Aditi Kumar1, Sadhna B Lal1, Anmol Bhatia2, Ashim Das3. 1. Divisions of Paediatric Gastroenterology, Hepatology and Nutrition. 2. Paediatric Radiology. 3. Department of Pathology, Post Graduate Institute of Medical Education and Research, Chandigarh, India.
Abstract
BACKGROUND: Noninvasive tools (NITs) for predicting varices in children with portal hypertension (PHTN) are infrequently used. METHODS: Eighty-five consecutive, treatment-naïve children with PHTN and 97 controls were enrolled study from July 2017 to November 2018. Each case was evaluated by esophagogastroduodenoscopy (EGD) and various NITs: platelet spleen size Z (PSZ), clinical prediction rule (CPR), King's variceal prediction rule (KVaPS), Splenic stiffness (SSM) and liver stiffness measurement (LSM) by point shear wave elastrography (pSWE). RESULTS: Had PHTN due to extra hepatic portal vein obstruction (EHPVO) and 70% due to cirrhosis [chronic liver disease (CLD)]. Sixty-five percent of PHTN cases had varices. Children with varices had lower platelet counts, lower albumin and larger spleens. SSM and LSM were significantly higher in cases as compared with controls. SSM was significantly higher in cases with varices than those without. SSM and LSM, at cutoffs of 3.8 and 3.2 kPa, respectively, discriminated PHTN cases from controls with an area under the curve (AUROC) of 0.67 (0.59-0.74). Both SSM and LSM predicted varices in CLD, but in EHPVO, only SSM predicted varices. SSM of 5.2 and 12.8 kPa, in CLD and EHPVO subgroups, respectively, had AUROC of 0.73 and 0.94 for variceal prediction. Blood-based NITs performed better in the CLD subgroup: aspartate aminotransferase platelet ratio index, CPR and KVPS predicted severity of PHTN with AUROC of 0.81, 0.92 and 0.93, respectively. CONCLUSIONS: Blood-based NITs outperform elastography for prediction of PHTN/varices in children with CLD. SSM by pSWE is a better predictor of varices than LSM, especially in the EHPVO subgroup.
BACKGROUND: Noninvasive tools (NITs) for predicting varices in children with portal hypertension (PHTN) are infrequently used. METHODS: Eighty-five consecutive, treatment-naïve children with PHTN and 97 controls were enrolled study from July 2017 to November 2018. Each case was evaluated by esophagogastroduodenoscopy (EGD) and various NITs: platelet spleen size Z (PSZ), clinical prediction rule (CPR), King's variceal prediction rule (KVaPS), Splenic stiffness (SSM) and liver stiffness measurement (LSM) by point shear wave elastrography (pSWE). RESULTS: Had PHTN due to extra hepatic portal vein obstruction (EHPVO) and 70% due to cirrhosis [chronic liver disease (CLD)]. Sixty-five percent of PHTN cases had varices. Children with varices had lower platelet counts, lower albumin and larger spleens. SSM and LSM were significantly higher in cases as compared with controls. SSM was significantly higher in cases with varices than those without. SSM and LSM, at cutoffs of 3.8 and 3.2 kPa, respectively, discriminated PHTN cases from controls with an area under the curve (AUROC) of 0.67 (0.59-0.74). Both SSM and LSM predicted varices in CLD, but in EHPVO, only SSM predicted varices. SSM of 5.2 and 12.8 kPa, in CLD and EHPVO subgroups, respectively, had AUROC of 0.73 and 0.94 for variceal prediction. Blood-based NITs performed better in the CLD subgroup: aspartate aminotransferase platelet ratio index, CPR and KVPS predicted severity of PHTN with AUROC of 0.81, 0.92 and 0.93, respectively. CONCLUSIONS: Blood-based NITs outperform elastography for prediction of PHTN/varices in children with CLD. SSM by pSWE is a better predictor of varices than LSM, especially in the EHPVO subgroup.