1. When to start CRRT during ECMO?
Matthew L. Paden, MD
Associate Professor of Pediatric Critical Care
Director, Pediatric ECMO

2. Disclosures
Everything in ECMO is off label use
I’m a believer…

3. When do YOU start CRRT during ECMO?
On every pediatric respiratory patient at initiation of ECMO, regardless of creatinine/UOP.

4. When to start CRRT during ECMO?
No one knows.

5. Concomitant ECMO and CRRT
Question
Evidence
Optimal population for therapy?
None
What indication?
Survey results about what people are doing
Timing of initiation?
Expert opinion, ELSO Guidelines
Optimal mode of therapy?
Optimal method of therapy? (Device vs. in-line)
Optimal dose of therapy?
Effectiveness of therapy? (URR, Kt/V)
Outcome?
4 single center reports, Registry data

6. Objectives Discuss why it might be helpful
Discuss what people are doing
Review the small amount of outcome literature
Discuss where future work needs to be done

7. Why CRRT Could Provide Potential Benefit on ECMO
Treatment of Acute Kidney Injury which is common
Impact on Reducing Inflammatory Response of ECMO/Underlying Disease Process
Impact on Decreasing Fluid Overload

8. AKI Prevalence in ECMO Neonates - 25% (Askenazi 2011)
ELSO registry ~8000 non-cardiac neonates
Congenital diaphragmatic hernia - 71% (Gadepalli 2011)
Congenital hearts - 72% (Smith 2009)
Pediatric respiratory - 63% (ELSO DB 2011)
AKI on ECMO is associated with increased mortality, controlling for confounders (Askenazi 2012)
AKI on adult ECMO: OR 12.1 (2.5-59)
AKI on pediatric ECMO: OR 24.0 ( )

9. CRRT Associated with Increased Risk of Death in Pediatric ECMO
ELSO Registry ( )
Adjusting for other risk factors: Patients with AKI and CRRT had higher mortality
Neonates (25% AKI incidence)
OR with AKI: 3.2
OR with RRT: 1.9
Children (46% AKI incidence)
OR with AKI: 1.7
OR with RRT: 2.5
Therapies to prevent/ameliorate AKI and optimize RRT could improve outcomes
-Askenazi et al., Pediatr Crit Care Med, 2011

10. Impact of ECMO on Inflammation
Circulation of blood across synthetic surfaces escalates a pro-inflammatory response (already activated by disease)
Early elevation of TNF-alpha, IL-1beta, IL-6, IL-8 within 3-4 hours post-ECMO cannulation; associated with lung changes (Fortenberry et al., J Peds 1996; Massoudy et al., Chest 2001)

11. CRRT Decreases Cytokine mRNA Expression in VV ECMO-Healthy Pig Model
-Shen et al. (Nanjing U), Inflammation, 2013

12. Improvement in ECMO-Induced Mitochondrial Dysfunction with CRRT in Healthy Pigs
-Shen et al. (Nanjing U), Inflammation, 2013

13. CRRT Modulates Renal Inflammatory Cytokines During ECMO
Porcine model
Control, sham vs. VV ECMO, ECMO+CRRT
CRRT addition decreased renal TNF, IL-1, IL-6 expression
Decreased Nf-KB transcription gene expression
TNF- a
Nf-KB
-Hu Yimin et al., (Nanjing), J Cardiothoracic Surgery 2013

14. Fluid Overload and Outcome
Strong body of evidence showing fluid overload is associated with:
Acute kidney injury
Increased mortality
Increased ventilator days
Increased ICU LOS
In both children and adults
Conservative fluid approach (fluid restriction, diuretics) associated with improved ventilator days, ICU LOS (FACTT Trial, NEJM 2006)

17. ELSO Guidelines
…spontaneous or pharmacologic diuresis should be instituted until patient is close to dry weight and edema has cleared. This will enhance recovery from heart or lung failure and decrease the time on ECLS.
The goal of fluid management is to return the extracellular fluid volume to normal (dry weight) and maintain it there.
As with all critically ill patients, full caloric and protein nutritional support is essential.

18. ELSO Guidelines
The hourly fluid balance goal should be set and maintained until normal extracellular fluid volume is reached (no systemic edema, within 5% of “dry” weight). Renal replacement therapy use to enhance fluid removal allowing adequate nutritional support is often performed.
Despite the literature surrounding fluid overload (>10%) as a risk factor for death, review of the ELSO registry also finds that use of renal replacement therapy is also a risk factor for poor outcome.
Even if acute renal failure occurs with ECLS, resolution in survivors occurs in >90% of patients without need for long-term dialysis.

19. So we should do it, but how do we do it?
No guidance.

20. In-line hemofiltration during ECMO
Hemofilter placed post-pump, preoxygenator
Returns pre-pump
Control UF with IV pump’s resistance
Measure UF with urometer
Deliver replacement fluids/dialysate with IV pump

21. Advantages of in-line hemofiltration
Cheap
Does not require stand alone machine
Only need dialysis filter, IV pumps, and urometer
No need for dialysis/CRRT trained nurse
Can provide SCUF, CVVH, CVVHD
Smaller addition of extracorporeal volume compared to stand alone CRRT device

22. Disadvantages of in-line hemofiltration
Creates a shunt
Pump flow ≠ delivered flow
Less effective with diffusive techniques
Incredibly inaccurate fluid control
IV pump error rates – between %
ECMO conditions – 34 mL/hour (>800 mL/day) difference between prescribed and actual ultrafiltration.

23. Pediatric ECMO / In-line CRRT Warning
Sucosky et al., J Med Devices (2), 2008
IV pumps
Your I/O’s are not accurate
Delivers less replacement fluid than ordered.
10 kg child with 300 mL/hour UF rate – negative 288 mL per day (28 ml/kg)
45 kg adolescent with 2000 ml/hour UF rate – negative 1.9 L/day (42 ml/kg)

24. Using a CRRT device during ECMO
CRRT device placed pre-pump
Returns pre-pump
UF controlled by CRRT device

25. Advantages of CRRT device
Engineered to provide CRRT *
Provides multiple modalities of therapy
More accurate fluid balance control than in-line hemofiltration
No shunt

26. Disadvantages of CRRT device
Not engineered to provide CRRT on ECMO
Alarm modes
Default access pressure alarms are
usually negative
Accuracy
Better than inline over time, but still can
be a problem over short time periods
-30
-10 -5
-40 35 10
Hourly Fluid Balance

27. CRRT outcomes NOT on ECMO
ppCRRT (344 multicenter patients)
overall survival 58% (Symons ClinJAmSocNeph 2007)
MODS 51% survival (Goldstein Kid Int 2005)
<10 kg 43% (Askenazi J Peds 2012)
Multiple other single center reports
40-55% survival

28. RRT/ECMO Outcomes – ELSO Registry
Survival
Neonatal respiratory
2696/ (51%)
Pediatric respiratory
1010/ (40%)
Adult respiratory
815/ (46%)
Cardiac 0-30d
527/ (24%)
Cardiac 31d – 364d
364/ (30%)
Cardiac 1y-16 y
437/ (40%)
Cardiac >16 y
366/ (26%)

29. Concomitant ECMO and RRT – Neo/Peds
Author (year)
Patient population
Number of patients
Survivors
Outcome reported
Bartlett (1986)
Neonates 10 2
Not reported
Weber (1990) 43 8
Sell (1987)
Neo/Ped 6
“Normal renal function at discharge”
Adolph (1991)
Pediatric 3
N/A
Weber (1998) 38
“No survivors with renal injury at discharge”
Swaniker (2000) 18
Meyer (2001) 35 15
***
Paden (2011)
154 68

30. Concomitant ECMO/CRRT Renal Outcomes
University of Michigan
35 CRRT/ECMO patients – 15 survivors (43%)
14/15 (93%) with full renal recovery at D/C
Wegeners – ultimately transplanted
Children’s Healthcare of Atlanta
154 CRRT/ECMO patients - 68 survivors (44%)
65/68 (96%) with full renal recovery at D/C
1 nosocomial enterococcus sepsis at transfer – normal 1 month later
2 primary renal disease (Wegeners/polyangiitis) – Cr 13.7/6.5
One ultimately transplanted / one with elevated Cr, no RRT

31. ECMO and CRRT Outcome Reporting
Under reporting in the literature
Medical literature survivors – 161
ELSO Registry survivors – 6215
2.6% of all survivors
Biased towards bad outcomes
Medical literature - 28% survival of 567 patients
ELSO Registry – 40% survival of 15,486 patients
Highly biased to a few centers experience
83% of the medical literature patients come from 3 centers
201 centers reporting to ELSO in 2012

32. KIDMO
Multiple centers – Michigan, UAB, Vanderbilt, Cincinnati, Emory, McGill
Standard definition of AKI (KDIGO modification of AKIN)
Analysis of ~900 patients

33. AKI is common

34. AKI - longer ECMO and increased mortality

35. Fluid overload at ECMO initiation

36. Peak Fluid overload

37. More in depth review of the topic

38. Angelica Hale