Strategies for initiating RRT in AKI

Strategies for initiating RRT in AKI
Stéphane Gaudry
Réanimation médico-chirurgicale
Hôpital Louis Mourier, Colombes
Sorbonne-Paris-Cité University
Conflict of interest
Educational grants from Xenios France
No conflict of interest regarding RRT
AKIKI study funded by a grant from French
Ministry of Health
Invention of the artificial kidney
Willem Johan "Pim" Kolff
February 14, 1911, Leiden
February 11, 2009
•
The father of artificial organs
– Hemodialysis, first life saving (1945)
– Membrane oxygenators (1948)
– Artificial heart (with Jarvik) (1957)
•
He immigrated to USA (1950)
•
Albert Lasker award (2002)
•
He was recognized Righteous Among the
Nations by Yad Vashem Jerusalem center (2012)
for hiding jews during WW2
Kolff WJ
First clinical experience with artificial kidney
Ann Intern Med 62: 608-619;
1965
Reversible renal failure
?
Reversibility
Rapid
(hours)
Slow
(-- weeks)
Kidney International
1972
Control
Prophylactic
RRT
Early RRT initiation
• Pros:
– Better fluid balance control
– Improved acid-base status
– Better control of electrolytes abnormalities
– Removal of toxins or cytokines in sepsis ??
• Cons:
– Potential harm (catheter, hypotension, metabolic)
– Costs
Available data on the timing of RRT
Observational studies
Randomized controlled trials
RCTs
N=270
Cohort studies
1961
1963
1963
1966
1965
1970
1971
1972
N=2118
Early RRT
Delayed RRT
Paul M. Palevsky
Crit Care Med 2008
Patients with AKI
Patients managed with RRT
Early RRT
initiation
Late RRT
initiation
No
RRT
Paul M. Palevsky
Crit Care Med 2008
Patients with AKI
Patients managed with RRT
Early RRT
initiation
Late RRT
initiation
Observational studies
No
RRT
Paul M. Palevsky
Crit Care Med 2008
study design to adequately answer
clinical question of timing of RRT in AKI
Patients with AKI
Early RRT strategy
Receive RRT
Late RRT strategy
Receive RRT
No RRT
When to start ?
ICM March 2016
When to start ?
Why start it ?
Criteria ?
ICM March 2016
To compare strategies
we need RCTs
ICM March 2016
RCTs on the timing of RRT
• Bouman et al
• Jamale et al
• Zarbock et al (ELAIN)
• Wald et al (Pilot STARRT-AKI)
• Combes et al (HEROICS)
• Gaudry et al (AKIKI)
RCTs on the timing of RRT
Patients
n
Center(s)
Population
Bouman
106
Single-center Surgical
(cardiac surgery +++)
Jamale
208
Single-center Medical
ELAIN
231
Single-center Surgical
(cardiac surgery +++)
Pilot
STARRT-AKI
100
Multicenter
Mixed
HEROICS
224
Multicenter
Cardiac surgery
AKIKI
620
Multicenter
Mixed
Time to RRT initiation
in the delayed strategy
Time to RRT
initiation
AKI
Time difference (h)
ELAIN
BOUMAN
JAMALE
HEROICS
STARRT
(Pilot)
AKIKI
20
35
??
36
42
57
Time to RRT initiation
in the delayed strategy
Time to RRT
initiation
AKI
ELAIN
BOUMAN
JAMALE
HEROICS
STARRT
(Pilot)
AKIKI
Time difference (h)
20
35
??
36
42
57
Free of RRT
(delayed group)
9%
17%
17%
43%
37%
49%
RCTs on the timing of RRT
• Bouman et al
• Jamale et al
• Zarbock et al (ELAIN)
• Wald et al (Pilot STARRT-AKI)
• Combes et al (HEROICS)
• Gaudry et al (AKIKI)
Bouman et al:
Crit Care Med 2002 Vol. 30, No. 10
RRT within 12h
28-Day Survival
100%
74.3%
68.8%
RRT if:
urea >40 mmol/L
K>6.5 mmol/L
severe pulmonary edema
75.0%
80%
60%
40%
n=35
n=35
n=36
20%
0%
EarlyHV
EarlyLV
LateLV
RCTs on the timing of RRT
• Bouman et al
• Jamale et al
• Zarbock et al (ELAIN)
• Wald et al (Pilot STARRT-AKI)
• Combes et al (HEROICS)
• Gaudry et al (AKIKI)
Mortality
JAMA 2016
Delayed RRT
Early RRT
Mortality
JAMA 2016
Delayed RRT
Early RRT
Early RRT
Delayed RRT
JAMA 2016
Mortality
P=0.03
Delayed RRT
P=0.07
Early RRT
Early RRT
Delayed RRT
JAMA 2016
Early RRT: KDIGO 2
Delayed RRT: KDIGO 3
JAMA 2016
Early RRT: KDIGO 2
Delayed RRT: KDIGO 3
Observational data have suggested that
in practice, very few patients with stage 2 AKI and only a
minority of patients with stage 3 AKI receive RRT
Hoste EA Intensive Care Med 2015
Hoste EA et al Crit Care 2006
Liu and Pavlesky CJASN 2016
JAMA 2016
Early RRT: KDIGO 2
Delayed RRT: KDIGO 3
Only 9% in the “delayed
RRT” group did not receive
RRT
Time to RRT (h)
Early RRT
Delayed RRT
6.0
25.5
JAMA 2016
Early RRT: KDIGO 2
Delayed RRT: KDIGO 3
The high rate of progression from stage 2 to stage 3 AKI
contrasts with observational data
(less than 50% of critically ill patients with the equivalent of
stage 2 AKI progress to stage 3 AKI)
Hoste EA et al Crit Care 2006
Liu and Pavlesky CJASN 2016
JAMA 2016
ns
ns
p 0.03
ELAIN Fragility index
3
Fragility index
Number of “nonevents” in the treatment group with the lowest
event rate that must be changed to “events” in order for the p
value calculated by the Fisher exact test to equal or exceed 0.05 .
JAMA 2016
Duration of RRT (d)
Hospital stay (d)
Early
Delayed
p
9
51
25
82
.04
<.001
“It is difficult to reconcile how such a relatively
small difference in the timing of initiation of RRT
(< 24 hours) resulted in not only 34 % reduction in
the hazard for death but also reductions in
median duration of RRT of > 2 weeks and median
duration of hospital stay of > 4 weeks”
• The results from single-center trials must be viewed with caution
• Center-specific effects (particular systems of care and background
therapies used) may preclude the generalizability of the findings
Results from single-center studies should rarely serve as the
basis for changing guidelines unless the results have been
validated in subsequent multicenter trials.
Similarity between ELAIN and HEROICS
• Surgery (cardiac surgery)
• Early group : “very early”
Baseline creatinine:
98 µmol/l (ELAIN)
150 µmol/l (HEROICS)
But HEROICS is a multicenter trial
Delayed Strategy = “Standard Care”
•
•
•
•
Severe hyperkalemia
Urine output < 0.3 ml/Kg/h 24h
Urea > 36 mmol/l
Creatinine > 352 µmol/l (or > X 3 baseline)
43% in the “Standard
Care” group did not
receive RRT
RCTs on the timing of RRT
• Bouman et al
• Jamale et al
• Zarbock et al (ELAIN)
• Wald et al (Pilot STARRT-AKI)
• Combes et al (HEROICS)
• Gaudry et al (AKIKI)
Kidney international
2015
Delayed strategy = “Standard RRT initiation”
• Hyperkalemia (>6 mmol/l)
• Severe metabolic acidosis
• Pulmonary edema (PaO2/FiO2 <200)
Kidney international
2015
37 % in the “Standard
RRT initiation” group
did not receive RRT
Kidney international
2015
Accelerated
RRT
Death by day 90
18 (38)
Standard
RRT
19 (37)
p
0.92
We need Large RCTs !
Large RCTs
RCT
n patients
Status
AKIKI
620
Published
IDEAL-ICU
864
Recruitment
terminated
STARRT-AKI
2800
Recruiting
July 2016
Wednesday November 2, 2016
5528 Had AKI and received vasoactive
agent and/or invasive MV
3430 Had AKI stage 3
of KDIGO classification
620 Underwent randomization
Early RRT Strategy
n=312
Delayed RRT Strategy
n=308
1 patient refused the
use of data
619 Were included in the analysis
Delayed Strategy Group
Pre-specified criteria
• Severe hyperkalemia
• Severe acidosis (pH <7.15)
• Acute pulmonary edema due to fluid overload
Responsible for severe hypoxemia
•
•
Oliguria/Anuria >72 hours
Serum urea concentration > 40mmol/l
B
Proportion of patients free of RRT
1
Patients free of RRT
0.8
0.6
0.4
Early RRT strategy
Delayed RRT strategy
0.2
p−value: <0.001
0
0
1
2
3
4
5
6
7
8
12 16 20 24 28
3
1
Days
311
7
4
4
4
4
3
3
308
268
229
192
153
135
118
105
1
0
0
0
92 61 39 28 21 13
A
Proportion of survivors
1
Survival
0.8
0.6
0.4
Early RRT strategy
Delayed RRT strategy
0.2
p−value: 0.79
0
0
7
14
21
28
35
42
49
56
60
Days
311
241
207
194
179
172
167
161
158 157
308
239
204
191
178
165
161
156
156 155
A
Proportion of survivors
1
Survival
0.8
0.6
0.4
Early RRT strategy
Delayed RRT strategy
0.2
p−value: 0.79
0
0
7
14
21
28
35
42
49
56
60
Days
311
241
207
194
179
172
167
161
158 157
308
239
204
191
178
165
161
156
156 155
B
To detect an effect size of 1.2 %
Proportion of patients free of RRT
1
(i.e., the difference in mortality that we found between the two groups)
0.8
0.6
with a power of 90%:
0.4
Early RRT strategy
Delayed RRT strategy
0.2
p−value: <0.001
> 70,000 patients would be required
0
0
1
2
3
4
5
6
Days
7
8
12 16 20 24 28
Secondary outcomes
Outcomes
RRT-free days – median (IQR) day 28
Catheter-related bloodstream infection– no. (%)
Early RRT strategy
(N=311)
Delayed RRT strategy
(N=308)
P
value
17 (2-26)
19 (5-29)
<0.001
31 (10)
16 (5)
0.03
Secondary outcomes
Outcomes
Early RRT strategy
(N=311)
Delayed RRT strategy
(N=308)
P
value
17 (2-26)
19 (5-29)
<0.001
31 (10)
16 (5)
0.03
RRT-free days – median (IQR) day 28
Probability of adequate diuresis
with no need for renal replacement therapy
Catheter-related bloodstream infection– no. (%)
Adequate urine output with no need for RRT
1
0.8
Delayed strategy
more rapid renal function recovery
0.6
0.4
Early RRT strategy
Delayed RRT strategy
0.2
p−value: <0.001
0
0
7
14
21
28
Days
311
99
42
27
10
308
68
29
14
7
Take home message
• We need to compare strategies with large RCTs
• November 2016:
A conservative strategy might be considered as the standard
• Benefits of the conservative strategy
 Obviated the need for RRT in almost 50% of cases
 Mortality did not differ significantly
 Renal function recovery was more rapid
• Perspective: IDEAL-ICU/ STARRT-AKI
(accumulation of evidence base data from Large RCTs)
Thank you
“Wait and see” approach:
safe but careful surveillance is mandatory
Is it more dangerous to draw several blood samples for
potassium and ABGs determination or to indwell a
dialysis catheter and initiate an unnecessary RRT in
unstable patients ?
Patients
n
Early group
Delayed group
Bouman
106
UO < 30ml/h +
Cr cl < 20 ml/min
•
•
•
Jamale
208
Urea > 25 mmol/l
Creat > 616 µmol/l
Clinically indicated as judged by
physician
ELAIN
231
KDIGO 2
KDIGO 3
Pilot STARRT-AKI
100
~ KDIGO 2
+/- NGAL ≥ 400
ng/ml
•
•
•
K > 6 mmol/l
Severe acidosis
Severe pulmonary edema
HEROICS
224
Severe post-cardiac
surgery shock
•
•
•
•
Severe hyperkalemia
UO < 0.3 ml/Kg/h 24h
Urea > 36 mmol/l
Creatinine > 352 µmol/l
Urea> 40 mmol/L
K > 6.5 mmol/L
Severe pulmonary edema
(or > X 3 baseline)
AKIKI
620
KDIGO 3
• Severe hyperkalemia
• Severe acidosis
• Severe pulmonary edema
• Oliguria/Anuria >72 hours
Criteria to start RRT in the delayed groups
Bouman
Jamale
ELAIN
Severe pulmonary
edema
Urea or Creatinine
concentration
Oligoanuria
(duration)
KDIGO 3
Severe acidosis
Judged by physician
Severe Hyperkalemia
Pilot
STARRT
HEROICS
AKIKI