Skip to content

8.5 — The Control of Breathing

Hold your breath and the urge to breathe builds until it becomes irresistible. That urge is not caused by lack of oxygen. It is caused by rising carbon dioxide, and the two can be separated experimentally — breathing pure oxygen and then holding your breath produces the same urge on almost the same schedule, even though blood oxygen is fine.

Your brain monitors carbon dioxide continuously and oxygen only as a backup. Understanding why explains hyperventilation, altitude sickness, sleep apnoea, why a swimmer can black out underwater, and why opioid overdose kills.

Where the rhythm comes from

The respiratory centres are in the brainstem — in the medulla and pons — and they generate the breathing rhythm automatically.

The medullary centres contain the pattern generator. A group of neurons called the pre-Bötzinger complex appears to be the core oscillator: isolated slices of it continue producing a rhythm in a dish, which is about as direct a demonstration as physiology gets.

Two groups then shape the output. The dorsal respiratory group drives quiet inspiration. The ventral respiratory group is mostly quiet during quiet breathing and is recruited for forced breathing, driving both active inspiration and active expiration.

The pontine centres fine-tune the switch from inspiration to expiration, smoothing the transition.

And breathing is under dual control, which is unusual.

Automatic control from the brainstem, running whether you are awake, asleep or anaesthetised.

Voluntary control from the cerebral cortex, allowing you to speak, sing, hold your breath, blow out a candle, or breathe in a deliberate pattern.

The two pathways run down the spinal cord separately, and this is not a technicality — it produces a real and striking clinical picture.

Ondine's curse (congenital central hypoventilation syndrome) is a failure of automatic control with voluntary control intact. The person breathes normally while awake and stops breathing when they fall asleep. It is named after a myth in which a water nymph curses a faithless lover so that he must remember to breathe. Treatment is lifelong ventilatory support during sleep, and it is one of the clearest demonstrations that the two control systems are genuinely separate.

The reverse also occurs: certain brainstem strokes abolish voluntary control while leaving automatic breathing intact, so the person breathes but cannot deliberately hold their breath or coordinate speech.

What is being measured

Three sets of sensors, and their relative importance is the key to the whole chapter.

Central chemoreceptors — carbon dioxide, indirectly

On the surface of the medulla, and they are responsible for about 80 percent of the drive to breathe.

They do not detect carbon dioxide directly. They detect the pH of the cerebrospinal fluid.

The mechanism is worth following, because it explains their behaviour. The blood–brain barrier (Chapter 11.10) is impermeable to hydrogen ions and bicarbonate, but freely permeable to carbon dioxide.

So CO₂ crosses into the cerebrospinal fluid, and there it reacts with water:

\mathrm{CO_2} + \mathrm{H_2O} \rightarrow \mathrm{H^+} + \mathrm{HCO_3^-}

The hydrogen ions produced lower the pH, and the receptors detect that.

Two consequences follow directly.

Cerebrospinal fluid has very little protein and therefore very little buffering capacity, so a given amount of CO₂ produces a larger pH change there than in blood. The system is exquisitely sensitive — a rise of just 2 to 3 mmHg in arterial CO₂ measurably increases ventilation.

And a metabolic acidosis in the blood does not stimulate these receptors directly, because the hydrogen ions cannot cross the barrier. It stimulates the peripheral receptors instead.

Peripheral chemoreceptors — oxygen, and the backup

In the carotid bodies at the carotid bifurcation, and the aortic bodies on the aortic arch.

They respond to low oxygen, and also to high CO₂ and low pH.

But the oxygen response is not linear, and this is the crucial point. They barely respond until the arterial oxygen falls below about 60 mmHg, and then they fire increasingly steeply.

That threshold is exactly the shoulder of the dissociation curve (Chapter 8.4). Above 60 mmHg, saturation is over 90 percent and oxygen delivery is essentially fine, so there is nothing to respond to. Below it, saturation falls fast, and the receptors kick in hard.

So oxygen is a backup sensor with a threshold, and carbon dioxide is the continuous one.

The carotid bodies are also astonishing in one respect: they have the highest blood flow per gram of any tissue in the body, about 2 litres per 100 grams per minute — some 40 times higher than brain. They need it, because they must sample arterial blood before it has had any oxygen extracted from it.

Mechanoreceptors and others

Stretch receptors in the airways trigger the Hering–Breuer reflex, inhibiting inspiration when the lungs are strongly inflated. In adults this only operates at large tidal volumes — over about 1.5 litres — so it is a protective limit rather than part of normal breathing. It is more active in infants.

Irritant receptors trigger coughing, sneezing and bronchoconstriction.

J receptors in the alveolar walls respond to fluid in the interstitium, producing rapid shallow breathing. This is thought to be part of why pulmonary oedema and pneumonia make people breathe fast.

Joint and muscle receptors contribute to the immediate increase in breathing at the start of exercise.

And higher centres matter more than textbooks suggest. Pain, emotion, anxiety and temperature all change breathing directly, through connections from the limbic system and hypothalamus.

Exercise: the unsolved bit

During exercise, ventilation increases up to 20-fold, and it matches carbon dioxide production so precisely that arterial CO₂ stays essentially constant up to moderate intensity.

And that is the puzzle. If CO₂ is the signal and CO₂ does not change, what is driving the increase?

The honest answer is that it is not fully settled, and it is worth saying so rather than presenting a tidy story. The contributions appear to be:

Anticipatory drive from the motor cortex — the same signal that commands the muscles also drives breathing. Ventilation rises before the exercise starts, which cannot be a response to anything metabolic.

Feedback from moving joints and muscles — passive movement of a limb increases ventilation.

Increased sensitivity of the chemoreceptors during exercise, so small oscillations in CO₂ produce larger responses.

Potassium released from working muscle, which stimulates the carotid bodies.

At high intensity, above the point where lactate accumulates, the acid produced stimulates the peripheral chemoreceptors directly and ventilation rises out of proportion — which is measurable and is used in fitness testing as the ventilatory threshold.

Hyperventilation

Breathing more than metabolic need requires, which blows off carbon dioxide and raises blood pH — respiratory alkalosis.

The symptoms are specific and worth knowing, because they are frequently mistaken for something worse.

Tingling around the mouth and in the fingers, and muscle spasms of the hands and feet. The reason is calcium: alkalosis increases the binding of calcium to albumin, so the free ionised calcium falls, and nerves become hyperexcitable even though total calcium is unchanged.

Light-headedness and visual disturbance, from a different mechanism: low carbon dioxide constricts cerebral blood vessels. A drop of 1 kPa in CO₂ reduces cerebral blood flow by roughly 30 percent, and vigorous hyperventilation can halve it.

Chest tightness, from the sustained effort of the respiratory muscles — which is often interpreted as a heart problem, which increases the anxiety, which increases the hyperventilation.

Treatment is reassurance and slowed breathing. Breathing into a paper bag is no longer recommended — it has caused deaths in people whose breathlessness was actually from a heart attack, asthma or pulmonary embolism, and rebreathing dangerously low-oxygen air in those situations is harmful. The correct approach is to make the diagnosis first, and then to coach slow breathing.

And this matters because hyperventilation is a diagnosis of exclusion. Diabetic ketoacidosis, pulmonary embolism, sepsis, aspirin overdose and asthma can all present as a young person breathing fast and feeling anxious.

One specific and preventable danger: hyperventilating before breath-hold swimming. Deliberately over-breathing lowers CO₂ so the urge to breathe is delayed — but it does not increase oxygen stores at all, because haemoglobin was already nearly saturated. So the swimmer uses up their oxygen while the warning signal is still absent, and loses consciousness underwater without ever feeling short of breath. This is shallow water blackout, it kills competent swimmers every year, and it is entirely avoidable by not hyperventilating before a breath-hold. Chapter 23.10.

Altitude

At 5,500 metres, atmospheric pressure is about half sea level, so the partial pressure of oxygen is halved even though the air is still 21 percent oxygen.

The immediate response is hyperventilation, driven by the peripheral chemoreceptors once oxygen falls below their threshold.

And that response is self-limiting, which is the interesting part. Hyperventilating blows off CO₂, causing alkalosis, and the central chemoreceptors then inhibit breathing. The two systems oppose each other, and ventilation settles at less than it should be.

Acclimatisation over days resolves the conflict. The kidney excretes bicarbonate, correcting the alkalosis, so the central receptors stop objecting and ventilation can rise further. This takes several days, and it is why gradual ascent works and rapid ascent does not.

Acetazolamide accelerates it by causing bicarbonate loss in the urine directly, producing a mild metabolic acidosis that permits greater ventilation. It is not a symptom treatment — it accelerates the actual adaptation, which is why it is taken starting a day or two before ascent.

Other adaptations follow over weeks: raised 2,3-BPG within a day or two (Chapter 8.4), raised erythropoietin and red cell count over weeks, and increased capillary density and mitochondrial content over months.

Altitude illness comes in three forms, and the rule underlying all of them is simple.

Acute mountain sickness — headache, nausea, fatigue, poor sleep, above about 2,500 metres. Common and usually self-limiting.

High altitude pulmonary oedema (HAPE) — fluid in the lungs, from uneven hypoxic pulmonary vasoconstriction over-pressurising the regions that remain open. Breathlessness at rest, cough, pink frothy sputum.

High altitude cerebral oedema (HACE) — brain swelling. Confusion, unsteadiness, and a characteristic inability to walk heel to toe in a straight line.

Both severe forms are fatal without descent, and both are treatable by descent. The one rule that matters is: descend. Oxygen, dexamethasone and nifedipine help, and none of them substitutes for going down. A descent of 500 to 1,000 metres frequently produces dramatic improvement within hours.

And the prevention rule is equally simple: ascend slowly and do not sleep more than 300 to 500 metres higher than the previous night above 3,000 metres, with a rest day every 3 to 4 days. Climb high, sleep low.

Sleep apnoea

Breathing changes during sleep in everyone: the drive falls, CO₂ rises by 3 to 7 mmHg, and airway muscle tone relaxes.

Obstructive sleep apnoea — the upper airway collapses repeatedly during sleep, despite continued respiratory effort. The person stops breathing, oxygen falls, CO₂ rises, and they briefly wake enough to restore airway tone — often hundreds of times a night, without any memory of it.

Affects perhaps 10 to 25 percent of middle-aged adults to some degree, and most are undiagnosed.

Risk factors: obesity, particularly neck circumference; male sex; increasing age; a small or set-back jaw; large tonsils; alcohol and sedatives before bed.

And the consequences go far beyond tiredness, which is why it matters more than its reputation suggests. Untreated obstructive sleep apnoea causes hypertension, and is independently associated with atrial fibrillation, heart failure, stroke, type 2 diabetes and depression. The repeated surges of sympathetic activity with each arousal are a plausible mechanism.

It also causes a two to threefold increase in road traffic accident risk, and this has legal implications for drivers in most countries.

Treatment is genuinely effective, which is the encouraging part. CPAP — continuous positive airway pressure — splints the airway open with a stream of pressurised air through a mask. It abolishes the apnoeas immediately, and people frequently describe the first night as the best sleep they have had in years. Adherence is the main limitation.

Mandibular advancement devices work for milder cases. Weight loss is effective, and even a 10 percent reduction produces measurable improvement. Avoiding alcohol in the evening helps. And in children, removing enlarged tonsils and adenoids is often curative (Chapter 8.1).

Central sleep apnoea is different — the effort itself stops, because the brainstem does not send the signal. Seen in heart failure, after stroke, at altitude, and with opioids.

And Cheyne–Stokes breathing — the cycle of deepening then fading breaths described in Chapter 4.7 — is an oscillating control loop, caused by a delay between the lungs changing blood gases and the brain sensing it.

Opioids

Opioids suppress the respiratory centre directly, reducing both the rate and the response to carbon dioxide. The person does not feel short of breath, because the sensor itself has been silenced.

This is the mechanism of death in opioid overdose: breathing slows and becomes shallow, CO₂ rises, oxygen falls, and there is no distress signal to prompt any response.

Naloxone reverses it within minutes by competitively displacing the opioid from its receptor (Chapter 1.3). Naloxone is safe, has essentially no effect on someone who has not taken opioids, and can be given by a bystander as a nasal spray. Community naloxone distribution programmes have prevented very large numbers of deaths, and this is one of the clearest cases where an ordinary person carrying a simple device saves lives. Chapter 23.6.

What the next page fixes

The system so far operates at sea level and at rest. Chapter 8.6 pushes it to its limits — what happens at altitude, under water, and when you hold your breath — and closes the Part with what the respiratory system can do that is genuinely remarkable.