Neuro: ICP and Stroke

In depth

The one-page summary above is what the PDF prints. These sections go further.

Monro-Kellie, compliance, CPP and autoregulation

The skull holds brain (about 80% of its volume), blood (10%) and CSF (10%). By the Monro-Kellie doctrine, added volume (hematoma, edema, hydrocephalus) must push an equal volume out: CSF shifts first into the spinal subarachnoid space, then venous blood leaves through the jugular veins. CSF is made at about 20 mL/h (500 mL/day) against a total of about 150 mL.

Compliance (change in volume ÷ change in pressure) stays high while these buffers last, then collapses: the pressure–volume curve is flat, then steep. On the steep part a few milliliters raise ICP 10–20 mm Hg, so suctioning, coughing or fever can tip a patient into crisis.

CPP = MAP − ICP. Cerebral blood flow is normally about 50 mL/100 g/min; neurons stop signaling below about 20 and infarct below about 10. Intact autoregulation holds flow steady across a CPP of roughly 50–150 mm Hg, shifted right by chronic hypertension. After TBI, SAH or ICH it often fails: flow becomes pressure-passive, low CPP causes ischemia and high CPP adds hyperemia and edema. A positive pressure reactivity index (PRx, the moving MAP–ICP correlation) signals failed autoregulation and can define an individual optimal CPP. Zero the arterial line at one documented level (tragus in a 2015 UK consensus, phlebostatic axis in many US units); at 30° they differ by 10 mm Hg or more.

ICP monitors and the ICP waveform

EVD placement shows new hemorrhage on CT in about 6% but clinically significant bleeding in under 1% (Binz 2009 meta-analysis). SIBICC (2020) treats PbtO₂ below 20 by raising CPP, FiO₂ or hemoglobin; outcome benefit is unproven.

Each beat gives three peaks: P1 (percussion), the arterial pulse from the choroid plexus; P2 (tidal), reflecting brain compliance; P3 (dicrotic), aortic valve closure. Normally P1 > P2 > P3. When P2 equals or exceeds P1 the trace rounds off: compliance is poor even if the mean is under 20. A flat trace with no peaks points to a kinked, clotted or air-filled system, not a real fall in ICP.

EVD management: leveling, drainage, clamping, infection

Zero the transducer and set the drip chamber at the same point: the tragus as the landmark for the foramen of Monro (BTF describes the midbrain; follow one unit standard). Drain height is ordered in cm H₂O above it, commonly 10–20. CSF drains only when ICP exceeds that height; 1 mm Hg = 1.36 cm H₂O, so 20 cm H₂O lets ICP reach about 15 mm Hg.

Orders specify open drainage at a set height, with ICP read by clamping briefly, or a clamped drain opened above an ICP threshold. An open drain's number mostly reflects chamber height. Many protocols want a report above about 20–25 mL/h or when drainage stops. Overdrainage can tear bridging veins (subdural hematoma), rebleed an unsecured aneurysm, or cause upward herniation with a posterior fossa mass.

Ventriculitis complicates roughly 5–10% of drains; risk rises with IVH or SAH, CSF leak, frequent access and days in place. The NCS 2016 EVD consensus recommends antimicrobial-impregnated catheters, one periprocedural antibiotic dose rather than continuous prophylaxis, CSF sampling only when clinically indicated, and no routine catheter exchange. Blood and fever are common anyway, so watch trends: falling CSF glucose, rising protein, a rising cell index, positive Gram stain or culture. Empiric therapy (IDSA 2017): vancomycin plus cefepime, ceftazidime or meropenem.

Weaning: raise the drain in steps or clamp for about 24 h while ICP, exam and CT are followed; failure means shunt-dependent hydrocephalus. Rapid clamping matched gradual weaning for shunt rates (Klopfenstein 2004). After removal, watch for CSF leak.

Neuro assessment and herniation syndromes

Use a graded stimulus: voice, touch, then pressure. Central pressure (trapezius, supraorbital notch) tests localization; nail-bed pressure tests flexion. A patient localizes when the hand crosses midline or rises above the clavicle toward the stimulus. Score the best motor response but record each side, because asymmetry localizes. The FOUR score (0–16) has no verbal part and suits intubated patients. Record pupil size in mm and reactivity; an oval or newly sluggish pupil can be the first sign of third-nerve compression. A pupillometer NPi below 3 is abnormal.

In uncal herniation the opposite cerebral peduncle can hit the tentorium (Kernohan notch), giving weakness on the same side as the blown pupil. Central herniation progresses through Cheyne-Stokes, central neurogenic hyperventilation, apneustic and ataxic breathing, and can cause diabetes insipidus. After craniectomy, a sunken flap with decline when upright or after lumbar drainage is paradoxical herniation: lay flat, give fluids, stop lumbar drainage.

TBI and tiered ICP therapy

A single SBP below 90 or PaO₂ below 60 roughly doubled mortality (Traumatic Coma Data Bank). Tranexamic acid within 3 h helped mild to moderate TBI only (CRASH-3, 2019).

SIBICC (2019) tiers ICP care; treatments within a tier are not ranked.

1. Tier 0: ICU, intubation, HOB 30–45°, analgesia and sedation, prevent fever, CPP ≥60, Hb above 7, SpO₂ ≥94%, avoid low sodium.
2. Tier 1: CPP 60–70, deeper sedation, PaCO₂ 35–38, mannitol or hypertonic saline boluses, CSF drainage, consider EEG.
3. Tier 2: PaCO₂ 32–35, neuromuscular blockade if a test dose works, MAP challenge to test autoregulation, then raise CPP if intact.
4. Tier 3: barbiturate coma if a test dose works, secondary decompressive craniectomy, mild hypothermia (35–36 °C).

SIBICC avoids continuous mannitol, lumbar drainage, PaCO₂ below 30 and steroids. Pentobarbital (classic regimen 10 mg/kg over 30 min, 5 mg/kg/h for 3 h, then 1 mg/kg/h to burst suppression) causes hypotension, ileus, immunosuppression and hypothermia, and can leave pupils small and unreactive.

Mannitol first expands plasma and lowers viscosity, triggering reflex arteriolar constriction within minutes, then draws water from brain over 15–30 min; the effect lasts about 1.5–6 h. Replace its diuresis with isotonic fluid. Redose when the osmolar gap is below about 20; holding at osmolality 320 does not predict kidney injury (NCS 2020) but is still the exam answer. Hypertonic saline expands volume and suits hypotensive patients. Protocols cap sodium at 155–160, use acetate blends against hyperchloremic acidosis, and wean slowly to avoid rebound edema.

Aneurysmal subarachnoid hemorrhage

Rebleeding peaks in the first 24 h. Until the aneurysm is secured, treat pain, reverse anticoagulants, place an EVD for hydrocephalus and avoid hypertension; most centers keep SBP below 140–160, and the 2023 AHA/ASA guideline sets no single number. Secure the aneurysm early, ideally within 24 h (2023 AHA/ASA), by coiling or clipping; ISAT (2002) favored coiling when both were possible.

Angiographic vasospasm appears in about 70% in classic series, mostly days 3–14, with risk peaking at days 6–10 (AHA/ASA 2023); about 30% develop delayed cerebral ischemia (DCI): a new focal deficit or GCS drop of 2 or more for at least 1 h with no other cause. Hourly exams are the main monitor; transcranial Doppler adds MCA mean velocity (above 120 cm/s suggests spasm, above 200 severe) and a Lindegaard ratio above 3. Keep euvolemia; triple-H therapy is abandoned. For DCI, induce hypertension stepwise, then intra-arterial vasodilators or angioplasty. Magnesium, statins and clazosentan did not improve outcome.

ICH and ischemic stroke after reperfusion

The ICH score (GCS 3–4 = 2 or 5–12 = 1; volume ≥30 mL; IVH; infratentorial; age ≥80) gave 30-day mortality from 0% at 0 to 72% at 3 (Hemphill 2001). Early limits on care make such scores self-fulfilling (2022 AHA/ASA). Up to a third expand, mostly within 6 h. Priorities: smooth BP control (variability predicts worse outcome), glucose, normothermia, a swallow screen, compression devices from admission (CLOTS 3) and low-dose heparin after 24–48 h. Cerebellar ICH with decline, brainstem compression or hydrocephalus needs evacuation. Minimally invasive evacuation of lobar ICH improved function in ENRICH (2024).

After a lytic, symptomatic hemorrhage occurs in about 6% (6.4% with alteplase in NINDS) and orolingual angioedema in about 1–5%, more with ACE inhibitors. After thrombectomy, check the access site and distal pulses with each neuro check; back or flank pain with hypotension suggests retroperitoneal bleeding. Keep BP below 180/105 for 24 h; the 2026 AHA/ASA guideline rates an SBP target below 140 after successful thrombectomy as harmful.

Large MCA infarcts swell most on days 2–5. Hemicraniectomy within 48 h in patients 60 or younger cut mortality from about 71% to 22% (pooled DECIMAL, DESTINY, HAMLET); over 60 (DESTINY II) it lowered mortality but most survivors were severely disabled. Keep glucose 140–180.

Seizures, cEEG, prognosis after arrest and brain death

In Claassen 2004, 19% of ICU patients on continuous EEG had seizures, 92% purely electrographic; comatose patients often needed 48 h of recording. ACNS 2021 defines electrographic status as 10 continuous minutes, or 20% of any hour, of seizure activity. Paralyzed patients need cEEG to show seizures at all.

In ESETT (2019) levetiracetam 60 mg/kg (max 4,500 mg), fosphenytoin 20 mg PE/kg (max 1,500) and valproate 40 mg/kg (max 3,000) each stopped status in about half. Give fosphenytoin no faster than 150 mg PE/min on a cardiac monitor. Refractory status needs midazolam, propofol, ketamine or pentobarbital infusions titrated on cEEG for 24–48 h, then a slow wean.

After cardiac arrest, prognosticate no earlier than 72 h after ROSC, and only after sedatives, blockers, hypothermia and metabolic derangement are excluded (AHA 2020, ERC-ESICM 2021). ERC-ESICM 2021 considers poor outcome likely when a patient with a motor score of 3 or less at 72 h has two or more of: absent pupillary and corneal reflexes, bilaterally absent N20 SSEP, a highly malignant EEG after 24 h, NSE above 60 µg/L at 48–72 h, status myoclonus within 72 h, or diffuse anoxic injury on CT or MRI.

The 2023 AAN/AAP/CNS/SCCM guideline sets one brain death standard for adults and children. Prerequisites: a known irreversible cause, 24 h or more after a hypoxic-ischemic injury, core temperature ≥36 °C, SBP ≥100 and MAP ≥75 (MAP alone on VA-ECMO), drugs cleared, four twitches on train-of-four, and metabolic disorders corrected. The exam shows coma, non-reactive pupils, and absent corneal, oculocephalic, oculovestibular (50 mL ice water per ear), gag and cough reflexes. One exam suffices in adults; children need two.

• Apnea test: preoxygenate with FiO₂ 1.0 for 10 min, start at PaCO₂ 35–45, give O₂ by tracheal catheter or CPAP
• Positive with no breathing effort at PaCO₂ ≥60 and ≥20 above baseline, pH <7.30
• Abort and draw the ABG for hypotension, a progressive fall in SpO₂ below 85% or arrhythmia
• Ancillary tests only if the exam or apnea test cannot be completed: catheter angiography, radionuclide perfusion or TCD; EEG and CTA do not count

Evidence and controversies

BEST-TRIP is often misread: its control arm used a strict exam-and-CT protocol, so it showed that monitoring without a treatment plan adds nothing. DECRA operated after 15 minutes of ICP above 20; RESCUEicp waited longer, and at 12 months showed more upper severe disability or better (45% vs 32%).

Unsettled: the ICP threshold (BTF's 22 came from one cohort), individualized CPP from PRx, whether PbtO₂-guided care improves outcome, mannitol versus hypertonic saline, the EVD reference point, and whether to cluster or space nursing care in low compliance.

Exam pearls

• CPP = MAP − ICP; TBI target 60–70; treat ICP above 22 (BTF) or 20
• P2 above P1 means poor compliance; Lundberg A waves are always pathological
• Level the EVD at the tragus and re-level after every position change; never leave it clamped
• Hyperventilation is a brief bridge only; avoid it in the first 24 h of TBI
• Hold mannitol above osmolality 320 (exam rule); 23.4% saline needs a central line
• Nimodipine 60 mg every 4 h for 21 days, enteral only; vasospasm peaks days 6–10
• SAH hyponatremia, from salt wasting or SIADH (studies find SIADH more often): give saline, never fluid-restrict
• Apnea test is positive at PaCO₂ ≥60 and ≥20 above baseline