Principles of Osmotherapy

Intracranial pressure is fundamentally a problem of volume interacting with compliance, while cerebral perfusion pressure is only a pressure gradient and should never be mistaken for cerebral blood flow. The natural next step is cerebral edema and osmotherapy, because this is where those concepts become therapeutic.

Central idea: Hyperosmolar therapy is extremely good at transiently manipulating intracranial volume, but remarkably poorly proven as a way to improve long-term neurologic outcome. Those two statements are not contradictory.

Cerebral Edema Is Not Intracranial Hypertension

Cerebral edema means abnormal water accumulation within brain tissue. Intracranial hypertension is the pressure consequence of all intracranial volumes interacting with cranial compliance.

You can therefore have substantial edema with relatively normal ICP while compensatory reserve remains. Conversely, dangerous ICP can arise from hematoma, hydrocephalus, venous congestion, or increased cerebral blood volume without edema being the dominant problem.

Intervention	Primary physiologic target
Osmotherapy	Reduces tissue water and parenchymal volume.
CSF drainage	Reduces the CSF compartment.
PaCO₂ manipulation	Modifies cerebral blood volume.
Hematoma evacuation	Removes pathologic mass.
Decompressive craniectomy	Changes the pressure-volume relationship itself.

Calling all of these “ICP treatment” is clinically correct, but it obscures their very different physiology.

The Major Edema Phenotypes

Cytotoxic edema

Cytotoxic edema begins with cellular energy failure. ATP depletion impairs the sodium-potassium ATPase, sodium accumulates intracellularly, membrane potential fails, chloride and water follow, and astrocytes and neurons swell.

A subtle but important point is that pure early cytotoxic edema is initially a redistribution of water from extracellular to intracellular space rather than necessarily a large increase in total brain water. Net brain water increases as ionic gradients across the neurovascular unit break down and sodium and water enter the tissue from the circulation. This helps explain why malignant infarction evolves over hours rather than appearing instantly at maximal size when the artery occludes.

Vasogenic edema

Blood-brain barrier permeability increases and water plus plasma solutes enter the extracellular compartment. Tumor-associated edema is the archetype, but vasogenic components also occur in trauma, hemorrhage, inflammation, PRES, and later ischemia.

Interstitial edema

Interstitial edema is essentially hydrocephalic edema: an excessive ventricular-to-parenchymal pressure gradient drives CSF across the ependyma.

Osmotic edema

Osmotic edema occurs when plasma tonicity falls relative to brain tonicity, as in sufficiently rapid or severe hyponatremia.

Actual injured brains commonly contain mixtures of these forms. The distinctions nevertheless explain why one generic “anti-edema” therapy cannot work equally well in every condition. Dexamethasone can dramatically improve tumor-associated vasogenic edema, yet it is not a generic treatment for cerebral swelling and is contraindicated as a treatment strategy in severe TBI. CSF diversion can rapidly reverse the intracranial consequences of acute obstructive hydrocephalus without changing the molecular mechanisms of cytotoxic edema at all.

Effective Osmoles, the BBB, and Water Movement

Water moves according to effective osmotic gradients, not simply according to the laboratory number labeled serum osmolality. The ability of a solute to generate an osmotic force across a membrane depends partly on how effectively that membrane excludes the solute.

An intact blood-brain barrier therefore matters enormously. Raise the effective osmolality of plasma with an osmole largely excluded from brain, and water moves from brain toward blood. If the barrier becomes freely permeable to that osmole, the gradient collapses.

In real TBI, ICH, and ischemia, the BBB is spatially heterogeneous rather than simply “intact” or “destroyed.” This is why osmotherapy can still work very well in an injured brain even though the BBB is abnormal.

Why It Matters

The physiologic leverage comes from the intracranial compliance curve. You do not have to remove a dramatic quantity of brain water to produce a dramatic fall in ICP. On the steep portion of the pressure-volume curve, a relatively small reduction in parenchymal volume can move the system back toward a much more compliant region.

This is the essential Deranged Physiology framing of osmotherapy: it temporarily “desiccates” brain tissue enough to reduce its intracranial volume contribution. It has not treated the contusion, infarction, hematoma, mitochondrial failure, inflammation, or disrupted BBB that generated the swelling. It has bought pressure-volume reserve.

An osmotherapy response should therefore be viewed as the result of a physiologic intervention rather than a disease-modifying drug. If a patient is herniating from a malignant hemispheric infarction and hypertonic saline restores the pupil and drops the ICP from 35 to 16, that is an immediate treatment success. It does not mean the malignant infarction has been treated. You have created time in which definitive decisions about decompression, CSF diversion, hematoma evacuation, or control of the underlying process can occur.

Mannitol Versus Hypertonic Saline

Mannitol and hypertonic saline both exploit an osmotic gradient, but their systemic physiology differs substantially.

Feature	Mannitol	Hypertonic saline
Core action	Remains predominantly extracellular, raises plasma tonicity, and draws water from brain tissue toward the vascular compartment.	Creates an extracellular osmotic gradient that draws water from brain while generally expanding the intravascular compartment.
Early vascular effects	Initially expands intravascular volume. Reduced blood viscosity may improve flow and, when autoregulation is functioning, provoke compensatory arteriolar vasoconstriction that reduces cerebral blood volume.	Can support circulating volume and MAP while reducing parenchymal volume.
CPP implications	Osmotic diuresis can contract intravascular volume, lower MAP, and reduce CPP even when ICP improves.	May simultaneously reduce ICP, support MAP, and augment CPP, particularly in a hypotensive or relatively hypovolemic patient.
Major systemic liabilities	Water and electrolyte loss, hypotension, volume depletion, and accumulation when renal elimination fails.	Volume expansion, hypernatremia, substantial chloride exposure, hyperchloremia, and metabolic acidosis.
Potential poor fit	Hypotension, hypovolemia, or advanced renal failure with oliguria.	Severe heart failure, pulmonary edema, marked hypernatremia, or severe hyperchloremia.

Clinical Trap

“Hypertonic saline is good and mannitol is bad.” This is not evidence based. Both agents lower ICP. Systemic physiology often determines which one fits the patient. Giving a drug that drops ICP by 10 mm Hg but drops MAP by 20 mm Hg is not necessarily a cerebral perfusion victory.

Evidence Anchor

2020 Neurocritical Care Society cerebral-edema guideline: Conditionally suggested hypertonic sodium solutions over mannitol for initial management of elevated ICP or edema in TBI, while explicitly rating the evidence as low quality. It also preferred hypertonic saline over mannitol in ICH on very-low-quality evidence.
Brain Trauma Foundation severe-TBI guideline, fourth edition: States that hyperosmolar therapy lowers ICP, but that evidence about clinical outcomes is insufficient to recommend a specific agent. This remains the relevant general adult severe-TBI guideline; newer penetrating-TBI material addresses a separate population.
NCS ENLS 6.0: Describes mannitol and hypertonic saline as having equivalent efficacy for emergency ICP reduction.
Bernhardt and colleagues: A systematic review appearing in the 2024 volume of Neurocritical Care included 10 randomized trials and 760 patients with TBI. Compared with other ICP-lowering agents, predominantly mannitol, hypertonic saline did not demonstrate improved favorable Glasgow Outcome Scale outcome: risk ratio 0.82, 95% confidence interval 0.48–1.40. Mortality was also neutral: risk ratio 0.96, 95% confidence interval 0.60–1.55. The point estimate for uncontrolled ICP favored hypertonic saline but did not reach conventional statistical significance, while hypertonic saline increased hypernatremia. Evidence certainty was low to very low.
COBI trial: Continuous 20% hypertonic saline in moderate-to-severe TBI did not improve six-month neurologic outcome.
UK Sugar or Salt trial: Designed to compare equiosmolar boluses of mannitol and hypertonic saline in severe TBI requiring treatment for intracranial hypertension. Definitive outcome results were not identified as of August 2026, leaving the head-to-head outcome question genuinely unsettled.

These statements are easier to reconcile than they first appear. One asks which agent a panel would preferentially use given physiology and low-quality comparative data. Another asks whether rigorous outcome evidence proves superiority in severe TBI. ENLS addresses emergency ICP reduction. Hypertonic saline can therefore be a perfectly reasonable first choice without claiming that it has been demonstrated to produce superior six-month neurologic outcomes.

The COBI trial tested something conceptually different from rescue bolus therapy. Its neutral result should not be interpreted as evidence that hypertonic saline fails to abort an ICP crisis. It challenges the hypothesis that continuously maintaining a hyperosmolar state is itself neuroprotective. These are different hypotheses.

Clinical Trap

“Keep the sodium at 150–155 because induced hypernatremia improves neurologic outcome.”

There is no high-quality evidence that simply maintaining this sodium range improves neurologic outcome. Sodium is a means of producing tonicity, not the therapeutic endpoint.

In SAH, the NCS cerebral-edema guideline favors symptom-based hypertonic-saline boluses rather than sodium-target-based dosing. In ICH, either a symptom-based strategy or targeted sodium was considered reasonable, but this was a conditional recommendation resting on very-low-quality evidence. Disease state matters, and the existence of a serum sodium target in a local order set should never be mistaken for proof that induced hypernatremia is intrinsically protective.

ENLS 6.0 Emergency Dosing

Agent	ENLS 6.0 dose	Administration
Mannitol	0.5–1 g/kg IV	Approximately 5–15 minutes
3% saline	5 mL/kg	Approximately 5–20 minutes
5% saline	3 mL/kg	Per protocol
7.5% saline	2 mL/kg	Per protocol
23.4% saline	30 mL	Approximately 10–20 minutes

ENLS 6.0 specifies central access for emergent concentrations greater than 3%. These are guideline protocol doses; institutional protocols commonly differ, particularly in how 3% and 23.4% saline are delivered.

Peripheral 23.4% Saline: Evolving Safety Evidence

A 2025 Neurocritical Care retrospective cohort examined 863 peripheral administrations of 30 mL of 23.4% saline. Extravasation was identified after 56 administrations, or 6.4%, but there were no cases of tissue necrosis or surgical intervention and only one case requiring hyaluronidase. Diabetes was associated with greater extravasation risk.

Silva et al., 2025 provides useful evidence when confronting a true brain code without central access, but retrospective safety data do not replace the current ENLS recommendation or local medication policy. The appropriate interpretation is that the traditional belief that peripheral 23.4% saline invariably causes catastrophic tissue injury is too strong; evidence supporting carefully monitored emergency peripheral administration is becoming more reassuring.

Monitoring Mannitol: Osmolality Is Not Tonicity

The usual calculated serum osmolality is approximately:

Calculated serum osmolality ≈ (2 × sodium) + (glucose ÷ 18) + (BUN ÷ 2.8)

BUN contributes to measured osmolality but is a relatively ineffective osmole because urea equilibrates across cell membranes. Mannitol, conversely, is an effective extracellular osmole but is not included in that routine calculated formula.

Osmolar gap = measured osmolality − calculated osmolality

After mannitol administration, the osmolar gap becomes useful as a rough marker of unmeasured osmoles and, in the appropriate context, mannitol accumulation.

The old reflexive rule of “never give more mannitol when osmolality exceeds 320” is too crude. Modern guidance has moved away from treating 320 mOsm/kg as a magic toxicity boundary. ENLS 6.0 describes repeat mannitol dosing according to the osmolar gap and notes that clinicians may use gaps roughly between 20 and 55 mOsm/kg, while also stating that pushing toward the upper end is unlikely to provide additional therapeutic benefit. Renal function, hemodynamics, urine output, cumulative exposure, and whether the previous dose actually changed ICP are more informative than worshiping a single osmolality cutoff.

Monitoring Hypertonic Saline

ENLS 6.0 recommends checking sodium every four to six hours during therapy and keeping serum sodium below 160 mEq/L. This is a safety ceiling, not a goal of 159. There is no prize for maximizing serum sodium.

As sodium and chloride climb while the ICP response progressively diminishes, the risk-benefit ratio is worsening in front of you. The declining response may reflect loss of osmotic gradient, severe structural disease, BBB disruption, or simply that tissue water is no longer the dominant intracranial volume problem.

Renal Evidence

A July 2026 MIMIC-IV retrospective analysis examined 2,756 adult neurocritical-care patients who were alive, remained in the ICU, and were free of AKI at 48 hours. Early exposure to mannitol or hypertonic saline was associated with subsequent AKI. In a doubly robust overlap-weighted analysis, the relative risk was 1.30, with a 95% confidence interval of 1.08–1.57.

What Should NOT Be Concluded

This is an association, not proof that osmotherapy caused the AKI. Sicker patients with more severe cerebral edema are more likely to receive osmotherapy and are also more likely to develop systemic organ dysfunction. The investigators explicitly acknowledge residual confounding by indication.

Adjustment for early sodium and chloride burden did not materially explain the association. The practice implication is renal vigilance and avoidance of unnecessary exposure—not withholding lifesaving osmotherapy from a herniating patient. Du et al., 2026.

Rebound and Diminishing Returns

Mannitol can enter injured regions when BBB permeability is sufficiently abnormal. If mannitol accumulates within brain tissue while its plasma concentration subsequently falls, the osmotic gradient can diminish or theoretically reverse.

With prolonged hyperosmolar therapy, the brain also adapts by accumulating intracellular osmoles. This reduces the effectiveness of a persistent extracranial osmotic gradient and creates a physiologic rationale for rebound swelling when chronic hyperosmolarity is rapidly withdrawn. Clinical evidence for the magnitude of these effects is imperfect, but it is sufficient to reject the claim that repeated osmotherapy can be escalated indefinitely without diminishing returns.

Choosing the Agent at the Bedside

Clinical state	Physiologic reasoning
ICP 32, MAP 68, sodium 140, relatively hypovolemic	Hypertonic saline is physiologically attractive. It can reduce tissue volume while supporting intravascular volume, MAP, and CPP.
ICP 32, sodium 158, chloride 124, pulmonary edema, preserved renal function, reasonable blood pressure	Another large sodium-chloride load is less attractive. Mannitol may fit the systemic physiology better.
Advanced renal failure with oliguria	Mannitol becomes dangerous because it can accumulate. Hypertonic saline may again be preferable despite the need to manage sodium and volume carefully.

There is no contradiction here. The “best” osmotic agent is conditional on what threatens the brain and what the rest of the body can tolerate.

After Every Dose, Demand a Response

Did ICP fall?
Did the pupil improve?
Did CPP improve or deteriorate?
What happened to MAP?
Did urine output surge after mannitol?
What happened to sodium, chloride, osmolar gap, acid-base status, and renal function?

If three escalating osmotic doses are required to hold the ICP down for progressively shorter intervals, the correct conclusion is usually not that you have discovered a patient who needs a fourth, fifth, and sixth empiric bolus forever. You are being shown refractory intracranial pathophysiology.

Reassess:
→ Anatomy and interval imaging
→ CSF compartment
→ Venous drainage
→ Seizures
→ PaCO₂
→ Temperature
→ Sedation
→ Autoregulation
→ Surgical options
→ Goals of care



  What Should NOT Be Concluded
  
    Cerebral edema is not synonymous with intracranial hypertension.
    Serum sodium is not synonymous with brain water.
    CPP is not synonymous with cerebral blood flow.
    Successful ICP reduction is not synonymous with improved neurologic outcome.
  



  Osmotherapy remains indispensable precisely because a relatively small, rapid reduction in brain volume can move a critically noncompliant intracranial system away from herniation. But it is fundamentally a temporizing manipulation of water and volume.
  The attending-level question is never merely, “Should I give mannitol or 3%?” It is: which intracranial compartment is producing the pressure, what is the underlying edema physiology, how intact is the osmotic barrier likely to be, what systemic consequence will each agent produce, what physiologic response am I trying to achieve, and what definitive intervention am I buying time to reach?