Background: Could nutritional status serve as prognostic factors for coronavirus disease 2019 (COVID-19)? The present study evaluated the clinical and nutritional characteristics of COVID-19 patients and explored the relationship between risk for malnutrition at admission and in-hospital mortality. Methods: A retrospective, observational study was conducted in two hospitals in Hubei, China. Confirmed cases of COVID-19 were typed as mild/moderate, severe, or critically ill. Clinical data and in-hospital death were collected. The risk for malnutrition was assessed using the geriatric nutritional risk index (GNRI), the prognostic nutritional index (PNI), and the Controlling Nutritional Status (CONUT) via objective parameters at admission. Results: Two hundred ninety-five patients were enrolled, including 66 severe patients and 41 critically ill patients. Twenty-five deaths were observed, making 8.47% in the whole population and 37.88% in the critically ill subgroup. Patients had significant differences in nutrition-related parameters and inflammatory biomarkers among three types of disease severity. Patients with lower GNRI and PNI, as well as higher CONUT scores, had a higher risk of in-hospital mortality. The receiver operating characteristic curves demonstrated the good prognostic implication of GNRI and CONUT score. The multivariate logistic regression showed that baseline nutritional status, assessed by GNRI, PNI, or CONUT score, was a prognostic indicator for in-hospital mortality. Conclusions: Despite variant screening tools, poor nutritional status was associated with in-hospital death in patients infected with COVID-19. This study highlighted the importance of nutritional screening at admission and the new insight of nutritional monitoring or therapy.

Contrast-induced acute kidney injury (CI-AKI) contributes toward unfavorable clinical outcomes after primary percutaneous coronary intervention (pPCI). We will assess whether hyperuricemia is an independent predictor of CI-AKI and outcomes in patients undergoing pPCI. Our study is a secondary analysis for the database from ATTEMPT study, enrolling 560 ST-segment elevation myocardial infarction (STEMI) patients undergoing pPCI. Patients will be divided into 2 groups according to the admission serum uric acid (SUA) level. Hyperuricemia will be defined as a SUA level > 7 mg/dL (417 mmol/L) in males and > 6 mg/dL (357 mmol/L) in females. The primary endpoint was CI-AKI, defined as > 25% or 0.5 mg/dL increase in serum creatinine from baseline during the first 48–72 h post-procedurally. Multivariate analyses for CI-AKI and long-term mortality will be performed using the logistic regression and Cox regression analyses, respectively. This study will determine the predictive value of hyperuricemia for the development of CI-AKI and outcomes in patients with STEMI undergoing pPCI. We predict that hyperuricemia will be associated with a risk of CI-AKI in patients with pPCI. Furthermore, after adjusting for other variables, long-term mortality after pPCI may be higher in those with hyperuricemia than in those with normouricemia. Results of this study may provide scientific evidence for the effect of hyperuricemia on CI-AKI and long-term outcomes, thereby offering the potential possibility of lowering SUA on the development of CI-AKI and outcomes. ClinicalTrials.gov NCT02067195, Registered on 20 February 2014.

Background: Point-of-care ultrasonography has become a valuable tool for assessing diaphragmatic function in critically ill patients receiving invasive mechanical ventilation. However, conventional diaphragm ultrasound assessment remains highly operator-dependent and subjective. Previous research introduced automatic measurement of diaphragmatic excursion and velocity using 2D speckle-tracking technology. Objective: This study aimed to develop an artificial intelligence–multimodal learning framework to improve the prediction of weaning failure and guide individualized weaning strategies. Methods: This prospective study enrolled critically ill patients older than 18 years who received mechanical ventilation for more than 48 hours and were eligible for a spontaneous breathing trial in 2 intensive care units in Guangzhou, China. Before the spontaneous breathing trial, diaphragm ultrasound videos were collected using a standardized protocol, and automatic measurements of excursion and velocity were obtained. A total of 88 patients were included, with 50 successfully weaned and 38 experiencing weaning failure. Each patient record included 27 clinical and 6 diaphragmatic indicators, selected based on previous literature and phenotyping studies. Clinical variables were preprocessed using OneHotEncoder, normalization, and scaling. Ultrasound videos were interpolated to a uniform resolution of 224×224×96. Artificial intelligence–multimodal learning based on clinical characteristics, laboratory parameters, and diaphragm ultrasonic videos was established. Four experiments were conducted in an ablation setting to evaluate model performance using different combinations of input data: (1) diaphragmatic excursion only, (2) clinical and diaphragmatic indicators, (3) ultrasound videos only, and (4) all modalities combined (multimodal). Metrics for evaluation included classification accuracy, area under the receiver operating characteristic curve (AUC), average precision in the precision-recall curve, and calibration curve. Variable importance was assessed using SHAP (Shapley Additive Explanation) to interpret feature contributions and understand model predictions. Results: The multimodal co-learning model outperformed all single-modal approaches. The accuracy improved when predicted through diaphragm ultrasound video data using Video Vision Transformer (accuracy=0.8095, AUC=0.852), clinical or ultrasound indicators (accuracy=0.7381, AUC=0.746), and the multimodal co-learning (accuracy=0.8331, AUC=0.894). The proposed co-learning model achieved the highest score (average precision=0.91) among the 4 experiments. Furthermore, calibration curve analysis demonstrated that the proposed colearning model was well calibrated, as the curve was closest to the perfectly calibrated line. Conclusions: Combining ultrasound and clinical data for colearning improved the accuracy of the weaning outcome prediction. Multimodal learning based on automatic measurement of point-of-care ultrasonography and automated collection of objective clinical indicators greatly enhanced the practical operability and user-friendliness of the system. The proposed model offered promising potential for widespread clinical application in intensive care settings.