FACULTY OF MEDICINE, DENTISTRY AND HEALTH SCIENCES
Choice of Colloid Fluid Therapy in Septic Patients – A
Comparative Cost-Effectiveness Analysis of
Crystalloid, Albumin and Hydroxyethyl Starch
Albert Farrugia, PhD
Megha Bansal, MA Econ
ESICM LIVES 2013
26th Annual Congress
Paris, October 5-9, 2013
Background (1)
Current concepts of fluid exchange
Annu. Rev. Biomed. Eng. 2007. 9:121–67
J Physiol 557.3 (3004) p704
The University of Western Australia
Background (2)
Severe Sepsis
1800
1600
Severe sepsis cases
1400
1200
1000
US population
800
600
2001
2025
2050
600
550
500
450
400
350
300
250
300
250
200
150
Incidence
Mortality
100
50
0
Severe
Sepsis
CVA
Breast
CA
Lung
CA
• More than 750,000 cases/Y of severe sepsis in US
• Estimated annual healthcare costs due to severe sepsis in
$16 billion
U.S. exceed $
• Leading cause of death in non-coronary ICU
The University of Western Australia
Microcirculatory perfusion in sepsis
Potentially impairing factors
Spronk et al Critical Care 2004, 8:462-468
The University of Western Australia
Background (3)
Fluid Resuscitation of Shock
Crystalloid Solutions
Colloid Solutions
(universally used for
i iti l volume
initial
l
resuscitation)
(achieve hemodynamic goals
more quickly
i kl with
ith
significantly less volume)
• “Normal”
Normal saline
• “Balanced” solutions
– Ringers Lactate
solution
– Plasmalyte
• Hydroxyethyl
Starch(es)
• Blood products
(albumin, RBC,
plasma)
As sepsis
A
i proceeds,
d …significant
i ifi
t titissue accumulation
l ti off
resuscitation fluid occurs, resulting in adverse effects..
The University of Western Australia
Which Fluids?
 An ideal resuscitative fluid would maintain
intravascular volume without expanding the
interstitial space.
 Crystalloid solutions are universally used for initial
volume resuscitation in sepsis and septic
shock…...
 Colloid solutions achieve hemodynamic goals
more quickly than crystalloids with significantly
less volume.
 As sepsis proceeds,…..significant tissue
accumulation
l ti off resuscitation
it ti flfluid
id occurs, and
d thi
this
Elsevier 2010
may result in adverse effects..
The University of Western Australia
Fluid Therapy in Septic Patients
Albumin is retained intravascularly…
intravascularly
SAFE investigators
i
ti t
I t Care
Int
C
M
Med
d 2011
2011, 37
(1), 86-96
…and
and expands plasma volume
Ernest et al Critical Care Medicine:Volume
27(1)January 1999 pp 46-50
The University of Western Australia
Methods (1)
Indirect comparisons
Network meta
meta-analysis
analysis / mixed treatment comparisons
 Evidence from head to head comparison trials is often limited or unavailable
 Using
U i evidence
id
ffrom A vs B and
d A vs C ttrials
i l tto d
draw conclusions
l i
about
b t th
the
effect of B relative to C
 It allows all evidence to be combined in a single analysis
• Inference based on more evidence can (usually) provide more precision
• Treatments can be ranked
• Better informed decisions
Treatment B
(albumin)
Treatment A
(crystalloid)
Treatment C
(HES)
The University of Western Australia
Network MA for studies in sepsis
Search strategy
391 studies identified
through database
search and all other
sources
Crystalloid
6
336 excluded as not
relevant in first
screening
13 studies included in
network meta-analysis
55 full text articles were
reviewed in second
screening
Excluded: 42
Not relevant comparator
n=6
Not sepsis
Not a randomized clinical trial
No mortality results
Duplicates/Repeats
n=11
n=3
n=19
n=3
5
1
Albumin
1
HES
Bansal et al 2013
The University of Western Australia
Methods (2)
Decision tree to assess cost-effectiveness
Fluids in sepsis
The University of Western Australia
Variables used to populate the decision
analysis model
Base case
One-way sensitivity
value
analysis values
Cost of albumin US$ (cAlb)
270
250 – 1,000
1 000
Not applied
Cost of hydroxyethyl starch US$ (cHES)
269
±20%
Not applied
Cost of sepsis standard of care US$ (cSepsisGen)
20,133
±20%
Gamma~ (55.56, 0.003)
Cost of renal replacement therapy US$ (cRenal)
142,404
76,540 – 30,616
Normal~ (142404, 146792)
Cost of treatment for bleeding US$ (cBleeding)
1,732
1,044 – 2,366
Normal~ (1732, 705.6)
Life expectancy – general population at 65 years (LEgenpop)
18.60
Not applied
Not applied
In-hospital or 28 day mortality with crystalloid (pDeadSep)
33.27%
Not applied
Beta~ (16.52, 33.14)
In-hospital or 28 day mortality with albumin (pDeadAlb)
30.95%
Not applied
Beta~ (14.78, 32.98)
In-hospital or 28 day mortality with hydroxyethyl starch (pDeadHES)
38.54%
Not applied
Beta~ (20.45, 32.60)
90-day excess mortality with hydroxyethyl starch (pDeadHES90)
13 6%
13.6%
Not applied
Not applied
6.5%
3.5% - 19.5%
Uniform~(0.035, 0.195)
Excess p
probability
y of bleeding
g with hydroxyethyl
y
y y starch (p
(pBleeding)
g)
3.29%
Not applied
pp
Not applied
pp
Prob. of mortality in bleeding episodes (pDeadBleed)
7.3%
Not applied
Beta~ (45.6, 579.1)
Prob. of mortality after RRT (pDeadRenal)
54.1%
50.8% - 60.8%
Uniform (0.508, 0.608)
Variable (Abbreviation in the model)
Probability Distribution
Excess probability of renal replacement therapy with hydroxyethyl
starch (pRenalHES)
The University of Western Australia
Results (1)
Network MA for studies in sepsis
Ranking probabilities of competing fluid treatments
%p
probability to ran
nk at each place
100.00%
80.00%
60.00%
Albumin
40.00%
Crystalloid
HES
20.00%
0.00%
Albumin
Crystalloid
HES
Rank 1
96.38%
3 61%
3.61%
0.01%
Rank 2
3.60%
96 27%
96.27%
0.00%
Rank 3
0.00%
0 00%
0.00%
99.87%
Bansal et al 2013
The University of Western Australia
Results (2)
Results of cost-effectiveness analysis
Total Medical
Cost
Effectiveness *
Crystalloid
$18,199
Reference
Albumin
$18,469
0.23
HES
$24,196
-0.45
* Years saved/lost compared
p
to crystalloid.
y
The University of Western Australia
Results of One-way Sensitivity Analysis
Variable
Cost of Renal
Replacement
Therapy
Cost of Albumin
Cost of treatment
of bleeds
Range ($)
Fluid Treatment
Total Cost/Life
Year (Low) $
Total Cost/Life
Year (High) $
Crystalloid
9,086
9,086
Albumin
8,259
,
8,259
,
Hydroxyethyl Starch
13,775
19,639
Crystalloid
y
9,086
9,086
Albumin
8,259
8,259
Hydroxyethyl Starch
15,457
15,457
Crystalloid
9,086
9,086
Albumin
8,259
8,259
Hydroxyethyl Starch
15,450
15,470
76,540 – 306,160
250 – 1,000
1,193 – 2,693
The University of Western Australia
Perspectives
Regulatory Actions - USFDA
• Do not use HES solutions in critically ill adult patients including those with sepsis,
and
d th
those admitted
d itt d tto the
th ICU.
ICU
• Avoid use in patients with pre-existing renal dysfunction.
• Discontinue use of HES at the first sign of renal injury.
• Need for renal replacement
p
therapy
py has been reported
p
up
p to 90 days
y after HES
administration.
• Continue to monitor renal function for at least 90 days in all patients.
• Avoid use in patients undergoing open heart surgery in association with
cardiopulmonary bypass due to excess bleeding.
bleeding
• Discontinue use of HES at the first sign of coagulopathy.
The University of Western Australia
Regulatory Action - EMA
The University of Western Australia
Summary/Conclusions
 Colloids are required after initial crystalloid infusion
 Albumin has been shown to improve outcomes in sepsis
 HES has been shown to increase mortality and renal damage
 Network meta-analysis and Decision Modeling allow direct
comparison between fluid therapies
 In decision-making, a focus on total, long term costs needs to
supersede short term considerations of individual interventions
Disclosure:
The authors provide services to the pharmaceutical and biotechnology industry,
including the manufacturers of therapies described in this presentation.
The University of Western Australia