Keywords
septic shock - norepinephrine - epinephrine - phenylephrine - dopamine - angiotensin
II - vasopressin - terlipressin - selepressin
Vasopressors are the most commonly ordered drugs in vasodilatory shock, mostly septic
shock, but also vasodilatory shock post-cardiovascular surgery, post-acute myocardial
infarction (post-AMI), post-general/abdominal surgery, posttrauma, pancreatitis and
other conditions causing a severe systemic inflammatory response, and postanesthetic
and other drug administration. While evidence-based medicine emphasizes large, pivotal
randomized controlled trials (RCTs), there is a place for consideration of both small
and large well-controlled studies that are useful to guide clinical practice of vasopressor
use. Vasopressors have been evaluated in numerous small underpowered studies but fortunately
also in high quality multicenter RCTs, mainly in septic shock. These latter RCTs are
the main sources of evidence for practice guidelines,[1] prior reviews,[2] and recommendations for clinical use, while the former help us understand nuances
and adverse effects not seen in the constrained environment of an RCT.
Responses to Hypotension and Common Characteristics of Vasopressor Hormones
Responses to Hypotension and Common Characteristics of Vasopressor Hormones
There are strategically positioned pressure and metabolic sensors that are the first
responders to the existential threat of hypotension and tissue hypoxia. Pressure sensors
in the carotid are stimulated by hypotension and trigger sympathetically mediated
increases in heart rate to directly increase cardiac output and blood pressure. An
interconnected endocrine system has widespread, redundant stores of already-synthesized
synergistically acting vasopressor hormones that are rapidly secreted in response
to hypotension ([Fig. 1]). Pressure sensors in the renal vasculature increase renin secretion that increases
angiotensin I synthesis and angiotensin I is secreted and converted to angiotensin
II by angiotensin-converting enzyme (ACE) in the lung vasculature. The adrenal medulla
is stimulated to secrete and synthesize epinephrine and norepinephrine, while the
posterior pituitary is stimulated to secrete vasopressin.
Fig. 1 Vasopressors bind to receptors on vascular smooth muscle to induce vasoconstriction.
Norepinephrine (NE) binds to α-1 adrenergic receptors, β-2 receptors causing vasodilation,
and α-1 and β-2 adrenergic receptors on leukocytes to modulate immune response in
sepsis. NE downregulates α-1 and β-2 receptor density changing sensitivity to NE,
thereby leading to increased doses of NE and greater risk of adverse vascular and
immune effects. Vasopressin (AVP) binds to the AVPR1a receptor, dopamine (DA) binds
to DA1 and DA2 receptors, and angiotensin II (AG) binds to angiotensin II receptors
(AGTR1 and AGTR2), all causing vasoconstriction.
Virtually all therapeutic vasopressors are natural hormones (or hormone derivatives)
that are critical downstream effectors of the response system to hypotension. Vasopressor
hormones regulate blood pressure in health and disease by receptor binding and activation
of downstream intracellular signaling systems. The hormones and their relevant receptors
are norepinephrine/epinephrine: α1,β1, β2; angiotensin II: AGTR1, AGTR2; vasopressin:
AVPR1a, AVPR1b, AVPR2; and dopamine: DA1, DA2 ([Fig. 1]).
These hormone systems ([Fig. 2]) are complex, have numerous interactions, and converge on a limited number of receptors
that are widely dispersed across the arterial and venous blood vessels. Furthermore,
the receptors are up- and downregulated adding to variation in response to native
hormone levels and exogenous infusion. Finally, there are genetic variations of the
receptors and downstream signaling systems in cells, again adding heterogeneity of
patient response to these hormone vasopressors.
Fig. 2 The endocrine system is central to the homeostatic response to cardiovascular stress
and hypotension. The complex endocrine response to septic shock includes the following:
(1) release of norepinephrine and epinephrine from the adrenal medulla, (2) release
of adrenocorticotropic hormone (ACTH) from the anterior pituitary then stimulating
synthesis and release of cortisone and cortisol from the adrenal cortex, (3) release
of vasopressin from the posterior pituitary, and (4) release of renin (in response
to hypotension) from the kidney. Renin is converted to angiotensin I by angiotensinogen
(released from the liver), and then angiotensin I is converted to angiotensin II by
angiotensin converting enzyme (ACE) in the lung. Angiotensin II increases aldosterone
synthesis and release from the adrenal cortex; aldosterone increases sodium retention
in the kidney. Angiotensin II also increases release of vasopressin. Norepinephrine
and epinephrine bind to α-1 adrenergic receptors, vasopressin binds to AVPR1a receptors,
and angiotensin binds to angiotensin receptors (AG 1 and AG 2), all on vascular smooth
muscle. After ligand binding, intracellular signal transduction increases intracellular
calcium, thereby causing vascular smooth muscle contraction and vasoconstriction.
Corticosteroids have complex cardiovascular effects that are incompletely understood
but include modulation of α-1 receptor density and other actions that quickly increase
responsiveness to catecholamines, such as norepinephrine.
The Highly Variable Practice of Vasopressor Use in Shock
The Highly Variable Practice of Vasopressor Use in Shock
Vasopressor choice varies quite widely despite international guidelines (e.g., Surviving
Sepsis Campaign [SSC][1]) for the above patient heterogeneity reasons and also because of behavioral variation
of physician practice. The remarkably variable use of vasopressin in practice is illustrative.
In a U.S. observational study of over 500,000 patients,[3]
[4] patients in “high vasopressin use” hospitals were about three times more likely
to receive vasopressin than patients in “low vasopressin use” hospitals. Furthermore,
norepinephrine doses varied widely at baseline in norepinephrine control groups of
vasopressor RCTs (mean: 0.20–0.82 μg/kg/min)[5] and not entirely due to differences in severity of shock at baseline.
Vasopressors are indicated for patients with “inadequate” response to fluid resuscitation[1] but “inadequate” response to fluid resuscitation varies from patient to patient
in part because methods to determine volume status are relatively inaccurate and because
different physicians have different opinions about when a patient is adequately resuscitated.
Vasopressor Use and Guidelines
Vasopressor Use and Guidelines
In the SSC guidelines, norepinephrine is the first-line vasopressor.[1] However, there is no high quality RCT evidence of alternative vasopressors as first-line
vasopressors because RCTs of vasopressors include patients who are already nearly
all on norepinephrine at baseline to which a second vasopressor (e.g., vasopressin,
angiotensin II) is added. Recently, a Cochrane analysis concluded that there was insufficient
mortality evidence to declare that any vasopressor was superior to others.[6]
Giving a vasopressor for hypotension in the presence of a very high cardiac output
(vasoplegia) is very different than in the presence of a normal or low cardiac output.
In the former case, a pure vasopressor (vasopressin, phenylephrine, angiotensin II)
may be helpful, but in the latter, it can be very harmful.
Fluid administration is important and vasopressors and inotropic agents may be required,
but we emphasize that blood pressure is not the only variable to focus on. These potent
interventions require titration and monitoring of markers of perfusion such as mentation,
skin perfusion, urine output, and serum lactate to guide drug choice, dose, and titration.
A target mean arterial pressure (MAP) during vasopressor use of 65 mm Hg is recommended
by SSC[1] and this is a general target that can be adjusted based on individual patient characteristics,
i.e., one size does not fit all. Clinicians questioned whether a higher target MAP
may lead to better outcomes and was the primary hypothesis of a high-quality RCT of
usual versus high MAP targets.[7] Asfar and colleagues[8] found no difference in short-term mortality between “usual” MAP (65–70 mm Hg) and
high MAP (80–85 mm Hg) targets. The high-target MAP group had significantly less acute
kidney injury (AKI) in the subgroup of patients who had pre-existing hypertension.[8] However, a recent pooled analysis reported that lower blood pressure targets were
not associated with adverse events, including chronically hypertensive patients.[9]
Septic shock is the commonest cause of vasodilatory shock and this definition was
recently revised.[10] The new definition of septic shock 3.0 requires (1) infection, (2) use of vasopressor(s),
and (3) a serum lactate >2 mmol/L.[10] The addition of the increased lactate level was based on a review of outcomes of
patients in very large cohorts, and marks a higher mortality than in patients with
lactate <2 mmol/L[11] and also recognizing that shock is more than simple hypotension. The new definition
of septic shock 3.0 will have ramifications for practice and for clinical studies
by altering the inclusion criteria of RCTs of therapies for septic shock. When this
new septic shock 3.0 definition was applied to patients in a prior pivotal RCT of
norepinephrine versus vasopressin (Vasopressin and Septic Shock Trial [VASST]),[12] it was predicted that patients who had lactate >2 mmol/L had higher mortality rates
(∼10% higher) and the response to vasopressin depended on lactate at the baseline.
Vasopressin was most effective in patients who did not meet the new definition (i.e., patients who were on vasopressor(s) with lactate ≤
2 mmol/L).[13] Thus, the baseline lactate level was both a prognostic (for mortality) and a predictive
(for response to vasopressin) biomarker.
Vasopressor Trials: Limitations, “Misses,” and Shortfalls
Vasopressor Trials: Limitations, “Misses,” and Shortfalls
The septic shock vasopressor RCT field has been plagued by negative pivotal RCTs despite
many positive small proof-of-principle RCTs. Indeed, to date, no RCT has found significant
difference in short-term (e.g., 28-day) mortality rate between various vasopressors.
VASST did show decreased mortality in prespecified, prestratified and preset p-values, and low severity of shock group, but this is not generally recognized. This
result provides some evidence that RCTs in this area are not futile if properly powered.
In addition, several strategies added a second vasopressor (e.g., vasopressin,[12]
[14] angiotensin II[15]) to decrease the dose and duration of norepinephrine infusion, but to date, these
RCTs have not shown a significantly decreased 28-day mortality rate. One exception
discussed in more detail below is that corticosteroids routinely decrease the dose
requirements of norepinephrine and other catecholamine vasopressors and also decrease
mortality in some,[16]
[17] but not all,[18]
[19] pivotal RCTs.
In sum, evidence-based vasopressor use has changed: although norepinephrine remains
the first-line vasopressor, dopamine has a negative recommendation[1] because of its tachycardic and tachyarrhythmic adverse effects; angiotensin II[15]
[20]
[21] is now approved in the United States and by the European Union and is available
clinically in the United States; and a novel vasopressin derivative (i.e., selepressin[22]
[23]) has completed a large Phase 2B RCT[24] with no effect on the primary outcome and so will not be available clinically.
Herein we review pathophysiology of vasodilatory shock, major vasopressor RCTs, and
describe whether, when, and what vasopressor(s) to use specifically: norepinephrine,
epinephrine, phenylephrine, dopamine, vasopressin, terlipressin, selepressin, and
angiotensin II. We cover the pertinent pharmacology, guidelines, effects, adverse
effects, dosing, and outcomes.
We briefly consider what inotropic agents to infuse to complement vasopressors because
vasopressors can decrease the ventricular function and cardiac output, so an inotropic
agent(s) is added to vasopressors in 15 to 30% of patients on vasopressors.[25] Several potential candidate predictive biomarkers identify patients who are good
or poor responders to vasopressors. Although at first seemingly illogical, there is
in fact a rationale and some evidence for an ironic role for β1-blockers in septic
shock, but more RCTs are needed to extend the current incomplete evidence base in
septic shock. There is also a role for vasopressors in vasodilatory shock post-cardiovascular
surgery, post-general/abdominal surgery, and post-anesthetic and other drug injection.
We conclude with recommendations that we will help the practicing clinician.
The Complex Pathophysiology of Vasodilatory Shock
The Complex Pathophysiology of Vasodilatory Shock
Vasodilatory shock is characterized physiologically by excessive vasodilation (with
low systemic vascular resistance), hypotension, and inadequate perfusion (hyperlactatemia,
oliguria, confusion)[26] because of inappropriate vascular smooth muscle relaxation as the primary event
and continued vasodilation despite hypotension, the most potent stimulus for vasoconstriction.[27]
[28] Septic shock is also often complicated by ventricular dysfunction and hypovolemia
in septic shock.
The endocrine system has a redundant array of hormonal responses of very potent vasoconstrictor
hormones that are released from stores in early vasodilatory and septic shock, leading
to elevated plasma levels of norepinephrine, epinephrine, vasopressin, angiotensin
II, aldosterone, adrenomedullin, and cortisol. These hormones act synergistically
on their unique complementary receptors on vascular smooth muscle endeavoring to increase
vasomotor tone, and occupying cardiac myocyte receptors with a usual net effect of
increasing heart rate and contractility. Adrenergic receptor downregulation,[29] receptor genotype differences between patients,[30]
[31] decreased responsiveness in septic shock,[27] and variable hormone metabolism[32] all conspire to allow continued vasodilation and hypotension. There is also a profound
deficiency of vasopressin later in septic shock.[33]
Major mediators of vasodilation in septic shock are nitric oxide, prostaglandins,
and adrenomedullin. Endotoxin and cytokines induce inducible NO synthase (iNOS) to
simulate NO synthesis.[34]
[35] Endotoxin and inflammatory cytokines stimulate prostacyclin synthesis and release
by endothelial cells.[36]
[37] RCTs in septic shock and sepsis of an iNOS inhibitor and a prostaglandin synthesis
inhibitor (ibuprofen) showed that they actually worsened[38] or had no effect respectively on mortality.[39] Adrenomedullin, a vasodilating hormone and cardiac depressant in septic shock,[40] is associated with increased mortality[40]
[41] and antiadrenomedullin decreases mortality,[42] increases responsiveness to norepinephrine,[43] and improves renal function in animal models of sepsis.[42]
[43] Antiadrenomedullin is rational for RCTs in septic shock.
Vasopressor Role and Protocols in Shock Resuscitation
Vasopressor Role and Protocols in Shock Resuscitation
Airway, breathing, and circulation resuscitation are fundamental priorities in vasodilatory
shock. Intravascular volume status (jugular venous pressure) and perfusion (skin temperature,
mentation, urine output) evaluation are supplemented by arterial blood gases, lactate,
hematology, renal and hepatic function, and bedside echocardiography ([Fig. 3]). Rapid, accurate screening for sepsis accelerates recognition and earlier intervention
thereby improving outcomes.[44] The most recent sepsis 3.0 definition recommends screening for sepsis by using the
quick Sequential Organ Failure Assessment Score (qSOFA) because it can be done quickly
and does not require laboratory test results (respiratory rate > 22/min, altered mentation,
and systolic blood pressure < 100 mm Hg) for screening for sepsis[10] based on good evidence from very large sepsis cohorts.[11] However, qSOFA criteria apply to most patients with any form of shock, so are relatively
nonspecific and not useful for differentiating septic from other causes of shock.
Fig. 3 Algorithm for vasopressor management in vasodilatory and septic shock. The first
priority is airway, breathing, and circulation (ABC) resuscitation, while in parallel
doing laboratory evaluation (arterial blood gases, lactate, hematology, renal and
hepatic function) and evaluating the cause of vasodilatory shock. Initial fluids (30 mL/kg
initially and more as needed) should be crystalloid. In patients not responding to
adequate fluid resuscitation, norepinephrine is started. In patients unresponsive
to norepinephrine, vasopressin (terlipressin) or epinephrine is added. In profoundly
hypotensive patients, phenylephrine or angiotensin II may be considered. Regarding
the cause of shock, fever and leukocytosis suggest septic shock and the need to search
for source of sepsis and drainage of abscesses and empyema. Sepsis mimics include
post-acute myocardial infarction (AMI), post-cardiovascular surgery and other causes
(pancreatitis, aspiration, acute respiratory distress syndrome [ARDS], post-abdominal
surgery, trauma, and drugs [anesthetics and drug allergy/anaphylaxis]). 1In patients not responsive to norepinephrine, vasopressin, epinephrine, or angiotensin
II, cardiovascular evaluation is necessary. 2Cardiovascular evaluation should occur such as limited bedside echocardiograph, noninvasive
cardiac output, central venous pressure (CVP) or pulmonary capillary wedge pressure
(via pulmonary artery catheter). If there is decreased ventricular function (decreased
ejection fraction), then dobutamine should be added. 3Not responsive to norepinephrine or other vasopressors is not well defined but generally
means not responsive to a high dose. 4Vasopressin can be substituted with terlipressin but the randomized controlled trials
of terlipressin are much smaller than with vasopressin. Selepressin (a highly specific
AVPR1a agonist) is in development.
We emphasize that vasopressors should be administered simultaneously with fluid replacement
to prevent and decrease duration of hypotension in shock with vasodilation. Volume
resuscitation (30 mL/kg initially is recommended but more or less may be needed) with
crystalloid should precede or coincide with norepinephrine infusion, added if perfusion
remains inadequate.[1] Assessment of volume status is somewhat inaccurate; fluid overload is associated
with increased mortality of septic shock[45]
[46]
[47]; and a restrictive fluid practice is under investigation in septic shock.[48] Consideration of the differential diagnosis of vasodilatory shock occurs while initiating
resuscitation ([Fig. 3]). Fever or hypothermia, leukocytosis or leukopenia, and an obvious source of infection
(the commonest sources are pneumonia, abdominal infection/peritonitis, urinary tract
infection, and skin source) suggest sepsis. The sepsis source should be investigated
to ensure adequate source control of abscesses and empyema. Other prevalent conditions
that can present with vasodilatory shock are acute pancreatitis, aspiration, acute
respiratory distress syndrome, post-cardiovascular and other surgeries, post- AMI,
trauma, and drugs (anesthetics and drug allergy/anaphylaxis).
Patients receiving vasopressors may require invasive arterial pressure monitoring
by an arterial catheter (or by noninvasive automated cuff blood pressure) complemented
by central venous access for vasopressor administration and central venous pressure
monitoring.[1] In more severe shock, clinicians use a variety of invasive and noninvasive cardiovascular
assessment and monitoring.
Early treatment of septic shock is critical as illustrated by studies of early antibiotics[44]
[49] and early goal-directed therapy.[44]
[50] These studies and a recent RCT of early use of norepinephrine[51]
[52] align with an artificial intelligence (AI) study in which the AI clinician recommended
vasopressors be given more often (30 vs. 17%) than was used in the care of septic
patients.[53] However, uncontrolled observational data found that earlier vasopressor use is harmful,[54] suggesting equipoise regarding earlier, more frequent use of vasopressors in septic
shock.
See [Table 1] for a summary of vasopressors, their cognate receptors, actions, usual doses, and
potential predictive biomarkers.
Table 1
Vasopressors, their receptor binding, possible additional beneficial actions, dose,
and possible relevant biomarkers
|
Vasopressor
|
Receptor activity
|
Additional actions
|
Dose
(all intravenous)
|
Possible predictive biomarkers
|
|
Norepinephrine
|
α1 > β1, β2
|
Immune activity[179]
|
5–100 μg/min
|
β2 receptor SNP[30]
|
|
Epinephrine
|
α1 > β1, β2
More β1 than NE
|
Immune activity[179]
|
5–60 μg/kg/min[55]
|
β2 receptor SNP[30]
|
|
Phenylephrine
|
α1
|
Immune activity[179]
|
50–100 μg bolus
0.1–1.5 μg/kg/min
|
|
|
Dopamine
|
DA1, DA2
|
Immune activity[180]
[181]
|
1–5 μg/kg/min: “low dose.”
5–15 μg/kg/min: moderate dose
20–50 μg/kg/min: high dose
|
|
|
Vasopressin
|
AVPR1a, AVPR1b, AVPR2
|
Immune activity[182]
|
0.01–0.04 U/min[12]
[105]
|
LNPEP SNP[32]
Angiopoietin 1/2[128]
Vasopressin/copeptin
|
|
Terlipressin
|
AVPR1a (AVPR1b) > AVPR2
|
Immune activity
|
1.3 μg/kg·h[183]
20–160 μg/h,[113] bolus: 1 mg
|
LNPEP[32]
Vasopressin/copeptin
|
|
Selepressin
|
AVPR1a
|
Angiopoietin-2
Vascular leak
|
1.25–2.5 ng/kg/min in Phase 2[23]
1.25–5.0 ng/kg/min in Phase 3[22]
|
LNPEP SNP[32]
Angiopoietin 1/2[128]
Vasopressin/copeptin
|
|
Angiotensin-II
|
Angiotensin II receptors (AGTR1, AGTR2)
|
Vasopressin
Erythropoietin
|
5–200 ng/kg/min (first 3 h; 1.25–40 ng/kg/min up to 7 d[15]
|
AGTRAP SNP[31]
|
|
Methylene blue[184]
|
Inhibits GABAA receptors
|
Vascular leak
|
Bolus (2 mg/kg) then infusion—stepwise increasing rates: 0.25, 0.5, 1, 2 mg/kg/h
|
|
Abbreviations: AGTR1 and AGTR2, angiotensin II receptors 1 and 2; AGTRAP, angiotensin
II receptor associated protein; GABAA, gamma-aminobutyric acid; LNPEP, leucyl and
cystinyl aminopeptidase; SNP, single nucleotide polymorphism.
Norepinephrine is the first-line vasopressor in septic shock because it is superior
to dopamine and equivalent to vasopressin and epinephrine in pivotal RCTs of norepinephrine
versus epinephrine,[55] norepinephrine versus dopamine,[56] norepinephrine plus dobutamine versus epinephrine,[57] early vasopressin[14] versus norepinephrine, and vasopressin versus norepinephrine in septic shock[12] ([Table 2]).
Table 2
Pivotal randomized controlled trials of vasopressors in septic shock
|
Vasopressor intervention
|
Control (reference number)
|
Intervention
mortality (%)
|
Control
mortality (%)
|
AD (95% CI)
p-value
|
|
NE
|
AVP[12]
|
35.4%[a]
|
39.3%
|
3.9 (−2.9–10.7)
0.26
|
|
NE
|
AVP[14]
|
30.9%[a]
|
27.5%
|
3.4 (−5.4–12.3)
|
|
SE
|
Placebo
|
40.6[b]
15.0
|
39.4
14.5
|
1.1 (−6.5–8.8)
0.77
|
|
NE
|
DA[56]
|
48.5%[c]
|
52.5%
|
1.17 (0.97–1.42)
0.10
|
|
ANG II
|
Placebo[15]
|
46%[a]
|
54%
|
HR: 0.78 (0.57–1.07)
0.12
|
|
Epi
|
NE[55]
|
23%[a]
|
27%
|
HR: 0.87 (0.48–1.58)
0.65
|
|
Epi
|
NE + DOB[57]
|
40%[a]
|
34%
|
RR: 0.86 (0.65–1.14)
0.31
|
Abbreviations: AD, absolute difference; ANG II, angiotensin II; AVP, vasopressin;
DA, dopamine; DOB, dobutamine; Epi, epinephrine; HR, hazard ratio; NE, norepinephrine;
RR, relative risk; SE, selepressin.
a 28-day mortality.
b 90-day mortality.
c All causes of shock; 28-day mortality.
Adverse Effects and Risks of Vasopressors
Adverse Effects and Risks of Vasopressors
Digital and organ ischemia/dysfunction, decreased cardiac function, tachyarrhythmias,
and atrial fibrillation[8] (with increased risk of stroke in septic shock)[58] are the commonest serious adverse effects of vasopressors. Higher cumulative doses
of vasopressors are associated with more organ dysfunction and higher mortality,[59] but association studies may be confounded by indication and coexisting severity
of illness, so it is not clear that the higher cumulative doses of vasopressors caused more organ dysfunction and higher mortality. Similarly, a recent study of patients
on high-dose norepinephrine (>1.0 μg/kg/min) demonstrated that such high-dose norepinephrine
is a strong independent predictor of mortality in a multivariate model, but it is
impossible to assign causality in such an association study.[60]
Monitoring
Vasopressor use in shock requires continuous monitoring of MAP by an arterial line
or by noninvasive arm cuff blood pressure, evaluation of perfusion (mentation, urine
output, lactate), and noninvasive cardiovascular assessment (e.g., noninvasive cardiac
output, echocardiographic evaluation of ventricular function and volume status, i.e.,
inferior vena cava collapse). Assessment of the microcirculation by sublingual techniques
is used by some but not routinely recommended.
Goal-directed bedside echocardiography may be effective to guide fluid, vasopressor
and inotropic agent choice, and infusion dose. Volume status was often more than replete
in a case–control study of bedside echocardiography in resuscitated but still in shock
intensive care unit (ICU) patients. Further fluid restriction was recommended in 65%
of patients and initiation of dobutamine in 25% of patients[61] because of findings of decreased right or left ventricular dysfunction. Mortality
was lower in the limited bedside echocardiography group than in controls.[61] To date, there is no pivotal RCT of limited bedside echocardiography-guided resuscitation
versus usual care in shock.
Weaning of Vasopressors
The weaning of vasopressors has not been evaluated critically in RCTs and so the evidence
of optimal best practice is missing. Generally, patients are judged appropriate for
weaning (gradual, e.g., hourly, decrements of vasopressor dose) when “stable” (surprisingly
there is no universal definition of hemodynamic stability) and both volume status
and perfusion are adequate. Deterioration of MAP or perfusion necessitates titration
of vasopressor(s) back up to higher doses, attempting to re-establish stability, and
later by repeated decrements of vasopressor dose until vasopressor infusion is off.
The complexity and heterogeneity of vasopressor weaning is illustrated by the insight
that medical informatics accurately predicts successful vasopressor weaning both earlier
and more accurately than can clinicians using standard clinical practice.[62]
We recommend that norepinephrine be weaned first followed by weaning of the second
vasopressor (vasopressin or epinephrine) if the patient remains hemodynamically stable.
Weaning norepinephrine first has been shown to decrease the risk of hemodynamic instability
during vasopressor weaning.[63]
[64] Difficulties with weaning vasopressors has not been well studied but may be related
to ongoing septic shock, inadequate volume resuscitation and ongoing hypovolemia,
or decreased ventricular contractility due to sepsis or other causes.
Main Outcomes of Pivotal RCTs of Vasopressors
Main Outcomes of Pivotal RCTs of Vasopressors
The mortality of septic shock appears to be decreasing[65] and short-term mortality has been the commonest primary outcome for pivotal RCTs
of vasopressors in septic shock. Consequently, RCTs of vasopressors are evolving in
two complementary directions. Some now focus on improving long-term outcomes; others
aim to improve short-term organ dysfunction (e.g., days alive and free of vasopressors[22]) because short-term organ dysfunction is associated with long-term mortality.[66]
[67]
[68]
[69] For example, the pivotal RCT of selepressin in septic shock used vasopressor- and
ventilator-free days as the primary outcome.[22]
[24]
For each vasopressor we now review pharmacology, guidelines, effects, adverse effects,
and dosing ([Table 1]).
Norepinephrine: The First-Choice Vasopressor
Norepinephrine: The First-Choice Vasopressor
Since the initial discovery of epinephrine as the primary vasopressor hormone produced
in the adrenal medulla, the pharmacology of catecholamines has been elucidated.[70] Norepinephrine is structurally similar to epinephrine except that norepinephrine
lacks a methyl group on its nitrogen. Some clinicians may find it surprising that
in the adrenal gland, norepinephrine is converted into epinephrine via the enzyme
N-methyltransferase. This pharmacology may question the value of utilizing epinephrine
in a patient with refractory shock on high doses of norepinephrine, unless it is being
used for its inotropic effects. Norepinephrine has a very short half-life (<5 minutes).
Some advantages that norepinephrine has compared with other vasopressors include vasopressor
potency because of α1 agonism, minimal effect on β2 adrenergic receptors (that can
increase lactate levels), and the ability to increase cardiac index without increasing
heart rate or myocardial oxygen consumption. Unlike epinephrine, in the United States
norepinephrine is frequently programed into the intravenous pump as μg/min rather
than the weight-based dosing of μg/kg/min. Commonly norepinephrine is started at 5
to 10 μg/min and titrated up to the target MAP, usually 65 mm Hg.[1] An RCT of usual versus high MAP target showed that the high MAP target (80–85 mm
Hg) decreased AKI in previously hypertensive patients.[7] Similarly, two recent single-center studies suggested that higher MAP (75–85 mm
Hg) was associated with less AKI and higher survival.[71]
[72]
One of the first studies of norepinephrine in shock in 1953 was a case cohort study
(n = 32) with various forms of shock.[73] Twenty-six patients improved in their shock and the mortality rate (62.5%) was below
the “expected” level (>80% as per the authors). Since then, most studies have compared
norepinephrine to other vasopressor(s) (dopamine, epinephrine, and vasopressin). In
a recent meta-analysis of 43 trials (n = 5,767 patients) that assessed 17 vasopressors and inotropic agents,[74] the combination of norepinephrine and dobutamine was associated with the lowest
mortality while dopamine was associated with the highest incidence of arrhythmia.
The largest RCT of norepinephrine versus dopamine was performed in approximately 1,700
patients and reported no difference in mortality between the two vasopressors, but
in patients with cardiogenic shock, dopamine was associated with greater mortality.[56] Dopamine was associated with increased adverse events, particularly double the prevalence
of tachyarrhythmias compared with norepinephrine. In another RCT, there was no difference
in mortality or safety between a combination of norepinephrine and dobutamine versus
epinephrine.[57] Finally, an Australian RCT compared norepinephrine to epinephrine for shock and
demonstrated no difference between the two vasopressors in achieving blood pressure
control but significantly more patients randomized to epinephrine had the study drug
stopped for adverse events.[55]
As a result of these and other well-conducted RCTs, norepinephrine is recommended
as the first-line vasoactive drug for septic shock.[1] The widespread adoption of norepinephrine as a first-line vasopressor is also likely
due to its ease of use, rapid onset of action, familiarity, low cost, and the lack
of an alternative agent showing superiority to norepinephrine. A recent global web-based
survey (839 physicians from 82 countries, 65% main specialty/activity intensive care)
responded that the commonest first-line vasopressor was norepinephrine (97%), targeting
predominantly a MAP >60–65 mm Hg (70%), with higher targets in patients with chronic
arterial hypertension (79%).[75]
It is recommended that norepinephrine be administered via a central intravenous catheter
or a large bore peripheral intravenous catheter in the antecubital vein to avoid serious
complications if an extravasation occurs (https://www.accessdata.fda.gov/drugsatfda_docs/label/2007/007513Orig1s024lbl.pdf). If extravasation does occur, infusion needs to be stopped and the affected area
should be infiltrated with the α blocker phentolamine to reverse the severe vasoconstriction
that can cause ischemia and necrosis. A recent study of peripherally administered
norepinephrine demonstrated that extravasation occurred in only 0.035% of approximately
14,000 perioperative patients.[76] Another recent RCT of early, low dose (0.05 μg/kg/min) often started via peripheral
intravenous line showed improvement of the primary endpoint, control of shock (defined
by MAP > 65 mm Hg, plus urine output > 0.5 mL/kg/h or 10% decrease in serum lactate).[51]
The most common adverse effects of norepinephrine are due to activation of α1 receptors
causing excessive vasoconstriction and decreased end-organ perfusion. This complication
occurs more commonly when norepinephrine is infused without appropriately correcting
hypovolemia. Vasoconstriction secondary to α1 stimulation can cause reflex bradycardia
via the baroreceptor reflex, which is generally not compensated for by norepinephrine's
weak β1 activity. The overall result is that cardiac output may decrease or not change
despite β1 agonism. At the same time, the increase in systemic vascular resistance
increases the work of the heart by increasing afterload, thereby increasing myocardial
oxygen demand.
Norepinephrine administration may increase pulmonary vascular resistance which could
have negative sequelae in patients with pulmonary hypertension. Decreased hepatic
blood flow (secondary to α-mediated vasoconstriction) may alter hepatic metabolism
of drugs leading to a transient increase in drug levels. Generally, the use of norepinephrine
is contraindicated in patients with mesenteric or peripheral vascular thrombosis because
subsequent norepinephrine-mediated vasoconstriction could increase risk of ischemia
and infarction.
Epinephrine
Epinephrine has more β1 agonism than norepinephrine. Epinephrine is a second-line
agent in septic shock[1]
[55]
[74] in patients not responding to norepinephrine. Epinephrine is comparable to norepinephrine,[55] to norepinephrine plus dobutamine,[57] and to norepinephrine and vasopressin[77] in efficacy in RCTs and in a meta-analysis[74]; however, epinephrine has a greater risk of mesenteric ischemia, tachyarrhythmias,
and hyperlactatemia compared with norepinephrine.[1]
[55]
[57] Epinephrine is an optional first-choice vasopressor in countries where norepinephrine
is more costly.[78]
Phenylephrine
At the turn of the 19th century, following the identification of the adrenal medullary
hormones responsible for regulation of cardiovascular function, there was a surge
of interest into modifying the core structure common to norepinephrine and epinephrine
([Fig. 4]). In 1910 several potentially sympathomimetic chemicals were synthesized by Dale
and their cardiovascular effects characterized by Barger.[79] Apart from the endogenous epinephrine and norepinephrine, phenylephrine was found
to be the most potent vasopressor. Phenylephrine was later studied by Trendelenburg
using reserpine, a drug which depletes presynaptic noradrenaline, and cocaine, which
inhibits norepinephrine reuptake.[80]
[81] Using these agents Trendelenburg proved that phenylephrine is a direct-acting sympathomimetic
with direct stimulation of adrenergic α1 receptors. In contrast, methamphetamine stimulates
the release of endogenous norepinephrine.
Fig. 4 Chemical structure of commonly used catecholamines. Catecholamines have a catechol
group (benzene with two hydroxyl substitutions) and a side-chain amine. Phenylephrine
is a synthetic derivative, while epinephrine, norepinephrine, and dopamine are endogenous
catecholamines.
In healthy subjects intravenous infusion of phenylephrine acts via the α1 adrenergic
receptor, rapidly increasing diastolic and systolic blood pressure(s), decreasing
renal blood flow, and yielding vagally mediated reflex bradycardia and a slight drop
in cardiac output.[82]
[83] However, if the heart rate does not drop reflexively (using atropine or vagotomy),
phenylephrine increases cardiac output, likely through increased left ventricular
end diastolic volume (preload).
Phenylephrine has a much slower rate of uptake and clearance from extracellular fluid
than norepinephrine, which is rapidly taken up into adrenergic nerves following intravenous
injection. Additionally, unlike endogenous catecholamines, phenylephrine is not metabolized
by catechol-O-methyl-transferase in liver and other tissues. In healthy individuals,
this causes a vasopressor effect for 20 minutes, much longer than the 1 to 2 minutes
seen with norepinephrine bolus.[84] Like epinephrine and norepinephrine, phenylephrine is metabolized by monoamine oxidase
in mitochondria of nerves and liver.[84]
When compared with norepinephrine for continuous infusion in mixed shock states, phenylephrine
appears less potent, requiring 220% of the dose of norepinephrine to achieve a MAP
of 65 to 75 mm Hg following fluid resuscitation.[85] This likely relates to phenylephrine's lack of β adrenergic effects (positive inotropy
and chronotropy) rather than tachyphylaxis. At low doses phenylephrine has only α1
adrenergic activity, while higher doses cause some β1 adrenergic activity.[82]
[86] At 12 hours following infusion initiation, and at 220% of the dose of norepinephrine,
phenylephrine increased the heart rate similarly to norepinephrine.
A national shortage of norepinephrine in the United States set the stage for a natural
experiment comparing efficiency of norepinephrine versus phenylephrine in septic shock.
Phenylephrine was the commonest replacement during the norepinephrine shortage; in
a before/after observational cohort study of patients who had septic shock, phenylephrine
was associated with increased mortality compared with norepinephrine.[87]
Given the favorable (prolonged) kinetic profile of phenylephrine over epinephrine
and norepinephrine following a single bolus, there may be a niche for its use in early
resuscitation of shock in geographic areas such as the medical ward or prehospital
care where initiation of infusions is challenging and would take valuable time away
from transportation. Phenylephrine's routine use as a first-line agent for continuous
infusion to reverse shock states cannot be recommended because of lower potency than
norepinephrine and significant tachyphylaxis requiring the introduction of an additional
vasopressor.
Dopamine
Dopamine had great promise because it has potential to increase cardiac contractility
and stroke volume while augmenting renal perfusion and urine output.[88]
[89]
[90] Although structurally very similar to the other endogenous and synthetic catecholamines
([Fig. 4]), dopamine also stimulates dopaminergic receptors DA1, DA2, and DA4.[91]
In healthy volunteers intravenous dopamine stimulates dopaminergic receptors beginning
at 1 μg/kg/min, with a plateau of stimulation at 3 μg/kg/min;[92] at less than 3 μg/kg/min there is no adrenergic receptor stimulation. When healthy
volunteers are infused at or below 3 μg/kg/min, there is a large, physiologically
significant 128% increase in renal sodium excretion and a 43% rise in glomerular filtration
rate.[93] Dopamine also has nonreceptor-mediated renoprotective effects in profound renal
ischemic insult, protection mediated by mitigation of oxidative stress induced by
reactive oxygen species.[94] When low-dose (4 μg/kg/min) dopamine was added for 3 to 9 hours in neurologically
deceased donors (NDDs), it significantly decreased the need for posttransplant dialysis.[95] Therefore, while the use of any catecholamine vasopressor drug is recommended to
treat the NDD kidney donor,[96] the global recommendations suggest a combination of vasopressin and norepinephrine
to treat shock states, while dopamine would be preferred in the absence of shock.[95]
At doses exceeding 3 μg/kg/min in healthy volunteers, dopamine has additional α and
β adrenergic stimulation.[93] Consequently, in the 1980s dopamine was used as a first-line vasopressor for critically
ill patients because of its distinct dose-dependent effects, low-dose (3 μg/kg/min)
dopaminergic stimulation, and higher dose α and β adrenergic stimulation.
However, these uniquely titratable dopamine dose effects are reproducibly seen only
in subjects with normal cardiovascular physiology. Unfortunately, in patients with
shock the plasma clearance decreases by 50% compared with surgical controls, while
those with shock and renal dysfunction have a 75% reduction in plasma clearance.[97] The unpredictable relationship between infusion rate and plasma levels causes variable
receptor activation that can cause adverse events due to excessive α and β adrenergic
stimulation.
Dopamine is not recommended for treatment of shock because of the results of a large
RCT of dopamine versus norepinephrine (n = 1,629) as first-line vasopressor therapy for undifferentiated shock.[56] Dopamine-treated patients had a distinct cardiovascular profile. During the first
day of therapy, those treated with dopamine had much higher heart rates (excluding
arrhythmias) than norepinephrine-treated patients (102 vs. 94 beats per minute (BPM),
respectively). Furthermore, dopamine was associated with significantly double the
rate of tachyarrhythmias (24 vs. 12%). Urine output was higher in those treated with
dopamine (by 200 mL per day during the first day of infusion).[56] While there was no statistically significant difference in 28-day mortality between
dopamine and norepinephrine (52.5 vs. 48.5%), and in the prespecified cardiogenic
shock subgroup (n = 280), there was a 5% higher 28-day mortality with dopamine versus norepinephrine.
Thus, dopamine has more risk than benefit in the treatment of shock because of its
narrow therapeutic index and uncertain plasma levels as a result of decreased renal
excretion leading to a higher heart rate and double the frequency of tachyarrhythmias.
The use of dopamine in the critical care unit is therefore best reserved for renal
protection in the NDD organ donor, although this recommendation is based on evidence
from a single RCT[95] and has therefore not been widely advocated in guidelines for management of organ
donors.[96]
Vasopressin Analogues
Vasopressin
Vasopressin is an endogenous hormone released from the posterior pituitary gland.
In health it is mainly involved in osmoregulation because of binding to V2 receptors
in the distal convoluted tubules and promoting water retention (as antidiuretic hormone).
In shock states, circulating vasopressin levels rise because of rapid release of vasopressin
stores from the posterior pituitary gland, and vasopressin acts as a powerful vasoconstrictor,
binding to V1a receptors. Interest in administering exogenous vasopressin in septic
shock began after the seminal report of a relative vasopressin deficiency in septic
shock compared with other shock states.[98] Furthermore, vasopressin may maintain better perfusion of the kidney compared with
norepinephrine because of the heterogeneous distribution of V1a receptors in the kidney
(more V1a receptors in glomerular efferent than afferent arterioles).[99] In small clinical studies vasopressin infusion increased urine output and improved
creatine clearance compared with norepinephrine.[100]
[101]
However, larger RCTs have not demonstrated decreased mortality compared with norepinephrine
treatment of septic shock. There was no difference in overall mortality between vasopressin
and norepinephrine in VASST,[12] but vasopressin may have been more effective in less severe shock (baseline norepinephrine
dose < 15 μg/min). Vasopressin (<0.04 units/min) was associated with similar outcomes
to norepinephrine in a propensity-matched cohort study.[102] In VANISH,[14] early vasopressin had similar mortality to norepinephrine. This lack of mortality
reduction was confirmed in a recent individual patient data meta-analysis (IPDMA),
including over 1,400 patients, in which the relative risk (RR) for 28-day mortality
was 0.98 (95% confidence interval [CI]: 0.86–1.12).[103] Interestingly there was a signal to lower mortality at 90 days (RR: 0.91 95% CI:
0.81–1.01) with vasopressin suggesting that it is important to include long-term effects
of intensive care interventions in clinical trials. In this IPDMA there was less frequent
use of renal replacement therapy (RRT) with vasopressin (RR: 0.86, 95% CI: 0.74–0.99),
most notably in patients without significant AKI at the time of inclusion, supporting
the potential benefit of vasopressin to prevent deterioration in renal function, even
if there is no benefit on mortality.
Importantly this recent IPDMA demonstrated that the number of serious adverse events
was similar for both vasopressin and norepinephrine. However, vasopressin had a different
side-effect profile compared to norepinephrine, notably with a lower rate of arrhythmias
(absolute risk difference: −2.8%, 95% CI −0.2 to −5.3). This finding was supported
in another meta-analysis including more than 3,000 patients who had any form of vasodilatory
shock, not just sepsis, and also included treatment using other vasopressin analogues.[104] There was a marked reduction in the rate of atrial fibrillation with vasopressin
(RR: 0.77, 95% CI: 0.67–0.88) and also a reduction in mortality in this broader population
(RR: 0.89, 95% CI: 0.82–0.97). However, both meta-analyses found that treatment with
vasopressin analogues also led to a higher incidence of digital ischemia, an absolute
risk increase of approximately 2% in both studies.
In summary, this evidence supports the use of vasopressin as a safe adjunctive vasopressor
to use in addition to norepinephrine. There is no clear evidence to support an improvement
in patient survival in septic shock but there is good evidence that vasopressin can
reduce rates of arrhythmias and tachycardia[25] and could reduce the requirement for RRT. Vasopressin derivatives may limit complications
in distributive shock,[105] but not in distributive shock when the vasoplegia is not established. Vasopressin
is not just another vasopressor to increase blood pressure. Vasopressin and its derivatives
have favorable effects on hepatosplanchnic and renal perfusion,[106]
[107] pulmonary hypertension,[108] and mitigate edema formation and fluid loss due to increased permeability in several
sepsis models[106]
[107]
[109]
[110]
[111] and selepressin decreases fluid balance in a proof-of-principle RCT of selepressin.[23]
However, caution must be exercised because of the increased risk of digital ischemia.
Clinicians should consider this information when selecting which vasopressors to use
and select the combination that best balances the benefit/risk ratio of each individual
patient.
Terlipressin
Terlipressin is a synthetic analogue of vasopressin with a greater selectivity for
the V1a receptor than vasopressin and is used in patients who have liver failure in
an attempt to reduce hepatorenal syndrome.[112] It also has a longer half-life than vasopressin and so can be given by intermittent
bolus injection rather than as a continuous infusion.
There are less clinical trial data to support the use of terlipressin in septic shock.
One recent RCT was stopped for futility after the randomization and treatment of 535
patients, as it failed to show any difference in the primary outcome, 28-day mortality:
38 vs. 40% for norepinephrine- and terlipressin-treated patients respectively, p = 0.63.[113] However, there were markedly more serious adverse events in the terlipressin-treated
patients, especially digital ischemia (0.35 vs. 12.6% for norepinephrine- and terlipressin-treated
patients respectively, p < 0.0001). Why the rate of digital ischemia was so high in the terlipressin-treated
patients is not clear but could relate to unrecognized/untreated hypovolemia or possibly
the accumulation of terlipressin or its metabolites after administration of a continuous
infusion, particularly at higher doses (up to 160 µg/h was allowed in this RCT).
As a meta-analysis of RCTs of terlipressin in septic shock[114] reported no mortality benefit (RR: 1.00, 95% CI: 0.83–1.20) and there is concern
about its safety, terlipressin cannot be recommended ahead of vasopressin if both
are available. However, vasopressin is not available in all regions and so terlipressin
may be the only option, but care must be applied to first correct any hypovolemia
and avoid high doses to mitigate adverse effects of terlipressin.
Selepressin
Selepressin is another synthetic vasopressin analogue and is highly selective for
the V1a receptor. As well as being a potent vasoconstrictor, in both preclinical[107] and a Phase 2A proof-of-principle RCT,[23] selepressin reduced edema formation and intravenous fluid requirements. It was recently
tested in a large seamless Phase 2B/3 RCT, Sepsis-adaptive clinical trial, that had
four unique features for a pivotal RCT in septic shock.[24] First, there was a continuous response adaptive design,[22] meaning that assignment to a study drug group was determined in real time as new
patients were included and the primary endpoint was loaded into a prespecified computer
algorithm that then assigned a treatment group. Second, there was to be a seamless
transition from Phase 2 to Phase 3. Third, several doses of selepressin were to be
pooled for the decision to go to Phase 3. And fourth, the primary endpoint was vasopressor-
and ventilator-free days. However, the RCT was stopped after Phase 2B for futility,
because there was no difference in the primary outcome: the number of vasopressor-
and ventilator-free days[24] ([Table 2]). There were no differences in any of the other secondary outcomes or in any of
the prespecified subgroups. As selepressin is not currently approved for clinical
use, it remains to be seen if this drug can become a useful additional vasopressor
agent for the treatment of septic shock.
Angiotensin II
Angiotensin II has shown promise as a potent vasopressor for patients who have marked
hypotension due to vasodilatory shock based on preclinical studies and a series of
RCTs from the proof of principle to pivotal Phase 3. In response to hypotension, renin
is rapidly converted to angiotensin I in the kidney and then angiotensin I is secreted
and converted in the lung to angiotensin II by ACE. Angiotensin II is a crucial vasopressor
of the renin–angiotensin–aldosterone system (RAAS) that modulates vascular tone by
binding to angiotensin 1 and 2 receptors (AGTR1 and AGTR2), G-protein-coupled receptors
that increase cytosolic calcium concentrations to induce vasoconstriction (also aldosterone
synthesis and vasopressin release). AGTR1 is downregulated in models of sepsis, leading
to relative angiotensin II insensitivity.[15]
[20]
Angiotensin II was recently approved by the Food and Drug Administration[115] for treatment of vasodilatory hypotension based on the results of a clinical program
including a pilot RCT[20] and culminating in the Angiotensin II for the Treatment of High-Output Shock 3 (ATHOS-3)
trial.[15] ATHOS-3 added infusion of concealed angiotensin II or placebo in refractory vasodilatory
shock. Angiotensin II was initiated at 20 ng/kg/min and then increased to achieve
a MAP of 75 mm Hg with up to 200 ng/kg/min and after 3 hours until 48 hours angiotensin
II was adjusted to 1.25 to 40 ng/kg/min. At 48 hours angiotensin II study drug was
weaned off, but if a patient became unstable, the angiotensin II could be restarted
and maintained for up to 7 days. Angiotensin II more rapidly increased MAP over 3 hours,
decreased norepinephrine dose, and improved the cardiovascular SOFA score compared
with placebo.[15] Mortality (28-day) was 46 versus 54% (p = 0.12) in the angiotensin II and placebo groups, respectively.
A greater sensitivity to infused angiotensin II is associated with lower angiotensin
II levels prior to treatment and lower mortality. A prespecified analysis of the ATHOS-3
RCT was that there would be a difference in mortality according to whether patients
were titrated down from the initial dose of angiotensin II of 20 ng/kg/min to <6 ng/kg/min
at 30 minutes versus >5 ng/kg/min at 30 minutes.[116] The hypothesis was confirmed; mortality rates were 41 versus 67% (p = 0.0007) in the <6 ng/kg/min of angiotensin II at 30 minutes versus >5 ng/kg/min
at 30 minutes, respectively.[116] Interestingly, the mortality results were aligned with angiotensin II levels at
treatment initiation that were significantly lower in the <6 ng/kg/min of angiotensin
II at 30 minutes versus >5 ng/kg/min (128 pg/mL [199 pg/mL] vs. 421 pg/mL [680 pg/mL][mean,
standard deviation, p = 0.0009; normal range in health: 5–35 pg/mL[117]]).
There were no safety issues with angiotensin II in the pivotal ATHOS-3 RCT.[15] There was no difference in the frequency of serious adverse events between angiotensin
II (60.7%) and placebo (67.1%) including digital, gut, and myocardial ischemia and
arrhythmias.[15] Indeed, serious adverse events that led to discontinuation of study drug occurred
in 14.1% of angiotensin II-treated and 21.5% of placebo-treated patients in ATHOS-3.[15] Furthermore, cardiac disorders occurred in 16.6 versus 20.3%, respiratory disorders
in 10.4 versus 15.8%, and vascular disorders in 10.4 versus 9.5%, including thromboembolic
events (deep venous thrombosis 1.8 vs. 0%) (angiotensin II-treated vs. placebo-treated).[15]
Angiotensin II may be especially effective in patients who have AKI requiring RRT.
A post-hoc study of the subgroup of patients in the ATHOS-3 RCT who had AKI and were
on RRT at baseline found potential benefit of angiotensin II: lower mortality (47
vs. 70%, angiotensin II and placebo, respectively, p = 0.012) and more frequent recovery without the need for RRT by day 7 (38 vs. 17%,
angiotensin II and placebo, respectively, p = 0.007).[118] One interpretation is that these are patients who have more severe forms of septic
shock (AKI requiring RRT), have decreased renal perfusion pressure at onset, and have
shown improvement by angiotensin II, and so these patients benefit from angiotensin
II. Another interpretation is that the normal ratio of angiotensin I to angiotensin
II in health is 0.5, but it is 1.63 in vasodilatory shock,[15] indicating dysfunction of ACE in vasodilatory shock.[119] Infusion of angiotensin II improves the angiotensin I-to-II ratio. Genetic variations
of ACE are associated with worse renal function in sepsis,[120] further emphasizing the important role of angiotensin I and II in regulatory renal
function in septic AKI. Finally, an ovine model of septic AKI shows decreased renal
function and urinary oxygenation and angiotensin II infusion improves renal function
but does not worsen urinary oxygenation,[121] by angiotensin II preferentially constricting the efferent renal arteriole.[121]
[122]
Angiotensin II is especially effective in patients with a low ACE and high levels
of angiotensin I (120). Renin activity seems to correlate well with ACE. Normal levels
of renin activity are 0.5 to 2.0 ng/mL/h. Angiotensin II seems to be more effective
in patients with a high renin activity up to three times the normal levels. Plasma
renin activity could be a potential marker to select a population with a higher benefit.
Inotropic Agents to Complement Vasopressors in Septic Shock
Septic shock can decrease cardiac contractility and cardiac output and this decline
may be exacerbated by vasopressors. Therefore, inotropic agent(s) such as dobutamine
are commonly added to norepinephrine[57] and vasopressin[12]
[25] to increase cardiac output, but with side effects (tachyarrhythmias and increased
heart rate and myocardial oxygen consumption). The effect of the combination of dobutamine
plus norepinephrine was found to be equivalent to epinephrine alone in one large RCT.[57] Levosimendan, a positive nonadrenergic inotropic agent, was not effective in an
RCT in septic shock.[55]
[57] More patients on levosimendan had tachyarrhythmias and fewer patients on levosimendan
were successfully weaned from mechanical ventilation.[123] Thus, levosimendan is not recommended in septic shock.
Biomarkers to Guide Vasopressor Selection
Predictive biomarkers mark response to drugs and could improve the clinical efficacy
and safety of vasopressors in vasodilatory shock by improving patient selection for
therapy. Concentrations of proteins or drugs, RNA expression,[124]
[125] and rapid genotyping for single nucleotide polymorphisms (SNPs) could facilitate
personalized selection of vasopressor(s). Responders to norepinephrine could be identified
by a β2 SNP that marked increased mortality of septic shock.[30] Plasma angiopoietin-2[126] (selepressin-decreased plasma angiopoietin-2, a mediator of increased permeability),
leucyl/cystinyl aminopeptidase (the enzyme that catalyzes vasopressin; LNPEP) SNP
genotype (that altered vasopressin clearance and action),[32] and AVPR1a SNPs[32] could be possible predictive biomarkers of vasopressin, terlipressin, and selepressin.
Plasma levels of RAAS[127]
[128]
[129] and SNPs of angiotensin-II receptor associated protein (AGTRAP) are associated with
worse outcomes in septic shock and may be biomarkers for angiotensin II use.[31]
Corticosteroids and Their Interaction with Vasopressors
Corticosteroids have been used for many years in septic shock patients who are receiving
vasopressors. Indeed, recommendations for corticosteroid use remain within the most
recent SSC guidelines.[1] Some RCTs show modest benefit of corticosteroid use while others do not.[16]
[17]
[19]
[130] A pattern emerging from these RCTs is that very severely ill septic shock patients
treated with very high doses of catecholamine vasopressors may benefit more,[16]
[17] while septic shock patients treated with modest or low doses of catecholamines do
not appear to benefit.[19]
[130]
[131] Several recent reviews identify additional reasons (such as inclusion criteria,[132] prevalence of pneumonia in the RCTs,[133] corticosteroids used[132] [e.g., use of fludrocortisone or not], corticosteroid administration regimen [continuous
infusion[19]
[133] vs. intermittent bolus],[16] differences in concomitant use of vasopressin,[133] and patient gene expression differences[124]
[125]) why there is controversy among RCTs of steroids in septic shock.[124]
[132]
[133]
[134]
Corticosteroids potentiate catecholamine signaling by increasing the number of β-adrenergic
receptors expressed on the cell surface and also by enhancing coupling of adrenergic
receptors to adenylate cyclase.[135] This likely explains the observation common to both positive and negative corticosteroid
RCTs that catecholamine vasopressor use declines following administration of corticosteroids.[17]
[19]
[130] The catecholamine-sparing effect of corticosteroids in septic shock may then reduce
catecholamine-induced adverse events, including tachyarrhythmias,[56] increased myocardial oxygen demand and myocardial ischemic events, increased whole-body
oxygen consumption, increased glycolysis leading to elevated lactate levels,[136] alterations in immune function,[137] and other effects. RNA expression profiles effectively identified patients for steroids
in septic shock in one large study.[124]
Corticosteroids may also interact with vasopressin,[138] but a recent RCT has placed doubt on the clinical relevance of such an interaction.[14] In a case–control study, septic shock patients treated with vasopressin had decreased
mortality if they were also treated with corticosteroids.[139] A subsequent retrospective analysis of the VASST trial demonstrated a statistically
significant interaction where catecholamine-treated septic shock patients also treated
with vasopressin benefitted from additional corticosteroid administration while those
patients not treated with vasopressin did not (19237882). There was a trend to increased
vasopressin levels in patients treated with corticosteroids, providing insight into
the mechanism of a potential interaction. However, there was no interaction of vasopressin
with corticosteroid treatment on mortality in the VANISH trial.[14]
Is There a Limited Role for β1-Blockers in Septic and Vasodilatory Shock?
Vasopressors play a fundamental role in the management of hypotension in septic and
vasodilatory shock. Adrenergic vasopressors such as norepinephrine, epinephrine, and
dopamine have mixed α-adrenergic and β-adrenergic effects. α-Adrenergic agonists are
vasoconstrictors which raise arterial resistance and therefore help raise low arterial
pressures in septic and vasodilatory shock. In contrast, β-adrenergic stimulation
causes smooth muscle relaxation and may reduce arterial resistance. β-adrenergic stimulation
increases the rate of glycolysis which contributes to lactate production and β-adrenergic
stimulation is generally calorigenic so that oxygen demand by all tissues increases.[136] In particular, β1-adrenergic agonists increase heart rate and myocardial oxygen
demand, which may cause or worsen myocardial ischemia and cause or worsen cardiac
arrhythmias. These concerns are borne out in several large RCTs which suggest that
vasopressors with more β-adrenergic agonist effects result in increased lactate production,
increased events potentially due to inadequate myocardial oxygen delivery in relation
to demand, and increased incidence of cardiac arrhythmias.[12]
[56]
[103] For example, the SOAP II RCT of dopamine versus norepinephrine found that use of
dopamine compared with norepinephrine infusion resulted in a doubling of supraventricular
arrhythmias, particularly atrial fibrillation.[56] New-onset atrial fibrillation is important in septic shock because it can decrease
cardiac output and is associated with a significantly increased risk of stroke[58] and death.[140] Although the evidence is not yet strong, there is no benefit of anticoagulation
to decrease stroke risk in new-onset atrial fibrillation in sepsis, and there is a
significantly increased risk of bleeding.[141]
In view of potential problems associated with β-adrenergic stimulation, the use of
β-blockers in septic shock has been proposed. A small trial involving patients with
exceptionally severe septic shock suggested that β-blocker use may improve survival.[142] An esmolol infusion was titrated to maintain a heart rate between 80 and 96 BPM
for their ICU stay.[142] The esmolol group had a lower 28-day mortality (49.4%) compared with the surprisingly
high mortality of the control group (80.5%). In a systematic review, 14 of 14 trials
found that β-blockers decrease heart rates without a decrease in blood pressure in
septic shock patients.[143] Whether β-blockers decrease mortality in septic shock is less certain due to the
small number of trials, small number of patients, heterogeneity of the reported trials,
lack of blinding, and significant asymmetry of a funnel plot suggesting potential
publication bias.[143] Nevertheless, the reported trials suggest a decrease in mortality in β-blocker-treated
patients but further trials are ongoing (https://doi.org/10.1186/ISRCTN12600919).
Vasodilatory Shock Post-Cardiovascular Surgery
After cardiovascular surgery a minority of patients develop vasodilatory shock characterized
by hypotension and low systemic vascular resistance with or without high cardiac output.
Post-cardiovascular surgery vasodilatory shock is especially common in patients on
β-blockers or ACE inhibitors. Vasopressors are the main treatment after assuring adequate
(but not excessive) volume status.[105] If hypotension persists after adequate volume resuscitation, norepinephrine is the
vasopressor of first choice to increase blood pressure, vital organ perfusion while
limiting renal dysfunction.[144]
Vasopressin has also been effective in vasodilatory shock after cardiovascular surgery.
Landry's group[98]
[145]
[146]
[147]
[148] made two seminal discoveries: first, there is a relative vasopressin deficiency
post-cardiovascular surgery and second, there are short-term benefits of vasopressin
infusion including increasing MAP and decreasing norepinephrine dose requirements.
There were several trials of vasopressin versus norepinephrine post-cardiovascular
surgery but they were underpowered to evaluate patient-centered outcomes such as mortality
and organ dysfunction.[98]
[145]
[146]
[147]
[148]
[149]
There was a boost to the evidence of vasopressin after cardiovascular surgery in 2019.
A single center blinded RCT (VANCS; n = 300) in Brazil of vasopressin versus norepinephrine in vasodilatory shock post-cardiovascular
surgery[105] found that vasopressin infusion significantly decreased the rates of mortality or
severe complications, the primary endpoint and decreased the frequency of atrial fibrillation,
decreased norepinephrine dose, shortened ICU stay, and decreased frequency of AKI
and need for RRT. There was no difference in 28-day mortality—the vasopressin beneficial
effect on the primary endpoint was driven by “severe complications.” Sparing of norepinephrine
dose[150] or nonhemodynamic effects of vasopressin could explain the benefits of vasopressin.
Vasopressin appears to be more beneficial in vasodilatory syndrome post-cardiovascular
surgery than in septic shock[12]
[14] for several reasons. First, the primary outcome in pivotal RCTs differed in the
VANCS RCT than in RCTs of vasopressin in septic shock—“mortality and severe complications”[105] in the former versus 28-day mortality[12] or AKI[14] in septic shock. Plasma vasopressin doses were similar in VASST,[12] VANISH,[14] and VANCS,[105] and yet peak vasopressin levels were much lower in VANCS[105] (20–25 pmol/L) than in VASST (80–100 pmol/L) or VANISH pilot[14]
[151] (300 pmol/L). This raises the hypothesis that lower vasopressin levels are optimal
in vasodilatory shock. Finally, mortality rates were high (15.9 and 15.4% at 28-day
norepinephrine vs. vasopressin) in VANCS.[105] Unfortunately, mortality rates were not reported in prior RCTs of vasopressin versus
norepinephrine for vasodilatory shock[152] post-cardiovascular surgery.[98]
[147]
[148]
[149]
Vasodilatory Shock Post-Acute Myocardial Infarction
AMI can be complicated by vasodilatory shock and presents typical features of vasodilatory
shock (hypotension, tachycardia, and low systemic vascular resistance). After assuring
adequate volume status, norepinephrine or epinephrine infusion to supplement inotropic
and/or device support is most often used and recommended for AMI complicated by vasodilatory
shock. A recent RCT of norepinephrine versus epinephrine infusion after AMI complicated
by vasodilatory shock found no differences in effects on cardiac index (primary outcome),
systemic vascular resistance, or refractory shock, but the heart rate was significantly
higher with epinephrine than with norepinephrine.[153] The higher heart rate would be detrimental in AMI complicated by vasodilatory shock
because of the risk of worsening myocardial ischemia and extending the infarct size,
again suggesting that norepinephrine is the vasopressor of first choice in AMI complicated
by vasodilatory shock.
Dopamine has more chronotropy and had a significantly higher mortality rate compared
with norepinephrine and so is not recommended in cardiogenic shock.[56]
Pediatric Vasodilatory and Septic Shock
Septic shock in children may present with vasodilatation or myocardial failure in
varying combinations and may change over time depending on organisms, host characteristics,
and response to therapy. Thus recent adult definitions of septic shock are difficult
to apply in pediatric populations.[154] In most cases it is reasonable to begin vasoactive agents after 40 to 60 mL/kg of
fluid resuscitation if normal perfusion is not restored or signs of fluid overload
is present.[155] As suggested by the World Health Organization, early administration of vasopressors
coupled with smaller fluid boluses (10–20 mL/kg over 30–60 minutes) is reasonable
in resource-poor areas without mechanical ventilatory support or RRT to treat iatrogenic
fluid overload.[156] In the United States, early vasopressor use has been associated with shorter hospital
and ICU length of stay.[157]
[158]
The choice of vasoactive agents in children is not guided by robust evidence. For
instance, no studies directly compare epinephrine with norepinephrine although these
are two of the most commonly used vasoactive drugs. Epinephrine has been compared
with dopamine in two RCTs in children with fluid-resistant septic shock.[159]
[160] Across both studies, epinephrine was associated with a lower risk of mortality (RR:
0.63; 95% CI: 0.40–0.99) and more organ failure-free days among survivors by day 28.
A recent RCT found that epinephrine (0.2–0.4 μg/kg/min) and dopamine (10–20 μg/kg/min)
had comparable efficacy and safety in neonatal septic shock.[161] Norepinephrine has not been studied in children with septic shock, but in an RCT
of norepinephrine versus saline in sedated, mechanically ventilated children, mortality
was similar between groups (RR: 0.50; 95% CI: 0.10–2.43), but the norepinephrine group
had higher urine output (p = 0.016) and blood pressure (p = 0.04), suggesting improved perfusion relative to saline.[162]
Epinephrine is more commonly used in children than in adults. Epinephrine or norepinephrine
is the preferred first-line vasoactive drug in children; however, dopamine may be
used as the first line if neither epinephrine nor norepinephrine is available. All
may be administered through a peripheral vein (or intraosseous, if in place) if central
venous access is not readily accessible.
Vasopressin-receptor agonists (vasopressin or terlipressin) have been studied in three
RCTs in children. Vasopressin was compared with saline in one RCT of children with
vasodilatory shock[163] and in one study of children with severe lung disease.[164] Terlipressin was compared with usual care in children with septic shock.[165] The mortality rate (RR: 1.14; 95% CI: 0.80–1.62) and ischemic events (RR: 1.56;
95% CI: 0.41–5.91) were higher with vasopressin/terlipressin although not statistically
significant. There were fewer vasoactive-free days with vasopressin (median: 25.2
days in AVP [interquartile range, IQR: 0.0–28.3]) versus controls (median: 27.5 days
[IQR: 23.1–28.9]).
There are no RCTs of inodilators (including milrinone, dobutamine, or levosimendan)
in children with septic shock with persistent hypoperfusion and cardiac dysfunction.
Improvement in cardiac output with the addition of inodilators was reported in two
children.[166] There was improved core-to-peripheral temperature gradient, with stable blood pressure
and no change in acidosis in a case series of 10 children with meningococcal septic
shock treated with milrinone.[167] Despite scant evidence inodilators are frequently used in children with septic shock
who have evidence of persistent hypoperfusion and cardiac dysfunction despite other
vasoactive agents, especially in a pediatric ICU with advanced hemodynamic monitoring
available.
Directions for Future Research
Directions for Future Research
Regarding vasopressors, one could ask how much fluid defines “nonresponsive,” how
to identify fluid nonresponsive patients, how to predict vasopressor responders, how
to de-resuscitate, which patients need inotropic therapy, are certain combinations
of vasopressors better than others, and who to select for short-acting β1-blockade
with esmolol.
Patients may be less reactive to one vasopressor (e.g., vasopressin[32]) but more reactive to another(s) (e.g., norepinephrine[30]) in patients who have septic shock because of differences in host genotype,[30]
[31]
[32] variable organ-specific receptor expression and downregulation in different tissues,[29] and native plasma norepinephrine, epinephrine, vasopressin, and angiotensin II concentrations
prior to treatment ([Fig. 1]). Discovery and validation of biomarkers that predict response to vasopressors would
enable precision vasopressor therapeutics.[168]
[169] Perhaps patients will be started on several vasopressors with complementary mechanisms of action.[170] De-resuscitation to limit cumulative vasopressor toxicity deserves greater emphasis.[171] Better assessment of volume status is a priority.[61] Vitamin C has received attention[172] but a recent RCT found no benefit on cardiovascular or respiratory dysfunction in
sepsis.[173] Two other RCTs of hydrocortisone, ascorbic acid, and thiamine (HAT and VITAMINS)—three
readily available inexpensive agents—are underway.[174]
[175] The VITAMINS RCT of vitamin C, thiamine, and hydrocortisone versus hydrocortisone
alone as negative for the primary outcome, time alive and free of vasopressors up
to day 7.[176] Machine learning for earlier recognition[177] and better selection of vasopressors for individual patients could improve outcomes.
Finally, closer collaboration of academia with industry could accelerate discovery
of novel potent, safer vasopressors.[178]
Conclusion and Recommendations
Conclusion and Recommendations
Clinically available vasopressors are hormones that occupy and activate relevant receptors
(adrenergic: α1, α2, β1, β2; angiotensin II: AG1, AG2; vasopressin: AVPR1a, AVPR1B,
AVPR2; dopamine: DA1, DA2) inducing vasoconstriction but commonly have adverse effects.
Norepinephrine is the first-choice vasopressor in vasodilatory shock after adequate
volume resuscitation. Vasopressin or epinephrine may be added to norepinephrine-refractory
patients. Angiotensin II may be indicated for early resuscitation of profoundly hypotensive
patients. Vasopressors may decrease ventricular contractility, so an inotropic agent
(e.g., dobutamine) may be added. Future strategies could include initiation of several
vasopressors with complementary mechanisms of action titrated according to response
to each vasopressor. Predictive biomarkers would facilitate selection of patient-specific
vasopressors. Novel vasopressors may emerge with fewer adverse effects.
