THERAPEUTIC HYPOTHERMIA AFTER CARDIAC ARREST

1
THERAPEUTIC HYPOTHERMIA AFTER CARDIAC ARREST

• Adam Oster R3
• Resident Oral Presentation
• November 13, 2003

2
Therapeutic Hypothermia Post Cardiac Arrest

• Guiding questions
• Supporting science
• Preliminary studies
• Clinical trials
• Cooling technology
• Who to cool
• When to cool
• How long to cool

3
Therapeutic Hypothermia Post Cardiac Arrest

• Baseline cardiac arrest data
• Physiology of CA
• Preliminary studies of induced HT
• Recent clinical trials
• Cooling Technology
• The Future

4
Promising Therapies?

• Thiopental
• Steroids
• Calcium channel antagonists
• Glutamate channel antagonists
• Nimodipine
• Lidoflazine
• PEG-SOD
• Mg /-ativan

5
OPALS Data1991-1997

• 9273 out-of-hospital CA
• 38.6 VF/pVT
• 27 ROSC
• 21 admitted to hospital
• 9 survival to discharge
• 15 poor neurologic outcome
• not OPALS data
• 7 of all pre-hospital CA return home to
  independent living
• Eisenberg, M. Annals of Emergency Medicine, 1990.

6
VF/VTCognitive Sequelae

• Outcome of patients surviving to hospital
  post-VF/VT with GCS /lt9
• Best estimates (based on control group in two
  large trials)
• Mortality at 6 mo 55-68
• Neurologic outcome at 6 mo
• 26-40 poor outcome CPC 3/4

7
Cardiac ArrestCognitive Sequelae

• Graves, J. Resuscitation 1997
• Sweden 1980-1993
• N3754
• 9 survived to discharge
• 21 mortality at 1yr
• 56 by 5yrs
• 82 by 10yrs
• Cerebral Performance Category on discharge, N320
• 1 53
• 2 21
• 3 24
• 4 2

8
Cardiac ArrestCognitive Sequelae

• Bur, A. Intensive Care Medicine, 2001.
• Patients admitted post-VF CA, N276 (out of 1254)
• 50 mortality at 6mo
• 87 good neurologic outcome
• Age, duration of ROSC, time to EMS, time to 1st
  defib, and amount of epi all significantly
  related to CPC category.

9
Cardiac ArrestCognitive Sequelae and QOL

• Granja, C. Resuscitation, 2002.
• Compared CPC and QOL post-CA
• QOR survey administered at 6mo, N24/97
• N97 admitted after CA
• 36 (37) discharged from hospital
• 12 more died before 6mo
• 5 LTFU
• Questionnaire administered to 19
• No significant differences compared to other
  non-CA ICU survivors

10
Cardiac Arrest Physiology

• 4 stages
• Pre-arrest
• Arrest
• Resuscitation
• Post-resuscitation

11
How Effective is CPR?

• CPR cardiac output
• optimally carried-out up to 60
• realistically 20-30
• CO inversely proportional to duration of CA
  preceding initiation of CPR
• animal models
• 50 pre-arrest CBF if lt2mins
• 0 if gt10mins

12
Cardiac Arrest Physiology

• Arrest and Resuscitative Phases
• No to low-flow state tolerated for approx 5mins
• Brain O2 stores lost in 20secs
• ATP and glucose in 5 mins
• cells revert to anaerobic metabolism
• Major mechanism of injury is Ca influx
• multiple biochemical pathways are initiated
• loss of normal cellular ionic gradients
• tissues most susceptible --
• brain (esp. hippocampus, cerebral cortex and
  cerebellum)
• Ross. Journal of Cerebral Blood Flow and
  Metabolism, 1993.

13
Cerebral Blood Flow

• Post-arrest CBF
• Reperfusion injury
• After initial increase, CBF reduced to 50 normal
  for 90mins to 24hrs in normotensive pts
• Heterogenous CBF
• Increase in cerebral O2 uptake
• Bottiger, et al., Resuscitation 1997.
• Some evidence of raised ICP and cerebral edema
  post-ROSC.
• Morimoto, et.al., Critical Care Medicine, 1993.

14
Effects of Hyperthermia

• Hickey, R. Critical Care Medicine. 2003.
• Hyperthermia exacerbates histologic neuronal
  damage post-hypoxic arrest in rats.

15
Hypothermia Physiology

• How could hypothermia help?
• 7 reduction in cerebral metabolic rate (CMRO2)
  for every 1 degree reduction in brain temp.
• In part due to reduction in electric activity
• Critical Care Medicine, 1996
• Suppresses many chemical reactions
• Reduction in oxidative damage
• Reduces free calcium shifts
• Maintains mitochondrial function
• Reduces excitatory glutamate release
• Journal of Cerebral Blood Flow, 2000.

16
Hypothermia Physiology

• CNS effects of IH
• Cerebral metabolic rate for O2 is the major
  determinant of CBF
• May improve flow to selective ischemic areas of
  the brain
• Decreases ICP
• Likely due to global cerebral vasoconstriction
  and decreased IC blood volume
• Critical Care Medicine, 1984.
• Decreases amount of excitatory neurotransmitters
• Anaesthesia, 1994.

17
Hypothermia Physiology

• CVS Effects of IH
• With shivering mechanism blocked
• Decrease HR
• Increases SVR
• SV and MAP constant
• Osbourne wave at 33 deg

18
Hypothermia Physiology

• Respiratory Effects of IH
• ?increased risk of pneumonia
• Does not appear to if lt24hrs

19
Hypothermia Physiology

• Renal effects of IH
• Decreased resorbtion of solute causes osmotic
  diuresis
• K shifts into cells
• Decreased phosphate

20
Hypothermia Physiology

• Acid-Base/ABG correction
• When ABG corrected for temp, looks like a
  respiratory alkalosis
• Controversial whether ABGs should be corrected
  for temp but currently they are not corrected
• Some evidence for better outcome (animal studies)
  if you do correct for temp and manage pH ?
  decreased cerebral infarct volume and amount of
  edema formed.
• Anesthesiology, 2002.

21
Hypothermia Physiology

• GI effects of IH
• Decreased motility
• Decreased insulin release causes increase in
  glucose. All patients require insulin to avoid
  the complications of hyperglycemia.

22
Induced Hypothermia Trials

• Bigelow, 1950.
• Benson et al., 1955.
• Williams and Spencer, 1958.
• Bernard et al. Annals of Emergency Medicine,
  1997.
• Yanagawa et al. Resuscitation, 1998.
• Zeiner, et al. Stroke, 2000
• Holzer et al. NEJM, 2002.
• Bernard et al. NEJM, 2002.

23
Neurologic Outcome Measurements

• Glasgow Outcome Score
• Cerebral Performance Category
• Physiatrist assessment of best discharge location

24
Bernard et al., Annals of Emergency Medicine,
1997.

• Prospective, consecutive case series compared to
  consecutive historic control group
• ROSC post-CA (included non-VF/VT)
• Exclusion
• SBP lt90 with pressors
• Decreased LOC possibly due to trauma or CVA
• Age lt16, possibly pregnant
• N22

25
Bernard et al., Annals of Emergency Medicine,
1997.

• Intubation/paralysis and sedation
• Surface cooling with ice packs to 33deg for 12hrs
  then actively rewarmed
• Thrombolysis as indicated (no angioplasty)
• Similar protocoled ICU management
• Glasgow Outcome Scale estimated by unblinded
  chart review based on data at time of hospital
  discharge

26
Bernard et al., Annals of Emergency Medicine,
1997.

• Results
• 2 groups comparable at entry
• Similar incidences of witnessed collapse, time to
  CPR, ROSC, VF as presenting rhythm, brainstem
  reflexes
• None in NT group received thrombolysis vs 4 in MH
  group
• Mortality
• MH 10 vs NT 17 (45 vs 77 ARR 32 ? NNT 3), sig.
• Good neurologic outcomes (GOS1/2)
• MH 11 vs NT 3 (50 vs. 13.7, ARR 36 ? NNT 2.7),
  sig.
• Adverse Events
• No difference between groups

27
Bernard et al., Annals of Emergency Medicine,
1997.

• Study limitations
• Small numbers
• Historic controls
• Some pre-hospital data unavailable (eg EMS to
  ROSC
• Unclear if post-resuscitation protocols similar
• Non-blinded assessment of outcome ?
  classification bias
• Underpowered to find difference in adverse events
• Strengths
• MH feasible and likely safe
• May have effect on mortality and neurologic
  outcome

28
Yanagawa, et al. Resuscitation, 1998

• Consecutive, patients with ROSC post-CA, N13
• Compared to historic normothermic control group.
• Similar exclusion criteria
• Intubated/paralyzed/sedated as per protocol
• MH cooled to 33 deg for 48hrs using cooling
  blankets and EtOH on skin
• Passively rewarmed over 3-4 days
• GOS at 6 mo (not blinded to treatment)

29
Yanagawa, et al. Resuscitation, 1998

• Results
• Groups had different incidences of cardiac (vs
  pulmonary) etiology of arrest
• Stat sig difference in witnessed collapse (10 vs
  3, in MH group)
• No difference in mortality
• 3 vs 1 with GOS 1/2
• Stat sig. increase in pulmonary complications in
  MH group

30
Zeiner, A. et al., Stroke, 2000.

• Prospective, multicentered.
• Historic controls
• Included only post-VF
• Exclusion
• CA lt5 or gt15 mins or 60 mins without ROSC
• Post-resuscitation SBPlt60 or SaO2lt85
• Pts having subsequent CA within 6mo
• Cooled to 33deg via external head and body for
  24hrs then passively rewarmed
• CPC at 6mo

31
Zeiner, A. et al., Stroke, 2000

• Results
• 31 pts MH
• 4 excluded from analysis
• 11 died (mortality 41)
• CPC 1/2 14 (52)
• CPC 3/4 2 (7)
• No formal comparison with historic controls

32
Bernard et al. and Holzer et al., NEJM, 2002.

• Two (European and Australian) prospective,
  randomised controlled trials of MH post VF/VT CA.
• Similar inclusion and exclusion criteria
• Primary outcome was neurologic function at 6 mo
  or discharge from hospital
• Differences cooling methodology, initiation of
  IH, total duration of cooling and blinding of
  evaluators.

33
Bernard et al., NEJM, 2002.

• Australian Trial
• Only included VF-resuscitated out-of-hospital pts
  who remained unresponsive
• Did not specify duration of CA
• Exclusion criteria
• Odd-even day randomization
• Pre-hospital initiation of cooling
• Thrombolysis as indicated

34
Bernard et al., NEJM, 2002.

• Ice packs to head, neck, torso and limbs
• MH for 12hrs with sedation and paralysis
• Actively re-warmed with heating blanket at 18hrs
• After 24hrs patient care followed usual ICU
  protocols
• Blinded assessment by Physiatrist when pt ready
  for d/c from hospital (good vs poor outcome)

35
Bernard et al., NEJM, 2002.

• 84 pts eligible over 33mo
• 7 excluded from analysis
• 77 pts ? 43 (MH), 34 (NT)
• Groups statistically different in rates of
  bystander CPR (NTgtMH)
• 72 treated correctly
• Intention-to-treat analysis
• Median time to target temp from ROSC, 120min

36
Bernard et al., NEJM, 2002.

• Results
• Good neuro outcome at discharge (MH vs NT)
• 49 vs 26, p0.045 (n21 vs 9)
• ARR 23 ? NNT 4
• OR for good outcome with MH was 5.25 (1.47-18.5),
  p0.01
• Mortality (MH vs NT)
• 51 vs 68 (95 CI crosses 1)
• Complication rate
• Not stated

37
Bernard et al., NEJM, 2002.

• Take home
• Small study
• Randomization method
• Neurologic benefit
• Mortality benefit not statistically sig
• ?underpowered
• Unblinded treating physicians may have introduced
  treatment bias
• Unable to confirm that outcome assessors were
  blinded to treatment assignment
• Did not publish complication rate

38
Holzer et al. NEJM, 2002.

• Consecutive pts, with witnessed VF/VT CA, gt18yrs,
  CA durationgt5 and lt15mins, ROSClt60mins
• Exclusion criteria
• No thrombolysis
• Randomised to MH (33 deg) using a cooling blanket
  (TheraKool) /- ice packs if required
• Cooling for 24hrs, followed by passive rewarming
• Standard, protocoled intensive care

39
Holzer et al. NEJM, 2002.

• Primary Outcome
• Blinded assessment of neurologic status within
  6mo (Cerebral Performance Category)
• Secondary Outcome
• Mortality
• Rate of complications
• Intention-to-treat analysis for mortality outcome
  only

40
Holzer et al. NEJM, 2002.

• Results
• 3551 pts eligible
• 3426 did not meet inclusion criteria
• 30 excluded for other reasons
• 8 enrolled, 275 ? 175 MH, 138 NT.
• Groups different at baseline for DM/CAD and
  receipt of BLS (all higher in NT group), none
  stat sig.
• Median time to cooling 105mins
• Median time between ROSC and attainment of target
  temp, 8hrs
• Target temp not reached in 19pts
• Hypothermia discontinued early in 14 pts

41
Holzer et al. NEJM, 2002.

• Results
• Favorable neurologic outcome CPC1/2 (MH vs NT)
• 55 vs 39, (RR 1.47, 95 CI 1.09-1.82)
• ARR 16 ? NNT 6.25 (4-25)
• Mortality (MH vs NT)
• 41 vs 55, (RR 0.74, 0.58-0.95)
• ARR 14 ? NNT 7 (4-33)
• Complication rates different between groups but
  not statistically significant (approx 70 of
  patients in both groups)
• 22 more complications MH group (pneumonia
  NNH12, sepsis NNH14)

42
Holzer et al. NEJM, 2002.

• Take home
• Larger study
• Neurologic and mortality benefit
• NNT 6-7 for each end-point
• Establishes that there is a higher rate of
  complications
• Unblinded treating physicians
• Could not verify blinding of outcome assessments.

43
Holzer and Bernard
Trial Feature Holzer Bernard
inclusion differences witnessed VF/VT witnessed VF
Exclusion gt5 lt15 mins to first BLS no restriction to time to BLS
N 275 77
Randomization stratified according to centre even-odd day
Initiation of cooling hospital pre-hospital
Lytics? no yes
cooling device Mattress (TheraKool) ice-packs (CoolCare)
Duration of cooling 24 12
Goal temp 33 33 /-1
Rewarming passive actively at 18hrs
Primary end-point favourable outcome within 6mo at hospital discharge
Result 55 vs 39 49 vs 26
NNT 6.25 4
Mortality 41 vs 55 51 vs 68
Complications NS not stated
44
Lingering questions

• Were groups randomised for all important
  prognostic features?
• Ie brains stem reflexes, gluc
• Blinding of outcome evaluators
• How big a deal is not blinding the treating and
  outcome physicians?
• Optimal time of initiation of cooling
• Re-warming strategy
• Cooling technique

45
Were groups randomised for all important
prognostic features?

• Longstreth. NEJM, 1993.
• 4 criterion model that predicts neurologic
  recovery (awakening) after out-of-hospital VF or
  asystolic CA
• Retrospectively derived and tested
• Predictor variables from ICU admission note
• Median time 2.7days (longest 100days)
• N389
• 50 survived to discharge
• 209 awakened

46
Longstreth. NEJM, 1993.

• Predictor variables
• Motor response (0-4)
• absent, extensor, flexor, non-posturing,
  withdraws or localizes.
• Pupillary light response (3x)
• Spontaneous eye movements
• Glucose lt20mmol/L

47
Longstreth. NEJM, 1993.

• Test Cohort
• Cutoff of gt/4 maximized sensitivity (0.92) and
  specificity (0.65).
• NPV 0.84 PPV .80
• 44 errors in classification
• Majority were of predicted awakening in patients
  who never awakened
• 16 patients predicted not to awaken who did
  awaken
• 12 with severe neurologic defecits
• 4 awakened within 36hrs and made a good recovery
  and returned to pre-arrest functioning.

48
Non-blinding of treating physicians
introduction of bias?

• Schulz, K. Empirical Evidence of Bias. JAMA,
  1995.
• Observational meta-analysis which assessed the
  methodological quality of 250 controlled studies
  on a specific topic
• Determined the associations between those
  assessments and the published treatment effects.

49
Schulz, K. Empirical Evidence of Bias. JAMA, 1995.

• Controlling for allocation concealment
• Trials that were not double-blinded had OR that
  were 17 higher than those trials that were
  double-blinded

50
Timing of cooling.

• When should cooling be initiated?
• When is it too late for cooling to be beneficial?

51
When should cooling be initiated?

• Kuboyama et al. Critical Care Medicine, 1993.
• Prospective, randomised and controlled dog study
• N18 VF arrest 12.5min ? CPB ? defib lt5min ?
  randomised ? standard ICU care for 96hrs
• 3 groups
• NT control
• Immediate IH to 24 deg for 60mins
• Delayed IH for 15mins then maintained for 60mins
• better overall performance category in group 2
  (NS) and improved brain histologic score.

52
Clinical Feasibility Studies

• How to cool
• External blankets
• Ice-packs
• Cranial cooling
• Cold fluid IV infusion

53
Clinical Feasibility Studies

• Felberg, et al. Circulation, 2001.
• 2 Cooling blankets (RK-2000, Thermia)
• Initiation to goal temp took median time of
  300min (goal 120mins).
• Approximately 0.3C/hr

54
Clinical Feasibility Studies

• Callaway, et al., Resuscitation, 2002.
• External cranial cooling
• Ice applied to head and necks of subjects with
  out-of-hospital arrests during CPR
• Rate of temp decrease 0.06 /-0.06 C/min similar
  to control group without ice.
• Likely ineffective

55
Clinical Feasibility Studies

• Hachimi-Idrissi, et al. Resuscitation, 2001.
• Helmet device (Frigicap) containing aqueous
  glycerol placed around the head and neck. Kept in
  refrigerator prior to use. Changed Q1H.

56
Clinical Feasibility Studies

• Hachimi-Idrissi, et al. Resuscitation, 2001.
• Mean starting temp 35.7
• Target bladder temp reached at median time of
  180min after ROSC.
• (Target median tympanic temp reached at 60min)

57
Clinical Feasibility Studies

• Bernard, et al. Resuscitation, 2003.
• 30cc/kg 4 deg LR over 30min for induction
  followed by ice-packs
• N22 CA pts
• No adverse events (including pulmonary edema)
• Median decrease in core (bladder) temp 1.6 deg

58
ILCOR Advisory Statement

• Nolan, J. et al., Circulation, 2003.
• Recommends MH in witnessed, out-of-hospital VF
  arrests with spontaneous ROSC who remain
  unconscious
• Cool to 32-34 degrees for 12 to 24hrs

59
CCU (draft) Protocol

• Inclusion
• Witnessed
• VF or pVT
• 18-75 yrs
• lt15min est to BLS
• lt60mins to ROSC
• MAPgt60

• Exclusion
• Initial Tlt30
• GCSgt9
• Coma prior to CA
• Pregnant
• Terminal illness
• SaO2 lt 85 gt15min
• Shock (?)

60
CCU (draft) Protocol

• Minimum of 10 covered ice packs /- cooling
  blankets prn
• Maintain for 24hrs
• Allow for passive re-warming
• Intubation/sedation/paralysis
• Document neurologic exam prior to initiation

61
Take Home

• Appropriate patients are a highly selected
  sub-group (approx 8 of all-comers with CA)
• Arrest of primary cardiac origin
• Witnessed
• First attempt at resuscitation lt15mins
• ROSClt60mins
• SBPgt90 and SaO2gt90
• Remain unresponsive
• Reasonable pre-arrest quality of life and life
  expectancy
• Insufficient evidence to support ped use

62
Take Home

• Timing of Cooling
• As soon as possible after ROSC (animal studies)
• But, likely beneficial even if delayed upto 6hrs
• ?No longer than 16hrs after ROSC
• Target Temperature
• More research needed
• 33 deg /- 1 deg
• Duration of Cooling
• More research needed
• No longer than 24hrs (may not be any benefit from
  gt12hrs)

63
Take Home

• Cooling Technique
• Unsure if any benefit from rapid induction of MH
  vs gradual onset
• External cooling method (ie. Cooling blanket /-
  ice packs) are the simplest and have been tested.
• Other strategies are experimental (eg. cold RL
  bolus, frigicap) and may be helpful for induction

64
The Near Future

• Formalized relationship with Critical Care
  Medicine (ICU and CCU) with a shared protocol and
  QA strategy.
• Initiated cooling of appropriate patients in the
  ED
• Transfer to the receiving service with minimal
  delay

65
The Not-So-Near Future

• Further studies to define optimal implementation
  of induced hypothermia
• Study whether benefit in other arrest rhythms
• Pediatric population application

66
(No Transcript)