Safety of Albumin Revisited

1
Safety of Albumin Revisited

• Blood Products Advisory Committee Meeting
• March 17, 2005
• Laurence Landow MD, FRCPC

2
Agenda

• Introduction and background
• Laurence Landow MD
• Review of the Cochrane Report
• Paul Hebert MD, Vice-Chair of Research, Ottawa
  Health Research Institute, Ontario Canada
• Review of the SAFE study
• Simon Finfer MD, Senior Staff Specialist in
  Intensive Care, University of Sydney, Australia

3
Human albumin administration in critically ill
patients systematic review of randomized
controlled trialsCochrane Injuries Group. BMJ
1998317235-40

• Design Systematic review of all randomized
  controlled trials published by March 1998
  comparing administration of albumin with no
  administration or with crystalloid in ICU
  patients with hypovolemia, burns, or
  hypoalbuminemia
• Primary Endpoint All-cause mortality
• Conclusions There is no evidence that albumin
  reduces mortality and a strong suggestion that it
  may increase mortality
• Aggregate relative risk 1.68 (1.26, 2.23)
• Hypovolemia 1.46 (0.97,2.22)
• Burns 2.40 (1.11, 5.19)
• Hypoproteinemia 1.69 (1.07, 2.67)

4
FDA Letter to Healthcare Providers
August 19, 1998

• It is FDAs current view that the BMJ
  meta-analysis warrants serious consideration.
• FDA encourages additional controlled trials on
  the use of albumin
• Until the results of furtherstudies are
  available, the FDA urges treating physicians to
  exercise discretion in use of albumin based on
  their own assessment of these data.

5
Patient survival after human albumin
administration
Wilkes Navickis. Ann Intern Med 2001135149-164

• Design Systematic review of randomized
  controlled trials comparing administration of
  albumin with crystalloid, no albumin, or lower
  doses of albumin in ICU patients with trauma,
  hypovolemia, burns, hypoalbuminemia, ascites, and
  high-risk neonates
• Primary Endpoint All-cause mortality
• Conclusions No effect on mortality was
  detectedThis finding supports the safety of
  albumin
• Aggregate relative risk 1.11 (0.95, 1.28)
• Cochrane meta-analysis 1.68 (1.26, 2.23)
• Hypovolemia 1.59 (0.91, 2.78)
• Cochrane meta-analysis 2.40 (1.07, 2.67)
• Burns 1.76 (0.97, 3.17)
• Cochrane meta-analysis 2.40 (1.11, 5.19)
• Hypoproteinemia 1.59 (0.91, 2.78)
• Cochrane meta-analysis 1.69 (1.07, 2.67)

6
Colloid use for fluid resuscitation evidence
and spin
Cook Guyatt. Ann Intern Med 2001135205-8

• Wilkes and Navickis conclude that their findings
  should serve to allay concerns regarding the
  safety of albumin
• But these results are reassuring only insofar as
  they fail to show a statistically significant
  increase in mortality. In each case, the point
  estimate - the best estimate of the true effect
  of treatment - shows an increase in the relative
  risk for death of more than 10 overall
• Confidence intervals estimate the range within
  which the true effect plausibly lies. These
  confidence intervals indicate a relative overall
  increase in mortality of 28 (aggregate relative
  risk 1.11 (0.95, 1.28)
• Point estimates that suggest harm and confidence
  intervals that include important increases in
  mortality cannot allay concerns about the
  potentially harmful effects of albumin

7
A comparison of albumin and saline for fluid
resuscitation in the Intensive Care Unit
SAFE study investigators. N Engl J Med
20043502247-56

• Design Randomized, double-blind trial comparing
  resuscitation with albumin or saline in a
  heterogeneous population of hypovolemic ICU
  patients
• Subjects with burns, cardiac surgery, or liver
  transplantation were excluded
• Stratified randomization at baseline for trauma,
  severe sepsis, and ARDS
• Primary Endpoint All-cause mortality
• Conclusions Use of either albumin or normal
  saline results in similar outcomes at 28 days
• Aggregate relative risk 0.99 (0.91, 1.09)
• Cochrane meta-analysis 1.68 (1.26, 2.23)
• Wilkes and Navickis meta-analysis 1.11 (0.95,
  1.28)
• Trauma 1.36 (0.99, 1.86)
• Trauma TBI 1.62 (1.12, 2.34)
• Trauma without TBI 1.00 (0.56, 1.79)
• Severe sepsis 0.87 (0.74-1.02)
• ARDS 0.93 (0.61, 1.41)

8
Letter to FDA from Plasma Protein Therapeutics
Association
21-DEC-2004

• These SAFE study data prove clinically that
  albumin is a safe therapy and clearly refute the
  findings of a meta-analysis from the Cochrane
  Collaboration which was the subject of a 1998
  Dear Doctor letter from the Food and Drug
  Administration (FDA).
• A subsequent meta-analysis of Wilkes and
  Navickis could not replicate the finding of
  excess albumin-associated mortality reported by
  the Cochrane investigators. In fact, the results
  of the SAFE trial corroborate the conclusions of
  this second and more rigorous meta-analysis.

9
Question for the Committee

• 1. Have data from the SAFE study resolved the
  safety concerns that were raised in the
  meta-analysis by the Cochrane Group for
• a. critically ill patients in general?
• b. subgroups of critically ill patients with
  burns, hypovolemia, or hypoproteinemia?