Ultrasound measurement of the antral cross-sectional area allows a quantitative estimate of gastric contents in non-pregnant adults, but this relationship may be affected by compression of the stomach exerted by the gravid uterus during pregnancy. This study aimed to assess differences in quantitative (Perlas score) and qualitative (antral cross-sectional area) ultrasound assessments of the gastric antrum performed immediately before and after caesarean section. Forty-three women having elective caesarean section performed under spinal anaesthesia were studied in the semirecumbent and semirecumbent-right lateral positions. Thirty-nine women showed no change in stomach contents using the Perlas score between the two measurement periods; four women showed a change, but by one grade only. The median (IQR [range]) antral cross-sectional area was 323 (243-495 [103-908]) mm before, and 237 (165-377 [112-762]) mm after, caesarean section in the semirecumbent position (p = 0.001); the comparable values in the semirecumbent-right lateral position were 418 (310-640 [161-1238]) mm and 362 (280-491 [137-1231]) mm (p = 0.09). The distance between the skin and the antrum, and the aorta and the antrum, decreased significantly in both positions after surgery. We suggest that our results indicate that stomach contents remain largely unchanged in women having elective caesarean section, but antral cross-sectional area decreases, especially in the semirecumbent position, related to a change in the position of the stomach within the abdomen. This implies that the relationship of antral cross-sectional area to volume of stomach contents, which has been determined for non-pregnant subjects, may not apply in term pregnant women.