Serologic Responses to COVID-19 Vaccination in Children with History of Multisystem Inflammatory Syndrome (MIS-C)

Understanding the serological responses to COVID-19 vaccination in children with history of MIS-C could inform vaccination recommendations. We prospectively enrolled five children hospitalized with MIS-C and measured SARS-CoV-2 binding IgG antibodies to spike protein variants longitudinally pre- and post-Pfizer-BioNTech BNT162b2 primary series COVID-19 vaccination. We found that SARS-CoV-2 variant cross-reactive IgG antibodies waned following acute MIS-C, but were significantly boosted with vaccination and maintained for at least 3 months. We then compared post-vaccination binding, pseudovirus neutralizing, and functional antibody-dependent cell-mediated cytotoxicity (ADCC) titers to the reference strain (Wuhan-hu-1) and Omicron variant (B.1.1.529) among previously healthy children (n=6) and children with history of MIS-C (n=5) or COVID-19 (n=5). Despite the breadth of binding antibodies elicited by vaccination in all three groups, pseudovirus neutralizing and ADCC titers were reduced to the Omicron variant. Vaccination after MIS-C or COVID-19 (hybrid immunity) conferred advantage in generating pseudovirus neutralizing and functional ADCC antibodies to Omicron.

Evaluation of machine learning for predicting COVID-19 outcomes from a national electronic medical records database

ObjectiveWhen novel diseases such as COVID-19 emerge, predictors of clinical outcomes might be unknown. Using data from electronic medical records (EMR) allows evaluation of potential predictors without selecting specific features a priori for a model. We evaluated different machine learning models for predicting outcomes among COVID-19 inpatients using raw EMR data. Materials and MethodsIn Premier Healthcare Data Special Release: COVID-19 Edition (PHD-SR COVID-19, release date March, 24 2021), we included patients admitted with COVID-19 during February 2020 through April 2021 and built time-ordered medical histories. Setting the prediction horizon at 24 hours into the first COVID-19 inpatient visit, we aimed to predict intensive care unit (ICU) admission, hyperinflammatory syndrome (HS), and death. We evaluated the following models: L2-penalized logistic regression, random forest, gradient boosting classifier, deep averaging network, and recurrent neural network with a long short-term memory cell. ResultsThere were 57,355 COVID-19 patients identified in PHD-SR COVID-19. ICU admission was the easiest outcome to predict (best AUC=79%), and HS was the hardest to predict (best AUC=70%). Models performed similarly within each outcome. DiscussionAlthough the models learned to attend to meaningful clinical information, they performed similarly, suggesting performance limitations are inherent to the data. ConclusionPredictive models using raw EMR data are promising because they can use many observations and encompass a large feature space; however, traditional and deep learning models may perform similarly when few features are available at the individual patient level.