Transient splenic elastography predicts high-risk esophageal varices in patients with non-cirrhotic portal hypertension

BACKGROUND AND AIMS

Non-cirrhotic portal hypertension (NCPH) comprise a group of diseases that cause portal hypertension without cirrhosis, leading to a high risk of hemorrhage from esophageal varices. There are no non-invasive predictors of high-risk varices (HRV) described in the literature for NCPH. This study aimed to evaluate whether transient splenic elastography (TSE) can predict HRV in patients with NCPH.

METHODS

Prospective study of patients with NCPH who underwent a single timepoint evaluation with transient liver and spleen elastography, ultrasonography, upper endoscopy, and laboratory tests. The study was performed from January to September 2020. Patients were divided into two groups based on the presence of HRV. The relation between TSE, transient liver elastography (TLE), spleen size, and platelet count to the presence of HRV was evaluated.

RESULTS

Of 42 patients with NCPH, 50% (21/42) presented HRV. In univariate analysis, TSE (median, 58.4 vs. 28.3, p = 0.009) and spleen size (median, 17.5 vs. 14.5 cm, p = 0.013) were associated with HRV. No statistically significant relationship was found between the presence of HRV and platelet count or TLE. In multivariate analysis, TSE was the only variable related to HRV (OR 1.21, 95% CI 1.02-1.38). TSE had a good performance in predicting HRV in our population (AUROC 0.878; 95% CI 0.751-1000). TSE > 35.4 kPa presents 93.3% sensitivity, 60.0% specificity, and 90.9% negative predictive value.

CONCLUSION

In our population of patients with NCPH, TSE is useful in predicting HRV. TLE, spleen size, and platelet count were not related to HRV.

Feasibility and reproducibility of spleen transient elastography and its role in combination with liver transient elastography for predicting the severity of chronic viral hepatitis

Liver transient elastography (L-TE) is a reliable, noninvasive predictor of disease severity in chronic liver disease of viral aetiology (CLD). Owing to the relationships among severity of CLD, portal hypertension and spleen involvement, the assessment of splenic stiffness (S-TE) may have an added value in staging CLD. Of 132 CLD patients of viral aetiology, 48 with myeloproliferative disorders (MD) and 64 healthy volunteers (HV), were concurrently investigated by both L-TE and S-TE. Liver disease severity was staged by liver biopsy (LB; Metavir) taken concurrently with TE examination and upper gastrointestinal tract endoscopy for gastro-oesophageal varices. The S-TE inter-observer agreement was analysed by an intra-class correlation coefficient (ICC); L-TE and S-TE accuracy was evaluated by receiver operating characteristic (ROC) curve analysis. Logistic regression analysis assessed the independent effect of L-TE and S-TE as predictors of hepatic fibrosis stage. S-TE failed in 22 CLD (16.6%), 12 (25%) MD and 12 (18%) HV. In the three groups, the ICC was 0.89 (0.84-0.92), 0.90 (0.85-0.94) and 0.86(0.80-0.91), respectively. In the CLD group, L-TE and S-TE independently predicted significant fibrosis (OR 5.2 and 4.6) and cirrhosis (OR 7.8 and 9.1), but at variance from L-TE, S-TE was independent from liver necroinflammation and steatosis. The NPV of S-TE for gastro-oesophageal varices was 100% using a 48 kPa cut-off. In CLD, spleen stiffness alone or in combination with hepatic stiffness can be reliably and reproducibly assessed by TE with the added value of improving the noninvasive diagnosis of severe liver disease and excluding the presence of oesophageal varices.