ABSTRACT
BACKGROUND: Palpation and percussion are standard bedside techniques used to diagnose hepatomegaly. Ultrasonography is a noninvasive and accurate method for measurement of liver size, but many patients in developing countries have limited access to it. We compared the accuracy of palpation and percussion in a rural population in central India, using ultrasonography as a reference standard. METHODS: The study design was a blinded, cross-sectional analysis of a hospital-based case series. Three physicians, blind to clinical data and to each other's results, independently used palpation and percussion to detect hepatomegaly. Diagnostic accuracy was measured by computing sensitivity, specificity, and likelihood ratio values. Inter-physician agreement was assessed using the kappa statistic. RESULTS: Of the 180 study patients, 36 (20%) had enlarged liver on ultrasonography. The likelihood ratios for findings at both palpation (2.2, 3.0, and 2.5 for the three physicians, respectively) and percussion (1.1 for all three physicians) as predictors of true hepatomegaly were low. The kappa values for inter-observer agreement between three physicians for the presence of hepatomegaly at palpation (=0.44-0.53) and percussion (=0.17-0.33) were low, indicating poor reliability of these techniques. CONCLUSION: Clinical assessment of hepatomegaly by palpation and percussion lacks both accuracy and reliability.