Session Type: 1-hour Mini Oral Flash
Session Title: 1-hour Mini Oral Flash
Authors(s): M.P. Freire (1), M. Rinaldi (2), D.R.B. Terrabuio (3), D.P.N. Carlotti (4), Z. Pasquini (2), N.N. Nunes (5), G.T. Lemos (5), A. Maccaro (2), A. Sinischalchi (6), M. Cescon (7), L.A.C. D´albuquerque (8), M.C. Morelli (6), A.T.W. Song (9), E. Abdala (5), P. Viale (10), M. Giannella (10)
Authors Affiliations(s): (1) Working Committee for Hospital Epidemiology and Infection Control, University of São Paulo School of Medicine Hospital das Clínicas, Brazil, (2) Infectious Diseases Unit, Department of Medical and Surgical Sciences, Policlinico Sant'Orsola Malpighi, University of Bologna, Italy, (3) Division of Clinical Gastroenterology and Hepatology, Hospital das Clínicas, Department of Gastroenterology of University of São Paulo School of Medicine, Brazil, (4) Institute of Mathematics and Statistics, University of São Paulo, Brazil, (5) Department of Infectious Diseases, University of São Paulo School of Medicine Hospital das Clínicas, Brazil, (6) Department of General Surgery and Tran splantation, University of Bologna Sant’Orsola - Malpighi Hospital, Italy, (7) Department of Medical and Surgical Sciences, University of Bologn, Italy, (8) Division of Liver and Gastrointestinal Transplant, Hospital das Clínicas, Department of Gastroenterology, University of São Paulo School of Medicine, Brazil, (9) Department of Gastroenterology, Hospital das Clinicas HCFMUSP, Faculdade de Medicina, Universidade de Sao Paulo, Brazil, (10) Infectious Diseases Unit, IRCCS Azienda Ospedaliero-Universitaria di Bologna, Policlinico di Sant’Orsola, Italy
Background:
Carbapenem-resistant Enterobacteriaceae (CRE) colonization at liver transplantation (LT) increases the risk of CRE infection after LT which impacts on recipients’ survival. Generally, colonization status becomes evident only in the immediate post-transplant period. Thus predictive models can be useful to guide antibiotic prophylaxis in endemic centers. The aim of this study was to identify risk factors for CRE colonization at LT in order to build a predictive model.
Methods:Retrospective multicenter study including consecutive adult patients underwent LT, from 2010 to 2019, at two large teaching hospitals in Brazil and Italy. We excluded patients with missing data and those who had CRE infection within 90 days before LT. CRE screening was performed in all patients on the day of LT. Exposure variables were considered within 90 days before LT and included cirrhosis complications, underlying disease, time on waiting list, MELD and CLIF-SOFA scores, antibiotic use, ICU and hospital stay, and previous MDR non-CRE infections. A model of artificial intelligence was trained to detect the probability of a patient being colonized with CRE at LT. We used a technique to oversample the training dataset to balance the classes. This technique was not used in the test dataset. After training, the algorithm was calculated by logistic regression using LASSO for feature selection.
Results:A total of 905 patients were analyzed. Median age was 52 years (range 15 to 73), the most common cause of end-stage liver diseases was hepatitis (311 – 34.4%). CRE colonization at LT was found in 84 (9.3%) patients. The one-year survival after LT was 83.6%, and patients colonized with CRE at LT had a higher risk of one-year death (p <0.001, HR 2.09). Variables most correlated with pre-LT CRE colonization were antibiotic use, carbapenem use, number of bacterial infections before LT, infection by non-CRE multidrug-resistant bacteria, prophylaxis for spontaneous bacterial peritonitis, acute-on-chronic liver failure, ICU stay before LT and gastrointestinal bleeding (Graphic 1). The proposed algorithm had a sensitivity of 78% and a specificity of 82%.
Conclusions:We created a model able to predict CRE colonization at LT based on easy to obtain features that could guide antibiotic prophylaxis.
Session Type: 1-hour Mini Oral Flash
Session Title: 1-hour Mini Oral Flash
Authors(s): M.P. Freire (1), M. Rinaldi (2), D.R.B. Terrabuio (3), D.P.N. Carlotti (4), Z. Pasquini (2), N.N. Nunes (5), G.T. Lemos (5), A. Maccaro (2), A. Sinischalchi (6), M. Cescon (7), L.A.C. D´albuquerque (8), M.C. Morelli (6), A.T.W. Song (9), E. Abdala (5), P. Viale (10), M. Giannella (10)
Authors Affiliations(s): (1) Working Committee for Hospital Epidemiology and Infection Control, University of São Paulo School of Medicine Hospital das Clínicas, Brazil, (2) Infectious Diseases Unit, Department of Medical and Surgical Sciences, Policlinico Sant'Orsola Malpighi, University of Bologna, Italy, (3) Division of Clinical Gastroenterology and Hepatology, Hospital das Clínicas, Department of Gastroenterology of University of São Paulo School of Medicine, Brazil, (4) Institute of Mathematics and Statistics, University of São Paulo, Brazil, (5) Department of Infectious Diseases, University of São Paulo School of Medicine Hospital das Clínicas, Brazil, (6) Department of General Surgery and Tran splantation, University of Bologna Sant’Orsola - Malpighi Hospital, Italy, (7) Department of Medical and Surgical Sciences, University of Bologn, Italy, (8) Division of Liver and Gastrointestinal Transplant, Hospital das Clínicas, Department of Gastroenterology, University of São Paulo School of Medicine, Brazil, (9) Department of Gastroenterology, Hospital das Clinicas HCFMUSP, Faculdade de Medicina, Universidade de Sao Paulo, Brazil, (10) Infectious Diseases Unit, IRCCS Azienda Ospedaliero-Universitaria di Bologna, Policlinico di Sant’Orsola, Italy
Background:
Carbapenem-resistant Enterobacteriaceae (CRE) colonization at liver transplantation (LT) increases the risk of CRE infection after LT which impacts on recipients’ survival. Generally, colonization status becomes evident only in the immediate post-transplant period. Thus predictive models can be useful to guide antibiotic prophylaxis in endemic centers. The aim of this study was to identify risk factors for CRE colonization at LT in order to build a predictive model.
Methods:Retrospective multicenter study including consecutive adult patients underwent LT, from 2010 to 2019, at two large teaching hospitals in Brazil and Italy. We excluded patients with missing data and those who had CRE infection within 90 days before LT. CRE screening was performed in all patients on the day of LT. Exposure variables were considered within 90 days before LT and included cirrhosis complications, underlying disease, time on waiting list, MELD and CLIF-SOFA scores, antibiotic use, ICU and hospital stay, and previous MDR non-CRE infections. A model of artificial intelligence was trained to detect the probability of a patient being colonized with CRE at LT. We used a technique to oversample the training dataset to balance the classes. This technique was not used in the test dataset. After training, the algorithm was calculated by logistic regression using LASSO for feature selection.
Results:A total of 905 patients were analyzed. Median age was 52 years (range 15 to 73), the most common cause of end-stage liver diseases was hepatitis (311 – 34.4%). CRE colonization at LT was found in 84 (9.3%) patients. The one-year survival after LT was 83.6%, and patients colonized with CRE at LT had a higher risk of one-year death (p <0.001, HR 2.09). Variables most correlated with pre-LT CRE colonization were antibiotic use, carbapenem use, number of bacterial infections before LT, infection by non-CRE multidrug-resistant bacteria, prophylaxis for spontaneous bacterial peritonitis, acute-on-chronic liver failure, ICU stay before LT and gastrointestinal bleeding (Graphic 1). The proposed algorithm had a sensitivity of 78% and a specificity of 82%.
Conclusions:We created a model able to predict CRE colonization at LT based on easy to obtain features that could guide antibiotic prophylaxis.
Session Type: 1-hour Mini Oral Flash
Session Title: 1-hour Mini Oral Flash
Authors(s): M.P. Freire (1), M. Rinaldi (2), D.R.B. Terrabuio (3), D.P.N. Carlotti (4), Z. Pasquini (2), N.N. Nunes (5), G.T. Lemos (5), A. Maccaro (2), A. Sinischalchi (6), M. Cescon (7), L.A.C. D´albuquerque (8), M.C. Morelli (6), A.T.W. Song (9), E. Abdala (5), P. Viale (10), M. Giannella (10)
Authors Affiliations(s): (1) Working Committee for Hospital Epidemiology and Infection Control, University of São Paulo School of Medicine Hospital das Clínicas, Brazil, (2) Infectious Diseases Unit, Department of Medical and Surgical Sciences, Policlinico Sant'Orsola Malpighi, University of Bologna, Italy, (3) Division of Clinical Gastroenterology and Hepatology, Hospital das Clínicas, Department of Gastroenterology of University of São Paulo School of Medicine, Brazil, (4) Institute of Mathematics and Statistics, University of São Paulo, Brazil, (5) Department of Infectious Diseases, University of São Paulo School of Medicine Hospital das Clínicas, Brazil, (6) Department of General Surgery and Tran splantation, University of Bologna Sant’Orsola - Malpighi Hospital, Italy, (7) Department of Medical and Surgical Sciences, University of Bologn, Italy, (8) Division of Liver and Gastrointestinal Transplant, Hospital das Clínicas, Department of Gastroenterology, University of São Paulo School of Medicine, Brazil, (9) Department of Gastroenterology, Hospital das Clinicas HCFMUSP, Faculdade de Medicina, Universidade de Sao Paulo, Brazil, (10) Infectious Diseases Unit, IRCCS Azienda Ospedaliero-Universitaria di Bologna, Policlinico di Sant’Orsola, Italy
Background:
Carbapenem-resistant Enterobacteriaceae (CRE) colonization at liver transplantation (LT) increases the risk of CRE infection after LT which impacts on recipients’ survival. Generally, colonization status becomes evident only in the immediate post-transplant period. Thus predictive models can be useful to guide antibiotic prophylaxis in endemic centers. The aim of this study was to identify risk factors for CRE colonization at LT in order to build a predictive model.
Methods:Retrospective multicenter study including consecutive adult patients underwent LT, from 2010 to 2019, at two large teaching hospitals in Brazil and Italy. We excluded patients with missing data and those who had CRE infection within 90 days before LT. CRE screening was performed in all patients on the day of LT. Exposure variables were considered within 90 days before LT and included cirrhosis complications, underlying disease, time on waiting list, MELD and CLIF-SOFA scores, antibiotic use, ICU and hospital stay, and previous MDR non-CRE infections. A model of artificial intelligence was trained to detect the probability of a patient being colonized with CRE at LT. We used a technique to oversample the training dataset to balance the classes. This technique was not used in the test dataset. After training, the algorithm was calculated by logistic regression using LASSO for feature selection.
Results:A total of 905 patients were analyzed. Median age was 52 years (range 15 to 73), the most common cause of end-stage liver diseases was hepatitis (311 – 34.4%). CRE colonization at LT was found in 84 (9.3%) patients. The one-year survival after LT was 83.6%, and patients colonized with CRE at LT had a higher risk of one-year death (p <0.001, HR 2.09). Variables most correlated with pre-LT CRE colonization were antibiotic use, carbapenem use, number of bacterial infections before LT, infection by non-CRE multidrug-resistant bacteria, prophylaxis for spontaneous bacterial peritonitis, acute-on-chronic liver failure, ICU stay before LT and gastrointestinal bleeding (Graphic 1). The proposed algorithm had a sensitivity of 78% and a specificity of 82%.
Conclusions:We created a model able to predict CRE colonization at LT based on easy to obtain features that could guide antibiotic prophylaxis.
Session Type: 1-hour Mini Oral Flash
Session Title: 1-hour Mini Oral Flash
Authors(s): M.P. Freire (1), M. Rinaldi (2), D.R.B. Terrabuio (3), D.P.N. Carlotti (4), Z. Pasquini (2), N.N. Nunes (5), G.T. Lemos (5), A. Maccaro (2), A. Sinischalchi (6), M. Cescon (7), L.A.C. D´albuquerque (8), M.C. Morelli (6), A.T.W. Song (9), E. Abdala (5), P. Viale (10), M. Giannella (10)
Authors Affiliations(s): (1) Working Committee for Hospital Epidemiology and Infection Control, University of São Paulo School of Medicine Hospital das Clínicas, Brazil, (2) Infectious Diseases Unit, Department of Medical and Surgical Sciences, Policlinico Sant'Orsola Malpighi, University of Bologna, Italy, (3) Division of Clinical Gastroenterology and Hepatology, Hospital das Clínicas, Department of Gastroenterology of University of São Paulo School of Medicine, Brazil, (4) Institute of Mathematics and Statistics, University of São Paulo, Brazil, (5) Department of Infectious Diseases, University of São Paulo School of Medicine Hospital das Clínicas, Brazil, (6) Department of General Surgery and Tran splantation, University of Bologna Sant’Orsola - Malpighi Hospital, Italy, (7) Department of Medical and Surgical Sciences, University of Bologn, Italy, (8) Division of Liver and Gastrointestinal Transplant, Hospital das Clínicas, Department of Gastroenterology, University of São Paulo School of Medicine, Brazil, (9) Department of Gastroenterology, Hospital das Clinicas HCFMUSP, Faculdade de Medicina, Universidade de Sao Paulo, Brazil, (10) Infectious Diseases Unit, IRCCS Azienda Ospedaliero-Universitaria di Bologna, Policlinico di Sant’Orsola, Italy
Background:
Carbapenem-resistant Enterobacteriaceae (CRE) colonization at liver transplantation (LT) increases the risk of CRE infection after LT which impacts on recipients’ survival. Generally, colonization status becomes evident only in the immediate post-transplant period. Thus predictive models can be useful to guide antibiotic prophylaxis in endemic centers. The aim of this study was to identify risk factors for CRE colonization at LT in order to build a predictive model.
Methods:Retrospective multicenter study including consecutive adult patients underwent LT, from 2010 to 2019, at two large teaching hospitals in Brazil and Italy. We excluded patients with missing data and those who had CRE infection within 90 days before LT. CRE screening was performed in all patients on the day of LT. Exposure variables were considered within 90 days before LT and included cirrhosis complications, underlying disease, time on waiting list, MELD and CLIF-SOFA scores, antibiotic use, ICU and hospital stay, and previous MDR non-CRE infections. A model of artificial intelligence was trained to detect the probability of a patient being colonized with CRE at LT. We used a technique to oversample the training dataset to balance the classes. This technique was not used in the test dataset. After training, the algorithm was calculated by logistic regression using LASSO for feature selection.
Results:A total of 905 patients were analyzed. Median age was 52 years (range 15 to 73), the most common cause of end-stage liver diseases was hepatitis (311 – 34.4%). CRE colonization at LT was found in 84 (9.3%) patients. The one-year survival after LT was 83.6%, and patients colonized with CRE at LT had a higher risk of one-year death (p <0.001, HR 2.09). Variables most correlated with pre-LT CRE colonization were antibiotic use, carbapenem use, number of bacterial infections before LT, infection by non-CRE multidrug-resistant bacteria, prophylaxis for spontaneous bacterial peritonitis, acute-on-chronic liver failure, ICU stay before LT and gastrointestinal bleeding (Graphic 1). The proposed algorithm had a sensitivity of 78% and a specificity of 82%.
Conclusions:We created a model able to predict CRE colonization at LT based on easy to obtain features that could guide antibiotic prophylaxis.