Analysis of prognostic variables in the prediction of mortality, shunt failure, variceal rebleeding and encephalopathy following the transjugular intrahepatic portosystemic stent-shunt for variceal haemorrhage.

BACKGROUND/AIMS: The aim of this study was to analyse prognostic variables predicting mortality, shunt insufficiency, variceal rebleeding and encephalopathy following transjugular intrahepatic portosystemic stent-shunt for variceal haemorrhage.

METHODS

Sixty-eight patients with cirrhosis who successfully underwent transjugular intrahepatic portosystemic stent-shunt for variceal haemorrhage were studied. Mean age was 54.8 years (s.e. 1.5); 41 were males and 27 were females. Mean Child score was 8.3 (s.e. 0.3). Cirrhosis was alcohol related in 47. The patients have been followed up for a mean of 10.8 (s.e. 1.1) months. The univariate association between 30 prognostic variables and post-treatment encephalopathy was tested by the Chi-squared or Wilcoxon rank-sum tests and multiple logistic regression was used to test the significance of factors adjusted for one another. Cox's proportional hazard regression was used to test the univariate and multivariate significance of the other three outcomes.

RESULTS

Transjugular intrahepatic portosystemic stent-shunt was performed successfully in 89.5% of patients. Twenty-two patients have died and eight have been transplanted. Fourteen patients died within 30 days of transjugular intrahepatic portosystemic stent-shunt and death was predicted independently by hyponatraemia (p < 0.001) and by severe liver disease (Child C, p < 0.001). Eight patients died during follow up and survival in the long term was predicted independently by the presence of encephalopathy (p < 0.001) prior to transjugular intrahepatic portosystemic stent-shunt (p < 0.001). Shunt insufficiency was predicted by an initial portal pressure gradient of greater than 18 mmHg (p < 0.01). None of the variables analysed predicted variceal rebleeding. Encephalopathy following transjugular intrahepatic portosystemic stent-shunt was only predicted indepedently by its presence prior to transjugular intrahepatic portosystemic stent-shunt (p < 0.001).

CONCLUSIONS

The results of this study suggest that patients with severe liver disease and hyponatraemia are liable to die early, and the presence of encephalopathy prior to transjugular intrahepatic portosystemic stent-shunt independently determines long-term survival. Patients in these groups should be considered high risk and worked up for orthotopic liver transplantation early. Shunt function in patients with an initial portal pressure gradient of > 18 mmHg requires close supervision. Encephalopathic patients should have smaller shunts and prophylactic measures to prevent worsening encephalopathy.