Air embolism: diagnosis with single-photon emission tomographyand successful hyperbaric oxygen therapy

L. Droghetti1*, M. Giganti2, A. Memmo1 and R. Zatelli1

1Department of Anaesthesia and Intensive Care Medicine, S. Anna Hospital, I-44100 Ferrara, Italy.2Department of Nuclear Medicine, University of Ferrara, Ferrara, Italy

*Corresponding author

Venous air embolism may occur when the surgical ®eld is above the level of the heart. We pre-

sent a case of venous air embolism in a patient undergoing percutaneous nephrolithotripsy in

the prone position and presenting with blindness and neurological de®cits 8 h later. The clinical

diagnosis of paradoxical air embolism was con®rmed by early single-photon emission tomogra-

phy (SPET), whereas magnetic resonance imaging including diffusion-weighted imaging (DW-

MRI) was diagnostic only 30 h later. Hyperbaric oxygen therapy was successful. In this case,

early DW-MRI scan was inconclusive, but a SPET study of the brain appeared to be useful in

con®rming the clinical diagnosis. Early hyperbaric oxygen was demonstrated to be a successful

therapy.

Br J Anaesth 2002; 89: 775±8

Keywords: anaesthesia, urology; complications, air embolism; therapy, hyperbaric oxygen

Accepted for publication: July 4, 2002

During surgical procedures in which the surgical ®eld is

positioned above the level of the heart, venous air embolism

may occur. When the air passes through to the arterial

system, causing paradoxical air embolism, the conse-

quences can be devastating. We present a case of venous

air embolism occurring in percutaneous nephrolithotripsy in

the prone position, with features of paradoxical air

embolism appearing 8 h later.

Case report

A previously healthy 36-yr-old male, height 175 cm, weight

84 kg, was scheduled for nephrolithotripsy for an upper pole

staghorn calculus. Anaesthesia was induced with alfentanil

7 mg kg±1 and propofol 2 mg kg±1 i.v., and neuromuscular

block was achieved with vecuronium bromide 0.1 mg kg±1.

After tracheal intubation with a cuffed tube, arti®cial

ventilation was started and anaesthesia maintained with

60% nitrous oxide in oxygen and sevo¯urane at an end-

expiratory concentration of 1.5 vol%. Continuous intra-

operative monitoring included an electrocardiogram (ECG),

pulse oximetry (SpO2), end-tidal carbon dioxide concentra-

tion (PE¢CO2), end-tidal volatile anaesthetic agent concen-

tration, and ventilation volumes and pressures, while non-

invasive blood pressure was measured at 3-min intervals.

Cystoscopy was performed and a catheter was inserted into

the right ureter. The patient was then placed in the prone

position with his head and legs down. A nephroscope

connected to an ultrasound generator (Calculson-Storz,

Tuttlingen, Germany) at 26 000 Hz was inserted into the

renal pelvis, which was perfused constantly with saline

solution ¯owing from an irrigating bag placed 40 cm above

the surgical ®eld and drained by an aspirating pump. About

10 min after the beginning of ultrasound lithotripsy, the

PE¢CO2decreased abruptly from 4 to 2.9 then to 1.8 kPa. The

SpO2decreased from 100 to 86% and the systolic arterial

pressure from 120 to 96 mm Hg. The heart rate decreased

from 75 to 60 beats min±1. As the airway pressures and tidal

volume were unchanged, airway obstruction was excluded

and, with the strong suspicion of air embolism, the nitrous

oxide was discontinued and manual ventilation of the lungs

was performed with oxygen 100%. Within 2±3 min, the

patient was placed supine. The PE¢CO2increased to 4 kPa and

the SpO2increased to 100%. A catheter was inserted into the

left radial artery and arterial blood gas analysis showed a

PaCO2of 5.4 kPa, while the PE¢CO2

was still 3.8 kPa. The

surgery was stopped and the patient was awakened and his

trachea extubated. He was conscious, haemodynamically

stable and his SpO2was 100% while breathing air.

Neurological examination, biochemical and haematological

CASE REPORTS

British Journal of Anaesthesia 89 (5): 775±8 (2002)

Ó The Board of Management and Trustees of the British Journal of Anaesthesia 2002

by guest on Novem

ber 20, 2013http://bja.oxfordjournals.org/

Dow

nloaded from

pro®les, coagulation indices (indexed normalized ratio 1.12,

thromboplastin time 30.12 s, D dimers 269 ng ml±1 and

®brinogen 412 mg ml±1) and blood gas analysis were

normal.

Seven hours after the presumed venous air embolism, the

patient complained of nausea and retching. Metoclopramide

10 mg was administered i.m., with bene®t. At that time,

arterial blood gas analysis revealed mild hypoxaemia (PaO2

9 kPa). Two hours later the patient suddenly complained of

complete blindness, while his tendon re¯exes were in-

creased on the left side and a tremor was evident in the left

arm. Ophthalmological examination revealed bilateral

amaurosis without funduscopic abnormalities; the retinal

vessels and the optic disc were normal. The patient

underwent magnetic resonance imaging (MRI) including

diffusion-weighted imaging (DW-MRI), which showed only

a mild hyperintense area in the left cerebellar hemisphere

with cortical±subcortical extension. Nevertheless, despite

the negative MRI data, on the basis of a strong clinical

suspicion of paradoxical air embolism the patient was

transferred as rapidly as possible to a hyperbaric centre. It

took 5 h from the onset of blindness to instigate hyperbaric

therapy. During the ®rst hyperbaric session (Table 1) the

patient's sight started to recover (he could see shadows).

Thirty hours later, a new MRI showed two hyperintense

areas in the left cerebellar hemisphere with cortical±-

subcortical extension and bilateral hyperintensity in the

cortical occipital gyrus, more evident on the right side than

the left. A single-photon emission tomography (SPET)

study of the brain was performed (Fig. 1), and this

demonstrated a reduced uptake of isotopic tracer (99mTc-

HMPAO, 740 MBq) in the right cerebellar hemisphere

(Fig. 1A) and bilateral cerebral occipital (Fig. 1B) and left

frontal (Fig. 1C) defects. Immediately after the SPET study,

contrast transthoracic echocardiography was performed. No

evidence of air in the right side of the heart or of interatrial

Table 1 Scheme of the early hyperbaric treatment

Gas mixture Maximum pressurein atmospheres

Time(h:min)

(ATA)

Oxygen in helium 50% 4 8:00

Oxygen 100% 2.8 0:25

Air 2.2 0:5

Air 1.8 0:5

Fig 1 Three representative transverse slices of the SPET study of the brain with 99mTc-HMPAO, performed 31 h after paradoxical air embolism. (A)

The arrow indicates a right cerebellar hemisphere uptake defect with interhemispheric asymmetry. (B) The arrow indicates bilateral occipital defects.

(C) The arrow indicates a left frontal defect.

Fig 2 SPET study of the brain performed 13 days after injury. Mild interhemispheric cerebellar asymmetry (A) and completely normal tracer uptake in

the occipital (B) and frontal (C) cerebral regions.

Droghetti et al.

776

by guest on Novem

ber 20, 2013http://bja.oxfordjournals.org/

Dow

nloaded from

or interventricular shunt was found, even after a Valsalva

manoeuvre. Hyperbaric therapy was scheduled daily for 9

days (1.8 bar for 80 min), even though the patient recovered

completely about 48 h after the paradoxical air embolism.

Thirteen days after surgery, a further SPET study provided

evidence of complete recovery of the occipital (Fig. 2B) and

frontal (Fig. 2C) defects, only a mild interhemispheric

cerebellar asymmetry persisting (Fig. 2A).

Discussion

The phenomenon of pyelovenous back¯ow was ®rst

described in 1856 by Gigon, who noted the passage of

¯uids from the calyces into the renal veins.1 Later,

pyelovenous2 3 and tubulovenous4 back¯ow was described.

Air embolism as a complication of retrograde pyelography5

and air entering the hepatic veins after nephrolithotripsy

have been reported.6 Saline solution ¯owing from a bag

placed 40 cm above the surgical ®eld applies a pressure of

about 4 kPa to the pelvocaliceal space. This pressure is

reduced by an unknown amount by the aspirating pump.

Because of the high rate of saline ¯ow, accidental injection

of small amounts of air, in the form of microbubbles, cannot

be excluded. In addition, the high power of therapeutic

shock waves could induce cavitation,7 an ultrasound-related

mechanical effect. The prone position8 of the patient in this

case produced a signi®cant gravitational gradient between

the right side of the heart and the renal pelvis, and air could

thus have been drawn into open veins by negative pressure.

This case demonstrates the problems associated with the

combination of a gravitational gradient,9 decreased caval

pressure due to the position of the lower limbs, a likely high

pressure5 of irrigating ¯uid, possibly containing air

bubbles,10 11 and ultrasound shock waves.12

Air bubbles, the surfaces of which are covered by a

network of ®brin, platelets and fat globules, induce

neutrophil-mediated microvascular damage, activate the

intrinsic coagulation cascade and obstruct lung capillaries.

This will cause an increased physiological dead space,13 14

disordered ventilation±perfusion matching15 and reduced

cardiac output as a result of right ventricular out¯ow

obstruction, leading to decreased PE¢CO2, SpO2

and systolic

arterial pressure, together with an increased end-tidal arterial

carbon dioxide gradient. In our case, we assumed venous air

embolism rather than pulmonary thromboembolism, because

coagulation was normal and there was a rapid improvement

after ventilation with oxygen 100%. The oxygen concentra-

tion gradient could have facilitated the release from the air

embolus of both nitrogen and nitrous oxide.

Eight hours after the venous air embolism, the patient

displayed the features of the air having moved to the arterial

side of the circulation, a phenomenon described most

frequently in acute decompression illness after diving.

We used PE¢CO2to detect air embolism.16 Doppler

transthoracic ultrasound is more sensitive, but only early

transoesophageal echocardiography might have demon-

strated the presence of air which could progress through

either a patent foramen ovale or through pulmonary shunts16

to cause paradoxical air embolism. In our case, transoeso-

phageal echocardiography did not demonstrate right-to-left

interatrial shunting. However, the delay between venous and

arterial embolic episodes suggests transpulmonary passage,

which has been described in dogs17 and humans18 and has

also been demonstrated by transoesophageal echocardio-

graphy.19 20 Pathways involved in transatrial or transpul-

monary air transport become functionally open only during

episodes of venous air embolism in which signi®cant

elevation of pulmonary artery pressure occurs.19 In our

case, such an increase in pulmonary artery pressure could

have been induced during the retching episodes. While the

patient was in the head-up position, air crossing to the

systemic arterial circulation could have migrated up to the

carotid and cerebellar arteries and then to the cerebral and

cerebellar hemispheres.18

Although DW-MRI is widely recognized to be the earliest

imaging technique that detects brain ischaemia, in our case

the early MRI scan (performed immediately after the onset of

symptoms) did not show abnormalities consistent with

ischaemic tissue, as described in a similar case report.21

These ®ndings could be related to the small dimensions of the

injured areas21 at the time of the ®rst MRI scan. The later DW-

MRI scan demonstrated abnormalities consistent with

ischaemic brain damage in the occipital and cerebellar

cortex. The SPET study of the brain con®rmed these ®ndings

with enhanced sensitivity in indicating the extent and

localization of the brain damage, even though the total

number of counts recorded was only 9 000 000, with a reduced

acquisition time (Fig. 1). The second SPET study (Fig. 2),

which recorded a total of 13 000 000 counts with a standard

acquisition time, provided de®nitive evidence of recovery.

In neurosurgical procedures in which the patient is in the

sitting position, invasive monitoring to detect venous air

embolism and the insertion of a catheter in the right atrium

to aspirate air are used routinely. In other surgical proced-

ures,10 such as nephrolithotripsy, these precautionary

measures may not be so easily justi®ed. However, our

case suggests that transoesophageal echocardiography can

demonstrate the presence of air in the right heart or

pulmonary veins, and this should be removed from the

venous circulation through a right atrial catheter. In the case

of persistent air trapping, prophylactic hyperbaric oxygen

therapy should be performed. Until prospective studies

provide an estimate of the true incidence of venous air

embolism and paradoxical air embolism, it is dif®cult to

decide whether transoesophageal echocardiography and

central venous catheterization should be used routinely

during percutaneous lithotripsy.

References1 Lopez FA, Dalinka M, Doboy JG. Pyelovenous back¯ow in the

human kidney. Facts, fallacies and signi®cance. Urology 1973; 2:612±14

SPET and hyperbaric oxygen in air embolism

777

by guest on Novem

ber 20, 2013http://bja.oxfordjournals.org/

Dow

nloaded from

2 Hinman F, Lee-Braun RK. Pyelovenous back¯ow. JAMA 1924; 82:607±9

3 Olsson O. Studies on back-¯ow in excretion urography. ActaRadiol 1948; 29: 3±6

4 Baghavan BS, Wenk RE, Dutta D. Pathways of urinary back¯ow inobstructive uropathy. Hum Pathol 1979; 10: 669±83

5 Pyron CL, Segal AJ. Air embolism: a potential complication ofretrograde pyelography. J Urol 1983; 130: 125±6

6 Miller RA, Kellett MJ, Wickham JE. Air embolism, a newcomplication of percutaneous nephrolithotomy. J Urol (Paris)1984; 90: 337±9

7 American Institute of Ultrasound Medicine. Mechanical bioeffectsfrom diagnostic ultrasound. AIUM Consensus statements. JUltrasound Med 2000; 19: 69±168

8 Albin MS, Ritter RR, Pruett CE, Kalff K. Venous air embolismduring lumbar laminectomy in the prone position: report ofthree cases. Anesth Analg 1991; 73: 346±9

9 Albin MS, Ritter RR, Reinhart R, Erickson D, Rockwood A.Venous air embolism during radical retropubic prostatectomy.Anesth Analg 1992; 74: 151±3

10 Hofsess DW. Fatal air embolism during transurethral resection. JUrol 1984; 131: 355

11 Vacanti CA, Lodhia KL. Fatal massive air embolism duringtransurethral resection of the prostate. Anesthesiology 1991; 74:186±7

12 Delius M, Enders G, Xuan Z, et al. Biological effects of shock

waves: kidney damage by shock waves in dogsÐdosedependence. Ultrasound Med Biol 1988; 14: 117±22

13 Presson RG, Kirk KR, Hasselby KA, et al. Fate of air emboli in thepulmonary circulation. J Appl Physiol 1989; 67: 1898±902

14 Albertine KH, Wiener-Kronish JP, Koike K, et al. Quanti®cationof damage by air emboli to lung microvessels in anesthetizeddogs. J Appl Physiol 1984; 54: 1360±8

15 Orebaugh SL. Venous air embolism: clinical and experimentalconsiderations. Crit Care Med 1992; 20: 1169±77

16 Chang J, Albin MS, Bunegin L, Hunt T. Analysis and comparisonof venous air embolism detection methods. Neurosurgery 1980;7: 135±41

17 Buttler BD, Hills BA. Transpulmonary passage of venous airemboli. J Appl Physiol 1985; 59: 543±7

18 Tommasino C, Rizzardi R, Beretta L, Venturino M, Piccoli S.Cerebral ischemia after venous embolism in the absence ofcardiac defects. J Neurosurg Anesth 1996; 8: 30±4

19 Bedell EA, Berge KH, Losasso TJ. Paradoxic air embolism duringvenous embolism: transesophageal echocardographic evidenceof transpulmonary air passage. Anesthesiology 1994; 80: 947±50

20 Thrackeray NM, Murphy PM, McLean RF, DeLacy JL. Venous airembolism accompanied by echocardiographic evidence oftranspulmonary air passage. Crit Care Med 1996; 24: 359±61

21 Sayama T, Mitani M, Inamura T, Yagi H, Fukui M. Normaldiffusion weighted imaging in cerebral air embolism complicatingangiography. Neuroradiology 2000; 42: 192±4

British Journal of Anaesthesia 89 (5): 778±82 (2002)

Epidural abscess complicating insertion of epidural catheters

J. M. G. Phillips1, J. C. Stedeford2, E. Hartsilver and C. Roberts*

Gloucestershire Royal Hospital, Great Western Road, Gloucester GL1 3NN, UK

1Present address: Musgrove Park Hospital, Taunton, Somerset, UK

2Present address: Bristol Royal In®rmary, Bristol, UK

3Present address: Royal Devon and Exeter Hospital, Exeter, UK

*Corresponding author

We present three cases of epidural abscess, all in patients in whom an epidural catheter had

been inserted for postoperative pain management. In all three cases the infecting organism was

Staphylococcus aureus and two patients had diabetes. The diagnosis was made within 3 days of

epidural catheter removal in two cases, but in one the abscess did not present until after the

patient had been discharged from hospital. We have retrospectively calculated the incidence of

epidural abscess in our hospital over the 5-yr period 1993±98 to be 1 in 800 (0.12%). We

emphasize the importance of using techniques that minimize the risk of bacterial contamination

during both catheter placement and the management of infusion, and seek to raise awareness

of this relatively rare but signi®cant condition.

Br J Anaesth 2002; 89: 778±82

Keywords: anaesthesia; epidural; complications, epidural abscess

Accepted for publication: May 30, 2002

Epidural abscess complicating epidural catheter insertion

Ó The Board of Management and Trustees of the British Journal of Anaesthesia 2002

by guest on Novem

ber 20, 2013http://bja.oxfordjournals.org/

Dow

nloaded from