Intracranial Pressure Management in Traumatic Brain Injury
Severe traumatic brain injury (TBI) transforms the skull from a protective vault into a rigid trap: once the brain, blood, and cerebrospinal fluid (CSF) volumes plus any new mass lesion exceed the limited compensatory space, small further increases in volume produce disproportionate, dangerous rises in intracranial pressure (ICP). The tiered management strategy below — codified by the Brain Trauma Foundation (BTF) — balances aggressive ICP control against the equally real risk of impairing cerebral perfusion and oxygenation.
- > 22 mmHg: ICP treatment threshold (BTF 4th Ed. (2016) — sustained elevation)
- 60–70 mmHg: CPP target range (avoid <50 and routine >70)
- 5–15 mmHg: Normal ICP (supine adult) (CSF opening pressure equivalent)
- GCS 3–8: Monitoring indication (+ abnormal CT, or normal CT + risk factors)
The Monro–Kellie doctrine and the volume–pressure curve
The Monro–Kellie doctrine, first articulated by Alexander Monro (1783) and George Kellie (1824), states that the skull after cranial suture closure is a rigid, nearly incompressible container of fixed total volume. That volume is normally partitioned among three compartments: brain parenchyma (~80%, ~1,400 mL), CSF (~10%, ~150 mL), and intracranial blood (~10%, ~150 mL, mostly venous). Because the container cannot expand, any increase in the volume of one compartment — or the introduction of a new compartment such as an epidural, subdural, or intraparenchymal hematoma — must be offset by an equal decrease in the others, or intracranial pressure will rise.
Initially the brain compensates well: CSF is displaced from the cranial vault into the spinal thecal sac, and venous blood is displaced out through the jugular system. Over this early "compensated" phase, the pressure–volume curve is nearly flat — substantial volume can be added with only a small ICP rise. This flat region represents intact compensatory reserve. But CSF and venous blood reserves are finite. Once they are exhausted, the curve turns sharply upward: the relationship becomes exponential, and each additional milliliter of added volume produces a much larger rise in ICP than the one before. This nonlinear region is where small, previously trivial changes — a seizure, a coughing fit, a few degrees of head rotation obstructing jugular outflow — can precipitate a catastrophic pressure spike and herniation.
The steepness of this curve at any point is the intracranial elastance (ΔP/ΔV); its inverse, compliance, is the "compensatory reserve." Clinically, this reserve can be estimated at the bedside from the ICP waveform morphology (see below) or, invasively, from the pressure response to a small saline bolus injected into an intraventricular catheter (the pressure-volume index, PVI).
Indications for invasive ICP monitoring
The Brain Trauma Foundation recommends invasive ICP monitoring in patients with severe TBI (Glasgow Coma Scale 3–8 after resuscitation) and an abnormal admission CT scan — defined as showing hematomas, contusions, swelling, herniation, or compressed basal cisterns. ICP monitoring is also recommended in severe TBI patients with a normal CT scan if two or more of the following adverse features are present: age over 40 years, unilateral or bilateral motor posturing, or systolic blood pressure below 90 mmHg. This second criterion recognizes that a normal initial CT does not exclude subsequent diffuse swelling or delayed hematoma expansion in a high-risk patient who cannot be serially examined because of sedation and mechanical ventilation.
Invasive monitors fall into two main categories. An external ventricular drain (EVD), with its tip in the lateral ventricle connected to an external transducer, remains the gold-standard monitor because it provides highly accurate global ICP measurement and doubles as a therapeutic CSF drainage conduit. Intraparenchymal fiberoptic or strain-gauge microtransducers are easier to place (no ventricular cannulation required, useful with slit-like or effaced ventricles) but cannot be recalibrated once inserted and offer no therapeutic drainage capability. Choice of monitor depends on ventricular size, coagulation status, and the anticipated need for CSF diversion.
CPP-targeted versus ICP-targeted management philosophy
Two historically competing schools of thought have shaped modern neurocritical care protocols.
The Rosner (Lund-opposite / "CPP-targeted") approach, championed by Rosner and Daughton in the 1990s, treats cerebral perfusion pressure as the primary target, arguing that maintaining CPP above roughly 70 mmHg — using vasopressors and volume expansion to raise MAP — preserves cerebral blood flow and prevents secondary ischemic injury even in the face of moderately elevated ICP, since a well-perfused, appropriately autoregulating brain will itself help normalize ICP.
The Lund concept, developed in Sweden (Grände, Asgeirsson, Nordström), instead treats ICP control as primary and is more cautious about permissive high CPP. It aims to reduce cerebral blood volume and capillary hydrostatic pressure through modest arterial pressure control (avoiding hypertension), reduction of stress-hormone-driven vasoconstriction, colloid osmotic pressure support (albumin), and preservation of the blood-brain barrier — accepting a lower CPP (sometimes 50–60 mmHg) so long as ICP is well controlled, on the theory that aggressive CPP-elevation in an injured, poorly autoregulating brain simply drives more transcapillary fluid into already-swollen tissue and worsens edema.
Contemporary BTF-endorsed practice occupies a middle ground: target CPP 60–70 mmHg, individualized to the patient's autoregulatory status where possible, while avoiding both the harms of hypoperfusion (CPP <60, associated with ischemia) and of an overly aggressive CPP-elevation strategy (CPP >70, associated with a 5-fold increase in adult respiratory distress syndrome from the fluids and vasopressors required to sustain it, per Robertson et al.'s randomized CPP-management trial).
Hyperosmolar therapy — mannitol and hypertonic saline
Mannitol (20% solution) is dosed as an intermittent IV bolus of 0.25–1 g/kg, typically given over 15–20 minutes, and acts through two mechanisms with different time courses: an immediate rheologic effect (reducing blood viscosity and reflexively vasoconstricting resistance vessels within minutes, lowering cerebral blood volume) followed over 15–30 minutes by its classic osmotic effect — drawing free water across an intact blood-brain barrier from brain interstitium into the intravascular space, from which it is excreted by the kidney. Repeated dosing requires monitoring of serum osmolality (keep <320 mOsm/kg) and the osmolar gap, because mannitol is a potent osmotic diuretic that can produce hypovolemia, hypotension (which directly threatens CPP), and acute kidney injury if allowed to accumulate. Mannitol should be avoided or used cautiously in hypovolemic or hypotensive patients since its diuretic effect can precipitously drop MAP.
Hypertonic saline (commonly 3%, sometimes given as 23.4% for rapid effect) raises serum sodium and serum osmolality to create the same osmotic gradient without the diuresis, making it generally preferred in hypotensive or hypovolemic patients, and it may cause less rebound intracranial hypertension than mannitol on discontinuation. It can be given as intermittent boluses (e.g., 3% at 150–250 mL) or, in some protocols, as a continuous infusion titrated to a target serum sodium (typically 145–155 mEq/L). Risks include hypernatremia, and — chiefly with rapid correction of chronic hyponatremia rather than with therapeutic hypertonic saline itself — osmotic demyelination (central pontine myelinolysis) if serum sodium is allowed to rise too quickly; volume overload and heart failure are also concerns in patients with limited cardiac reserve.
Both agents work primarily by expanding the osmotic gradient across an intact blood-brain barrier — neither is effective, and both can theoretically worsen edema, in regions where the barrier has been destroyed by direct contusion.
Hyperventilation — a powerful but dangerous temporizing tool
Lowering PaCO₂ causes rapid cerebral arteriolar vasoconstriction: cerebral blood flow changes by roughly 3–4% for every 1 mmHg change in PaCO₂ around the normal range. This is the fastest ICP-lowering maneuver available — effects begin within a minute — which makes controlled hyperventilation (PaCO₂ to roughly 30–35 mmHg) a useful Tier-1 bridge and an essential immediate maneuver for acute clinical herniation (blown pupil, extensor posturing) while more definitive therapy (osmotherapy, surgery) is arranged.
The danger is that the same vasoconstriction that lowers cerebral blood volume also lowers cerebral blood flow, and can push already-vulnerable peri-contusional tissue into frank ischemia. Sustained or aggressive hyperventilation (PaCO₂ below ~25–28 mmHg) is specifically discouraged by BTF guidelines except as a brief bridge, because the vasoconstrictive effect attenuates within 24 hours (cerebral pH normalizes as bicarbonate re-equilibrates) while the ischemic risk does not — leaving a hyperventilated patient with reduced perfusion but no lasting ICP benefit. When hyperventilation beyond mild targets is used, BTF and most protocols recommend concurrent monitoring of cerebral oxygenation (jugular venous oxygen saturation, SjO₂, or brain tissue oxygen tension, PbtO₂) to detect ischemia before it becomes injurious.
Prophylactic hyperventilation to PaCO₂ ≤25 mmHg is not recommended by BTF guidelines. Hyperventilation should be reserved as a temporizing measure for acute neurological deterioration, ideally guided by brain oxygenation monitoring.
Decompressive craniectomy — trial evidence and the ICP-versus-outcome tradeoff
Decompressive craniectomy removes a large section of the skull (and opens the underlying dura) so that swollen brain can expand outward rather than compress itself and adjacent structures — mechanically flattening the pressure–volume curve and reliably lowering refractory ICP. Two major randomized trials shaped how this Tier-3 option is now used.
DECRA (Cooper et al., NEJM 2011) randomized patients with diffuse injury and early, moderately elevated ICP (>20 mmHg for 15 minutes within a 1-hour period, relatively early after injury) to bifrontotemporoparietal craniectomy versus continued medical management. Craniectomy reduced ICP and length of ICU stay but was associated with worse functional outcomes (higher rates of severe disability) at 6 months, raising concern that operating "too early" on relatively modest, still-treatable ICP elevation removed a protective mechanical constraint without net clinical benefit.
RESCUEicp (Hutchinson et al., NEJM 2016) took a different approach, enrolling patients with refractory ICP (>25 mmHg for 1–12 hours) after maximal medical therapy had failed — i.e., a genuinely last-tier population. Craniectomy in this trial reduced mortality substantially compared with continued medical management, but survivors had higher rates of vegetative state and severe disability; the proportion achieving a favorable outcome (upper moderate disability or good recovery) was similar between groups. Taken together, the trials suggest decompressive craniectomy reliably trades death for survival with disability in truly refractory ICP, while operating pre-emptively on earlier, still-medically-controllable ICP elevations may not improve — and can worsen — functional outcomes. This evidence underpins its placement as a Tier-3, refractory-ICP-only intervention rather than an early or prophylactic one.
Quick-reference: therapy tier, mechanism, and principal risk