Background/aims: It is recommended that patients with cirrhosis undergo endoscopic screening to rule out the presence of gastroesophageal varices: a noninvasive predictive method to identify cirrhotic patients with a very low risk of esophageal varices could potentially avoid unnecessary endoscopies.
Methods: We studied in 85 HIV-infected patients with cirrhosis the association between the absence of esophageal varices and portal hypertensive gastropathy, assessed by endoscopy, and liver stiffness measurement by transient elastography. We analyzed other parameters related to portal hypertension and hepatic function.
Results: Values of transient elastography and platelet count were significantly associated with absence of esophageal varices/portal hypertensive gastropathy. The area under the receiver operating characteristics curve [95% confidence interval (CI)] of transient elastography for the prediction of the absence of esophageal varices/portal hypertensive gastropathy was 0.7 (0.58-0.81). A liver stiffness measurement value less than 20 kPa was highly predictive of the absence of esophageal varices and portal hypertensive gastropathy. The combination of transient elastography (<20 kPa) and platelet count (>120 × 10 cells/l) had the highest negative predictive value (100%, CI 95% 77.2-100) for absence of esophageal varices and portal hypertensive gastropathy.
Conclusion: Transient elastography combined with platelet count is useful for predicting the absence of esophageal varices and portal hypertensive gastropathy and, therefore, avoid unnecessary diagnostic endoscopies in HIV-infected patients with liver cirrhosis.