Why do patients need extra oxygen during a general anaesthetic?

It is routine practice during general anaesthesia (GA) to administer more than the 21% oxygen in which we mostly spend our lives. It is essential to understand the physiology underlying this practice if we are to keep patients safe by avoiding both hypoxaemia and hyperoxia during GA.¹ Four pathophysiological changes occur, which explain the answer to the question posed in the title of this article.

• (i)Abnormal respiratory muscle activity

Induction of GA is followed within seconds by profound relaxation of all skeletal muscles, to a much greater extent than seen with physiological sleep. The effects on airway muscles are well known to all anaesthetists, and artificial airways are used routinely to bypass this part of the respiratory tract. The spinal muscles relax, increasing thoracic spinal curvature; relaxation of intercostal muscles leads to a reduction in the cross-sectional area of the ribcage; and the diaphragm is displaced in a cephalad direction, particularly in dependent regions because of the weight of the abdominal contents. Together, these three changes to the chest cavity shape cause an immediate reduction in functional residual capacity (FRC) of 15–20% compared to the value when awake and supine. In effect, the patient takes a prolonged expiration as they lose consciousness and when breathing recommences, this is from a lower lung volume. The reduction in FRC is similar in patients breathing spontaneously and those whose lungs are ventilated artificially, and is much greater in obese patients.²

• (ii)Formation of atelectasis

In older patients, the reduced FRC may be less than the closing capacity and so airway collapse will occur throughout the lung. Even if FRC remains above the closing capacity, the changes in chest wall and diaphragm shape commonly result in direct compression of lung tissue in the caudal and dependent regions behind the diaphragm. This leads to atelectasis in 75–90% of patients, which is easily detected by CT. There is some evidence that use of oxygen 100% at various stages of GA may exacerbate the formation of atelectasis. Blood flow through atelectasis constitutes an intrapulmonary pathological shunt and will adversely affect arterial oxygenation. The amount of shunt through these areas varies between individuals and is likely to be influenced by the efficacy of their hypoxic pulmonary vasoconstriction reflex.³ This becomes particularly important during one-lung ventilation. Increasing F_IO₂ will have no effect on this situation as the extra oxygen does not reach these closed lung regions. Atelectasis can be reduced at induction by avoiding oxygen 100%: in one study the use of 80 or 60% oxygen before induction reduced the atelectasis area on CT scans by 76 and 96%, respectively, compared with oxygen 100%.⁴ Alternatively, use of moderate amounts of CPAP (6 cm H₂O) before induction of anaesthesia prevents atelectasis formation.⁵ During maintenance of anaesthesia the use of PEEP helps to limit the amount of atelectasis that forms, but once formed, a recruitment manoeuvre with high airway pressures (30–40 cm H₂O) is required to re-expand the collapsed areas.

• (iii)Abnormal regional ventilation and perfusion matching in the lungs

In regions of the lung that are not collapsed, the normal ventilation (V˙) and perfusion (Q˙) relationships are disturbed. In an awake subject, both V˙ and Q˙ increase on moving through lung regions from non-dependent areas to dependent areas, with Q˙ increasing slightly more than V˙. During GA, with spontaneous ventilation these relationships remain mostly unchanged, but this is not the case with IPPV when ventilation becomes distributed much more to the ventral regions (Fig. 1) while perfusion distribution is unchanged. This leads to increased scattering of V˙/Q˙ ratios (i.e. an increase in lung areas with both high and low V˙/Q˙ ratios). In dependent regions, V˙/Q˙ ratios are usually less than one because of poor regional ventilation and maintained perfusion; so venous blood passing through these lung regions does not become fully oxygenated and arterial hypoxaemia occurs. This explains impaired oxygenation during GA even in the absence of atelectasis (where V˙/Q˙=0). The areas of lung with 0>V˙/Q˙<1 when combined with the shunt through areas of atelectasis result in an overall shunt fraction of approximately 10%. Unlike for areas of atelectasis, increasing F_IO₂ by even small amounts (30–40%) can correct the adverse effects of lung areas with very low V˙/Q˙ ratios.

Fig 1.

Electrical impedance tomography of ventilation distribution before and after induction of anaesthesia (white=well ventilated, blue=some ventilation, black=no ventilation). (A) Awake patient showing the normal situation of greater ventilation of the right lung and dependent regions on both sides. (B) Same patient 4 min after GA induction with positive pressure ventilation showing continued greater ventilation of the right lung, but now predominantly non-dependent regional ventilation. The figures show the percentage ventilation to the four quadrants.

In non-dependent lung regions, there is good ventilation but reduced perfusion, both from the increased ventilation reducing regional lung perfusion and from the overall lower cardiac output that normally follows induction of anaesthesia. This area of lung contributes to alveolar dead space: this impairs carbon dioxide removal and contributes to the development of hypercapnia (as described below) even when minute ventilation is normal.

• (iv)Development of hypercapnia

Minute ventilation is significantly reduced during anaesthesia if artificial ventilation is not used: opioids normally slow the ventilatory frequency and inhaled anaesthetics reduce tidal volume, so the usual combination of both drugs causes a profound reduction in ventilation. The normal reflex ventilatory responses to both hypoxia and hypercapnia are also attenuated.⁶ This hypoventilation, along with the increased alveolar dead space during GA, commonly results in hypercapnia in a spontaneously breathing patient. Hypercapnia is often more marked in laparoscopic or thoracoscopic surgery, when carbon dioxide absorption from the peritoneum or pleura, respectively, increases the amount of carbon dioxide that must be excreted by the lungs. Assuming a physiological ‘steady state’ (i.e. no recent change in inspired gas composition) a simple form of the alveolar gas equation clearly shows that any increase in alveolar PCO₂ will be accompanied by an almost identical decrease in alveolar PO₂:

$$P{\text{AO}}_2=P{\text{iO}}_2-\frac{P{\text{ACO}}_2}{\text{RQ}}$$

where PAO₂=alveolar PO₂, PiO₂=inspired PO₂, PACO₂=alveolar PCO₂ (normally assumed to equal arterial PCO₂), and RQ=respiratory quotient. Increasing F_IO₂ by a similar amount to the increase in PCO₂ will prevent this from affecting arterial oxygenation.

In summary, oxygenation is impaired during GA by four mechanisms: hypercapnia reducing alveolar PO₂, abnormal respiratory muscle activity, which leads to shunting of venous blood through areas of atelectasis, and mismatch of ventilation to perfusion leading to areas of lung with low V˙/Q˙ ratios. Abnormal muscle activity and atelectasis cannot be rectified by increasing F_IO₂, but the other two are easily corrected with quite small increases to around 30–40% inspired oxygen, which normally results in acceptable arterial saturation for most patients.

Declaration of interest

The author declares that they have no conflict of interest.

Biography

Andy Lumb FRCA is consultant anaesthetist at St James's University Hospital and honorary clinical associate professor at the University of Leeds, specialising in thoracic anaesthesia. His research focuses on respiratory physiology and he teaches at all levels including at postgraduate meetings in the UK and abroad. Dr Lumb has authored many chapters, reviews, editorials, and four editions of Nunn's Applied Respiratory Physiology. He is an editorial board member of BJA Education and an associate editor of the BJA.