ringer type, whereas the albumin patients received a 5% solution of human albumin.


OutcomesDeaths reported.


NotesLength of follow-up unspecified.


Risk of bias

BiasAuthors' judgementSupport for judgement

Allocation concealment (selection bias)Low risk

 
Characteristics of excluded studies [ordered by study ID]

StudyReason for exclusion

Artru 1989Intervention to control intracranial pressure not directed at fluid resuscitation.

Bellomo 2006Trial was subset of SAFE 2004 trial, thus offering no new data.

Brehme 1993Intervention directed at haemodilution, not at volume replacement.

Carlon 1979Randomised controlled trial of pre-operative volume expansion during anaesthesia.

Fiorica 1991Not a randomised trial. The first 18 patients received standard maintenance crystalloid solution. The next 10 consecutive patients received 100g of a concentrated 25% albumin solution.

Goslinga 1992Intervention directed at haemodilution, not volume replacement.

Grundmann 1985This was a randomised controlled trial of 220 patients, 106 were given albumin when their colloid osmotic pressure (COP) fell below 24 cm water and 114 were given albumin when their COP fell below 29 cm water. Patients were not randomised to albumin or no albumin, nor were they randomised to supplemental albumin versus normal amounts of albumin, rather, this was a trial of different criteria for albumin supplementation. It is unlikely therefore that the two arms of the trial were comparable and hence the trial is excluded.

Grundmann 1986This was a randomised controlled trial to examine whether postoperative human albumin supply is justified in intensive care patients in the case that the colloid osmotic pressure decreases below 26 centimetres of water. The therapy group received human albumin only if the colloid osmotic pressure dropped below 26 cm water. The control group also received albumin but only for resuscitation of cardiac output and central venous pressure. The trial was excluded because both intervention and control groups received albumin.

Hauser 1980Cross-over trial.

Lagonidis 1995Intervention was pre-loading for coronary artery bypass surgery.

Lennihan 2000Participants had suffered subarachnoid hemorrhage and therefore did not meet the inclusion criteria.

Magder 1999Participants were stable patients following cardiopulmonary bypass surgery and therefore did not meet the inclusion criteria.

Maitland 2005aTrial was subset of SAFE 2004 trial, thus offering no new data.

Martin 1999Intervention involved comparison of albumin with furosemide verus placebo therefore did not meet the inclusion criteria.

Metildi 1984Participants 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 and therefore did not meet the inclusion criteria for the review.

Steinberg 1989Cross-over trial.

Tomita 1994Randomised controlled trial of normal versus high oncotic pressure following head injury. Patients were not randomised to albumin or no albumin. Albumin and furosemide were used together to achieve high oncotic pressure.

 
Comparison 1. Supplemental albumin

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 deaths3810842Odds Ratio (M-H, Fixed, 95% CI)1.05 [0.95, 1.16]

    1.1 hypovolaemia
229880Odds Ratio (M-H, Fixed, 95% CI)1.02 [0.92, 1.13]

    1.2 burns
4205Odds Ratio (M-H, Fixed, 95% CI)2.93 [1.28, 6.72]

    1.3 hypoalbuminaemia
12757Odds Ratio (M-H, Fixed, 95% CI)1.26 [0.84, 1.88]