Pharmacology

Hypothalamic–pituitary– Negative feedback control of hypothalamic–pituitary

adrenal function: anaesthetic axis

implications
Hypothalamus
Negative feedback

Releasing
Justine K Elliott factor
George M Hall Anterior
pituitary
Posterior pituitary
Stimulatory
Abstract
hormone
The hypothalamic–pituitary–adrenal axis is activated by surgery, trauma Negative feedback
and critical illness. Anaesthesia can suppress adrenocortical secretion
either by an effect at the hypothalamus (e.g. by a decrease in neural
input with regional anaesthesia) or by a direct effect on the adrenal
Stimulatory hormone
cortex (e.g. by etomidate). For patients undergoing routine surgery an
released into
increase in cortisol secretion is unnecessary; uneventful recovery occurs systemic circulation
in the presence of circulating cortisol concentrations within the normal
range. Patients taking corticosteroids for many medical conditions often
present for surgery. A rationale for steroid supplementation based on
physiological principles and clinical evidence is presented in this article. Target gland
The use of steroids in critically ill patients remains contentious, with
little agreement about what is an appropriate circulating cortisol concen-
tration and the interpretation of the results of the short Synacthen test. Active hormone
Negative feedback
Keywords cortisol; etomidate; sepsis; surgery
Target tissue

Reproduced from Ashton N. Pituitary, adrenal and thyroid dysfunction. Anaesthesia and
Intensive Care Medicine 2005; 6: 346−49.

The hypothalamic–pituitary axis is the major neuroendocrine

organ of the body. Regulation of hormone secretion is under-
taken by feedback loops, both positive and negative. A typical
pathway involves secretion of a releasing factor from the hypo-
thalamus, such as a corticotrophin-releasing factor or hormone
(CRF or CRH), which stimulates the anterior pituitary to secrete
adrenocorticotropic hormone (ACTH) into the circulation. In
turn, this acts on the target organ, the adrenal cortex, to secrete
cortisol. An increase in the circulating concentration of cortisol
acts on the hypothalamus and pituitary to inhibit release of CRF
and ACTH, respectively, so restoring circulating cortisol levels
to normal physiological values (Figure 1). The feedback control
mechanisms fail in surgery, trauma and critical illness, resulting
in abnormal endocrine function. It has been postulated that cor-
rection of such hormonal disturbances may improve outcome,
but at present there is little evidence to support this view.
Justine K Elliott, FCARCSI, is a Specialist Registrar training in South-West
Thames, London, UK. She qualified at St George’s Hospital, London
and is persuing a career in anaesthetics with a special interest in
neuroanaesthesia and obstetric anaesthesia.
George M Hall, FRCA PhD, is Professor of Anaesthesia and St George’s
Hospital, University of London, UK. His main research interests are
in the area of endocrine and metabolic problems associated with
anaesthesia.
Figure 1
The hypothalamus is part of the anterior diencephalon and con-
tributes to the floor and lateral wall of the third ventricle. There are
vascular connections between the hypothalamus and the anterior
lobe of the pituitary, the portal hypophysial vessels, whereas the pos-
terior pituitary is linked to the hypothalamus by neural connections,
the hypothalamohypophysial tract, arising from cell bodies in the
supraoptic and paraventricular nuclei. The key difference between
the anterior and posterior pituitary results from their embryological
origins. The anterior pituitary is derived from Rathke’s pouch, an
evagination from the roof of the pharynx, while the posterior lobe
arises from an evagination of the third ventricle. Other than neu-
roendocrine control, the hypothalamus has important functions in
temperature regulation, appetite control and sexual behaviour.
The anterior pituitary secretes six hormones in response to
­releasing factors secreted into the hypophysial vessels in the
­hypothalamus:
• adrenocorticotropic hormone (ACTH)
• growth hormone (GH)
• prolactin (PRL)
• thyroid-stimulating hormone (TSH)
• luteinizing-stimulating hormone (LSH)
• follicle-stimulating hormone (FSH).
ACTH, GH and prolactin are polypeptides, whereas TSH, LSH and
FSH are glycoproteins. The posterior pituitary secretes only two hor-
mones, arginine vasopressin (AVP) and oxytocin. These ­hormones

ANAESTHESIA AND INTENSIVE CARE MEDICINE 9:10 454 © 2008 Elsevier Ltd. All rights reserved.

are synthesized in the cell bodies in the supraoptic and paraven-
tricular nuclei and transported down the axons of the neurons to
the posterior lobe. Oxytocin and AVP are, therefore, typical neural
hormones. The intermediate lobe of the pituitary is rudimentary.
ACTH is secreted in irregular bursts throughout the day, with
the greatest release occurring in the morning. It is a single-chain
polypeptide of 39 amino acids and is derived from a much larger
precursor molecule pro-opiomelanocortin (POMC). Like many
polypeptide hormones it has a short half-life of about 10 min-
utes. ACTH acts on the adrenal cortex to increase the release of
cortisol and aldosterone. Most of the synthesis of cortisol occurs
in the zona fasciculata of the adrenal cortex, with the outer zona
glomerulosa releasing mainly aldosterone and the zona reticu-
laris mainly androgens. Angiotensin II also stimulates the adre-
nal cortex to produce aldosterone (Figure 2).
Adrenocortical hormones and the gonadal hormones are
derived from cholesterol. The basic structure is a four-ringed
nucleus, the steroid nucleus, with different side-chains to dis-
tinguish each compound. Cortisol is the archetypal glucocorti-
coid hormone because of its effect on glucose metabolism, and
aldosterone is a typical mineralocorticoid as a consequence of
its effect on increasing reabsorption of sodium and water from
the nephron. Glucocorticoids have effects on protein and lipid
metabolism in addition to their actions in increasing circulat-
ing glucose. Secretion of cortisol shows a classic diurnal rhythm
with some episodic secretion within the underlying circadian
framework. There is typically a surge in early morning secre-
tion, with levels maintained during normal waking hours and a
decline during sleep to a nadir between 24:00 and 04:00 hours.
Total secretion of cortisol in an unstressed individual is about
25 mg/24 hours. This hormone has a longer half-life (about 90
minutes) than the peptide hormones and circulating cortisol is
mostly bound to cortisol-binding globulin. Typical circulating
concentrations are in the range 300–500 mmol/litre.
Inhibitors of aldosterone production
Mineralocorticoid pathway
ACTH + 17α-Hydroxlase
Cholesterol Pregnenolone
Angiotensin II +
Trilostane – 3β-Dehydrogenase
17α-Hydroxlase
Progesterone
21β-Hydroxylase
17α-Hydroxlase
11-Deoxycorticosterone
Etomidate – 11β-Hydroxylase
17α-Hydroxlase
Corticosterone
Angiotensin II + 18-Hydroxylase
Aldosterone
Figure 2
ANAESTHESIA AND INTENSIVE CARE MEDICINE 9:10
Pharmacology
Effects of surgery and anaesthesia
Surgical trauma evokes a rapid secretion of ACTH which is fol-
lowed by an increase in circulating cortisol concentrations. In
major surgery the amount of ACTH released is far greater than
that necessary to produce the maximal output of cortisol. Ini-
tially, there is also a failure of the usual feedback control mecha-
nism; the increased cortisol secretion fails to inhibit further ACTH
secretion. The reason(s) for this are not known. Circulating cor-
tisol concentrations may remain increased for several days after
major surgery, although not at the values seen immediately after
surgery (600–700 mmol/litre rather than 1000–1200 mmol/litre).
It is probable that the function of the cortisol response to surgery
is to limit the size and duration of the inflammatory changes that
occur in response to tissue damage. The use of etomidate, which
suppresses cortisol synthesis (see below), is associated with a
transient increase in inflammatory markers such as interleukin
6. In contrast, glucocorticoids inhibit cytokine gene expression
as part of their profound anti-inflammatory effects.
Activation of the hypothalamic–pituitary–adrenal axis in
response to surgery may be modified by anaesthesia. Regional
anaesthesia can prevent an increase in ACTH and cortisol secre-
tion but only if autonomic afferent fibre activity is blocked as
well as somatic afferent fibre activity. There are a limited num-
ber of operative sites at which this can be achieved: pelvis, limb
and eye. For pelvic surgery an extensive afferent blockade from
dermatomes T4 to S5 is necessary to prevent pituitary hormone
secretion. This upper dermatomal limit is often higher than
required for surgery, but is essential if sympathetic afferent
blockade is to be achieved. It is presumed that complete afferent
blockade of the surgical site markedly decreases the neural input
to the hypothalamus.
The release of CRH in the hypothalamus is influenced by a
variety of neurotransmitters. It is not surprising, therefore, that
Glucocorticoid pathway Androgen pathway
17α-Hydroxypregnenolone Dehydroepiandrosterone
3β-
3β-Dehydrogenase
Dehydrogenase
17α-Hydroxyprogesterone Androstenedione
21β-Hydroxylase
11-Deoxycortisol
Etomidate – 11β-Hydroxylase
Hydrocortisone (cortisol)
455 © 2008 Elsevier Ltd. All rights reserved.

several anaesthetic drugs have been shown to inhibit ACTH
release during surgery. Large doses of μ-opioid analgesics have
been shown repeatedly to block the ACTH/cortisol response
to surgery, but only at the expense of postoperative respira-
tory depression. Similarly, large doses of midazolam have been
found to partially inhibit ACTH secretion, presumably as a conse-
quence of activating γ-aminobutyric acid (GABA) receptors in the
hypothalamus. Although this is unlikely to be relevant clinically
when midazolam is used for short-term sedation, suppression of
cortisol will occur if it is infused for several days. The α2-agonist
dexmedetomidine has an interesting biphasic effect on hypo-
thalamic–pituitary function. It decreases ACTH secretion while
stimulating GH release. In addition to the opioid receptors, α-
adrenoceptors and GABA receptors, other neurotransmitters that
have been shown to be involved in controlling CRH release in
animals include acetylcholine and 5-hydroxytryptamine.
Etomidate achieved notoriety in the early 1980s when used
for prolonged sedation in critically ill patients in intensive care
units. Mortality was increased significantly and this was shown
to be associated with severe adrenocortical suppression. Subse-
quently, etomidate was withdrawn in many countries, although
it is still available for use as a single-dose induction agent in
UK. The mechanism underlying the adrenal suppression has
been well described; inhibition of 11β-hydroxylase with a lesser
effect on other β-hydroxylases and cholesterol cleavage enzyme
(Figure 2). Other commonly administered induction agents such
as propofol and thiopentone are approximately 1000 times less
potent than etomidate in inhibiting 11β-hydroxylase. Etomidate
is an imidazole derivative, and it is this moiety that is respon-
sible for enzyme inhibition. Indeed, etomidate is such a potent
inhibitor of adrenal steroidogenesis that it has been used to treat
inoperable malignant adrenocortical tumours. A low-dose infu-
sion markedly decreased cortisol secretion with only minimal
sedation.
In spite of the profound endocrinological effects of etomidate
this compound is still used for induction of anaesthesia. Because
of its safety profile – a 30-fold difference between the anaesthetic
dose and the lethal dose – and the minimal depressant effects
Steroid treatment regimens
Patients currently taking < 10 mg/day
steroids (prednisolone)
> 10 mg/day
High-dose immunosuppression
Patients stopped taking < 3 months
steroids
> 3 months
HPA, hypothalamic–pituitary axis
Table 1
ANAESTHESIA AND INTENSIVE CARE MEDICINE 9:10
Pharmacology
on the cardiovascular system, etomidate is often administered to
physiologically unstable patients undergoing elective surgery, for
example major vascular surgery. It has been shown repeatedly
that a single induction dose of etomidate inhibits cortisol and
aldosterone secretion for about 8 hours. Circulating cortisol con-
centrations are usually in the low–normal range during this time
(250–300 mmol/litre). Despite the absence of the typical cortisol
response to surgery these patients show no increase in mortality
or major morbidity. The inference, therefore, is that, for patients
undergoing routine elective surgery, cortisol is necessary only in
normal concentrations at least intraoperatively and for the first
hours after surgery.
Therapeutic use of steroids and anaesthesia
Corticosteroids are used widely in medicine, and after prolonged
administration there is failure of endogenous cortisol secretion as
a result of the negative feedback on ACTH and CRH (Figure 1).
It is now at least half a century since the notion was introduced
that patients taking steroids needed large doses of steroid cover
to survive surgery. During this time many patients have received
200–300 mg hydrocortisone per 24 hours for 2–3 days to prevent
the possibility of adrenal insufficiency. This practice is illogical
and unnecessary.
There are two principal arguments against the administration
of large doses of steroids perioperatively. First, work with etomi-
date has shown that only normal circulating values of cortisol
and aldosterone are necessary for routine surgery. Furthermore,
primate studies have confirmed that the normal daily cortisol
output, 25 mg/day in man, is sufficient for major upper abdomi-
nal surgery. Second, even if the anaesthetist wished to mimic the
usual cortisol response to surgery, the total secretion in the first
24 hours after surgery rarely exceeds 100 mg. Thus, the current
recommendations for steroid cover are based on these underly-
ing physiological principles (Table 1). The key points to note are
that if the patient is taking less than 10 mg prednisolone/day
(or an equivalent dose of another steroid) then additional hydro-
cortisone is unnecessary; if the patient has not taken steroids in
Assume normal HPA response Additional steroid cover not required
Minor surgery 25 mg hydrocortisone at induction
Moderate surgery Usual preoperative steroids + 25 mg
hydrocortisone at induction + 100 mg/day
for 24 hours
Major surgery Usual preoperative steroids + 25 mg
hydrocortisone at induction + 100 mg/day
for 48–72 hours
Give usual immunosuppressive doses during perioperative period
Treat as if taking steroids
No perioperative steroids necessary
456 © 2008 Elsevier Ltd. All rights reserved.
 Pharmacology
the past 3 months, again hydrocortisone is not required periop-
eratively; and if high-dose steroids are being given for immuno-
suppression then these must be maintained perioperatively.
Old regimens of 200–300 mg/day of hydrocortisone are of
historical interest only. The excessive administration of steroids
perioperatively is associated with the following complications:
hypertension from fluid retention, hyperglycaemia, immuno-
suppression, failure of wound healing, gastric erosions and psy-
chological disturbances. Perioperative hypotension in a patient
taking steroids is not uncommon. It has been shown that this is
only rarely associated with low circulating cortisol values and is
commonly the result of hypovolaemia. If hypotension persists
after fluid administration then 25 mg intravenous hydrocortisone
may be given. Ideally, a blood sample should be taken for corti-
sol estimation before the hydrocortisone bolus so that suspected
steroid deficiency can be confirmed or refuted.
Steroids and the critically ill
The ACTH/cortisol response to elective surgery has been well
defined and the rationale for hydrocortisone supplementation in
surgical patients taking glucocorticoids is now well established.
In critically ill patients in the intensive care unit the appropri-
ateness of the observed cortisol changes is unclear. Plasma cor-
tisol values tend to be increased in the early stages of critical
illness and correlate with the severity of other physiological
scoring systems and hence predict outcome. However, there is
much confusion about the magnitude of the cortisol response
necessary to maintain normal cardiovascular function in such
patients. Attempts have been made to assess adrenocortical
function by using 250 μg of synthetic ACTH intravenously to
stimulate cortisol production. Blood samples are collected before
and 30 and 60 minutes after ACTH administration for cortisol
determination (short Synacthen test; Alliance Pharmaceuticals
Ltd, Chippenham, UK). This test was devised originally to assess
adrenocortical responsiveness in endocrinological patients who
were taking, or had taken, glucocorticoids. Even in this rela-
tively straightforward clinical area interpretation of the results
of the short ­Synacthen test has been difficult. There have been
ANAESTHESIA AND INTENSIVE CARE MEDICINE 9:10
disputes about whether to use an absolute threshold value after
ACTH, such as 550 mmol/litre cortisol, or to use an incremental
increase. The problem with the latter approach is that the size
of the increment is inversely related to the magnitude of the
concentration of cortisol before ACTH.
It is not surprising that attempts to assess adrenocortical func-
tion in patients in intensive care units have proved difficult.
Nevertheless, there may be a group of patients who have adre-
nocortical insufficiency and in whom outcome may be improved
by the administration of hydrocortisone. Glucocorticoids have a
key role in maintaining the responsiveness of the vasculature to
catecholamines so that adrenocortical insufficiency may result in
the failure to respond to infusions of adrenergic drugs. Steroids
have been given to critically ill patients for several decades, par-
ticularly those with sepsis, and enthusiasm for this approach has
waxed and waned. Recommendations produced by the Surviving
Sepsis Campaign in 2004 stated that patients with septic shock
who, despite adequate fluid replacement, required inotropes
to maintain blood pressure should be given a ‘low-dose’ daily
replacement with 200–300 mg hydrocortisone. However, these
recommendations may be revised as a recent large randomized
control trial found no effect on mortality in patients with sepsis
given hydrocortisone 200 mg/24 hours and an increase in infec-
tions and hypertension in the steroid-treated patients. It is likely
that there will continue to be confusion about the role of hydro-
cortisone in critically ill patients until the pathophysiological
changes in hypothalamic–pituitary–adrenal function have been
elucidated. ◆
Further reading
Dellinger RP, Carelet JM, Masur H, et al. Surviving Sepsis Campaign
guidelines for management of severe sepsis and shock. Crit Care
Med 2004; 32: 858–73.
Desborough JP. The stress response to trauma and surgery. Br J
Anaesth 2000; 85: 109–17.
Nicholson G, Burrin JM, Hall GM. Peri-operative steroid
supplementation. Anaesthesia 1998; 53: 1091–104.
457 © 2008 Elsevier Ltd. All rights reserved.