MNM in the newborn, HIE, the cooling bundle, and the prognostic conversation

1. Three patient vignettes#

Vignette A. Canonical Sarnat 2 cooled HIE#

Layla, term (39+5 weeks), 3.4 kg, day 2 (post-natal hour 36). Born by emergency caesarean for prolonged decelerations; Apgars 2 / 4 / 6, cord pH 6.91, base deficit 18, lactate 9.2 mmol/L. Required positive-pressure ventilation, intubation, 5 minutes of chest compressions. Sarnat stage 2 encephalopathy at 1 hour (hypotonia, weak suck, abnormal Moro, intermittent agitation). Met TOBY / NICHD cooling-eligibility criteria. Cooling started at 4 hours of life; on day 2 currently at 33.5 degrees, paralysed and sedated for cooling. aEEG over 36 hours: initial isoelectric trace, evolved by hour 18 to discontinuous (bursts > 10 microvolts separated by < 5 microvolt interburst intervals), now showing early sleep-wake cycling at hour 36. NIRS rSO2 bilateral 78 / 79 (a touch on the high side; cooling raises rSO2 by ~3 to 5% via reduced CMRO2). TCD MCA: PSV 65, EDV 35, PI 0.65 (low, reflecting reduced cerebral vascular resistance and reperfusion). Pupils PERLA, NPi 4.0 / 4.0. The question: how does the bundle predict day-7 MRI and 18-month outcome? 1 2 3

Vignette B. Premature (33-week) NE-like presentation#

Khalid, born at 33 weeks, 1.6 kg, day 3. No clear perinatal asphyxia; intrauterine growth restriction with placental insufficiency. Sarnat-like clinical picture (hypotonia, abnormal feeding behaviour), but premature infants are generally excluded from therapeutic hypothermia (the cooling trials enrolled ≥ 36 weeks). aEEG patterns are different in preemies: more discontinuity is normal, sleep-wake cycling emerges later. Bilateral NIRS rSO2 65 / 62 (lower than term norms; preterm baseline is 55 to 75%). Head ultrasound: bilateral periventricular echo-densities suggestive of evolving PVL. The team focuses on neuroprotection (avoid hypotension, treat seizures aggressively, normothermia), serial HU surveillance for IVH and PVL, and MRI at term-equivalent age. The premature point: cooling is not standard, the MNM bundle is different (more weight on HU than MRI in the acute phase), and the prognostic conversation is more uncertain. 4 5

Vignette C. Atypical: severe HIE with luxury perfusion masking depth of injury#

Yusra, term (40+1), 3.2 kg, day 2. Sarnat stage 3 (deep coma, no spontaneous movement, no brainstem reflexes except weak pupillary, frequent seizures). Cooled. NIRS rSO2 bilateral 90 / 91% (very high, "luxury perfusion"). TCD: PI 0.32, EDV 65, MFV 95 (high diastolic flow, very low pulsatility). aEEG isoelectric since hour 0 with no evolution at hour 48. The paradox: the brain looks "well oxygenated" on NIRS, but is functionally dead on aEEG. Luxury perfusion in HIE is a bad prognostic sign: blood flows through a brain whose metabolic demand has collapsed; the high rSO2 reflects loss of extraction, not adequate function. Pair this with the aEEG and the picture is one of severe injury despite reassuring oxygenation numbers. MRI day 5: extensive bilateral basal ganglia, thalamus, perirolandic, hippocampal, and subcortical white-matter signal abnormality on DWI. The prognostic conversation: severe injury, very poor outcome anticipated; transition to compassionate care discussions in coordination with neonatal palliative care. The lesson: NIRS rSO2 alone is not reassuring in HIE; pair with aEEG and the clinical exam. 6 2

2. The clinical question#

In the term newborn with HIE under cooling, what does each modality contribute to acute neuroprotection, and how does the multimodal bundle frame the day 4 to 7 prognostic conversation with the family? The integration question is how aEEG, NIRS, TCD, clinical exam, and MRI combine into a single coherent prognostic narrative.

3. Pathophysiology refresher#

Hypoxic-ischaemic encephalopathy in the term newborn results from a perinatal sentinel event (placental abruption, prolapsed cord, severe shoulder dystocia, uterine rupture) that interrupts placental gas exchange. The fetus tolerates short interruptions through redistribution of cardiac output (brain-sparing); prolonged interruption (typically > 20 to 30 minutes) overwhelms compensation and produces a global hypoxic-ischaemic insult. Subsequent reperfusion injury in the first 6 to 48 hours involves excitotoxicity, oxidative stress, mitochondrial dysfunction, and delayed apoptosis. Therapeutic hypothermia interrupts this cascade by lowering cerebral metabolic rate (CMRO2 falls ~6 to 8% per degree), reducing excitotoxicity, and dampening apoptotic signalling. Cooling within 6 hours of birth for 72 hours at 33 to 34 degrees core temperature is the standard intervention. 1 7

The aEEG trajectory in HIE is the bedside backbone of acute monitoring. Toet 1999 established the four canonical background patterns:

1. Continuous normal voltage (CNV): lower border > 5 microvolts, upper border > 10 microvolts. Normal or mild HIE.
2. Discontinuous (DC): lower border < 5 microvolts, upper border > 10 microvolts. Moderate HIE.
3. Burst-suppression (BS): low-voltage background interrupted by high-amplitude bursts. Severe HIE.
4. Continuous low voltage (CLV) or isoelectric / flat: both borders < 5 microvolts or < 2 microvolts (flat). Very severe HIE.

The evolution matters more than the initial pattern: a previously isoelectric trace evolving to discontinuous by 24 to 48 hours is consistent with recovery; persistent isoelectric or burst-suppression at 36 to 48 hours indicates very severe injury. Sleep-wake cycling typically emerges by 24 to 72 hours in cooled infants with recovery potential. 2 3

NIRS rSO2 in HIE behaves counterintuitively. In severe HIE with luxury perfusion, rSO2 is paradoxically high (often 85 to 95%) because the brain has lost its ability to extract oxygen (collapsed metabolic demand). A "reassuring" high NIRS in a deeply encephalopathic newborn is therefore a bad sign, not a good one. The interpretation requires the aEEG and clinical context. Conversely, a low rSO2 (< 50%) suggests hypoperfusion or extraction beyond delivery; both are concerning. 8 9 6

TCD in HIE also has counterintuitive patterns. Low PI with high EDV = luxury perfusion (reduced cerebral vascular resistance from injured vessels), often seen days 1 to 3 in severe HIE and associated with poor outcome. High PI with low EDV = high cerebrovascular resistance, often raised ICP from oedema, also associated with poor outcome. Normal PI (0.65 to 0.95 in term newborns) with normal MFV is reassuring. 6

MRI day 4 to 7 is the single most predictive modality. The injury pattern reveals the mechanism and severity: basal ganglia and thalamus (BGT) pattern suggests acute profound hypoxia-ischaemia (worst prognosis); watershed pattern (parasagittal cortex and subcortical white matter) suggests partial prolonged hypoxia-ischaemia (better prognosis if isolated); diffuse cortical / subcortical pattern suggests severe global injury; normal MRI in a clinically encephalopathic infant has good prognostic implications. 1 10

The Sarnat clinical staging remains the primary clinical anchor: stage 1 (mild, hyperalert), stage 2 (moderate, lethargic, hypotonic, weak suck, frequent seizures), stage 3 (severe, deep coma, no spontaneous movement, absent brainstem reflexes). Stages 2 and 3 are cooling-eligible; stage 1 is generally not. 1

Seizures in HIE occur in 30 to 50%. Most are subclinical (cooling and sedation mask convulsive activity). cEEG is the gold standard; aEEG detects 60 to 80%. Treatment: phenobarbital 20 mg/kg loading is the historical first-line; levetiracetam 60 mg/kg is increasingly used (less sedation, better cognitive profile). 4 11

4. The multimodal picture table#

The most useful pairings in neonatal HIE: aEEG + clinical exam + MRI (the canonical prognostic triad), NIRS + aEEG (catches luxury perfusion), and cEEG + aEEG (seizure detection plus continuity trajectory).

5. Decision tree#

flowchart TD
  Start[Term newborn, perinatal asphyxia, Sarnat 2 or 3] --> Cool{Cool within 6 h?}
  Cool -->|Yes| Cooling[TH 33-34 deg for 72 h]
  Cool -->|No, > 6 h| LateCool[Late TH controversial; case-by-case]
  Cooling --> Bundle[Continuous aEEG + bilateral NIRS + TCD; daily exam]
  Bundle --> aEEG24{aEEG at 24 to 48 h}
  aEEG24 -->|Continuous, sleep-wake cycling| Good[Better prognosis; continue bundle]
  aEEG24 -->|Discontinuous, evolving| Inter[Intermediate; reassess at 72 h]
  aEEG24 -->|Persistent isoelectric or BS| Severe[Severe; alert family; plan MRI day 4 to 7]
  Bundle --> Seizures{Seizures on aEEG or cEEG?}
  Seizures -->|Yes| Treat[Phenobarbital 20 mg/kg or levetiracetam 60 mg/kg]
  Seizures -->|No| Continue1[Continue surveillance]
  Treat --> Recheck[cEEG endpoint: seizure control]
  Inter --> Day4[MRI day 4 to 7]
  Severe --> Day4
  Good --> Day4
  Day4 --> Pattern{Injury pattern}
  Pattern -->|Normal or minor| OutcomeGood[Likely favourable outcome]
  Pattern -->|Watershed| OutcomeMod[Mixed; cognitive and motor deficits possible]
  Pattern -->|BGT or global| OutcomeSevere[Severe disability or death; palliative discussion]
  Pattern -->|Diffuse cortical / subcortical| OutcomeSevere
  OutcomeSevere --> Family[Coordinated prognostic conversation with neonatal palliative care]

6. Step-by-step bedside actions#

1. Start cooling within 6 hours of birth for eligible infants (TOBY / NICHD: pH ≤ 7.00 OR base deficit ≥ 16 PLUS Apgars ≤ 5 at 10 minutes OR continued resuscitation PLUS Sarnat 2 to 3 encephalopathy). Target core 33 to 34 degrees for 72 hours.
2. Place continuous aEEG immediately. Two-channel (C3-P3, C4-P4) for amplitude; expand to multi-channel cEEG if seizures suspected. Record continuously through cooling and rewarming.
3. Bilateral NIRS pads on, frontal placement; document baseline rSO2 in the first 6 hours; expect cooling to raise rSO2 by 3 to 5%.
4. Daily TCD MCA for PI and MFV; document baseline; interpret in the context of clinical state.
5. Daily neurological exam by a senior physician: pupillary, gag, suck, posture, tone, spontaneous movement, primitive reflexes; document the Sarnat stage trajectory.
6. Treat seizures aggressively. Phenobarbital 20 mg/kg loading (may give second 10 mg/kg dose if persistent); levetiracetam 60 mg/kg an emerging alternative; midazolam 0.05 to 0.2 mg/kg/h infusion for refractory.
7. Avoid hyperthermia during rewarming (the rebound risk); aim for ≤ 0.5 degrees per hour rewarming rate; document.
8. Day 4 to 7 MRI with DWI, T1, T2, MR spectroscopy if available. Interpret with neuroradiology; the injury pattern is the primary prognostic input.
9. Coordinate the prognostic conversation with neonatal palliative care, the bedside nurse, and family liaison. Use the multimodal evidence (aEEG, MRI, clinical exam, NIRS) to support an honest evidence-based discussion.
10. Plan post-discharge follow-up at 6, 12, and 24 months with developmental assessment; the day 4 to 7 MRI predicts but does not perfectly determine outcome.

7. Management ladder and endpoints#

Success looks like: completed cooling without complications, evolving aEEG continuity, MRI without major injury or with limited watershed only, age-appropriate developmental milestones at 12 to 24 months.

Failure looks like: persistent isoelectric aEEG at 72 hours, BGT pattern on MRI, severe multi-organ failure, leading to compassionate care transitions.

8. Variant subsections#

8.1 The cooling bundle in detail#

The NICHD (Shankaran 2005), CoolCap, and TOBY trials established whole-body cooling at 33.5 degrees for 72 hours as the standard of care for moderate-to-severe HIE in term infants when started within 6 hours of birth. Number needed to treat: approximately 7 for prevention of death or severe disability at 18 months. Cooling reduces CMRO2, dampens excitotoxicity, and limits the delayed phase of apoptotic neuronal death. 1

8.2 aEEG interpretation in detail#

aEEG compresses 2 to 4 EEG channels into a single time-amplitude plot. The four canonical patterns (CNV, DC, BS, CLV / isoelectric) form a severity hierarchy. Sleep-wake cycling (the slow rhythmic 30 to 60 minute oscillation of the amplitude envelope) emerging by 36 to 72 hours predicts favourable outcome. Persistent isoelectric beyond 36 hours indicates severe injury. 2 3 12

8.3 NIRS interpretation in HIE#

Normal term newborn rSO2 is 65 to 75%. Luxury perfusion (> 85%, often > 90% in severe HIE) reflects collapsed metabolic demand; counterintuitively a bad sign. Hypoperfusion (rSO2 < 50%) signals inadequate delivery or excess extraction; also bad. NIRS asymmetry > 5 to 8% suggests focal pathology (uncommon in HIE; more common in stroke). The interpretation always requires the aEEG and clinical context. 8 13 9

8.4 TCD patterns in HIE#

Term newborn TCD MCA: PSV ~ 45, EDV ~ 15, MFV ~ 25 cm/s, PI ~ 0.7 to 0.95. Low PI with high EDV (PI < 0.5, EDV > 30) = luxury perfusion (bad). High PI with low EDV (PI > 1.4, EDV < 10) = high resistance (raised ICP from oedema; bad). Normal PI with low MFV = generalised hypoperfusion. Daily TCD is feasible through thin temporal windows and the anterior fontanelle. 6

8.5 Seizures in HIE#

30 to 50% of cooled HIE infants develop seizures, most subclinical. cEEG detects > 95%; aEEG ~ 60 to 80%. Phenobarbital 20 mg/kg load remains first-line by tradition; levetiracetam 60 mg/kg is increasingly used (less sedation, less aEEG depression, better neurodevelopmental signal in trials). NeoLEV2 and other trials are establishing the comparative evidence base. Seizure burden correlates with worse outcome independent of MRI. 4

8.6 The day 4 to 7 MRI#

The MRI is the single most predictive modality in neonatal HIE. Patterns:

• BGT (basal ganglia and thalamus) pattern: acute profound asphyxia; worst outcome; severe motor and cognitive disability or death.
• Watershed (parasagittal cortex and subcortical white matter) pattern: partial prolonged asphyxia; better than BGT; motor and cognitive deficits variable.
• Diffuse cortical / subcortical: severe global; very poor outcome.
• Focal infarct: not classic HIE; consider stroke aetiology.
• Normal MRI with severe clinical encephalopathy: rare; may indicate metabolic disease, mimic, or recovery beyond the MRI window.

MR spectroscopy adds lactate / NAA ratio in the basal ganglia as a quantitative prognostic marker. 1 10

9. Multimodal integration matrix#

10. Worked alternative scenarios#

10.1 What if the aEEG looks better than the clinical exam?#

A term newborn day 2 of cooling, aEEG shows discontinuous evolving toward continuous with early sleep-wake cycling; clinical exam still Sarnat 3 (deep coma, no movement). The discordance is informative: clinical exam may be lagging because of sedation; the aEEG is suggesting recovery. Continue cooling, defer prognostic conversation until day 4 to 7 MRI, expect clinical improvement as sedation wears off. The MRI is the tiebreaker.

10.2 What if the MRI is normal but the clinical exam is severe?#

A term newborn day 5, clinical Sarnat 3 (no improvement), but MRI shows no definite injury. The differential: metabolic encephalopathy (urea cycle defect, organic acidaemia, mitochondrial disease; check ammonia, lactate, plasma amino acids, urine organic acids), non-accidental injury (subdural collections may not show on day 5 MRI), prolonged ictal-postictal state (review cEEG), or mimic (channelopathy, hyperammonaemia). The MNM bundle now redirects toward metabolic workup and genetic testing. 14

10.3 What if cooling is started after 6 hours?#

A term newborn transferred from a peripheral hospital at 9 hours of life. Outside the conventional cooling window. Late cooling (6 to 24 hours after birth) is controversial; some evidence suggests benefit in selected infants, but the trials were inconsistent. The team should follow local protocols and individualise. MNM bundle remains as for in-window cooling; the prognostic implications are similar by aEEG / clinical exam / MRI trajectory.

11. Outcome data#

• Shankaran 2005 NICHD: whole-body cooling reduced death or moderate-to-severe disability at 18 to 22 months from 62% to 44% (NNT 7). The foundational pediatric HIE trial. 1
• Toet 1999: aEEG patterns at 3 to 6 hours of life predict outcome in HIE; CNV very good, DC intermediate, BS or CLV very poor. 2
• Hellstrom-Westas 2006: review and reference patterns for aEEG in term and preterm newborns. 3
• Moler 2015, 2017 (THAPCA): pediatric cardiac arrest cooling trial; provides comparator evidence for the broader HIE / post-arrest population. 7 10
• Pressler 2017: neonatal seizure detection and treatment guidelines; aEEG and cEEG roles. 4
• Sansevere 2023: neonatal cEEG; subclinical seizures common; aEEG inadequate for some seizure types. 5
• Kirschen 2020: pediatric TCD in HIE / post-arrest; luxury perfusion patterns and prognosis. 6
• Naim 2023: pediatric brain injury MNM update; neonatal applications. 15

12. Pitfalls#

• Trusting NIRS rSO2 in isolation. Luxury perfusion looks reassuring; the aEEG and clinical exam must be paired.
• Missing subclinical seizures. aEEG detects 60 to 80%; switch to cEEG when seizures suspected or clinical evolution is poor.
• Rewarming too fast. Aim ≤ 0.5 degrees per hour; faster rewarming can precipitate seizures and may worsen injury.
• Skipping the day 4 to 7 MRI. It is the most predictive single test; image even if the clinical course looks favourable, because the MRI changes the conversation about long-term follow-up.
• Phenobarbital sedation confounding aEEG. Use the lowest effective dose; consider levetiracetam if available.
• Treating only convulsive seizures. Most HIE seizures are subclinical; aEEG / cEEG is the standard for detection.
• Cooling complications missed. Hypotension, bradycardia, coagulopathy, thrombocytopenia, electrolyte abnormalities are common during cooling; monitor and treat.
• Prognosticating too early. Day 1 to 2 trajectory is informative but not definitive; the family conversation should wait for the MRI and a coherent multimodal picture.
• Forgetting metabolic differential. A "normal MRI" in severe encephalopathy mandates a metabolic workup; do not assume HIE without confirming the perinatal asphyxia history.

13. Pediatric considerations#

14. Combine with#

• aEEG and continuous EEG: the bedside electrophysiology backbone.
• NIRS: rSO2 interpretation including luxury perfusion.
• TCD: neonatal PI and MFV patterns.
• Pupillometry: brainstem function in encephalopathy.
• MNM on ECMO integration: neonatal ECMO MNM bundle overlaps.
• Brain death MNM integration: when HIE progresses to brain death.
• Family communication MNM: the prognostic conversation framework.
• Inborn errors of metabolism integration: the metabolic differential.

15. Evidence summary and recent literature (2022 to 2025)#

Foundational#

Recent literature (2022 to 2025)#

• Sansevere 2023: neonatal cEEG detects subclinical seizures missed by aEEG; ECMO and HIE both cited. 5
• Naim 2023 PCCM: pediatric brain injury monitoring update; integrates neonatal HIE applications. 15
• Helbok 2024 pediatric MMM update: neonatal HIE in the broader pediatric MMM framework. 17
• Figaji 2025 pediatric MMM consensus: neonatal HIE bundle recommendations; aEEG, NIRS, TCD, MRI. 18
• Tasker 2023 PCCM review: integrative pediatric MMM review; neonatal HIE chapter. 19
• Plomgaard 2024 SafeBoosC III: NIRS-guided care in preterm infants; relevant to broader neonatal NIRS evidence base. 20

16. Self-check#