On this page
Study tools
Write your answer
Saved on this device. No marking — you are the marker.
Stem
A 70-year-old retired smoker with known severe COPD (FEV1 38 percent predicted, on tiotropium and a salbutamol inhaler, long-term oxygen at home for chronic type 2 respiratory failure with a baseline PaCO2 of 7.4 kPa) is brought to the emergency department. She has been more breathless for three days with increasing green sputum, and over the last 12 hours has become drowsy. On arrival she is cyanosed, drowsy (GCS 13), using accessory muscles and breathing at 32 per minute with a prolonged expiratory phase and wheeze. SpO2 on room air is 74 percent. BP 152/88, pulse 112, temperature 37.9 degrees C. Her arterial blood gas on room air shows pH 7.24, PaCO2 11.2 kPa (84 mmHg), PaO2 5.6 kPa (42 mmHg), HCO3 34 mmol/L, base excess plus 9, lactate 2.4.
Questions
a) Classify her respiratory failure using the arterial blood gas and explain what the bicarbonate and base excess add to your reading of the gas. (2 marks)
This is Type 2 (hypercapnic) respiratory failure: PaO2 below 8 kPa (she is 5.6 kPa, 42 mmHg) with PaCO2 above 6 kPa (she is 11.2 kPa, 84 mmHg), so it is a VENTILATORY (pump) failure. The raised bicarbonate (34 mmol/L) and positive base excess (plus 9) indicate renal bicarbonate retention — evidence that she has a CHRONICALLY COMPENSATED hypercapnia at baseline (her home PaCO2 was 7.4 kPa). The fact that the pH is now 7.24 (acidaemic) despite this chronic compensation means she has an ACUTE DECOMPENSATION on top of her chronic type 2 state — i.e. acute-on-chronic hypercapnic respiratory failure. This is the classic COPD exacerbation scenario, and it is the indication for controlled oxygen and early non-invasive ventilation.
b) She is put on a non-rebreather mask at 15 L/min and her SpO2 rises to 98 percent, but she becomes more drowsy. A repeat gas shows pH 7.18, PaCO2 13.5 kPa, PaO2 11.0 kPa. Explain the mechanism by which the supplemental oxygen worsened her hypercapnia and what oxygen strategy you should have used. (3 marks)
She has developed oxygen-induced hypercapnia progressing toward CO2 narcosis. Two mechanisms explain the rise in PaCO2 from 11.2 to 13.5 kPa after high-flow oxygen:
- The Haldane effect. Deoxyhaemoglobin carries carbon dioxide (as carbamino compounds) and buffers hydrogen ion far more effectively than oxyhaemoglobin. Oxygenating her venous blood in the lung releases CO2 from haemoglobin into plasma, raising mixed-venous and arterial CO2. In a patient whose ventilation is already marginal (fatigue, airflow limitation), this extra CO2 load cannot be cleared and PaCO2 rises.
- Loss of hypoxic pulmonary vasoconstriction. In COPD, poorly ventilated alveolar units are kept locally vasoconstricted by hypoxia, diverting blood toward better-ventilated units. Supplemental oxygen relieves that hypoxia, relaxes the vasoconstriction, and re-perfuses poorly ventilated units, worsening V/Q mismatch and increasing dead-space-equivalent CO2 retention.
Rising PaCO2 produces CSF acidosis, which first causes cerebral vasodilatation (headache) and then depresses the central respiratory drive, producing the drowsiness, asterixis and progressive coma of CO2 narcosis.
The oxygen strategy she should have had from the start is controlled oxygen via a fixed-performance Venturi mask (24 or 28 percent), targeting SpO2 88 to 92 percent, with the arterial blood gas repeated at 30 to 60 minutes (BTS oxygen guideline; BTS/ICS 2016). Hypoxia kills faster and more certainly than hypercapnia, so oxygen must not be withheld — but it must be controlled in a known CO2 retainer.
c) Outline the immediate definitive management of this acute-on-chronic acidotic hypercapnic exacerbation, naming the ventilatory mode, its settings, and the trial evidence that supports it. (3 marks)
After switching to controlled oxygen (Venturi 28 percent, target SpO2 88 to 92 percent) and reversing the oxygen-induced hypercapnia, she should start non-invasive ventilation (NIV) with bilevel positive airway pressure (BiPAP) without delay, because she has persistent acidotic hypercapnic respiratory failure (pH below 7.35) despite standard medical therapy.
BiPAP setup: IPAP 10 to 15 cmH2O, EPAP 4 to 5 cmH2O, oxygen entrained to target SpO2 88 to 92 percent, full-face mask (interface of first choice); titrate IPAP upward by 2 to 5 cmH2O every 10 minutes until the pH rises above 7.35 or the PaCO2 falls; reassess clinically and with the blood gas at 1, 4 and 12 hours. Escalate to intubation if NIV fails (rising CO2, falling pH, exhaustion) or the airway is threatened.
Trial evidence: the Plant 2000 multicentre randomised controlled trial (Lancet) showed that early NIV on general respiratory wards for COPD exacerbations with pH 7.25 to 7.35 reduced in-hospital mortality, the need for intubation, and length of stay. The BTS/ICS 2016 guideline codifies NIV as first-line therapy in this scenario. Concurrent specific therapy for the exacerbation is nebulised salbutamol 5 mg plus ipratropium 500 micrograms, prednisolone 30 to 40 mg orally (or hydrocortisone 100 mg IV), and antibiotics (amoxicillin-clavulanate or doxycycline) for the infective trigger.
d) Two days later her sister is admitted with sudden pleuritic breathlessness and a SpO2 of 88 percent; her gas shows a NORMAL A-a gradient with mild hypoxaemia and a normal PaCO2. What does a normal A-a gradient tell you about the mechanism of her hypoxaemia, and why does this matter for her management? (2 marks)
A normal A-a gradient with hypoxaemia indicates PURE ALVEOLAR HYPOVENTILATION — the lung itself is normal and the problem is the respiratory pump (the drive or the bellows). Of the five mechanisms of hypoxaemia (hypoventilation, V/Q mismatch, shunt, diffusion impairment, low inspired PaO2), only hypoventilation and low inspired PaO2 keep the A-a gradient normal; the other three raise it. The alveolar gas equation explains why: PAO2 equals 150 minus PaCO2/0.8 on room air, so a rising PaCO2 directly lowers alveolar (and arterial) oxygen without any lung disease, and the gradient stays flat.
This matters for management because it redirects the whole work-up away from the lung and toward the pump — think opiate or sedative toxicity, brainstem stroke, neuromuscular weakness (Guillain-Barre, myasthenia, motor neuron disease), or fatigue. A normal A-a gradient in a hypoxaemic patient should prompt a drug history and toxin screen, a neurological examination, and respiratory-muscle testing (FVC, MIP/MEP) — not a hunt for pneumonia. (Note: a normal A-a gradient also fits altitude and a hypoxic gas mixture, ruled out by history.)
(Bonus mark-worthy nuance: in suspected pulmonary embolism the A-a gradient is usually RAISED from increased dead space and V/Q mismatch; a normal A-a gradient makes clinically significant PE considerably less likely, although it does not exclude it.)