Methods

Quasi‐RCT

Parallel design

Single centre

Participants

Total number of randomised participants: 180

Inclusion criteria: patients admitted to MIU

Exclusion criteria: no details

Participant condition: head, chest, abdominal injuries

Baseline characteristics

Colloids group

• Age, mean (range): 28 (21‐60) years

• Gender, M:F: 81:9

• BP, mean (range): SBP: 95 (35‐130); DBP: 49 (10‐70) mmHg

Crystalloids group

• Age, mean (range): 27 (21‐59) years

• Gender, M:F: 81:9

• BP, mean (SD): SBP: 97 (40‐127); DBP: 51 (12‐75) mmHg

Country: UK

Setting: MIU

Interventions

Colloids group

• Participants: n = 90; losses = 0; analysed = 90

• Details: 7.5% NaCl in 4.2% dextran 70; 4 mL/kg up to a maximum 250 mL

• Additional details: further fluid infusions continued with Hartmann's or blood transfusions, if required

Crystalloids group

• Participants: n = 90; losses = 0; analysed = 90

• Details: we have assumed that crystalloid solution was RL from other information in the study report

• Additional details: further fluid infusions continued with Hartmann's (RL) or blood transfusions if required

Outcomes

Outcomes measured/reported: haemodynamic analysis; urine outputs; recovery; LoS

Outcomes relevant to the review: mortality (time not reported)

Notes

Funding/declarations of interest: none apparent

Study dates: not reported

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

High risk

Alternate participants added to each group based on odd/even numbers

Allocation concealment (selection bias)

High risk

Alternate allocation used and therefore unlikely to be concealed

Blinding of participants and personnel (performance bias): mortality

Low risk

No details of blinding; unlikely to introduce bias for this outcome

Blinding of outcome assessment (detection bias): mortality

Low risk

No details of blinding; not likely to introduce bias for this outcome

Incomplete outcome data (attrition bias)
All outcomes

Low risk

No apparent losses

Selective reporting (reporting bias)

Unclear risk

No details of clinical trials registration; not feasible to assess risk of selective reporting bias

Baseline characteristics

Unclear risk

The proportion of participants in each arm with chest injuries differed. It is unclear whether this influenced results.

Other bias

Low risk

No other sources of bias identified

Methods

RCT

Parallel design

Multicentre

Participants

Total number of randomised participants: 2857

Inclusion criteria: no prior fluid resuscitation in ICU; required fluid resuscitation for acute hypovolaemia

Exclusion criteria: received fluid resuscitation in ICU; anaesthesia‐related hypotension; advanced chronic liver disease; acute anaphylactic reaction; inherited coagulation disorders; do‐not‐resuscitate order; pregnant; burned > 20% of TBSA; allergy to study drug; refused consent; dehydrated; brain death or organ donor; other (not specified)

Participant condition: acute hypovolaemia, sepsis, and trauma

Baseline characteristics

Colloids group

• Age, median (IQR): 63 (50‐76) years

• Gender, M:F: 880:534

• Weight, median (IQR): 70 (60‐81) kg

• BP, median (IQR): SBP: 92 (80‐112) mmHg

• SAPS II, median (IQR): 48 (35‐64)

Crystalloids group

• Age, median (IQR): 50 (36‐65) years

• Gender, M:F: 902:541

• Weight, median (IQR): 70 (61‐81) kg

• BP, median (IQR): SBP: 94 (80‐113) mmHg

• SAPS II, median (IQR): 50 (36‐65)

Country: France, Belgium, Canada, Algeria, Tunisia

Setting: ICU

Interventions

Colloids group

• Participants: n = 1414; losses = 0; analysed = 1414

• Details: colloids, any type from 4% gelatin, 5% albumin, dextrans, HES, 20% or 25% albumin; at discretion of local investigators; not > 30 mL/kg/d; median in first 7 days 2000 mL (IQR, 1000 mL‐3502 mlL; median 2 d duration

• Additional details: participants received colloids or crystalloids prior to ICU

Crystalloids group

• Participants: n = 1443; losses = 0; analysed = 1443

• Details: crystalloids, any type; at discretion of local investigator; median for first 7 days 3000 mL (IQR, 500 mL‐5200 mL); median 2 d duration

• Additional details: isotonic saline or HS, any buffered solutions; participants received colloids or crystalloids prior to ICU

Outcomes

Outcomes measured/reported: mortality at 28 days; mortality at 90 days and at ICU and hospital discharge; number of days alive and not receiving renal replacement therapy, mechanical ventilation or vasopressor therapy; days not in ICU or hospital; days without organ failure

Outcomes relevant to the review: mortality (at 28 days, 90 days, and at end of follow‐up); renal replacement therapy; requiring blood transfusion

Notes

Funding/declarations of interest: funded by French Ministry of Health. Study sponsors not involved in design and conduct of study

Study dates: Febuary 2003‐November 2012

Note: study was stopped early because study authors noted no difference in 28‐day mortality rates.

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

Computer‐generated sequence

Allocation concealment (selection bias)

Low risk

Sealed envelopes were used at the bedside to allow randomisation of eligible participants without any delay and was done blinded to block size

Blinding of participants and personnel (performance bias): mortality

Low risk

Clinicians were not blinded because of immediate need for resuscitation; unlikely to introduce bias for this outcome

Blinding of participants and personnel (performance bias): transfusion/renal replacement therapy/adverse events

High risk

Clinicians were not blinded because of immediate need for resuscitation; could introduce bias for this outcome

Blinding of outcome assessment (detection bias): mortality

Low risk

Quote: "mortality end‐points were collected and assessed by study members blinded to treatment assignment." Unlikely to introduce bias for this outcome

Blinding of outcome assessment (detection bias): transfusion/renal replacement therapy/adverse events

Unclear risk

No details

Incomplete outcome data (attrition bias)
All outcomes

Low risk

No losses

Selective reporting (reporting bias)

Low risk

Prospective clinical trials registration (NCT00318942); all outcomes listed on registration site were reported

Baseline characteristics

Low risk

Baseline characteristics appear comparable

Other bias

High risk

47.5% of participants in the crystalloid group were given colloids within 12 h before the start of the study and this may have influenced study results

Methods

RCT

Parallel design

Multicentre (2 x level 1 adult trauma centres)

Participants

Total number of randomised participants: 64

Inclusion criteria: coma, with a loss of consciousness because of isolated blunt head trauma or a GCS score ≤ 8

Exclusion criteria: primary penetrating injury; previous IV therapy ≥ 50 mL; time of arrival at scene to IV access > 4 h; < 16 years of age; burn or amputation; presumed to be pregnant; vital signs absent prior to randomisation

Participant condition: blunt trauma head injury

Baseline characteristics

Colloids group

• Age, mean (SD): 42.5 (± 20.9) years

• Gender, M:F: 18:13

• APACHE II, mean (SD): 13.2 (± 5.6)

Crystalloids group

• Age, mean (SD): 42.3 (± 20.7) years

• Gender, M:F: 23:10

• APACHE II, mean (SD): 14.4 (± 5.2)

Country: Canada

Setting: ambulatory prior to adult‐designated level 1 trauma centres

Interventions

Colloids group

• Participants: n = 31; losses = 0; analysed = 31

• Details: 7.5% HS in 6% dextran 70; 250 mL

• Additional details: emergency medical service personnel administered the study solution prehospital; after administration of study fluid participants were treated according to ATLSG; participants received additional crystalloid for ongoing resuscitation per existing protocols

Crystalloids group

• Participants: n = 33; losses = 0; analysed = 33

• Details: 0.9% isotonic NS; 250 mL

• Additional details: same as colloid group

Outcomes

Outcomes measured/reported: neurological outcomes at hospital discharge (or 30 days) using various scales; mortality; biomarkers

Outcomes relevant to the review: mortality (at 28 days)

Notes

Funding/declarations of interest: funded by Defence Research and Development Canada

Study dates: September 2004‐January 2006

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

Computer‐generated sequence

Allocation concealment (selection bias)

Low risk

Computer randomisation was used to assign sequentially numbered identical IV bags to the ambulance

Blinding of participants and personnel (performance bias): mortality

Low risk

Participants and personnel were blinded to treatment allocation

Blinding of outcome assessment (detection bias): mortality

Low risk

Paramedics, physicians and study co‐ordinators were blinded to treatment allocation

Incomplete outcome data (attrition bias)
All outcomes

Low risk

No apparent losses

Selective reporting (reporting bias)

Unclear risk

No details of clinical trials registration; not feasible to assess risk of selective reporting bias

Baseline characteristics

Low risk

Baseline characteristics appear comparable

Other bias

High risk

Study authors report that participants could receive additional fluid resuscitation during standard care and this could influence outcome results for this study

Methods

RCT

Parallel design

Single centre

Participants

Total number of randomised participants: 48

Inclusion criteria: ≥ 16 years of age with second‐ or third‐degree acute burn injuries and > 15% of body surface area burned

Exclusion criteria: expected to die within 24‐36 h (i.e. burn victims with whole body burn trauma); in situations of palliative care; pregnancy; lack of informed consent; known allergy to HES; contraindications for balanced 6% HES 130/0.04; intracerebral bleeding; acute renal failure; severe hypernatraemia and other severe electrolyte disorders; severe von Willebrand Syndrome; acute liver failure

Participant condition: burns; TBSA > 15%

Baseline characteristics

Colloids group

• Age, median (IQR): 49 (22‐69) years

• Gender, M:F: 17:6

• Weight, median (IQR): 75 (70‐83) kg

• BP, median (IQR): SBP: 109 (93‐130); DBP: 60 (55‐65) mmHg

• TBSA, median burned (IQR): 31% (21‐47)

Crystalloids group

• Age, median (IQR): 47 (26‐61) years

• Gender, M:F: 17:5

• Weight, median (IQR): 80 (70‐80) kg

• BP, median (IQR): SBP: 123 (104‐150) mmHg; DBP: 68 (59‐76) mmHg

• TBSA, median burned (IQR): 32% (20%‐50%)

Country: Switzerland

Setting: tertiary burns unit

Interventions

Colloids group

• Participants: n = 24; losses = 0; analysed = 24

• Details: 6% HES 130/0.4; 500 mL; each participant first received 2 bags of unblinded RL solution (500 mL each bag); after each bag of study solution, all participants again received 2 bags of unblinded RL solution, before a next bag of study solution from the blinded box was infused; maximum to be given as 50 mL/kg/24 h

• Additional details: fluid was administered until target variables were met; 2 bags of unblinded RL (500 mL each bag); then 1 bag of HES; then 2 bags of unblinded RL

Crystalloids group

• Participants: n = 24; losses = 0; analysed = 24

• Details: RL solution; 500 mL; each participant first received 2 bags of unblinded RL solution (500 mL each bag); after each bag of study solution, all participants again received 2 bags of unblinded RL solution, before a next bag of study solution from the blinded box was infused

• Additional details: as for colloids group but given RL in blinded bags in between unblinded bags

Outcomes

Outcomes measured/reported: group difference in administration of fluid with 72 h; creatinine levels; urine output; ARDS; LoS in ICU; LoS in hospital; in‐hospital mortality and at 28 days; post‐hoc 90‐day mortality; RRT

Outcomes relevant to the review: mortality (28 days; and 90 days); RRT (collected as a 90‐day post‐hoc analysis)

Notes

Funding/declarations of interest: funding from manufacturer of HES, which supplied study fluids; 2 of the authors have vocationally been members of advisory board meetings. No competing interests declared. Funders reported as having no input in study design and interpretation of results

Study dates: November 2009‐January 2013

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

Randomisation completed using minimisation technique, conducted by a third party

Allocation concealment (selection bias)

Low risk

A third party not involved in conduction of study, performed the randomisation process

Blinding of participants and personnel (performance bias): mortality

Low risk

All personnel blinded. Fluids prepared externally, and concealed in bags of black plastic

Blinding of participants and personnel (performance bias): transfusion/renal replacement therapy/adverse events

Low risk

All personnel blinded. Fluids prepared externally, and concealed in bags of black plastic

Blinding of outcome assessment (detection bias): mortality

Low risk

No details in study report. However, trial registration report states that outcome assessors were blinded

Blinding of outcome assessment (detection bias): transfusion/renal replacement therapy/adverse events

Low risk

No details in study report. However, trial registration report states that outcome assessors were blinded

Incomplete outcome data (attrition bias)
All outcomes

Low risk

2 participants were retrospectively excluded because of meeting exclusion criteria. Data missing from 1 additional participant because of early discharge. Overall, < 10% dropout/exclusion; data reported for 45/48 randomised participants

Selective reporting (reporting bias)

High risk

Prospective clinical trials registration (NCT01012648). Only primary outcome (fluid volume administered) was listed on the trial registration site

Baseline characteristics

Low risk

Baseline characteristics comparable

Other bias

Low risk

No other sources of bias identified

Methods

RCT

Parallel design

Single centre

Participants

Total number of randomised participants: 22

Inclusion criteria: ICU patients with an acute, spontaneous subarachnoid haemorrhage, with stable ICP in the range of 10 mmHg‐20 mmHg; > 18 years of age; sedated; mechanically ventilated; stable haemodynamics; serum sodium of < 160 mmol/L

Exclusion criteria: no details

Participant condition: spontaneous subarachnoid haemorrhage

Baseline characteristics

Colloids group

• Age, mean (SD): 50.1 (± 10.5) years

• Gender, M:F: 3:8

• SAPS II, mean (SD): 40.5 (± 11.1)

Crystalloids group

• Age, mean (SD): 55.2 (± 10.8) years

• Gender, M:F: 1:10

• SAPS II, mean (SD): 47.0 (± 12.1)

Country: Norway

Setting: ICU

Interventions

Colloids group

• Participants: n = 11; losses = 0; analysed = 11

• Details: 7.2% saline in 6% HES 200/0.5; 2 mL/kg over 30 min

• Additional details: participants monitored from 10 min before to 210 min after start of infusion; need for rescue treatment was defined by treatment failure limits for ICP (> 20 mmHg) and CPP (< 60 mmHg). Otherwise, no changes to study fluid regimen

Crystalloids group

• Participants: n = 11; losses = 0; analysed = 11

• Details: 0.9% saline solution; 2 mL/kg over 30 min

• Additional details: same as colloid group

Outcomes

Outcomes measured/reported: ICP; CPP; extravascular lung water; serum sodium levels

Outcomes relevant to the review: none

Notes

Funding/declarations of interest: not reported

Study dates: April 2002‐October 2004

Participant condition not reported by group; "A total of 21 patients had haemorrhaged because of a ruptured aneurysm, and one patient was diagnosed with a fusiform dilation of the left vertebral artery."

Methods

RCT

Parallel design

Multicentre

Participants

Total number of randomised participants: 537

Inclusion criteria: patients with severe sepsis or septic shock; ≥ 18 years of age; onset of the syndrome < 24 h before admission to the ICU or < 12 h after admission if the condition developed in the ICU

Exclusion criteria: treatment with > 1000 mL of HES within 24 h before study inclusion; pre‐existing renal failure requiring dialysis or a serum creatinine level ≥ 320 μmoL/L (3.6 mg/dL); < 18 years of age; pregnancy; known allergy against HES; intra‐cerebral haemorrhage; heart failure with NYHA IV; requirement of an inspiratory oxygen fraction of at least 0.7; immunosuppression from cytostatic chemotherapy; high dosage of steroids or AIDS; participation in another interventional trial; moribund due to coexisting disease; order to withhold or withdraw therapy

Participant condition: severe sepsis or septic shock

Baseline characteristics

Colloids group

• Age, mean (SD): 64.4 (± 13.3) years

• Gender, M:F: 158:104

• APACHE II, mean (SD): 20.1 (± 6.7)

Crystalloids group

• Age, mean (range): 64.9 (± 14.1) years

• Gender, M:F: 164:111

• APACHE II, mean (SD): 20.3 (± 6.7)

Country: Germany

Setting: ICU; 18 tertiary hospitals

Interventions

Colloids group

• Participants: n = 262; losses = 0; analysed = 262

• Details: 10% pentastarch; HES 200/0.5 with 0.9% NS; to achieve CVP 8 mmHg, MAP > 70 mmHG or central venous oxygen saturation > 70%; given for up to 96 h

• Additional details: participants to be excluded if they had received > 1000 mL HES in 24 h prior to randomisation; all participants given different insulin therapies in a 2 x 2 factorial design

Crystalloids group

• Participants: n = 275; losses = 0; analysed = 275

• Details: RL; to achieve CVP 8 mmHg, MAP > 70 mmHg or central venous oxygen saturation > 70%; given for up to 96 h

• Additional details: same as colloids group

Outcomes

Outcomes measured/reported: mortality (at 28 days and 90 days); morbidity (according to SOFA scores); need for blood transfusion; renal failure (to include need for RRT); time to haemodynamic stabilisation; frequency of vasopressor therapy; need for red‐cell transfusion; duration of mechanical ventilation, LoS in the ICU; adverse events (worsening of oxygenation, bleeding complications, allergic reaction, any event judged to occur in relation to study fluid)

Outcomes relevant to the review: mortality (at 28 days and 90 days; need transfusion of a blood product; need for renal replacement therapy

Notes

Funding/declarations of interest: supported by a grant (01 KI 0106) from the German Federal Ministry of Education and Research and by unrestricted grants from B Braun, HemoCue and Novo Nordisk

Study dates: April 2003‐June 2005

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

No details

Allocation concealment (selection bias)

Unclear risk

No details

Blinding of participants and personnel (performance bias): mortality

Low risk

Open‐label design; unlikely to introduce bias for this outcome

Blinding of participants and personnel (performance bias): transfusion/renal replacement therapy/adverse events

High risk

Open‐label design; could introduce bias for this outcome

Blinding of outcome assessment (detection bias): mortality

Low risk

No details; unlikely to introduce bias for this outcome

Blinding of outcome assessment (detection bias): transfusion/renal replacement therapy/adverse events

Unclear risk

No details

Incomplete outcome data (attrition bias)
All outcomes

Low risk

No losses

Selective reporting (reporting bias)

Unclear risk

No details of clinical trials registration; not feasible to assess risk of selective reporting bias

Baseline characteristics

Low risk

Baseline characteristics appear comparable

Other bias

High risk

Note: 26.6% of participants in the crystalloid group were given colloids during the study period and this may have influenced study results

Methods

RCT

Parallel design

Single centre

Participants

Total number of randomised participants: 209

Inclusion criteria: blunt trauma; > 17 years of age (or adult size if age was unknown); at least 1 prehospital SBP measurement ≤ 90 mmHg; transported directly to a single level 1 trauma centre from the site of injury

Exclusion criteria: ongoing cardiopulmonary resuscitation; isolated penetrating trauma; known or suspected pregnancy; receipt of > 2000 mL of crystalloid before availability of study fluid

Participant condition: blunt trauma

Baseline characteristics

Colloids group

• Age, mean (SD): 41 (± 18) years

• Gender, M:F: 69:41

• BP, mean (SD): SBP: 71 (± 27) mmHg

Crystalloids group

• Age, mean (SD): 38 (± 19) years

• Gender, M:F: 68:31

• BP, mean (SD): SBP: 72 (± 25) mmHg

Country: USA

Setting: prehospital (ambulatory) prior to admission to a single level 1 trauma centre

Interventions

Colloids group

• Participants: n = 110; losses = 0; analysed = 110

• Details: 7.5% HS and 6% dextran 70 (HSD); 250 mL; followed by additional RL as necessary during transport

Crystalloids group

• Participants: n = 99; losses = 0; analysed = 99

• Details: 250 mL followed by additional RL as necessary during transport

Outcomes

Outcomes measured/reported: incidence of ARDS; mortality; multiple organ failure syndrome; nosocomial infections; length of hospital and ICU stay; ventilator‐free days; adverse events; non‐infectious complications

Outcomes relevant to the review: mortality (28 days)

Notes

Funding/declarations of interest: grant R01 HL073233‐01 from the National Institutes of Health

Study dates: October 2003‐August 2005

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

Computer‐generated randomisation

Allocation concealment (selection bias)

Low risk

A random number (computer‐generated by pharmacist) was applied to each bag and kept by the pharmacist. Ambulance crew did not have access to allocation sequence

Blinding of participants and personnel (performance bias): mortality

Low risk

All contents of fluid bags were blinded by research pharmacists. Therefore, personnel and participants were blinded to treatment assignment

Blinding of outcome assessment (detection bias): mortality

Low risk

No details of blinding; unlikely to introduce bias for mortality outcomes

Incomplete outcome data (attrition bias)
All outcomes

Unclear risk

21 participants did not meet eligibility criteria once randomisation had taken place but remained in the results using ITT analysis. Three participants lost to follow‐up, explanations reported by study authors

Selective reporting (reporting bias)

Unclear risk

Clinical trials registration ID: NCT01012648. All outcomes specified on clinical trials registration site were reported. However, we noted that the outcomes were only added to the trials registration site after the study start date

Baseline characteristics

Unclear risk

We noted higher injury severity scores for those in the colloids group, and we could not be certain whether this could influence outcome data

Other bias

Low risk

No other sources of bias identified

Methods

RCT

Parallel design

Multicentre

Participants

Total number of randomised participants: 1331

Inclusion criteria: blunt mechanism of injury; ≥ 15 years of age; GCS score ≤ 8; ineligibility for enrolment in the haemorrhagic shock cohort

Exclusion criteria: known or suspected pregnancy; < 15 years of age; out‐of‐hospital cardiopulmonary resuscitation; administration of > 2000 mL of crystalloid or any amount of colloid or blood products prior to enrolment; severe hypothermia (28 °C); drowning; asphyxia because of hanging; burns on > 20% of TBSA; isolated penetrating head injury; inability to obtain IV access; > 4 h between receipt of dispatch call to study intervention

Participant condition: traumatic brain injury

Baseline characteristics

Colloids group

• Age, mean (SD): 38.5 (± 18.6) years

• Gender, M:F: 86:273

• BP, mean (SD): SBP: 141.2 (± 33.1) mmHg

Crystalloids group (NS)

• Age, mean (SD): 39.5 (± 19.2) years

• Gender, M:F: 156:426

• BP, mean (SD): SBP: 139.1 (± 33.1) mmHg

Crystalloids group (HS)

• Age, mean (SD): 38.6 (± 17.3) years

• Gender, M:F: 64:277

• BP, mean (SD): SBP: 136.9 (± 33.5) mmHg

Country: USA and Canada

Setting: 11 regional clinical centres

Interventions

Colloids group

• Participants: n = 373; losses = 14 (5 did not meet inclusion criteria; 3 met an exclusion criterion; 4 had no IV access; 1 fluid bag sterility broke; 1 EMS responder unsure of inclusion/exclusion criteria); analysed = 359

• Details: 7.5% saline in 6% dextran 70; 250 mL

• Additional details: single bolus; all conducted out‐of‐hospital; participants may have been given fluid before attendance of study personnel but must have only received < 2 L of crystalloid and no colloid, mannitol or blood products

Crystalloids group (NS)

• Participants: n = 603; losses = 21 (8 did not meet inclusion criteria; 4 had inadequate time to administer; 2 met an exclusion criterion; 4 had no IV access; 2 fluid bag sterility broke; 1 EMS responder unsure of inclusion/exclusion criteria); analysed = 582

• Details: 0.9% saline; 250 mL

• Additional details: same as colloid group

Crystalloids group (HS)

• Participants: n = 355; losses = 14 (5 did not meet inclusion criteria; 1 met an exclusion criterion; 6 had no IV access; 1 fluid bag sterility broke; 1 EMS responder unsure of inclusion/exclusion criteria); analysed = 341

• Details: 7.5% saline; 250 mL

• Additional details: same as colloid group

Outcomes

Outcomes measured/reported: 6‐month neurologic status (Glasgow Outcome Score); 28‐day survival; survival to discharge; ICP; interventions required to manage intracranial hypertension; fluid and bolus requirements in first 24 h; physiologic parameters of organ dysfunction; 28‐day ARDS‐free survival; MODS; nosocomial infections

Outcomes relevant to the review: mortality (28 days)

Notes

Funding/declarations of interest: National Heart, Lung and Blood Institute plus partners

Study dates: May 2006‐May 2009

Study terminated after futility criteria met at 6 months

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

Randomly generated numeric code used at central location

Allocation concealment (selection bias)

Low risk

Randomisation scheme conducted externally and all personnel unaware of allocation

Blinding of participants and personnel (performance bias): mortality

Low risk

Quote: "Study fluids were provided in identical intravenous bags and shipped to a single distribution center, where they were labelled with a randomly generated numeric code"

Participants, caregivers, and outcome assessors were blinded to treatment

Blinding of outcome assessment (detection bias): mortality

Low risk

Participants, caregivers, and outcome assessors were blinded to treatment

Incomplete outcome data (attrition bias)
All outcomes

Low risk

Mortality data reported for 359/373 (HSD), 341/355 (HS), and 582/603 (NS). < 5% dropout/loss in each group

Selective reporting (reporting bias)

Low risk

Prospective clinical trials registration: NCT00316004. All outcomes were prespecified

Baseline characteristics

Low risk

Baseline characteristics appear comparable

Other bias

Low risk

No other sources of bias identified

Methods

RCT

Parallel design

Multicentre

Participants

Total number of randomised participants: 895

Inclusion criteria: ≥ 15 years of age; in significant haemorrhagic shock (out‐of‐hospital SBP ≤ 70 mmHg or 71‐90 mmHg with concomitant HR ≤ 108 bpm)

Exclusion criteria: known or suspected pregnancy; < 15 years of age; out‐of‐hospital cardiopulmonary resuscitation; administration of > 2000 mL crystalloid, colloid, or blood products before enrolment; severe hypothermia (< 28 °C); drowning; asphyxia because of hanging; burns > 20% TBSA; isolated penetrating head injury; inability to obtain IV access; time of dispatch call received to study intervention > 4 h; known prisoners

Participant condition: traumatic hypovolaemic shock

Baseline characteristics

Colloids group

• Age, mean (SD): 37.7 (± 17.3) years

• Gender, M:F: 170:50

• BP, mean (SD): SBP: 59.1 (± 35.5) mmHg

• GCS, mean (SD): 10.0 (± 4.9)

Crystalloids group (NS)

• Age, mean (SD): 36.2 (± 16.4) years

• Gender, M:F: 291:85

• BP, mean (SD): SBP: 58.1 (± 32.2) mmHg

• GCS, mean (SD): 9.8 (± 5.0)

Crystalloids group (HS)

• Age, mean (SD): 36.8 (± 16.1) years

• Gender, M:F: 205:52

• BP, mean (SD): SBP: 54.1 (± 35.3) mmHg

• GCS, mean (SD): 10.0 (± 5.0)

Country: USA and Canada

Setting: out‐of‐hospital

Interventions

Colloids group

• Participants: n = 231; losses = 0; analysed = 231

• Details: 7.5% saline in 6% dextran 70 (HSD); 250 mL bolus

• Additional details: bolus given in out‐of‐hospital setting; once study fluid had been administered, additional fluids could be given as guided by local EMS protocols

Crystalloids group (NS)

• Participants: n = 395; losses = 0; analysed = 395

• Details: 0.9% NS; 250 mL bolus

• Additional details: same as colloid group

Crystalloids group (HS)

• Participants: n = 269; losses = 0; analysed = 269

• Details: 7.5% HS

• Additional details: same as colloid group

Outcomes

Outcomes measured/reported: 28‐day survival; physiologic parameters of organ dysfunction; ARDS criteria met in the first 28 days after injury; MODS; presence of nosocomial infection

Outcomes relevant to the review: mortality (28 days); participants having transfusion (0‐9 units); participants having transfusion (> 10 units)

Notes

Funding/declarations of interest: The National Heart, Lung and Blood Institute. Study authors declare no financial conflicts of interest

Study dates: May 2006‐August 2008

Note: the previous version of this review did not include participants in the HS group (Perel 2013). We have included outcome data for these participants, and in analysis we have combined both crystalloid groups.

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

Quote: "Randomly generated numeric code was applied to each bag and a randomization list kept by the Data Co‐ordinating Center"

Information taken from study protocol

Allocation concealment (selection bias)

Low risk

Randomisation list kept by study investigators (taken from study protocol)

Blinding of participants and personnel (performance bias): mortality

Low risk

Care providers, investigators, and participants were blinded to treatment assignment, study fluids concealed

Blinding of participants and personnel (performance bias): transfusion/renal replacement therapy/adverse events

Low risk

Care providers, investigators, and participants were blinded to treatment assignment, study fluids concealed

Blinding of outcome assessment (detection bias): mortality

Low risk

All personnel were blinded

Blinding of outcome assessment (detection bias): transfusion/renal replacement therapy/adverse events

Low risk

All personnel were blinded

Incomplete outcome data (attrition bias)
All outcomes

Low risk

42/895 participants were not included in analysis but reasons were clearly provided (most of these losses were because of inclusion/exclusion criteria)

Selective reporting (reporting bias)

Unclear risk

A protocol was published for this study; publication of protocol was retrospective and it was not feasible to use this to assess risk of selective reporting bias

Baseline characteristics

Low risk

Baseline characteristics appear comparable

Other bias

High risk

We noted that participants may have received up to 2000 mL of crystalloid or colloid before randomisation (as part of exclusion criteria). Study authors did not report how many participants received fluid resuscitation before randomisation, or which fluid was given, and this may influence outcome data for this study

Methods

RCT

Parallel design

Multicentre

Participants

Total number of randomised participants: 1810

Inclusion criteria: ≥ 18 years of age; severe sepsis within previous 24 h

Exclusion criteria: < 18 years of age; terminal state; known adverse reaction to albumin administration; severe sepsis or septic shock after proved or suspected head injury; clinically active; congestive heart failure (NYHA class 3 or 4); pathological conditions in which albumin administration was clinically indicated (hepatic cirrhosis with ascites, intestinal malabsorption syndrome, nephrotic syndrome, burns); > 24 h since inclusion criteria were met; religious objection to the administration of human blood products; inclusion in other experimental studies

Participant condition: severe sepsis

Baseline characteristics

Colloids group

• Age, median (IQR): 70 (57‐77) years

• Gender, M:F: 543:360

• SAPS II, median (IQR): 48 (37‐59)

Crystalloids group

• Age, median (IQR): 69 (59‐77) years

• Gender, M:F: 550:357

• SAPS II, median (IQR): 48 (37‐60)

Country: Italy

Setting: ICU

Interventions

Colloids group

• Participants: n = 903; losses = 0; analysed = 903

• Details: 20% albumin; 300 mL; fluids administered according to the “early‐goal directed therapy” protocol; administered from day 1 until day 28 or ICU discharge to maintain serum concentration ≥ 30 g/L; given crystalloids whenever clinically indicated by attending physician

• Additional details: all conducted out‐of‐hospital. Participants may have been given fluid before attendance of study personnel, but had to have received < 2 L of crystalloid and no colloid, mannitol or blood products

Crystalloids group

• Participants: n = 907; losses = 0; analysed = 907

• Details: no details of crystalloid solution or administration

• Additional details: same as colloid group

Outcomes

Outcomes measured/reported: death from any cause (28 days); death from any cause (90 days); number of participants with organ dysfunction; length of ICU and hospital stay

Outcomes relevant to the review: mortality (at 28 days, and at 90 days); RRT

Notes

Funding/declarations of interest: Italian Medicines Agency

Study dates: Aug 2008‐Feb 2012

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

Quote: "Randomization was performed centrally, with the use of the computer‐generated and blinded assignment sequence. Randomization was stratified according to the participating ICU and the interval between the time that the patient met the clinical criteria for severe sepsis and randomization"

Allocation concealment (selection bias)

Low risk

Central allocation, blinded

Blinding of participants and personnel (performance bias): mortality

Low risk

Open‐label study; lack of blinding unlikely to introduce bias for this outcome

Blinding of participants and personnel (performance bias): transfusion/renal replacement therapy/adverse events

Unclear risk

No details

Blinding of outcome assessment (detection bias): mortality

Low risk

No details of blinding; unlikely to introduce bias for this outcome

Blinding of outcome assessment (detection bias): transfusion/renal replacement therapy/adverse events

Unclear risk

No details

Incomplete outcome data (attrition bias)
All outcomes

Low risk

Few losses; unlikely to affect analysis

Selective reporting (reporting bias)

Low risk

Prospective clinical trials registration (NCT00707122). All outcomes listed were reported

Baseline characteristics

Unclear risk

Quote: "Baseline characteristics were similar between the two study groups, except for a slight imbalance in the number of patients with organ dysfunction and values of central venous oxygen saturation".

It was not reported if these differences between groups were at a level of statistical significance. We were uncertain whether these differences might influence the results

Other bias

Low risk

No other sources of bias identified

Methods

RCT

Parallel design

Single centre

Participants

Total number of randomised participants: 49

Inclusion criteria: SBP ≤ 90 mmHg for < 1 h; normal ECG; written consent by participant or first‐degree relative

Exclusion criteria: pregnancy; renal, cardiac, or neurological diseases

Participant condition: haemorrhagic shock

Baseline characteristics

Colloids group

• Age, mean (range): 42 (22‐76) years

• Gender, M:F: 18:8

• BP, mean (SD): SBP: 64 (± 21); DBP: 32 (± 14) mmHg

Crystalloids group

• Age, mean (range): 42 (52‐58 [sic]) years

• Gender, M:F: 14:9

• BP, mean (SD): SBP: 75 (± 18); DBP: 40 (± 12)

Country: Mexico

Setting: hospital

Interventions

Colloids group

• Participants: n = 26; losses = 0; analysed = 26

• Details: 7.5% NaCl in 6% dextran 60; 250 mL

• Additional details: all solutions were administered as soon as possible; 16 participants by peripheral vein; 10 participants via the intraosseous route; supplementary isotonic saline fluid given to achieve SBP > 100 mmHg

Crystalloids group

• Participants: n = 23; losses = 0; analysed = 23

• Details: conventional RL

• Additional details: by peripheral vein; supplementary isotonic saline fluid given to achieve SBP > 100 mmHg; dextran 40 given if necessary according to medical judgement

Outcomes

Outcomes measured/reported: haemodynamic variables; urinary output; GCS; mortality (within 24 h)

Outcomes relevant to the review: mortality

Notes

Funding/declarations of interest: none reported

Study dates: not reported

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

Participants were assigned to groups using random numbers but no Additional details provided

Allocation concealment (selection bias)

Unclear risk

No details

Blinding of participants and personnel (performance bias): mortality

Low risk

No details of blinding; unlikely to introduce bias for this outcome

Blinding of outcome assessment (detection bias): mortality

Low risk

No details of blinding; unlikely to introduce bias for this outcome

Incomplete outcome data (attrition bias)
All outcomes

Low risk

No apparent losses

Selective reporting (reporting bias)

Unclear risk

No details of clinical trials registration; not feasible to assess risk of selective reporting bias

Baseline characteristics

Low risk

Baseline characteristics appear comparable

Other bias

High risk

Quote: "Dextran 40 was administered to the control group if necessary according to medical judgement." Study authors did not report the number of participants in the crystalloid group who received additional colloids and this may influence outcome data for this study

Methods

Quasi‐RCT

Parallel design

Single centre

Participants

Total number of randomised participants: 26

Inclusion criteria: admitted to a children's medical centre; fever lasting 2‐7 days; haemorrhagic manifestations; evidence of consumptive coagulopathy, a fall in platelet count, prolonged bleeding, prolonged prothrombin time, or prolonged partial thromboplastin time; evidence of plasma leakage; evidence of circulatory failure

Exclusion criteria: severe infection other than dengue haemorrhagic fever; protein‐deficient abnormalities; bleeding diathesis; given multiple plasma substitutes

Participant condition: DSS

Baseline characteristics

Colloids group

• Age, mean (SD): 56.2 (± 22.86) months

• Gender, M:F: 6:5

Crystalloids group

• Age, mean (SD): 73.88 (± 28.66) months

• Gender, M:F: 10:6

Country: Philippines

Setting: ICU

Interventions

Colloids group

• Participants: n = 11; losses = 1 (withdrawn from study because different fluid management was required); analysed = 11 for mortality data; 10 for blood transfusion data

• Details: 6% Haes‐Steril given in doses of 10 mL/kg‐20 mL/kg; doses repeated ≥ 2‐3 times until vital signs were restored to normal

• Additional details: once vital signs were restored, participants were given fluids according to hospital ICU hydration protocol

Crystalloids group

• Participants: n = 16 ; losses = 3 (withdrawn from study because different fluid management was required); analysed = 16 for mortality data; 13 for blood transfusion data

• Details: RL given in doses of 10 mL/kg‐20 mL/kg; doses repeated ≥ 2‐3 times until vital signs were restored to normal

• Additional details: once vital signs were restored, participants were given fluids according to hospital ICU hydration protocol

Outcomes

Outcomes measured/reported: duration of control of shock, haematocrit level, length of ICU stay, transfusion of blood products, frequency of recurrence of shock, mortality

Outcomes relevant to the review: mortality (time point not reported); transfusion of blood products (FFP or packed red blood cells)

Notes

Funding/declarations of interest: not reported

Study dates: June 2001‐July 2001

Note: 3 out of 16 participants in the crystalloid group (18.75%) also received colloids during the study period and this may have influenced study results

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

High risk

Quasi‐randomised method to allocate participants, using alternating allocation to each group

Allocation concealment (selection bias)

High risk

Not possible to conceal allocation because of methods used to allocate participants

Blinding of participants and personnel (performance bias): mortality

Low risk

Personnel were not blinded; however, unlikely to introduce bias for this outcome

Blinding of participants and personnel (performance bias): transfusion/renal replacement therapy/adverse events

Unclear risk

No details

Blinding of outcome assessment (detection bias): mortality

Low risk

No details of blinding; unlikely to introduce bias for this outcome

Blinding of outcome assessment (detection bias): transfusion/renal replacement therapy/adverse events

Unclear risk

No details

Incomplete outcome data (attrition bias)
All outcomes

Low risk

Four participants were excluded from some analysis. Because mortality data were reported for these participants we included these in analysis.

Selective reporting (reporting bias)

Unclear risk

No details of clinical trials registration; not feasible to assess risk of selective reporting bias

Baseline characteristics

Low risk

Baseline characteristics appear comparable

Other bias

High risk

3 out of 16 participants in the crystalloid group (18.75%) also received colloids during the study period and this may have influenced study results

Methods

RCT

Parallel design

Multicentre

Participants

Total number of randomised participants: 42

Inclusion criteria: thermal burn of ≥ 20% TBSA; time elapsed since injury ≤ 12 h; written informed consent from the participant or a suitable substitute decision maker; availability of data regarding fluids administered before arrival at the study centre

Exclusion criteria: unlikely survival, defined as APACHE II score > 30 or predicted mortality ≥ 90%; ventricular fibrillation; ventricular tachycardia; unstable angina; known congestive heart failure or myocardial infarction within the month before thermal injury; electrical or chemical burn injury; pregnancy

Participant condition: burns

Baseline characteristics

Colloids group

• Age, median (95% CI): 36 (24‐45) years

• Gender, M:F: 15:4

• Weight, median (95% CI): 80 (70‐100) kg

• APACHE II, median (95% CI): 15 (11‐27)

Crystalloids group

• Age, median (95% CI): 31 (25‐39) years

• Gender, M:F: 21:2

• Weight, median (95% CI): 82 (75‐90) kg

• APACHE II, median (95% CI): 10 (10‐14)

Country: Canada

Setting: hospital units

Interventions

Colloids group

• Participants: n = 19; losses = 0; analysed = 19

• Details: 5% albumin; participants initially given basal rate of saline using calculation; then given stabilisation rate (2 mL × body weight × TBSA%)/24 mL/h; within first 24 h, followed by stabilisation phase for > 24 h until wound closure

• Additional details: participants received fluids through two independently controlled infusions (BR and AFR) over two periods: not > 24 h after injury (resuscitation phase) and > 24 h and injury (stabilisation phase); the use of synthetic colloid starches for volume resuscitation was not permitted; conservative red cell and blood product transfusion strategies were also recommended

Crystalloids group

• Participants: n = 23; losses = 0; analysed = 23

• Details: RL as BR = (2 mL × body weight × TBSA%) − TFV/24 mL/h; and as additional flow rate (2 mL x body weight x TBSA%)/24 mL/h; within first 24 h, followed by stabilisation phase for > 24 h until wound closure

• Additional details: same as colloid group

Outcomes

Outcomes measured/reported: MODS; mortality; duration of mechanical ventilation; LoS in ICU; local infection events; systemic infection events; percentage of graft take; oxygenation failure (PaO₂‐to‐FiO₂ ratio) (all evaluated up to and including Day 28)

Outcomes relevant to the review: mortality (28 days); blood transfusion

Notes

Funding/declarations of interest: funded by Bayer Biologics, Canada

Study dates: June 1999‐June 2001

Trial stopped early due to slow enrolment

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

Computer‐generated randomisation list

Allocation concealment (selection bias)

Low risk

Sequentially numbered, sealed, opaque envelopes

Blinding of participants and personnel (performance bias): mortality

Low risk

No details of blinding; unlikely to introduce bias for this outcome

Blinding of participants and personnel (performance bias): transfusion/renal replacement therapy/adverse events

High risk

Quote: "Treatment fluid was given in an open label fashion owing to differences in the physical properties (color, tendency to bubble) and medium of delivery (glass vials vs. polymer bags)"

Could introduce bias for blood transfusion outcome

Blinding of outcome assessment (detection bias): mortality

Low risk

No details; lack of blinding unlikely to influence data for mortality

Blinding of outcome assessment (detection bias): transfusion/renal replacement therapy/adverse events

Unclear risk

No details

Incomplete outcome data (attrition bias)
All outcomes

Low risk

No losses

Selective reporting (reporting bias)

Unclear risk

No details of clinical trials registration; not feasible to assess risk of selective reporting bias

Baseline characteristics

Unclear risk

Baseline characteristics and demographics were comparable between groups except for predicted mortality, which was greater in the colloid group (18.6%) compared with the crystalloid group (9.4%)

Other bias

Low risk

No other sources of bias

Methods

RCT

Parallel design

Single centre

Participants

Total number of randomised participants: 41

Inclusion criteria: adults; male and female; with hospital diagnosis of severe acute pancreatitis

Exclusion criteria: history of allergy to HES; history of cardiac dysfunction or renal insufficiency; pregnancy, malignancy or immunoinsufficiency; other colloids within 24 h; serum albumin < 25 g/L; likely death within 48 h. Also excluded those who died within 72 h; received surgery during treatment period; severe adverse effects to HES

Participant condition: severe acute pancreatitis

Baseline characteristics

Colloids group

• Age, mean (SD): 45.6 (± 10.8) years

• Gender, M:F: 12:8

• APACHE II, mean (SD): 12.1 (± 10.8)

• BP, mean (SD): SBP: 134.9 (± 12.8) mmHg

Crystalloids group

• Age, mean (SD): 45.7 (± 11.1) years

• Gender, M:F: 12:9

• APACHE II, mean (SD): 10.7 (± 4.1)

• BP, mean (SD): SBP: 128.6 (± 12.2) mmHg

Country: China

Setting: university hospital

Interventions

Colloids group

• Participants: n = 20; losses = 0; analysed = 20

• Details: 6% HES 130/0.4, plus RL; RL given to both groups at 1 mL/kg/h‐2 mL/kg/h; HES infused at volume ratio of 1:3 compared with saline solution

• Additional details: rate and volume given to maintain haemodynamic stability

Crystalloids group

• Participants: n = 21; losses = 0; analysed = 21

• Details: RL; given at 1 mL/kg/h‐2 mL/kg/h

• Additional details: same as colloid group

Outcomes

Outcomes measured/reported: mortality (within‐hospital stay); intra‐abdominal pressure; fluid balance; major organ complications; use of respirator; APACHE II score; serum levels of inflammatory mediators

Outcomes relevant to the review: mortality (within‐hospital stay)

Notes

Funding/declarations of interest: supported by Sichuan Province of Science and Technology Department Technology Support Project

Study dates: January 2008‐November 2009

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

Computer‐derived random number table

Allocation concealment (selection bias)

Unclear risk

No details

Blinding of participants and personnel (performance bias): mortality

Low risk

No details of blinding; unlikely to introduce bias for this outcome

Blinding of outcome assessment (detection bias): mortality

Low risk

No details of blinding; unlikely to introduce bias for this outcome

Incomplete outcome data (attrition bias)
All outcomes

Low risk

We have included 1 participant that was excluded from the study as this participant died therefore providing relevant data for this review

Selective reporting (reporting bias)

Unclear risk

No details of clinical trials registration; not feasible to assess risk of selective reporting bias

Baseline characteristics

Low risk

Baseline characteristics appear comparable

Other bias

High risk

57.1% of participants in the crystalloid group were given colloids during the study period and we noted that this may have influenced study results

Methods

RCT

Parallel design

Multicentre

Participants

Total number of randomised participants: 25

Inclusion criteria: ≥ 18 years of age; confirmed or suspected infection plus ≥ 2 signs of the systemic inflammatory response syndrome (definition of sepsis by American College of Chest Physicians/Society of Critical Care Medicine criteria), and tissue hypoperfusion (MAP < 65 mmHg despite a crystalloid fluid challenge of 20 mL/kg or blood lactate concentration of ≥ 4 mmol/L)

Exclusion criteria: impossible to perform sublingual video‐microscopy; < 18 years of age; pregnancy; stroke; acute coronary syndrome; hydrostatic pulmonary oedema; status asthmaticus; cardiac arrhythmias (as a main diagnosis); contraindication for central venous catheterisation; active gastrointestinal haemorrhage; seizures; drug intoxications; burns; trauma; need of immediate surgery; terminal cancer; immunosuppression (organ transplant or systemic illness); no resuscitation order; delayed admission to the intensive care unit from the emergency department (> 4 h); or previous resuscitation with > 1500 mL of fluids

Participant condition: sepsis

Baseline characteristics

Colloids group

• Age, mean (SD): 62 (± 21) years

• Gender, M:F: 8:4

Crystalloids group

• Age, mean (SD): 65 (± 12) years

• Gender, M:F: 7:6

Country: Argentina

Setting: 2 teaching ICUs

Interventions

Colloids group

• Participants: n = 12; losses = 0; analysed = 12

• Details: 6% HES 130/0.4 (Voluven); early goal‐directed therapy; administered to achieve CVP 8‐12 mmHg, MAP > 65 mmHg, and ScV0₂ ≥ 70%

• Additional details: up to 1500 mL fluids permitted

Crystalloids group

• Participants: n = 13; losses = 0; analysed = 13

• Details: 0.9% saline; early goal directed therapy: administered to achieve CVP 8‐12 mmHg, MAP > 65 mmHg, and ScV02 ≥ 70%

• Additional details: same as colloid group

Outcomes

Outcomes measured/reported: heart rate; MAP; CVP; central venous gases and oxygen saturations; microcirculatory variables; mortality

Outcomes relevant to the review: mortality (time point not reported)

Notes

Funding/declarations of interest: supported by the grant PICT‐2007‐00912, Agencia Nacional de Promoción Científica y Tecnológica, Argentina

Study dates: January 2006‐August 2009

Note: data for mortality were not clearly reported in the study report. We have included deaths of participants within 24 h and combined these with deaths reported in the study report outcome table. The previous version of this review did not include mortality outcome data (Perel 2013).

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

Quote: "Simple randomization by the use of sealed envelopes"

Insufficient details to allow judgement

Allocation concealment (selection bias)

Unclear risk

Sealed envelopes, but no mention of opaqueness, or whether they were numbered sequentially

Blinding of participants and personnel (performance bias): mortality

Low risk

No details of blinding; unlikely to introduce bias for this outcome

Blinding of outcome assessment (detection bias): mortality

Low risk

No details of blinding; unlikely to introduce bias for this outcome

Incomplete outcome data (attrition bias)
All outcomes

Low risk

Study authors reported a small number of losses because of death. We included these as data for the mortality outcome

Selective reporting (reporting bias)

Unclear risk

Clinical trials registration occurred after the start of the study (NCT00799916); not feasible to assess risk of selective outcome reporting bias

Baseline characteristics

Low risk

Baseline characteristics appear comparable

Other bias

Low risk

No other sources of bias identified

Methods

RCT

Parallel design

Single centre

Participants

Total number of randomised participants: 50

Inclusion criteria: 5‐15 years of age; DSS; had not received IV fluid therapy during their current illness

Exclusion criteria: no details

Participant condition: DSS

Baseline characteristics

Colloids group (dextran)

• No baseline characteristics reported

Colloids group (gelatin)

• No baseline characteristics reported

Crystalloids group (RL)

• No baseline characteristics reported

Crystalloids group (NS)

• No baseline characteristics reported

Country: Vietnam

Setting: hospital

Interventions

Colloids group (dextran)

• Participants: n = 12; losses = 0; analysed = 12

• Details: dextran 70 (60 g dextran in 0.9% saline); 20 mL/kg for first hour; 10 mL/kg for next hour; IV; in packs of 500 mL; study fluids only given for 2 h then subsequent fluid given according to physician preference and WHO guidelines

Colloids group (gelatin)

• Participants: n = 13; losses = 0; analysed = 13

• Details: Gelafundin, 35,000 Da; 20 mL/kg for first hour; 10 mL/kg for next h; IV; in packs of 500 mL; study fluids only given for 2 h then subsequent fluid given according to physician preference and WHO guidelines

Crystalloids group (RL)

• Participants: n = 13; losses = 0; analysed = 13

• Details: RL solution; 20 mL/kg for first hour; 10 mL/kg for next h; IV; in packs of 500 mL; study fluids only given for 2 h then subsequent fluid given according to physician preference and WHO guidelines

Crystalloids group (NS)

• Participants: n = 12; losses = 0; analysed = 12

• Details: 0.9% w/v saline and chloride; 20 mL/kg for first hour; 10 mL/kg for next h; IV; in packs of 500 mL; study fluids only given for 2 h then subsequent fluid given according to physician preference and WHO guidelines

Outcomes

Outcomes measured/reported: recovery from shock; duration of shock and number of episodes of shock; improvements in cardiac output and haematocrit values; requirements for further fluid resuscitation

Outcomes relevant to the review: none

Notes

Funding/declarations of interest: B Braun provided the fluids used in this study. Financial support from The Wellcome Trust of Great Britain

Study dates: all participants admitted between July and November 1995

Methods

RCT

Parallel design

Single centre

Participants

Total number of randomised participants: 18

Inclusion criteria: septic; critically ill; fluid infusion clinically indicated; pulmonary catheter already in place; patient not overtly bleeding

Exclusion criteria: no details

Participant condition: sepsis

Baseline characteristics

Colloids group

• Age, mean (SD): 51 (± 21) years

• Gender, M:F: 5:4

• APACHE II, mean (SD): 19 (± 8)

Crystalloids group

• Age, mean (SD): 55 (± 17) years

• Gender, M:F: 6:3

• APACHE II, mean (SD): 16 (± 7)

Country: Canada

Setting: ICU

Interventions

Colloids group

• Participants: n = 9; losses = 0; analysed = 9

• Details: 5% albumin; fluid infusion to meet PAOP determined by clinician, which was mostly 15 mmHg

Crystalloids group

• Participants: n = 9; losses = 0; analysed = 9

• Details: NS; fluid infusion to meet PAOP determined by clinician, which was mostly 15 mmHg

Outcomes

Outcomes measured/reported: MAP, PAOP, cardiac index, arterial oxygen content, plasma albumin concentration, PV and ECFV

Outcomes relevant to the review: none

Notes

Funding/declarations of interest: none reported

Study dates: not reported

Methods

RCT

Parallel design

Single centre

Participants

Total number of randomised participants: 25

Inclusion criteria: > 16 years of age; blunt or penetrating trauma; requiring IV fluid resuscitation; arrival at trauma unit within 2 h of injury; RL as the only prehospital infusion; no underlying illness or medication that would affect the patient's coagulating system

Exclusion criteria: no details

Participant condition: trauma

Baseline characteristics

Colloids group

• Age, median (IQR): 30 (29‐38) years

• Gender, M:F: 9:2

Crystalloids group

• Age, median (IQR): 30 (25‐39) years

• Gender, M:F: 12:2

Country: South Africa

Setting: hospital

Interventions

Colloids group

• Participants: n = 11; losses = 0; analysed = 11

• Details: Haemaccel; given fluid until fully resuscitated, with end point as stable vital signs

Crystalloids group

• Participants: n = 14; losses = 0; analysed = 14

• Details: RL; given fluid until fully resuscitated, with end point as stable vital signs

Outcomes

Outcomes measured/reported: bleeding times, prothrombin, thrombin, partial thromboplastin times, platelet count, secondary resuscitation

Outcomes relevant to the review: mortality (data from personal communication with study authors; time point unknown)

Notes

Funding/declarations of interest: "Hoechst SA for their independent grant and sponsorship for this research project"

Study dates: not reported

Note: we used mortality data reported in the previous version of this review (Perel 2013). These data were collected from personal communication with the study authors

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

Described as randomised but no further details

Allocation concealment (selection bias)

Unclear risk

No details

Blinding of participants and personnel (performance bias): mortality

Low risk

No details of blinding; unlikely to introduce bias for this outcome

Blinding of outcome assessment (detection bias): mortality

Low risk

No details of blinding; unlikely to introduce bias for this outcome

Incomplete outcome data (attrition bias)
All outcomes

Low risk

No apparent losses

Selective reporting (reporting bias)

Unclear risk

No details of clinical trials registration; not feasible to assess risk of selective reporting bias

Baseline characteristics

Low risk

Baseline characteristics appear comparable

Other bias

Low risk

No other sources of bias identified

Methods

RCT

Parallel design

Multicentre

Participants

Total number of randomised participants: 6997

Inclusion criteria: ≥ 18 years of age; treating clinician judged to require fluid administration to maintain or increase intravascular volume

Exclusion criteria: people admitted to ICU after cardiac surgery; liver transplantation; treatment of burns

Participant condition: various ICU admissions (to include trauma, sepsis, ARDS)

Baseline characteristics

Colloids group

• Age, mean (SD): 58.6 (± 19.1) years

• Gender, M:F: 2073:1424

• APACHE II, mean (SD): 18.7 (± 7.9)

Crystalloids group

• Age, mean (SD): 58.5 (± 18.7) years

• Gender, M:F: 2124:1376

• APACHE II, mean (SD): 19.0 (± 8.0)

Country: Australia and New Zealand

Setting: hospital ‐ 16 ICUs

Interventions

Colloids group

• Participants: n = 3497; losses = study authors reported loss of 26 participants mostly because of withdrawal of surrogate consent; analysed = 3473

• Details: 4% albumin; volume determined by treating clinicians

• Additional details: until discharge, death or 28 days from randomisation

Crystalloids group

• Participants: n = 3500; losses = study authors reported loss of 41 participants mostly because of withdrawal of surrogate consent; analysed = 3460

• Details: 0.9% NaCl; volume determined by treating clinicians

• Additional details: until discharge, death or 28 days from randomisation

Outcomes

Outcomes measured/reported: all‐cause mortality within 28 days, survival time during first 28 days, proportion of participants with organ failure, duration of mechanical ventilation, duration of renal‐replacement therapy, duration of ICU and hospital stay

Outcomes relevant to the review: mortality (28 days), RRT (for subgroup of participants with severe sepsis)

Notes

Funding/declarations of interest: Auckland District Health Board and the Health Research Council of New Zealand

Study dates: November 2001‐June 2003

Note: in the previous version of the review (Perel 2013), this study was called SAFE 2004.

This study reports a subgroup of participants who had severe sepsis (1218 participants; 603 in the albumin group, and 615 in the saline group). Data were available for RRT for these participants and we have included this subgroup of participants in analysis.

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

Randomisation was carried out centrally with the use of a minimisation algorithm; service accessed through a secure website

Allocation concealment (selection bias)

Low risk

Used central randomisation by a third party

Blinding of participants and personnel (performance bias): mortality

Low risk

Blinding was maintained by use of identical 500 mL bottles and cartons designed to mask fluid type and administration sets

Blinding of participants and personnel (performance bias): transfusion/renal replacement therapy/adverse events

Low risk

Blinding was maintained by use of identical 500 mL bottles and cartons designed to mask fluid type and administration sets

Blinding of outcome assessment (detection bias): mortality

Low risk

No details of blinding provided; unlikely to introduce bias for this outcome

Blinding of outcome assessment (detection bias): transfusion/renal replacement therapy/adverse events

Unclear risk

No details

Incomplete outcome data (attrition bias)
All outcomes

Low risk

Data (on vital status) were missing for 1% of randomised participants at 28 days, which is acceptable. Some discrepancies with reported numbers of participants analysed, but not significant

Selective reporting (reporting bias)

Low risk

Prospective clinical trials registration (ISRCTN76588266); all outcomes listed on registration site were reported

Baseline characteristics

Unclear risk

We noted that the albumin group had a higher CVP at baseline; we could not be certain whether this imbalance might influence results. No other baseline imbalances were noted

Other bias

High risk

Study authors reported that 3.9% of participants in the saline group were given albumin in the previous 72 h; this represents few participants and it is not likely to have introduced significant bias. However, some participants were given additional resuscitation fluids during the study period according to clinician preference, and numbers for this were not reported. This may influence outcome data for this study

Methods

RCT

Parallel design

Single centre

Participants

Total number of randomised participants: 79

Inclusion criteria: control of resuscitation obtained within 4 h of injury; all participants admitted within 12 h of injury

Exclusion criteria: none stated

Participant condition: burns

Baseline characteristics

Colloids group

• Age, mean (SD): 28 (± 7) years

Crystalloids group

• Age, mean (SD): 28 (± 8) years

Country: USA

Setting: Brooke Army Medical Center

Interventions

Colloids group

• Participants: n = 40; losses = 0; analysed = 40

• Details: 2.5% albumin RL; during the first 24 h, fluid was administered at a rate sufficient to stabilise vital signs and to produce a urinary output of 30 mL/h‐50 mL/h

• Additional details: plasma volume was replaced on the second postburn day by colloid equivalent to plasma in a dosage of 0.3 mL/kg body weight/% TBSA to 0.5 mL/kg body weight/% TBSA

Crystalloids group

• Participants: n = 39; losses = 0; analysed = 39

• Details: RL

• Additional details: same as colloids group

Outcomes

Outcomes measured/reported: haemodynamic responses; mortality at end of follow‐up

Outcomes relevant to the review: mortality at end of follow‐up

Notes

Funding/declarations of interest: study authors state, "the opinions or assertions contained herein are the private views of the authors and are not to be construed as official or as reflecting the views of the Department of the Army or the Department of Defense"

Study dates: not reported

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

Participants randomised using random numbers table

Allocation concealment (selection bias)

Unclear risk

No details

Blinding of participants and personnel (performance bias): mortality

Low risk

No details of blinding; unlikely to introduce bias for this outcome

Blinding of outcome assessment (detection bias): mortality

Low risk

No details of blinding; unlikely to introduce bias for this outcome

Incomplete outcome data (attrition bias)
All outcomes

Low risk

No losses reported

Selective reporting (reporting bias)

Unclear risk

No details of clinical trials registration; not feasible to assess risk of selective reporting bias

Baseline characteristics

Low risk

Baseline characteristics are comparable

Other bias

High risk

All participants in the crystalloid group received colloids after 24 h and this may have influenced study results

Methods

RCT

Parallel design

Single centre

Participants

Total number of randomised participants: 50

Inclusion criteria: trauma patients who met the criteria for haemorrhagic‐hypovolaemic shock, with definitive signs of external or internal haemorrhage in a prehospital setting; aged 18‐60 years

Exclusion criteria: no details

Participant condition: haemorrhagic‐hypovolaemic shock

Baseline characteristics

Colloids group

• No baseline characteristics reported

Crystalloids group

• No baseline characteristics reported

Country: Croatia

Setting: prehospital

Interventions

Colloids group

• Participants: n = 25; losses = 0; analysed = 25

• Details: 10% HES plus 7.5% NaCl solution; 4 mL/kg 7.5% NaCl followed by 500 mL HES

Crystalloids group

• Participants: n = 25; losses = 0; analysed = 25

• Details: 0.9% NaCl solution; 2000 mL 0.9% NaCl

Outcomes

Outcomes measured/reported: BP, pulse rate, peripheral oxygen saturation, and respiration rate

Outcomes relevant to the review: none

Notes

Funding/declarations of interest: none reported

Study dates: not reported

Methods

RCT

Parallel design

Multicentre

Participants

Total number of randomised participants: 196

Inclusion criteria: ≥ 18 years of age; required fluid resuscitation; clinically defined severe sepsis

Exclusion criteria: serum creatinine > 300 µmol/L; chronic renal failure; anuria lasting > 4 h; requirement for renal support

Participant condition: severe sepsis

Baseline characteristics

Colloids group

• Age, mean (SD): 65.8 (± 15.4) years

• Gender, M:F: 64:36

• SOFA, mean: 7.9

Crystalloids group

• Age, mean (SD): 65.9 (± 14.7) years

• Gender, M:F: 57:39

• SOFA, mean: 9.1

Country: France and Germany

Setting: hospital

Interventions

Colloids group

• Participants: n = 100; losses = 0; analysed = 100

• Details: 6% HES 130/0.4; maximum dose 50 mL/kg/day on day 1, then 25 mL/kg/day from day 2‐day 4; to ensure sufficient hydration, additional crystalloid infusions given in ratio of 1:2

Crystalloids group

• Participants: n = 96; losses = 1; analysed mortality = 95; analysed RRT = 96

• Details: 0.9% NaCl; maximum dose 50 mL/kg/day on day 1, then 25 mL/kg/day from day 2‐day 4; to ensure sufficient hydration, additional crystalloid infusions given in ratio of 1:2

Outcomes

Outcomes measured/reported: amount of study drug to achieve haemodynamic stabilisation; time to achieve initial haemodynamic stabilisation; quantity of study drug infused over 4 consecutive days; LoS in ICU and hospital; SOFA scores; kidney injury (RIFLE and AKIN scores); mortality (28 days and 90 days); blood transfusion; adverse events (itching)

Outcomes relevant to the review: mortality (28 days); blood transfusion (red blood cells); RRT (score of 3 using AKIN); adverse events (itching)

Notes

Funding/declarations of interest: supported by grant from Fresenius Kabi, Germany. The pharmaceutical company was involved in the study design, analysis and preparation of the report

Study dates: not reported

Note: the previous version of this review (Perel 2013) used mortality data at 90 days; in this review we have analysed mortality data at 28 days

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

Refers to reference from Myburgh 2012 to describe randomisation technique. Used external web‐based randomisation

Allocation concealment (selection bias)

Low risk

Use of web‐based system ensured that allocation code was kept concealed

Blinding of participants and personnel (performance bias): mortality

Low risk

Participants and personnel were blinded. Study drugs were kept in identical packaging

Blinding of participants and personnel (performance bias): transfusion/renal replacement therapy/adverse events

Low risk

Participants and personnel were blinded. Study drugs were kept in identical packaging

Blinding of outcome assessment (detection bias): mortality

Low risk

Reference from Myburgh 2012 suggests that all personnel were blinded, including outcome assessors

Blinding of outcome assessment (detection bias): transfusion/renal replacement therapy/adverse events

Low risk

Reference from Myburgh 2012 suggests that all personnel were blinded, including outcome assessors

Incomplete outcome data (attrition bias)
All outcomes

Low risk

One participant was unaccounted for in saline group for mortality outcome only, unlikely to influence outcome data overall

Selective reporting (reporting bias)

High risk

Prospective clinical trials registration (NCT00464204). Clinical trials registration documents do not list mortality, transfusion of blood products, or RRT as study outcomes. Clinical trials registration documents title of the study is "Effects of voluven on hemodynamics and tolerability of enteral nutrition in patients with severe sepsis" and some outcomes relate to assessment of caloric intake

Baseline characteristics

Low risk

Baseline characteristics appear comparable

Other bias

Low risk

No other sources of bias identified

Methods

RCT

Parallel design

Single centre

Participants

Total number of randomised participants: 172

Inclusion criteria: admitted during acute phase, with burns for which treatment for shock was indicated; adults and children

Exclusion criteria: no details

Participant condition: burns

Baseline characteristics

Colloids group

• Age, median (5% and 95% percentiles): 20 (1 and 71) years

• Weight, median (5% and 95% percentiles): 54 (10 and 85) kg

Crystalloids group

• Age, median (5% and 95% percentiles): 24 (1 and 66) years

• Weight, median (5% and 95% percentiles): 65 (11 and 90) kg

Country: Denmark

Setting: hospital

Interventions

Colloids group

• Participants: n = 86; losses = 0; analysed = 86

• Details: 6% dextran 70 in 0.9% NaCl; 120 mL/% TBSA; in first 48 h

• Additional details: plus metabolic water requirements (orally or IV); participants could drink freely during the shock phase

Crystalloids group

• Participants: n = 86; losses = 0; analysed = 86

• Details: RL; 4 mL RL/% TBSA/kg body weight in first 24 h; during next 24 h, indicator formula for fluid administration was 10% of body weight before the burn

• Additional details: same as colloid group

Outcomes

Outcomes measured/reported: fluid input and output, haemoglobin levels, mortality

Outcomes relevant to the review: mortality (48 h)

Notes

Funding/declarations of interest: supported by grant from Danish Medical Research Council

Study dates: not reported, the last participant was recruited in December 1975

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

Participants were stratified according to burn severity and type and then lots were used to determine which treatment the first participant in each stratum received

Allocation concealment (selection bias)

Unclear risk

No details

Blinding of participants and personnel (performance bias): mortality

Low risk

No details of blinding; unlikely to introduce bias for this outcome

Blinding of outcome assessment (detection bias): mortality

Low risk

No details of blinding; unlikely to introduce bias for this outcome

Incomplete outcome data (attrition bias)
All outcomes

Low risk

Data are reported for all randomised participants

Selective reporting (reporting bias)

Unclear risk

No details of clinical trials registration or pre‐published protocol; not feasible to assess risk of selective outcome reporting bias

Baseline characteristics

Low risk

Baseline characteristics appear comparable

Other bias

Low risk

No other sources of bias identified

Methods

RCT

Parallel design

Single centre

Participants

Total number of randomised participants: 19

Inclusion criteria: witnessed cardiac arrest with probable cardiac cause; advanced medical life support within 15 min; return of spontaneous circulation within 60 min; comatose when admitted to the hospital; aged 18‐80 years

Exclusion criteria: terminal illness; strongly in need of nursing; primary coagulopathy; prehospital fluid load > 2000 mL

Participant condition: postcardiac arrest

Baseline characteristics

Colloids group

• Age, median (range): 60 (48‐74) years

• Gender, M:F: 8:2

Crystalloids group

• Age, median (range): 60 (22‐75) years

• Gender, M:F: 8:1

Country: Norway

Setting: hospital

Interventions

Colloids group

• Participants: n = 10; losses = 0; analysed = 10

• Details: hypertonic colloid 7.2% NaCl with 6% HES 200/0.5 (volume ratios not reported); fluid given to achieve standardised treatment parameters

• Additional details: HS with HES limited to 500 mL/24 h (20 mL/h); further needs for fluid were met by Ringer's acetate/saline 9 mg/mL; all participants who returned to spontaneous circulation and remained unconscious were cooled to 33 °C using a Coolgard catheter.

Crystalloids group

• Participants: n = 9; losses = 0; analysed = 9

• Details: Ringer's acetate and saline 9 mg/mL; fluid given to achieve standardised treatment parameters

• Additional details: further needs for fluid were met by Ringer's acetate/saline 9 mg/mL; all participants who returned to spontaneous circulation and remained unconscious were cooled to 33 °C using a Coolgard catheter.

Outcomes

Outcomes measured/reported: fluid volume required to achieve treatment goals; oedema; haemodynamics; adverse events (to include renal failure); survival after 1 year

Outcomes relevant to the review: survival after 1 year

Notes

Funding/declarations of interest: supported by grant from the Regional Centre for Emergency Medical Research and Development and Development and Section of Emergency Medicine, Dept of Anaesthesia and Intensive Care, Haukeland University Hospital

Study dates: September 2005‐March 2007

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

Study authors report use of stratified randomisation, with allocation generated by study authors. We could not be certain whether this method was sufficient

Allocation concealment (selection bias)

Unclear risk

Numbered envelopes were distributed and opened by a physician after participant enrolment. Study authors do not report whether envelopes were sealed or opaque

Blinding of participants and personnel (performance bias): mortality

Low risk

No details of blinding of physician; unlikely to introduce bias for this outcome

Blinding of outcome assessment (detection bias): mortality

Low risk

No details of blinding; unlikely to introduce bias for this outcome

Incomplete outcome data (attrition bias)
All outcomes

Low risk

No losses

Selective reporting (reporting bias)

Unclear risk

No details of clinical trials registration; not feasible to assess risk of selective reporting bias

Baseline characteristics

Low risk

Baseline characteristics are comparable

Other bias

Low risk

No other sources of bias identified

Methods

RCT

Parallel design

Single centre

Participants

Total number of randomised participants: 115

Inclusion criteria: penetrating or blunt trauma; requiring > 3 L of volume resuscitation; 18‐60 years of age

Exlusion criteria: fluid overload pulmonary oedema; known allergy to HES; known pre‐existing renal failure with oliguria or anuria; receiving dialysis treatment before the injury; severe hypernatraemia or hyperchloraemia on admission; severe head injury from which recovery was unlikely; severe intracranial bleeding; severe crush injury; arterial pressure unresponsive to 2 L IV fluid loading which could not be recorded; clinically obvious cardiac tamponade; neurogenic shock (high spinal cord injury); known AIDS or AIDS‐related complex; admitted > 6 h after injury; people who had already received any colloid before randomisation; taking part in another clinical trial at the same time; refused consent

Participant condition: penetrating or blunt trauma

Baseline characteristics

Colloids group (penetrating trauma HES)

• Age, mean (range): 27.6 (18‐49) years

• Gender, M:F: 33:3

• Weight, mean (SD): 72.2 (± 7.6) kg

Crystalloids group (penetrating trauma saline)

• Age, mean (range): 32.6 (21‐56) years

• Gender, M:F: 27:4

• Weight, mean (SD): 77.4 (± 13.7) kg

Colloids group (blunt trauma HES)

• Age, mean (range): 33 (18‐50) years

• Gender, M:F: 15:5

• Weight, mean (SD): 76.8 (± 14.4) kg

Crystalloids group (blunt trauma saline)

• Age, mean (range): 35.7 (20‐58) years

• Gender, M:F: 15:7

• Weight, mean (SD): 78.8 (± 13.6) kg

Country: South Africa

Setting: hospital, level 1 trauma centre

Interventions

Colloids group (penetrating trauma HES + blunt trauma HES)

• Participants: randomised = 58; losses = 2 (prior colloids = 1; severe head injury = 1 (died)); analysed for mortality = 58; analysed for RRT = 56

• Details: 6% HES 130/0.4; given according to predetermined algorithm; resuscitation complete when haemodynamic and renal targets achieved and sustained

• Additional details: severely injured participants received a maximum of 2 L of crystalloids before randomisation; participants given adrenaline (epinephrine) for vasoactive support if required

Crystalloids group (penetrating trauma saline + blunt trauma saline)

• Participants: randomised = 57; losses = 4 (under age = 2; protocol violation = 1; unresponsive BP = 1 (died)); analysed for mortality = 57; analysed for RRT = 53

• Details: 0.9% NS; given according to predetermined algorithm; resuscitation complete when haemodynamic and renal targets achieved and sustained

• Additional details: same as colloid group

Outcomes

Outcomes measured/reported: volumes of study fluid in first 24 h; number of participants achieving normal gastrointestinal function by day 5; mortality; serious adverse events; acute renal injury; dialysis; use of blood products; biochemical abnormalities; days in ICU; days on ventilator support, SOFA scores, TEG measurements, skin itching

Outcomes relevant to the review: mortality (time point unknown), dialysis

Notes

Funding/declarations of interest: funding from Fresenius‐Kabi, who also supplied study fluids. Funders had no input into study design, analysis, interpretation etc. Also funds from TEG and laboratory investigations derived from Dept of Anaesthesia, UCT, research funds

Study dates: not reported

Study authors stratified data according to whether participants had penetrating or blunt trauma injuries. We have combined both types of injuries in analysis

Note: we used mortality data reported in the previous version of this review (Perel 2013). These data were collected from personal communication with the study authors

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

Used random numbers in blocks of 8 for each category of trauma

Allocation concealment (selection bias)

Low risk

Fluids prepacked by pharmacy, and we have assumed that, therefore, allocation was concealed from personnel

Blinding of participants and personnel (performance bias): mortality

Low risk

Study fluids were presented in identical black bags

Blinding of participants and personnel (performance bias): transfusion/renal replacement therapy/adverse events

Low risk

Study fluids were presented in identical black bags

Blinding of outcome assessment (detection bias): mortality

Low risk

No details of blinding; unlikely to introduce bias for this outcome

Blinding of outcome assessment (detection bias): transfusion/renal replacement therapy/adverse events

Unclear risk

No details

Incomplete outcome data (attrition bias)
All outcomes

Low risk

Few losses, and reasons were reported by study authors

Selective reporting (reporting bias)

Unclear risk

Retrospective clinical trials registration (ISRCTN 42061860); so not feasible to assess risk of selective reporting bias from these documents

Baseline characteristics

Unclear risk

Injury severity scores were higher in the colloids group. We could not be certain whether this could influence outcome data

Other bias

Low risk

No other sources of bias identified

Methods

RCT

Parallel design

Single centre

Participants

Total number of randomised participants: 19

Inclusion criteria: 20%‐98% TBSA; selected when, within 15 min, precise time of burn injury and intake and output experienced by patient from time of injury to time of admission was known

Exclusion criteria: no details

Participant condition: burns

Baseline characteristics

Colloids group

• Age, mean (SE): 47 (± 5.6) years

• Weight, mean (SE): 97 (± 7.4) kg

Crystalloids group (RL)

• Age, mean (SE): 34 (± 5.3) years

• Weight, mean (SE): 83 (± 1.4) kg

Crystalloids group (HS)

• Age, mean (SE): 52 (± 12.7) years

• Weight, mean (SE): 72 (± 6.1) kg

Country: USA

Setting: hospital

Interventions

Colloids group

• Participants: n = 7; losses = 0; analysed = 7

• Details: hypertonic solution with albumin; hypertonic solution ‐ 240 mEq sodium and 120 mEq each of chloride and lactate; 12.5 g albumin added to each litre; to maintain MAP ≥ 60 to ≤ 110 mmHg with a urine flow of 30 mL/h‐50 mL/h

• Additional details: resuscitation complete when MAP stable at 70 mmHg‐110 mmHg; urine output stable at 40 mL/h‐50 mL/h; lactic acid was ≤ 2 mg or fluid needs could be met by mouth; absolute BP and pulse rate were not criteria of concern for this group

Crystalloids group (RL)

• Participants: n = 7; losses = 0; analysed = 7

• Details: RL; to maintain MAP ≥ 60 to ≤ 110 mmHg with a urine flow of 30 mL/h‐50 mL/h

• Additional details: resuscitation complete when urine flow of at least 40 mL/h; pulse rate ≤ 110/min and elevation of SBP and DBP into premorbid normal range for participant

Crystalloids group (HS)

• Participants: n = 5; losses = 0; analysed = 5

• Details: 240 mEq Na 120 mEq Cl; to maintain MAP ≥ 60 mmHg to ≤ 110 mmHg with a urine flow of 30 mL/h‐50 mL/h

• Additional details: resuscitation complete when MAP stable at 70 mmHg‐110 mmHg; urine output stable at 40 mL/h‐50 mL/h; lactic acid was ≤ 2 mg or fluid needs could be met by mouth

Outcomes

Outcomes measured/reported: fluid volume; clinical results; laboratory results; urine variables (including renal failure); serum osmolality; sodium and potassium levels; cardiorespiratory and haemodynamic variables

Outcomes relevant to the review: mortality (time point not reported)

Notes

Funding/declarations of interest: supported in part by National Institutes of Health Grant

Study dates: January 1977‐March 1978

In the previous version of the review (Perel 2013), the study ID was Jelenko 1978

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

High risk

Described as randomised. No additional details but significant details in baseline demographics which would suggest an insufficient method of randomisation

Allocation concealment (selection bias)

Unclear risk

No details

Blinding of participants and personnel (performance bias): mortality

Low risk

No details; lack of blinding unlikely to influence data for this outcome

Blinding of outcome assessment (detection bias): mortality

Low risk

No details; lack of blinding unlikely to influence data for this outcome

Incomplete outcome data (attrition bias)
All outcomes

Low risk

No apparent losses

Selective reporting (reporting bias)

Unclear risk

No details of clinical trials registration or prepublished protocol; not feasible to assess risk of selective outcome reporting bias

Baseline characteristics

High risk

Statistically significant differences between groups for baseline characteristics

Other bias

Low risk

No other sources of bias identified

Methods

RCT

Parallel design

Single centre

Participants

Total number of randomised participants: 84

Inclusion criteria: 18‐85 years of age; meet criteria for septic shock; resuscitation within 6 h with crystalloid or HES ≥ 30 mL/kg; within 24 h no packed red blood cells, plasma or other blood products that would affect coagulation and fibrinolysis significantly; no unauthorised drugs; no previous coagulation disorders

Exclusion criteria: severe heart failure; bleeding occurring during resuscitation and requiring the use of blood products; serious renal insufficiency

Participant condition: septic shock

Baseline characteristics

Colloids group

• Age, mean (SD): 65.7 (± 15.1) years

• Gender, M:F: 16:28

• Weight, mean (SD): 65.9 (±12.0) kg

• APACHE II, mean (SD): 27.9 (± 5.9)

Crystalloids group

• Age, mean (SD): 64.7 (± 13.7) years

• Gender, M:F: 14:26

• Weight, mean (SD): 66.6 (± 11.3) kg

• APACHE II, mean (SD): 25.3 (± 4.5)

Country: China

Setting: hospital

Interventions

Colloids group

• Participants: randomised = 44; losses = 0; analysed = 44

• Details: HES 130/0.4

Crystalloids group

• Participants: randomised = 40; losses = 0; analysed = 40

• Details: RL

Outcomes

Outcomes measured/reported: prothrombin time, tissue factor, tissue factor pathway inhibitor, active protein C, LoS in ICU, mortality

Outcomes relevant to the review: mortality (time point unknown)

Notes

Funding/declarations of interest: none reported

Study dates: November 2009‐October 2014

Article in Chinese. Data for study characteristics taken from English abstract, and from study report tables, with translation using Google Translate

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

Described as randomised. Data for 'Risk of bias' assessment taken from English abstract only

Allocation concealment (selection bias)

Unclear risk

No details. Data for 'Risk of bias' assessment taken from English abstract only

Blinding of participants and personnel (performance bias): mortality

Low risk

No details of blinding; unlikely to introduce bias for this outcome

Blinding of outcome assessment (detection bias): mortality

Low risk

No details of blinding; unlikely to introduce bias for this outcome

Incomplete outcome data (attrition bias)
All outcomes

Low risk

No apparent losses

Selective reporting (reporting bias)

Unclear risk

Data for 'Risk of bias' assessment taken from English abstract only. No details of clinical trials registration in English abstract

Baseline characteristics

Low risk

Baseline characteristics appeared largely comparable

Other bias

Unclear risk

We could not be certain of other risks of bias because 'Risk of bias' assessments were made from the English abstract only

Methods

RCT

Parallel design

Single centre

Participants

Total number of randomised participants: 105

Inclusion criteria: perforation peritonitis; 18‐60 years of age

Exclusion criteria: pregnancy; known allergies or manifesting symptoms of possible anaphylaxis with test dose of HES; major coagulation disorders; renal failure because of medical renal disease; severe hepatic insufficiency; congestive cardiac failure at admission; traumatic perforation cases; < 18 years of age or > 60 years of age; people who had been resuscitated before reaching emergency surgical unit; denied consent

Participant condition: perforation peritonitis

Baseline characteristics

Colloids group

• Age, mean (SD): 35.75 (± 11.84) years

• Gender, M:F: 50:5

• Physiological score, mean (SD): 27.73 (± 7.50)

Crystalloids group

• Age, mean (SD): 33.44 (± 13.08) years

• Gender, M:F: 47:5

• Physiological score, mean (SD): 18.33 (± 7.37)

Country: India

Setting: hospital

Interventions

Colloids group

• Participants: n = 55; losses = 0; analysed = 55

• Details: 6% HES 130/0.4; at a rate of 15 mL/kg body weight/h; up to a total dose of 30 mL/kg body weight; after presentation for surgery, before start of emergency laparotomy

• Additional details: test dose 10 mL‐20 mL HES given slowly whilst observing for possible anaphylactic response; participants who had anuria or oliguria were given 1 L crystalloids IV within 30‐60 min to improve urine output; if urine output did not improve, participants were given 40 mg furosemide, and if this did not improve urine output then participants were excluded; also given crystalloids as required

Crystalloids group

• Participants: n = 52; losses = 0; analysed = 52

• Details: RL; amount and rate determined by participant condition

• Additional details: participants who had anuria or oliguria were given 1 L crystalloids IV within 30‐60 min to improve urine output; if urine output did not improve, participants were given 40 mg furosemide, and if this did not improve urine output then participants were excluded; also given crystalloids as required

Outcomes

Outcomes measured/reported: time to achieve goals of fluid resuscitation, morbidity, mortality, length of hospital stay, complications attributable to type of fluid administration

Outcomes relevant to the review: mortality (up to 30 days from hospital discharge)

Notes

Funding/declarations of interest: no funding and no conflicts of interest

Study dates: October 2006‐April 2009

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

Quote: "Patients were randomised with the help of computer‐generated random table"

Allocation concealment (selection bias)

Unclear risk

No details

Blinding of participants and personnel (performance bias): mortality

Low risk

No details of blinding; unlikely to introduce bias for this outcome

Blinding of outcome assessment (detection bias): mortality

Low risk

Quote: "Administered the fluid therapy according to randomisation without knowledge of the observer"

Incomplete outcome data (attrition bias)
All outcomes

Low risk

No apparent losses

Selective reporting (reporting bias)

Unclear risk

No details of clinical trials registration or prepublished protocol; not feasible to assess risk of selective outcome reporting bias

Baseline characteristics

Unclear risk

We noted differences in physiological scores between groups. We could not be certain whether this difference could influence the outcome data

Other bias

High risk

Note the length of time since completion of trial, and publication of full study report. Also, note that the study was reported by a single author

Methods

RCT

Parallel design

Single centre

Participants

Total number of randomised participants: 60

Inclusion criteria: not reported in abstract

Exclusion criteria: not reported in abstract

Participant condition: patients with septic shock

Baseline characteristics

Colloids group (HES)

• Age, mean (SD): 44.8 (± 23.7) years

• Gender, M:F: 10:5

• BP, mean (SD): SBP: 78.33 (± 10.03) mmHg; DBP: 47.87 (± 8.84) mmHg

Colloids group (HES with HS)

• Age, mean (SD): 46.0 (± 22.2) years

• Gender, M:F: 10:5

• BP, mean (SD): SBP: 78.80 (± 8.94) mmHg; DBP: 43.53 (± 6.35) mmHg

Crystalloids group (NS)

• Age, mean (SD): 38.6 (± 19.5) years

• Gender, M:F: 11:4

• BP, mean (SD): SBP: 80.93 (± 4.35) mmHg; DBP: 40.93 (± 6.22) mmHg

Crystalloids group (HS)

• Age, mean (SD): 50.2 (± 28.4) years

• Gender, M:F: 10:5

• BP, mean (SD): SBP: 79.60 (± 5.41) mmHg; DBP: 42.00 (± 4.42) mmHg

Country: China

Setting: hospital

Interventions

Colloids group (HES)

• Participants: n = 15; losses = 0; analysed = 15

• Details: HES

• Additional details: no additional details in abstract

Colloids group (HES with HS)

• Participants: n = 15; losses = 0; analysed = 15

• Details: hypertonic sodium chloride HES 40 solution

• Additional details: no additional details in abstract

Crystalloids group (NS)

• Participants: n = 15; losses = 0; analysed = 15

• Details: NS

• Additional details: no additional details in abstract

Crystalloids group (HS)

• Participants: n = 15; losses = 0; analysed = 15

• Details: 4% NaCl

• Additional details: no additional details in abstract

Outcomes

Outcomes measured/reported: haemodynamic parameters, blood lactate clearance, mortality (at 28 days)

Outcomes relevant to the review: mortality

Notes

Funding/declarations of interest: not reported in abstract

Study dates: not reported in abstract

Article in Chinese. Data for study characteristics taken from English abstract, and from study report tables, with translation using Google Translate

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

'Risk of bias' assessment made using English abstract only. Described as randomised, no additional detail

Allocation concealment (selection bias)

Unclear risk

No details. 'Risk of bias' assessment made using English abstract only

Blinding of participants and personnel (performance bias): mortality

Low risk

No details. 'Risk of bias' assessment made using English abstract only. However, lack of blinding unlikely to introduce bias for mortality

Blinding of outcome assessment (detection bias): mortality

Low risk

No details. 'Risk of bias' assessment made using English abstract only. However, lack of blinding unlikely to introduce bias for mortality

Incomplete outcome data (attrition bias)
All outcomes

Low risk

No apparent losses

Selective reporting (reporting bias)

Unclear risk

No details of clinical trials registration or prepublished protocol; not feasible to assess risk of selective outcome reporting bias. 'Risk of bias' assessment made using English abstract only

Baseline characteristics

Low risk

Baseline characteristics appeared largely comparable

Other bias

Unclear risk

We could not be certain about other risks of bias because 'Risk of bias' assessment were made using English abstract only

Methods

RCT

Parallel design