LAUSR

We read with interest the article by Abraldes et al. reporting on several prediction methods to assess noninvasively the presence of esophageal varices needing treatment (VNT) that were retrospectively tested in various cohorts of patients with compensated advanced chronic liver disease in Europe and Canada. The authors observed that in patients with VNT, the platelet/spleen ratio (PSR) had a diagnostic accuracy (0.74; 0.67-0.79) that was not different from the accuracy of the ratio between transient elastography value and PSR (liver stiffness to spleen/platelet score [LSPS]) (Area under the curve [AUC] 0.79; 0.720.84, 95% confidence interval [95%CI]) and was better than liver stiffness measurement alone (AUC5 0.67; 0.59-0.72, 95% CI), suggesting that both PSR and LSPS could be used to identify a subset of patients with compensated advanced chronic liver disease where endoscopy could be safely avoided. We feel that the results reported in this study are relevant for the management of patients and decisions regarding the most appropriate allocation of limited resources, such as endoscopy. However, we also feel that when assessing the accuracy of a diagnostic test, all patients who could benefit from application of the test should be included in the analysis. Transient elastography, which is part of the LSPS as well as of all the other noninvasive tests considered in this study (except the PSR), is a diagnostic test with a nonnegligible rate of unreliable measurements or failures. In particular, a French study including 13,369 examinations revealed a 3.1% failure rate and 15.8% unreliability rate for liver stiffness measurements, figures confirmed by a Chinese study that recorded failure and unreliability rates of 2.7% and 11.6%, respectively. Therefore, we suggest that the results reported by Abraldes et al. should be corrected in an “intention-totest” analysis in which unreliable measurements and failure of transient elastography are included in the analyses and are considered failures of the diagnostic test, given that these patients are recommended for endoscopy from a practical perspective. On the contrary, spleen bipolar diameter can always be determined and is highly reproducible in patients with compensated cirrhosis.