MNM
CPPopt · DERIVED

Optimal CPP (CPPopt)

The CPP at which the brain's pressure reactivity is most negative, derived from a 4-hour parabolic fit of PRx versus CPP. The individualised perfusion target.

ReactivityBedside + researchPeds + adultInvasiveEmerging
BLast reviewed 2026-05-1714-min read

1. Bedside vignettes: why this matters in the PICU

Vignette A. The 8-hour COGiTATE loop in a 12-year-old severe TBI

A 12-year-old severe TBI day 3. ICP 14 mmHg, MAP 80, CPP 66. Default age-based CPP floor is 60. The bedside team runs a CPPopt loop using the unit's ICM+ workstation:

  • Hour 0–4: accumulate data, PRx oscillating between 0 and +0.3. CPPopt fit gives vertex at CPP 72 mmHg with a clear parabola. Lift MAP with noradrenaline, CPP rises to 71.
  • Hour 4–8: PRx falls to 0.05 within the new operating range. CPPopt vertex shifts to CPP 73, stable. Continue.
  • Hour 8–12: a small subdural extension on the right (re-CT for sudden pupil change) raises ICP to 22. CPPopt recalculates to 78 because the brain is now less compliant; team raises MAP further.

The loop followed the patient through a physiological change rather than holding a static number. Over the 24-hour window the ICP-dose accumulated was lower than the matched-protocol cohort, and the family reported sustained improvement in the patient's GOS at 6 months.

Vignette B. CPPopt unobtainable after 4 hours

A 6-year-old severe TBI day 1. After 4 hours of clean recordings the (CPP, PRx) plot is flat: PRx hovers near +0.15 across every CPP bin from 50 to 90. No vertex, no parabola. What to do?

  • First check: signal quality. Re-zero the arterial line; confirm ICP transducer is reading a clean three-peak waveform. Both pass.
  • Second check: did the patient experience MAP variation across the 4-hour window? If MAP was tightly controlled around a narrow range (say 65–72), the algorithm has no data outside that band and cannot fit a U-curve.
  • Third check: is the patient on a high-dose vasopressor that pegs MAP within ±5 mmHg of a fixed target? Consider a deliberate small MAP excursion (lower noradrenaline slightly for 30 minutes, then raise) to give the fit data.
  • Fallback: revert to age-based CPP floor (CPP > 50 for a 6-year-old) until a fittable CPPopt is available.

CPPopt is a fitted parameter; it is not always fittable. Do not invent a CPPopt.

Vignette C. Pediatric difference: CPPopt 50 in a 3-year-old vs default 60

A 3-year-old severe TBI day 2. ICP 11 mmHg, MAP 65, CPP 54. Default team protocol says "CPP < 60 means action": team starts a noradrenaline infusion to lift MAP. The PRx, computed over the past 4 hours, shows a clean U-curve with CPPopt at CPP 48 mmHg. The team's noradrenaline took CPP to 64, where PRx rose to +0.30. The intervention pushed the patient above CPPopt. They wean noradrenaline back to baseline; CPP returns to 54; PRx falls to −0.10.

This is the most important pediatric lesson from the pediatric CPPopt data (Velle 2023; Lewis 2015): younger children have lower CPPopt and the adult-style default of 60–70 mmHg can be harmful by causing over-perfusion. Trust the data, not the default.

2. What CPPopt is, and what it is not

CPPopt is the CPP coordinate at the vertex of a parabolic fit through (CPP, PRx) data over the past 4 hours.

CPPopt=arg⁡min⁡CPP[PRx(CPP)]\mathrm{CPPopt} = \arg\min_{CPP} \left[ \mathrm{PRx}(CPP) \right]

where PRx(CPP) is the binned mean of PRx values in each 5-mmHg CPP bucket.

The logic. PRx is minimised when autoregulation is most efficient: small CPP changes do not drive ICP changes. Below CPPopt the brain is under-perfused and autoregulation fails by passive collapse (vessels dilate but cannot maintain flow); above CPPopt the brain is over-perfused and autoregulation fails by hyperaemia (vessels saturated at maximum constriction). The shape of PRx-vs-CPP is therefore a U-curve, and the vertex is the patient's biological sweet spot.

Two things follow.

CPPopt requires PRx as the input. Without continuous high-quality ICP and MAP signals, CPPopt cannot be derived. Non-invasive surrogates (Mx with TCD, COx with NIRS) can substitute as input, with differing sample volumes and noise profiles.

CPPopt is dynamic. A patient's CPPopt drifts over hours-to-days. The 4-hour moving window is a compromise between responsiveness and noise suppression. Within-patient CPPopt is not fixed and should be re-computed continuously rather than set once.

Clinical pearl

CPPopt is not a number to set and forget. It is a moving target that responds to the brain's state. Re-fit every hour and update the operational target accordingly.

In children

Pediatric CPPopt is age-dependent and often below the adult-style default. The 3-year-old in Vignette C is the canonical example: CPPopt 48 vs default 60 means a 12 mmHg over-perfusion if the default is held rigidly. Younger brains have lower baseline cerebral metabolic rate and lower autoregulatory floor; trust the data.

3. The five-step workflow

Fig. 1
THE CPPopt WORKFLOWaccumulate, compute, bin, fit, re-target, then loopStep 1 · accumulate4 h ICP + MAP at100-200 Hz, synced toone clockStep 2 · compute PRx5-min windows, updateevery 60 s (~200+usable after artefactreject)Step 3 · bin5-mmHg CPP buckets;mean PRx in each(drop <~5 points)Step 4 · fitparabola throughbucket means; vertex= CPPopt (flat = notavailable)Step 5 · re-targetadjust MAP so CPPsits within CPPopt+/-5 mmHgloop every 30-60 minMNM-Edu schematic · adapted from Aries 2012
The CPPopt workflow. Step 1: accumulate 4 hours of synchronous ICP and MAP recordings at 100–200 Hz. Step 2: compute PRx in 5-minute moving windows updating every minute (288 windows over 4 hours, after artefact rejection typically 200+ remain). Step 3: bin the (CPP, PRx) points into 5-mmHg CPP buckets, take the mean PRx in each bucket. Step 4: fit a parabola (or smoothed U-curve) to the binned means; the vertex is CPPopt. Step 5: re-target MAP so CPP sits within ±5 mmHg of CPPopt. Repeat the loop every 30–60 minutes.
MNM-Edu, original schematic adapted from Aries 2012.

Step 1: Accumulate

Continuous ICP and MAP at 100–200 Hz from the patient monitor, synchronised to a common clock. ICM+ or equivalent software handles this. Four hours of data is the conventional window for first-fit; longer windows reduce noise but slow responsiveness.

Step 2: Compute PRx

5-minute moving windows of 30 paired 10-second averages. Update every 60 seconds. Reject windows containing flagged artefacts (motion, suction, transducer flush, low MAP variance). See PRx page for the math.

Step 3: Bin

Place each 5-minute (CPP, PRx) point into a 5-mmHg CPP bucket (e.g., CPP 60–64 mmHg, CPP 65–69 mmHg, …). Compute the mean PRx in each bucket. Buckets with fewer than ~5 points are excluded as undersampled.

Step 4: Fit

Fit a parabola (least-squares) through the bucket means. The vertex of the parabola is CPPopt. A clean U-curve has a clear vertex; a flat curve has no fittable vertex and CPPopt is "not available".

Step 5: Re-target

Set the operational CPP target as CPPopt ± 5 mmHg. Adjust MAP (via fluid, vasopressor, or sedation lightening) to bring CPP into this band. Re-evaluate the loop every 30–60 minutes.

Caveat

A flat curve is not a "wide CPPopt", it is "CPPopt unobtainable". Setting the target at the midpoint of a flat curve is inventing a number. Revert to age-based default and try again next hour.

4. Try it: interactive widgets

CPPoptUCurve
Loading widget…
Interactive tool (view online).
MAPoptUCurve
Loading widget…
Interactive tool (view online).
PRxCalculator
Loading widget…
Interactive tool (view online).
MxCalculator
Loading widget…
Interactive tool (view online).
MultimodalDiscordance
Loading widget…
Interactive tool (view online).

5. Dose-response: time outside CPPopt versus outcome

Fig. 2
OUTCOME vs DEVIATION FROM CPPoptAries 2012 Fig 3 · best near CPPopt · worse in BOTH directions (asymmetrical)-15-10-50+5+10+15Median CPP - CPPopt (mmHg)02040606-month favourable GOS (%)at CPPoptbelow CPPopt:rising mortalityabove CPPopt:rising severe disabilityTHE ASYMMETRYOutcome is best whenmedian CPP sits near theindividual CPPopt.left armCPP below CPPopt:rising mortality.right armCPP above CPPopt:rising severe disability.n = 299 · adult severe TBIcohort CPPopt approx 75-80paediatric CPPopt is lowerMNM-Edu schematic · Aries 2012 (Fig 3) · 6-month GOS · adult severe TBI, paediatric CPPopt lower and age-dependent
Outcome versus deviation from CPPopt. In the Aries 2012 severe-TBI cohort (several hundred adults), six-month GOS outcome was best when median CPP sat close to the individualised CPPopt and worsened with deviation in either direction, an asymmetrical U. The arms differ in mechanism: CPP below CPPopt tracked rising mortality, CPP above CPPopt tracked rising severe disability. This is adult severe TBI; paediatric CPPopt targets are lower and age-dependent. Aries 2012 (Fig 3).
MNM-Edu schematic, adapted from Aries 2012 (Fig 3).

The Aries 2012 cohort showed:

  • Time spent with CPP below CPPopt predicted mortality and severe disability.
  • Time spent with CPP above CPPopt predicted severe disability (over-perfusion injury).
  • The dose-response was continuous; no safe "outside CPPopt" threshold below which outcome was unchanged.

Depreitere 2014 extended this with high-resolution data showing the ICP-dose framework integrates with CPPopt-dose. COGiTATE (Tas 2021) moved from observational to interventional: the phase II trial demonstrated feasibility and safety of actively targeting CPPopt and increased time within target band (46.5% vs a 36% feasibility target).

6. Trial evidence: COGiTATE, the follow-up, and pediatric extension

6.1 COGiTATE phase II (Tas 2021)

The first randomised feasibility trial of PRx-CPPopt-guided care in adult severe TBI.

  • Design: multicentre RCT, ~60 adults severe TBI, randomised to CPPopt-guided vs standard CPP 60–70 care for the first 72 hours.
  • Primary endpoint: feasibility. Met: the CPPopt-guided arm kept CPP within ±5 mmHg of CPPopt for 46.5% of monitored time (vs a 36% feasibility target), and spent less time below CPPopt than controls (19.1% vs 34.6%).
  • Safety endpoints: no excess fluid balance, no excess vasopressor use, no increase in therapeutic intensity.
  • Outcome: not powered for functional outcome; no mortality benefit was demonstrated.

The result establishes the framework as safe and feasible at scale.

6.2 Pediatric CPPopt (Velle 2023, STARSHIP)

Pediatric evidence is observational. Velle 2023 (single-centre, n=57) and Svedung Wettervik 2023 (n=61) showed CPPopt is computable in children with severe TBI, and STARSHIP (Agrawal 2023) is the major multicentre pediatric autoregulation study with published outcome analyses. A lower CPPopt in younger children and an association between deviation from CPPopt and outcome have been described; specific pediatric threshold values should be taken from these primary sources.

The pediatric evidence base is observational; no pediatric RCT exists yet.

6.3 Donnelly 2017 MAPopt extension

When ICP varies throughout the day (e.g., during decompression recovery, or with CSF drainage), holding CPP constant requires constant MAP adjustment. The MAPopt extension keeps the optimisation target on MAP directly (which is the variable the team controls), simplifying the bedside loop. Useful when ICP is being managed via EVD drainage.

7. Clinical contexts: CPPopt across acute brain injuries

7.1 Severe TBI (the validated indication)

Adult severe TBI is the canonical context. Pediatric severe TBI (Velle 2023; Lewis 2015) extends the framework with age-specific CPPopt values. The BTF 4 pediatric guidelines and a pediatric MMM review (Laws 2022) place PRx-based CPPopt as a specialist-centre modality where resources allow.

7.2 SAH (growing use)

CPPopt in SAH is less validated. The principle applies: if ICP is being monitored (via EVD for hydrocephalus or DCI), adding PRx-CPPopt is straightforward. AHA/ASA 2023 SAH guidelines acknowledge MMM including autoregulation indices as tier-2 modalities.

7.3 Pediatric severe TBI (Tas et al)

Velle 2023 (n=57): CPPopt is computable in children with severe TBI, with a lower CPPopt in younger children. STARSHIP (Agrawal 2023) is the multicentre pediatric autoregulation study. A pediatric MMM review (Laws 2022) places PRx-based CPPopt among specialist-centre modalities.

7.4 HIE / post-cardiac arrest (research only)

PRx signal quality is poor in early post-arrest (deep sedation, hypothermia, low slow-wave power). COx (NIRS-MAP) is more usable; a "COxopt" framework is in development. Not yet validated for routine use. Sparse

7.5 ECMO (research only)

Non-pulsatile VA-ECMO destroys PRx interpretability. COx-based CPPopt has been demonstrated in case series. Sparse

8. Multimodal integration: CPPopt inputs and endpoints

Input modalityOutput (CPPopt)When to use
PRx (ICP + MAP)Canonical CPPoptSevere TBI, SAH with EVD; whole-brain ICP available
Mx (TCD-MFV + CPP)Non-invasive CPPoptNo ICP placed; pediatric without invasive monitoring; resource-limited
COx (NIRS-rSO2 + MAP)Regional cortical, NIRS-derived autoregulationECMO, low-pulsatility flow, post-arrest, cardiac surgery, CHD
Pair with…What you gainWorked scenario
PbtO2Endpoint validation: target CPPopt and confirm tissue O2 is in rangePbtO2 page
MicrodialysisMetabolic endpoint: L/P ratio falls when CPPopt achievedMicrodialysis page
EEG / aEEGReactivity returns when CPPopt achieved in HIEEEG page
TCD / MxCross-validate CPPopt with Mx-CPPoptTCD page
NIRS / COxCross-validate with COx-CPPoptNIRS page
ICPThe input signal; quality is everythingICP page

9. Pitfalls and artefacts

  • Signal quality: a noisy ICP or arterial line poisons the fit. Validate every shift.
  • MAP tightly controlled: if MAP varies less than the bin width (5 mmHg), the algorithm has no data outside the operating range and cannot fit. Allow physiological MAP variation across the recording window.
  • Vasopressor dose pegging MAP: noradrenaline infusions held at a fixed target can produce a flat fit. A deliberate small MAP excursion (with bedside consent and senior oversight) provides data.
  • Flat curve: not a wide CPPopt; CPPopt is unobtainable. Revert to default and try again.
  • Yes-bias of fitting: any parabola-fit algorithm will produce a vertex from any data, even when the underlying curve is flat. Use quality metrics (residual variance, R², bin coverage) to reject low-confidence fits.
  • Time-varying CPPopt: CPPopt drifts over hours. The 4-hour window is a compromise; do not over-react to single-window changes.
  • Trust trends over snapshots: a CPPopt that has been stable at 70 for 12 hours is more trustworthy than one that has jumped 15 mmHg in the last 30 minutes.
  • Sedation changes: PRx changes with sedation depth; re-baseline after major changes.
  • Pediatric default trap: holding rigid to adult-style CPP 60–70 over a fitted pediatric CPPopt of 50 is the most common error. Trust the data.
  • ECMO and low-pulsatility flow: switch to COx or Mx as input.

10. Combine with…

  • PRx: the canonical input signal.
  • Mx: the non-invasive (TCD) fallback input.
  • COx: the regional cortical (NIRS) fallback input.
  • CPP: the action target.
  • ICP: the input signal pair.
  • PbtO2: endpoint validation.
  • Microdialysis: metabolic endpoint validation.

11. Evidence summary and recent literature

11.1 Evidence summary

TopicSourceGrade
Aries 2012 original CPPopt cohortB
PRx-mortality association (Steiner 2002)B
Sorrentino 2012 CPPopt outcomeB
Depreitere 2014 ICP-dose + CPPoptB
MAPopt (Donnelly 2017)B
Pediatric CPPopt (Velle 2023)B
Pediatric autoregulation (STARSHIP, Agrawal 2023)B
Lewis 2015 pediatric CPPopt foundationalC
COGiTATE phase II (Tas 2021)A
Pediatric MMM review (Laws 2022)B
BTF 4 pediatricexpert
NCS MMM consensusexpert
Autoregulation methods reviewreview

11.2 Recent literature (2022–2025)

  • Tas 2021 COGiTATE phase II: feasibility and safety of CPPopt-guided care; 46.5% within band vs a 36% target; not powered for outcome. The single most important interventional study to date.
  • Velle 2023 pediatric CPPopt: CPPopt is computable in children with severe TBI (n=57); a lower CPPopt in younger children has been described.
  • Laws 2022 pediatric MMM review: places PRx-based CPPopt among the specialist-centre modalities of pediatric multimodal monitoring.
  • Donnelly 2017 MAPopt: framework extension to MAP optimisation when ICP varies.
  • STARSHIP (Agrawal 2023): the multicentre pediatric autoregulation study (protocol).
  • Toward phase III: a full-scale phase III efficacy trial of CPPopt-guided therapy has been called for but is not yet registered.

12. Self-check

Retrieval check
A 12-year-old severe TBI day 3. ICP 14, MAP 80, CPP 66. CPPopt fit over the past 4 hours shows a clean U-curve with vertex at CPP 72 mmHg. PRx in the current window is +0.20. Most appropriate next step?
After 4 hours of clean recording in a 6-year-old, the (CPP, PRx) plot is flat across CPP 50–90 with PRx near +0.15. Signal quality is good; MAP has varied between 60 and 78 over the window. What does this mean?
A 3-year-old severe TBI day 2. ICP 11, MAP 65, CPP 54. The team starts noradrenaline to "lift CPP above 60 per protocol". PRx was −0.10 at baseline; after CPP rises to 64, PRx rises to +0.30. CPPopt fit gives a vertex at CPP 48. Most appropriate next step?

References

  1. Aries MJ, Czosnyka M, Budohoski KP, et al.. Continuous determination of optimal cerebral perfusion pressure in traumatic brain injury. Critical Care Medicine 2012;40(8):2456-2463. doi:10.1097/CCM.0b013e3182514eb6 link
  2. Beqiri E, Smielewski P, Robba C, et al.. Feasibility of individualised severe traumatic brain injury management using an automated assessment of optimal cerebral perfusion pressure: the COGiTATE phase II study protocol. BMJ Open 2019;9(9):e030727. doi:10.1136/bmjopen-2019-030727 link
  3. Velle F, Lewen A, Howells T, et al.. Cerebral pressure autoregulation and optimal cerebral perfusion pressure during neurocritical care of children with traumatic brain injury. Journal of Neurosurgery: Pediatrics 2023;31(5):503-513. doi:10.3171/2023.1.PEDS22352 link
  4. Lewis PM, Czosnyka M, Carter BG, et al.. Cerebrovascular pressure reactivity in children with traumatic brain injury. Pediatric Critical Care Medicine 2015;16(8):739–749. doi:10.1097/PCC.0000000000000471 link
  5. Tas J, Beqiri E, van Kaam RC, et al.. An Update on the COGiTATE Phase II Study: Feasibility and Safety of Targeting an Optimal Cerebral Perfusion Pressure as a Patient-Tailored Therapy in Severe Traumatic Brain Injury. Acta Neurochirurgica Supplement 2021;131:181-185. doi:10.1007/978-3-030-59436-7_29 link
  6. Brady KM, Mytar JO, Lee JK, et al.. Monitoring cerebral blood flow pressure autoregulation in pediatric patients during cardiac surgery. Stroke 2010;41(9):1957–1962. doi:10.1161/STROKEAHA.109.575167 link
  7. Depreitere B, Guiza F, Van den Berghe G, et al.. Pressure autoregulation monitoring and cerebral perfusion pressure target recommendation in patients with severe traumatic brain injury based on minute-by-minute monitoring data. Journal of Neurosurgery 2014;120(6):1451-1457. doi:10.3171/2014.3.JNS131500 link
  8. Tas J, Beqiri E, van Kaam RC, et al.. Targeting Autoregulation-Guided Cerebral Perfusion Pressure after Traumatic Brain Injury (COGiTATE): A Feasibility Randomized Controlled Clinical Trial. Journal of Neurotrauma 2021;38(20):2790-2800. doi:10.1089/neu.2021.0197 link
  9. Svedung Wettervik T, Velle F, Hånell A, et al.. ICP, PRx, CPP, and ΔCPPopt in pediatric traumatic brain injury: the combined effect of insult intensity and duration on outcome. Child's Nervous System 2023;39(9):2459-2466. doi:10.1007/s00381-023-05982-5 link
  10. Agrawal S, Placek MM, White D, et al.. Studying Trends of Auto-Regulation in Severe Head Injury in Paediatrics (STARSHIP): protocol to study cerebral autoregulation in a prospective multicentre observational research database study. BMJ Open 2023;13(9):e071800. doi:10.1136/bmjopen-2023-071800 link
  11. Donnelly J, Czosnyka M, Adams H, et al.. Individualizing thresholds of cerebral perfusion pressure using estimated limits of autoregulation. Critical Care Medicine 2017;45(9):1464-1471. doi:10.1097/ccm.0000000000002575 link
  12. Kochanek PM, Tasker RC, Carney N, et al.. Guidelines for the Management of Pediatric Severe Traumatic Brain Injury, Third Edition: Update of the Brain Trauma Foundation Guidelines, Executive Summary. Pediatric Critical Care Medicine 2019;20(3S):S1-S82. doi:10.1097/PCC.0000000000001736 link
  13. Smith CA, Rohlwink UK, Mauff K, et al.. Cerebrovascular Pressure Reactivity Has a Strong and Independent Association With Outcome in Children With Severe Traumatic Brain Injury. Critical Care Medicine 2023;51(5):573-583. doi:10.1097/CCM.0000000000005815 link
  14. Laws JC, Jordan LC, Pagano LM, Wellons JC, Wolf MS. Multimodal Neurologic Monitoring in Children With Acute Brain Injury. Pediatric Neurology 2022;129:62-71. doi:10.1016/j.pediatrneurol.2022.01.006 link
  15. Hoh BL, Ko NU, Amin-Hanjani S, et al.. Guideline for the management of patients with aneurysmal subarachnoid hemorrhage: a guideline from the American Heart Association/American Stroke Association. Stroke 2023;54(7):e314-e370. doi:10.1161/str.0000000000000449 link
  16. Rass V, Helbok R. How to diagnose delayed cerebral ischaemia and symptomatic vasospasm and prevent cerebral infarction in patients with subarachnoid haemorrhage. Current Opinion in Critical Care 2021;27(2):103-114. doi:10.1097/mcc.0000000000000798 link
  17. Rivera-Lara L, Zorrilla-Vaca A, Geocadin R, et al.. Cerebral Autoregulation-oriented Therapy at the Bedside. Anesthesiology 2017;126(6):1187-1199. doi:10.1097/ALN.0000000000001625 link
  18. Topjian AA, Scholefield BR, Pinto NP, et al.. Pediatric Post-Cardiac Arrest Care: A Scientific Statement From the American Heart Association. Circulation 2019;140(6):e194-e233. doi:10.1161/CIR.0000000000000697 link
  19. Lorusso R, Taccone FS, Belliato M, et al.. Brain monitoring in adult and pediatric ECMO patients: the importance of early and late assessments. Minerva Anestesiologica 2017;83(10):1061-1074. doi:10.23736/s0375-9393.17.11911-5 link
  20. Pisano DV, Ortoleva JP, Wieruszewski PM. Short-Term Neurologic Complications in Patients Undergoing Extracorporeal Membrane Oxygenation Support: A Review on Pathophysiology, Incidence, Risk Factors, and Outcomes. Pulmonary Therapy 2024. doi:10.1007/s41030-024-00265-z link
  21. Brady KM, Easley RB, Kibler K, et al.. Real-time continuous monitoring of cerebral blood flow autoregulation using near-infrared spectroscopy in patients undergoing cardiopulmonary bypass. Stroke 2010;41(9):1951–1956. doi:10.1161/STROKEAHA.109.575159 link
  22. Le Roux P, Menon DK, Citerio G, et al.. Consensus summary statement of the international multidisciplinary consensus conference on multimodality monitoring in neurocritical care. Intensive Care Medicine 2014;40(9):1189-1209. doi:10.1007/s00134-014-3369-6 link
  23. Steiner LA, Czosnyka M, Piechnik SK, et al.. Continuous monitoring of cerebrovascular pressure reactivity allows determination of optimal cerebral perfusion pressure in patients with traumatic brain injury. Critical Care Medicine 2002;30(4):733–738. doi:10.1097/00003246-200204000-00002 link
  24. Sorrentino E, Diedler J, Kasprowicz M, et al.. Critical thresholds for cerebrovascular reactivity after traumatic brain injury. Neurocritical Care 2012;16(2):258–266. doi:10.1007/s12028-011-9630-8 link