POSTOPERATIVE OXYGEN ADMINISTRATION

Br. J. Anaesth. (1975), 47, 108

POSTOPERATIVE OXYGEN ADMINISTRATION JULIAN M. LEIGH

A MODE OF RESUSCITATING PATIENTS AFTER INHALING THE VAPOUR OF ETHER. To the Editor of THE LAHCET.

Sin,—For the last week I liavq been using, as a meang of resuscitating patients, after inhaling the vapour of ether, pure oxygen gas, with the most perfect success. To-day, I operated in nine cases on the teeth: to each patient I gave a full dose of the ether vapour, and subsequently, a few inhalations of oxygen. In not one case did the patient complain of debility, and all recovered perfectly in less than a .minute and a half, timed by the medical" men present. I will, by your permission, in a future number of your journal, furnish the details of these and other experiments with oxygen.—I remain, Sir, your obedient Servant, Gower-etreet, Bedford-square, March 1847.

JAMES ROBINSOH

FIG. 1. Historical postoperative oxygen administration.

After every single anaesthetic, of anything but very brief duration, the following series of questions presents itself to the anaesthetist: Should I give oxygen, or not? And, if so, How much?; By what method?; and For how long? Underlying these clinical issues is a series of more fundamental questions: Is there hypoxaemia?; What is its degree?; Is it important?; How long will it last?; and finally—How do the risks of hypoxaemia of that degree, importance and duration equate with the undesirable effects of oxygen, if any? This paper will attempt to explore these problems. POSTOPERATIVE HYPOXAEMIA

The subject has been reviewed extensively (Sellery, 1968; Marshall and Wyche, 1972). Postoperative hypoxaemia may be subdivided into early, lasting a few hours; and late, when hypoxaemia occurs for more than 24 hr. The former may be considered a "patho-physiological J. M. LEIGH, M.D., F.F.A.RX.S., Anaesthetic Department,

St Luke's Hospital and Department of Physics, University of Surrey, Guildford, Surrey.

inevitability" even in normal individuals, and the latter when this state becomes prolonged by upper abdominal pain or is superseded by the onset of pulmonary complications. Late hypoxaemia should be suspected and diagnosed both clinically and by "blood gases", and treated according to the precepts employed when oxygen is prescribed in other spheres of clinical practice. It is the early type of hypoxaemia which is of interest in the present context. Early hypoxaemia. On returning to air-breathing, after an anaesthetic technique employing nitrous oxide, "diffusion hypoxia" can occur as this very soluble agent dilutes the alveolar gases. This is maximal in the first 5 min after withdrawal, and is easily overcome by raising the oxygen concentration of the inspired gas given from the anaesthetic machine for an appropriate period before permitting air breathing. In addition, where a technique employing hyperventilation has been used, there is a consequent deficit in the body carbon dioxide stores. Unless this is made up by the administration of carbon dioxide, there will be a period of relative alveolar hypoventilation as the losses are made good from endogenously produced carbon dioxide. If the patient is breathing air, there will be a consequent reduction in PA 0 2 and hence Pa02. This effect is mild and short-lived, however, and is only of significance as a contributing factor if other complications occur in the initial period due to residual effects of anaesthetic agents. The remaining component of early postoperative hypoxaemia may be recognized as a decrease in PaO2 which is unaccompanied by, or disproportionate to, a rise in Pa c02 and is therefore not attributable to the residual effects of anaesthetics, narcotics or muscle relaxants. It is an area of consuming interest and might reasonably be termed "spurious hypoxaemia" to distinguish it from diffusion hypoxaemia, post-hyperventilation hypoxaemia, and late hypoxaemia due to complications. There is a well-known negative correlation between Pa 0 , and age, such that Pa 02 decreases approximately 3.5 mm Hg per decade. After opera-

Downloaded from http://bja.oxfordjournals.org/ at University of California, San Francisco on March 18, 2015

Oxygen administration following surgery is not new (Robinson, 1847, fig. 1; Hooper, 1847), but knowledge governing its use for less empiric purposes has certainly accumulated. Indeed, investigation into the causes of postoperative hypoxaemia has provided anaesthetists with an intellectual challenge for almost two decades.

109

POSTOPERATIVE OXYGEN ADMINISTRATION

THE SIGNIFICANCE OF SPURIOUS POSTOPERATIVE HYPOXAEMIA

Prys-Roberts (1968) has expressed himself in a way which is worthy of elaboration: "It is easy in these days of preoccupation with the measurement of oxygen tension to forget the overall scheme of oxygen transport . • .". During the respiration of mixtures with an oxygen concentration like air, the following approximates to the amount of oxygen available per minute: O2 availability=K (Hb X Sa02 X Qt), where Hb=haemoglobin level in g/100 ml whole blood; SaO2=percentage oxygen saturation of the haemoglobin in arterial blood; Qt=cardiac output in ml/min; K = a constant, derived from the oxygen capacity of haemoglobin and appropriate decimal corrections, with a resulting value of 0.000139. Considering some normal values: O2 availability=K(14.5x 100x5,000) ===1,000 ml/min As basal oxygen consumption is of the order of 250 ml/min, a true perspective of the relative contributions of the various factors is obtained from the following interpretation of the preceding figures: Each litre change in cardiac output, when the haemoglobin level is normal and there is full

saturation, accounts for a change of 200 ml/min in oxygen availability; Each gram change in haemoglobin, when the cardiac output is 5 litre/min and there is full saturation, accounts for a change of 69 ml/min in oxygen availability; Each per cent change in saturation, when the cardiac output is 5 litre/min and the haemoglobin level is normal, accounts for 10 ml/min in oxygen availability. The correlations between age and oxygenation are usually expressed with Po2 on the y axis, and the slopes look reasonably daunting. However, figure 2 shows the regression of saturation against age taken from the regression lines for PaO2 published by Marshall and Wyche (1972), and using standard saturation curves. A value for pH of 7.4 was used for the preoperative regression (P60===29 mm Hg), and values for pH of 7.6 and 7.2 for the postoperative regressions (P5Or^43 and 19 mm Hg respectively); the higher pH represents the state of affairs which might exist immediately after operation following a period of hyperventilation, and the lower the state of affairs which might exist after some severe complication. The values used are given in table I.

90SATURATION

80-

70 J —1 CO AGE (YEARS)

80

FIG. 2. Regression of haemoglobin saturation against age using the values given in table I. The unbroken line shows the situation for normal subjects (assumed to have a pH of 7.4). The dashed line and broken line show the regressions in the presence of postoperative hypoxaemia, at pH values of 7.6 and 7.2 respectively. For further explanation, see text.

Downloaded from http://bja.oxfordjournals.org/ at University of California, San Francisco on March 18, 2015

tion, the regression curve is depressed, and although most authorities have shown this depression to be symmetrical (Marshall and Wyche, 1972; Davis and Spence, 1972), Kitamura, Sawa and Ikezono (1972) found that the hypoxaemia increased with increasing age, that is the line was steeper. The mechanism of the negative correlation with age is thought to be due to an increasing approach of the airways closing point to the end-tidal point (Alexander et al., 1972). The exacerbation after anaesthesia, manifest as spurious hypoxaemia, could be caused by V A / Q mismatch as a result of an extension of the above process due to a reduction in functional residual capacity (FRC), and also to a temporary disorder of the hypoxic pulmonary pressor response (Sykes et al., 1973). Obviously, in patients with poor cardiopulmonary function or who have pulmonary procedures carried out upon them, the state of affairs is intensified further. More detailed discussion on postoperative lung function will be found elsewhere in this issue (Hewlett and Branthwaite, 1975). The essential consideration is whether this hypoxaemia is important.

BRITISH JOURNAL OF ANAESTHESIA

110 TABLE I. Values of Pao, (mm Hg), and haemoglobin

saturation per cent, used in the construction of figure 2. (Taken from the regression lines published by Marshall and Wyche (1972), and using saturation curves with P 60 values as shown.) Normals pH 7.4

Postoperative

that full arterial saturation is no guarantee of adequate tissue oxygenation. POSSIBLE DETRIMENTAL EFFECTS OF OXYGEN THERAPY

Postoperative oxygen administration may be harmpH 7.6 pH 7.2 Pa02 % sat. Pao2 % sat. ful if the Pa 0 , is raised above normal for a parti-

Pa<,2

% sat.

0 40 80

103 89 75

100 99 96

87 75 68

99 98 97

87 75 68

93 89 84

Pso

29

50

43

50

19

50

Bearing in mind that the line at pH 7.6 looks better than it is and the pH 7.2 line worse than it is, because haemoglobin has a higher affinity for oxygen in the former than in the latter, spurious postoperative hypoxaemia per se constitutes a relatively innocuous threat. Provided that there is no gross reduction in Qt, no gross haemorrhage, and that no other acute problem arises, postoperative arterial hypoxaemia of the order described is of little consequence. In postoperative non-shivering patients, values for pH of 7.35 + 0.07 (SD), and in shivering patients values of 7.32 ± 0.09, have been found (calculated from data given by Bay, Nunn and Prys-Roberts, 1968). The demonstration of arterial hypoxaemia in no way indicates tissue hypoxia because the role of the cardiac output is not revealed by measurements of Pa 02 . Only mixed venous oxygen measurements, together with values of base deficit, will demonstrate whether oxygen consumption exceeds availability. No such demonstration appears in the literature, indeed in only one of the above authors' shivering patients did gross venous desaturation occur when cardiac output did not rise enough for the increase in oxygen consumption. It is evident from this discussion that postoperative hypoxaemia should be avoided in situations where the cardiac output, or cardiovascular reactivity, may be depressed. Such a state of affairs probably always exists after deliberate hypotension has been induced with drugs other than sodium nitroprusside, or when there has been gross haemorrhage. In the latter case there will also be a reduced haemoglobin level. The increased alveolar deadspace accompanying a reduced cardiac output causes an increase in PAr02 and a consequent decrease in P A 0 2 , thus providing an additional reason for oxygen therapy under these circumstances. However, it must be stressed once more

cular patient, or if PA 02 is in excess of 0.5 atm abs. The former can cause and exacerbate respiratory depression, and the latter may lead to pulmonary oxygen toxicity. Respiratory depression. The danger of oxygen therapy in causing further respiratory depression in patients with chronic respiratory disease has been recognized since the 1930's (Barach, 1938), and is of equal importance when such patients have been anaesthetized. The depressant effect of a high Pa02 on ventilation in normal subjects was implicit in the breath-holding experiments of Vernon (1909), and further elaborated by Lindhard (1911). Although it was confirmed several times subsequently, a definitive study was performed by Lloyd, Jukes and Cunningham (1958) who investigated the linear part of the V-PE'COU response curve of healthy individuals at differing PE' O 2 (range of the vast majority of readings 37.8-160 mm Hg). These authors showed that the relation between ventilation and carbon dioxide is inversely related to the PE' O 2 (and the PaO2 by inference), while the intercept on the Pco2 axis is independent of Po 2 ; that is, the slope of the response curve is depressed. Remembering also that the ^-Pco s response curve is shifted to the right by anaesthetics, or other supplements, the combination of the residual effects of these drugs and the effects of too high a Pa 0 , may lead to a potentially dangerous decrease in alveolar ventilation (Lambertson, 1964, 1965). Pulmonary oxygen toxicity. This subject has been reviewed by Hedley-Whyte and Winter (1967), and more recently by Winter and Smith (1972). In man, pulmonary oxygen toxicity does not develop in less than 24 hr even with an Fi o , of 1.0 and never with an Fi O2 of less than 0.4-0.5. If much higher inspired oxygen concentrations are required to produce a satisfactory Par,,, then physiotherapy and IPPV, possibly with PEEP, will cause a dramatic improvement in Qs/Qt fraction without necessitating any increase in Fi 02 . Winter and Smith (1972) succinctly sum-

Downloaded from http://bja.oxfordjournals.org/ at University of California, San Francisco on March 18, 2015

Age

POSTOPERATIVE OXYGEN ADMINISTRATION

gen is given, that the guiding principle should be that the Pa02 must be sufficiently high to combat hypoxaemia, but not so high as to cause respiratory depression in the presence of the residual narcotic effects of anaesthesia, or where there is pre-existing pulmonary disease. The anaesthetist should make the decision after surgery whether to give 2 1 % oxygen (air), or prescribe a dose accurately based upon the nature of the individual patient, taking into account his condition before operation, age, and the needs of the surgical procedure and predicting the duration of significant hypoxaemia. There is little doubt that avoidable postoperative anaesthetic morbidity and mortality is more often due to unrecognized respiratory depression and/or airway obstruction than to hypoxaemia alone. Equally, there is little doubt that if there is arterial hypoxaemia preceding one of these episodes the chances of a favourable outcome will be diminished. On the other hand, the author has encountered several cases of respiratory depression leading to carbon dioxide narcosis due to a relative overdose of oxygen in the early postoperative period, without meeting a single case where spurious hypoxaemia has produced any obvious deleterious effects. Though there may be no demonstrable morbidity from spurious hypoxaemia, it does not necessarily follow that none exists. As it is the objective of the modern anaesthetist to reduce the possibility of morbidity after anaesthesia, it would seem wise for oxygen therapy to be given within the constraints indicated in this article. The work of Davis and Spence (1972) indicates that the lower the Pa 02 the greater will be the Fi 0 3 required to return the Pa02 to normal. With this in mind, it seems that under ideal circumstances, spurious postoperative hypoxaemia could be managed with existing venturi devices according to the age of the patient as follows: CHOICE OF TECHNIQUE OF OXYGEN THERAPY 0-20 yr 21% It has been shown, except with HAFOE Venti20-40 yr 24% masks or similar devices, that there is a breath-to40-60 yr 28% breath and subject-to-subject variation in oxygen 60-80 yr 35% concentration (Leigh, 1970), besides the high peaks In the event of complications, a higher concenin oxygen concentration referred to above. The tration should be used, but the necessity would result is that the dose given to any individual best be judged by measurements. Indeed, in sumpatient cannot be predicted at all. It seems remark- ming up the whole subject the author wishes to able that although millions of words have been echo the sentiments of the above authors, and of written by hundreds of anaesthetists on the minu- Hedley-Whyte and Winter (1967), in a plea for tiae of spurious postoperative hypoxaemia, the same more control by measurements, and especially for attention has not been given to the prescription of further measurements on mixed venous blood; oxygen for its management. It seems clear, if oxy- further, that measurements should take place when

Downloaded from http://bja.oxfordjournals.org/ at University of California, San Francisco on March 18, 2015

marized the relative dangers as follows: "That patients should be exposed to dangerous levels of hypoxia for fear of developing oxygen toxicity is a form of therapeutic nihilism which is obviously self-defeating" (author's italics). However, the question remains as to what constitutes dangerous levels of hypoxia. Before dismissing prematurely the possibility of pulmonary oxygen toxicity in the context of postoperative oxygen therapy, the question must be asked: "Are levels as high as this given to patients after surgery?" The answer is: "Virtually always". Investigations have shown that, although the net inspired oxygen concentration may be of the order of 35%, the concentration of oxygen in inspired gas varies within each breath when variable performance devices are in use, so that high oxygen concentrations are consistently inhaled at the beginning of each breath (Leigh, 1974). Experiments have shown that, with the MC mask (representing a small-capacity device), at a flow of 5 litre/min there are peaks of oxygen concentration of 60%, and at 10 litre/min these can be as high as 95%. With a polymask (representing a large capacity system), the corresponding peaks were 65% and 82% (Leigh, 1973). Since the first part of the inspirate consistently ventilates the same parts of the lung, these will be ventilated consistently with a potentially toxic oxygen concentration. Although it may not be the case that therapy will be continued for long enough for any effects to develop, nevertheless the possibility still exists when postoperative oxygen therapy is prolonged. With HAFOE devices (high air flow with oxygen enrichment) such as the Ventimasks, this possibility does not exist as the inspired gas consists uniformly of the stated concentration throughout inspiration.

111

112

BRITISH JOURNAL OF ANAESTHESIA

Downloaded from http://bja.oxfordjournals.org/ at University of California, San Francisco on March 18, 2015

known doses of oxygen are given. The latter condi- "srKitamura, H., Sawa, T., and Ikezono, E. (1972). Postoperative hypoxemia: the contribution of age to the tion did not apply to the original study of the effect maldistribution of ventilation. Anesthesiology, 36, 244. of oxygen therapy on postoperative hypoxaemia by Lambertson, C. J. (1964). Effects of drugs and hormones on the respiratory response to carbon dioxide; in Conway and Payne (1963), nor indeed to the recent Handbook of Physiology (eds. Fenn, W. O., and Rahn, work of Davis and Spence (1972). H.), Section 3. Vol. I, p. 505. Washington: American Physiological Society. While the opinions on postoperative oxygen (1965). Effects of oxygen at high partial pressure; administration expressed above may be controin Handbook of Physiology (eds. Fenn, W. O., and versial, the admitted criticism is that they constiRahn, H.), Section 3. Vol. II, p. 1027. Washington: tute a counsel of perfection. It is the opinion of the American Physiological Society. author that only the further application of such Leigh, J. M. (1970). Variation in the performance of oxygen therapy devices. Anaesthesia, 25, 210. standards will continue to maintain advances in (1973). M. D. Thesis, University of London. the practice of anaesthesia as it enters the last (1974). Oxygen therapy at ambient pressure; in Scientific Foundations of Anaesthesia (eds. Scurr, quarter of the 20th century. C. F., and Feldman, S. A.), 2nd edn., p. 253. London: Heinemann. ACKNOWLEDGEMENTS Lindhard, J. (1911). On the excitability of the respiraThe author thanks Barbara Lister for secretarial assisttory centre. J. Physiol. (Lond.), 42, 337. ance and Hilary Knight for drawing figure 2. Lloyd, B. B., Jukes, M. G. M., and Cunningham, D. J. C. (1958). The relation between alveolar oxygen REFERENCES pressure and the respiratory response to carbon dioxide Alexander, J. I., Horton, P. W., Millar, W. T., Parikh, in man. Q. J. Exp. Physiol, 43, 214. R. K., and Spence, A. A. (1972). The effect of upper Marshall, B. E., and Wyche, M. Q., jr. (1972). Hypoxabdominal surgery on the relationship of airway closemia during and after anesthesia. Anesthesiology, 37, ing point to end tidal position. Clin. Set., 43, 137. 178. Barach, A. L. (1938). Physiological methods in the diag- Prys-Roberts, C. (1968). Postanesthetic shivering; in nosis and treatment of asthma and emphysema. Ann. Common and Uncommon Problems in Anesthesiology Intern. Med., 12, 454. (ed. Jenkins, M. T.), p. 357. Oxford: Blackwell. Bay, J., Nunn, J. F., and Prys-Roberts, C. (1968). Fac- Robinson, J. (1847). A mode of resuscitating patients tors influencing arterial Po2 during recovery from after inhaling the vapour of ether. Lancet, 1, 371. anaesthesia. Br. J. Anaesth., 40, 398. Conway, C. M., and Payne, J. P. (1963). Postoperative Sellery, G. R. (1968). A review of the causes of postoperative hypoxia. Can. Anaesth. Soc. J., IS, 142. hypoxaemia and oxygen therapy. Br. Med. J., 1, 844. Davis, A. G., and Spence, A. A. (1972). Postoperative Sykes, M. K., Davies, D. M., Chakrabarti, M. K., and Loh, L. (1973). The effect of inhalational anaesthetic hypoxemia and age. Anesthesiology, 37, 663. agents on the pulmonary vasculature of the isolated Hedley-Whyte, J., and Winter, P. M. (1967). Oxygen perfused cat lung. Br. J. Anaesth., 45, 655. therapy. Clin. Pharmacol. Ther., 8, 696. Hewlett, A. M., and Branthwaite, M. (1975). Postopera- Vernon, H. M. (1909). The production of prolonged apnoea in man. J. Physiol. (Lond.), 38, xviii. tive pulmonary function. Br. J. Anaesth., 47, 102. Hooper, W. (1847). Inhalation of oxygen for resuscitating Winter, P. M., and Smith, G. (1972). The toxicity of etherised patients. Pharmac. J., 6, 508. oxygen. Anesthesiology, 37, 210.