Pre-infection 25-hydroxyvitamin D3 levels and association with severity of COVID-19 illness.

OBJECTIVE: Studies have demonstrated a potential correlation between low vitamin D status and both an increased risk of infection with SARS-CoV-2 and poorer clinical outcomes. This retrospective study examines if, and to what degree, a relationship exists between pre-infection serum 25-hydroxyvitamin D (25(OH)D) level and disease severity and mortality due to SARS-CoV-2. PARTICIPANTS: The records of individuals admitted between April 7th, 2020 and February 4th, 2021 to the Galilee Medical Center (GMC) in Nahariya, Israel, with positive polymerase chain reaction (PCR) tests for SARS-CoV-2 (COVID-19) were searched for historical 25(OH)D levels measured 14 to 730 days prior to the positive PCR test. DESIGN: Patients admitted to GMC with COVID-19 were categorized according to disease severity and level of 25(OH)D. An association between pre-infection 25(OH)D levels, divided between four categories (deficient, insufficient, adequate, and high-normal), and COVID-19 severity was ascertained utilizing a multivariable regression analysis. To isolate the possible influence of the sinusoidal pattern of seasonal 25(OH)D changes throughout the year, a cosinor model was used. RESULTS: Of 1176 patients admitted, 253 had records of a 25(OH)D level prior to COVID-19 infection. A lower vitamin D status was more common in patients with the severe or critical disease (<20 ng/mL [87.4%]) than in individuals with mild or moderate disease (<20 ng/mL [34.3%] p < 0.001). Patients with vitamin D deficiency (<20 ng/mL) were 14 times more likely to have severe or critical disease than patients with 25(OH)D ≥40 ng/mL (odds ratio [OR], 14; 95% confidence interval [CI], 4 to 51; p < 0.001). CONCLUSIONS: Among hospitalized COVID-19 patients, pre-infection deficiency of vitamin D was associated with increased disease severity and mortality.

COVID-19 classification of X-ray images using deep neural networks.

OBJECTIVES: In the midst of the coronavirus disease 2019 (COVID-19) outbreak, chest X-ray (CXR) imaging is playing an important role in diagnosis and monitoring of patients with COVID-19. We propose a deep learning model for detection of COVID-19 from CXRs, as well as a tool for retrieving similar patients according to the model's results on their CXRs. For training and evaluating our model, we collected CXRs from inpatients hospitalized in four different hospitals. METHODS: In this retrospective study, 1384 frontal CXRs, of COVID-19 confirmed patients imaged between March and August 2020, and 1024 matching CXRs of non-COVID patients imaged before the pandemic, were collected and used to build a deep learning classifier for detecting patients positive for COVID-19. The classifier consists of an ensemble of pre-trained deep neural networks (DNNS), specifically, ReNet34, ReNet50¸ ReNet152, and vgg16, and is enhanced by data augmentation and lung segmentation. We further implemented a nearest-neighbors algorithm that uses DNN-based image embeddings to retrieve the images most similar to a given image. RESULTS: Our model achieved accuracy of 90.3%, (95% CI: 86.3-93.7%) specificity of 90% (95% CI: 84.3-94%), and sensitivity of 90.5% (95% CI: 85-94%) on a test dataset comprising 15% (350/2326) of the original images. The AUC of the ROC curve is 0.96 (95% CI: 0.93-0.97). CONCLUSION: We provide deep learning models, trained and evaluated on CXRs that can assist medical efforts and reduce medical staff workload in handling COVID-19. KEY POINTS: • A machine learning model was able to detect chest X-ray (CXR) images of patients tested positive for COVID-19 with accuracy and detection rate above 90%. • A tool was created for finding existing CXR images with imaging characteristics most similar to a given CXR, according to the model's image embeddings.