addition to these deaths, two patients have suffered severe
disabling cerebral ischaemia. Although these five adverse
outcomes were all associated with the use of double-
lumen endobronchial intubation, there is no evidence
to suggest that this technique is inherently dangerous.
Double-lumen tube anaesthesia is still employed in many
centres, particularly in the UK. The collapse of the lung
allows excellent visualization of the chest cavity, without
interference from a moving, ventilated lung. It is therefore
ideal when learning or teaching ETS. However, accurate
placement of the double-lumen tube is not easy and may
require checking by bronchoscopy. The necessary skills
may be lacking in units in which thoracic surgery is not
performed regularly. In such circumstances there may be
an argument for operating on the two sides on separate
occasions

5

.

It is worthy of note that in a worldwide context

double-lumen intubation has gradually given way to
standard two-lung anaesthesia. For example, a survey of
1556 procedures from 12 centres in Taiwan reported

that single-lumen endotracheal intubation was the usual
method, although surgeons occasionally requested the
double-lumen approach

6

. With increasing experience

there is a move to simpler anaesthetics. In a recent report
of 734 bilateral cases from Brazil, the authors initially used
Robertshaw tubes placed under bronchoscopic control but
the use of double-lumen catheters was slowly discontinued
during the course of the study in favour of simple
orotracheal catheters

7

.

The last of the nine deaths remains unexplained. A

patient collapsed several hours after an uneventful ETS.
Although post-mortem examination showed thrombosis in
the cerebral circulation, the exact cause of the death could
not be established.

Short-term morbidity

The most common perioperative complication is pneu-
mothorax; up to 75 per cent of patients have some residual
gas in the thorax at the end of the procedure. Although this
mostly resolves spontaneously, temporary tube drainage
is required in 0

·4-2·3 per cent of patients

8,9

. Of 44

patients having thoracoscopic splanchnicectomy for pain,
two needed a drain, but this was an extensive operation
involving resection from T5 to T12

10

. The incidence of

tension pneumothorax after surgery is unknown, but fortu-
nately this problem seems rare. The cause of postoperative
pneumothorax is usually either direct trauma to the lung at
the time of trocar insertion or tearing of an apical adhesion
as the lung is depressed. Occasionally apical bullae are seen;
it is possible that rupture of a bulla may occur as a conse-
quence of anaesthesia, especially if high inflation pressures
are used at the end of the procedure. Prevention involves
adequate reinflation of the lungs at the end of the operation
and chest radiography approximately 4 h later to rule out
significant pneumothorax. If present, chest intubation and
underwater seal drainage for 24 h should suffice.

Surgical emphysema is another fairly common periop-

erative complication. It occurs in up to 2

·7 per cent of

patients, with or without a pneumothorax

11

. It is usually

noted around the site of trocar insertion and is confined to
the chest wall, but rarely it involves the mediastinum and
tracks retroperitoneally, even as far as the scrotum in men.
Emphysema is usually obvious clinically, but chest radio-
graphy is required to rule out associated pneumothorax.
If absent, management is conservative. Segmental collapse
or atelectasis occurs occasionally. Lin and Fang

12

reported

four such problems in a series of 1360 patients. Recov-
ery was rapid with chest physiotherapy. Pleural effusion
is also encountered occasionally. Reports vary from 0 to

Copyright

 2004 British Journal of Surgery Society Ltd

www.bjs.co.uk

British Journal of Surgery 2004; 91: 264-269

Published by John Wiley & Sons Ltd

266

T. A. Ojimba and A. E. P. Cameron

1 per cent but, as patients tend to be discharged quickly,
the incidence may be much higher.

Apart from the deaths mentioned above, reports of

serious intraoperative bleeding are rare. Gossot et al.

8

described one laceration of the subclavian artery in
940 sympathectomies. It was managed by immediate
thoracotomy and suture. In the same series there was
a 5

·3 per cent incidence of significant bleeding (blood

loss of 300-600 ml) that was controlled thoracoscopically.
No patient needed transfusion. In a collected series of
7017 cases from 50 Japanese institutions, Ueyama et al.

13

reported a 0

·3 per cent incidence of intraoperative bleeding

(amount not defined), with six patients (0

·1 per cent)