Position, position, position – terminology during stomach ultrasound in pregnant women

The estate agents’ mantra ‘Location, location, location’ indicates the paramount importance of a property's geographical situation – many other aspects of a building may be changed, but this is immutable. Body position is crucial for many physiological functions, especially the stomach, however confusion in terminology is widespread. This is especially pertinent when interpreting research in pregnant women, as clinicians already use a number of modifications of body position to reduce supine aortocaval compression. Anaesthetists have an interest in the volume (and type) of stomach contents, as these may be regurgitated or vomited peri-operatively, and aspirated into the lungs. Ultrasound can be used to image the stomach, and various quantitative and qualitative measurements have been used to assess stomach contents. However, gastric contents move freely under the influence of gravity, and hence these investigations should be carried out in precisely defined positions to allow reliable imaging and consistent measurement. In probably the first published study using ultrasound to image the stomach, Bolondi et al. used a single measurement of antral cross-sectional area to investigate stomach emptying in non-pregnant subjects. They used the supine position in all subjects, but in about half of the subjects they also compared this to measurements in an unspecified ‘upright’ position, noting that food contents move into the antrum in the latter position 1. The antrum is also physically dependent in the right lateral decubitus position, and the correlation between volume of stomach contents and antral cross-sectional area is better in this position than supine; by administering known volumes of water to fasted subjects, Perlas et al. showed that there was a linear relationship for volumes up to 300 ml 2. This group have refined their model, using measurements from only the right lateral position 3. Anaesthetists often want a simple qualitative measure to determine whether a patient is at risk of a particular outcome or complication – blood pressure thresholds for cancellation of surgery being a well-known example 4. Bouvet et al. measured antral cross-sectional area pre-operatively in surgical patients, and then aspirated the stomach contents after induction of general anaesthesia 5. They proposed that an antral cross-sectional area of > 340 mm2 was likely to indicate a patient who had gastric contents of > 0.8 ml.kg−1, or whose stomach contained solid particles, which they termed a ‘risk stomach’ (Table 1). Of note, the position used for scanning was described as ‘semi-upright, with the head of the bed increased to 45°’ 5, misattributing this protocol to Bolondi et al. (quoted above) 1. Another approach has been used by Perlas et al., foregoing quantitative measurements altogether 9. They developed a simplified procedure to combine differences in gastric contents between the supine and right lateral recumbent positions into a qualitative score, indicating risk during general anaesthesia. Grade 0 is assigned when no contents are demonstrated in either supine or right lateral; grade 1 demonstrates contents collected in the antrum in the right lateral position, but no contents when dispersed in the supine position; grade 2 is applied when contents are found in both the right lateral as well as supine positions. They noted that this correlated well with their more complicated cross-sectional area measurements 9. This approach is likely to require less expertise than antral cross-sectional area measurements, but does require successful scans in two positions, as well as losing the ‘weighting’ inherent in a quantitative measurement. The applications of ultrasound in obstetric anaesthesia have been reviewed recently 10. Carp et al. were the first to describe gastric ultrasound to estimate stomach contents in pregnant women. They intentionally modified the standard assessment position in this group: “To displace the gravid uterus away from the ultrasound beam, pregnant patients were scanned in the right lateral position, with the head of the bed elevated to 45°, rather than in the sitting position…” 11. The use of bed-elevation has been followed in all studies thereafter 6, 8, 12-14. A value of 320–340 mm2 for antral cross-sectional area indicating risk was first proposed (Table 1). However, Arzola et al. recently performed a study on women just before elective caesarean section, who had been kept nil-by-mouth for > 2 h for clear fluids and > 6 h for food. They found a mean (SD) antral cross-sectional area with lateral positioning of 450 (320) mm2, contrasting with Bouvet et al. and Bataille et al. 5-7, though consistent with other authors 13, 14; they proposed a considerably higher threshold for the upper limit of normal of 1030 mm2 (Table 1). Of note, in their methodology they reported “Subjects were first placed supine and then in the right lateral decubitus, always in a 45° semi-recumbent position”; however, in the abstract they state merely that the supine and right lateral decubitus positions were used – a notable lack of precision 8. This delineation of the literature sets the background for a paper in this issue of Anaesthesia. Rouget et al. describe a study of gastric ultrasound performed before and after elective caesarean section 15. Following Arzola et al.'s findings that the antral cross-sectional area in pregnant women who have been kept nil-by-mouth for an adequate length of time might be larger than in non-pregnant subjects, they postulate that this might occur not because of a difference in the volume of stomach contents, but because of a change in the anatomical orientation of the antrum. They measured antral cross-sectional area in the semirecumbent and semirecumbent-right lateral positions, presence / absence of fluid in the antrum in two positions, and the distance between the skin to the antrum and aorta to the antrum. The ‘Perlas score’ of antral contents (applied here to women who were not horizontal) showed good correlation over the two measurement periods. Although some degree of change in stomach contents from before to after major surgery might well occur because of gastric secretion or emptying, we might equally expect such change to be relatively minor – so this is probably an ‘expected’ result. On the other hand, and inconsistent with this first finding, there was a reduction in antral cross-sectional area – statistically significant in the case of measurements in the semirecumbent position, not significant for the semirecumbent-right lateral position. The distance between the skin to the antrum and aorta to the antrum was significantly lower in both positions after surgery, showing that the position of the antrum in late-pregnant women cannot be assumed to be the same as in the non-pregnant woman. The conclusion drawn was that the change in measured antral cross-sectional area is unlikely to indicate a real change in gastric contents 15. What are the implications for assessment of stomach contents using ultrasound in pregnant women? Antral cross-sectional area measurements assume that the area measured is the smallest possible, because the ultrasound beam crosses the antrum at right angles (sagittally). If the antrum is orientated differently, this assumption may be wrong. It is important, therefore, that research is undertaken to clarify what the correct angulation should be for the ultrasound probe during stomach ultrasound in pregnant women. The next stage is correlations of fluid in (drink) or fluid out (gastric aspiration) versus antral area in this group, to construct specific regression relationships. Besides the extrapolation of results from investigations in non-pregnant subject to pregnant women in the literature, there has also been a careless cross-over of findings from studies using different positions, given the repeated observations by various authors that stomach contents move freely according to gravity. The ‘Perlas score’ was established during investigations in subjects who were horizontal. It requires to be confirmed whether this score can be used equally for subjects, pregnant or not, in a semirecumbent position. Finally, I submit a plea for consistent and standard terminology in positioning. Recumbent and decubitus have the same definition, although usually the former word is used with ‘supine’, and the latter with ‘lateral’, to indicate the patient is horizontal. The semirecumbent position has variously been referred to as sitting up at 45° 12, semisitting 13, 14, semiupright with the head of the bed elevated to 45° 5, 7, 13, 14, supine position with a head-of-bed elevation of 45°” 6, supine in a 45° semi-recumbent position 8 and possibly ‘upright’ 1. It should be that simple…. No external funding or competing interests declared. SMK is an editor of Anaesthesia.