Brain Edema, Osmotherapy, and Osmotic Gradients

Brain edema develops when the distribution or amount of tissue water changes through disturbed cellular transport, vascular permeability, or fluid drainage. Its consequences depend on where that water accumulates and how much intracranial reserve remains. Osmotherapy changes water movement by altering effective osmotic gradients, while also changing the systemic circulation and renal solute handling. Understanding these linked mechanisms explains both an immediate pressure response and its limitations. The central distinction throughout this chapter is between changing brain water or intracranial pressure and demonstrating improved neurological recovery.

62.1 Brain water, brain volume, and intracranial pressure describe different properties

Edema is excess tissue water, but cellular swelling can initially redistribute water already inside the brain. Additional brain volume develops when water enters from blood or cerebrospinal fluid faster than it leaves. A tumor, hematoma, or increased vascular volume can also enlarge an intracranial compartment without being edema. These processes often coexist, so a swollen brain should not automatically be assigned one water-transport mechanism.

Intracranial pressure, or ICP, is the mechanical consequence of volume interacting with available accommodation. The same added volume can produce a small pressure rise when compensatory reserve is available and a much larger rise when that reserve is exhausted. Conversely, displacement between compartments or a focal lesion can threaten tissue before one monitored ICP value captures the entire problem. Chapter 58 develops these pressure–volume and spatial relationships.

Even the expression of tissue water requires attention to its denominator. Experimental studies often report water as a percentage of wet tissue mass. Consider a constructed specimen containing 80 g of water and 20 g of dry material. Its wet mass is 100 g and its water fraction is 80%. If dry material stays fixed while the water fraction becomes 81%, dry material now represents 19% of wet mass. The new wet mass is therefore 20/0.19 = 105.3 g, containing 85.3 g of water.

A one-percentage-point increase in water fraction has added approximately 5.3 g of water in this model. It is not simply a 1% increase in specimen mass. Translating mass into volume would additionally require density assumptions and separation of vascular, extracellular, and cellular contributions. The example explains why apparently small changes in experimental brain-water percentages can represent meaningful swelling without supplying a clinical volume estimate.

Clinical imaging describes the distribution, mass effect, and evolution of injury. A pressure monitor describes a local mechanical signal. Neither independently measures every compartment's water content or identifies the permeability of each vascular region. Interpretation begins by specifying which property has actually been observed.

62.2 Cytotoxic swelling and ionic edema connect cellular failure to net water entry

Cells maintain volume by regulating their osmotically active solutes. The sodium–potassium ATPase and other transport systems sustain ion gradients across the membrane. During energy failure, these gradients deteriorate while channels, cotransporters, and excitatory signaling permit net solute entry. Water follows the increase in intracellular effective osmoles, producing swelling of astrocytes, neurons, and other cells.

This cellular process is commonly called cytotoxic edema. In an initially isolated tissue compartment, movement of water from extracellular space into cells reduces extracellular volume without necessarily increasing total tissue water. The cells can be swollen while the brain has not yet gained the same amount of water from outside. In living injury, however, redistribution rapidly interacts with continued vascular supply and evolving endothelial transport.

Ionic edema identifies net accumulation of ions and water from the circulation without requiring unrestricted passage of plasma proteins. Depletion of extracellular sodium and water by cellular uptake helps establish gradients that favor their replenishment from blood. Endothelial transport and permeability determine how that replenishment occurs. Thus a barrier that still excludes albumin can nevertheless permit abnormal salt and water accumulation.

O'Donnell and colleagues investigated endothelial sodium–potassium–chloride cotransport in a rat middle cerebral artery occlusion model. Their anatomical and intervention findings supported participation of luminal endothelial ion transport in early ischemic edema. This is mechanistic evidence from an experimental preparation, not evidence that transporter inhibition has an established clinical benefit in human brain injury. 1

The distinction between cellular redistribution and net vascular influx helps explain why the cellular and blood–brain barriers must be considered separately. Perfused tissue around an ischemic region can continue receiving salt and water while cellular handling is abnormal. Reperfusion can change that supply again. Neither “intact barrier” nor “cytotoxic edema” means that all relevant transport processes remain normal.

These categories describe overlapping mechanisms rather than a compulsory sequence with a fixed timetable. Ischemia, trauma, seizures, inflammation, and metabolic injury can affect different cells and regions simultaneously. The resulting swelling depends on both cellular solute accumulation and the pathways through which additional water becomes available.

62.3 Vasogenic and interstitial edema use different entry pathways

Vasogenic edema develops when vascular barrier dysfunction permits increased passage of fluid and normally restricted constituents into brain extracellular space. Altered endothelial junctions, transcellular transport, inflammatory signaling, and structural injury can contribute. Protein leakage is an important distinction from purely ionic edema, but barrier dysfunction is graded and heterogeneous rather than one all-or-none event.

Once water enters extracellular space, tissue architecture and pressure gradients influence where it spreads. White matter pathways and the size of extracellular spaces can favor movement in some directions. Cellular swelling can narrow those spaces, changing the resistance to extracellular fluid movement. Consequently, the distribution of edema does not simply reproduce the distribution of the original vascular lesion.

Interstitial edema associated with hydrocephalus involves a different source. Ventricular fluid can move across the ependymal region into periventricular tissue when pressure and drainage conditions favor that route. Ventricular enlargement, altered tissue mechanics, and impaired disposal can coexist. This mechanism should not be equated with generalized endothelial leakage, and its detailed cerebrospinal fluid dynamics belong to Chapter 61.

Water channels influence the rate of movement across cellular membranes without determining its direction independently of gradients. Aquaporin-4 is prominent in astroglial membranes, including perivascular endfeet. It does not actively pump water or itself explain endothelial selectivity. Manley and colleagues found reduced swelling in aquaporin-4-deficient mice during experimental water intoxication and ischemic stroke. 2

The effect was different in experiments designed around extracellular fluid accumulation. Papadopoulos and colleagues found worse fluid accumulation in aquaporin-4-deficient mice in several vasogenic-edema preparations, supporting a role in clearance as well as entry. 3 The comparison shows why a pathway that facilitates swelling in one state may facilitate water removal in another. It does not establish an aquaporin-directed treatment for critically ill patients.

These mechanisms often mix within one lesion. A contusion can contain cellular swelling, protein-rich extracellular fluid, hemorrhage, and adjacent tissue with different permeability. Describing the dominant mechanism is useful, but treating the label as a uniform map of the entire brain can misrepresent both water movement and the response to osmotherapy.

62.4 Effective osmotic pressure depends on the barrier as well as the particles

Water movement responds to differences in water chemical potential. A solute concentration difference can draw water across a membrane when that solute does not equilibrate freely across it on the relevant timescale. Total measured osmolality counts particles; effective osmotic pressure weights their ability to sustain a gradient across the particular barrier.

The reflection coefficient, σ, expresses how effectively a solute contributes to osmotic pressure relative to the ideal impermeant-solute case. It is dimensionless. A value near one approaches full osmotic effectiveness, whereas a value near zero implies little sustained osmotic effect under the modeled conditions. It should not be interpreted literally as the percentage of molecules that bounce off a membrane.

For a simplified dilute-solution model, let Δπeffective be the osmotic pressure favoring water movement from tissue toward plasma, R the gas constant in J/mol/K, T absolute temperature in K, and ΔCi the plasma-minus-tissue particle concentration difference for solute i in mol/m³. Positive differences favor water movement toward plasma:

Consider an illustrative concentration difference of 5 mOsm/L, equivalent to 5 osmoles/m³. At 310 K, using R = 8.314 J/mol/K and an assumed σ of 0.9, the pressure is 8.314 × 310 × 0.9 × 5 = 11,598 Pa, or approximately 87 mmHg using 133.3 Pa per mmHg. If the assumed coefficient were 0.2, the same particle difference would generate approximately 19 mmHg. These coefficients illustrate different barrier properties; they are not assigned bedside values for sodium or mannitol.

The calculation explains why a small effective concentration difference can exert a substantial osmotic force. It does not specify how many milliliters will leave the brain. Water redistribution changes the concentrations, while transport, vascular supply, and changing barrier properties alter the gradient. The blood–brain barrier and brain-cell membranes also have different solute transport characteristics. A solute that is ineffective after systemic equilibration may still generate a transient brain gradient when its movement across these barriers lags behind a plasma change.

62.5 Hydrostatic and oncotic forces contribute according to barrier properties

Hydrostatic pressure favors fluid movement from a higher-pressure compartment toward a lower-pressure one. In the cerebral microcirculation, capillary pressure, surrounding tissue pressure, and downstream venous conditions help determine that force. The arterial pressure displayed on a monitor is not itself the pressure inside every capillary. Resistance changes and venous outflow conditions modify pressure transmission along the vascular pathway.

Protein oncotic pressure is the osmotic contribution of macromolecules such as albumin. In many peripheral capillary beds, relatively permeable small solutes equilibrate rapidly, leaving protein-related forces prominent in transvascular water exchange. The blood–brain barrier restricts small-solute movement much more strongly. Small-solute osmotic gradients can therefore be important in the brain even when protein concentration has not changed.

Zornow and colleagues separately altered plasma osmolality and colloid oncotic pressure in anesthetized, neurologically normal rabbits. Reducing osmolality increased cortical water, whereas a substantial isolated oncotic-pressure reduction did not produce the same response in that preparation. The experiment demonstrates why peripheral edema reasoning cannot simply be transferred to an intact cerebral barrier. 4

An injured barrier changes the balance again. Increased hydraulic permeability can make hydrostatic filtration more important. Passage of proteins changes the effective protein gradient, while altered ion permeability can weaken the osmotic gradient created by a small solute. The result depends on which permeability properties changed, in which region, and over what interval. Neither low plasma albumin nor protein leakage alone determines the net water flux.

Venous congestion can increase cerebral vascular volume and downstream pressure while also affecting perfusion. Increasing systemic arterial pressure may support flow in one setting yet transmit more hydrostatic pressure to vulnerable microvessels in another. Lowering it may reduce filtration but jeopardize supply. These competing pathways connect edema physiology with the autoregulatory and outflow conditions discussed in Chapter 59.

The practical implication is to assess perfusion, barrier function, and systemic fluid conditions together. Increasing plasma oncotic pressure is not automatically a brain-dehydrating treatment, and a plasma osmotic change cannot be interpreted without considering the circulation through which it acts.

62.6 Regional permeability and transport make osmotherapy spatially uneven

Osmotherapy raises the effective osmotic activity of blood relative to accessible brain water. Water can then move toward the vascular compartment, reducing tissue volume where the barrier and gradients permit. This mechanism requires access to perfused vessels and sufficient restriction of the administered osmole. It does not imply uniform dehydration of every injured region.

Relatively preserved tissue may retain a stronger gradient than tissue with severe barrier disruption. Where an agent penetrates more readily, its plasma-to-tissue concentration difference may diminish. Ischemic regions with poor vascular delivery may respond differently again. Thus a global plasma measurement can coexist with several local transport states inside the same brain.

Once extracellular fluid accumulates, its subsequent movement depends on tissue pathways as well as vascular exchange. Reulen and colleagues studied cold-injury edema in cats, using tracers and interstitial pressure measurements. Their findings supported pressure-driven bulk movement through white matter rather than diffusion alone as the main explanation for edema spread in that model. 5 The model does not define all clearance pathways in an injured human brain, but it demonstrates that tissue transport resistance matters.

An early fall in ICP can also include vascular effects. Osmotic intravascular expansion, altered blood rheology, and a vascular-tone response can change cerebral blood volume alongside tissue water redistribution. The relative contributions depend on the agent, systemic hemodynamics, and autoregulatory state. A pressure response therefore cannot be translated directly into a measured mass of water removed.

Different compartments also equilibrate at different speeds. Plasma, ventricular cerebrospinal fluid, extracellular fluid near a lesion, and intracellular water do not share one instantaneous concentration. A cerebrospinal fluid sample is consequently not a direct measurement of the osmotic environment around every injured cell.

These distinctions explain why repeated assessment requires both anatomical and temporal information. The distribution of edema on imaging, the behavior of a verified pressure signal, and the systemic response describe complementary parts of the process. Improvement in one of them does not establish that every region has recovered its normal transport conditions.

62.7 Osmolyte adaptation changes the response to the next osmotic change

Cells exposed to increased extracellular tonicity initially lose water. They can then accumulate inorganic ions and organic osmolytes, helping restore volume despite the continued external osmotic condition. With sustained hypotonicity, the corresponding response includes loss of cellular solutes. These adaptations protect volume but alter the starting state for a later change.

A simple constructed water-compartment model makes the sequence explicit. Let V1 and V2 be osmotically responsive cell-water volumes in liters, and C1 and C2 the external effective particle concentrations in mOsm/L. If intracellular effective osmole amount stays fixed and water equilibrates without a significant opposing mechanical pressure:

Suppose the initial cell-water volume is 1.00 L at an effective concentration of 300 mOsm/L. Raising the external concentration to 310 gives 1.00 × 300/310 = 0.968 L, a water loss of approximately 32 mL. If solute accumulation subsequently restores volume to 1.00 L at 310, the effective intracellular osmole amount is now 310 mOsm. Returning the external concentration to 300 before those osmoles leave would give 310/300 = 1.033 L.

The second expansion reflects retained cellular solute, not failure of osmosis. The model omits nonosmotic cell volume, membrane transport during the transition, mechanical constraints, and heterogeneous tissue. Its concentrations and volumes are illustrative rather than human treatment thresholds or predictions for total brain volume.

Lien and colleagues directly measured organic solutes in hypernatremic rats and documented accumulation and unequal recovery of different osmolytes. 6 Their related chronic-hyponatremia experiments demonstrated depletion and delayed reaccumulation of organic solutes during correction. 7 These studies support a changing cellular composition, while providing no universal safe human transition rate.

Prior osmotic exposure therefore matters when interpreting both continued therapy and its withdrawal. Current sodium or osmolality does not reveal how much cellular adaptation has occurred. The degree of injury, underlying dysnatremia, renal function, and duration and pattern of exposure all influence the next response.

62.8 Mannitol links transient extracellular osmoles to renal elimination

Mannitol adds an extracellular osmole that normally crosses intact cell membranes and the blood–brain barrier relatively slowly. Its initial distribution raises plasma osmotic activity, drawing water from cells into extracellular fluid and expanding the vascular compartment. Where cerebral exclusion is sufficient, the plasma-to-brain gradient can reduce brain water and volume.

The agent is filtered by the kidneys and is poorly reabsorbed, so its presence in tubular fluid promotes osmotic diuresis. The sequence can therefore move from initial intravascular expansion to substantial urinary water and electrolyte loss. The final circulatory state depends on renal filtration, the amount retained, concurrent fluids, and other losses. Diuresis does not demonstrate that the excreted water came specifically from edematous brain.

Rudehill and colleagues measured plasma and cerebrospinal fluid mannitol during intracranial tumor surgery. They observed distribution and elimination phases, variation in cerebrospinal fluid concentrations, and early circulatory changes. The small surgical study illustrates compartmental kinetics and intravascular effects; it does not establish one elimination time for patients with renal failure or diffuse brain injury. 8

Reduced renal clearance can prolong systemic mannitol exposure. Retained osmoles can sustain water movement into extracellular fluid, contributing to vascular expansion and pulmonary congestion in a susceptible circulation. Conversely, effective renal elimination with ongoing diuresis can reduce circulating volume and arterial pressure, potentially impairing cerebral perfusion. Renal injury can itself alter subsequent clearance, making repeated exposure different from the first administration.

Measured sodium may fall during the initial extracellular water shift without proving sodium depletion. Later water and electrolyte losses can change the direction of the sodium response. Thus sodium, urine output, and measured osmolality must be interpreted together with fluid balance and kidney function.

The intended brain gradient is also time dependent. As plasma concentration falls, its relation to local tissue concentration changes, especially where the barrier is abnormal. A persistent osmotic effect in blood is not synonymous with a persistent favorable gradient across every brain region. No fixed plasma half-life or urine-output response can replace assessment of the current compartments and circulation.

62.9 Hypertonic sodium solutions change salt balance as well as water distribution

Hypertonic sodium solutions increase extracellular effective osmoles. The accompanying anion contributes particles and maintains electroneutrality. The intact cerebral barrier restricts rapid passive equilibration of sodium salts while also regulating active ion transport. Increasing plasma effective tonicity can therefore favor water movement from brain toward blood, but the effect still depends on regional permeability and changing tissue composition.

Systemically, sodium distributes within extracellular fluid rather than remaining exclusively inside blood vessels. Water drawn from cells initially expands that compartment, while vascular–interstitial exchange and renal handling determine its later distribution. Changes in arterial pressure, cardiac output, and venous pressure can accompany the osmotic effect. Volume expansion may support perfusion in one setting and aggravate pulmonary or venous congestion in another.

Unlike mannitol, the administered salt becomes part of regulated sodium and anion balance. Natriuresis and water excretion can reduce its persistence, while impaired renal excretion can prolong salt and fluid accumulation. A maintained plasma sodium concentration may therefore require a changing balance of input, renal loss, and water movement. It is not a direct measurement of ongoing water removal from the brain.

The formulation also influences acid–base physiology. A chloride load can lower strong-ion difference and contribute to hyperchloremic metabolic acidosis. Sodium concentration alone does not describe that effect. Solutions containing other anions add their own metabolism and buffering consequences, so equal nominal osmotic loads need not have identical systemic effects. Chapters 53–55 explain these relationships in detail.

The comparison with mannitol cannot be reduced to equal solution volume or equal mass. The number of effective particles, speed of distribution, renal elimination, and accompanying fluid all matter. Even equal calculated osmolar loads do not guarantee identical local gradients or circulatory responses.

These mechanisms explain the need to follow sodium, chloride, acid–base state, renal function, and fluid balance alongside neurological observations. They do not establish a sodium concentration that is universally effective or safe, nor do they prove that a sustained sodium increase prevents clinically important swelling.

62.10 Rebound and laboratory monitoring require a compartmental explanation

An osmotic gradient can weaken or reverse if the administered osmole enters injured tissue and then clears from plasma more quickly than from that tissue. Kaufmann and Cardoso demonstrated mannitol accumulation and worsening vasogenic edema after repeated administration in a cold-injury cat model. 9 This supports one rebound mechanism under abnormal-barrier conditions; it does not establish its frequency in every clinical setting.

Cellular osmolyte adaptation provides another mechanism for water return when external tonicity falls. However, rising ICP after a therapy change is not automatically osmotic rebound. Progressing injury, hemorrhage, venous congestion, altered vascular tone, or impaired cerebrospinal fluid drainage can produce a similar pressure trajectory. Establishing the mechanism requires the associated time course and measurements.

Measured osmolality includes both effective and less effective solutes. For a constructed laboratory example, let sodium, glucose, and urea be 145, 6, and 5 mmol/L, respectively. A conventional calculated osmolarity is 2 × 145 + 6 + 5 = 301 mOsm/L. If measured osmolality is 321 mOsm/kg, the approximate osmolal gap is 321 − 301 = 20 mOsm/kg, using the customary dilute-plasma comparison of per-kilogram and per-liter values.

The gap indicates particles not adequately represented by that calculation, plus analytical and formula effects. Mannitol can contribute during exposure, but the gap does not uniquely identify its concentration, brain penetration, or regional reflection coefficient. Excluding equilibrated urea from a simple systemic tonicity estimate gives 2 × 145 + 6 = 296 mOsm/L; that estimate still omits administered effective osmoles and cannot describe transient brain urea gradients.

A randomized postcraniotomy measurement study by Li and colleagues demonstrated that agreement between measured and calculated osmolality depended on the agent and formula. It studied measurement performance rather than neurological outcome. 10

Serial renal function, urine losses, fluid balance, electrolytes, and neurological observations therefore remain necessary. Neither one osmolal gap nor one total-osmolality cutoff measures the desired cerebral effect or establishes universal renal safety.

62.11 Pressure reduction and neurological benefit require different evidence

Small physiological studies can establish an immediate response without establishing functional benefit. Cottenceau and colleagues randomized severe-TBI patients to equiosmolar mannitol or hypertonic saline and examined ICP, circulation, and cerebral metabolic measurements. Both agents reduced ICP, while responses in other variables differed. The selected study supports physiological activity, not a universal comparative advantage for long-term neurological recovery. 11

COBI tested a different question: adding a continuous hypertonic saline strategy to standard care after moderate to severe traumatic brain injury. Standard care could include rescue osmotherapy. The primary outcome was the distribution of six-month Extended Glasgow Outcome Scale scores, with a dimensionless common odds ratio above one favoring the continuous strategy. Among 370 randomized patients, 359 had primary outcome data. The adjusted common odds ratio was 1.02, with a 95% confidence interval of 0.71–1.47. The trial did not demonstrate improved functional outcome, and the interval retained clinically relevant uncertainty. 12

That result cannot determine the value of every rescue administration, compare all sodium formulations with mannitol, or prove that lowering dangerous pressure has no physiological importance. The population, continuous preventive strategy, comparator, and outcome must remain attached to the finding. Similarly, the trial's time-varying pressure observations do not directly measure cellular osmolytes or prove a particular rebound mechanism.

Contemporary observational findings require a different interpretation. Ahmed and Kuo's 2026 propensity-matched severe-TBI analysis associated earlier hyperosmolar therapy with shorter ICU and ventilation durations, without significant differences in mortality or operative intervention. The design cannot establish that timing caused those associations; residual differences in injury, treatment selection, and the evolving indication remain possible. 13

Protocols and registry entries, including those describing an agent-comparison trial, specify planned research rather than treatment results. They cannot settle comparative efficacy before the actual outcome evidence is available and assessed.

The evidentiary sequence is therefore explicit: an agent may change blood osmolality, alter brain water or vascular volume, and reduce ICP. Whether a strategy improves survival or functional recovery depends on additional benefits, harms, and clinical evidence.

62.12 Integrate the edema mechanism, the gradient, and systemic tolerance

The first question is what is enlarging and where. Cellular swelling, extracellular vascular leakage, ventricular-fluid movement, and vascular engorgement can require different explanations even when ICP rises in each case. Imaging, examination, pressure trends, and the underlying disease help identify these contributions. A single edema label cannot replace that anatomical assessment.

The next question is whether an effective osmotic gradient is likely to remain available. Relevant factors include current and prior tonicity, regional barrier dysfunction, vascular delivery, renal clearance, and cellular adaptation. A rise in plasma osmolality confirms a blood change; it does not confirm uniform removal of water from the lesion. An observed pressure fall should be interpreted alongside perfusion, systemic circulation, and the evolution of mass effect.

Guidance also distinguishes populations and outcomes. The Neurocritical Care Society guideline currently listed by the society conditionally favors hypertonic sodium solutions over mannitol for initial ICP or edema management in TBI, with low-quality evidence, while advising against expecting either agent to improve neurological outcome on that basis. For acute ischemic stroke, it allows either agent initially and identifies insufficient outcome evidence. These are pathology-specific recommendations, not interchangeable conclusions about every cause of brain swelling. 14

Systemic tolerance remains part of the cerebral assessment. Initial vascular expansion, later diuresis, retained salt or mannitol, acid–base changes, and renal dysfunction can alter arterial supply and venous drainage. A favorable ICP response accompanied by deteriorating circulation or pulmonary function describes competing physiological effects, rather than an unqualified success.

Finally, response must be followed through time. Changing permeability, ongoing injury, osmole clearance, and recovery of cellular solute balance can alter the effect of the same intervention. Apparent loss of response or later pressure elevation should trigger a renewed explanation of those mechanisms rather than an assumption that one laboratory target defines the solution.

This framework explains osmotherapy without supplying a dose, sodium or osmolality target, or universal transition rate. Its purpose is to connect the observed brain response with the gradients and systemic conditions that produced it, while keeping mechanistic plausibility separate from proven neurological benefit.

References

1. O'Donnell ME, Tran L, Lam TI, Liu XB, Anderson SE. Bumetanide inhibition of the blood-brain barrier Na-K-Cl cotransporter reduces edema formation in the rat middle cerebral artery occlusion model of stroke. J Cereb Blood Flow Metab. 2004;24:1046–1056. https://doi.org/10.1097/01.WCB.0000130867.32663.90

2. Manley GT, Fujimura M, Ma T, et al. Aquaporin-4 deletion in mice reduces brain edema after acute water intoxication and ischemic stroke. Nat Med. 2000;6:159–163. https://doi.org/10.1038/72256

3. Papadopoulos MC, Manley GT, Krishna S, Verkman AS. Aquaporin-4 facilitates reabsorption of excess fluid in vasogenic brain edema. FASEB J. 2004;18:1291–1293. https://doi.org/10.1096/fj.04-1723fje

4. Zornow MH, Todd MM, Moore SS. The acute cerebral effects of changes in plasma osmolality and oncotic pressure. Anesthesiology. 1987;67:936–941. https://doi.org/10.1097/00000542-198712000-00010

5. Reulen HJ, Graham R, Spatz M, Klatzo I. Role of pressure gradients and bulk flow in dynamics of vasogenic brain edema. J Neurosurg. 1977;46:24–35. https://doi.org/10.3171/jns.1977.46.1.0024

6. Lien YHH, Shapiro JI, Chan L. Effects of hypernatremia on organic brain osmoles. J Clin Invest. 1990;85:1427–1435. https://doi.org/10.1172/JCI114587

7. Lien YHH, Shapiro JI, Chan L. Study of brain electrolytes and organic osmolytes during correction of chronic hyponatremia. Implications for the pathogenesis of central pontine myelinolysis. J Clin Invest. 1991;88:303–309. https://doi.org/10.1172/JCI115292

8. Rudehill A, Gordon E, Ohman G, Lindqvist C, Andersson P. Pharmacokinetics and effects of mannitol on hemodynamics, blood and cerebrospinal fluid electrolytes, and osmolality during intracranial surgery. J Neurosurg Anesthesiol. 1993;5:4–12. https://doi.org/10.1097/00008506-199301000-00002

9. Kaufmann AM, Cardoso ER. Aggravation of vasogenic cerebral edema by multiple-dose mannitol. J Neurosurg. 1992;77:584–589. https://doi.org/10.3171/jns.1992.77.4.0584

10. Li Q, Chen H, Hao JJ, Yin NN, Xu M, Zhou JX. Agreement of measured and calculated serum osmolality during the infusion of mannitol or hypertonic saline in patients after craniotomy: a prospective, double-blinded, randomised controlled trial. BMC Anesthesiol. 2015;15:138. https://doi.org/10.1186/s12871-015-0119-4

11. Cottenceau V, Masson F, Mahamid E, et al. Comparison of effects of equiosmolar doses of mannitol and hypertonic saline on cerebral blood flow and metabolism in traumatic brain injury. J Neurotrauma. 2011;28:2003–2012. https://doi.org/10.1089/neu.2011.1929

12. Roquilly A, Moyer JD, Huet O, et al. Effect of continuous infusion of hypertonic saline vs standard care on 6-month neurological outcomes in patients with traumatic brain injury: the COBI randomized clinical trial. JAMA. 2021;325:2056–2066. https://doi.org/10.1001/jama.2021.5561

13. Ahmed N, Kuo YH. Association of early hyperosmolar therapy and operative interventions in severe traumatic brain injury. Injury. 2026;57:113671. https://doi.org/10.1016/j.injury.2026.113671

14. Cook AM, Jones GM, Hawryluk GWJ, et al. Guidelines for the acute treatment of cerebral edema in neurocritical care patients. Neurocrit Care. 2020;32:647–666. https://doi.org/10.1007/s12028-020-00959-7