Reliability of epigastric auscultation to detect gastric insufflation

J. Brimacombe*,1, C. Keller2, S. Kurian1 and J. Myles1

1Department of Anaesthesia and Intensive Care, Cairns Base Hospital, The Esplanade, Cairns 4870, Australia 2Department of Anaesthesia and Intensive Care Medicine, Leopold-Franzens University, A-6020 Innsbruck, Austria*Corresponding author

{dagger} LMA® is the property of Intavent Limited.

Accepted for publication: September 7, 2001


    Abstract
 Top
 Abstract
 Introduction
 Methods and results
 Comment
 References
 
Background. We studied the reliability of epigastric auscultation to detect gastric insufflation in 30 anaesthetized, paralysed intubated patients.

Methods. A 16FG gastric tube was positioned with the tip in the mid-oesophagus with the proximal end attached to an injection port with a one-way valve. Four observers participated in the study. Observers were paired and each pair studied 15 patients. Each patient underwent four test sequences in random order, two by each observer. Each test sequence comprised one observer injecting different volumes of air (0.25 ml, 0.5 ml, 1 ml, 2 ml, 3 ml, 4 ml, 5 ml, 10 ml, 15 ml and 0 ml as a control) in random order whilst the second blinded observer listened with a stethoscope over the epigastrium. Each randomized volume was injected rapidly at 5 s intervals for 1 min. The number of injections required to detect air entering the stomach was recorded. The stomach was deflated between each test sequence.

Results. To detect air entering the stomach with 95% confidence, 11 injections were required for 0.25 ml; 7 for 0.5 ml; 3 for 1 ml; 2 for 2 ml and 3 ml, and 1 for >=4 ml. The mean (range) inter- and intraobserver reliability was 0.73 (0.71–0.75) and 0.76 (0.76–0.89), respectively. The incidence of false positives was 21% (25/120) and the incidence of false negatives was 10% (103/1080), making the specificity and sensitivity 79% and 91%, respectively.

Conclusions. We conclude that epigastric auscultation can detect gastric insufflation of 0.25 ml air after 11 breaths and >=4 ml air after one breath with 95% confidence. Inter- and intraobserver reliability is moderate to excellent. Epigastric auscultation should be repeated to reduce the risk of false positives.

Br J Anaesth 2002; 88: 127–9

Keywords: complications, gastric insufflation; ventilation, positive pressure


    Introduction
 Top
 Abstract
 Introduction
 Methods and results
 Comment
 References
 
Gastric insufflation is a major hazard of positive-pressure ventilation in non-intubated patients and can predispose to regurgitation, impaired ventilation, reduced cardiac output and visceral rupture.1 In a survey of almost 40 000 general anaesthetics, 15% were with positive-pressure ventilation using the laryngeal mask airway (LMA{dagger})2 and gastric insufflation has been reported in this context.3 Epigastric auscultation is a clinical test commonly used to detect air entering the stomach during positive-pressure ventilation and it has been suggested that volumes of as little as 5 ml can be detected,4 but this has not been formally tested. We determined the reliability of epigastric auscultation to detect gastric insufflation.


    Methods and results
 Top
 Abstract
 Introduction
 Methods and results
 Comment
 References
 
With ethics committee approval and written informed consent, we studied 30 ASA I–II adult patients requiring tracheal intubation for elective surgical procedures in the supine position. Patients were excluded if they were pregnant, had a body mass index >35 kg m–2 or had oesophagogastric disease. Anaesthesia was induced with fentanyl 1 µg kg–1 and propofol 2 mg kg–1 and maintained with 1–3% sevoflurane in 100% oxygen. Muscle relaxation was with rocuronium 0.5 mg kg–1. After laryngoscope-guided tracheal intubation, a 16FG gastric tube was positioned in the oesophagus by feeding the tip 10 cm beyond the upper oesophageal sphincter using Magill’s forceps. The proximal end of the gastric tube was attached to an injection port with a one-way valve so that known volumes of air could be injected. Pre-prepared syringes were used to inject the following volumes (syringe size): 0.25 ml (1 ml), 0.5 ml (1 ml), 1 ml (1 ml), 2 ml (3 ml), 3 ml (3 ml), 4 ml (5 ml), 5 ml (5 ml), 10 ml (10 ml) and 15 ml (20 ml). A sham injection of 0 ml functioned as a control. Four observers (A, B, C and D) participated in the study. Observers were paired (A with B, C with D) and each pair studied 15 patients. To determine inter- and intraobserver variation, each patient underwent four test sequences in random order, two by each observer. Each test sequence comprised one observer injecting each of the nine volumes plus the sham volume in random order whilst the second blinded observer listened with a stethoscope over a fixed location 5 cm inferolateral to the xiphisternum. Each randomized volume was injected rapidly at 5 s intervals for 1 min or until the blinded observer indicated that air had entered the stomach. The observer was verbally notified immediately before each injection. Observations were made when the operating theatre was quiet, before or after surgery. The time when air was first heard entering the stomach was recorded. The observer injecting air noted if there were any oropharyngeal leaks during each injection by listening over the mouth. At the end of each test sequence, the stomach was deflated by advancing the gastric tube 30 cm into the stomach, applying suction, and withdrawing it 30 cm. Blinding was accomplished by a large opaque drape separating the two observers.

Sample size (n=30) was based on data from a pilot study with eight patients with an injected volume of 0.5 ml (reporting a difference of the mean of 1.5 and a SD of 2.7) for a type I error of 0.05 and a power of 0.9. The distribution of data was determined using Kolmogoroff–Smirnov analysis.5 Inter- and intraobserver reliability was analysed using intraclass correlation coefficient (ICC).6 Scores for statistical measurements with the ICC range from 0 to 1 where 0 shows no reliability and 1 shows perfect reliability. A score >=0.75 is graded as excellent reliability, 0.41–0.74 as moderate reliability and <0.40 as poor reliability.7

The mean (range) age, height and weight were 38 (18–81) yr, 172 (151–193) cm and 69 (48–110) kg, respectively. The male:female ratio was 16:14. There were no missing data points. To detect air entering the stomach with 95% confidence, 11 injections were required for 0.25 ml; 7 for 0.5 ml; 3 for 1 ml; 2 for 2 ml and 3 ml; and 1 for >=4 ml. The interobserver reliability between A and B was 0.71 and between C and D was 0.75, with an overall reliability of 0.73. The intraobserver reliability was 0.89, 0.76, 0,76 and 0.76 for observer A, B, C and D, respectively, with an overall value of 0.76. The incidence of false positives was 21% (25/120) and the incidence of false negatives was 10% (103/1080), making the specificity and sensitivity 79% and 91%, respectively. The incidence of false negatives for volumes >=4 ml after one breath was 4% (16/360) and 0% after four breaths. No oropharyngeal leaks were detected.


    Comment
 Top
 Abstract
 Introduction
 Methods and results
 Comment
 References
 
We found that epigastric auscultation is a reliable technique for detecting gastric insufflation. It is more reliable than chest auscultation in detecting the position of a double-lumen tube8 and in assessing breath sounds,9 but less reliable than neck stethoscopy for detecting oropharyngeal leak of gas during positive-pressure ventilation with the LMA.10 Our finding that >=4 ml air can be detected after one injection with 95% confidence is similar to the figure of 5 ml alluded to by Lawes et al.4 The smaller the volume of air entering the oesophagus, the greater the number of injections required before detection. We speculate that gas accumulates in the oesophagus and then vents into the stomach once a critical volume is reached. Our data suggest that this critical volume is 3–4 ml. If true, volumes smaller than 0.25 ml should be detectable provided adequate listening time is allowed. We found that inter- and intraobserver reliability was moderate to excellent. One individual had a higher intraobserver variability (0.89) than the others, suggesting that some individuals are better at detecting gastric insufflation than others. Interestingly, we detected no oropharyngeal leaks during the test sequence. This suggests that once air enters the oesophagus in anaesthetized paralysed patient it tends to enter the stomach.

Our experimental model does not precisely mimic the clinical situation in which gastric insufflation occurs because patients were intubated and a gastric tube was in situ. However, tracheal intubation was necessary for airway protection and to ensure that no additional air leaked into the oesophagus, as might have occurred if a face mask or LMA was used. The gastric tube was necessary to allow known volumes of air to be injected into the oesophagus. Although unlikely, it is also possible that the gastric tube may have interfered with the dynamics of the lower oesophageal sphincter and thus influenced the quality of sound generated as air passed across it into the stomach.

We found a false-positive rate of 21% and a false-negative rate of 10%. This implies that in approximately one in five occasions the clinician will hear air entering the stomach when in fact it is not, and in approximately 1 in 10 occasions will not hear air entering the stomach when in fact it is. Repeated auscultation is likely to reduce the false-positive rate. The clinical importance of small volumes of air entering the stomach with every breath is unknown. We speculate that the healthy stomach will be able to cope with some gas load without insufflation through absorption and movement of the gas into the small bowel. Cricoid pressure is an effective method of preventing gastric insufflation during ventilation with the face mask and probably with other extraglottic airway devices.1

We conclude that epigastric auscultation can detect gastric insufflation of 0.25 ml air after 11 breaths and >=4 ml air after one breath with 95% confidence. Inter- and intraobserver reliability is moderate to excellent. Epigastric auscultation should be repeated to reduce the false-positive rate.


    References
 Top
 Abstract
 Introduction
 Methods and results
 Comment
 References
 
1 Brimacombe J, Berry A. Cricoid pressure. Can J Anaesth 1997; 44: 414–25[Abstract]

2 Verghese C, Brimacombe J. Survey of laryngeal mask airway usage in 11910 patients: safety and efficacy for conventional and nonconventional usage. Anesth Analg 1996; 82: 129–33[Abstract]

3 Wittmann PH, Wittmann FW. Laryngeal mask and gastric dilatation. Anaesthesia 1991; 46: 1083

4 Lawes EG, Campbell I, Mercer D. Inflation pressure, gastric insufflation and rapid sequence induction. Br J Anaesth 1987; 59: 315–8[Abstract]

5 Sachs L. Der Kolmogoroff-Smirnov-Test fuer die Guete der Anpassung. In: Angewandte Statistik. Berlin: Springer Verlag, 1992; 426–30

6 Rosner B. Hypothesis testing: categorical data. In: Fundamentals of Biostatistics. Belmont: Wadsworth Publishing Company, 1995; 423–6

7 Fleiss JL. Reliability of measurements. In: The Design and Analysis of Clinical Experiments. New York: John Wiley and Sons, 1986; 2–32

8 Alliaume B, Coddens J, Deloof T. Reliability of auscultation in positioning of double-lumen endobronchial tubes. Can J Anaesth 1992; 39: 687–90[Abstract]

9 Brooks D, Thomas J. Interrater reliability of auscultation of breath sounds among physical therapists. Phys Ther 1995; 75: 1082–8[ISI][Medline]

10 Keller C, Brimacombe J, Keller K, Morris R. A comparison of four methods for assessing airway sealing pressure with the laryngeal mask airway in adult patients. Br J Anaesth 1999; 82: 286–7[Abstract/Free Full Text]