Nomogram-based prediction model for survival of COVID-19 patients: A clinical study

Prediction models based on machine learning algorithms for COVID-19 severity risk

BACKGROUND

The World Health Organization has highlighted the risk of Disease X, urging pandemic preparedness. Coronavirus disease 2019 (COVID-19) could be the first Disease X; therefore, understanding the epidemiological experiences of COVID-19 is crucial while preparing for future similar diseases.

METHODS

Prediction models for COVID-19 severity risk in hospitalized patients were constructed based on four machine learning algorithms, namely, logistic regression, Cox regression, support vector machine (SVM), and random forest. These models were evaluated for prediction accuracy, area under the curve (AUC), sensitivity, and specificity as well as were interpreted using SHapley Additive exPlanation.

RESULTS

Data were collected from 1,485 hospitalized patients across 6 centers, comprising 1,184 patients with severe or critical COVID-19 and 301 patients with nonsevere COVID-19. Among the four models, the SVM model achieved the highest prediction accuracy of 98.45%, with an AUC of 0.994, a sensitivity of 0.989, and a specificity of 0.969. Moreover, oxygenation index (OI), confusion, respiratory rate, and age were found to be predictors of COVID-19 severity risk.

CONCLUSIONS

SVM could accurately predict COVID-19 severity risk; thus, it can be prioritized as a prediction model. OI is the most critical predictor of COVID-19 severity risk and can serve as the primary and independent evaluation indicator.

A nomogram to predict long COVID risk based on pre- and post-infection factors: Results from a cross-sectional study in South China

OBJECTIVES

Long COVID has received much attention as a complex multi-system disease due to its serious impact on quality of life. However, there remains inconsistent results in terms of risk factors, and a prediction model for the accurate prediction of long COVID is still lacking.

STUDY DESIGN

Cross-sectional study.

METHODS

In this retrospective study, a community population from the Futian District of Shenzhen, Guangdong Province, China, were included. Data were collected from September to December 2023 using an electronic questionnaire. Logistic regression analyses were used to identify predictors of long COVID. Pre-infection and post-infection prediction models (with/without post-infection characteristics) were developed, and the C-index was used to evaluate accuracy.

RESULTS

In total, 420 patients infected COVID-19 were included. The prevalence of long COVID was 32.9 %. The most common symptoms of long COVID were weakness/fatigue, persistent cough and cognitive dysfunction. Independent predictors of long COVID included in the pre-infection model were age, long-term medication, and psychological problems such as stress and doing things without enthusiasm/interest before COVID-19 infection (C-index: 0.721). Independent predictors included in the post-infection model were age, inability to concentrate before COVID-19 infection, and symptoms of weakness/fatigue, abnormal smell/taste, diarrhoea, eye conjunctivitis and headache/dizziness during the acute-phase (C-index: 0.857).

CONCLUSIONS

Age, psychological problems before COVID-19 infection and acute-phase symptoms were important risk factors of long COVID. Results from the pre-infection model provide guidance for non-infected individuals on how to prevent long COVID. Results from the post-infection model can be used to accurately predict individuals who are at high risk of long COVID and help design treatment plans for patients in the acute phase.

Presepsin as a prognostic biomarker in COVID-19 patients: combining clinical scoring systems and laboratory inflammatory markers for outcome prediction

BACKGROUND

There is still limited research on the prognostic value of Presepsin as a biomarker for predicting the outcome of COVID-19 patients. Additionally, research on the combined predictive value of Presepsin with clinical scoring systems and inflammation markers for disease prognosis is lacking.

METHODS

A total of 226 COVID-19 patients admitted to Beijing Youan Hospital's emergency department from May to November 2022 were screened. Demographic information, laboratory measurements, and blood samples for Presepsin levels were collected upon admission. The predictive value of Presepsin, clinical scoring systems, and inflammation markers for 28-day mortality was analyzed.

RESULTS

A total of 190 patients were analyzed, 83 (43.7%) were mild, 61 (32.1%) were moderate, and 46 (24.2%) were severe/critically ill. 23 (12.1%) patients died within 28 days. The Presepsin levels in severe/critical patients were significantly higher compared to moderate and mild patients (p < 0.001). Presepsin showed significant predictive value for 28-day mortality in COVID-19 patients, with an area under the ROC curve of 0.828 (95% CI: 0.737-0.920). Clinical scoring systems and inflammation markers also played a significant role in predicting 28-day outcomes. After Cox regression adjustment, Presepsin, qSOFA, NEWS2, PSI, CURB-65, CRP, NLR, CAR, and LCR were identified as independent predictors of 28-day mortality in COVID-19 patients (all p-values < 0.05). Combining Presepsin with clinical scoring systems and inflammation markers further enhanced the predictive value for patient prognosis.

CONCLUSION

Presepsin is a favorable indicator for the prognosis of COVID-19 patients, and its combination with clinical scoring systems and inflammation markers improved prognostic assessment.

HB-EGF Plasmatic Level Contributes to the Development of Early Risk Prediction Nomogram for Severe COVID-19 Cases

(1) Background: Heparin-Binding Epidermal Growth Factor-like Growth Factor (HB-EGF) is involved in wound healing, cardiac hypertrophy, and heart development processes. Recently, circulant HB-EGF was reported upregulated in severely hospitalized COVID-19 patients. However, the clinical correlations of HB-EGF plasma levels with COVID-19 patients' characteristics have not been defined yet. In this study, we assessed the plasma HB-EGF correlations with the clinical and paraclinical patients' data, evaluated its predictive clinical value, and built a risk prediction model for severe COVID-19 cases based on the resulting significant prognostic markers. (2) Methods: Our retrospective study enrolled 75 COVID-19 patients and 17 control cases from May 2020 to September 2020. We quantified plasma HB-EGF levels using the sandwich ELISA technique. Correlations between HB-EGF plasma levels with clinical and paraclinical patients' data were calculated using two-tailed Spearman and Point-Biserial tests. Significantly upregulated parameters for severe COVID-19 cases were identified and selected to build a multivariate logistic regression prediction model. The clinical significance of the prediction model was assessed by risk prediction nomogram and decision curve analyses. (3) Results: HB-EGF plasma levels were significantly higher in the severe COVID-19 subgroup compared to the controls (p = 0.004) and moderate cases (p = 0.037). In the severe COVID-19 group, HB-EGF correlated with age (p = 0.028), pulse (p = 0.016), dyspnea (p = 0.014) and prothrombin time (PT) (p = 0.04). The multivariate risk prediction model built on seven identified risk parameters (age p = 0.043, HB-EGF p = 0.0374, Fibrinogen p = 0.009, PT p = 0.008, Creatinine p = 0.026, D-Dimers p = 0.024 and delta miR-195 p < 0.0001) identifies severe COVID-19 with AUC = 0.9556 (p < 0.0001). The decision curve analysis revealed that the nomogram model is clinically relevant throughout a wide threshold probability range. (4) Conclusions: Upregulated HB-EGF plasma levels might serve as a prognostic factor for severe COVID-19 and help build a reliable risk prediction nomogram that improves the identification of high-risk patients at an early stage of COVID-19.