Autonomic Regulation and the Stress Response

Autonomic regulation links sensed threats to changes in circulation, ventilation, metabolism, and visceral function. These responses can preserve perfusion during an acute disturbance, yet their persistence or loss of appropriate coordination can become harmful. Understanding the response requires distinguishing afferent information, central integration, neural output, circulating mediators, and the tissues' ability to respond.

64.1 Autonomic Output Is an Integrated Pattern, Not a Single Stress Signal

The autonomic nervous system continuously adjusts organ function rather than remaining inactive until an emergency occurs. Cardiovascular, respiratory, visceral, thermal, and nociceptive information enters circuits that also receive influences from arousal, emotion, cognition, and anticipated activity. The resulting output depends on the combination of inputs and the organism's current state.

Brainstem networks, including the nucleus tractus solitarius and interconnected medullary regions, integrate important cardiovascular and respiratory afferents. Hypothalamic, limbic, and cortical influences modify that processing. A painful procedure, hypoxemia, falling arterial pressure, and anxiety can therefore interact rather than generating independent responses that can simply be added together.

Sympathetic and parasympathetic divisions describe major efferent organizations, but their activities are not constrained to move as opposite ends of one scale. Withdrawal of cardiac vagal influence can accelerate the heart without equivalent sympathetic activation in every vascular bed. Sympathetic activity can also increase in one region while remaining relatively unchanged elsewhere.

Victor and colleagues recorded muscle sympathetic nerve activity during arm exercise in conscious humans. Some exercise conditions increased heart rate and pressure without increasing the recorded leg muscle sympathetic activity; other conditions produced a pronounced neural response. The findings oppose the idea that a challenge necessarily produces a uniform whole-body sympathetic discharge. 1

The distinction between central command and feedback is useful. Anticipation and motor drive can initiate cardiovascular adjustments before a major metabolic deficit develops. Muscle and visceral afferents then report the consequences of activity or disease, modifying the pattern. In critical illness, pain, mechanical loading, inflammation, and disordered gas exchange provide continuing inputs even when the patient is immobile.

The endocrine stress response operates alongside these neural mechanisms. Adrenal and pituitary pathways can extend their effects in time and space, but hormone concentrations do not reproduce the detailed pattern of neural traffic. Their regulation is developed in Chapter 65.

64.2 Baroreceptor Afferents Report Mechanical Conditions

Arterial baroreceptors are mechanosensitive sensory endings, particularly in the carotid sinus and aortic arch. Carotid afferents travel predominantly through the glossopharyngeal pathway, while aortic afferents travel through the vagus. Their central projections help regulate sympathetic and cardiac parasympathetic output.

The immediate stimulus is deformation of the receptor-bearing vessel wall. Intravascular pressure matters through its relationship to surrounding pressure, wall geometry, stiffness, and pulsation. The same measured arterial pressure need not produce the same receptor stretch in a stiff artery, a differently pressurized surrounding compartment, or a different pulse pattern.

An acute increase in receptor loading generally increases afferent traffic, promoting cardiac slowing and reduced sympathetic vasomotor output. Unloading generally favors vagal withdrawal and sympathetic activation. These responses oppose the original pressure disturbance and form a rapid negative-feedback system.

Cardiopulmonary mechanoreceptors add information about central filling and mechanical conditions within the heart and pulmonary circulation. Their signals are not direct measurements of total blood volume. Chamber compliance, intrathoracic pressure, distribution of blood, and cardiac function influence the relationship between measured pressure, stretch, and effective filling.

Ogoh and colleagues used a neck-pressure apparatus to alter carotid sinus loading in eight healthy subjects at rest and during exercise. Changes in vascular conductance made a major contribution to the pressure response. This demonstrates why the cardiac response alone does not describe the entire arterial baroreflex. The experimental pressure pulses are a research method, not a bedside diagnostic maneuver. 2

The reflex also operates through a mechanical circulation that can constrain its success. Increasing sympathetic output cannot restore adequate flow if venous return is severely restricted or ventricular outflow is obstructed. Increasing heart rate may not increase cardiac output when filling or contractile reserve is limited.

An apparently appropriate neural response can consequently coexist with persistent hypotension. The observed arterial pressure reflects both regulation and the cardiovascular system being regulated.

64.3 Reflex Gain, Operating Point, and Resetting Are Different Properties

A reflex response curve has a range, an operating point, and a slope. Gain describes how much the measured output changes for a change in input near the conditions tested. It is not the same as the prevailing level of sympathetic or parasympathetic activity.

For the cardiac baroreflex, one possible estimate is the change in R–R interval divided by the accompanying change in systolic arterial pressure:

RR is the interval between successive normal ventricular beats in milliseconds, and SBP is systolic arterial pressure in mmHg. The resulting sensitivity is expressed in ms/mmHg. This describes a selected cardiac input–output relationship, not the gain of all autonomic control pathways.

For a constructed example, suppose systolic pressure increases from 100 to 120 mmHg while the appropriately paired R–R interval increases from 700 to 780 ms. The interval change is 80 ms and the pressure change 20 mmHg, giving 80 ÷ 20 = 4 ms/mmHg. The longer interval represents cardiac slowing under those assumed conditions.

That number is not a universal normal value or a treatment target. Direction of pressure change, timing, respiration, baseline heart rate, medications, and the method used to pair pressure with intervals influence the estimate. Spontaneous sequences and controlled perturbations do not necessarily measure identical properties.

Resetting means that the reflex operates around a different pressure or activity state. During a physiological challenge, pressure can remain higher while the reflex continues to buffer fluctuations around its new operating point. A higher prevailing pressure therefore does not prove that feedback has stopped functioning.

Conversely, a reduced cardiac slope does not prove absent vascular control. As illustrated by the exercise measurements in Ogoh's study, the relative cardiac and vascular contributions can change. A response can also appear small near a saturated part of its curve despite useful regulation over another range. 2

Interpretation requires asking which branch was measured, where on its response curve it was operating, and whether the system was sufficiently stable for a slope to be meaningful.

64.4 Chemoreflexes Interact with Ventilation and Circulation

Peripheral arterial chemoreceptors respond to the chemical environment of arterial blood, including low oxygen partial pressure and changes in carbon dioxide and hydrogen ion activity. Central chemosensitive networks respond importantly to carbon dioxide-related changes in the brain's chemical environment. These inputs interact with the respiratory and cardiovascular control systems.

Oxygen content and oxygen partial pressure are different signals. Anemia can substantially lower oxygen content while arterial oxygen partial pressure remains relatively preserved. A peripheral chemoreceptor response to hypoxemia therefore cannot be treated as a direct meter of whole-body oxygen delivery.

Chemoreflex activation can increase ventilatory drive and sympathetic vasomotor activity, but the resulting heart-rate and pressure responses depend on concurrent breathing. Lung inflation, respiratory timing, changes in thoracic pressure, and cardiac vagal influences modify the integrated response. Apnea and active ventilation are consequently different physiological conditions even at similar blood-gas values.

Somers and colleagues recorded peroneal sympathetic activity in 13 healthy subjects during hypoxic and hypercapnic challenges. Their results showed interactions among gas stimulus, ventilation, and apnea, including a greater response to combined hypoxia and hypercapnia than a simple additive interpretation predicted. These controlled human observations do not provide a fixed neural-response coefficient for critically ill patients. 3

In intensive care, gas abnormalities rarely occur in isolation. Hypoxemia can coexist with hypotension, acidosis, pain, fever, sedation, and altered respiratory mechanics. Changes in any of these can modify the response to the others. A tachycardic, vasoconstricted patient may therefore be expressing several overlapping reflex demands.

Ventilatory assistance can alter more than gas exchange: it changes effort, lung inflation, and intrathoracic pressure. Sedation can change afferent perception and central integration while also affecting vascular tone directly. The before-and-after response cannot be assigned automatically to one neural pathway.

The chemoreflex is best understood as part of a coupled respiratory–circulatory network. The stimulus, the breathing pattern, and the cardiovascular capacity together determine the observed state.

64.5 Efferent Signals Must Be Translated by Receptors and Cells

Autonomic preganglionic neurons use acetylcholine at nicotinic receptors within autonomic ganglia. Most sympathetic postganglionic neurons release norepinephrine at their target tissues, whereas many parasympathetic postganglionic neurons release acetylcholine acting through muscarinic receptors. Important exceptions, including sympathetic cholinergic control of sweating, prevent transmitter identity from being a complete division label.

Neural norepinephrine acts near release sites before much of it reaches plasma. Reuptake, metabolism, diffusion, and presynaptic regulation influence the local exposure. Adrenal catecholamines provide circulating signals with a different distribution and time course. Infused catecholamines add another source that cannot be separated from endogenous secretion by measuring the total concentration alone.

Receptor subtype and intracellular coupling determine the response. Cardiac β₁ receptors generally couple through stimulatory G proteins to adenylyl cyclase and cyclic AMP, altering pacemaker activity, calcium handling, contraction, and relaxation. β₂ signaling can relax smooth muscle in some beds and influence metabolism. A shared second messenger therefore does not imply a shared physiological effect across tissues.

α₁ receptors commonly promote smooth-muscle contraction through pathways involving phospholipase C and intracellular calcium. α₂ receptors include presynaptic sites that limit further transmitter release, alongside central and peripheral effects. Muscarinic M₂ signaling slows nodal activity through inhibitory signaling and potassium-channel effects. These descriptions establish mechanisms rather than mutually exclusive actions of a particular circulating concentration.

The response also depends on receptor abundance, coupling, cellular energetics, and competing local influences. Strong neural traffic can coexist with a weak organ response when receptor signaling or the tissue itself is impaired. Conversely, increased sensitivity can amplify modest input.

Parasympathetic activity is not a universal direct vasodilator system opposing sympathetic constriction in every systemic vessel. Cardiac effects are prominent, while vascular innervation and endothelial responses vary by region. Inferring whole-body vasodilation from a presumed increase in cardiac vagal influence therefore loses important anatomy.

The causal chain runs from neural signal through transmitter exposure and receptor response to organ mechanics. Observing its endpoint does not identify which earlier step changed.

64.6 Cardiac Acceleration Does Not Guarantee More Flow

Cardiac sympathetic effects can accelerate pacemaker activity, enhance conduction, increase contractile performance, and hasten relaxation. Cardiac parasympathetic effects prominently slow nodal activity and influence conduction. Their balance is filtered through intrinsic pacemaker behavior, conduction disease, loading conditions, and medications.

The immediate flow relationship is:

Heart rate is in beats/min and stroke volume in mL/beat, giving cardiac output in mL/min; dividing by 1000 converts the result to L/min. The equation is an accounting identity, not proof that changing heart rate leaves stroke volume unchanged.

Suppose an illustrative patient has a heart rate of 110 beats/min and stroke volume of 55 mL/beat. Output is 110 × 55 = 6050 mL/min, or 6.05 L/min. If rate later increases to 140 while stroke volume falls to 40 mL/beat, output becomes 5600 mL/min, or 5.60 L/min. The faster heart is delivering less blood per minute.

The example does not establish that tachycardia caused the stroke-volume reduction. Hypovolemia, ventricular dysfunction, afterload, or altered thoracic pressure could affect both. It shows why heart rate alone cannot establish the adequacy of the cardiac response.

Faster rate shortens cycle duration, particularly the available diastolic interval. Depending on relaxation, coronary conditions, and loading, this can constrain filling and myocardial oxygen supply while increasing oxygen demand. Enhanced relaxation and contractility may compensate in some circumstances, so the effect is conditional rather than uniformly harmful.

Tachycardia can be an effective adaptation to a low stroke volume, a marker of continuing afferent stress, or part of a process worsening efficiency. Bradycardia likewise can reflect several mechanisms, including vagal influence, nodal disease, drugs, or metabolic disturbance. Neither rate pattern diagnoses autonomic state by itself.

The meaningful assessment joins rate to rhythm, stroke volume, pressure, perfusion, and the stimulus driving the response. A numerically more familiar rate is not automatically a better circulatory state.

64.7 Vascular Resistance and Venous Capacitance Are Distinct Effectors

Sympathetic vascular control influences both resistance vessels and capacitance vessels. Arteriolar constriction changes the pressure–flow relationship into tissue beds. Venous constriction changes the volume accommodated at a given pressure and can redistribute blood toward a stressed volume that supports venous return.

This does not mean that vasoconstriction creates blood volume. It changes how existing volume is distributed and how much contributes to vascular distending pressure. Its consequences also depend on cardiac function and downstream pressure.

A simplified venous-return relation helps organize those effects:

Mean systemic filling pressure, Pmsf, and right atrial pressure, PRA, are in mmHg. Resistance to venous return, RVR, is in mmHg/(L/min), giving venous return in L/min. The relationship assumes an interpretable effective flow gradient and does not replace the limits imposed by vascular collapse or dynamic pressure transmission.

Choose constructed values of Pmsf 14 mmHg, PRA 6 mmHg, and RVR 2 mmHg/(L/min). Flow is approximately (14 − 6) ÷ 2 = 4 L/min. If vascular recruitment and cardiac interaction produce values of 18 and 8 mmHg with the same resistance, flow becomes 5 L/min. If resistance instead increases to 2.5, flow remains 4 L/min despite the higher filling pressures.

These are explanatory parameters, not a normal range, a fluid estimate, or a vasopressor target. They show why the net result depends on both the driving gradient and resistance.

Persichini and colleagues studied norepinephrine changes in 16 patients with septic shock using estimated venous-return relationships. Both mean systemic pressure and resistance changed. The findings support considering these simultaneous effects, while the estimation method and drug intervention prevent a direct numerical equivalence with endogenous nerve activity. 4

Arterial pressure can therefore rise through several combinations of flow and resistance. A favorable pressure response does not by itself prove improved regional perfusion, recruited preload, or increased total blood volume.

64.8 Regional, Metabolic, and Immune Effects Extend Beyond Blood Pressure

Autonomic responses are regionally patterned, and local tissue conditions modify their effects. Metabolic vasodilation can oppose adrenergic constriction in active tissue, while skin circulation also responds to thermal demands. The same central challenge can consequently produce different flow responses across organs.

Anderson and colleagues measured arm and leg muscle sympathetic activity during mental stress in healthy men. Simultaneous recordings demonstrated different regional responses, reinforcing that an accessible nerve signal cannot be assigned to every vascular bed. The experiment concerns controlled mental stress, not a quantitative map of sympathetic output during septic shock. 5

Stress signaling also changes substrate availability and cellular work. Adrenergic influences can support glycogen breakdown, lipolysis, and accelerated glycolytic flux, while interacting with insulin and other hormones. Increased lactate production can therefore accompany adequate oxygen availability in some tissues, although hypoperfusion and impaired disposal remain important possibilities.

James and colleagues studied isolated rat skeletal muscle in oxygenated medium. Epinephrine-stimulated lactate production was linked to sodium–potassium pump activity, with inhibition experiments supporting that relationship. This experimental mechanism does not quantify its contribution in human septic shock or identify a treatment for hyperlactatemia. 6

Neural signaling also interacts with inflammatory responses. Borovikova and colleagues found that vagal stimulation altered cytokine responses in rat endotoxemia, alongside experiments in human macrophage cultures. This provides mechanistic evidence for neuroimmune communication, not proof that bedside cardiac vagal indices quantify the pathway or that stimulation improves outcomes in general intensive care populations. 7

These wider effects explain why an autonomic intervention may change more than the intended pressure or rate. They also caution against attributing a metabolic or inflammatory observation to one branch of the nervous system. Local flow, receptor expression, hormonal context, and disease alter the result.

Detailed endocrine regulation and inflammatory dysfunction are developed in Chapters 65 and 66. Here, the essential principle is that autonomic output coordinates multiple organs and cellular processes with different response thresholds and time courses.

64.9 Persistent Activation Can Coexist with Reduced Responsiveness

During acute volume loss or another immediate threat, autonomic activation can preserve pressure and redistribute limited circulatory resources. The same pattern sustained over time may increase cardiac work, impair regional flow, disrupt sleep and recovery, and interact with inflammatory and metabolic stress. Benefit depends on the demand being met and the physiological cost of meeting it.

Persistent catecholamine exposure can alter receptor regulation and intracellular coupling. Receptor phosphorylation, internalization, changes in signaling proteins, and downstream metabolic constraints can reduce responsiveness. These processes need not affect all receptor subtypes or organs equally.

Silverman and colleagues studied patients with sepsis and septic shock, measuring catecholamine responses and β-adrenergic signaling in circulating lymphocytes. Reduced cyclic AMP responses were associated with impaired cardiac chronotropic responsiveness in septic shock. Because the cellular assay used lymphocytes, it was an indirect indicator rather than a direct measurement of every myocardial signaling defect. 8

A high circulating catecholamine concentration with a weak cardiovascular response therefore does not prove absent neural drive. Conversely, a strong response to an agonist does not establish that endogenous signaling was previously inadequate. Delivery, clearance, receptor coupling, loading conditions, and tissue capacity can each explain part of the observation.

Critical illness can also impair afferent signaling or central coordination. Neuropathy, brain or spinal injury, sleep disruption, metabolic derangement, and sedative or vasoactive drugs modify the control system. The resulting state may combine excessive output in one pathway with inadequate responsiveness or loss of modulation in another.

It is therefore misleading to describe every persistent tachycardia as excessive sympathetic drive or every reduced variability as autonomic exhaustion. Those labels compress different mechanisms into an interpretation that the observation alone cannot support.

The distinction between necessary compensation and costly persistence requires examining the original threat, present flow and perfusion, organ reserve, and changes over time. An intervention that reduces one signal may remove compensation, reduce harmful exposure, or do both in different organs.

64.10 Measurements Sample Different Parts of the Control System

Plasma norepinephrine reflects entry into plasma relative to removal. Much neural transmitter is taken back up near release sites, so plasma appearance is only a fraction of total release. Sampling location, blood flow, metabolism, renal and hepatic function, and infused drugs further affect the measured concentration.

In a simplified steady state without exogenous input:

Concentration can be expressed in µg/L, spillover in µg/min, and clearance in L/min. Spillover means the rate reaching the sampled plasma compartment, not the total amount released at all sympathetic synapses.

For illustration, spillover of 1.2 µg/min and clearance of 3 L/min produce 0.4 µg/L. If clearance falls to 1.5 L/min with spillover unchanged, concentration doubles to 0.8 µg/L. The concentration change does not require increased neural activity. These assumed values are not diagnostic reference limits.

In 12 human subjects undergoing cognitive challenge, Esler and colleagues found that cardiac spillover and plasma concentrations from different sites changed differently. A peripheral venous sample could underrepresent selective cardiac activation. Tracer kinetics improve mechanistic separation but retain their own modeling assumptions. 9

Microneurography records activity in an accessible sympathetic nerve fascicle. It provides direct regional neural information, but muscle and skin recordings differ, and neither automatically describes renal, cardiac, or splanchnic output. Burst frequency and burst incidence also express different quantities and depend on heart rate and analysis.

Heart-rate variability describes fluctuations in successive normal-beat intervals. Respiratory modulation, baroreflexes, baseline rate, sinus-node responsiveness, and recording conditions shape its components. Arrhythmia, pacing, ectopy, drugs, mechanical ventilation, and nonstationary illness can invalidate simple interpretations.

Kingwell and colleagues compared spectral variability with cardiac norepinephrine spillover and muscle sympathetic activity across several human physiological and disease states. Low-frequency variability did not provide a direct quantitative measure of sympathetic firing. Neither its value nor a low-frequency/high-frequency ratio should be treated as a universal measure of sympathetic–parasympathetic balance or a proven treatment target. 10

64.11 A Physiological Rationale Does Not Establish an Outcome Benefit

Reducing adrenergic stimulation can plausibly decrease cardiac work and metabolic cost. It can also reduce contractility or remove a compensatory rate response. Increasing adrenergic stimulation can support pressure and flow while imposing other costs. The net clinical effect depends on patient selection, timing, the specific intervention, and the accompanying management strategy.

Morelli and colleagues' single-center, open-label phase 2 trial randomized 154 patients with septic shock and persistent tachycardia to esmolol or standard care. Heart-rate control was the primary endpoint; favorable mortality findings were secondary and required further investigation. The trial did not establish that every tachycardic patient benefits from reducing sympathetic effects. 11

In the later multicenter STRESS-L trial, 126 adults with established, vasopressor-dependent septic shock and tachycardia were randomized to landiolol or standard care. The trial was stopped early for lack of expected benefit and concern about possible harm. It did not show improvement in the primary organ-failure outcome. Different drugs, populations, and protocols prevent treating this as a simple repetition of the earlier study. 12

Landi-SEP subsequently showed improved achievement of a combined heart-rate-control and vasopressor-requirement endpoint with landiolol. Its analysis included 196 patients, and clinical secondary outcomes, including mortality, did not differ significantly. Achieving the physiological endpoint therefore did not establish improved patient-centered outcomes. 13

These studies illustrate a general problem: a measurable autonomic or cardiovascular change is not necessarily a validated therapeutic mediator. An intervention can change heart rate without correcting the initiating stressor or improving regional perfusion. A beneficial mechanism can coexist with an adverse effect elsewhere.

The appropriate inference is consequently specific to the studied strategy and population. The evidence does not yield a universal heart-rate target, a dose, or a mandate to suppress an assumed sympathetic state. Demonstrating benefit requires outcomes and tolerability in the relevant clinical setting, beyond showing that a receptor or physiological variable changed.

64.12 Interpreting the Response Through Its Cause and Consequences

Consider a constructed patient with infection, fever, discomfort, and reduced effective arterial filling. Heart rate is 130 beats/min, stroke volume 35 mL/beat, and arterial pressure is low. Cardiac output is 130 × 35 = 4550 mL/min, or 4.55 L/min. Several afferent inputs could contribute to the rate, while limited filling constrains flow.

The first explanatory task is to identify which stresses remain active: impaired gas exchange, pain, temperature, low circulating volume, vasodilation, ventricular dysfunction, or another process. The observed rate does not decide among them, and more than one can be present.

Suppose that after the clinical state changes, heart rate is 110 and stroke volume 50 mL/beat. Output is now 5500 mL/min, or 5.50 L/min. Rate has fallen while flow has increased. This pattern is compatible with reduced afferent stress and improved filling or performance, but it does not prove which intervention or autonomic mechanism produced the change.

If rate instead falls to 100 while stroke volume remains 35, output becomes 3.50 L/min. The lower rate now accompanies lower flow. Whether that state is tolerable depends on metabolic demand and perfusion, not on rate reduction alone. These are alternative illustrations, not sequential predictions or treatment targets.

A higher catecholamine concentration or lower variability would add information only in context. Exogenous administration, clearance, rhythm, respiration, and receptor responsiveness would need consideration before assigning the change to neural drive. Regional perfusion and organ function could still diverge from global pressure and output.

The useful before-and-after assessment therefore links stimulus, integrated response, flow, pressure, and physiological cost. Adaptation is effective when it meets the required demand with acceptable consequences; persistence becomes concerning when the original demand has changed or the response itself adds injury.

Autonomic physiology supplies that causal framework. It helps explain the response and its limits while leaving detailed endocrine, inflammatory, and therapeutic questions to their respective evidence and chapters.

References

1. Victor RG, Seals DR, Mark AL. Differential control of heart rate and sympathetic nerve activity during dynamic exercise: insight from intraneural recordings in humans. J Clin Invest. 1987;79:508–516. Original human physiology study.

2. Ogoh S, Fadel PJ, Nissen P, et al. Baroreflex-mediated changes in cardiac output and vascular conductance in response to alterations in carotid sinus pressure during exercise in humans. J Physiol. 2003;550:317–324. Original human physiology study.

3. Somers VK, Mark AL, Zavala DC, Abboud FM. Contrasting effects of hypoxia and hypercapnia on ventilation and sympathetic activity in humans. J Appl Physiol. 1989;67:2101–2106. Original human physiology study.

4. 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 clinical physiology study.

5. Anderson EA, Wallin BG, Mark AL. Dissociation of sympathetic nerve activity in arm and leg muscle during mental stress. Hypertension. 1987;9:III114–III119. Original human physiology study.

6. James JH, Wagner KR, King JK, et al. Stimulation of both aerobic glycolysis and Na+-K+-ATPase activity in skeletal muscle by epinephrine or amylin. Am J Physiol. 1999;277:E176–E186. Original experimental study.

7. Borovikova LV, Ivanova S, Zhang M, et al. Vagus nerve stimulation attenuates the systemic inflammatory response to endotoxin. Nature. 2000;405:458–462. Original experimental study.

8. Silverman HJ, Penaranda R, Orens JB, Lee NH. Impaired β-adrenergic receptor stimulation of cyclic adenosine monophosphate in human septic shock: association with myocardial hyporesponsiveness to catecholamines. Crit Care Med. 1993;21:31–39. Original clinical physiology study.

9. Esler M, Jennings G, Lambert G. Measurement of overall and cardiac norepinephrine release into plasma during cognitive challenge. Psychoneuroendocrinology. 1989;14:477–481. Original human physiology study.

10. Kingwell BA, Thompson JM, Kaye DM, McPherson GA, Jennings GL, Esler MD. Heart rate spectral analysis, cardiac norepinephrine spillover, and muscle sympathetic nerve activity during human sympathetic nervous activation and failure. Circulation. 1994;90:234–240. Original human physiology study.

11. Morelli A, Ertmer C, Westphal M, et al. Effect of heart rate control with esmolol on hemodynamic and clinical outcomes in patients with septic shock: a randomized clinical trial. JAMA. 2013;310:1683–1691. Original randomized trial.

12. Whitehouse T, Hossain A, Perkins GD, et al. Landiolol and organ failure in patients with septic shock: the STRESS-L randomized clinical trial. JAMA. 2023;330:1641–1652. Original randomized trial.

13. Rehberg S, Frank S, Černý V, et al. Landiolol for heart rate control in patients with septic shock and persistent tachycardia: a multicenter randomized clinical trial (Landi-SEP). Intensive Care Med. 2024;50:1622–1634. Original randomized trial.