Oxygen at 4 litres in a known COPD patient who is now drowsy

The pH of 7.26 is acidaemia, and the raised PaCO2 of 72 mmHg makes this a respiratory acidosis. The bicarbonate of 31 mmol/L is the key to timing. Against his own baseline of 29 mmol/L, it has risen by only 2 mmol/L while the PaCO2 has risen by 22 mmHg above his baseline of 50 mmHg. In an acute rise, bicarbonate increases by roughly 1 mmol/L for every 10 mmHg of CO2, so 2 mmol/L is exactly what an acute change would produce - the 29 mmol/L baseline, not the current 31, represents his chronic compensation.

So this is acute-on-chronic respiratory acidosis: a chronically hypercapnic patient with a fresh acute deterioration, not a patient with a new metabolic alkalosis. The quantification matters clinically. If you read the bicarbonate as evidence of prolonged compensation, you may underestimate how acute and how unstable this is. His PaO2 of 78 mmHg on 4 L/min is numerically adequate oxygenation, but the saturation of 99% tells you he is being over-oxygenated relative to a target of 88-92%, and the drowsiness in a hypercapnic patient is a warning sign of CO2 narcosis rather than something to attribute to fatigue.

Several mechanisms operate together. In chronically hypercapnic COPD, the respiratory drive depends more on hypoxia than on CO2, so a high inspired oxygen concentration can reduce ventilation. More importantly, oxygen reverses hypoxic pulmonary vasoconstriction, so blood is redirected to poorly ventilated lung units, worsening ventilation-perfusion mismatch and increasing dead space. The Haldane effect adds to this: oxygenated haemoglobin carries less carbon dioxide, so a sudden rise in saturation releases CO2 into the blood. Nitrogen washout and absorption atelectasis contribute further. The result is a rising PaCO2 in a patient who cannot increase his minute ventilation.

The response is not to remove oxygen, which would be dangerous - hypoxaemia kills faster than hypercapnia. It is to switch to controlled low-flow oxygen, a 24-28% Venturi mask or 1-2 L/min nasal prongs, and titrate to a saturation of 88-92%, rechecking the blood gas within 30-60 minutes. Importantly, oxygen is only one contributor: this man has an infective exacerbation with purulent sputum, and treating that is what will lower his CO2. Sedatives should be avoided entirely, as they will compound the hypoventilation.

In the first hour: controlled oxygen to target; nebulised salbutamol with ipratropium, repeated as needed; systemic corticosteroids; antibiotics for purulent sputum with increased volume and breathlessness; a chest radiograph and ECG to look for pneumonia, pneumothorax or arrhythmia; electrolytes including potassium after repeated bronchodilators; and continuous oximetry with regular assessment of consciousness.

Escalation is decided on the blood gas combined with the clinical picture. A pH persistently below 7.35 with hypercapnia despite optimal medical therapy and controlled oxygen is the standard indication for non-invasive ventilation, and this patient already qualifies on his presenting gas. Non-invasive ventilation needs a cooperative patient and a protected airway, so his drowsiness is the deciding factor: if he can be roused to cooperate and the unit can intubate if he fails, a closely monitored trial is reasonable and often prevents intubation. If he cannot protect his airway, is deteriorating despite NIV, has haemodynamic instability, or develops a rising respiratory rate with exhaustion, then intubation and intensive care are needed.

The clinical point is that drowsiness in this context is a sign of impending failure, not a reason to wait and see. The place of care, and whether intubation is possible there, should be settled before the trial of non-invasive ventilation, not after it fails.