Intravenous BCAA Infusion Does Not Lead to a Sustained Recovery From Overt HE in ACLF - An Open Label Randomized Clinical Trial.

Background: Hepatic encephalopathy (HE) in acute-on-chronic liver failure (ACLF) is associated with significant morbidity and mortality. We conducted a prospective, randomized controlled clinical trial to study the efficacy of intravenous branched chain amino acids (IV-BCAA) with lactulose versus lactulose alone for improvement in HE at 24 h, day 3, and day 7. The primary outcome was an improvement in encephalopathy by ≥ 1 grade at 72 h. Patients and methods: European association for study of liver (EASL) defined ACLF patients with overt HE were assessed and randomized into the experimental arm (IV-BCAA - 500 mL/day for 3 days + Lactulose; n = 39) and the comparator arm (Lactulose alone; n = 37). Six patients developed COVID-19 after randomization and were excluded (4-experimental arm and 2-comparator arm). Results: Of 222 screened patients, 70 (35 in each arm) were included in the analysis. Baseline characteristics, including HE grade (2.9 ± 0.7 vs 2.8 ± 0.7; P = 0.86) and (chronic liver failure) CLIF-C ACLF score (54.2 ± 5.6 vs 54.8 ± 5.7; P = 0.65), were similar. Overall survival was 40% at 28 days (48.5% vs 31.4%; P = 0.14). Improvement in hepatic encephalopathy scoring algorithm (HESA) by ≥ 1 grade at 24 h occurred in 14 patients (40%) in the BCAA arm and 6 patients (17.1%) in the control group (P = 0.03) which translated to a shorter intensive care unit (ICU) stay. The median change in HESA at 24 h was greater in the BCAA arm than the control arm (P = 0.006), which was not sustained at days 3 or 7. Ammonia levels did not correlate with the grade of HE (Spearman's correlation coefficient (ρ) = - 0.0843; P = 0.29). Conclusion: Intravenous BCAA does not lead to a sustained improvement in HE grade in ACLF. Trial registration no: NCT04238416 (clinicaltrials.gov).

Transfer Learning Based Lightweight Ensemble Model for Imbalanced Breast Cancer Classification.

Automated classification of breast cancer can often save lives, as manual detection is usually time-consuming & expensive. Since the last decade, deep learning techniques have been most widely used for the automatic classification of breast cancer using histopathology images. This paper has performed the binary and multi-class classification of breast cancer using a transfer learning-based ensemble model. To analyze the correctness and reliability of the proposed model, we have used an imbalance IDC dataset, an imbalance BreakHis dataset in the binary class scenario, and a balanced BACH dataset for the multi-class classification. A lightweight shallow CNN model with batch normalization technology to accelerate convergence is aggregated with lightweight MobileNetV2 to improve learning and adaptability. The aggregation output is fed into a multilayer perceptron to complete the final classification task. The experimental study on all three datasets was performed and compared with the recent works. We have fine-tuned three different pre-trained models (ResNet50, InceptionV4, and MobilNetV2) and compared it with the proposed lightweight ensemble model in terms of execution time, number of parameters, model size, etc. In both the evaluation phases, it is seen that our model outperforms in all three datasets.

Neurological monitoring and sedation protocols in the Liver Intensive Care Unit.

Patients with liver disease often have alteration of neurological status which requires admission to an intensive care unit. Patients with acute liver failure (ALF), acute-on-chronic liver failure (ACLF) and rarely cirrhosis are at risk of cerebral edema. These patients require prompt assessment of neurological status including assessment of intra-cranial pressure (ICP) and monitoring metabolic parameters like arterial/venous ammonia levels, serum creatinine and serum electrolytes so that timely specific therapy for raised ICP can be instituted to prevent permanent neurological dysfunction. The overall aims of neuromonitoring and sedation protocols in a liver intensive care unit are to identify the level of multifactorial metabolic encephalopathy, individualize sedation and analgesia requirements for patients on mechanical ventilation, institute specific therapy to correct the neurological insult in ALF and ACLF, provide clear physiological data for guided therapy of drugs like muscle relaxants, antiepileptics, and cerebral edema reducing agents, and assist with overall prognostication. In this review article we will outline the clinical scenarios related to liver disease requiring intensive care and neuromonitoring, current techniques of neurological assessment, sedation protocols and point of care tests which enable the treating physician and intensivist guide therapy for raised ICP.

Hypercoagulability on thromboelastography after living donor hepatectomy-The true side of the coin.

BACKGROUND AND AIMS: Coagulation dynamics after donor hepatectomy are complex. Having complete knowledge of the actual changes in the coagulation status during donor hepatectomy is important to prevent complications such as pulmonary embolism, deep vein thrombosis, and bleeding. Hence, the present study aimed to study the coagulation dynamics following open donor hepatectomy both by thromboelastography (TEG) and conventional coagulation tests (CCT). METHODS: A total of 50 prospective liver donors were included. TEG and CCT [activated partial thromboplastin time (aPTT), prothrombin time (PT), international normalised ratio (INR), fibrinogen, and platelet counts] were performed for each patient before surgery (baseline), on postoperative day (POD) 0, 1, 2, 3, 5, and 10. RESULTS: TEG showed hypercoagulability in 28%, 38%, 30%, 46%, 42%, and 48% patients; in contrast INR showed hypocoagulability in 58%, 63%, 73%, 74%, 20%, and 0% patients on POD 0,1,2,3,5, and 10, respectively. Patients demonstrating hypercoagulability on TEG had significantly decreased reaction time (P = 0.004), significantly increased maximum amplitude (P < 0.001), and alpha angle value (P < 0.001). Postoperatively, INR, PT, and aPTT values increased significantly, while platelets and fibrinogen levels decreased significantly when compared to their baseline values. There was no coagulation-related postoperative complication in any of the patients. CONCLUSION: Hypercoagulability after donor hepatectomy is common. TEG showed hypercoagulability and did not show any hypocoagulability as suggested by the CCT. In patients undergoing donor hepatectomy, CCT may not reflect the actual changes incoagulation status and tests such as TEG should be performed to know the correct nature of changes in coagulation following donor hepatectomy.