A Retrospective Analysis of Fluid Balance in Patients With AKI and Respiratory Failure due to COVID-19

Background: The optimal amount of hydration for patients with severe COVID-19 infection and AKI is unknown. This study aims to investigate the impact of fluid management strategy and outcomes in patients with AKI and respiratory failure due to COVID-19. Method(s): Data was gathered from a retrospective chart review of patients with hypoxia due to COVID-19 infection and stage 2 or greater AKI. Primary outcome was the difference in net fluid balance between patients who were successfully weaned to lower levels of oxygen support and discharged compared to those who died or remained ventilator dependent. Result(s): Of 58 cases, 41 died, 3 remained ventilator-dependent, and 14 were discharged without supplemental oxygen. The groups differed in net fluid balance (-10,065 cc vs +7,980 cc, p <0.001) and daily fluid balance (-367 vs. 515 cc/day, p <0.001) with a substantially lower mean fluid balance in patients who survived with minimal requirement for supplemental oxygen. Patients who maintained a positive fluid balance were significantly more likely to become ventilator dependent or die (OR: 40.7, 95% CI: 5.3 - 312.9). A fluid restrictive strategy did not reduce the likelihood of recovery from AKI or increase the need for renal replacement therapy. Conclusion(s): In our cohort, patients with COVID-19 and AKI who survived with minimal or no oxygen requirements tended to have negative fluid balance in contrast to those who died or remained ventilator-dependent. A fluid restrictive strategy with judicious volume removal using diuretics or dialysis may lead to improved outcomes in COVID-19 patients with AKI.

Analyzing the Sentiments by Classifying the Tweets Based on COVID-19 Using Machine Learning Classifiers

In the current scenario, almost all the countries face one of the biggest disasters in COVID-19. This paper has to analyze the tweets related to COVID 19 and discuss the various machine learning algorithms and their performance analysis on the tweets associated with COVID-19. The implemented classification algorithms are applied to classify the sentiments to predict whether they relate to COVID-19 or non-COVID-19. Ten most popular classification algorithms implemented. The Linear Support Vector Machine (LSVM) achieved the highest test accuracy in these algorithms with 90.3%. Logistic regression has performed better in recall with 96.06%, F1 score of 90.46%, ROC_AUC with 90.48%. Random forest classifier has achieved the better specificity and precision of 99.16% and 96.3%, respectively. Out of all, stochastic gradient descent (SGD) has attained better results in all the computational parameters. © 2021 IEEE.

Generation of free fatty acid during lockdown and its effect on the corrosion in rolling emulsion tank

Corrosion in roll coolant tank is a rare phenomenon in the steel industry. The ability of the cold rolling emulsion to prevent corrosion on the cold rolled steel, the mild steel coolant tank, the mill housing and pipelines stems from the presence of oil soluble corrosion inhibitors, anti-oxidants or anti-wear additives present in rolling oil formulation. During normal operational conditions, the mill coolant (mixture of oil and water) along with the generated iron fines is continuously recirculated through the rolling mill application system. The concentration of the coolant is maintained by periodic addition of oil and demineralized water in the system. During Covid-19 unprecedented lockdown, voluminous corrosion was observed in the inner walls of coolant tanks, pipelines and other accessories, which remained in contact with the emulsion. Though normal rolling operations were halted, the bath temperature and recirculation of the mill coolant were maintained to sustain the emulsion health against bacterial attack. It has been earlier observed that during indefinite mill shutdowns, if emulsion is left without heating, circulation and stirring the health of emulsion deteriorates leaving a characteristic rancid odor which resembles bacterial attack. This study primarily highlights the role of acidic by-products formed due to depletion of anti-oxidants leading to rapid deterioration of emulsion triggering corrosion within the roll coolant tank. Techniques, like FTIR, SEM-EDS and Raman Spectroscopy were employed to understand the morphology and nature of the corrosion. Corrosion rust formed after simulation under accelerated laboratory conditions was finally subjected to Raman Spectroscopy, and Protective Ability Index (PAI) was applied to verify the variation in rust protection ability of the fresh and lockdown period rolling oil. It clearly shows the connection between the loss of protective nature of the rust and increase in acidity due to lockdown leading to accelerated corrosion of the tank system. © 2021 Elsevier Ltd