The stress response

What it normally does

• Stress - infection, surgery, trauma, a low blood sugar - makes the hypothalamus release CRH, which makes the pituitary release ACTH, which makes the middle layer of the adrenal cortex (zona fasciculata) make cortisol [hypothalamic-pituitary-adrenal axis]. Cortisol then switches the hypothalamus and pituitary back off [negative feedback]. Normally the axis fires in pulses, peaking around waking and bottoming near midnight; in serious illness the pulses merge, the rhythm disappears, and circulating cortisol climbs broadly with severity - driven partly by ACTH and partly by slowed cortisol breakdown, so in prolonged critical illness ACTH is often low while cortisol stays high. The axis cannot tell where a steroid came from - a tablet occupies the same receptors and closes the same feedback loop as the hormone.

  Explains why a cortisol level means nothing without the time of day and how sick the patient is beside it, and why weeks of prednisolone leave an adrenal gland that has forgotten how to answer.

• Cortisol holds up the blood pressure mainly by letting other things work rather than by constricting vessels itself [permissive action]. It keeps alpha-1 adrenoceptors expressed on blood vessels and keeps their signalling coupled, so adrenaline and noradrenaline can constrict. It keeps the liver making glucose (gluconeogenesis) and blunts insulin action in muscle and fat. It lets the kidney get rid of free water, partly by restraining vasopressin. It holds inflammation in check. In pharmacological excess it does raise blood pressure in its own right, largely by spilling onto the mineralocorticoid receptor. In illness the carrier protein (cortisol-binding globulin) falls, so the free, active fraction rises more than a total cortisol measurement suggests.

  Explains the whole picture of adrenal crisis - shock that ignores fluid and vasopressors, low glucose, low sodium, fever with no source - and predicts most of the side effects of glucocorticoid drugs.

• The adrenal medulla is a sympathetic ganglion that lost its axons [chromaffin cells]. Preganglionic sympathetic fibres release acetylcholine onto nicotinic receptors and the cells squirt adrenaline (about 80%) and noradrenaline straight into the blood. Effects are on in seconds and gone in minutes: alpha-1 constricts vessels, beta-1 drives heart rate and force, beta-2 opens bronchi, dilates muscle vessels and pushes potassium into cells. Cortisol takes an hour because it works through gene transcription. The medulla also depends on the cortex - the very high cortisol in blood draining past it induces the enzyme (PNMT) that converts noradrenaline to adrenaline.

  Explains why catecholamines own the first minutes and cortisol owns the next hours, and why a noradrenaline infusion works poorly when cortisol is missing.

• The outer layer of the cortex (zona glomerulosa) makes aldosterone and is not sustained by ACTH - it answers to angiotensin II and to potassium (ACTH gives only a small, short-lived stimulus). ACTH itself is cut from a larger precursor (pro-opiomelanocortin, POMC) that also contains the melanocyte-stimulating hormone sequence, so a very high ACTH also darkens the skin.

  Explains why destroying the gland gives high potassium and pigmented skin, while switching off the pituitary with steroid tablets gives neither.

What goes wrong

• Adrenal crisis← from “Cortisol holds up the blood pressure mainly by…”

  Cortisol demand overtakes supply - known adrenal insufficiency plus gastroenteritis, sepsis or surgery, or a first presentation in someone never diagnosed. Without cortisol the vessels answer catecholamines poorly, so vasodilatory shock develops and fluid does not fix it. The liver stops making glucose. Free water is retained, so sodium falls. If the whole cortex is gone, aldosterone is gone too, so sodium is lost in the urine and potassium climbs. The immune brake lifts, so fever can appear with nothing else to explain it.

  Treated on suspicion. Take blood for cortisol and ACTH if it costs no time, then give parenteral hydrocortisone with intravenous saline, checking the glucose and correcting it if it is low. Waiting for a cortisol result is how people die of this. Around 6-8 crises occur per 100 patient-years in people already known to have adrenal insufficiency; gastrointestinal infection is the commonest trigger, and stress dosing is often either not taken or not kept down.

  You would find: Hypotension that does not respond to fluid and needs escalating vasopressor. Vomiting and abdominal pain that can pass for a surgical abdomen. Fever, drowsiness, sometimes coma. Bloods: sodium low, glucose low, potassium high if the cause is primary, and eosinophils present when acute stress should have wiped them out. Ask two questions early - is this person on steroids, and have they recently stopped any?

• Primary adrenal insufficiency (Addison disease)← from “The outer layer of the cortex (zona glomerulos…”

  The whole cortex is destroyed, so cortisol and aldosterone both go. Around 80-90% of cases in Australia are autoimmune (21-hydroxylase antibodies, often alongside autoimmune thyroid disease or type 1 diabetes); the rest is tuberculosis, bilateral haemorrhage, metastases or infiltration, adrenoleukodystrophy in young men, or drugs that block steroid synthesis. Losing aldosterone means losing sodium in the urine while holding onto potassium and hydrogen. Losing cortisol removes the brake on the pituitary, so ACTH and its POMC siblings rise several-fold and pigment the skin.

  High ACTH with low cortisol puts the fault in the gland; low ACTH with low cortisol puts it above the gland. Pigmentation and hyperkalaemia point to primary disease - neither is expected when the problem is the pituitary, because the zona glomerulosa never depended on ACTH. Prevalence is about 100-140 per million. Tuberculous adrenalitis is uncommon in Australia but remains an important cause in high-burden settings, so it is worth considering in people who have lived in or come from those settings.

  You would find: Months of fatigue, weight loss, nausea and postural dizziness, then a crisis when something else goes wrong. Pigmentation where a tan does not belong - buccal mucosa, palmar creases, old scars, pressure points. Sodium low, potassium high, mild metabolic acidosis, occasionally a raised calcium. Diagnosis is a morning cortisol with a paired ACTH, confirmed by a short Synacthen (tetracosactide) test in which cortisol fails to climb past roughly 400-550 nmol/L, the exact cut-off depending on the assay.

• Secondary adrenal insufficiency from glucocorticoid withdrawal← from “Stress - infection, surgery, trauma, a low blo…”

  Any glucocorticoid taken above replacement level for more than about three weeks feeds back on CRH and ACTH. Corticotrophs go quiet, and the zona fasciculata shrinks without ACTH to drive it. Recovery after the drug stops commonly takes 6-12 months and sometimes longer. The zona glomerulosa is untouched because it never needed ACTH, so aldosterone keeps working. Stop the drug abruptly, or leave the dose alone while pneumonia triples the demand, and there is no cortisol to meet it.

  This is now the commonest cause of adrenal insufficiency in Australia - roughly 1% of adults in Australia and comparable countries are on long-term oral glucocorticoids. Two rules fall out of the physiology: come down gradually rather than stopping abruptly, and put the dose up when sick. Tapering does not itself make the axis recover faster; it lowers the exogenous dose slowly enough that a still-suppressed axis is not caught out, and it is not needed after courses of under about three weeks. Aches and lethargy on tapering may instead be a steroid withdrawal syndrome with a perfectly normal axis; the two are separated by testing, not by how bad the patient feels.

  You would find: Someone on long-term prednisolone who becomes hypotensive, nauseated and hypoglycaemic during an illness - with a normal potassium and no pigmentation, because neither aldosterone nor ACTH is the problem. Sodium can still fall, since cortisol normally restrains vasopressin. A Cushingoid face on the same patient is a clue the axis is asleep. Non-oral routes count: high-dose inhaled fluticasone, potent topical steroid over large areas, repeated joint injections.

• Catecholamine-resistant shock in critical illness← from “The adrenal medulla is a sympathetic ganglion …”

  In prolonged septic shock the alpha-1 receptors on vascular smooth muscle are downregulated and uncoupled, nitric oxide and prostacyclin production is high, and ATP-sensitive potassium channels open and hold the muscle cell hyperpolarised. Noradrenaline binds and less happens [vasoplegia]. Cortisol is what maintains receptor number and coupling, and although total cortisol is usually high in these patients, the tissue response can still fall short of the demand [critical illness-related corticosteroid insufficiency].

  The Australian and New Zealand ADRENAL trial (about 3800 patients) gave a continuous hydrocortisone infusion in septic shock: shock resolved faster and patients came off ventilators sooner, with no difference in 90-day mortality. The French APROCCHSS trial, using hydrocortisone plus fludrocortisone, did show a 90-day mortality difference. Practice reserves corticosteroids for shock that keeps demanding vasopressor - an adjunct, never the resuscitation itself.

  You would find: A vasopressor requirement that keeps climbing in a patient who is volume-replete and whose source is controlled. There is no good test: a random cortisol and a Synacthen test both perform poorly in critical illness, because binding globulin has fallen and tissue resistance is not something a blood level measures. This is a clinical situation, not a number.

What we give, and how it works

Open a drug to see what it binds, what that does to the cell, and what you then see in the patient.