External Validation of COVID-19 Risk Scores during Three Waves of Pandemic in a German Cohort-A Retrospective Study

Utility of ED triage tools in predicting the need for intensive respiratory or vasopressor support in adult patients with COVID-19

BACKGROUND

Serum and radiological parameters used to predict prognosis in COVID patients are not feasible in the Emergency Department. Due to its damaging effect on multiple organs and lungs, scores used to assess multiorgan damage and pneumonia such as Pandemic Medical Early Warning Score (PMEWS), National Early Warning Score 2 (NEWS2), WHO score, quick Sequential Organ Failure Assessment (qSOFA), and DS-CRB 65 can be used to triage patients in the Emergency Department. They can be used to predict patients with the highest risk of seven-day mortality and need for intensive respiratory or vasopressor support (IRVS).

PURPOSE

The primary purpose was to find the score with the highest AUC in predicting IRVS and mortality at seven days. Additional objective was to find out any independent factors associated with IRVS and mortality.

METHODS

The data of adult patients who presented to the Emergency Department (ED) between April 1, 2021 and June 30, 2021 were collected. The WHO score, CRB-65, DS-CRB 65, PMEWS, NEWS2, and qSOFA score were calculated for all patients. Statistical analysis was done and an ROC curve was calculated for all the tools for mortality and need for IRVS at seven days.

FINDINGS

677 patients presented to the Emergency Department with COVID-19 during the period above. Presence of Diabetes Mellitus (p = 0.001), Hypertension (p = 0.001), and chronic kidney disease(CKD) (p = 0.04) was significantly associated with need for IRVS. Age, duration of symptoms, pulse rate, respiratory rate, room air saturation, mental status at admission, and time to IRVS need were identified as independent predictors of in-hospital mortality. The longer the time to IRVS need from ED arrival, the higher the likelihood of mortality. PMEWS (0.830) had the highest AUC, followed by NEWS2 (0.805). A PMEWS cut-off of 6.5 was 74.2% sensitive and 78.3% specific in predicting the need for IRVS. ROC analysis to predict 7-day mortality showed that PMEWS had an AUC of 0.802 (0.766-0.839). QSOFA performed poorly in predicting IRVS (AUC 0.645) and 7-day mortality (AUC 0.677).

CONCLUSION

PMEWS may be used for triaging patients presenting to the Emergency Department with COVID-19 and accurately predicts the need for IRVS and seven day mortality.

Comparison of Prognostic Scores for Patients with COVID-19 Presenting with Dyspnea in the Emergency Department

BACKGROUND

Easy-to-use bedside risk assessment is crucial for patients with COVID-19 in the overcrowded emergency department (ED).

OBJECTIVE

The aim of this study was to explore the prognostic ability of ratio of percutaneous oxygen saturation (SpO2) to fraction of inspired oxygen (FiO2) (S/F); ratio of SpO2/FiO2 to respiratory rate (ROX); National Early Warning Score (NEWS); quick Sequential Organ Failure Assessment (qSOFA); and confusion, respiratory rate, blood pressure, and age ≥ 65 years (CRB-65) in patients with COVID-19 presenting with dyspnea to the ED.

METHODS

In this retrospective observational study, clinical and demographic details of patients with COVID-19 were obtained at ED admission. S/F, ROX, NEWS, CRB-65, and qSOFA scores were calculated at the time of ED arrival. Accuracy of these five indices to predict the need for invasive mechanical ventilation (IMV) within 48 h, intensive care unit (ICU) admission, and early (7-day) mortality were determined using receiver operating characteristic curves.

RESULTS

A total of 375 patients were included in this study. Fifty patients (13.3%) required IMV within 48 h and 58 patients (15.5%) were transferred to the ICU. Seven-day mortality was 6.7% and 28-day mortality was 18.1%. Among all five scores determined from patient data on ED admission, ROX, S/F, and NEWS presented greater discriminatory performance than CRB-65 and qSOFA in predicting IMV within 48 h, ICU admission, and early mortality.

CONCLUSIONS

Emergency physicians can effectively use S/F, ROX, and NEWS scores for rapid risk stratification of patients with COVID-19 infection. Moreover, from the perspective of simplicity and ease of calculation, we recommend the use of the S/F ratio.

ANCOC Score to Predict Mortality in Different SARS-CoV-2 Variants and Vaccination Status

BACKGROUND

More than three years after the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) pandemic outbreak, hospitals worldwide are still affected by coronavirus disease 19 (COVID-19). The availability of a clinical score that can predict the risk of death from the disease at the time of diagnosis and that can be used even if population characteristics change and the virus mutates can be a useful tool for emergency physicians to make clinical decisions. During the first COVID-19 waves, we developed the ANCOC (age, blood urea nitrogen, C-reactive protein, oxygen saturation, comorbidities) score, a clinical score based on five main parameters (age, blood urea nitrogen, C-reactive protein, oxygen saturation, comorbidities) that accurately predicts the risk of death in patients infected with SARS-CoV-2. A score of less than -1 was associated with 0% mortality risk, whereas a score of 6 was associated with 100% risk of death, with an overall accuracy of 0.920. The aim of our study is to internally validate the ANCOC score and evaluate whether it can predict 60-day mortality risk independent of vaccination status and viral variant.

METHODS

We retrospectively enrolled 843 patients admitted to the emergency department (ED) of our hospital with a diagnosis of COVID-19. A total of 515 patients were admitted from July 2021 to September 2021, when the Delta variant was prevalent, and 328 in January 2022, when the Omicron 1 variant was predominant. All patients included in the study had a diagnosis of COVID-19 confirmed by polymerase chain reaction (PCR) on an oropharyngeal swab. Demographic data, comorbidities, vaccination data, and various laboratory, radiographic, and blood gas parameters were collected from all patients to determine differences between the two waves. ANCOC scores were then calculated for each patient, ranging from -6 to 6.

RESULTS

Patients infected with the Omicron variant were significantly older and had a greater number of comorbidities, of which hypertension and chronic obstructive pulmonary disease (COPD) were the most common. Immunization was less common in Delta patients than in Omicron patients (34% and 56%, respectively). To assess the accuracy of mortality prediction, we constructed a receiver operating characteristic (ROC) curve and found that the area under the ROC curve was greater than 0.8 for both variants. These results suggest that the ANCOC score is able to predict 60-day mortality regardless of viral variant and whether the patient is vaccinated or not.

CONCLUSION

In a population with increasingly high vaccination rates, several parameters may be considered prognostic for the risk of fatal outcomes. This study suggests that the ANCOC score can be very useful for the clinician in an emergency setting to quickly understand the patient's evolution and provide proper attention and the most appropriate treatments.

[Risk factors for mortality from COVID-19 Omicron variant: Retrospective analysis in elderly from the Canary Islands]

BACKGROUND AND AIMS

Since the beginning of the COVID-19 pandemic, the elderly population has had the highest rates of complications and mortality. This study aimed to determine the influence of different risk factors on deaths due to the Omicron variant in the Canary Islands.

MATERIALS AND METHODS

A retrospective observational study of 16,998 cases of COVID-19 over 40 years of age was conducted in the Canary Islands between August 1, 2022, and January 31, 2023. We extracted sociodemographic data (age and sex) and clinical data (death, vaccination history, hospital admission, previous diseases, and treatments).

RESULTS

Among the deaths, there was a higher proportion of males aged over 70 years, with diabetes, cardiovascular, renal, respiratory, and systemic diseases, and nursing home residents. Significant differences were observed in the number of doses of the vaccine. The multiple regression model showed that male sex (OR [95% CI]=1.92 [1.42-2.58]), age (70-79 years, 9.11 [4.27-19.43]; 80-89 years, 21.72 [10.40-45.36]; 90-99 years, 66.24 [31.03-141.38]; 100 years or older, 69.22 [12.97-369.33]), being unvaccinated (6.96, [4.01-12.08]), or having the last dose administered at least 12 months before the diagnosis (2.38, [1.48-3.81]) were significantly associated with mortality.

CONCLUSIONS

Multiple factors may increase the risk of mortality due to COVID-19 in the elderly population. In our study, we found that only three predictors can effectively explain the variability: older age, male sex, and not being vaccinated or last vaccination date prior to one year.

Machine learning models for predicting severe COVID-19 outcomes in hospitals

The aim of this observational retrospective study is to improve early risk stratification of hospitalized Covid-19 patients by predicting in-hospital mortality, transfer to intensive care unit (ICU) and mechanical ventilation from electronic health record data of the first 24 h after admission. Our machine learning model predicts in-hospital mortality (AUC = 0.918), transfer to ICU (AUC = 0.821) and the need for mechanical ventilation (AUC = 0.654) from a few laboratory data of the first 24 h after admission. Models based on dichotomous features indicating whether a laboratory value exceeds or falls below a threshold perform nearly as good as models based on numerical features. We devise completely data-driven and interpretable machine-learning models for the prediction of in-hospital mortality, transfer to ICU and mechanical ventilation for hospitalized Covid-19 patients within 24 h after admission. Numerical values of. CRP and blood sugar and dichotomous indicators for increased partial thromboplastin time (PTT) and glutamic oxaloacetic transaminase (GOT) are amongst the best predictors.