A Comparison of Albumin and Saline for Fluid Resuscitation in the Intensive Care Unit

1
A Comparison of Albumin and Saline for Fluid
Resuscitation in the Intensive Care Unit
The SAFE Study Investigators The New England
Journal of Medicine, May 27, 2004. 350222247-225
6
2
Methods - Study Design and Treatment Protocol

• Patient selection
• gt18 yrs old
• ICU in 16 hospitals in Australia and NZ
• Between Nov 2001 and June 2003
• Required fluid administration
• Excluded post-op cardiac surgery, liver
  transplant or burns
• Intervention
• Randomly assigned to receive either 4 Albumin or
  N/S

3
Method Non-study fluids

• Patients also received
• Maintenance fluids
• Specific replacement fluids
• Enteral/parental nutrition
• Blood products

4
Design

• () Double-blinded
• () Randomised groups
• () Groups had similar baseline characteristics
• () Prospective trial
• () Multi-centre
• () Large study population
• (-) Multidisciplinary ICU patient type not
  defined

5
Method - Primary Outcome Measures

• Death from any cause within 28 days
• Subgroups
• Presence/Absence of trauma
• Presence/Absence of severe sepsis
• Presence/Absence of Acute Respiratory Distress
  Syndrome

6
Method - Primary Outcome Measures

• Proportion with 1-5 new organ failures
• Cardiovascular
• Respiratory
• Renal
• Haematologic
• Hepatic
• Duration of mechanical ventilation
• Duration of renal-replacement therapy
• Duration of ICU and hospital stay

7
Results

• 6997 patients
• 3497 Albumin
• 3500 Saline
• () Large power study
• (-) Three patients underwent randomisation twice
• Counted in group allocated first

8

• Consent
• Delayed from patient or surrogate
• Prior consent from surrogate
• Outcomes not available in 67 patients (1 of
  patients)
• 56 out of 67 due to withheld or withdrawn consent

9

• The only statistically significant difference at
  baseline
• Mean CVP - higher in the albumin group

10
Results - Parameters

• Study fluids
• Albumin group received less than saline group
• Non-study fluids
• Day 12 albumin group received more packed
  cells than saline
• Day 2 saline group received more non-study
  fluids
• Day 4 - no further differences in volume of
  study fluids

11

• Study fluids
• Not administed to 197 patients (2.8 of patients)
• Non-study resuscitation fluids
• Most commonly due to error
• Clinicians preference

12

• No significant difference in mean arterial
  pressure
• Albumin group had lower heart rate at Day 1
• Albumin group had significantly higher CVP
  throughout D1-4
• Serum albumin higher than albumin group

13
Results - Outcomes

• Death after 28 days
• 20.9 of albumin group
• 21.1 of saline group
• Absolute difference in mortality was -0.2
• Remaining in ICU after 28 days
• 3.2 of albumin group
• 2.5 of saline group
• Remaining in Hospital after 28 days
• 22.8 of albumin group
• 24.5 of saline group
• Survival time after 28 days
• No significant difference between groups

14

• Mean length of stay in ICU
• 6.5 /- 6.6 days for albumin group
• 6.2 /- 6.2 days for saline group
• Mean length of stay in Hospital
• 15.3 /- 9.6 days for albumin group
• 15.6 /- 9.6 days for saline group
• M.O.F. similar number

15
Results Subgroup Analysis

• Presence of Trauma
• Death relative risk 1.36
• Absence of Trauma
• Deaths relative risk 0.96
• Why?
• Greater number of trauma patients also had
  associated brain injury in albumin group

16
Results Subgroup Analysis

• Presence of Trauma WITH Brain injury
• Deaths relative risk 1.62
• 24.5 in albumin group
• 15.1 in saline group
• Presence of Trauma WITHOUT Brain injury
• Deaths relative risk 1.00 (no difference)

17
Results Subgroup Analysis

• Presence of Sepsis
• Deaths relative risk 0.87
• 30.7 in Albumin group
• 35/3 in Saline group
• Absence of Sepsis
• Deaths relative risk 1.05

18
Results Subgroup Analysis

• Presence of ARDS
• Death relative risk 0.93
• Absence of ARDS
• Death relative risk 1.00

19
Discussion

• No difference in
• Death from any cause
• Requirements for mechanical ventilation
• Requirements for renal-replacement
• Time spent in ICU
• Time spent in Hospital
• Time until death
• Similar rates of new single/multiple organ
  failures

20

• Albumin group were resuscitated with less fluid
  than the saline group
• But no difference in MAP
• Only small difference in CVP and HR

21

• Albumin group receive greater volume of packed
  cells
• Why?
• ? Greater haemodilution with albumin
• ? Greater blood loss due to transient alterations
  in coagulation with albumin
• Is this the cause of higher mortality with
  albumin?
• Herbert et al 3U packed cells associated with
  higher mortality
• Albumin group had 71ml extra fluid (1/4U)
  enough for an effect??

22

• ?better result for trauma patients with saline
• Choi colloid associated with increased
  mortality in patients with trauma
• Trauma with brain injury
• But
• only 7 of study group had brain injury with only
  21 more deaths in albumin group than saline
  ?chance result
• For brain injury, should not measure 28 day death
  rate, but rather mortality and functional
  neurological status at 6 months

23

• Patients with sepsis
• Differences could be chance needs further study

24

• Patients had differing concurrent interventions
  across institutions
• ? But randomised within institution

25
Conclusion

• Albumin and saline are clinically equivalent
  treatments for intravascular volume replacement
  in a heterogeneous population of patients in the
  ICU