Maximally informative feature selection using Information Imbalance: Application to COVID-19 severity prediction.

Clinical databases typically include, for each patient, many heterogeneous features, for example blood exams, the clinical history before the onset of the disease, the evolution of the symptoms, the results of imaging exams, and many others. We here propose to exploit a recently developed statistical approach, the Information Imbalance, to compare different subsets of patient features and automatically select the set of features that is maximally informative for a given clinical purpose, especially in minority classes. We adapt the Information Imbalance approach to work in a clinical framework, where patient features are often categorical and are generally available only for a fraction of the patients. We apply this algorithm to a data set of ∼ 1300 patients treated for COVID-19 in Udine hospital before October 2021. Using this approach, we find combinations of features which, if used in combination, are maximally informative of the clinical fate and of the severity of the disease. The optimal number of features, which is determined automatically, turns out to be between 10 and 15. These features can be measured at admission. The approach can be used also if the features are available only for a fraction of the patients, does not require imputation and, importantly, is able to automatically select features with small inter-feature correlation. Clinical insights deriving from this study are also discussed.

Potential role of IgM-enriched immunoglobulin as adjuvant treatment in severe SARS-CoV-2 infection.

BACKGROUND: Severe COVID-19 patients are characterized by a dysregulated host response to an infection, with uncontrolled pro- and anti- inflammatory pathway activation. Consistent proportion of patients require admission in intensive care units and are at risk of progression to severe forms of disease. These patients are generally admitted during later stages of the disease, when effective antiviral and monoclonal antibody are not indicated. We aimed to assess the potential role of IgM-enriched intra venous immunoglobulins (IGAM) preparations in this setting. METHODS: This retrospective, observational case-controlled study was conducted at a single-center University Hospital of Udine in the Friuli Venezia Giulia Region of Italy. Patients referring to the center between March 2020 and April 2021 was included. During the study period, patient who received Pentaglobin® IGAM treatment (N.=56), administered as compassionate use, was compared with a control group (N.=169) to assess, by propensity score analysis, clinical outcome. RESULTS: Untreated controls required, respect to patient treated with IGAM therapy, longer time to hospitalization with no significant differences in death and orotracheal intubation requirement. Significant differences in the two cohort were in: SOFA was higher in treated, while D-dimer and P/F ratio was better in the treatment cohort. Multivariate logistic regression analysis performed on the "matched sample," obtained by a weighting propensity score approach, identify, as significant protective factor for death outcome, the Pentaglobin® treatment (0.820 [0.698-0.963], P=0.016) and low C-reactive protein (1.001 [1.000-1.002], P=0.031) value while the delay of onset hospitalization is associate with a worst outcome (0.983 [0.967-0.999], P=0.041). CONCLUSIONS: The present study offers a significant insight concerning the use of IgM-enriched immunoglobulin preparations in patients with SARS-CoV-2 severe infection and also could identifying the specific immunological and biochemical profile of the patient who can more benefit from this therapeutic option.

Post-COVID-19 syndrome and humoral response association after 1 year in vaccinated and unvaccinated patients.

OBJECTIVES: This study aimed to describe the impact of vaccination and the role of humoral responses on post-COVID-19 syndrome 1 year after the onset of SARS coronavirus type 2 (CoV-2). METHODS: This prospective study was conducted through interviews to investigate post-COVID-19 syndrome 6 and 12 months after disease onset in all adult in- and outpatients with COVID-19 at Udine Hospital (March-May 2020). Vaccination status and two different serological assays to distinguish between response to vaccination (receptor-binding domain (RBD) SARS-CoV-2 IgG) and/or natural infection (non-RBD-SARS-CoV-2 IgG) were also assessed. RESULTS: A total of 479 patients (52.6% female; mean age: 53 years) were interviewed 13.5 months (standard deviation: 0.6 months) after acute infection. Post-COVID-19 syndrome was observed in 47.2% of patients (n = 226) after 1 year. There were no significant differences in the worsening of post-COVID-19 symptoms (22.7% vs. 15.8%; p = 0.209) among vaccinated (n = 132) and unvaccinated (n = 347) patients. The presence of non-RBD SARS-CoV-2 IgG induced by natural infection showed a significant association with post-COVID-19 syndrome (OR: 1.35; 95% CI, 1.11-1.64; p = 0.003), and median non-RBD SARS-CoV-2 IgG titres were significantly higher in long haulers than in patients without symptoms (22 kAU/L (interquartile range, 9.7-37.2 kAU/L) vs. 14.1 kAU/L (interquartile range, 5.4-31.3 kAU/L); p = 0.009) after 1 year. In contrast, the presence of RBD SARS-CoV-2 IgG was not associated with the occurrence of post-COVID-19 syndrome (>2500 U/mL vs. 0.9-2500 U/mL; OR: 1.36; 95% CI, 0.62-3.00; p = 0.441), and RBD SARS-CoV-2 IgG titres were similar in long haulers as in patients without symptoms (50% values > 2500 U/mL vs. 55.6% values > 2500 U/mL; p = 0.451). DISCUSSION: The SARS-CoV-2 vaccination is not associated with the emergence of post-COVID-19 symptoms more than 1 year after acute infection. The persistence of high serological titre response induced by natural infection, but not vaccination, may play a role in long-haul COVID-19.