Machine learning and predictive models: 2 years of Sars-CoV-2 pandemic in a single-center retrospective analysis

Background Since January 2020, coronavirus disease 19 (COVID-19) has rapidly spread all over the world. An early assessment of illness severity is crucial for the stratification of patients in order to address them to the right intensity path of care. We performed an analysis on a large cohort of COVID-19 patients (n=581) hospitalized between March 2020 and May 2021 in our intensive care unit (ICU) at Policlinico Riuniti di Foggia hospital. Through an integration of the scores, demographic data, clinical history, laboratory findings, respiratory parameters, a correlation analysis, and the use of machine learning our study aimed to develop a model to predict the main outcome. Methods We deemed eligible for analysis all adult patients (age >18 years old) admitted to our department. We excluded all the patients with an ICU length of stay inferior to 24 h and the ones that declined to participate in our data collection. We collected demographic data, medical history, D-dimers, NEWS2, and MEWS scores on ICU admission and on ED admission, PaO2/FiO2 ratio on ICU admission, and the respiratory support modalities before the orotracheal intubation and the intubation timing (early vs late with a 48-h hospital length of stay cutoff). We further collected the ICU and hospital lengths of stay expressed in days of hospitalization, hospital location (high dependency unit, HDU, ED), and length of stay before and after ICU admission;the in-hospital mortality;and the in-ICU mortality. We performed univariate, bivariate, and multivariate statistical analyses. Results SARS-CoV-2 mortality was positively correlated to age, length of stay in HDU, MEWS, and NEWS2 on ICU admission, D-dimer value on ICU admission, early orotracheal intubation, and late orotracheal intubation. We found a negative correlation between the PaO2/FiO2 ratio on ICU admission and NIV. No significant correlations with sex, obesity, arterial hypertension, chronic obstructive pulmonary disease, chronic kidney disease, cardiovascular disease, diabetes mellitus, dyslipidemia, and neither MEWS nor NEWS on ED admission were observed. Considering all the pre-ICU variables, none of the machine learning algorithms performed well in developing a prediction model accurate enough to predict the outcome although a secondary multivariate analysis focused on the ventilation modalities and the main outcome confirmed how the choice of the right ventilatory support with the right timing is crucial. Conclusion In our cohort of COVID patients, the choice of the right ventilatory support at the right time has been crucial, severity scores, and clinical judgment gave support in identifying patients at risk of developing a severe disease, comorbidities showed a lower weight than expected considering the main outcome, and machine learning method integration could be a fundamental statistical tool in the comprehensive evaluation of such complex diseases.

Effects of Positive End-Expiratory Pressure in "High Compliance" Severe Acute Respiratory Syndrome Coronavirus 2 Acute Respiratory Distress Syndrome.

OBJECTIVES: Clinical observation suggests that early acute respiratory distress syndrome induced by the severe acute respiratory syndrome coronavirus 2 may be "atypical" due to a discrepancy between a relatively unaffected static respiratory system compliance and a significant hypoxemia. This would imply an "atypical" response to the positive end-expiratory pressure. DESIGN: Single-center, unblinded, crossover study. SETTING: ICU of Bari Policlinico Academic Hospital (Italy), dedicated to care patients with confirmed diagnosis of novel coronavirus disease 2019. PATIENTS: Eight patients with early severe acute respiratory syndrome coronavirus 2 acute respiratory distress syndrome and static respiratory compliance higher than or equal to 50 mL/cm H2O. INTERVENTIONS: We compared a "lower" and a "higher" positive end-expiratory pressure approach, respectively, according to the intervention arms of the acute respiratory distress syndrome network and the positive end-expiratory pressure setting in adults with acute respiratory distress syndrome studies. MEASUREMENTS AND MAIN RESULTS: Patients were ventilated with the acute respiratory distress syndrome network and, subsequently, with the ExPress protocol. After 1 hour of ventilation, for each protocol, we recorded arterial blood gas, respiratory mechanics, alveolar recruitment, and hemodynamic variables. Comparisons were performed with analysis of variance for repeated measures or Friedman test as appropriate. Positive end-expiratory pressure was increased from 9 ± 3.5 to 17.7 ± 1.7 cm H2O (p < 0.01). Alveolar recruitment was 450 ± 111 mL. Static respiratory system compliance decreased from 58.3 ± 7.6 mL/cm H2O to 47.4 ± 14.5 mL/cm H2O (p = 0.018) and the "stress index" increased from 0.97 ± 0.03 to 1.22 ± 0.07 (p < 0.001). The PaO2/FIO2 ratio increased from 131 ± 22 to 207 ± 41 (p < 0.001), and the PaCO2 increased from 45.9 ± 12.7 to 49.8 ± 13.2 mm Hg (p < 0.001). The cardiac index went from 3.6 ± 0.4 to 2.9 ± 0.6 L/min/m (p = 0.01). CONCLUSIONS: Our data suggest that the "higher" positive end-expiratory pressure approach in patients with severe acute respiratory syndrome coronavirus 2 acute respiratory distress syndrome and high compliance improves oxygenation and lung aeration but may result in alveolar hyperinflation and hemodynamic alterations.

A Lower Global Lung Ultrasound Score Is Associated with Higher Likelihood of Successful Extubation in Invasively Ventilated COVID-19 Patients.

Lung ultrasound (LUS) can be used to assess loss of aeration, which is associated with outcome in patients with coronavirus disease 2019 (COVID-19) presenting to the emergency department. We hypothesized that LUS scores are associated with outcome in critically ill COVID-19 patients receiving invasive ventilation. This retrospective international multicenter study evaluated patients with COVID-19-related acute respiratory distress syndrome (ARDS) with at least one LUS study within 5 days after invasive mechanical ventilation initiation. The global LUS score was calculated by summing the 12 regional scores (range 0-36). Pleural line abnormalities and subpleural consolidations were also scored. The outcomes were successful liberation from the ventilator and intensive care mortality within 28 days, analyzed with multistate, competing risk proportional hazard models. One hundred thirty-seven patients with COVID-19-related ARDS were included in our study. The global LUS score was associated with successful liberation from mechanical ventilation (hazard ratio [HR]: 0.91 95% confidence interval [CI] 0.87-0.96; P = 0.0007) independently of the ARDS severity, but not with 28 days mortality (HR: 1.03; 95% CI 0.97-1.08; P = 0.36). Subpleural consolidation and pleural line abnormalities did not add to the prognostic value of the global LUS score. Examinations within 24 hours of intubation showed no prognostic value. To conclude, a lower global LUS score 24 hours after invasive ventilation initiation is associated with increased probability of liberation from the mechanical ventilator COVID-19 ARDS patients, independently of the ARDS severity.

A Pilot Study on Electrical Impedance Tomography During CPAP Trial in Patients With Severe Acute Respiratory Syndrome Coronavirus 2 Pneumonia: The Bright Side of Non-invasive Ventilation.

Background: Different severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) pneumonia phenotypes were described that match with different lung compliance and level of oxygenation, thus requiring a personalized ventilator setting. The burden of so many patients and the lack of intensive care unit (ICU) beds often force physicians to choose non-invasive ventilation (NIV) as the first approach, even if no consent has still been reached to discriminate whether it is safer to choose straightforward intubation, paralysis, and protective ventilation. Under such conditions, electrical impedance tomography (EIT), a non-invasive bedside tool to monitor lung ventilation and perfusion defects, could be useful to assess the response of patients to NIV and choose rapidly the right ventilatory strategy. Objective: The rationale behind this study is that derecruitment is a more efficient measure of positive end expiratory pressure (PEEP)-dependency of patients than recruitment. We hypothesized that patients who derecruit significantly when PEEP is reduced are the ones that do not need early intubation while small end-expiratory lung volume (ΔEELV) variations after a single step of PEEP de-escalation could be predictive of NIV failure. Materials and Methods: Consecutive patients admitted to ICU with confirmed SARS-CoV-2 pneumonia ventilated in NIV were enrolled. Exclusion criteria were former intubation or NIV lasting > 72 h. A trial of continuos positive airway pressure (CPAP) 12 was applied in every patient for at least 15 min, followed by the second period of CPAP 6, either in the supine or prone position. Besides standard monitoring, ventilation of patients was assessed by EIT, and end-expiratory lung impedance (ΔEELI) (%) was calculated as the difference in EELI between CPAP12 and CPAP6. Tidal volume (Vt), Ve, respiratory rate (RR), and FiO2 were recorded, and ABGs were measured. Data were analyzed offline using the dedicated software. The decision to intubate or continue NIV was in charge of treating physicians, independently from study results. Outcomes of patients in terms of intubation rate and ICU mortality were recorded. Results: We enrolled 10 male patients, with a mean age of 67 years. Six patients (60%) were successfully treated by NIV until ICU discharge (Group S), and four patients failed NIV and were intubated and switched to MV (Group F). All these patients died in ICU. During the supine CPAP decremental trial, all patients experienced an increase in RR and Ve. ΔEELI was < 40% in Group F and > 50% in Group S. In the prone trial, ΔEELI was > 50% in all patients, while RR decreased in Group S and remained unchanged in Group F. Conclusion: ΔEELI < 40% after a single PEEP de-escalation step in supine position seems to be a good predictor of poor recruitment and CPAP failure.

Time course of risk factors associated with mortality of 1260 critically ill patients with COVID-19 admitted to 24 Italian intensive care units.

PURPOSE: To evaluate the daily values and trends over time of relevant clinical, ventilatory and laboratory parameters during the intensive care unit (ICU) stay and their association with outcome in critically ill patients with coronavirus disease 19 (COVID-19). METHODS: In this retrospective-prospective multicentric study, we enrolled COVID-19 patients admitted to Italian ICUs from February 22 to May 31, 2020. Clinical data were daily recorded. The time course of 18 clinical parameters was evaluated by a polynomial maximum likelihood multilevel linear regression model, while a full joint modeling was fit to study the association with ICU outcome. RESULTS: 1260 consecutive critically ill patients with COVID-19 admitted in 24 ICUs were enrolled. 78% were male with a median age of 63 [55-69] years. At ICU admission, the median ratio of arterial oxygen partial pressure to fractional inspired oxygen (PaO2/FiO2) was 122 [89-175] mmHg. 79% of patients underwent invasive mechanical ventilation. The overall mortality was 34%. Both the daily values and trends of respiratory system compliance, PaO2/FiO2, driving pressure, arterial carbon dioxide partial pressure, creatinine, C-reactive protein, ferritin, neutrophil, neutrophil-lymphocyte ratio, and platelets were associated with survival, while for lactate, pH, bilirubin, lymphocyte, and urea only the daily values were associated with survival. The trends of PaO2/FiO2, respiratory system compliance, driving pressure, creatinine, ferritin, and C-reactive protein showed a higher association with survival compared to the daily values. CONCLUSION: Daily values or trends over time of parameters associated with acute organ dysfunction, acid-base derangement, coagulation impairment, or systemic inflammation were associated with patient survival.

Rationale for Polyclonal Intravenous Immunoglobulin Adjunctive Therapy in COVID-19 Patients: Report of a Structured Multidisciplinary Consensus.

INTRODUCTION: Adjunctive therapy with polyclonal intravenous immunoglobins (IVIg) is currently used for preventing or managing infections and sepsis, especially in immunocompromised patients. The pathobiology of COVID-19 and the mechanisms of action of Ig led to the consideration of this adjunctive therapy, including in patients with respiratory failure due to the SARS-CoV-2 infection. This manuscript reports the rationale, the available data and the results of a structured consensus on intravenous Ig therapy in patients with severe COVID-19. METHODS: A panel of multidisciplinary experts defined the clinical phenotypes of COVID-19 patients with severe respiratory failure and, after literature review, voted for the agreement on the rationale and the potential role of IVIg therapy for each phenotype. Due to the scarce evidence available, a modified RAND/UCLA appropriateness method was used. RESULTS: Three different phenotypes of COVID-19 patients with severe respiratory failure were identified: patients with an abrupt and dysregulated hyperinflammatory response (early phase), patients with suspected immune paralysis (late phase) and patients with sepsis due to a hospital-acquired superinfection (sepsis by bacterial superinfection). The rationale for intravenous Ig therapy in the early phase was considered uncertain whereas the panelists considered its use in the late phase and patients with sepsis/septic shock by bacterial superinfection appropriate. CONCLUSION: As with other immunotherapies, IVIg adjunctive therapy may have a potential role in the management of COVID-19 patients. The ongoing trials will clarify the appropriate target population and the true effectiveness.

Use of low-molecular weight heparin, transfusion and mortality in COVID-19 patients not requiring ventilation.

It is still debated whether prophylactic doses of low-molecular- weight heparin (LMWH) are always effective in preventing Venous Thromboembolism (VTE) and mortality in COVID-19. Furthermore, there is paucity of data for those patients not requiring ventilation. We explored mortality and the safety/efficacy profile of LMWH in a cohort of Italian patients with COVID-19 who did not undergo ventilation. From the initial cohort of 422 patients, 264 were enrolled. Most (n = 156, 87.7%) received standard LMWH prophylaxis during hospitalization, with no significant difference between medical wards and Intensive Care Unit (ICU). Major or not major but clinically relevant hemorrhages were recorded in 13 (4.9%) patients: twelve in those taking prophylactic LMWH and one in a patient taking oral anticoagulants (p: n.s.). Thirty-nine patients (14.8%) with median age 75 years. were transfused. Hemoglobin (Hb) at admission was significantly lower in transfused patients and Hb at admission inversely correlated with the number of red blood cells units transfused (p < 0.001). In-hospital mortality occurred in 76 (28.8%) patients, 46 (24.3%) of whom admitted to medical wards. Furthermore, Hb levels at admittance were significantly lower in fatalities (g/dl 12.3; IQR 2.4 vs. 13.3; IQR 2.8; Mann-Whitney U-test; p = 0.001). After the exclusion of patients treated by LMWH intermediate or therapeutic doses (n = 32), the logistic regression showed that prophylaxis significantly and independently reduced mortality (OR 0.31, 95% CI 0.13-0.85). Present data show that COVID-19 patients who do not require ventilation benefit from prophylactic doses of LMWH.

Prone position in intubated, mechanically ventilated patients with COVID-19: a multi-centric study of more than 1000 patients.

BACKGROUND: Limited data are available on the use of prone position in intubated, invasively ventilated patients with Coronavirus disease-19 (COVID-19). Aim of this study is to investigate the use and effect of prone position in this population during the first 2020 pandemic wave. METHODS: Retrospective, multicentre, national cohort study conducted between February 24 and June 14, 2020, in 24 Italian Intensive Care Units (ICU) on adult patients needing invasive mechanical ventilation for respiratory failure caused by COVID-19. Clinical data were collected on the day of ICU admission. Information regarding the use of prone position was collected daily. Follow-up for patient outcomes was performed on July 15, 2020. The respiratory effects of the first prone position were studied in a subset of 78 patients. Patients were classified as Oxygen Responders if the PaO2/FiO2 ratio increased ≥ 20 mmHg during prone position and as Carbon Dioxide Responders if the ventilatory ratio was reduced during prone position. RESULTS: Of 1057 included patients, mild, moderate and severe ARDS was present in 15, 50 and 35% of patients, respectively, and had a resulting mortality of 25, 33 and 41%. Prone position was applied in 61% of the patients. Patients placed prone had a more severe disease and died significantly more (45% vs. 33%, p < 0.001). Overall, prone position induced a significant increase in PaO2/FiO2 ratio, while no change in respiratory system compliance or ventilatory ratio was observed. Seventy-eight % of the subset of 78 patients were Oxygen Responders. Non-Responders had a more severe respiratory failure and died more often in the ICU (65% vs. 38%, p = 0.047). Forty-seven % of patients were defined as Carbon Dioxide Responders. These patients were older and had more comorbidities; however, no difference in terms of ICU mortality was observed (51% vs. 37%, p = 0.189 for Carbon Dioxide Responders and Non-Responders, respectively). CONCLUSIONS: During the COVID-19 pandemic, prone position has been widely adopted to treat mechanically ventilated patients with respiratory failure. The majority of patients improved their oxygenation during prone position, most likely due to a better ventilation perfusion matching. TRIAL REGISTRATION: clinicaltrials.gov number: NCT04388670.

Mortality and Transfusion Requirements in COVID-19 Hospitalized Italian Patients According to Severity of the Disease.

There is paucity of data on the transfusion need and its impact on the overall mortality in patients with COVID-19. We explored mortality in hospitalized patients with COVID-19 who required transfusions. Information on clinical variables and in-hospital mortality were obtained from medical records of 422 patients admitted to medical wards or the Intensive Care Unit (ICU). In-hospital mortality occurred in 147 (34.8%) patients, 94 (63.9%) of whom were admitted to the ICU. The median fatalities age was 77 years (IQR 14). Overall, 100 patients (60 males) received transfusion during hospitalization. The overall mortality was significantly and independently associated with age, ICU admission, Chronic Kidney Disease (CKD), and the number of transfused Red Blood Cell (RBC) units. Specifically, CKD was associated with mortality in patients admitted to medical wards, whereas the number of transfused RBC units predicted mortality in those admitted to the ICU. Transfusion strongly interacted with the admission to ICU (OR: 9.9; 95% CI: 2.5-40.0). In patients with COVID-19, age is one of the strongest risk factors in predicting mortality independently of the disease's severity. CKD confers a higher risk of mortality in patients admitted to medical wards. In those admitted to the ICU, the more RBC units are transfused, the more mortality increases.

SARS-CoV-2 pneumonia succesfully treated with cpap and cycles of tripod position: a case report.

BACKGROUND: Pneumonia induced by 2019 Coronavirus (COVID-19) is characterized by hypoxemic respiratory failure that may present with a broad spectrum of clinical phenotypes. At the beginning, patients may have normal lung compliance and be responsive to noninvasive ventilatory support, such as CPAP. However, the transition to more severe respiratory failure - Severe Acute Respiratory Syndrome (SARS-CoV-2), necessitating invasive ventilation is often abrupt and characterized by a severe V/Q mismatch that require cycles of prone positioning. The aim of this case is to report the effect on gas exchange, respiratory mechanics and hemodynamics of tripod (or orthopneic sitting position) used as an alternative to prone position in a patient with mild SARS-CoV-2 pneumonia ventilated with helmet CPAP. CASE PRESENTATION: A 77-year-old awake and collaborating male patient with mild SARS-CoV-2 pneumonia and ventilated with Helmet CPAP, showed sudden worsening of gas exchange without dyspnea. After an unsuccessful attempt of prone positioning, we alternated three-hours cycles of semi-recumbent and tripod position, still keeping him in CPAP. Arterial blood gases (PaO2/FiO2, PaO2, SaO2, PaCO2 and A/a gradient), respiratory (VE, VT, RR) and hemodynamic parameters (HR, MAP) were collected in the supine and tripod position. Cycles of tripod position were continued for 3 days. The patient had a clinically important improvement in arterial blood gases and respiratory parameters, with stable hemodynamic and was successfully weaned and discharged to ward 10 days after pneumonia onset. CONCLUSIONS: Tripod position during Helmet CPAP can be applied safely in patients with mild SARS-CoV-2 pneumonia, with improvement of oxygenation and V/Q matching, thus reducing the need for intubation.

SARS-COV-2 Serological Profile in Healthcare Professionals of a Southern Italy Hospital.

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is the first coronavirus that has caused a pandemic. Assessing the prevalence of anti-SARS-CoV-2 in healthcare worker groups offers a unique opportunity to study the correlation between seroconversion and immunization because of their occupational exposure and a higher risk of contagion. The study enrolled 3242 asymptomatic employees of "Policlinico Riuniti", Foggia. After the first screening, we collected sequential serum samples for up to 23 weeks from the same subjects. In order to perform a longitudinal follow-up study and get information about the titration of IgG levels, we analyzed data from subjects (33) with at least two consecutive serological IgG-positive tests; 62 (1.9%; 95% CI: 1.4-2.3) tested positive for at least one anti-SARS-CoV-2 antibody. The seroprevalence was lower in the high-risk group 1.4% (6/428; 95% CI: 0.5-2.6) vs. the intermediate-risk group 2.0% (55/2736; 95% CI: 1.5-2.5). Overall, within eight weeks, we detected a mean reduction of -17% in IgG levels. Our data suggest a reduction of about 9.27 AU/mL every week (R2 = 0.35, p = 0.0003). This study revealed the prevalence of SARS-CoV-2 antibodies among Foggia's hospital healthcare staff (1.9%). Moreover, the IgG level reduction suggests that the serological response fades fast in asymptomatic infections.