Comparison of the diagnosis of oesophageal varices by oesophagoscopy and transhepatic portography

We compared the diagnosis of oesophageal varices obtained by oesophagoscopy and percutaneous transhepatic portography in 78 patients with hepatic cirrhosis who were suspected of having oesophageal varices because of ascites and/or preceding upper gastrointestinal haemorrhage. Portograms were evaluated independently of the clinical and endoscopical findings. Endoscopy showed varices in 59 patients and portography in 54 patients. The two methods agreed on the presence of varices only in 46 patients (59%), on the absence of varices in 11 patients (14%), and disagreed in 21 patients (27%). Among the patients showing varices by endoscopy, portography agreed on the diagnosis in 78%, and among patients without endoscopical varices, portography agreed in 58%. The corresponding rates of agreement expected by chance alone are 69% and 31%, respectively. The free portal pressure was lowest in patients with no varices according to both techniques, highest in patients with varices by both techniques and intermediate in patients in whom there was disagreement on the diagnosis. Modification of the criteria for endoscopical and portographical diagnosis of varices and distinction between patients with and without ascites and preceding haemorrhage had no appreciable effect on the agreement. We conclude that endoscopy and portography show only moderate agreement in the diagnosis of oesophageal varices.