Longer duration of transient elastography predicts unreliable liver stiffness measurements

Abstract

Transient elastography (TE) is a noninvasive, validated method to assess liver fibrosis by obtaining liver stiffness measurements (LSM). However, TE can be limited by unreliable measurement (UM). The relationship between the time taken to perform TE (duration) and UM has not been studied. To determine whether the duration of TE correlates with UM. We prospectively studied the frequency and predictors of UM over a 5-year period. UM was defined as follows: less than 10 successful measurements, success rate less than 60%, or interquartile range more than 30% of the median LSM value (IQR/LSM>30%). Among the 2834 patients with LSM analysed, UM occurred in 19.0%. Duration [odds ratio (OR) 4.2, 95% confidence interval (CI) 2.8-6.4; P<0.0001] was the strongest predictor of UM, followed by BMI more than 28 kg/m (OR 2.1, 95% CI 1.5-3.0; P<0.0001), age more than 52 (OR 1.6, 95% CI 1.1-2.3; P=0.007) and non-HBV aetiology (OR 1.6, 95% CI 1.1-2.3; P=0.02). An optimal cut-off of 3 min 47 s was calculated for predicting UM (sensitivity 70%, specificity 65%, OR 4.2, 95% CI 2.7-6.6, P<0.0001). Examinations that took longer than 8 min 10 s had a 90% chance of UM. In experienced hands, duration is a strong predictor of UM in patients undergoing TE. Examinations longer than 4 min are more likely to be unreliable. Examinations longer than 8 min are unlikely to yield a valid result and should be considered a futility endpoint. Older age and increased BMI and nonhepatitis B aetiology are independent, albeit weaker, predictors of UM.