Anaesthesia 1979, Volume 34, pages 578-582 CASE REPORT
Detection of venous air embolism A clinical report using end-tidal carbon dioxide monitoring during neurosurgery
DAVID HURTER
AND
Venous air embolism is a recognised hazard to patients undergoing operations in the sitting position. A widely varying incidence has been reported, from 1.6%’ to 9 3 x 2 with a mortality from 0%’ to 73%.’ This wide range almost certainly represents a gross disparity in the sensitivity of the monitoring methods used to detect air embolism. The various methods of detection have been compared in dogs4 and in man5 and the pathological physiology has been described.6 Because this complication can prove rapidly fatal, its early detection is of vital importance. Until recently, anaesthetists have had to rely on comparatively crude methods of monitoring, namely pulse rate and electrocardiography (ECG) * blood pressure, central venous p r e s s ~ r e ,respiratory ~ pattern l o * and heart sounds heard through an oesophageal s t e t h o s c ~ p e . ~Changes ~ * ’ ~ in these caused by venous air embolism are variable and, with the exception of heart sounds, non-specific. In recent years, two newer and more specific methods of detection have been introduced, the l 5 and Doppler ultrasonic flow the continuous carbon dioxide (CO,) analyser. l 6 The Doppler transducer is an extremely sensitive device which detects air bubbles in the heart in very small amounts. It may be difficult to position the probe satisfactorily on the chest, especially in the obese patient. The frequency with which it raises a false alarm by detecting
P E T E R S. S E B E L
insignificant amounts of air has proved a great disadvantage in its clinical use.’ Air in the pulmonary vessels interferes with gaseous exchange and causes retention of C O z , giving rise to a rapid fall in end-tidal CO,. Breath-by-breath C O , analysis permits the early detection of this fall. The amount of air necessary to cause a fall in end-tidal CO, has been variously reported. 1.5 ml/kg of intravenous air in dogs gave an obvious fa11I7and 0.18 ml/kg in a dog gave a fall of 0.3 volumes p e r ~ e n t . ~ For the diagnosis of venous air embolism in the Regional Neurosurgical Unit at the Brook Hospital, the authors have, until recently, relied on monitoring pulse rate, blood pressure, electrocardiagram, respiratory pattern and heart sounds. In February, 1977 an infrared COz analyser became available and this paper describes clinical experience in its use. Materials and methods
Between February and November 1977 twenty patients were anaesthetised for neurosurgical operations in the sitting position. The age range was from 23 to 71 years, and the operations carried out were cervical laminectomy six patients, posterior fossa explorations eleven patients, and Torkildsen’s operation three patients. All the patients were unpremedicated. Anaesthesia was induced with a sleep dose of
David Hurter, MB, BS, FFARCS, Consultant Anaesthetist, Neurosurgical Unit, Brook Hospital. Peter S. Sebel, MB BS, FFARCSI, Registrar in Anaesthetics, King’s College Hospital. Correspondence to Dr Hurter, Brook General Hospital, Shooter’s Hill Road, London, S.E.18.
578
0003-2409/79/0600-0S78$02.0001979 Blackwell Scientific Publications
Detection of venous air embolism
579
580
D. Hurter and P.S.Sebel
thiopentone, followed immediately by 100 mg of suxamethonium. When paralysis was complete, the lungs were ventilated with oxygen via a face-piece. A disposable 18-gauge Portex oesophageal stethoscope was then passed and its position adjusted to give minimum intensity of breath sounds, with maximum intensity of heart sounds. The larynx and trachea were sprayed with 4% lignocaine and a wide bore, cuffed, nylon reinforced, latex orotracheal tube passed under direct vision. Anaesthesia was maintained with nitrous oxide 7 litrelmin, oxygen 3 litrelmin and halothane 1 %via a Magill circuit (Mapleson A). The patients were allowed to breathe spontaneously. Ten milligram doses of pethidine were given intravenously as required. One patient (uide infra patient 2). for whom at the start of anaesthesia it was not anticipated that the operation would be carried out in the sitting position, was ventilated with nitrous oxide 7 litrelmin, oxygen 3 litrelmin, 0.5% halothane and incremental doses tubocurarine with a tidal volume of 700 ml. An intravenous infusion of Hartmann’s solution was established. Either a dorsalis pedis or radial artery was cannulated with an 18gauge Medicut cannula and connected with manometer tubing to an aneroid gauge for the continuous display of mean arterial pressure. The patients were placed in the sitting position and lead I1 of the ECG monitored as continuously as diathermy use would allow. The rubber cap of a Magill suction union connector was pierced by a 12-gauge Medicut cannula. A plastic catheter from a Bard-e-Cath was passed through this until it lay just short of the tip of the tracheal tube. This was connected with Portex manometer tubing via a water trap to the sampling inlet of a C 0 2 analyser (P.K. Morgan Ltd), which had been previously
calibrated using room air. Calibration with anaesthetic gas mixtures was not carried out as changes in end-tidal C 0 2 were considered more relevant than absolute values. The sampling rate was adjusted to between 0.5 and 1 litrelmin to give an acceptable suction pump noise level. A continuous permanent record of COz concentration was also obtained. Neck compression was cariied out at intervals throughout the operations. The line of skin incision was infiltrated with 1% lignocaine, with adrenaline 1 in 400,000, in all patients. Figure 1 shows the arrangement of the anaesthetic machine and monitoring devices. The apparatus is arranged to display everything the anaesthetist needs to see within one field of vision, and as nearly as possible at eye level. Case reports
Of the twenty patients undergoing operations in the sitting position, two patients had apparent air embolism.
Patient 1 Figure 2 is the end-tidal CO, tracing from a patient aged 44 undergoing cervical laminectomy to relieve spinal cord compression. While the muscles were being stripped from the spine, there was a sudden slight fall in end-tidal C 0 2 (Point A). Neck compression was immediately instituted and the surgeon alerted. At this time a splashing sound coincidental with each heart beat was heard via the oesophageal stethoscope, confirming the diagnosis of air embolism. The site of air entry was not found but, as the endtidal COa was returning to its former level and heart sounds had returned to normal, the surgeon proceeded with the operation. At Point
Fig. 2. End-tidal C02 trace from patient 1.
Detection of venous air embolism B, a further sudden and dramatic fall in endtidal CO, occurred. Air was heard splashing in the heart and crepitus was felt on neck compression, indicating air in the jugular veins. Apart from a minimal rise in pulse rate (less than 10 beatslmin) and a minimal fall in mean arterial pressure (less than 10 mmHg) there were no other changes in monitored signs. An open vein was found and diathermied. At Point C the patient stopped breathing. The operation was discontinued and the patient ventilated manually until Point D, when spontaneous respiration returned and all signs of venous air embolism had disappeared. The rest of the operation was uneventful and the patient made a full recovery, with no sequelae of the air embolism.
Patient 2 Figure 3 is the tracing from the only patient in the series who was ventilated. The patient, a man aged 40 years, had undergone lumbar pneumoencephalography under general anaesthesia immediately prior to the operation of ventriculocisternostomy (Torkildsen's operation), to relieve internal hydrocephalus caused by a brain stem tumour. This patient was initially thought to have a cerebello-pontine angle tumour, which it was planned to remove with him in the lateral position. He was, therefore, ventilated from the outset and it was considered unwise to change to spontaneous respiration for the sitting up part of the procedure. At Point A, while muscle was being stripped from the occipital bone, there was a slight fall in end-tidal CO,. The presence of air in the heart could not be confirmed owing to technical difficulties with the oesophageal stethoscope. The surgeon did find and diathermy an open vein, after which the end-tidal CO, returned to its former level. It was assumed that the fall in end-tidal CO, was due to an air embolus because, allowing for the difference in the
581
tracing between spontaneous and controlled ventilation, this tracing is similar to that of Figure 2, Point A. The patient recovered following the procedure and showed no detectable sequelae of the assumed air embolism.
Discussion This series of patients is too small to draw any far reaching conclusions. It does, however, appear that end-tidal CO, measurement is a reliable method for the early detection of air embolus. No changes occurred in other monitored signs suggestive of air embolus without an associated fall in the end-tidal CO,. Some authorities'8 consider that it is mandatory to insert a central venous catheter in these patients. This is not the authors' practice for the following reasons. Firstly, there is no point in introducing a central venous catheter unless the position of its tip is in the right place for aspirating air and the air may be in either the right atrium or ventricle. Secondly, the rationale is in doubt. A small volume of air is rapidly dispersed in the circulation so that it is unnecessary, and probably impossible, to aspirate it via a central venous catheter. It is only those patients who have a massive air embolism and who develop signs of gross circulatory failure from whom it is necessary to aspirate the offending air. It is hard to believe that enough air would be aspirated, at sufficient speed, via a central venous catheter, to have any beneficial effect on the circulation. In such a desperate situation the only appropriate course of action is to open the chest and aspirate directly from the heart. Discussions on the pros and cons of operating on patients in the sitting position continue wherever and whenever anaesthetists and neurosurgeons meet. There is no doubt that for ease of surgical access and comfort of the surgeons the position carries distinct advantages over all others. There are, however, some who maintain that the hazard of venous air
Fig. 3. End-tidal COz trace from patient 2.
582
D.Hurter and P.S. Sebel
embolism outweighs the advantages of the position and precludes its use. The authors believe that provided all members of the team, surgeons, anaesthetists and theatre staff, are fully aware of the hazard and the importance of early recognition and prompt treatment, there are some operations on some patients for which the sitting position carries the best prognosis. The decision to use the sitting position is one that should be taken by surgeons and anaesthetists together on each occasion. Having decided on its use, it is the duty of all members of the team to cooperate in reducing the hazard to a minimum. In recent years the introduction of more reliable and sensitive means of monitoring those vital signs which may change when air embolus occurs have further reduced the hazard. The introduction of continuous C o t monitoring has proved an important, essential addition to the anaesthetists' armamentarium which should reduce the mortality and morbidity previously associated with venous air embolism.
References MICHENFELDER, J.D., MARTIN,J.T., ALTENBURG, B.M. & REHDER,K. (1969) Air embolism during neurosurgery. An evaluation of right-atrial catheters for diagnosis and treatment. Journal of the American Medical Association 208, 1353-1358. MICHENFELDER, J.D., MILLER,R.H. & GRONERT, G.A. (1972) Evaluation of an ultrasonic device (Doppler) _ _ for the diagnosis of venous air embolism. Anesthesiology 36, 164-167. J.A., GOTTLIEB,J.D. & SWEET,R.B. 3. ERICSSON, (1964) Closed-chest cardiac massage in the treatment of venous air embolism. New England Journal of Medicine, 270, 1353-1354. J., PRYS-ROBERTS, C. & ADAMS, 4. EDMONDS-SEAL, A.P. (1971) Air embolism. A comparison of various methods of detection. Anaesthesia 26, .
202-208. 5. BUCKLAND, R.W. & MANNERS. J.M. (1976) Venous
air embolism during neurosurgery. A comparison of various methods of detection in man. Anaesthesia, 31, 633-643. V.L., BETHLINE,R.W.M. & SOLDO, 6. BRECHNER, N.J. (1967) Pathological physiology of air embolism. Anesthesiology 28, 240-241. 7. WHITEY,J.D. (1963) Electrocardiography during
8.
Summary
9.
A clinical report on the detection of venous air embolism during neurosurgery is presented. The use of end-tidal CO, monitoring is described.
10.
11.
12.
Key words ANAESTHESIA; neurosurgical. EMBOLISM; air. MONITORING; end-tidal carbon dioxide.
Acknowledgments The authors would like to record thanks to Messrs G.B. Northcroft, J.R. Bartlett, G. Neil-Dwyer and R. Maurice-Williams, the neurosurgeons who carried out the operations, for their interest and co-operation. They would also like to thank Mrs Shelagh Levy for her secretarial assistance and Mr Laing for the photographs.
I
~
13.
posterior fossa operations. British Journal of Anaesthesia 35, 624-630. LEWIS,J.M. & REES,G.A.D. (1964) Electrocardiography during posterior fossa operations. British Journal of Anaesthesia 36, 63. DURANT,T.M., LONO,J. & OPPENHEIMER, M.J. (1947) Pulmonary (venous) air embolism. American Heart Journal, 33,269-281. HUNTER, A.R. (1962) Air embolism in the sitting position. Anaesthesia, 17, 467-472. MCCOMISH,P.B. & THOMPSON,D.E.A. (1968) Respiratory disturbances in air embolism. Anaesthesia, 23, 259-263. SMITHC. (1954) An endo-oesophageal stethoscope. Anesthesiology, 15, 566. MARSHALL, B.M. (1965) Air embolus in neurosurgical anaesthesia, its diagnosis and treatment. Canadian Anaesthetists' Society Journal, 12, 255-
261. J.C., GOODMAN, J.M., HORNER, T.G. & 14. MAROON, CAMPBELL,R.L. (1968) Detection of minute
15.
16.
17.
18.
venous air emboli with ultrasound. Surgery, Gynecology and Obstetrics 127, 1236-1238. EDMONDS-SEAL, J. & MAROON,J.C. (1969) Air embolism diagnosed with ultrasound. A new monitoring technique. Anaesthesia, 24, 438-439. BETHUNE, R.W.M. & BRECHNER, V.L. (1968) Detection of venous air embolism by carbon dioxide monitoring. Anesthesiology, 29, 178. BRECHNER,V.L. & BETHUNE,R.W.M. (1971) Recent advances in monitoring pulmonary air embolism. Anesthesia and Analgesia: Current Reseorches, 50, 255-261. HORTON, J. (1977) Personal communication.
Detection of venous air embolism A clinical report using end-tidal carbon dioxide monitoring during neurosurgery
DAVID HURTER
AND
Venous air embolism is a recognised hazard to patients undergoing operations in the sitting position. A widely varying incidence has been reported, from 1.6%’ to 9 3 x 2 with a mortality from 0%’ to 73%.’ This wide range almost certainly represents a gross disparity in the sensitivity of the monitoring methods used to detect air embolism. The various methods of detection have been compared in dogs4 and in man5 and the pathological physiology has been described.6 Because this complication can prove rapidly fatal, its early detection is of vital importance. Until recently, anaesthetists have had to rely on comparatively crude methods of monitoring, namely pulse rate and electrocardiography (ECG) * blood pressure, central venous p r e s s ~ r e ,respiratory ~ pattern l o * and heart sounds heard through an oesophageal s t e t h o s c ~ p e . ~Changes ~ * ’ ~ in these caused by venous air embolism are variable and, with the exception of heart sounds, non-specific. In recent years, two newer and more specific methods of detection have been introduced, the l 5 and Doppler ultrasonic flow the continuous carbon dioxide (CO,) analyser. l 6 The Doppler transducer is an extremely sensitive device which detects air bubbles in the heart in very small amounts. It may be difficult to position the probe satisfactorily on the chest, especially in the obese patient. The frequency with which it raises a false alarm by detecting
P E T E R S. S E B E L
insignificant amounts of air has proved a great disadvantage in its clinical use.’ Air in the pulmonary vessels interferes with gaseous exchange and causes retention of C O z , giving rise to a rapid fall in end-tidal CO,. Breath-by-breath C O , analysis permits the early detection of this fall. The amount of air necessary to cause a fall in end-tidal CO, has been variously reported. 1.5 ml/kg of intravenous air in dogs gave an obvious fa11I7and 0.18 ml/kg in a dog gave a fall of 0.3 volumes p e r ~ e n t . ~ For the diagnosis of venous air embolism in the Regional Neurosurgical Unit at the Brook Hospital, the authors have, until recently, relied on monitoring pulse rate, blood pressure, electrocardiagram, respiratory pattern and heart sounds. In February, 1977 an infrared COz analyser became available and this paper describes clinical experience in its use. Materials and methods
Between February and November 1977 twenty patients were anaesthetised for neurosurgical operations in the sitting position. The age range was from 23 to 71 years, and the operations carried out were cervical laminectomy six patients, posterior fossa explorations eleven patients, and Torkildsen’s operation three patients. All the patients were unpremedicated. Anaesthesia was induced with a sleep dose of
David Hurter, MB, BS, FFARCS, Consultant Anaesthetist, Neurosurgical Unit, Brook Hospital. Peter S. Sebel, MB BS, FFARCSI, Registrar in Anaesthetics, King’s College Hospital. Correspondence to Dr Hurter, Brook General Hospital, Shooter’s Hill Road, London, S.E.18.
578
0003-2409/79/0600-0S78$02.0001979 Blackwell Scientific Publications
Detection of venous air embolism
579
580
D. Hurter and P.S.Sebel
thiopentone, followed immediately by 100 mg of suxamethonium. When paralysis was complete, the lungs were ventilated with oxygen via a face-piece. A disposable 18-gauge Portex oesophageal stethoscope was then passed and its position adjusted to give minimum intensity of breath sounds, with maximum intensity of heart sounds. The larynx and trachea were sprayed with 4% lignocaine and a wide bore, cuffed, nylon reinforced, latex orotracheal tube passed under direct vision. Anaesthesia was maintained with nitrous oxide 7 litrelmin, oxygen 3 litrelmin and halothane 1 %via a Magill circuit (Mapleson A). The patients were allowed to breathe spontaneously. Ten milligram doses of pethidine were given intravenously as required. One patient (uide infra patient 2). for whom at the start of anaesthesia it was not anticipated that the operation would be carried out in the sitting position, was ventilated with nitrous oxide 7 litrelmin, oxygen 3 litrelmin, 0.5% halothane and incremental doses tubocurarine with a tidal volume of 700 ml. An intravenous infusion of Hartmann’s solution was established. Either a dorsalis pedis or radial artery was cannulated with an 18gauge Medicut cannula and connected with manometer tubing to an aneroid gauge for the continuous display of mean arterial pressure. The patients were placed in the sitting position and lead I1 of the ECG monitored as continuously as diathermy use would allow. The rubber cap of a Magill suction union connector was pierced by a 12-gauge Medicut cannula. A plastic catheter from a Bard-e-Cath was passed through this until it lay just short of the tip of the tracheal tube. This was connected with Portex manometer tubing via a water trap to the sampling inlet of a C 0 2 analyser (P.K. Morgan Ltd), which had been previously
calibrated using room air. Calibration with anaesthetic gas mixtures was not carried out as changes in end-tidal C 0 2 were considered more relevant than absolute values. The sampling rate was adjusted to between 0.5 and 1 litrelmin to give an acceptable suction pump noise level. A continuous permanent record of COz concentration was also obtained. Neck compression was cariied out at intervals throughout the operations. The line of skin incision was infiltrated with 1% lignocaine, with adrenaline 1 in 400,000, in all patients. Figure 1 shows the arrangement of the anaesthetic machine and monitoring devices. The apparatus is arranged to display everything the anaesthetist needs to see within one field of vision, and as nearly as possible at eye level. Case reports
Of the twenty patients undergoing operations in the sitting position, two patients had apparent air embolism.
Patient 1 Figure 2 is the end-tidal CO, tracing from a patient aged 44 undergoing cervical laminectomy to relieve spinal cord compression. While the muscles were being stripped from the spine, there was a sudden slight fall in end-tidal C 0 2 (Point A). Neck compression was immediately instituted and the surgeon alerted. At this time a splashing sound coincidental with each heart beat was heard via the oesophageal stethoscope, confirming the diagnosis of air embolism. The site of air entry was not found but, as the endtidal COa was returning to its former level and heart sounds had returned to normal, the surgeon proceeded with the operation. At Point
Fig. 2. End-tidal C02 trace from patient 1.
Detection of venous air embolism B, a further sudden and dramatic fall in endtidal CO, occurred. Air was heard splashing in the heart and crepitus was felt on neck compression, indicating air in the jugular veins. Apart from a minimal rise in pulse rate (less than 10 beatslmin) and a minimal fall in mean arterial pressure (less than 10 mmHg) there were no other changes in monitored signs. An open vein was found and diathermied. At Point C the patient stopped breathing. The operation was discontinued and the patient ventilated manually until Point D, when spontaneous respiration returned and all signs of venous air embolism had disappeared. The rest of the operation was uneventful and the patient made a full recovery, with no sequelae of the air embolism.
Patient 2 Figure 3 is the tracing from the only patient in the series who was ventilated. The patient, a man aged 40 years, had undergone lumbar pneumoencephalography under general anaesthesia immediately prior to the operation of ventriculocisternostomy (Torkildsen's operation), to relieve internal hydrocephalus caused by a brain stem tumour. This patient was initially thought to have a cerebello-pontine angle tumour, which it was planned to remove with him in the lateral position. He was, therefore, ventilated from the outset and it was considered unwise to change to spontaneous respiration for the sitting up part of the procedure. At Point A, while muscle was being stripped from the occipital bone, there was a slight fall in end-tidal CO,. The presence of air in the heart could not be confirmed owing to technical difficulties with the oesophageal stethoscope. The surgeon did find and diathermy an open vein, after which the end-tidal CO, returned to its former level. It was assumed that the fall in end-tidal CO, was due to an air embolus because, allowing for the difference in the
581
tracing between spontaneous and controlled ventilation, this tracing is similar to that of Figure 2, Point A. The patient recovered following the procedure and showed no detectable sequelae of the assumed air embolism.
Discussion This series of patients is too small to draw any far reaching conclusions. It does, however, appear that end-tidal CO, measurement is a reliable method for the early detection of air embolus. No changes occurred in other monitored signs suggestive of air embolus without an associated fall in the end-tidal CO,. Some authorities'8 consider that it is mandatory to insert a central venous catheter in these patients. This is not the authors' practice for the following reasons. Firstly, there is no point in introducing a central venous catheter unless the position of its tip is in the right place for aspirating air and the air may be in either the right atrium or ventricle. Secondly, the rationale is in doubt. A small volume of air is rapidly dispersed in the circulation so that it is unnecessary, and probably impossible, to aspirate it via a central venous catheter. It is only those patients who have a massive air embolism and who develop signs of gross circulatory failure from whom it is necessary to aspirate the offending air. It is hard to believe that enough air would be aspirated, at sufficient speed, via a central venous catheter, to have any beneficial effect on the circulation. In such a desperate situation the only appropriate course of action is to open the chest and aspirate directly from the heart. Discussions on the pros and cons of operating on patients in the sitting position continue wherever and whenever anaesthetists and neurosurgeons meet. There is no doubt that for ease of surgical access and comfort of the surgeons the position carries distinct advantages over all others. There are, however, some who maintain that the hazard of venous air
Fig. 3. End-tidal COz trace from patient 2.
582
D.Hurter and P.S. Sebel
embolism outweighs the advantages of the position and precludes its use. The authors believe that provided all members of the team, surgeons, anaesthetists and theatre staff, are fully aware of the hazard and the importance of early recognition and prompt treatment, there are some operations on some patients for which the sitting position carries the best prognosis. The decision to use the sitting position is one that should be taken by surgeons and anaesthetists together on each occasion. Having decided on its use, it is the duty of all members of the team to cooperate in reducing the hazard to a minimum. In recent years the introduction of more reliable and sensitive means of monitoring those vital signs which may change when air embolus occurs have further reduced the hazard. The introduction of continuous C o t monitoring has proved an important, essential addition to the anaesthetists' armamentarium which should reduce the mortality and morbidity previously associated with venous air embolism.
References MICHENFELDER, J.D., MARTIN,J.T., ALTENBURG, B.M. & REHDER,K. (1969) Air embolism during neurosurgery. An evaluation of right-atrial catheters for diagnosis and treatment. Journal of the American Medical Association 208, 1353-1358. MICHENFELDER, J.D., MILLER,R.H. & GRONERT, G.A. (1972) Evaluation of an ultrasonic device (Doppler) _ _ for the diagnosis of venous air embolism. Anesthesiology 36, 164-167. J.A., GOTTLIEB,J.D. & SWEET,R.B. 3. ERICSSON, (1964) Closed-chest cardiac massage in the treatment of venous air embolism. New England Journal of Medicine, 270, 1353-1354. J., PRYS-ROBERTS, C. & ADAMS, 4. EDMONDS-SEAL, A.P. (1971) Air embolism. A comparison of various methods of detection. Anaesthesia 26, .
202-208. 5. BUCKLAND, R.W. & MANNERS. J.M. (1976) Venous
air embolism during neurosurgery. A comparison of various methods of detection in man. Anaesthesia, 31, 633-643. V.L., BETHLINE,R.W.M. & SOLDO, 6. BRECHNER, N.J. (1967) Pathological physiology of air embolism. Anesthesiology 28, 240-241. 7. WHITEY,J.D. (1963) Electrocardiography during
8.
Summary
9.
A clinical report on the detection of venous air embolism during neurosurgery is presented. The use of end-tidal CO, monitoring is described.
10.
11.
12.
Key words ANAESTHESIA; neurosurgical. EMBOLISM; air. MONITORING; end-tidal carbon dioxide.
Acknowledgments The authors would like to record thanks to Messrs G.B. Northcroft, J.R. Bartlett, G. Neil-Dwyer and R. Maurice-Williams, the neurosurgeons who carried out the operations, for their interest and co-operation. They would also like to thank Mrs Shelagh Levy for her secretarial assistance and Mr Laing for the photographs.
I
~
13.
posterior fossa operations. British Journal of Anaesthesia 35, 624-630. LEWIS,J.M. & REES,G.A.D. (1964) Electrocardiography during posterior fossa operations. British Journal of Anaesthesia 36, 63. DURANT,T.M., LONO,J. & OPPENHEIMER, M.J. (1947) Pulmonary (venous) air embolism. American Heart Journal, 33,269-281. HUNTER, A.R. (1962) Air embolism in the sitting position. Anaesthesia, 17, 467-472. MCCOMISH,P.B. & THOMPSON,D.E.A. (1968) Respiratory disturbances in air embolism. Anaesthesia, 23, 259-263. SMITHC. (1954) An endo-oesophageal stethoscope. Anesthesiology, 15, 566. MARSHALL, B.M. (1965) Air embolus in neurosurgical anaesthesia, its diagnosis and treatment. Canadian Anaesthetists' Society Journal, 12, 255-
261. J.C., GOODMAN, J.M., HORNER, T.G. & 14. MAROON, CAMPBELL,R.L. (1968) Detection of minute
15.
16.
17.
18.
venous air emboli with ultrasound. Surgery, Gynecology and Obstetrics 127, 1236-1238. EDMONDS-SEAL, J. & MAROON,J.C. (1969) Air embolism diagnosed with ultrasound. A new monitoring technique. Anaesthesia, 24, 438-439. BETHUNE, R.W.M. & BRECHNER, V.L. (1968) Detection of venous air embolism by carbon dioxide monitoring. Anesthesiology, 29, 178. BRECHNER,V.L. & BETHUNE,R.W.M. (1971) Recent advances in monitoring pulmonary air embolism. Anesthesia and Analgesia: Current Reseorches, 50, 255-261. HORTON, J. (1977) Personal communication.
