Can artificial intelligence predict COVID-19 mortality?

OBJECTIVE

COVID-19 infection rapidly spread across the globe and evolved into a pandemic. Artificial intelligence (AI) has been used to predict the spread of the virus and diagnose and treat the disease. In this study, we aimed to predict whether patients admitted to the intensive care unit (ICU) due to COVID-19 infection will result in mortality.

PATIENTS AND METHODS

Ninety parameters were used for each 589 ICU patient. The nine parameters with the highest effect on mortality were determined. Four hundred seventy-one patients were used to train the AI with these nine parameters. AI was tested with 118 patient data.

RESULTS

AI estimated mortality with 83% sensitivity, 84% specificity, 84% accuracy, and 0.81 F1 score. Therefore, the area under the curve was calculated as 0.91. The results indicate that mortality among COVID-19 patients admitted to the ICU can be predicted based on their laboratory parameters on the first day.

CONCLUSIONS

These findings underscore the potential benefits of utilizing AI in managing pandemics.

The aggregate index of systemic inflammation may predict mortality in COVID-19 patients with chronic renal failure

OBJECTIVE

Coronavirus disease 2019 (COVID-19), caused by severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2), was first detected in December 2019 and then spread globally, resulting in a pandemic. Initially, it was unknown if chronic kidney disease (CKD) contributed to the mortality caused by COVID-19. The immunosuppression associated with this disease may minimize the COVID-19-described hyper-inflammatory state or immunological dysfunction, and a high prevalence of comorbidities may lead to a poorer clinical prognosis. Patients with COVID-19 have abnormal circulating blood cells associated with inflammation. Risk stratification, diagnosis, and prognosis primarily rely on hematological features, such as white blood cells and their subpopulations, red cell distribution width, mean platelet volume, and platelet count, in addition to their combined ratios. In non-small-cell lung cancer, the aggregate index of systemic inflammation (AISI), (neutrophils x monocytes x platelets/lymphocytes) is evaluated. In light of the relevance of inflammation in mortality, the objective of this study is to determine the impact of AISI on the hospital mortality of CKD patients.

PATIENTS AND METHODS

This study is an observational retrospective study. Data and test outcomes of all CKD patients, stages 3-5, hospitalized for COVID-19 and followed between April and October 2021 were analyzed.

RESULTS

Patients were divided into two groups according to death (Group 1-Alive, Group 2-Died). Neutrophil count, AISI and C-reactive protein (CRP) levels were increased in Group-2 [10.3±4.6 vs. 7.65±4.22; p=0.001, 2,084.1 (364.8-2,577.5) vs. 628.9 (53.1-2,275); p=0.00 and 141.9 (20.5-318) vs. 84.75 (0.92-195); p=0.00; respectively]. Receiver operating characteristic (ROC) analysis demonstrated 621.1 as a cut-off value for AISI to predict hospital mortality with 81% sensitivity and 69.1% specificity [area under ROC curve 0.820 (95% CI: 0.733-0.907), p<.005]. Cox regression analysis was used to analyze the effect of risk variables on survival. In survival analysis, AISI and CRP were identified as important survival predictors [hazard ratio (HR): 1.001, 95% CI: 1-1.001; p=0.00 and HR: 1.009, 95% CI: 1.004-1.013; p=0.00].

CONCLUSIONS

This study demonstrated the discriminative effectiveness of AISI in predicting disease mortality in COVID-19 patients with CKD. Quantification of AISI upon admission might assist in the early detection and treatment of individuals with a bad prognosis.

Negativity of the electromechanical window: relation to frequent premature ventricular complexes

OBJECTIVE

The electromechanical window (EMW) was investigated as a new predictor of arrhythmia in the presence of long QT. However, the use of EMW to predict idiopathic frequent ventricular premature complexes (PVCs) in those with normal QT intervals has not been clarified.

PATIENTS AND METHODS

This single-center study included consecutive patients who presented to the Cardiology Clinic with palpitations and were found to have idiopathic PVC on 24-hour Holter monitoring. Those with a PVC/24-hour frequency of < 1% were defined as group 1, 1-10% as group 2, and > 10% as group 3. The EMW was defined as the time difference (in ms) between the aortic valve closure and the end of the QT interval, measured from an ECG on the concurrent echocardiogram.

RESULTS

A total of 148 patients were included in the study, 64% (n = 94) of which were female. The patients' mean age was 50.11 ± 14.7. The groups were similar in terms of the patients' age, BMI, and comorbidities. There was a statistically significant difference between the three groups in terms of the EMW measurements (group 1: 3.78 ± 19.6, group 2: -7 ± 30.9, group 3: -34.83 ± 55.2 ms: p < 0.001). In the multivariate regression analysis, the EMW (OR 0.971, p = 0.007) and every 10-ms decrease in the EMW (OR 1.254, p = 0.011) were thus determined to be independent predictors of PVC > 10%. An EMW value of ≤ -15 ms was associated with the frequency of 24-h PVC > 10%, with a sensitivity of 70% and a specificity of 70% (AUC 0.716, 95% CI: 0.636-0.787 p < 0.001).

CONCLUSIONS

The results showed that a negative increase in the EMW may be associated with frequent idiopathic PVCs.

The effect of thrombosis-related laboratory values on mortality in COVID-19 infection

OBJECTIVE

COVID-19 may cause thrombosis in both venous and arterial systems. Familiarity with the signs and symptoms of thrombosis and its treatment is essential in treating COVID-19 infection and its complications. D-Dimer and mean platelet volume (MPV) are measurements related to the development of thrombosis. This study investigates whether MPV and D-Dimer values could be used to determine the risk of thrombosis and mortality in the COVID-19 early stages.

PATIENTS AND METHODS

424 patients who were COVID-19 positive, according to the World Health Organization (WHO) guidelines, were randomly and retrospectively included in the study. Demographic and clinical characteristics such as age, gender, and length of hospitalization were obtained from the digital records of participants. Participants were divided into living and deceased groups. The patients' biochemical, hormonal, and hematological parameters were analyzed retrospectively.

RESULTS

White blood cells (WBC), neutrophils, and monocytes were significantly different in the two groups (p-value <0.001), and their values were lower in the living group than in the deceased group. MPV median values did not differ according to prognosis (p-value = 0.994). While the median value was 9.9 in the survivors, it was 10 in the deceased. Creatinine, procalcitonin, ferritin, and the number of hospitalization days in living patients were significantly lower than in patients who died (p-value <0.001). Median values of D-dimer (mg/L) differ according to prognosis (p-value <0.001). While the median value was 0.63 in the survivors, it was found as 438 in the deceased.

CONCLUSIONS

Our results did not show any significant relationship between the mortality of COVID-19 patients and their MPV levels. However, a significant association between D-Dimer and mortality in COVID-19 patients was observed.

The significance of a novel inflammatory biomarker, presepsin, in predicting disease prognosis in patients with COVID-19

OBJECTIVE

This study aims at determining the significance of a novel inflammatory biomarker, presepsin, in predicting disease prognosis in patients with COVID-19.

PATIENTS AND METHODS

This retrospective study was concluded at the University Hospital between April and August 2020. The study involved 88 COVID-19 patients (48 men and 40 women). The patients were categorized into two groups: the patients admitted to the COVID-19 clinic, described as the moderate COVID-19 patients (Group-1; n=44), and those admitted to the internal medicine outpatient clinic, who were the mild COVID-19 patients (Group-2; n=44). The groups were compared using inflammatory markers: presepsin, C-Reactive Protein to Albumin Ratio, Neutrophil to Lymphocyte Ratio, and procalcitonin.

RESULTS

Serum presepsin levels (195.29 vs. 52.12 pg/ml) were significantly higher in the Group-1 compared to the Group-2 (p=0.001). The gender distribution and average age were similar in both groups (p > 0.05). While ferritin, lactate dehydrogenase, D-Dimer, platelet lymphocyte ratio, C-Reactive Protein to Albumin Ratio (p=0.001), erythrocyte sedimentation ratio, C-Reactive Protein and presepsin were significantly higher in the Group-1 compared to Group-2 (p<0.05), while hemoglobin and lymphocyte were significantly lower in the Group-1 than in Group-2 (p<0.05).

CONCLUSIONS

Serum presepsin levels were found to be significantly higher in moderate clinical group COVID-19 patients compared to mild group. Presepsin, a new inflammatory biomarker, may be useful in predicting the prognosis and early treatment of COVID-19 infection.

Predictors of disease severity, clinical course, and therapeutic outcome in COVID-19 patients: our experience with 1,700 patients

OBJECTIVE

Our study aimed at investigating the impacts of demographic, hematological, and biochemical factors on the clinical course and the prognostic outcome in adult COVID-19 patients.

PATIENTS AND METHODS

This retrospective study was performed in the internal medicine departments of two hospitals, and data were extracted from the medical files of 1,700 adult COVID-19 patients (836 females, 49.2%; 864 males, 50.8%) with an average age of 48.23 ± 16.68 (range: 18-93). Clinical data included baseline descriptives, prior medical history, admission date, treatment, and hematological and biochemical blood test results. The relationship between the survival, length of hospitalization, hematological, and biochemical parameters was investigated.

RESULTS

Advanced age (p<0.001), presence of at least on comorbid disease (p=0.045), increased length of hospitalization (p=0.006), elevated white blood cell (p=0.001) and neutrophil (p=0.002) counts, increased serum levels of glucose (p=0.027), blood urea nitrogen (p<0.001), AST (p=0.006), LDH (p<0.001), CRP (p>0.001), and D-dimer (p=0.001). In contrast, diminution of serum levels of albumin (p<0.001), ALT (p=0.028), calcium (p=0.022), and platelet count (p=0.010) were associated with increased mortality. There was a positive and weak relationship between serum D-dimer levels and length of hospitalization.

CONCLUSIONS

Our data imply that identifying and validating indicators that predict COVID-19 disease progression to improve health outcomes is crucial. Age, comorbidities, immunological response, radiographic abnormalities, laboratory markers, and signs of organ dysfunction may all predict poor outcomes individually or collectively. Identifying characteristics that predict COVID-19 problems is critical to guiding clinical management, improving patient outcomes, and allocating limited resources.

The novel biomarker, neopterin, can predict the severity of COVID-19

OBJECTIVE

SARS-CoV-2 infection primarily affects T-lymphocytes, particularly CD4+ and CD8+ T cells. However, there is a need for simpler and less expensive laboratory tests with predictive values comparable to CD4+ cell counts. Thus, the goal of this study was to investigate the role of neopterin levels in predicting intensive care and mortality in coronavirus disease patients in 2019.

PATIENTS AND METHODS

This retrospective study included 87 hospitalized patients who were diagnosed with COVID-19. Patients were divided into two groups: those receiving intensive care (Severe COVID-19; S-COVID-19) and those receiving non-intensive care (Moderate COVID-19; M-COVID-19). Patients' clinical characteristics, serum neopterin levels, and other laboratory data were compared across groups.

RESULTS

The average age was 63.9±155.2 years, and 44 (%) of the participants were male. WBC (p = 0.008), neutrophil (p = 0.002), HDL (p = 0.009), ferritin, calcium, albumin, LDH, APTT, lymphocyte, INR, D-dimer, troponin, prothrombin time sedimentation, and PaO2 (p = 0.001) were all associated with death. The neopterin level in the M-COVID-19 group was 3 (min-max; 3.1-5.9) and 3.2 (2.3-7) in the S-COVID-19 group, with no statistically significant difference (p = 0.456). Gender differences between groups were not significant (p = 0.183). According to the ROC analysis, if parameters such as age, D-Dimer, troponin, ferritin, albumin, LDH, CRP, procalcitonin, and PaO2 exceed the cut-off values and lymphocyte levels are below, it can predict the need for intensive care and mortality in COVID-19 patients.

CONCLUSIONS

Although we did not find statistically significant results with neopterin in terms of mortality in COVID-19 individuals in our study, more thorough, prospective, randomized controlled studies with expanded patient populations at various phases of the disease are needed.

The role of neutrophil albumin ratio in predicting the stage of non-small cell lung cancer

OBJECTIVE

Inflammation forms the basis of cancer development and progression. It causes changes in complete blood count parameters, such as neutrophil counts. Low albumin levels are associated with poor prognosis in cancer patients. We aimed to investigate the association between neutrophil to albumin ratio (NAR) and the stage of non-small cell lung cancer (NSCLC).

PATIENTS AND METHODS

257 NSCLC patients (24 females and 198 males) were included in the study. Patients were divided into two groups. Group 1 (n=61) included patients with early stage cancer (stage 1 and 2), while group 2  (n=196) included those with advanced stage cancer (stage 3 and 4). Demographic data, neutrophil, lymphocyte, platelet, white blood cell counts (WBC), C-reactive protein (CRP), erythrocyte sedimentation rate (ESR), ferritin and albumin levels at the time of diagnosis were recorded. The NAR of 2 groups were compared.

RESULTS

There were no significant differences between the lymphocyte count (2.0 vs. 2.0 103/mm3) and platelet count (291 vs. 311 103/mm3) of the two groups (p > 0.05). ESR (38.8 vs. 57.5 mm/h), CRP (158 vs. 57 mg/l), ferritin (85 vs. 261 ng/ml), WBC count (8.6 vs. 10.6 103/mm3), neutrophil count (5.6 vs. 7.5 103/mm3), albumin values (2.9 vs. 3.7 gr/dl), and (p < 0.05) NAR levels (1.6 vs. 2.3) (p < 0.05) were significantly higher in group 2.

CONCLUSIONS

NAR can be used in predicting the stage of NSCLC.

THE USEFULNESS OF MONITORING THE NEUTROPHIL TO LYMPHOCYTE RATIO IN PATIENTS WITH PERIPHERAL VERTIGO

The purpose of this study was to evaluate the usefulness of the neutrophil to lymphocyte ratio (NLR) in a differential diagnosis and follow-up of patients with peripheral vertigo. Twenty patients with benign positional paroxysmal vertigo (BPPV) and 20 patients diagnosed with vestibular neuritis (VN) were included in the study. Serum samples were analysed at the initial presentation and on the seventh day of admission retrospectively. The WBC (white blood cell) count was 10500±2100 /mm3, the neutrophil count was 4700±1100/mm3, the lymphocyte count was 5000±1200/mm3 and the NLR was 0.9±0.2 in the VN group. In patients with BPPV, the WBC count was 9200±1300/mm3, the neutrophil count was 5200±1200/mm3, the lymphocyte count was 3100±1200/mm3 and the NLR was 1.9±0.9. The NLR was lower in patients with VN than in patients diagnosed with BPPV. The WBC and lymphocyte count was significantly higher in the patients with VN than in the patients diagnosed with BPPV. Within the first week of admission, the WBC and lymphocyte counts in patients with VN decreased, and the NLR was more elevated than at the admission. It is highly recommended that NLR is used in the diagnosis and follow-up of the most commonly observed aetiological factors of peripheral vertigo, BPPV and VN.

EFFECTS OF VITAMIN D ANALOGS ON ERYTHROPOIESIS-STIMULATING AGENT DASAGE AND SECONDARY ANEMIA IN HEMODIALYSIS PATIENTS

To investigate the effects of Vitamin D Analogs, paricalcitol and alphacalcidol, on hemoglobin levels and erythropoietin-stimulating agents' dosage in hemodialysis patients with chronic renal failure. A total of 310 patients under hemodialysis treatment for chronic renal failure were included in this retrospective multicenter study. Data on serum parathormone and hemoglobin levels, erythropoietin-stimulating agents' doses, C-reactive protein, calcium and phosphate levels were collected from medical records to comparatively evaluate paricalcitol, alphacalcidol and no treatment groups. Apart from significantly higher levels for hematocrit in patients treated with paricalcitol compared to pre-treatment values (32.3(3.8) vs. 34.1(3.1) p=0.007), pre-treatment and post-treatment values for biochemical parameters were similar in paricalcitol and alphacalcidol groups including ESA dose. A significant increase in parathormone levels (p=0.000 for each) while a significant decrease in calcium (p=0.003 and 0.040, respectively), Hb (p=0.001 and 0.009, respectively) and hematocrit (p=0.001 and 0.021, respectively) levels were determined in paricalcitol and alphacalcidol treated patients compared with untreated patients. Also, phosphate levels in alphacalcidol treated patients were significantly higher (p=0.018) than untreated patients. Our findings revealed insufficient suppression of parathormone levels and there of lower hemoglobin and hematocrit levels, but similar ESA dosage among CRF patients treated with Vitamin D analogs compared with untreated patients.

Induction of T-cell apoptosis by Actinobacillus actinomycetemcomitans mutants with deletion of ltxA and cdtABC genes: possible activity of GroEL-like molecule

The pathogenic bacterium Actinobacillus actinomycetemcomitans expresses a leukotoxin (Ltx) and cytolethal distending toxin (CDT) with cytolytic properties. CDT also has cytostatic properties, inducing a G2 cell cycle block. The extent of the contribution of these, as well as other toxins, to the cytolytic and cytostatic activities of this microorganism have not been defined and the aim of this study was to determine their contribution. To that end, a naturally transformable A. actinomycetemcomitans clinical strain (D7S-smooth) was used to construct a series of deletion mutants (DeltacdtA, DeltacdtB, DeltacdtC, DeltacdtABC, DeltaltxA, DeltaltxA/DeltacdtABC). Human peripheral blood mononuclear cells were incubated with cell-associated and extracellular bacterial preparations. The ability of wild type and isogenic mutants to induce T-cell apoptosis and cell cycle arrest was compared. The expression of ltxA and each of the cdt gene loci partially contributed to A. actinomycetemcomitans apoptosis, since each of the isogenic mutants exhibited reduced ability to induce T-cell apoptosis. Conversely, the ability to induce cell cycle block was abolished in each of the cdt isogenic mutants. A mutant with simultaneous deletion of ltxA and cdtABC genes retained potent ability to induce apoptosis in its cell-associated, but not extracellular, preparation. Neutralization with Escherichia coli anti-GroEL monoclonal antibody, lead to significant diminution of apoptosis-inducing activity of the DeltaltxA/DeltacdtABC cell-associated preparation. These data provide evidence for the expression of other A. actinomycetemcomitans cytolytic molecule(s) distinct from CDT and leukotoxin, with a possible role for GroEL-like molecule in T-cell apoptosis.

Actinobacillus actinomycetemcomitans induces apoptosis of T lymphocytes by the Fas and Fas ligand pathway

Actinobacillus actinomycetemcomitans expresses a number of toxins capable of inducing apoptotic cell death of T lymphocytes. However, the exact mechanism(s) has not been elucidated. The present study investigated the involvement of the Fas (CD95)-mediated apoptotic pathway in A. actinomycetemcomitans-induced T-cell apoptosis. To that end, peripheral blood mononuclear cells (PBMC) were cultured with or without A. actinomycetemcomitans cell-free culture supernatant (CFCS) for 0-96 h. The cells were then labeled with specific monoclonal antibodies and flow cytometry was performed. Results demonstrated up-regulation of Fas and activation of caspase-3 in T cells in response to A. actinomycetemcomitans CFCS. Monocytes were the only cells analyzed to express Fas ligand (FasL) constitutively, and this was further up-regulated in response to A. actinomycetemcomitans CFCS, while T cells expressed FasL only after this stimulation. Depletion of monocytes prior to stimulation with A. actinomycetemcomitans CFCS led to a marked decline in apoptosis. Blocking of Fas-FasL interactions with anti-Fas monoclonal antibody or Fas:Fc fusion protein lead to a significant decline, but not abolition, of T-cell apoptosis. Nearly all T cells expressed Bcl-2 at the outset of culture, and Bcl-2 expression declined in T cells stimulated with A. actinomycetemcomitans CFCS. Collectively, these data provide evidence for the induction of T-cell apoptosis by A. actinomycetemcomitans via the Fas-mediated pathway, involving caspase-3 and Bcl-2. Moreover, this apoptotic response was dependent on the presence of monocytes.

Large-scale early in vitro response to actinobacillus actinomycetemcomitans suggests superantigenic activation of T-cells

The mode of T-cell response to Actinobacillus actinomycetemcomitans is largely unknown. The present study sought to investigate the hypothesis that A. actinomycetemcomitans expresses superantigens, capable of antigen-non-specific T-cell activation. To that end, peripheral blood mononuclear cells were stimulated with A. actinomycetemcomitans, and T-cell expression of the early activation marker, CD69, was determined by flow cytometry. Results showed that A. actinomycetemcomitans activated a large number of T-cells with magnitude similar to that of staphylococcal enterotoxin superantigens. A. actinomycetemcomitans sonicate preferentially activated T-cells expressing Vbeta5.1 and Vbeta8, while the extracellular preparation activated Vbeta5.1+, Vbeta8+, and Vbeta12+ T-cells. T-cell response to A. actinomycetemcomitans was observed in the presence of autologous, as well as heterologous, antigen-presenting cells, suggesting a MHC-non-restricted response. Thus, the in vitro response to A. actinomycetemcomitans is characterized by large-scale T-cell activation in a Vbeta-specific and MHC-non-restricted manner, consistent with the involvement of superantigens.

Evidence for apoptosis of the majority of T cells activated in vitro with Actinobacillus actinomycetemcomitans

Our previous studies had demonstrated that nearly half of all T cells stimulated with Actinobacillus actinomycetemcomitans are activated within a few hours. However, it was not known whether all of these T cells survive. The aim of the present study was to determine whether the T cells activated in response to A. actinomycetemcomitans undergo apoptosis. To that end, peripheral blood mononuclear cells were cultured at different time points in the presence of A. actinomycetemcomitans. Flow cytometric analysis demonstrated that, following exposure to a preparation of A. actinomycetemcomitans, T cells progressively externalized their plasma membrane phosphatidylserine, as measured by annexin V binding. Approximately half of all T cells bound annexin V by 96 h. During this period, Annexin V-positive T cells also incorporated propidium iodide suggesting loss of membrane integrity. The externalization of phosphatidylserine occurred at a higher rate among activated (CD69+) T cells, where roughly two-thirds became Annexin V-positive. Flow cytometric analysis also demonstrated shrinkage of the Annexin V-positive and propidium iodide-positive T cells. The data presented here provides evidence for the induction of apoptosis among the majority of the T cells responding to A. actinomycetemcomitans.