Gastric pull‐up with cervical oesophagogastric anastomosis is a common procedure after oesophageal resection. Contrast studies are used by many surgeons for detection of anastomotic leaks but may be unnecessary. A prospective study was undertaken to compare gastrografin study and test feeding using water for detection of cervical anastomotic leaks. In 25 patients. gastrografin study showed three leaks, two of which were clinically silent and patients did not have any problems. One patient had aspiration of contrast and the study could not be completed. Two of the leaks detected clinically were missed by contrast study (one because of an incomplete study and the other was a satisfactory study). Delayed leaks occurred in two patients. All leaks healed spontaneously. A contrast study may thus be unnecessary for evaluation of a cervical oesophagogastric anastomosis and can be replaced with the simpler and safer technique of ‘test feeding’ using