Distributive Shock and Loss of Vascular Tone

Distributive shock involves the circulation's organization as well as its pressure. Arterial resistance, venous capacity, endothelial exchange, cardiac performance, and regional flow can change together. High output may accompany inadequate tissue oxygen availability, while the syndrome can also present with low output. Interpretation connects pressure, venous return, pump reserve, oxygen content, and demand through the patient's changing course.

71.1 Distributive Shock Is a Circulatory Pattern with Several Possible Causes

Vascular tone helps maintain pressure and allocate circulating blood between regions and vascular compartments. When its regulation fails, arteriolar resistance may fall, venous capacity may increase, and flow may become poorly matched to local requirements. Shock develops when the resulting circulation cannot sustain adequate cellular function; hypotension is an important observation within this process rather than its complete definition.

Sepsis is a major setting, but distributive physiology also occurs in anaphylaxis, loss of sympathetic vascular control after selected neurological injuries, and some drug-related states. Their initiating mechanisms, permeability changes, cardiac effects, and time courses differ. Clinical trial results from septic shock cannot automatically be transferred to all causes of low vascular tone.

In sepsis, altered endothelial and smooth-muscle signaling can reduce vascular responsiveness while simultaneously disturbing barrier function and microvascular flow. Those molecular mechanisms are developed in Chapter 66. Here, the essential consequence is that the circulation no longer behaves as an appropriately coordinated set of resistance vessels, capacitance vessels, and exchange networks.

The familiar warm, high-output phenotype represents one possible state. Low resistance can permit a large flow if venous return and cardiac reserve support it. Warm skin and a wide pulse pressure may accompany this pattern, yet neither proves adequate kidney, gut, myocardial, or cellular perfusion. Conversely, cool extremities can occur as low output, regional vasoconstriction, or administered vasopressors modify the phenotype.

Volume loss, myocardial dysfunction, pulmonary vascular loading, and mechanical ventilation can coexist with vasodilation. These contributions do not require abandoning the distributive component when output becomes low. They require describing a mixed physiological state rather than forcing every patient into a single exclusive category.

The roles of pooling, leakage, cardiac reserve, support, and demand evolve over hours. Measurements after substantial resuscitation describe a different operating state from the initial illness. The explanation must include both the disease and the interventions affecting the circulation.

71.2 Arterial Resistance Helps Generate Pressure but Does Not Specify Flow

Small arteries and arterioles normally regulate the resistance encountered by systemic flow. Reduced smooth-muscle tone can lower the pressure generated for a given cardiac output. The effect is distributed across many vascular beds, whose changes need not be equal. A whole-body resistance estimate compresses those differences into one number.

Let MAP be mean arterial pressure and RAP right atrial pressure in mmHg with a common reference, CO cardiac output in L/min, and SVR systemic vascular resistance in dyn·s/cm⁵. For an averaged lumped systemic circulation, the unit-conversion factor is 80 dyn·s/cm⁵ for each mmHg·min/L.

In a constructed state with RAP = 5 mmHg, CO = 8 L/min, and SVR = 500 dyn·s/cm⁵, the pressure difference is 8 × 500 ÷ 80 = 50 mmHg. MAP is therefore approximately 55 mmHg. A high output has not prevented low arterial pressure because resistance is low.

Now consider a second observed state with RAP unchanged, CO = 6 L/min, and SVR = 800 dyn·s/cm⁵. The difference is 6 × 800 ÷ 80 = 60 mmHg, giving MAP approximately 65 mmHg. Pressure is higher while flow is lower. These are illustrative states, not a prediction of one drug's effect or a choice of treatment targets.

The equation also explains a measurement limitation. Calculated SVR uses the same pressure and output measurements from which it is interpreted. If output falls while pressure is maintained, calculated resistance rises; this does not independently identify the molecular cause or prove uniform constriction of all tissues.

The arterial waveform adds information but remains conditional. Diastolic pressure depends on runoff, heart rate, arterial properties, and upstream conditions. Pulse pressure depends on stroke volume, compliance, wave reflection, and measurement location. Their interpretation can support a hypothesis about tone or flow without becoming direct measurements of either.

Local venous and tissue pressures, autoregulation, geometry, and regional resistance determine how systemic pressure becomes tissue flow. The lumped gradient cannot represent all organs.

71.3 Venous Capacity Can Increase Without External Blood Loss

Most systemic blood is contained in compliant venous compartments. Their pressure–volume relationship determines how much volume can be accommodated at low distending pressure and how much contributes to the pressure driving blood toward the heart. Vascular tone consequently regulates circulation through capacity as well as resistance.

Stressed and unstressed volume describe a mechanical partition, not two permanently separated pools of blood. Unstressed volume fills a vascular compartment without producing the distending pressure represented in the model; stressed volume contributes to that pressure. Changes in tone, compliance, regional distribution, and surrounding pressure can change the partition at the same total blood volume.

When venous capacity increases, less of the existing volume may support the driving pressure for venous return. This creates relative underfilling even without hemorrhage or another external loss. A patient can therefore have substantial total-body fluid and edema while the effective circulatory distribution remains unfavorable.

Splanchnic veins are particularly relevant because of their capacity and responsiveness. Regional tone, local inflow, abdominal pressure, and cardiac function interact. Blood shifted toward the thorax must still pass through both sides of the heart and the lungs before becoming sustained systemic output.

Venoconstriction can redistribute blood toward pressure-generating compartments without adding fluid. Volume administration can also increase stressed volume, but how much it does so depends on the pressure–volume relationship and where the added volume distributes. Neither intervention is equivalent to a guaranteed increase in forward flow.

This distinction separates total amount from functional availability. A low filling pressure can reflect reduced blood volume, increased capacity, or altered surrounding pressures. A high filling pressure can coexist with impaired return or limited cardiac accommodation. Edema and a low effective filling state are therefore not contradictory observations.

Changing tone and capacity mean that a previous fluid response need not persist or predict the next response. The current state includes venous distribution, outflow resistance, and cardiac accommodation.

71.4 Venous Return and Cardiac Function Determine the Operating Flow Together

In a steady circulation, venous return equals cardiac output. The heart and systemic vessels settle at a common operating point; one does not supply an unlimited flow independently of the other. Raising a venous driving pressure can increase output only to the extent allowed by venous resistance, downstream pressure, and cardiac function.

Let Pms be an effective mean systemic filling pressure, RAP intracavitary right atrial pressure, and Rvr resistance to venous return. Pressures share a reference in mmHg and Rvr is in mmHg·min/L, giving flow Qvr in L/min in the simplified model.

For illustrative values of Pms = 12 mmHg, RAP = 7 mmHg, and Rvr = 1 mmHg·min/L, flow is (12 − 7) ÷ 1 = 5 L/min. In another state, Pms = 15 mmHg, RAP = 9 mmHg, and Rvr = 1.2 mmHg·min/L give (15 − 9) ÷ 1.2 = 5 L/min. Upstream and atrial pressures have risen without increased flow because the other conditions changed.

The model clarifies why venoconstriction has competing effects. It can raise filling pressure by changing capacity, while also increasing resistance along the return pathway. Persichini and colleagues demonstrated changes in both estimated mean systemic pressure and venous-return resistance during norepinephrine reduction in selected septic-shock patients. Their ventilatory-occlusion estimates support coupled vascular effects rather than a fixed response to every change in tone.1

Pms is a model-derived effective pressure, not a routinely sampled point in a single vein. Its estimates rely on assumptions about flow, vascular behavior, and the method used. Collapsible vessels and changing intrathoracic or abdominal pressure can introduce flow limitation that the simple linear equation does not capture.

RAP also has different roles in different questions. Intracavitary pressure is relevant to the downstream pressure for venous return, whereas myocardial distension depends on pressure relative to the local external environment. Positive-pressure ventilation can alter both. A measured CVP change is therefore not an automatic measure of increased preload or a universal correction for intrathoracic pressure.

71.5 Permeability and Edema Change the Meaning of Added Volume

Distributive shock often includes changes in endothelial exchange as well as vascular tone. Increased permeability allows more fluid and macromolecules to cross the microvascular barrier, while hydrostatic forces and lymphatic drainage determine the net accumulation. The degree of leakage varies between diseases, organs, and phases of illness.

Loss from the vascular space can contribute to absolute circulatory volume depletion even as total-body fluid increases. Meanwhile, venous dilation can cause relative underfilling. These mechanisms can coexist, so neither the presence of edema nor an assumption of venous pooling establishes the current blood volume.

Added fluid redistributes over time. An initial increase in filling and flow can diminish as fluid moves between compartments or vascular conditions change. A transient pressure or stroke-volume response therefore does not quantify how long an intervention will improve tissue perfusion or how much fluid will remain intravascular.

Edema has physiological costs beyond its visible amount. Increased diffusion distance, higher tissue pressure, impaired organ compliance, and compression of small vessels can interfere with exchange and drainage. Rising venous pressure can simultaneously increase filtration and reduce an organ's effective arterial-to-venous gradient. The consequence depends on the organ and surrounding compartment.

The relation between global and microvascular effects is also variable. In a nonrandomized study of severe sepsis, Ospina-Tascón and colleagues found that the sublingual microvascular response to fluid differed between their early and later illness groups and was not explained by the accompanying global hemodynamic changes. The study demonstrates context dependence; its enrollment windows are not a universal clock for fluid benefit.2

Responsiveness and tolerance must consequently be separated. A circulation may increase output when filling rises while the lungs or venous-congested organs experience an unfavorable cost. Conversely, a pressure change without increased flow may mainly reflect altered vascular or chamber mechanics.

Additional effective filling must be connected to a relevant flow or tissue deficit and its physiological cost. Fluid balance, edema, pressure, or a previous response alone cannot supply that account.

71.6 Cardiac Reserve Determines Whether Vasodilation Produces High Output

Low arterial load can permit a large stroke volume when filling, rate, and myocardial reserve are sufficient. It can also conceal impaired contractile function by making ejection easier. Consequently, high output or a preserved ejection fraction does not establish normal myocardium, while a low ejection fraction does not independently quantify the forward output being delivered.

Sepsis can disturb systolic contraction, relaxation, right-ventricular function, and the response to loading. Established heart disease, ischemia, arrhythmia, medications, and ventilation add further influences. These mechanisms may coexist with marked vasodilation rather than replacing the distributive diagnosis.

Early human serial observations by Parker and colleagues demonstrated reduced ejection fraction and increased ventricular volumes in some patients with septic shock despite normal or elevated cardiac index. Changes were reversible in selected survivors. This small observational study established important pump involvement without making reduced ejection fraction a universal phenotype or a favorable prognostic sign.3

Changing arterial tone can expose reserve limitations. If afterload rises, a ventricle may eject less unless greater filling, contractility, or coronary perfusion compensates. Conversely, restoring very low arterial pressure can improve coronary driving conditions and thereby support pump performance. These opposing effects help explain why the same pressure response can accompany increased, unchanged, or decreased output.

The right heart is part of this coupling. Hypoxemia, acidosis, lung injury, pulmonary vascular changes, and positive-pressure ventilation can raise right-ventricular load. Greater systemic venous filling may then increase right-sided pressure more than useful pulmonary flow. Interdependence can impair left-ventricular filling as well.

Tachycardia also has competing effects: it can maintain output when stroke volume falls but increases myocardial work and shortens filling and diastolic coronary-perfusion time. The ability to sustain a hyperdynamic state therefore depends on reserve rather than on low resistance alone.

Detailed pump-failure and obstructive mechanisms follow in Chapters 72 and 73. Here, vascular conditions must be connected to biventricular capacity; a vascular initiating cause does not imply an unaffected heart.

71.7 Regional Flow Can Remain Abnormal Despite Improved Global Measurements

Systemic output is divided among organs and then among networks within each organ. Distributive shock can disturb both levels. Some regions receive relatively abundant flow while others have reduced perfused capillary density, intermittent flow, or prolonged diffusion distances. A global average cannot reveal how oxygen-carrying red cells are distributed beside individual cells.

Flow heterogeneity also alters exchange. Blood passing rapidly through well-perfused channels may leave with substantial oxygen while adjacent poorly perfused tissue remains under-supplied. This need not represent a single anatomical shunt. It can emerge from unequal transit, functional capillary loss, local resistance, and altered diffusion.

De Backer and colleagues directly visualized the sublingual microcirculation in patients with severe sepsis and comparison groups. Reduced small-vessel perfusion and greater abnormalities in nonsurvivors demonstrated that clinically important heterogeneity was present despite what global variables alone could show. The association did not establish that correcting the observed images would improve survival.4

Pressure can still matter when organ flow is pressure dependent. However, increasing systemic pressure does not guarantee recruitment of the threatened networks. A small prospective norepinephrine study by Dubin and colleagues found no overall improvement in sublingual microcirculation as pressure increased, with considerable variation between patients. The result qualifies the pressure-to-perfusion inference; it does not show that pressure support never helps.5

Sublingual measurements are themselves regional observations. They do not directly report renal medullary, intestinal, or myocardial flow, and image acquisition and analysis have limitations. Their value is to expose the incompleteness of systemic measurements, not to substitute one sampling site for the whole organism.

Venous and tissue pressures further shape local gradients. An apparently adequate arterial pressure can coexist with impaired drainage or high compartment pressure. Autoregulatory capacity, metabolic vasodilation, endothelial signaling, and local obstruction determine whether the available pressure becomes useful flow.

Global measurements may recover before, alongside, or without regional recovery. The changing relationship between systemic circulation and tissue observations is itself physiological information.

71.8 Oxygen Content and Demand Determine What a Given Output Can Supply

Cardiac output transports blood, while arterial oxygen content determines how much oxygen that blood carries. Anemia, incomplete saturation, and changing pulmonary gas exchange can reduce content even when flow is high. Metabolic demand simultaneously changes with fever, shivering, respiratory effort, adrenergic activity, seizures, and the underlying disease.

Let ḊO₂ be systemic oxygen delivery in mL O₂/min, CO output in L/min, and CaO₂ arterial content in mL O₂/dL. The factor 10 converts liters into deciliters. Let V̇O₂ be whole-body consumption in mL O₂/min and E the dimensionless global extraction fraction under steady-state conditions.

With an invented CO of 8 L/min and CaO₂ of 12 mL/dL, delivery is 8 × 12 × 10 = 960 mL/min. If consumption is 300 mL/min, extraction is 300 ÷ 960 = 0.3125, or about 31%. If output is instead 6 L/min with content and demand unchanged, delivery is 720 mL/min and required extraction becomes 300 ÷ 720 ≈ 42%.

The second state requires greater extraction if consumption is maintained. These values are illustrative, not critical thresholds. Regional maldistribution or impaired diffusion can prevent the necessary adjustment even when the global calculation appears possible. Conversely, a lower measured consumption may reflect reduced demand or impaired tissue use rather than successful restoration of balance.

Mixed venous oxygen measurements reflect this combined system. A high venous saturation can accompany abundant supply, reduced demand, unequal perfusion, or impaired extraction. It does not independently exclude regional hypoxia. Central venous and true mixed venous samples are also different observations.

Lactate adds evidence about production, disposal, and distribution. Adrenergic metabolism can increase lactate, as human epinephrine experiments demonstrate. Those healthy-volunteer observations do not quantify the cause in a particular septic patient.6 Impaired hepatic or renal disposal and regional hypoperfusion remain possible simultaneous contributors.

71.9 Tests of Responsiveness Do Not Independently Establish Need or Tolerance

Bedside measurements sample different aspects of the circulation. Arterial pressure describes a driving condition; measured flow describes transport; echocardiography shows loading and pump behavior; skin findings describe a peripheral response. Urine output, lactate, and mental status reflect organ function through additional mechanisms and different time courses.

No single observation can answer every question. A low CVP does not quantify the blood volume missing from the system, while a high CVP can reflect external pressure, right-heart limitation, or congestion. Ejection fraction is load dependent. Oliguria can persist because of established kidney injury even when global flow improves.

Dynamic tests ask a more specific question: whether a temporary change in effective filling produces a measurable flow response in the current state. Passive leg raising can redistribute venous blood toward the central circulation without permanently administering fluid. Its interpretation requires a valid maneuver, a sufficiently responsive flow measurement, and attention to the conditions limiting the induced volume shift.

Monnet and colleagues compared the flow response to passive leg raising with the response to subsequent volume expansion in critically ill patients. Their findings established predictive physiological information, including in selected patients with spontaneous respiratory activity or arrhythmia. The experiment did not show that every predicted responder needs fluid or gains a patient-centered benefit from receiving it.7

Responsiveness means that stroke volume or output can rise with increased effective preload. Need concerns whether an important deficit would improve. Tolerance concerns pulmonary, venous, tissue, and cardiac consequences. These questions overlap but remain distinct. A patient can be responsive while already vulnerable to worsening edema or congestion.

Respiratory variations of pressure or flow likewise depend on the mechanics generating them. Ventilator conditions, spontaneous effort, rhythm, right-ventricular loading, and abdominal pressure can affect validity. A change in arterial pulse pressure is not automatically the same as a measured change in output.

Serial assessment links a proposed mechanism to the actual response. Reproducible improvement in one variable requires interpretation alongside the other consequences, rather than establishing complete circulatory recovery.

71.10 Vasoactive Effects Extend Beyond Raising Arteriolar Resistance

Vasoactive agents interact with arterial resistance, venous capacity, cardiac receptors, regional perfusion, and metabolism. Their effect on pressure is the visible result of this combined response. The same pressure increase can arise with different changes in stroke volume, output, and tissue perfusion.

Norepinephrine illustrates these interactions. Arterial constriction can increase perfusion pressure while increasing ventricular ejection load. Venous effects can increase stressed volume and the driving pressure for return, but can also increase resistance to return. Cardiac stimulation and improved coronary driving conditions may support pump performance. The balance depends on the starting vascular and cardiac state.

In a human observational study of severe septic hypotension, Hamzaoui and colleagues found increased cardiac output after early norepinephrine introduction or escalation, accompanied by changes in loading and cardiac-function indices. The study supports the possibility of a favorable flow response in that setting, not a guaranteed response or proof that its individual mechanisms caused an outcome benefit.8

Different receptor pathways can produce different combinations of effects, while acid–base state, inflammation, prior exposure, and tissue responsiveness modify them. The underlying molecular changes and endocrine contributions were developed in Chapters 65 and 66. A high administered dose is therefore not a pure measure of endogenous vascular tone: it also reflects the chosen pressure goal, distribution, clearance, receptor response, and concurrent physiology.

There are costs to consider in the same causal account. Greater vasoconstriction can improve a pressure-dependent bed while reducing flow elsewhere. Increased cardiac stimulation can raise demand or provoke arrhythmia. Excessive loading can expose ventricular limitations. A rising lactate during adrenergic treatment can have metabolic as well as perfusion-related explanations.

These possibilities do not justify assuming that vasopressors necessarily reduce tissue flow or that pressure restoration necessarily improves it. They require observing what happened to output, regional findings, demand, and organ function in that patient.

A vasoactive response is therefore a physiological perturbation that reveals reserve and coupling. Its clinical value depends on the net consequences and on evidence for the strategy used, rather than on the achievement of one arterial pressure alone.

71.11 Strategy Trials Preserve the Difference Between Physiological Response and Outcome

The competing effects of fluids, vasoactive support, and pressure goals require outcome trials. These evaluate strategies, including monitoring, timing, and cointerventions. Their results neither isolate every physiological component nor establish one universal protocol.

ARISE FLUIDS, published in 2026, compared restricted fluid volumes with earlier vasopressors against greater fluid volumes with later vasopressors in adults presenting to emergency departments with septic shock. Among 1,000 randomized patients, 963 entered the intention-to-treat population and 960 had primary-outcome follow-up. The earlier-vasopressor strategy did not improve days alive and out of hospital through day 90. Pulmonary edema was less frequent with that strategy, a distinct finding that should not be converted into a demonstrated survival benefit.9

ANDROMEDA-SHOCK-2 tested a personalized resuscitation strategy organized around capillary refill time and assessments of pressure patterns, fluid responsiveness, and cardiac function. Its hierarchical composite of mortality, duration of vital support, and hospital stay favored the intervention, primarily through shorter vital-support duration. The result supports the tested integrated strategy, not an isolated effect of measuring capillary refill or a demonstrated mortality reduction.10

OPTPRESS examined a higher-pressure strategy in older Japanese adults with septic shock. The trial stopped early, and 90-day mortality was higher in the higher-target group. This challenges the assumption that more pressure is automatically better, including when chronic hypertension is present. Its population and strategy still matter; the result does not establish that every increase in pressure is harmful in every patient.11

These studies answer related but different questions. A neutral primary outcome does not make all physiological approaches equivalent, and a favorable composite does not establish benefit for every component. Likewise, a harmful strategy does not identify the sole biological mechanism of harm. The disciplined interpretation connects the actual intervention and endpoint to the clinical setting while retaining uncertainty about individual response.

71.12 Integrating the State Requires Following Both Cause and Consequences

Return to the constructed pressure example: MAP rises from approximately 55 to 65 mmHg while output falls from 8 to 6 L/min. The pressure response is real, but its meaning remains unresolved until oxygen content, demand, regional perfusion, and organ responses are considered. The second state could improve a pressure-dependent organ while reducing transport elsewhere.

If arterial content is the 12 mL/dL used earlier, delivery falls from 960 to 720 mL/min. Whether that is adequate depends on demand and distribution rather than on the pressure increase alone. A contemporaneous rise in venous pressure could further change local gradients, while improved coronary perfusion or reduced excessive flow might have other consequences. The constructed states are alternatives for interpretation, not a treatment prediction.

The same approach applies to filling. A higher estimated mean systemic pressure or CVP without increased flow may reflect venous resistance, cardiac limitation, or altered external pressures. A positive preload response may identify remaining flow reserve without establishing durable tissue benefit or tolerance of further volume.

Serial observations help distinguish these possibilities. The pattern of pressure, measured output, oxygen content, peripheral findings, lactate trajectory, urine output, congestion, and echocardiographic changes can strengthen or weaken a causal explanation. Different signals recover on different time scales, so temporary disagreement should prompt interpretation rather than automatic escalation toward a single number.

The underlying cause remains central. Vascular support does not remove infection, stop mediator release in anaphylaxis, or restore neural control. Each cause has distinct disease-directed requirements. Hemorrhage, pump disease, obstruction, and treatment effects can remain active after one component improves.

Distributive shock is therefore best understood as a changing interaction of vascular regulation, effective filling, cardiac reserve, regional exchange, and cellular demand. This framework explains why a pressure can be restored before tissue function recovers, why a high output can remain inadequate, and why a useful intervention must be judged through its full consequences.

References

1. Persichini R, Silva S, Teboul JL, et al. Effects of norepinephrine on mean systemic pressure and venous return in human septic shock. Crit Care Med. 2012;40:3146–3153. Original human physiological study.

2. Ospina-Tascón G, Neves AP, Occhipinti G, et al. Effects of fluids on microvascular perfusion in patients with severe sepsis. Intensive Care Med. 2010;36:949–955. Original nonrandomized physiological study.

3. Parker MM, Shelhamer JH, Bacharach SL, et al. Profound but reversible myocardial depression in patients with septic shock. Ann Intern Med. 1984;100:483–490. Original serial human study.

4. De Backer D, Creteur J, Preiser JC, et al. Microvascular blood flow is altered in patients with sepsis. Am J Respir Crit Care Med. 2002;166:98–104. Original human microvascular study.

5. Dubin A, Pozo MO, Casabella CA, et al. Increasing arterial blood pressure with norepinephrine does not improve microcirculatory blood flow: a prospective study. Crit Care. 2009;13:R92. Original human physiological study.

6. Clutter WE, Bier DM, Shah SD, Cryer PE. Epinephrine plasma metabolic clearance rates and physiologic thresholds for metabolic and hemodynamic actions in man. J Clin Invest. 1980;66:94–101. Original human metabolic experiments.

7. Monnet X, Rienzo M, Osman D, et al. Passive leg raising predicts fluid responsiveness in the critically ill. Crit Care Med. 2006;34:1402–1407. Original preload-response study.

8. Hamzaoui O, Georger JF, Monnet X, et al. Early administration of norepinephrine increases cardiac preload and cardiac output in septic patients with life-threatening hypotension. Crit Care. 2010;14:R142. Original observational human study.

9. ARISE FLUIDS Investigators, ANZICS Clinical Trials Group, ACEM Clinical Trials Network. Vasopressors or fluids in early septic shock. N Engl J Med. 2026;395:1387–1398. Original randomized strategy trial.

10. Hernández G, Ospina-Tascón GA, Kattan E, et al. Personalized hemodynamic resuscitation targeting capillary refill time in early septic shock: the ANDROMEDA-SHOCK-2 randomized clinical trial. JAMA. 2025;334:1988–1999. Original randomized strategy trial.

11. Endo A, Yamakawa K, Tagami T, et al. Efficacy of targeting high mean arterial pressure for older patients with septic shock (OPTPRESS): a multicentre, pragmatic, open-label, randomised controlled trial. Intensive Care Med. 2025;51:883–892. Original randomized strategy trial.