When Brain Injury Happens in Newborns with Heart Disease Opportunities to Improve Outcome
Steven P. Miller, MDCM FRCPC
CIHR Clinician Scientist & MSFHR Scholar Assistant Professor, Pediatrics (Neurology) BCCH & CFRI University of British Columbia
Newborn Brain
31 Weeks
42 Weeks
Objectives
• When brain injury happens in newborns with congenital heart disease • Timing of brain injury informs mechanism in newborns with congenital heart disease • Implications for studying brain protection
and clinical trial design
Brain Injury in Congenital Heart Disease
• 6-8/1000 term newborns
– 50% require surgery as newborns
• Long-term neurodevelopmental disability is common in newborns with congenital heart disease
Brain Injury in Congenital Heart Disease
Bypass Birth
2000: Neurobehavioral abnormalities in >50% preoperative
Limperopoulos JPediatr (2000) 137
2003-2004: At school age, outcomes of different bypass strategies (Circulatory Arrest, Low Flow, Full Flow) not different but whole cohorts below expectations
Bellinger JTCS(2003)126 Karl JTCS(2004)127
You can observe a lot just by watchin’
Yogi Berra
PreOp
Bypass
PostOp
Term Birth
Outcome MRI NIRS MRI
PreOp
Bypass
PostOp
Term Birth
Outcome MRI NIRS MRI
McQuillen PS et al. Circulation 2006
Pre-Operative Injury: Stroke
Pre-operative Injury
Median (Range) or Number (%) Number of patients Balloon Atrial Septostomy SNAP-PE Lowest O2 saturation recorded 14 (9 - 26) 70 (26 - 82) 19.5 (9 - 30) 50 (20 - 70) 0.1 0.05 No Brain Injury Acquired Brain Injury P
17 (59%)
7 (41%)
12 (41%)
12 (100%) 0.001
Not route of BAS
No Brain Injury Umbilical Femoral Both 2 (29%) 4 (57%) 1 (14%) Acquired Brain Injury 2 (17%) 8 (67%) 2 (17%) 0.8 P
PreOp
Bypass
PostOp
Term Birth
Outcome MRI NIRS MRI
McQuillen PS et al. Stroke 2007
Term newborn
Congenital Heart Disease
Premature Newborn
Perioperative white matter injury is common
WMI Preoperative N=62 Postoperative N=53 Totals Stroke Any
11 (18%) 13 (21%) 23 (37%) 14 (26%) 5 (9%) 19 (36%)
23 (40%) 15 (26%) 32 (56%)
Impaired Global Oxygen Delivery
Median (Range) or Number (%)
No Postop Injury
New Postop Injury 19 7 (37%) -2.7 (-10 - 6.1) 85 (17 - 197) 8 (42%) 8 (42%) 3 (16%) 41 (33 - 63)
P
Number of patients Preoperative brain injury Pump base deficit Lowest flow ml/kg/min Bypass strategy Full Flow RCP Other Postop Day #1 Lowest MBP
34 13 (38%) -0.65 (-4.8 - 8.4) 142 (22 - 229) 26 (76%) 2 (6%) 6 (18%) 45 (33 - 66)
0.6 0.01 0.05
0.01 0.09
Intraoperative oxygen delivery
When brain injury happens in newborns with heart disease?
• Pre-operative stroke
– Therapeutic catheterization procedures
• Post-operative white matter injury
– Bypass strategy – Low cardiac output state
• Different modifiable risk factors
Newborn Brain
31 Weeks
42 Weeks
PreOp
Bypass
PostOp
Term Birth
Outcome MRI MRI
Normal Term Controls
• 41 Newborns with CHD
– 29 with TGA – 12 with Single Ventricle
• 16 Normal Controls
• 3D MRSI • DTI
– 1.4 mm in-plane – Dav – Fractional Anisotropy
Newborns with CHD relative to Controls MRSI NAA/Choline -10%
95% CI
p-value
-15% to -3%
0.003
Lactate/Choline
28%
-3% to 68%
0.08
DTI Average Diffusivity Fractional Anisotropy -12% -6% to -18% <0.001 4.5% 0.2% to 6.9% <0.001
Fetal Birth
PreOp
Bypass
PostOp
Outcome MRI MRI
aCCENT
A Congenital Heart Disease Center of Excellence for Neuroprotective Trials
University of British Columbia University of Alberta University of California, San Francisco
Goal
• To evaluate novel brain protection strategies to improve the neurodevelopmental outcomes of newborns with congenital heart disease
• Multidisciplinary and Multicentre • Brain Imaging
An Ideal Outcome
• Rapidly available • Objective or Quantifiable
Trial Designs
• Conventional Randomisation • “Play the winner” • Bayesian Adaptive Randomisation
First Proposed Trial
Bayesian Adaptive Randomization
Objective
• Decrease the occurrence of stroke in term newborns with transposition of the great arteries (TGA) requiring a balloon atrial septostmy (BAS)
Peer Pressure
• Anti-coagulate the next 10 babies needing a septostomy • If no strokes, make this clinical practice
Research Question
• Does therapeutic anticoagulation with Heparin decrease the occurrence of stroke on pre-operative MRI by 50% in term newborns with TGA undergoing a balloon atrial septostomy?
Cohort
• Term newborns with TGA needing BAS
– Defined Criteria for BAS Need
• Exclusion criteria
– Birth Asphyxia (Defined) – Late BAS (> 1 week) – Coagulopathy
Primary Outcome Measure
• Stroke on Pre-Operative MRI
– Standard Neuroradiological Criteria – Conventional MRI and Diffusion Tensor Imaging
Intervention
• Anticoagulation with Heparin
– Stratify for presence of umbilical vein catheter
Trial Design
• Conventional Design – Sample Size Requirements
• Baseline Risk: 50% incidence of brain injury • Targeted Risk Reduction: to 25% incidence • 100 subjects overall
– two-sided alpha level of .053, with power of.74 for 50% vs 25% and .90 for 50% versus 10% injury rates (control vs. heparin arm).
• Two-three years recruitment
Trial Design
• Bayesian Adaptive Randomization
Trial De sign Pr imary endpoint Numbe r of arms Min imum patient accrual Max imum patient accrual Numbe r of patients to randomiz e fairly Min imum randomiz ation probability Binary 2 30 100 20 0.1
Trial Design
• Bayesian Adaptive Randomization • Stopping Rules
Suspend accrual to an arm if Pr( qi < qj≠ i | Data ) < PL Stop the trial and select an arm as superior if Pr( qi < qj≠ i | Data ) > PU At the final analysis select an arm as superior if Pr( qi < qj≠ i | Data ) > PU*
Arm Name Control Exper imental
PL 0.025 0.025
PU 0.975 0.975
PU* 0.975 0.975
Trial Design
Arm Control Heparin True Prob (Success) 0.5 0.5 Pr( Sele cted ) 0.0295 0.023 Pr( Sele cted Earl y ) 0.029 0.0225 Pr( Stopped Earl y ) 0.023 0.0295 # Pa tients (2.5%,97.5%) 49.1 ( 25, 68 ) 48.9 ( 24, 68 )
Trial Design
Arm Control Heparin True Prob (Success) 0.5 0.5 Pr( Sele cted ) 0.0295 0.023 Pr( Sele cted Earl y ) 0.029 0.0225 Pr( Stopped Earl y ) 0.023 0.0295 # Pa tients (2.5%,97.5%) 49.1 ( 25, 68 ) 48.9 ( 24, 68 )
Arm Control Heparin
True Prob (Success) 0.5 0.25
Pr( Sele cted ) 0 0.749
Pr( Sele cted Earl y ) 0 0.742
Pr( Stopped Earl y ) 0.749 0
# Pa tients (2.5%,97.5%) 22.4 ( 11, 44 ) 46.1 ( 18, 75 )
Trial Design
Arm Control Heparin True Prob (Success) 0.5 0.5 Pr( Sele cted ) 0.0295 0.023 Pr( Sele cted Earl y ) 0.029 0.0225 Pr( Stopped Earl y ) 0.023 0.0295 # Pa tients (2.5%,97.5%) 49.1 ( 25, 68 ) 48.9 ( 24, 68 )
Arm Control Heparin
True Prob (Success) 0.5 0.25
Pr( Sele cted ) 0 0.749
Pr( Sele cted Earl y ) 0 0.742
Pr( Stopped Earl y ) 0.749 0
# Pa tients (2.5%,97.5%) 22.4 ( 11, 44 ) 46.1 ( 18, 75 )
Arm Control Heparin
True Prob (Success) 0.5 0.2
Pr( Sele cted ) 0 0.91
Pr( Sele cted Earl y ) 0 0.905
Pr( Stopped Earl y ) 0.91 0
# Pa tients (2.5%,97.5%) 17.9 ( 10, 36 ) 38.4 ( 18, 72 )
Conclusions
• Timing informs mechanism • Surprising patterns of brain injury • Important for implementing brain protection • Opportunities to consider new trial designs
UBC
Neurology UCSF Steven Miller MD Vann Chau MD Wendy Soulikias RN Vesna Popovska MD Alan Hill MD PhD Elke Roland MD Bruce Bjornson MD Biostatistics Rollin Brant PhD Neonatology Anne Syness MD Ruth Grunau PhD Radiology Ken Poskitt MD Michael Sargent MD PICU/Cardiac Surgery Andrew Campbell MD Norbert Froese MD
Neonatal Brain Disorders Center Patrick McQuillen MD Donna Ferriero MD A. James Barkovich MD Shannon Hamrick MD David V Glidden PhD MR Science Center Daniel Vigneron PhD Pratik Mukherjee MD PhD
Support CIHR- workshop planning grant Michael Smith Foundation March of Dimes NIH:NS35902, RR01271(PCRC) AHA Hillblom Foundation BC Children’s Hospital Foundation