IntroductionPatients with Coronavirus Disease 2019 (COVID-19) frequently experience a hyperinflammatory syndrome, that leads to unfavorable outcomes. This condition resembles Secondary Hemophagocytic Lymphohistiocytosis (sHLH) described in neoplastic, rheumatic and other infectious diseases. However, it has not been prospectively studied on these patients. A scoring system (HScore) has been validated for sHLH, and recently proposed to evaluate hyperinflammation in COVID-19.Methods143 patients aged ≥18 years admitted because of COVID-19 were enrolled in a prospective, single-center, cohort study. HScore was calculated within the 72 hours since admission. The incidence of sHLH during hospitalization was evaluated. Additionally, the relationship between HScore ≥130 points and either the requirement of mechanical ventilation or 60-days mortality was explored.ResultsThe median age of enrolled patients was 57 (21-100), and 63.6% were male. The median HScore was 96 (33-169). One patient was diagnosed with sHLH (incidence 0,7%), due to a HScore of 169. After adjusting for age, sex, comorbidities and obesity, HScore ≥130 was independently associated with the composite clinical outcome (HR 2.13, p=0.022).ConclusionsHLH is not frequent among COVID-19 patients. HScore can efficiently predict the risk for poor outcomes.

In this paper, we propose a method for the automatic maintenance of a directory of Web documents, based on the analysis of the query logs of a search engine. Our method allows automatic maintenance and continuous improvement of Web directories without the need of human editors. The method finds relevant documents to each topic of the Web directory and estimates the relevance of the documents to the topic from user clicks registered in the query log. We also enrich the Web directory by recommending queries to the topics of the directory. We show experiments to illustrate the effectiveness of our approach.

No abstract