Unraveling complex relationships between COVID-19 risk factors using machine learning based models for predicting mortality of hospitalized patients and identification of high-risk group: a large retrospective study.

Background: At the end of 2019, the coronavirus disease 2019 (COVID-19) pandemic increased the hospital burden of COVID-19 caused by the SARS-Cov-2 and became the most significant health challenge for nations worldwide. The severity and high mortality of COVID-19 have been correlated with various demographic characteristics and clinical manifestations. Prediction of mortality rate, identification of risk factors, and classification of patients played a crucial role in managing COVID-19 patients. Our purpose was to develop machine learning (ML)-based models for the prediction of mortality and severity among patients with COVID-19. Identifying the most important predictors and unraveling their relationships by classification of patients to the low-, moderate- and high-risk groups might guide prioritizing treatment decisions and a better understanding of interactions between factors. A detailed evaluation of patient data is believed to be important since COVID-19 resurgence is underway in many countries. Results: The findings of this study revealed that the ML-based statistically inspired modification of the partial least square (SIMPLS) method could predict the in-hospital mortality among COVID-19 patients. The prediction model was developed using 19 predictors including clinical variables, comorbidities, and blood markers with moderate predictability (Q2 = 0.24) to separate survivors and non-survivors. Oxygen saturation level, loss of consciousness, and chronic kidney disease (CKD) were the top mortality predictors. Correlation analysis showed different correlation patterns among predictors for each non-survivor and survivor cohort separately. The main prediction model was verified using other ML-based analyses with a high area under the curve (AUC) (0.81-0.93) and specificity (0.94-0.99). The obtained data revealed that the mortality prediction model can be different for males and females with diverse predictors. Patients were classified into four clusters of mortality risk and identified the patients at the highest risk of mortality, which accentuated the most significant predictors correlating with mortality. Conclusion: An ML model for predicting mortality among hospitalized COVID-19 patients was developed considering the interactions between factors that may reduce the complexity of clinical decision-making processes. The most predictive factors related to patient mortality were identified by assessing and classifying patients into different groups based on their sex and mortality risk (low-, moderate-, and high-risk groups).

Role of chest CT scan in patients with preexisting cancer and COVID-19 pneumonia.

BACKGROUND: Detection of COVID-19 in cancer patients is challenging due to probable preexisting pulmonary infiltration caused by many infectious and non-infectious etiologies. We evaluated chest CT scan findings of COVID-19 pneumonia in cancer patients and explored its prognostic role in mortality. METHODS: We studied 266 COVID-19 patients with a history of cancer diagnosis between 2020 and 2022. Chest CT images were reported based on Radiological Society of North America (RSNA) structural report and the CT score and pattern of involvement were noted. We used multivariate logistic regression models to determine the association between CT scan findings and mortality of the cancer COVID-19 patients. RESULTS: The mean age was 56.48 (± 18.59), and 53% were men. Gastrointestinal (29.3%), hematologic (26.3%), and breast (10.5%) cancers were the most frequent types of cancer. The prevalence of atypical or indeterminate findings in the chest CT was 42.8%. Most radiologic findings were consolidation mixed with ground-glass opacity (44.4%), pleural effusion (33.5%), and pure ground-glass opacity (19.5%). The risk of death was higher among those who had typical chest CT for COVID-19 (OR 3.47; 95% CI 1.14-8.98) and those who had a severity of score higher than 18 (OR 1.89; 95% CI 1.07-3.34). Also, presence of consolidation (P value 0.040), pleural effusion (P value 0.000), centrilobular nodules (P value 0.013), and architectural distortion (P value 0.005) were associated with a poorer prognosis. CONCLUSION: Less than half of COVID-19 patients with a history of cancer had typical imaging features of COVID-19. Radiologists should be aware of atypical, rare, or subtle chest CT findings in patients with pre-existing cancer.

Outcomes of coronavirus disease 19 patients with a history of rheumatoid arthritis: A retrospective registry-based study in Iran.

BACKGROUND: We aimed to study the outcomes of coronavirus disease 2019 (COVID-19) in patients with a history of rheumatoid arthritis (RA) in Iran, where most patients receive corticosteroids and are at high risk for COVID-19 infection. METHOD: We collected the demographic, diagnostic, and treatment data of all COVID-19 patients by the clinical COVID-19 registry system. We recruited 38 RA patients and 2216 non-RA patients from the COVID-19 registry. The primary outcome was mortality due to COVID-19. We also studied the risk of intensive care unit admission and intubation in RA patients compared to non-RA patients. We used multiple logistic regression analysis to study the association between RA and the risk of COVID-19 outcomes. RESULT: We recruited 38 RA patients and 2216 non-RA patients from the COVID-19 registry. The RA patients had a higher mean age (59.9 years) than the non-RA patients (57.7 years). The group of RA patients had a larger proportion of women (76.3%) than the non-RA patients (40.8%). The death rate due to COVID-19 was significantly higher in RA patients than non-RA patients (odds ratio [OR] = 2.69, 95% confidence interval [CI] = 1.24-5.81). The OR was higher among those who received prednisolone than among those who did not (OR = 3.59, 95% CI = 1.54-7.81). The odds of intubation were statistically significant among patients who received corticosteroid therapy (OR = 2.58, 95% CI = 1.07-6.18). CONCLUSION: The risk of COVID-19 outcomes was higher in RA patients than non-RA patients, especially for RA patients who received a low dose of prednisolone. The results of this study can be used to triage RA patients who get infected by COVID-19. Further studies with larger sample sizes are required to more precisely define the high-risk groups.

Incidence of symptomatic COVID-19 in close contacts of patients after discharge from hospital.

BACKGROUND: There is a little evidence about the infectiousness of recovered COVID-19 patients. Considering that the circumstance of the isolation of the COVID-19 patients after-discharge is not always optimal, it is not very unlikely that viral transmission still occurs after hospital discharge. This study aims to investigate the incidence of symptomatic COVID-19 in close contacts of recovered patients after discharge from hospital. METHODS: Four hundred fifty discharged COVID-19 patients discharged from the largest public treatment center in Tehran, capital city of Iran, were followed up. Demographic and clinical data of participants were collected from medical records. Follow-up data were acquired via telephone call interviews with patients or their main caregivers at home. RESULTS: The study's response rate was 93.77% (422 participated in the study). 60.90% patients were male and 39.10% were female (sex ratio = 1.55 male). The most prevalent comorbidities in these patients were hypertension (29.68%) and diabetes (24.80%). The mean of home isolation after discharge was 25.85. Forty-one (9.71%) patients had at least one new case in their close contacts, up to 3 weeks after they were discharged. There was a significant association between having at least a comorbidity with the odds of getting infected in close contacts [OR (CI) 2.22 (1.05-4.68)]. Density of inhabitant per room in a house' and the quality of isolation had significant associations with observing new cases in the patients' close contacts [high to moderate; OR (CI) 2.44 (1.06-5.61], [bad to good; OR (CI) 2.31 (1.17-4.59)], respectively. CONCLUSION: After hospital discharge, COVID-19 transmission can still occur, when a large number of people lives together in a single house. Another explanation can be that the less precaution measures are taken by recovered patients' cohabitants. Such conditions are also likely to happen when the recovered patient has other chronic diseases and requires additional care.

Differential Antibody Response to SARS-CoV-2 Antigens in Recovered and Deceased Iranian COVID-19 Patients.

The coronavirus infectious disease 2019 (COVID-19), which is initiated by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), has imposed critical challenges to global health. Understanding the kinetic of SARS-CoV-2-specific IgM and IgG responses in different subsets of COVID-19 patients is crucial to get insight into the humoral immune response elicited against the virus. We investigated IgM and IgG responses against SARS-CoV-2 nucleocapsid (N) and receptor-binding domain (RBD) of spike protein in two groups of recovered and deceased COVID-19 patients. The levels of IgM and IgG specific to N and RBD proteins were detected by ELISA. N- and RBD-specific IgM was higher in deceased patients in comparison with recovered patients, while there was no significant difference in N- and RBD-specific IgG between the two groups. A significant correlation was observed between IgG and IgM titers against RBD and N, in both groups of patients. These results argue against impaired antibody response in deceased COVID-19 patients.

Clinical Characteristics and Outcomes of 905 COVID-19 Patients Admitted to Imam Khomeini Hospital Complex in the Capital City of Tehran, Iran.

BACKGROUND: We studied the clinical characteristics and outcomes of 905 hospitalized coronavirus disease 2019 (COVID-19) patients admitted to Imam Khomeini Hospital Complex (IKHC), Tehran, Iran. METHODS: COVID-19 patients were recruited based on clinical symptoms and patterns of computed tomography (CT) imaging between February 20 and March 19. All patients were tested for the presence of COVID-19 RNA. The Poisson regression model estimated the incidence rate ratio (IRR) for different parameters. RESULTS: The average age (± standard deviation) was 56.9 (±15.7) years and 61.77% were male. The most common symptoms were fever (93.59%), dry cough (79.78%), and dyspnea (75.69%). Only 43.76% of patients were positive for the RT-PCR COVID-19 test. Prevalence of lymphopenia was 42.9% and more than 90% had elevated lactate dehydrogenase (LDH) or C-reactive protein (CRP). About 11% were severe cases, and 13.7% died in the hospital. The median length of stay (LOS) was 3 days. We found higher risks of mortality in patients who were older than 70 years (IRR = 11.77, 95% CI 3.63-38.18), underwent mechanical ventilation (IRR = 7.36, 95% CI 5.06-10.7), were admitted to the intensive care unit (ICU) (IRR = 5.47, 95% CI 4.00-8.38), tested positive on the COVID-19 test (IRR = 2.80, 95% CI 1.64-3.55), and reported a history of comorbidity (IRR = 1.76, 95% CI 1.07-2.89) compared to their corresponding reference groups. Hydroxychloroquine therapy was not associated with mortality in our study. CONCLUSION: Older age, experiencing a severe form of the disease, and having a comorbidity were the most important prognostic factors for COVID-19 infection. Larger studies are needed to perform further subgroup analyses and verify high-risk groups.