Carbon dioxide carriage
Metabolising tissue makes carbon dioxide continuously; the blood carries it to the lungs mostly as bicarbonate, and how fast you breathe sets how much stays behind — which is why PaCO2 measures ventilation, not oxygenation.
What it normally does
CO2 travels three ways: roughly 70% converted inside the red cell to bicarbonate (HCO3-), roughly 23% stuck to the terminal amino groups of haemoglobin itself (carbamino compounds), and only about 7% simply dissolved in plasma. The enzyme carbonic anhydrase in the red cell drives the conversion, and the bicarbonate is then swapped out into plasma for a chloride ion (the chloride shift).
the PaCO2 you measure on a blood gas is only the small dissolved fraction — a shallow pool that moves fast, so PaCO2 responds to a change in breathing within minutes.
That conversion is an acid reaction: CO2 + H2O gives carbonic acid, which splits into a hydrogen ion and bicarbonate. Carbon dioxide is an acid in solution.
CO2 that is not blown off is acid retained — the pH falls (respiratory acidosis). Blow off too much and the pH rises (respiratory alkalosis).
PaCO2 is set by how much CO2 the body makes divided by how much fresh air reaches the alveoli each minute (alveolar ventilation). Alveolar ventilation is minute ventilation minus the part wasted on dead space, so it can fall either because the patient breathes less or because more of each breath is wasted. Production is usually steady but rises with fever, sepsis, agitation and overfeeding. Halve alveolar ventilation and PaCO2 roughly doubles. Central chemoreceptors in the medulla sense the H+ generated as CO2 crosses into CSF, and drive breathing; the carotid bodies sense arterial H+ and PaO2.
in a patient breathing ordinary air, a high PaCO2 means alveolar ventilation is inadequate for the CO2 being produced. Usually that is the bellows failing, but it is also how severe obstructive disease raises CO2 — a big dead space means a high minute ventilation can still deliver too little fresh air to alveoli. What a high PaCO2 is essentially never explained by is a diffusion problem: CO2 diffuses too easily for that.
CO2 crosses the alveolar membrane about 20 times faster than O2, and the CO2 content of blood rises almost in a straight line with PaCO2 across the physiological range — no plateau. Haemoglobin, by contrast, is already about 97% saturated with O2 leaving a normal alveolus and cannot be filled much further.
an alveolus that is working can blow off extra CO2 for a sick neighbour, but it cannot load much extra O2 for it. Diseased lung therefore drops the O2 while the CO2 stays normal or low — type 1 failure.
What goes wrong
- Type 1 respiratory failure (low oxygen, normal or low CO2)← from “CO2 crosses the alveolar membrane about 20 tim…”
Disease fills or collapses alveoli (pneumonia, pulmonary oedema) or narrows the airways feeding them (asthma), so blood flowing past poorly ventilated or unventilated units is not oxygenated (V/Q mismatch, or shunt where there is no ventilation at all). Pulmonary embolus gets there less directly: the obstructed units become dead space, while blood diverted into the remaining lung overperfuses it and creates low V/Q regions, with reflex atelectasis adding shunt. The low O2 drives the patient to breathe harder. The healthy alveoli respond by dumping extra CO2, because the CO2 content curve has no plateau, so PaCO2 falls. They cannot return the favour for oxygen: their haemoglobin is already close to fully saturated, so the extra breathing adds almost no O2 to make up for the shunted blood.
V/Q mismatch → hypoxaemia. Compensatory hyperventilation lowers CO2 but cannot fix O2 (Hb already near-saturated). PaO2 <60 with PaCO2 normal or low.
You would find: PaO2 below 60 mmHg with a PaCO2 that is normal or low (often below 35 mmHg). The patient is visibly working — tachypnoeic, low SpO2, often anxious — and the gas does not show CO2 retention. That combination is the fingerprint of a gas exchange problem, not a pump problem.
- Type 2 respiratory failure (low oxygen with high CO2)← from “PaCO2 is set by how much CO2 the body makes di…”
Alveolar ventilation becomes inadequate. Most often the pump fails at some level: drive (opioids, sedatives, brainstem stroke), nerve and muscle (Guillain-Barre, motor neurone disease, myasthenia, exhaustion) or chest wall (severe obesity, kyphoscoliosis). Severe airways disease (COPD, severe asthma) gets there differently — dead space rises and the work of breathing rises, so even a high minute ventilation delivers too little fresh gas to alveoli, and fatigue then finishes the job. Either way CO2 accumulates in proportion to the shortfall. The rising CO2 takes up space in the alveolus and pushes alveolar O2 down (the alveolar gas equation), so in pure hypoventilation the hypoxaemia follows the hypercapnia rather than causing it.
Inadequate alveolar ventilation — pump failure or a large dead space — → raised PaCO2 with hypoxaemia. Drowsy, flap, bounding pulse. A normalising CO2 in acute asthma means the patient is tiring.
You would find: PaCO2 above 45 mmHg (many texts use above 50) with PaO2 below 60 mmHg. Look for the signs of CO2 itself: drowsiness, headache, a flapping tremor of the outstretched hands (asterixis), bounding pulse and warm hands — CO2 dilates vessels, including cerebral ones. In an exhausted asthmatic, a NORMAL PaCO2 is the alarm, because it means the hyperventilation has stopped.
- Acute versus chronic respiratory acidosis (the CO2 retainer)← from “That conversion is an acid reaction: CO2 + H2O…”
Because CO2 is an acid, a sudden rise drops the pH by roughly 0.08 for every 10 mmHg rise in PaCO2, buffered only slightly by a bicarbonate rise of about 1 mmol/L per 10 mmHg. The kidney starts retaining bicarbonate within hours, but takes about three to five days to complete the job. After that the bicarbonate is high — of the order of 3-4 mmol/L up per 10 mmHg of CO2 — and the pH has moved back towards normal (falling only about 0.03 per 10 mmHg), though compensation is rarely complete and the pH stays slightly acidaemic.
Bicarbonate is the clock. Normal HCO3- with high CO2 = acute (HCO3- up ~1 per 10 mmHg). Near-normal pH with HCO3- up ~3-4 per 10 mmHg = chronic compensation (days of renal work).
You would find: Read the bicarbonate to date the problem. High PaCO2 + low pH + normal bicarbonate = this happened today, and the patient is in trouble. High PaCO2 + near-normal pH + high bicarbonate = this is their baseline. In Australia the chronic retainer is most often severe COPD; COPD and chronic suppurative lung disease hospitalise and kill Aboriginal and Torres Strait Islander people at several times the rate of other Australians, and at younger ages.
- Oxygen-induced hypercapnia in COPD← from “PaCO2 is set by how much CO2 the body makes di…”
A high inspired oxygen concentration relaxes the hypoxic pulmonary vasoconstriction that was diverting blood away from poorly ventilated alveoli. Blood flows back into units that ventilate badly, so ventilation and perfusion match less well and effective dead space rises — more of each breath is wasted and CO2 clearance falls. At the same time, fully oxygenated haemoglobin holds CO2 less well, releasing carbamino CO2 and buffering fewer hydrogen ions (the Haldane effect), which pushes CO2 back into the dissolved compartment you measure. Reduced hypoxic drive to breathe is a real but minor third contributor.
High FiO2 in COPD raises CO2 mainly by worsening V/Q matching and by the Haldane effect, not just by knocking out hypoxic drive. Target sats 88-92%.
You would find: A COPD patient put on high-flow oxygen by reservoir mask who becomes progressively drowsy over the next half hour, with a repeat gas showing a climbing PaCO2 and falling pH. Target SpO2 is 88-92% in anyone at risk of hypercapnia, not 100%.
- Hyperventilation and respiratory alkalosis← from “That conversion is an acid reaction: CO2 + H2O…”
Over-breathing from anxiety, pain, sepsis, pulmonary embolus or salicylate poisoning strips CO2 out. Less acid means fewer hydrogen ions, so pH rises. Albumin, no longer occupied by H+, binds more calcium, and the free (ionised) calcium falls even though total calcium is unchanged. Alkalosis also constricts cerebral arteries.
Low CO2 → alkalosis → albumin grabs calcium → tetany and paraesthesia. Cerebral vasoconstriction gives the dizziness.
You would find: Tingling around the mouth and in the fingers, cramping of the hands (carpopedal spasm), light-headedness, chest tightness. Gas shows low PaCO2 with high pH. Do not diagnose panic until you have excluded pulmonary embolus and sepsis, which can present exactly this way.
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.
- Binds
- No molecular target. The alveolar-capillary partial pressure gradient for O2.
- Which does
- Raising inspired O2 raises alveolar PO2, which pushes more oxygen into capillary blood in alveoli that are still ventilated. It does little for blood passing alveoli with no ventilation at all (true shunt), beyond the small amount of extra dissolved O2 carried by the rest.
- So you see
- SpO2 and PaO2 rise. The work of breathing is barely changed and the underlying V/Q mismatch is not corrected — at high FiO2 it can be made worse.
- And the same mechanism causes
- The same rise in alveolar PO2 abolishes hypoxic pulmonary vasoconstriction and saturates haemoglobin, so in COPD it worsens V/Q matching and releases CO2 from haemoglobin (Haldane effect) — the PaCO2 climbs and the patient becomes drowsy. Washing nitrogen out of poorly ventilated units also lets them collapse (absorption atelectasis). The treatment for the O2 causes the CO2 problem.
- Handling
- Delivered concentration is only reliable with a fixed-performance device such as a Venturi mask; with nasal prongs the actual FiO2 varies with the patient's own breathing.
Catches people out: Oxygen treats a number, not a disease, and it does not lower CO2. A patient with type 2 failure whose sats look fine on oxygen can still be deteriorating from CO2 — what that needs is ventilatory support, not more oxygen.
- Binds
- Mu opioid receptors on medullary and pontine respiratory neurones (pre-Botzinger complex, Kolliker-Fuse/parabrachial region and the chemosensitive areas); it also antagonises kappa and delta receptors less potently
- Which does
- Competitive antagonism displaces the opioid, restoring the neurones' normal sensitivity to rising CO2 and H+.
- So you see
- Respiratory rate and tidal volume typically return within a minute or two of an intravenous dose, alveolar ventilation rises, PaCO2 falls and the patient rouses.
- And the same mechanism causes
- It is the same receptor everywhere else: blocking it strips away analgesia and, in a dependent patient, can precipitate acute withdrawal — pain, vomiting, agitation, sweating, tachycardia and hypertension. These follow directly from the mechanism rather than being unrelated side effects.
- Handling
- Its duration of action (roughly 30-90 minutes) is shorter than that of most opioids, especially slow-release preparations and methadone, so the patient can stop breathing again once it wears off and must be observed.
Catches people out: This is the clearest example of a type 2 failure with a specific antidote. Most other causes are managed by supporting ventilation rather than by a drug.
- Binds
- Beta-2 adrenoceptors (salbutamol) and muscarinic receptors, chiefly M3 (ipratropium), on bronchial smooth muscle
- Which does
- Salbutamol is an agonist at the Gs-coupled beta-2 receptor: cyclic AMP rises, protein kinase A phosphorylates and inactivates myosin light chain kinase and the muscle relaxes. Ipratropium is a non-selective muscarinic antagonist that blocks acetylcholine at airway M3 receptors, removing vagal bronchoconstrictor tone.
- So you see
- Airways widen, expiratory flow improves, more of each breath reaches alveoli rather than being trapped, so alveolar ventilation rises and PaCO2 falls.
- And the same mechanism causes
- The beta-2 receptor is not confined to the airway. On skeletal muscle it causes the tremor, on the heart tachycardia, and by stimulating the Na+/K+ ATPase it drives potassium into cells and produces hypokalaemia (with a rise in lactate and glucose from the same metabolic drive). Ipratropium's antimuscarinic action gives a dry mouth, and if nebulised mist reaches the eye, blurred vision and pupil dilatation — which can tip a susceptible eye into acute angle closure, so a mouthpiece or eye protection is used.
Catches people out: These raise ventilation only to the extent that the obstruction is reversible. They do little or nothing for a patient whose pump is failing from sedation, weakness or an unmoving chest wall.
- Binds
- Carbonic anhydrase (isoenzymes II and IV in the proximal tubule, and the enzyme in red cells, ciliary body and elsewhere)
- Which does
- Inhibiting the enzyme in the proximal tubule stops bicarbonate being reclaimed, so bicarbonate is lost in the urine. The plasma bicarbonate falls and a mild metabolic acidosis develops.
- So you see
- The acidosis is sensed mainly by the peripheral chemoreceptors in the carotid bodies (H+ crosses into CSF poorly, so the central chemoreceptors are not the immediate sensor here), and ventilation increases — the drug raises breathing by making the blood slightly acidic, not by acting on the lungs.
- And the same mechanism causes
- The deliberate acidosis brings tingling in the fingers and around the mouth, and the bicarbonate loss drags sodium, potassium and water with it, so hypokalaemia and dehydration follow. Alkaline urine with low urinary citrate favours calcium phosphate stones. Inhibiting red-cell carbonic anhydrase also slows tissue CO2 loading and lung unloading, which is why it is not a way of clearing CO2. The enzyme also sits in the ciliary body of the eye and in taste tissue, so aqueous production and eye pressure fall (useful in glaucoma) and carbonated drinks taste flat.
- Handling
- It also alkalinises the urine and causes a diuresis, and it is renally cleared. It is a sulfonamide derivative, so ask about sulfonamide hypersensitivity.
Catches people out: It is a niche drug with weak evidence in respiratory failure, not a treatment for it. Correcting a chronic retainer's compensating bicarbonate is usually the wrong move — that bicarbonate is what keeps their pH survivable.
PaCO2 is a ventilation number, not an oxygen number. Ask two separate questions of every blood gas. Is the exchanger broken? Look at PaO2. Is enough fresh air reaching alveoli? Look at PaCO2. Low O2 with low CO2 is a lung problem the patient is compensating for (type 1). Low O2 with high CO2 means alveolar ventilation has fallen short — the pump is failing or the dead space has grown (type 2) — and oxygen alone will not fix it. Then check the bicarbonate to see whether the CO2 has been high for days or for minutes.
Now test whether it stuck
Reading this through is not the same as being able to reconstruct it. Every question in the bank is free, with a full debrief on each option.