Accuracy of non-endoscopic predictors of oesophageal varices in liver cirrhosis using platelet count, splenic size and portal vein diameter.

Background: Endoscopy has been recommended for all patients with liver cirrhosis to detect varices, but it is expensive, invasive and uncomfortable. There is therefore, need to find non-endoscopic means to predict oesophageal varices.

Aim: To determine the sensitivity and specificity of platelet count, splenic size and portal vein diameter to predict oesophageal varices in patients with liver cirrhosis.

Methodology: Subjects were patients with liver cirrhosis and controls without liver disease aged 18 years and above. Platelet count was determined using Mindray BC-3000plus auto-analyzer. Portal vein diameter was measured at a point where it crosses the hepatic artery. Splenic length was measured through the hilum, as the distance between the dome and the tip. All the patients had upper gastrointestinal endoscopy. Varices were graded into I,II, III based on the Japanese classification of oesophageal vatices.

Results: The patients comprised 59 (81%). males and 14 (19%) females, while controls comprised 29 (73%) males and 11 (27%) females. The mean±(S.D.) age of the patients and controls was 44±12.6 and 40± 13 years respectively. There was statistically significant difference in the means of platelet count and platelet count/splenic size between patients with large oesophageal varices and those with small or no varices (p=0.00), while no such difference in the means of splenic size, portal vein diameter and the presence/size or absence of varices. (p=0.06).

Conclusion: Platelet count has the best sensitivity and specificity among the three values in predicting both small and large varices in patients with liver cirrhosis.