Is your patient hypoxaemic as well as breathless? In most emergency departments these days arterial oxygen saturation is a routine measurement and a significant proportion of breathless patients will have arterial blood gases taken as well. However, in my experience, the interpretation of these tests is seldom complete. Let us take an example (a real case) of a 23-year-old, breathless woman whose arterial oxygen saturation (SaO₂) breathing air was 96 per cent.

Can we be completely reassured by the normal oxygen saturation? The answer is ‘No’ for two reasons. First, the shape of the oxygen dissociation curve means that a SaO₂ of 96 per cent can be achieved with an arterial oxygen tension (paO₂) that may be far from normal in a 23-year-old person. Second, both oxygen saturation and oxygen tension cannot be assessed in isolation because they are both affected by the
degree of ventilation – both will increase with increasing ventilation (if the lungs are normal) and hypoventilation will have the opposite effect.

In the example quoted, blood gases were checked, revealing a normal paO₂ at 12.5 kPa, a low arterial carbon dioxide tension (paCO₂) of 3 kPa and a pH of 7.49. The low paCO₂ indicated that the patient was hyperventilating and, because the paO₂ was normal, primary hyperventilation was diagnosed. This conclusion was dangerously wrong because no objective calculation had been made of the paO₂ that might be expected for the degree of hyperventilation if oxygen exchange was normal.

A fundamental point is that the measured paO₂ is affected by the inspired oxygen concentration (FIO₂) and the degree of alveolar ventilation (VA), as well as the effectiveness or integrity of lung function, which dictates the ability to transfer oxygen from the alveolar space to the pulmonary capillaries.

If lung function is normal, an increase in FIO₂, VA or both will result in an increase in paO₂. A fall in FIO₂ or VA will have the opposite effect. It follows that paO₂ cannot provide an accurate assessment of lung function on its own; it must be examined in relation to what is happening in the alveolar space where the alveolar concentration of oxygen (FAO₂) and therefore its partial pressure (pAO₂), is influenced by both FIO₂ and VA. This young woman’s measured arterial carbon dioxide tension (paCO₂) was low and, because the transport of carbon dioxide from pulmonary capillary to lung alveolus is highly efficient, we know that her alveolar carbon dioxide tension (pACO₂) would have approximated very closely to paCO₂ at 3 kPa. Second, the total pressure of alveolar gases must equal atmospheric pressure and if the partial pressure of carbon dioxide in the alveoli falls, the contribution from another gas must rise in order to compensate (otherwise the individual would implode!). The alveolar nitrogen component of inspired air remains constant and so in normal situations this leaves only oxygen to make up for the deficit – in other words, as pACO₂ falls, pAO₂ will rise. The opposite is also true; an increase in pACO₂ (usually due to hypoventilation) results in a fall in pAO₂.

Going back to the example of the young woman, we are right to infer that a paCO₂ of 3 kPa would have resulted in a pAO₂ of more than 12.5 kPa and that her paO₂ should have been correspondingly higher if lung function had been normal. Unlike CO₂, the transfer of oxygen is never perfect (because of an unavoidable degree of ventilation-perfusion mismatch) and this means that there is always an alveolar-arterial oxygen difference (pAO₂ – paO₂), which increases with age. Rather than guessing, we should be precise about the paO₂ that can be expected for a particular paCO₂ (which reflects the degree of VA) and the first step is to calculate pAO₂. This is straightforward because pAO₂ is easy to calculate using the modified alveolar gas equation as follows:

where R is the respiratory quotient (0.8 under most conditions) and pIO₂ is 21 per cent of atmospheric pressure (101 kPa). Saturated water vapour pressure is 6.5 kPa and therefore the partial pressure of oxygen in inspired air is 21 per cent of 94.5 kPa, i.e. 19.845 kPa (or 20 for simplicity).

If we return to the example quoted:

Blood gases measure paO₂ and the alveolar-arterial oxygen difference (pAO₂ – paO₂) can therefore be calculated and compared with normal values:

pAO₂ – paO₂ correlates with the degree of ventilation-perfusion mismatch and therefore provides direct information on the integrity of lung function. A gradient of more than 2 kPa is very questionable in a 23 year old and one of 3.75 is definitely abnormal. In other words, by completing this simple calculation, we have shown that the patient is hyperventilating in order to compensate for abnormal oxygen exchange and to avoid hypoxaemia. This is not primary hyperventilation, there is something wrong with her lung function. Moreover, the abnormal physiology is not matched by any pathology that is visible on the chest radiograph and this discrepancy has to be explained.

This is a short but important step. Consider alternative blood gases in a young man: paO₂16.5 kPa, paCO₂2 kPa, pH 7.37

The A-a gradient calculates out at 1 kPa, which is completely normal. The patient is definitely hyperventilating and his lung function is good. However, before diagnosing primary hyperventilation, the parameters of acid-base balance must be examined. This was a real example in a 19-year-old man who had taken a salicylate overdose and was hyperventilating in order to correct for a metabolic acidaemia. The fall in paCO₂

Stay updated, free articles. Join our Telegram channel

Join