Unrecognized oesophageal intubation is a complication of anaesthesia  that can have disastrous consequences if undetected . Many devices and methods [3–5] have been described to avoid it, but currently capnography provides the best method of confirming tracheal intubation. However, it is not regularly available in wards, to paramedics in the field or in many developing countries. The oesophageal detector device described by Wee is a safe and very reliable method which is often used when capnography is not available . Colourimetric devices are inexpensive and disposable and have proven to be an effective method when used to confirm the correct placement of endotracheal tubes . The devices are small and fit easily onto the endotracheal tube. Clinical diagnosis by auscultation is not always reliable .
This study was designed as a blinded, randomized, controlled trial to test the hypothesis that deliberately advancing an endotracheal tube to 28 cm or until resistance is felt would be a reliable way of distinguishing tracheal from oesophageal intubation. If the tube is correctly placed in the trachea and then advanced, bronchial intubation will occur. Hand ventilation of the lungs would then produce unilateral breath sounds and a unilateral chest movement. It was postulated that oesophageal intubation would never result in unilateral signs.
The study protocol was approved by the Hospital Medical Ethics Committee and written patient consent was obtained. We enrolled patients between the ages of 18 and 65 yr (ASA I—II) requiring muscle relaxation and tracheal intubation for elective procedures. Patients were excluded if they were anticipated to pose a ‘difficult airway’, had gastrointestinal reflux or respiratory disease. Patients were then randomized to either a tracheal/bronchial group or to an oesophageal group. Anaesthetic management was standardized. Anaesthesia was induced with propofol 2–3 mg kg−1 and fentanyl 1 μg kg−1. If successful hand ventilation was achieved, vecuronium 0.1 mg kg−1 was given to facilitate orotracheal intubation. Lung ventilation was controlled and anaesthesia maintained with isoflurane 0.5 minimum alveolar concentration in 66% nitrous oxide and oxygen for 5 min. In both groups, the trachea was then intubated with a size 6 cuffed endotracheal tube under direct laryngoscopy by the investigator (G. B. or J. S.) and confirmed by capnography. A size 6 endotracheal tube was chosen because it was decided to use the smallest size that would allow safe ventilation and make insertion of the second tube possible. In the tracheal/bronchial group, the tube was then advanced to 28 cm or until resistance was felt and the cuff gently inflated. In the oesophageal group, the tracheal tube was fixed and a second size 6 tube was placed in the oesophagus and advanced to 28 cm. The anaesthetic breathing system was attached to the trachea in the tracheal/bronchial group and to the oesophageal tube in the oesophageal group. The patient's head was then concealed under a drape.
A blinded observer was brought into the anaesthetic room and asked specifically to decide whether there were unilateral breath sounds, bilateral breath sounds or no breath sounds. Two hand-ventilated breaths were delivered to each lung. The observer was then asked specifically to decide if there was unilateral, bilateral or no chest movement following a further three hand-ventilated breaths. In our test, unilateral breath sounds or movement suggest bronchial (and therefore tracheal) intubation, whereas bilateral or no breath sounds/chest movement would suggest oesophageal intubation.
Before removal of the oesophageal tube, suction was applied to aspirate the introduced gas despite there never being any visible signs of gastric distension.
In the tracheal/bronchial group, the endotracheal tube was withdrawn until a satisfactory position was obtained.
The test was performed in 42 patients by differently qualified observers. Five Senior House officers, 10 Specialist Registrars, 10 Consultant Anaesthetists and 17 other non-medical staff, including operating department assistants and theatre nurses, took part. The tracheal/bronchial group had 14 males and 8 females, and the oesophageal group 8 males and 12 females.
The endotracheal tube positions of 20 of 22 patients in the tracheal/bronchial group were correctly identified following auscultation (Table 1). According to our test, if unilateral breath sounds were heard after tube advancement, bronchial and therefore tracheal intubation must have occurred with a specificity of 1.0 (95% CI 0.83–1.0) and a sensitivity of 0.91 (0.71–0.99). The converse is not quite true as on two of 22 occasions a negative test occurred when the tube had been advanced in the trachea. The negative test both times was ‘bilateral’ breath sounds and not ‘no’ breath sounds.
In the oesophageal group, 100% (20/20) of breath sounds were not unilateral and therefore correctly identified as being in the oesophagus. In 11 patients, breath sounds were bilateral and in 9 patients no breath sounds were heard.
Chest movement was less diagnostic (Table 2). In the tracheal/bronchial group, 10 of 22 placements were correctly identified. Thus, in this group the presence of unilateral chest movement had a sensitivity of 0.45 (95% CI 0.24–0.67) and specificity of 0.80 (0.56–0.94) for endotracheal intubation. In the oesophageal group, 4 patients were seen to have unilateral chest movement, 4 patients had bilateral chest movement and 12 revealed no movement. There were no adverse events.
We have demonstrated that this test is a valid way of excluding oesophageal intubation in the absence of capnography. The specificity of 1.0, with no falsepositives, emphasizes its clinical usefulness. It is inexpensive and possible to perform whenever intubation has been attempted. Where there is doubt over the placement of the tube in the trachea, the presence of unilateral breath sounds after the tube is advanced reliably excludes oesophageal intubation.
Following confirmation of unilateral breath sounds, the endotracheal tube should be withdrawn until unilateral breath sounds are obtained.
On two occasions in the tracheal group, bilateral sounds were heard and this was probably due to inadequate advancement and therefore failure of endobronchial intubation. We also confirmed Scott's finding  from the oesophageal group that bilateral breath sounds can be heard when oesophageal intubation has occurred. Fifty-five per cent of patients in the oesophageal group had bilateral breath sounds demonstrating that auscultation on its own is not an infallible way to diagnose tracheal intubation.
Capnography is still the best method of diagnosing correct placement and if the correct position is in doubt the tube should be removed. This is safest practice when time and equipment are available. In resuscitation situations and where capnography is not available, we feel that our test might be of assistance in helping one to make the correct diagnosis. It is simple to perform and, as demonstrated from our observers, can be used by most non-medical personnel.
1. Edwards G, Morton HJV, Pask EA, Wylie WD. Deaths associated with anaesthesia. A report on 1000 cases. Anaesthesia
2. Peterson AW, Jacker LM. Death following inadvertent oesophageal intubation
: a case report. Anestb Analg
3. Wee MYK. The oesophageal detector device. Assessment of a new method to distinguish oesophageal from tracheal intubation
4. Andersen KH, Hald A. assessing the position of the tracheal tube. The reliability of different methods. Anaesthesia
5. Birmingham PK, Cheney FW, Ward RJ. Esophageal intubation
: a review of detection techniques. Anesth Analg
6. MacLeod BA, Heller MB, Gerard J, Yealy DM, Menegazzi JJ. Verification of endotracheal tube placement with colorimetric end-tidal CO2
detection. Ann Emerg Med
7. Andersen KH, Schultz-Lebahn T. Oesophageal intubation
can be undetected by auscultation of the chest. Acta Anaesth Scand
8. Scott DB. Endotracheal intubation
: friend or foe. BMJ