Cerebral Perfusion Pressure and Cerebral Autoregulation

Cerebral perfusion pressure summarizes a driving-pressure relationship. Cerebral autoregulation describes how the vascular bed changes its resistance as that relationship changes. Neither quantity independently establishes adequate tissue perfusion. The clinically useful interpretation connects pressure reference, arterial inflow, venous outflow, vascular responsiveness, and the condition of the brain being supplied. This chapter develops those connections before examining what bedside indices and clinical trials can establish. Chapter 58 addresses intracranial pressure and compliance; Chapter 60 develops cerebral blood flow, oxygen delivery, and metabolism in greater detail.

59.1 A pressure difference provides the opportunity for flow

Blood reaches the brain because pressure decreases along its vascular pathways. The conventional bedside estimate of cerebral perfusion pressure, CPP, subtracts mean intracranial pressure, ICP, from mean arterial pressure referenced to the brain, MAPbrain. All three pressures are expressed in mmHg. When ICP reasonably represents the effective downstream constraint, the estimate is:

An illustrative MAPbrain of 85 mmHg and ICP of 20 mmHg therefore give CPP of 65 mmHg. This identifies a pressure difference, not a measured blood-flow rate. The same gradient can produce different flows through vessels of different caliber, length, viscosity, or patency. Conversely, vascular adjustment can preserve flow despite a change in pressure difference.

For a deliberately simplified steady-flow model, let Q be cerebral blood flow in mL/min per 100 g of tissue and R effective vascular resistance in mmHg per (mL/min per 100 g). If CPP represents the relevant driving gradient:

Suppose CPP is 70 mmHg and Q is 50 mL/min per 100 g. The inferred resistance is 70/50 = 1.4 in the stated resistance units. If CPP falls to 50 mmHg while resistance remains 1.4, flow becomes 50/1.4 = 35.7 mL/min per 100 g. If resistance instead decreases to 1.0, the model preserves flow at 50. This is the basic purpose of autoregulatory vasodilation: changing resistance can buffer the flow consequence of reduced pressure.

These values are constructed examples, not physiological limits or treatment targets. Cerebral vessels form a branching, pulsatile, collapsible network; their resistance is neither spatially uniform nor necessarily constant during the measurement. A calculated resistance also inherits errors in the assumed pressure gradient and flow measurement. The model remains useful because it separates two questions that a CPP number alone cannot answer: how much driving pressure is available, and how does the vascular bed respond to it?

59.2 Pressure reference and height change the numerical CPP

An arterial pressure measured at heart level is not the same pressure as that within an artery above the heart. Gravity produces a hydrostatic pressure difference. An elevated head can therefore have a lower local arterial pressure even when the heart-level monitor displays an unchanged MAP. Subtracting a brain-referenced ICP from heart-referenced arterial pressure mixes reference levels and can overstate the local driving gradient.

For an illustrative blood-column calculation, let ΔP be the pressure decrease with elevation in pascals, ρ blood density in kg/m³, g gravitational acceleration in m/s², and Δh the vertical height difference in meters. Then ΔP = ρgΔh. Using density 1060 kg/m³, acceleration 9.81 m/s², and 133.3 Pa per mmHg gives approximately 0.78 mmHg per vertical centimeter. This coefficient describes blood; the water-like fluid column in an external ICP measuring system has a slightly different coefficient. The vertical distance, rather than bed angle alone, determines the correction.

Consider a particular position in which the chosen brain reference is 25 cm above the heart. If heart-level MAP is 90 mmHg, the modeled hydrostatic decrease is 0.78 × 25 = 19.5 mmHg, giving MAPbrain of 70.5 mmHg. With a brain-referenced ICP of 20 mmHg, CPP is approximately 70.5 − 20 = 50.5 mmHg. Combining the uncorrected heart-level MAP with that ICP would instead give 70 mmHg. The discrepancy is a reference problem, not evidence that cerebral physiology suddenly changed between calculations.

Lele and colleagues directly compared measurement techniques and transducer levels in neurocritically ill patients, documenting materially different MAP and calculated CPP values at heart and external-auditory-meatus levels. Their observation confirms the practical importance of reporting the reference. It does not provide one universal correction for every patient's geometry. 1

The external auditory meatus or tragus is a convenient anatomical surrogate, not the location of every cerebral vessel. Liquid-filled measuring systems also require correct leveling and zeroing. Head elevation can change ICP, venous drainage, and systemic circulation as well as arterial hydrostatics; the example holds the specified measured state for arithmetic rather than predicting the entire response to repositioning. Trials and clinical records should therefore identify their pressure reference before their numerical CPP values are compared.

59.3 Intracranial pressure is an approximation to the downstream constraint

The cerebral venous circulation is not a rigid tube connected directly to the right atrium. Thin-walled veins can narrow where their intravascular pressure approaches surrounding pressure. In a vascular-waterfall model, an upstream flow-limiting segment partly separates cerebral venous pressure from a lower downstream sinus or extracranial venous pressure. Under those conditions, lowering central venous pressure further may not proportionately increase cerebral flow.

Luce and colleagues investigated cerebral venous outflow in anesthetized dogs during changes in positive end-expiratory pressure, venous pressures, and experimentally raised ICP. Their results supported a Starling-resistor mechanism between cerebral veins and the sagittal sinus. This is direct mechanistic evidence for conditional pressure transmission, not a universal human bedside formula. 2

If venous pressure becomes sufficiently elevated, the outflow situation changes. Venous congestion can increase upstream pressure and intracranial blood volume, while reducing the available arterial-to-venous gradient. A heart-level CVP cannot simply be substituted for local cerebral venous pressure: height, venous collapse, collateral drainage, and intrathoracic pressure transmission all intervene.

A constructed example isolates the limitation. Let MAPbrain be 80 mmHg and ICP 15 mmHg. Conventional CPP is 80 − 15 = 65 mmHg. If a particular vascular pathway instead has an effective local downstream pressure of 25 mmHg, its modeled driving gradient is 80 − 25 = 55 mmHg. The example illustrates why elevated local venous pressure could matter despite an apparently unchanged conventional CPP. It does not propose measuring that local pressure from CVP or replacing clinical CPP with an unvalidated calculation.

Arterial wall tone adds another possible constraint. Critical closing pressure is an estimated arterial pressure at which modeled flow ceases; it may incorporate surrounding pressure and active vascular tension. In head-injured patients, Czosnyka and colleagues found that such estimates related to ICP and autoregulation but could not accurately substitute for measured ICP or CPP. 3 Conventional CPP is therefore a useful common descriptor whose physiological boundary assumptions must remain visible.

59.4 Autoregulation changes vascular caliber and intracranial blood volume

Pressure autoregulation arises from coordinated changes in vascular tone. An increase in arterial transmural pressure can stimulate smooth-muscle contraction, whereas a decrease favors relaxation. Membrane electrical behavior, calcium-dependent contraction, endothelial influences, and local metabolic signals contribute. The response concerns pressure across the vessel wall, not simply the systemic arterial pressure displayed elsewhere in the circulation.

Increased ICP can therefore influence vascular behavior in two ways. It reduces the conventional perfusion gradient and changes the external pressure surrounding vessels. Autoregulatory dilation may initially reduce resistance and help maintain flow. However, dilation also changes vascular volume. In a poorly compliant intracranial compartment, an increase in blood volume can raise ICP further, narrowing the driving gradient and encouraging additional dilation. This potential feedback links autoregulation with the pressure–volume relationships developed in Chapter 58.

Direct cranial-window experiments by Kontos and colleagues in anesthetized cats showed that cerebral arterial and arteriolar caliber responses depended on vessel size and the pressure range tested. Larger surface arteries as well as smaller arterioles contributed to regulation. The vascular response was distributed across the network rather than confined to one identical resistance segment. 4

This distribution matters when interpreting simple radius arguments. Narrowing one vessel can substantially increase its resistance, but the whole network also changes segmental pressure, collateral flow, and the behavior of other vessels. A fourth-power radius relationship for a rigid cylindrical tube cannot predict the complete living cerebral response without specifying those assumptions.

When vascular compensation is exhausted or disrupted, pressure changes are transmitted more directly into flow. Increased arterial pressure may then increase blood volume and capillary hydrostatic pressure, with consequences that depend on blood–brain barrier integrity and venous outflow. Conversely, low driving pressure may reduce flow despite maximal dilation. These are mechanisms of vulnerability, not proof that every pressure-associated ICP increase identifies failed autoregulation or that every pressure increase necessarily causes edema.

59.5 The static curve is a conditional relationship, not a universal plateau

Static autoregulation compares relatively settled flow states at different pressures. Its familiar diagram contains a middle region in which resistance adjustments limit flow changes, bordered by regions where flow becomes more pressure dependent. The diagram expresses reserve and its eventual limitation. It should not be read as proof that every brain has perfectly constant flow between two fixed numerical pressures.

The apparent curve depends on which pressure and flow were measured, how pressure was changed, how long the system was allowed to settle, and which other determinants remained constant. An arterial-pressure curve in a healthy person with low, stable ICP is not automatically a CPP curve for an injured brain with changing ICP. Measurements from different individuals also cannot establish the entire pressure-response curve of one individual.

Lucas and colleagues studied healthy volunteers during controlled pharmacological changes in arterial pressure while accounting for carbon dioxide. Middle cerebral artery velocity changed across the investigated range rather than displaying a completely horizontal plateau. Their experiment challenges the assumption that a finite pressure–velocity slope necessarily represents pathological failure. It also retains the limitations of vasoactive interventions and velocity as a flow surrogate. 5

Vascular history can modify the response. In the classic study by Strandgaard and colleagues, patients with severe arterial hypertension had higher lower autoregulatory limits than normotensive controls during the experimental conditions. 6 That finding supports adaptation in a selected population; it does not establish that every person with a hypertension diagnosis has the same right-shifted curve or needs a fixed extra increment of pressure.

Brain injury introduces further variation. Vascular responsiveness may differ between hemispheres, between injured and less affected tissue, and over time. A region with exhausted dilation downstream from an arterial obstruction may be pressure dependent while another region retains reserve. The useful physiological question is therefore how effectively the relevant vascular territory buffers pressure changes under current conditions. A remembered textbook plateau cannot answer that question or identify one universal lower or upper CPP limit.

59.6 Dynamic autoregulation describes the response to a changing pressure

Pressure can change faster than vascular resistance adjusts. Dynamic autoregulation examines the time course of flow recovery following a transient disturbance or during ongoing fluctuations. Effective regulation does not require an absolutely unchanged instantaneous flow. A brief pressure-related flow change followed by recovery can demonstrate active adjustment, whereas a sustained change may indicate less effective buffering under the conditions tested.

Aaslid and colleagues studied abrupt pressure decreases induced by release of inflated thigh cuffs in healthy volunteers while recording middle cerebral artery velocity. Recovery differed with carbon dioxide state, demonstrating that the speed of regulation depended on the vascular condition at the time of testing. 7 The experiment established a dynamic measurement approach, not a universal time within which every patient's cerebral circulation must recover. Interpreting velocity changes as proportional flow changes also assumes that the insonated artery's cross-sectional area remains sufficiently stable.

Dynamic and static assessments interrogate related but distinct features. Tiecks and colleagues found close agreement between them in a small, controlled anesthetic study in which autoregulation was experimentally altered. 8 In contrast, de Jong and colleagues found no linear correlation between dynamic and steady-state indices in healthy older adults studied with several methods. 9 The combined evidence discourages assuming that one index completely characterizes all autoregulatory performance.

Frequency-domain methods describe how pressure fluctuations relate to flow or velocity fluctuations at different frequencies. Gain quantifies relative fluctuation amplitude, phase describes their timing relation, and coherence reflects how consistently a linear relationship accounts for the signals at the frequency examined. These quantities require the relevant units, normalization, frequency range, and analysis method to be reported. A low gain can be compatible with buffering, but poor signal quality or weak physiological excitation can also affect it.

Autoregulatory indices derived from fitted response models likewise summarize model behavior rather than measuring a vessel directly. Breathing, carbon dioxide, heart-rate changes, and concurrent systemic responses can alter both input and output signals. Dynamic analysis is most informative when these influences and the clinical state are documented, rather than treating one dimensionless score as a complete map of pressure tolerance.

59.7 Carbon dioxide, metabolic state, and injury modify the available reserve

Cerebral autoregulation operates within a vascular system already responding to chemical and metabolic demands. Hypercapnia generally favors cerebral vasodilation, while hypocapnia favors constriction. A bed that is already substantially dilated may have less remaining dilatory reserve when pressure falls. A change in carbon dioxide can therefore alter baseline flow, blood volume, and the measured response to pressure without changing the nominal CPP.

The human cuff-release observations discussed above directly demonstrated carbon dioxide dependence of dynamic regulation. They do not establish that inducing hypocapnia is beneficial: a faster measured response can coexist with reduced baseline flow and different tissue oxygen availability. The physiological endpoint being improved must be specified before an intervention can be described as favorable.

Anesthetic and sedative conditions also matter. Changes in neuronal activity and metabolic demand alter the flow required by the tissue, while some agents have direct vascular effects. The controlled anesthetic comparison by Tiecks and colleagues illustrates that a response measured under one drug condition need not persist under another. It should not be converted into a general ranking of clinical anesthetic regimens without the relevant outcome evidence.

Hypoxemia, anemia, temperature, seizures, and altered cardiac output can further modify the circulation being tested. Their effects include changes in oxygen content, metabolic demand, vascular tone, and systemic pressure. Chapter 60 develops these determinants of flow–metabolism matching. Here the essential implication is that pressure reactivity cannot be interpreted independently of the physiological state that supplies its baseline tone.

After injury, vascular obstruction, microvascular dysfunction, edema, and disturbed signaling may coexist. Preserved responsiveness in one monitored territory does not prove adequate flow in another, and an arterial segment supplying a threatened region may already be maximally dilated. Similarly, positive-pressure ventilation can change arterial pressure, cerebral venous outflow, oxygenation, and carbon dioxide together. Any observed change in CPP or autoregulation must be reconciled with those simultaneous pathways rather than assigned to one pressure effect by assumption.

59.8 Pressure reactivity is an indirect window into vascular behavior

Continuous invasive monitoring permits analysis of slow changes in arterial pressure and ICP. If increased arterial pressure produces compensatory vasoconstriction, the resulting reduction in vascular volume may limit or reverse an ICP increase. If vessels instead behave more passively, arterial pressure and ICP may rise together. This reasoning underlies the pressure reactivity index, PRx.

Let MAPslow and ICPslow represent appropriately averaged slow-wave pressure signals, both originally measured in mmHg. PRx is their moving correlation coefficient and is dimensionless:

Positive values indicate that the two signals tend to move together over the analyzed interval. Negative values indicate opposing movement, while values near zero indicate little linear association. These descriptions concern the recorded signals. Their physiological interpretation requires the assumed pathway from arterial pressure through vascular tone and volume to ICP.

Czosnyka and colleagues developed and studied continuous pressure-reactivity assessment in patients with head injury. The index related to another autoregulatory measure and to clinical outcome, supporting physiological and prognostic relevance. It was not a trial showing that changing treatment to normalize PRx improves recovery. 10

The method also depends on the intracranial pressure–volume environment. A vascular-volume change may produce a different ICP response when compliance changes, after decompression, or when cerebrospinal fluid drainage affects the signal. Coughing, suction, movement, transducer artifacts, and simultaneous treatment changes can create associations unrelated to isolated pressure autoregulation. Insufficient spontaneous pressure variation can make a correlation unstable or uninformative.

PRx is consequently a surrogate for pressure-related vascular reactivity, not a direct blood-flow measurement. A favorable value cannot exclude regional ischemia, low oxygen content, or metabolic failure. Other indices based on Doppler velocity, oxygenation, or optical signals interrogate different compartments and assumptions. Agreement between methods can strengthen an interpretation, but the indices are not interchangeable simply because each is labeled an assessment of autoregulation.

59.9 CPPopt identifies an estimated reactivity optimum, not a proven treatment optimum

Pressure reactivity can be analyzed across the CPP values observed during monitoring. In some patients, plotting PRx against CPP produces a U-shaped relationship: reactivity appears less favorable at lower and higher pressures, with a minimum between them. CPPopt names the CPP associated with the minimum according to the calculation method. The word “optimal” refers to that measured reactivity criterion, not a directly demonstrated optimum for survival or neurological function.

Steiner and colleagues retrospectively analyzed prospectively collected head-injury monitoring data and identified CPPopt in a substantial proportion of patients. Deviation from the estimated value was associated with outcome, motivating prospective investigation. The study was an observational analysis, so injury severity, treatment selection, and the ability to maintain pressure could contribute to the relationship. 11

More recent work by Pelah and colleagues retrospectively examined multicenter traumatic-brain-injury recordings. Aggregate brain-tissue oxygenation patterns were favorable near calculated CPPopt, but individual relationships varied, and the estimated value had not been targeted during treatment. The physiological association supports further investigation while leaving the effect of a CPPopt-directed intervention unresolved. 12

Several technical limitations follow from the construction. A useful curve requires sufficient high-quality data across an informative pressure range. A minimum at the edge of that range may reflect incomplete observation rather than a fully characterized optimum. Flat, irregular, or monotonic relationships may not provide a defensible minimum. Algorithms also differ in averaging windows, fitting requirements, artifact rejection, and the persistence demanded before a recommendation is displayed.

The estimate can change as vascular physiology evolves. Moreover, CPP itself includes ICP, while PRx also uses ICP, introducing shared-variable dependence that must be considered alongside the physiological explanation. Treatment alters the pressures sampled and can influence the curve being used to guide later treatment. These features make uncertainty, availability, and stability as important as the displayed CPPopt number. A software output should therefore be interpreted as an estimate generated under current measurement conditions.

59.10 Equal calculated CPP values can arise through different interventions

Raising arterial pressure and lowering ICP can increase conventional CPP, but they are not physiologically equivalent. An arterial-pressure increase changes cardiac workload and arterial transmural pressure. Lowering ICP changes the external mechanical environment of cerebral vessels and tissue. Both may alter resistance, blood volume, venous drainage, and systemic physiology, with different consequences depending on the intervention used.

Consider a constructed starting state with MAPbrain of 80 mmHg and ICP of 25 mmHg, giving CPP of 55 mmHg. Raising MAPbrain to 95 while ICP stays 25 gives 95 − 25 = 70 mmHg. Alternatively, keeping MAPbrain at 80 while ICP falls to 10 gives 80 − 10 = 70 mmHg. The final calculated gradients match, but absolute arterial and surrounding tissue pressures differ. The underlying interventions may also change oxygen content, cardiac output, fluid balance, and vascular tone.

With effective autoregulatory constriction, increased arterial pressure may reduce vascular volume and sometimes ICP. With impaired reactivity, the same arterial-pressure increase may raise flow and vascular volume more directly. Neither response can be assumed from the starting CPP. Likewise, an ICP reduction does not guarantee that injured microvascular pathways reopen or that metabolic function recovers.

Robertson and colleagues compared bundled management strategies in a randomized trial using assigned time blocks after severe head injury. The strategy emphasizing higher CPP and avoiding more marked hypocapnia reduced jugular venous desaturation, but did not improve neurological outcome and was associated with more respiratory distress syndrome. Because multiple management components differed, the study cannot isolate one pressure threshold as the sole cause of benefit or harm. 13

The broader lesson is that pressure is reached through a treatment with its own effects. Fluid administration, vasoactive support, ventilation changes, cerebrospinal fluid drainage, and other ICP-directed interventions carry different systemic and intracranial consequences. A favorable change in the calculated gradient should be considered alongside the response in ICP, circulation, pulmonary function, and relevant cerebral measurements. Arithmetic equivalence is not clinical equivalence.

59.11 Guideline ranges and individualized strategies answer different evidence questions

Physiological variability does not erase the need for clinical evidence. Population-based guidance provides a practical framework based on the available studies, whereas an individualized index proposes that a patient's current response may help refine that framework. These are different kinds of information. Neither should be represented as a directly measured universal boundary of cerebral viability.

The Brain Trauma Foundation's currently posted fourth-edition severe-TBI guidance gives a Level IIB recommendation for a CPP target range of 60–70 mmHg. It acknowledges uncertainty about the appropriate minimum within that range and the possible role of autoregulatory status. A separate Level III recommendation suggests considering avoidance of aggressive attempts to maintain CPP above 70 with fluids and pressors because of respiratory risk. These recommendations concern severe traumatic brain injury, rather than defining pressure tolerance in every neurological disease or healthy cerebral circulation. 14

COGiTATE randomized 60 monitored TBI patients to standard guideline-range management or an autoregulation-guided strategy. The study demonstrated feasibility, with no significant between-group difference in its principal treatment-intensity safety comparison. Its size and design addressed feasibility and safety signals rather than establishing improved neurological outcome or excluding uncommon harms. In particular, success against its predefined time-in-target feasibility criterion should not be mistaken for a proven survival effect. 15

The trial also illustrates the importance of implementation details. Arterial pressure was referenced to the tragus, the automated estimate was not available continuously for every observation, and clinical judgment remained necessary when a calculated target was unavailable. A target derived under those conditions cannot be transferred to a differently referenced monitor without accounting for the difference.

A 2026 observational study compared MAPopt estimated from noninvasive cranial-deformation waveforms with MAPopt derived from invasive pressure reactivity. Their reported association and agreement address a method-comparison question, rather than demonstrating that treatment guided by either estimate improves recovery. Moreover, an estimated MAPopt does not by itself supply CPP when the relevant ICP is unknown. 16

An observational association between time below an estimated boundary and poor outcome may reflect a biologically important exposure, more severe injury, treatment resistance, or several of these together. A causal treatment strategy requires prospective comparison of benefits and harms, including systemic effects and functional recovery. Therefore, individualized monitoring should be described with its demonstrated role, its uncertainty, and the population studied rather than presented as proof that a newly calculated number supersedes outcome evidence.

59.12 Interpret the pressure, the response, and the tissue together

A useful assessment begins by establishing that the numbers refer to the intended compartments. The arterial transducer reference, ICP device, waveform quality, body position, and recent interventions determine what can be inferred from calculated CPP. A change caused by leveling or sampling conditions should be recognized before it is interpreted as altered cerebral vascular behavior.

Next, reconstruct the trajectory. A falling CPP can result from declining arterial pressure, rising ICP, or both. The same average value can conceal very different pressure pulsations, durations, and fluctuations. Concurrent changes in carbon dioxide, oxygenation, sedation, temperature, cardiac output, or venous pressure can change both the vascular response and the meaning of an autoregulatory index.

Two constructed observations illustrate the reasoning without prescribing a pressure challenge. Suppose an initial MAPbrain of 85 mmHg and ICP of 20 gives CPP of 65 mmHg. A later state with MAPbrain 95 and ICP 18 gives CPP of 77 mmHg; the direction of ICP change is compatible with a reduction in vascular volume, but does not prove its cause. A different later state with MAPbrain 95 and ICP 26 gives CPP of 69 mmHg. Here ICP has absorbed much of the arterial-pressure increase, prompting assessment of reactivity, compliance, venous conditions, and other simultaneous events. A pair of observations is not enough to calculate a reliable continuous reactivity index.

Finally, connect pressure information with the condition of the tissue being supplied. Flow, oxygen content, extraction, and metabolic demand determine whether a given perfusion state is adequate. A reassuring global pressure or surrogate index cannot exclude a threatened region. Conversely, an abnormal local measurement requires consideration of probe location, technical quality, and the surrounding injury before it is generalized to the whole brain.

The purpose of CPP and autoregulation monitoring is to make these mechanisms visible. Pressure arithmetic identifies a potential constraint; vascular-response assessment explores reserve; complementary measurements examine consequences. Their integration supports a more defensible physiological explanation while preserving the distinction between a plausible mechanism, an observed association, and a treatment proven to improve patient outcomes.

References

1. Lele AV, Wilson D, Chalise P, Nazzaro J, Krishnamoorthy V, Vavilala MS. Differences in blood pressure by measurement technique in neurocritically ill patients: a technological assessment. J Clin Neurosci. 2018;47:97–102. https://doi.org/10.1016/j.jocn.2017.10.079

2. Luce JM, Huseby JS, Kirk W, Butler J. A Starling resistor regulates cerebral venous outflow in dogs. J Appl Physiol Respir Environ Exerc Physiol. 1982;53:1496–1503. https://doi.org/10.1152/jappl.1982.53.6.1496

3. Czosnyka M, Smielewski P, Piechnik S, Al-Rawi PG, Kirkpatrick PJ, Matta BF, Pickard JD. Critical closing pressure in cerebrovascular circulation. J Neurol Neurosurg Psychiatry. 1999;66:606–611. https://doi.org/10.1136/jnnp.66.5.606

4. Kontos HA, Wei EP, Navari RM, Levasseur JE, Rosenblum WI, Patterson JL Jr. Responses of cerebral arteries and arterioles to acute hypotension and hypertension. Am J Physiol. 1978;234:H371–H383. https://doi.org/10.1152/ajpheart.1978.234.4.H371

5. Lucas SJE, Tzeng YC, Galvin SD, Thomas KN, Ogoh S, Ainslie PN. Influence of changes in blood pressure on cerebral perfusion and oxygenation. Hypertension. 2010;55:698–705. https://doi.org/10.1161/HYPERTENSIONAHA.109.146290

6. Strandgaard S, Olesen J, Skinhøj E, Lassen NA. Autoregulation of brain circulation in severe arterial hypertension. Br Med J. 1973;1:507–510. https://doi.org/10.1136/bmj.1.5852.507

7. Aaslid R, Lindegaard KF, Sorteberg W, Nornes H. Cerebral autoregulation dynamics in humans. Stroke. 1989;20:45–52. https://doi.org/10.1161/01.STR.20.1.45

8. Tiecks FP, Lam AM, Aaslid R, Newell DW. Comparison of static and dynamic cerebral autoregulation measurements. Stroke. 1995;26:1014–1019. https://doi.org/10.1161/01.STR.26.6.1014

9. de Jong DLK, Tarumi T, Liu J, Zhang R, Claassen JAHR. Lack of linear correlation between dynamic and steady-state cerebral autoregulation. J Physiol. 2017;595:5623–5636. https://doi.org/10.1113/JP274304

10. Czosnyka M, Smielewski P, Kirkpatrick P, Laing RJ, Menon D, Pickard JD. Continuous assessment of the cerebral vasomotor reactivity in head injury. Neurosurgery. 1997;41:11–17; discussion 17–19. https://doi.org/10.1097/00006123-199707000-00005

11. Steiner LA, Czosnyka M, Piechnik SK, Smielewski P, Chatfield D, Menon DK, Pickard JD. Continuous monitoring of cerebrovascular pressure reactivity allows determination of optimal cerebral perfusion pressure in patients with traumatic brain injury. Crit Care Med. 2002;30:733–738. https://doi.org/10.1097/00003246-200204000-00002

12. Pelah AI, Kazimierska A, Czosnyka M, Hawryluk GWJ. Optimal cerebral perfusion pressure in brain injury: physiological relationships and outcome. Neurosurgery. 2025;97:863–872. https://doi.org/10.1227/neu.0000000000003411

13. Robertson CS, Valadka AB, Hannay HJ, et al. Prevention of secondary ischemic insults after severe head injury. Crit Care Med. 1999;27:2086–2095. https://doi.org/10.1097/00003246-199910000-00002

14. Brain Trauma Foundation. Guidelines for the management of severe TBI, fourth edition: cerebral perfusion pressure thresholds. Official guidance, accessed November 10, 2026. https://braintrauma.org/coma/guidelines/severe-tbi

15. 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. J Neurotrauma. 2021;38:2790–2800. https://doi.org/10.1089/neu.2021.0197

16. Brasil S, Frigieri G, Taccone FS, et al. A noninvasive method for assessing optimal cerebral perfusion pressure. Crit Care. 2026;30:190. https://doi.org/10.1186/s13054-026-05926-w