Pediatric CRRT: Outcome

1
Pediatric CRRT Outcome

• Stuart L. Goldstein, MD

2
Ronco et al. Lancet 2000 351 26-30
3
Ronco et al. Lancet 2000 351 26-30

• Conclusions
• Minimum UF rates should reach at least 35
  ml/kg/hr
• (2000ml/1.73m2/hr when adapted for children)
• Survivors in all their groups had lower BUNs than
  non-survivors prior to commencement of
  hemofiltration
• Begs the question does early CRRT effect outcome?

4
Pediatric Acute Renal FailureIdeal Study Design

• Prospective protocol driven entry criteria to
  ensure that patients and their respective disease
  receive similar treatment
• Control for severity of illness, primary and
  co-morbid diseases
• Adequate power to detect effect of an
  intervention on or an association of a clinical
  variable with outcome

5
Pediatric Acute Renal FailureIdeal Study Design

• Prospective protocol driven entry criteria to
  ensure that patients and their respective disease
  receive similar treatment --- Do not exist!
• Control for severity of illness, primary and
  co-morbid diseases --- Some information
• Adequate power to detect effect of an
  intervention on or an association of a clinical
  variable with outcome --- Do not exist!

6
Renal Replacement Therapy in the PICUPediatric
Outcome Literature

• Few pediatric studies (all single center) use
  severity of illness measure to evaluate outcomes
  in pediatric RRT
• Lane noted that mortality was greater after bone
  marrow transplant who had gt 10 fluid overload at
  the time of HD initiation
• Smoyer2 found higher mortality in patients on
  pressors
• Faragson3 found PRISM to be a poor outcome
  predictor in patients treated with HD
• Zobel4 demonstrated that children who received
  CRRT with worse illness severity by PRISM score
  had increased mortality
• Did not stratify by modality

1. Bone Marrow Transplant 13613-7, 1994 2. JASN
61401-9, 1995 3. Pediatr Nephrol 7703-7,
1994 4. Child Nephrol Urol 1014-7, 1990
7
Pediatric ARF Modality and Survival
Plt0.01
Plt0.01
Survival
Bunchman TE et al Ped Neph 161067-1071, 2001
8
Pediatric ARF Modality and Survival

• Patient survival on pressors (35) lower than
  without pressors (89) (plt0.01)
• Lower survival seen in CRRT than in patients who
  received HD for all disease states

Bunchman TE et al Ped Neph 161067-1071, 2001
9
CRRT and Outcome in Children

• Retrospective review of all patients who received
  CVVH(D) in the Texas Childrens Hospital PICU
  from February 1996 through September 1998 (32
  months)
• Pre-CVVH initiation data
• Age
• Primary disease leading to need for CVVH
• Co-morbid diseases
• Reason for CVVH
• Fluid intake (Fluid In) from PICU admission to
  CVVH initiation
• Fluid output (Fluid Out) from PICU admission to
  CVVH initiation
• GFR (Schwartz formula) at CVVH initiation

Goldstein SL et al Pediatrics 2001
Jun107(6)1309-12

10
CRRT and Outcome in Children

• PRISM scores at PICU admission and CVVH
  initiation calculated by same nurse
• PICU Course Data
• Maximum number of pressors used
• Pressors completely weaned (y/n)
• Mean Airway Pressure (Paw) at CVVH initiation and
  termination
• ICU length of stay (days)
• CVVH complications
• Outcome (death or survival)

Goldstein SL et al Pediatrics 2001 1071309-12
11
CRRT and Outcome in Children

• Survival curve demonstrates that nearly 75 of
  deaths occurred less than 25 days into the ICU
  course

Goldstein SL et al Pediatrics 2001 1071309-12
12
CRRT and Outcome in Children

• Lesser FO at CVVH (D) initiation was associated
  with improved outcome (p0.03)
• Lesser FO at CVVH (D) initiation was also
  associated with improved outcome when sample was
  adjusted for severity of illness (p0.03
  multiple regression analysis)

Goldstein SL et al Pediatrics 2001 1071309-12
13
Fluid Overload as a Risk Factor
N113
p0.02 p0.01
Foland et al, CCM 2004 321771-1776
14
Kaplan-Meier survival estimates, by percentage
fluid overload category
Gillespie et al, Pediatr Nephrol (2004)
191394-1999
15
Fluid Overload as a Risk Factor
N 77
Gillespie et al, Pediatr Nephrol 2004
191394-1999
16
The Evolution of Idea to Practice Paradigm
Registry
Single center study
Randomized Trial
17
The Prospective Pediatric CRRT (ppCRRT) Registry

• No single pediatric center cares for enough CRRT
  patients annually to analyze the effect of more
  than a few variables on patient outcome
• Mitigate geographical and institutional effects
  on
• Patient demographics
• CRRT practice patterns
• SHARE INFORMATION
• Generate hypotheses for future RCTs

18
ppCRRT Experience

• First patient enrolled on 1/1/01
• 370 patients entered into database as of 07/12/05
• Currently 13 active participating pediatric
  centers

• Texas Childrens
• Boston Childrens
• Seattle Childrens
• UAB
• University of Michigan
• Mercy Childrens, KC
• Egleston Childrens, Atlanta

• All Childrens, St. Petersburg
• DC Childrens
• Columbus Childrens
• Packard Childrens, Palo Alto
• DeVos Childrens, Grand Rapids
• Cleveland Clinic

19
Patient Demographics

• Newborn to 25 years
• 59 males
• Weights 1.3 160kg (mean 33.5 kg)
• Mean 6.5 days in ICU prior to CRRT
• (range 0 135 days, median 2)
• Modality
• CVVH (33)
• CVVHD (54)
• CVVHDF (13)

20
ppCRRT Data Size Distribution
21
ppCRRT MODS Data
22
ppCRRT MODS Data

• BASELINE DEMOGRAPHICS
• 157 patients entered (1/1/2001 to 5/31/04)
• 116 with MODS (2 organs involved)
• Mean age 8.5 6.8 years (2 days to 25.1 years)
• Mean weight 33.7 25.1 kg (1.9 to 160 kg)
• Median 3 ICU days prior to CRRT initiation
• Range 0 to 103 days
• 67less than 7 days

Goldstein SL et al Kidney International 2005
23
ppCRRT MODS Data Clinical Variables
Goldstein SL et al Kidney International 2005
24
ppCRRT MODS Data Other Analyses

• 77 of non-survivors die within 3 weeks of ICU
  admission
• Survival rates similar by CRRT modality (H 57),
  (DF 53), (HD 50)
• Survival rates similar for patients on 0-1
  (53), 2 (54) or 3 (39) pressors
• Survival rates better for patients with lt20 FO
  (59) versus gt20 FO (35) at CRRT initiation
  (plt0.001)

Goldstein SL et al Kidney International 2005