Osmotherapy in pediatric raised ICP, what ICP and NIRS show in real time

1. Three patient vignettes#

Vignette A. Canonical school-age severe TBI#

Yusuf, 9 years old, 28 kg. Severe TBI from a high-speed bicycle vs car collision. GCS 6 at scene; intubated, sedated, paralysed for transport. Day 1 of admission. Right frontal Camino bolt (ICP only), bilateral frontal NIRS pads, arterial line for MAP and PRx, pupillometry every 4 hours. Current state at hour 6: ICP 28, MAP 80, CPP 52, PRx +0.32. Pupils symmetric, NPi 4.0 / 4.1. NIRS L 58% / R 60% (down from baseline 65 / 66). CPP is below the age-band floor (50 mmHg for 7 to 12 years; PBTF recommendation is to keep it comfortably above the floor in severe TBI). The team decides on 3% NaCl 5 mL/kg = 140 mL over 20 minutes through the central line. Expected: ICP fall by 30 minutes, NIRS rise modestly, CPP rise, PRx improve. What actually happens at hour 2 changes the case. 1 2

Vignette B. Infant severe TBI#

Ahmad, 14 months, 9 kg. Severe TBI from an unwitnessed fall (mechanism uncertain; non-accidental injury workup ongoing). GCS 5 at admission. Bifrontal contusions on CT. Anterior fontanelle still open; bedside cranial ultrasound is feasible. Frontal Camino bolt placed (pediatric-sized). Bilateral neonatal NIRS pads (smaller optode footprint). Current state day 1, hour 8: ICP 22, MAP 62, CPP 40, PRx +0.18. The age-banded CPP floor is 40 mmHg for an infant; he is on the floor. Sodium baseline 138; urine output 0.5 mL/kg/h (adequate). The team escalates to 3% NaCl 3 mL/kg = 27 mL over 10 minutes (a smaller per-kg dose given the rapid rise of sodium in infants). Mannitol is the alternative; here HTS is preferred to avoid osmotic diuresis in a young infant whose volume status is fragile. Expected: ICP fall by 20 minutes. The infant-specific point: the open fontanelle allows some compensation, so ICP may not be as informative as in older children; the bulging fontanelle on bedside exam is a complementary sign of evolving raised ICP. 1 3

Vignette C. Atypical: global ICP falls, regional NIRS reveals an expanding contusion#

Yusuf at hour 2 post-bolus. Initial response was textbook: ICP 28 → 19 over 30 minutes; NIRS rose to 63 / 64; CPP 67; PRx fell to +0.18; pupillometry unchanged. The team was content. Sodium drawn at 1 hour: 144 mmol/L (target 145 to 150). At hour 2, L NIRS drops from 63% to 57% over 30 minutes; R unchanged at 64%. Asymmetry now 7%, approaching the 8 to 10% clinical threshold. ICP holds at 19, CPP 64, PRx +0.12, pupils NPi 4.0 / 4.1. Differential: sedation re-deepening (would raise rSO2, not lower it: inconsistent), MAP drift (checked, stable), HTS rebound (uncommon with HTS, reflection coefficient ~1.0), or regional pathology evolving under the L optode. The team repositions the L pad (excludes sensor drift), checks pupillometry (unchanged), confirms PRx stable, then calls for urgent CT. CT shows an evolving left frontal contusion with 4 mm midline shift, under the L NIRS pad but on the contralateral side from the right-sided Camino bolt. The bolt did not see it. NIRS did. Decision: decompressive craniectomy. The lesson: a global metric that looks fine does not exclude regional pathology; NIRS asymmetry is the only modality that would have caught this contusion in time. 4 5

2. The clinical question#

In a child with raised ICP receiving osmotherapy, what does each modality show during and after the bolus, and how do you tell a successful response from a masked complication? The integration question is whether the multimodal trio (ICP + CPP + NIRS, with PRx and pupillometry as supports) confirms uniform brain improvement or reveals a hidden regional problem.

3. Pathophysiology refresher#

Hyperosmolar therapy works by establishing an osmotic gradient across the blood-brain barrier. A bolus of hypertonic saline or mannitol raises plasma osmolality by 5 to 10 mOsm/L; free water shifts from brain interstitium and intracellular compartments into plasma, reducing brain volume and lowering ICP within minutes. The reflection coefficient of the solute matters: sodium has a reflection coefficient near 1.0 (essentially impermeable across an intact BBB), mannitol approximately 0.9. A higher reflection coefficient preserves the osmotic gradient longer and reduces rebound oedema when the BBB is mildly disrupted, the typical post-TBI state. 6 7 8

Time course. ICP begins to fall within 5 to 15 minutes of HTS (faster than mannitol's 15 to 30 minutes) and the effect lasts 4 to 6 hours. Mannitol's effect lasts 2 to 4 hours and is accompanied by an osmotic diuresis that can dehydrate, drop MAP, and worsen CPP, the principal reason most modern pediatric protocols favour HTS over mannitol. 1 9

Why does NIRS rise? Falling ICP raises CPP at constant MAP. Higher CPP, in a brain with at-least-partially-preserved autoregulation, results in higher cerebral blood flow and more oxygen delivery to the cortex under the optode. The rise is usually 3 to 6% in rSO2 and is bilaterally symmetric in global ICP rises. A unilateral rise or fall (asymmetry > 5 to 8%) suggests focal pathology under one optode, the Vignette C scenario. 4 5 10

Why does PRx improve? PRx is the moving-window correlation between MAP and ICP. When ICP is high and CPP is below the lower autoregulation limit (LLA), vessels are maximally dilated; small MAP changes pass directly into ICP changes (positive PRx). When CPP rises into the autoregulatory plateau, vessels can buffer MAP changes; PRx falls toward zero or negative. A PRx that fails to improve after a successful ICP drop suggests autoregulation is broken at any CPP, a worse prognostic sign. 11 12

Sodium and renal load. HTS raises plasma sodium predictably (approximately 1 mmol/L per 1 mL/kg of 3%). The ceiling is 155 to 160 mmol/L; beyond this, the risk of osmotic demyelination on the way down and of renal injury rises. Mannitol does not raise sodium but raises serum osmolality directly and drives an osmotic diuresis. Check sodium 1 hour post-bolus and 4-hourly thereafter during active osmotherapy. 1

Pupillometry adds the brainstem sentinel. Asymmetric pupil dilation or NPi < 3.0 unilaterally is a marker of uncal herniation; bilateral NPi drop to 0 is end-stage. Pupillometry should be unchanged or improving during a successful osmotherapy response; any deterioration despite an ICP fall is a red flag for ongoing herniation independent of the global ICP. 13 14

4. The multimodal picture table#

The most useful pairings: ICP + bilateral NIRS (catches regional masked complications), ICP + CPP + PRx (confirms the haemodynamic mechanism), ICP + pupillometry (brainstem safety net), and ICP + sodium (safety ceiling).

5. Decision tree#

flowchart TD
  Trigger[ICP ≥ 20 sustained > 5 min] --> Tier1{First-line measures done?}
  Tier1 -->|No| First[Head-up 30 deg, sedation, normocapnia]
  Tier1 -->|Yes| Choose{HTS or mannitol?}
  Choose -->|Stable Na, central line ready| HTS[3% NaCl 3 to 5 mL/kg over 10-20 min]
  Choose -->|No central line, hyperNa already| Mannitol[Mannitol 0.5 to 1 g/kg over 10-20 min]
  HTS --> Watch[Watch ICP, NIRS, PRx, pupillometry]
  Mannitol --> Watch
  Watch -->|ICP falls, NIRS rises symmetric, PRx improves| Success[Sustain; recheck Na in 1 h]
  Watch -->|ICP falls but NIRS asymmetric drop| Regional[Urgent CT for focal lesion]
  Watch -->|ICP falls but PRx stays positive| AutoLost[Autoregulation lost; reconsider CPP target]
  Watch -->|ICP does not fall| Refractory{Refractory ICP}
  Refractory --> Tier3[Tier 3: hyperventilation bridge, barbiturate, DC]
  Regional -->|Expanding lesion| Surgery[Decompressive craniectomy or evacuation]
  Success --> Sustain[Recheck Na 4-hourly; Na ceiling 155 to 160]
  Sustain --> NextDose{ICP recurrent?}
  NextDose -->|Yes within 4 h| Repeat[Repeat HTS or transition to infusion]
  NextDose -->|No| Maintain[Continue baseline measures]

6. Step-by-step bedside actions#

1. Verify first-line measures: head-up 30 degrees, neck neutral (no jugular venous obstruction), adequate sedation (RASS −4 to −5), normocapnia (PaCO2 35 to 40), normoxia, normothermia. Do not skip these; they often resolve mild to moderate ICP rises without escalation.
2. Confirm trigger criteria: ICP > 20 mmHg sustained > 5 minutes, or rising trend with falling CPP, or pupillary change.
3. Choose the agent:
   • HTS first if sodium < 155, central line in place, no major heart failure or pulmonary oedema concern.
   • Mannitol if no central line and HTS unavailable, or if volume-overloaded (mannitol's diuresis is helpful here).
4. Calculate the dose:
   • 3% NaCl: 3 to 5 mL/kg over 10 to 20 minutes (max single dose typically 250 mL for adolescents). Yusuf at 28 kg: 84 to 140 mL.
   • Mannitol: 0.5 to 1 g/kg over 10 to 20 minutes. Yusuf at 28 kg: 14 to 28 g (= 70 to 140 mL of 20% mannitol).
5. Document baseline multimodal state in the 5 minutes before the bolus: ICP, MAP, CPP, NIRS L and R, PRx, NPi, sodium.
6. Administer through a central line ideally; peripheral 3% is acceptable for one rescue dose but risks extravasation injury.
7. Recheck multimodal state at 15 min, 30 min, 60 min post-bolus. Expected: ICP fall by 15 min, plateau by 60 min.
8. Send sodium at 1 hour post-bolus, and 4-hourly thereafter during sustained osmotherapy. Stop or switch if sodium > 155 to 160.
9. If NIRS asymmetry develops, walk the differential (sensor drift, sedation, MAP, regional pathology) before reaching for imaging. Reposition the pad first; if asymmetry persists despite stable global signals, image.
10. Plan the next dose or transition to infusion if ICP rebounds. Continuous 3% NaCl infusion at 0.1 to 1 mL/kg/h is an alternative to repeated bolus.

7. Management ladder and endpoints#

Success looks like: ICP < 20, CPP within age-band or CPPopt band, PRx near zero, NIRS symmetric, sodium 145 to 155, no new infarct on imaging.

Failure looks like: refractory ICP, sodium > 160 with no further effect, expanding focal lesion identified late, sustained NIRS asymmetry not explained by sensor or MAP.

8. Variant subsections#

8.1 HTS vs mannitol#

The pediatric tilt toward HTS is driven by three things: better haemodynamic profile (no diuresis), higher reflection coefficient (less rebound), and easier dosing (mL/kg of a stocked 3% bag). Mannitol remains useful in volume-overloaded patients or where HTS is unavailable. 6 7 1

8.2 Age-banded dosing#

• Neonate / infant (< 1 yr): HTS 3% at 2 to 3 mL/kg over 10 min; sodium watch every hour for the first 4 hours.
• Toddler (1 to 3 yr): HTS 3% at 3 to 4 mL/kg over 10 to 15 min.
• School-age (4 to 12 yr): HTS 3% at 3 to 5 mL/kg over 15 to 20 min.
• Adolescent (13 to 18 yr): HTS 3% at 3 to 5 mL/kg or 250 to 500 mL adult-style bolus over 20 min.

Mannitol pediatric dose: 0.25 to 1 g/kg, common starting dose 0.5 g/kg over 15 to 20 min.

The younger the child, the smaller the per-kg dose and the more frequent the sodium check. 1

8.3 Severe TBI bolus protocol#

In severe TBI, osmotherapy is tier-2 in the pediatric ICP escalation ladder. Sequence: head-of-bed elevation, sedation, CSF drainage if EVD present, then HTS bolus. The bolus is a trial; if ICP does not fall by 30 minutes, escalate rather than re-dosing. Maintain sodium 145 to 155 with continuous HTS infusion if recurrent doses needed.

8.4 DKA pre-emptive osmotherapy#

When DKA cerebral oedema is suspected (Glaser criteria), HTS or mannitol is given before CT confirmation. See the DKA cerebral oedema integration for the full Asher case. The dose is the same (3% NaCl 5 mL/kg); the urgency is higher because the trajectory from headache to herniation can be 1 to 2 hours.

8.5 Post-stroke malignant MCA syndrome#

Large MCA infarcts cause swelling and herniation in days 2 to 5. HTS can buy time pending decompressive craniectomy in adolescents and adults; in young children, DC is more often the first option. Pediatric malignant MCA is rare; adult evidence and centre-specific protocols dominate. NIRS asymmetry over the infarct hemisphere is informative. 9

8.6 Refractory case: sodium ceiling reached#

When sodium hits 158 to 160 and ICP is still elevated, options narrow: switch to mannitol (different osmotic mechanism, no further sodium rise), barbiturate coma (deep CMRO2 suppression), or DC. The multimodal monitoring continues to inform: a brain with PRx +0.5 will not benefit from raising CPP; one with stable PRx might. Always include the neurosurgeon in the conversation early.

9. Multimodal integration matrix#

10. Worked alternative scenarios#

10.1 What if ICP does not fall after the bolus?#

Yusuf at 30 min: ICP still 27. CPP 53. NIRS unchanged. Differential: dose too small (rare with 5 mL/kg), HTS leakage (the line was patent), severe BBB disruption (rebound osmotic gradient lost quickly), or a non-osmotic-responsive cause like an expanding mass lesion. The action: confirm line patency, repeat sodium (was the bolus actually delivered?), and escalate to imaging to look for a surgical lesion. Refractory ICP after a correctly-administered HTS bolus is an indication for urgent CT.

10.2 What if NIRS falls instead of rises?#

ICP fell from 28 to 19 as expected, but NIRS dropped from 60% bilaterally to 55%. This is paradoxical and worth investigating. Possible causes: a vasovagal MAP dip (check MAP), a sedation dose change (rare to fall this way), or a global cerebral perfusion event coincident with the bolus. Walk the differential before assuming HTS caused the drop; HTS rarely lowers NIRS in a brain with intact perfusion.

10.3 What if pupillometry deteriorates despite ICP falling?#

Yusuf's NPi drops from 4.0 / 4.1 to 3.0 / 4.1 (asymmetric, R now lower) at hour 2, while ICP is holding at 19. Uncal herniation can progress despite a falling global ICP if a focal lesion is expanding and herniating against the brainstem. The asymmetric NIRS in Vignette C, plus deteriorating pupillometry, plus stable global ICP, is a near-classic picture of a missed contralateral expanding contusion. Urgent CT, then surgical decision.

11. Outcome data#

• Roberts 2019 Cochrane review: pediatric HTS in severe TBI lowers ICP effectively; no clear outcome benefit over mannitol on pooled data, but adverse event profile favours HTS. 9
• Cottenceau 2011 RCT (adult): HTS vs mannitol in 47 severe TBI; both lower ICP, HTS lowers it faster and with greater amplitude. 6
• Mortazavi 2012 meta-analysis: 14 studies (largely adult); HTS more effective than mannitol for refractory ICP. 7
• Kochanek 2019 PBTF guidelines (4th edition): HTS recommended as first-line osmotherapy in pediatric severe TBI; mannitol acceptable alternative; sodium ceiling 155 to 160 mmol/L. 1 2 3
• Fisher 1992: foundational paper on the reflection-coefficient concept that underlies all osmotherapy. 8
• Davies 2017: NIRS in pediatric TBI; sensitivity for regional ischaemia, limitations of optode footprint. 4
• Andresen 2014: NIRS validation in pediatric ICU; asymmetric drop > 5 to 8% is the actionable threshold. 5

12. Pitfalls#

• Treating mild ICP rises with osmotherapy. Tier 0 measures (head-of-bed, sedation, normocapnia) resolve most transient rises; osmotherapy is tier 2 for sustained ICP > 20.
• Skipping the sodium check. Sodium climbs predictably; missing it leads to hypernatraemia, with brain shrinkage on the way up and demyelination risk on the way down.
• Ignoring NIRS asymmetry. A focal contusion under the unsampled hemisphere is the canonical missed lesion; bilateral NIRS catches it.
• Repeating boluses without reassessing. Each bolus is a trial; if the first does not work, investigate rather than repeat. Refractory ICP needs imaging, not more dose.
• Mannitol diuresis crashes MAP. Mannitol's osmotic diuresis can drop MAP by 10 to 20 mmHg over the first hour; in a patient with borderline CPP this can cause secondary insult. HTS avoids this.
• Peripheral HTS in the wrong line. 3% NaCl can cause extravasation injury through small peripheral cannulae; use a large-bore peripheral or a central line.
• Ignoring pupillometry while watching the bolt. The brainstem can herniate even with a falling global ICP; pupillometry is the brainstem-specific check.
• Hyperventilation as default tier 2. Sustained hyperventilation worsens outcome (vasoconstriction-induced ischaemia); use only as a brief bridge for impending herniation, not as a chronic ICP-control strategy. 16

13. Pediatric considerations#

14. Combine with#

• ICP monitoring: the foundation for the bolus trigger.
• NIRS: bilateral regional surveillance.
• PRx and autoregulation: why PRx improvement matters.
• Pupillometry: brainstem sentinel.
• CPPopt targeting integration: how the post-bolus CPP relates to the CPPopt band.
• DKA cerebral oedema integration: pre-emptive osmotherapy.
• Meningitis / encephalitis integration: non-invasive ICP triggers for osmotherapy.
• Brain death MNM: when osmotherapy is no longer indicated.

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

Foundational#

Recent literature (2022 to 2025)#

• Helbok 2024 pediatric MMM update: HTS as first-line osmotherapy; NIRS as regional surveillance; sodium ceiling 155 to 160. 17
• Figaji 2025 pediatric MMM consensus: same recommendations; emphasises the multimodal stack rather than ICP alone for treatment decisions. 18
• Tasker 2023 PCCM review: osmotherapy in the broader pediatric MMM framework; tier-2 escalation with a focus on convergent multimodal endpoints. 19
• Hawthorne 2014 ICP review (still cited): ICP-dose framework that underpins the rationale for aggressive control. 20
• Hartings 2024 SD intervention review: spreading depolarisations as a parallel mechanism that osmotherapy may also dampen indirectly. 21
• Kochanek 2019 PBTF (still the operative guideline): the most current pediatric severe TBI bundle; HTS, sedation, CSF drainage, and DC in escalating tiers. 1

16. Self-check#