Burden and Management of Multi-Drug Resistant Organism Infections in Solid Organ Transplant Recipients Across the World: A Narrative Review.

Solid organ transplant (SOT) recipients are particularly susceptible to infections caused by multidrug-resistant organisms (MDRO) and are often the first to be affected by an emerging resistant pathogen. Unfortunately, their prevalence and impact on morbidity and mortality according to the type of graft is not systematically reported from high-as well as from low and middle-income countries (HIC and LMIC). Thus, epidemiology on MDRO in SOT recipients could be subjected to reporting bias. In addition, screening practices and diagnostic resources may vary between countries, as well as the availability of new drugs. In this review, we aimed to depict the burden of main Gram-negative MDRO in SOT patients across HIC and LMIC and to provide an overview of current diagnostic and therapeutic resources.

Survey on the approach to antibiotic prophylaxis in liver and kidney transplant recipients colonized with "difficult to treat" Gram-negative bacteria.

BACKGROUND: Performance of active screening for multidrug-resistant Gram-negative bacteria (MDR-GNB) and administration of targeted antibiotic prophylaxis (TAP) in colonized patients undergoing liver (LT) and/or kidney transplantation (KT) are controversial issues. METHODS: Self-administered electronic cross-sectional survey disseminated from January to February 2022. Questionnaire consisted of four parts: hospital/transplant program characteristics, standard screening and antibiotic prophylaxis, clinical vignettes asking for TAP in patients undergoing LT and KT with prior infection/colonization with four different MDR-GNB (extended-spectrum cephalosporin-resistant Enterobacterales [ESCR-E], carbapenem-resistant Enterobacterales [CRE], multidrug-resistant Pseudomonas aeruginosa [MDR-Pa], and carbapenem-resistant Acinetobacter baumannii [CRAb]). RESULTS: Fifty-five respondents participated from 14 countries, mostly infectious disease specialists (69%) with active transplant programs (>100 procedures/year for 34.5% KT and 23.6% LT), and heterogeneous local MDR-GNB prevalence from <15% (30.9%), 15%-30% (43.6%) to >30% (16.4%). The frequency of screening for ESCR-E, CRE, MDR-Pa, and CRAb was 22%, 54%, 17%, and 24% for LT, respectively, and 18%, 36%, 16%, and 11% for KT. Screening time-points were mainly at transplantation 100%, only one-third following transplantation. Screening was always based on rectal swab cultures (100%); multi-site sampling was reported in 40% of KT and 35% of LT. In LT clinical cases, 84%, 58%, 84%, and 40% of respondents reported TAP for prior infection/colonization with ESCR-E, CRE, MDR-Pa, and CRAb, respectively. In KT clinical cases, 55%, 39%, 87%, and 42% of respondents reported TAP use for prior infection/colonization with ESCR-E, CRE, MDR-Pa, and CRAb, respectively. CONCLUSION: There is a large heterogeneity in screening and management of MDR-GNB carriage in LT and KT.

Pre-transplant multidrug-resistant infections in liver transplant recipients-epidemiology and impact on transplantation outcome.

BACKGROUND: Cirrhotic patients are highly exposed to healthcare services and antibiotics. Although pre-liver transplantation (LT) infections are directly related to the worsening of liver function, the impact of these infections on LT outcomes is still unclear. This study aimed to identify the effect of multidrug-resistant microorganism (MDRO) infections before LT on survival after LT. METHODS: Retrospective study that included patients who underwent LT between 2010 and 2019. Variables analyzed were related to patients' comorbidities, underlying diseases, time on the waiting list, antibiotic use, LT surgery, and occurrences post-LT. Multivariate analyses were performed using logistic regression, and Cox regression for survival analysis. RESULTS: A total of 865 patients were included; 351 infections were identified in 259 (30%) patients, of whom 75 (29%) had ≥1 pre-LT MDRO infection. The most common infection was spontaneous bacterial peritonitis (34%). The agent was identified in 249(71%), 53(15%) were polymicrobial. The most common microorganism was Klebsiella pneumoniae (18%); the most common MDRO was ESBL-producing Enterobacterales (16%), and carbapenem-resistant (CR) Enterobacterales (10%). Factors associated with MDRO infections before LT were previous use of therapeutic cephalosporin (p = .001) and fluoroquinolone (p = .001), SBP prophylaxis (p = .03), ACLF before LT (p = .03), and days of hospital stay pre-LT (p < .001); HCC diagnosis was protective (p = .01). Factors associated with 90-day mortality after LT were higher MELD on inclusion to the waiting list (p = .02), pre-LT MDRO infection (p = .04), dialysis after LT (p < .001), prolonged duration of LT surgery (p < .001), post-LT CR-Gram-negative bacteria infection (p < .001), and early retransplantation (p = .004). CONCLUSION: MDRO infections before LT have an important impact on survival after LT.

Meet in the middle: Could cell mediated-immunity assays be the answer for ideal Cytomegalovirus prophylaxis after lung transplantation? Observational study from a single center with intermittent antiviral prophylaxis.

BACKGROUND: Cytomegalovirus (CMV) can cause tissue-invasive disease and indirect effects after lung transplantation (LTx) such as acute rejection episodes and chronic lung allograft dysfunction. Monitoring CMV-specific cell immune recovery (CMV-CIR) after LTx can individualize CMV risks and establish better antiviral approach. This study evaluated the dynamics of CMV-CIR, using QuantiFERON-CMV assay (Qiagen Group), in the first year after LTx. METHODS: Prospective observational cohort study included lung transplant recipients from December/2015 to December/2016. Universal antiviral prophylaxis with intravenous ganciclovir 5 mg/kg/day 3 days/week for 3 months was given for CMV-seropositive recipients (R+) and only CMV-seropositive donor and negative recipient (D+/R-) received a 6-month-prophylaxis with ganciclovir and valganciclovir, on alternate days, in the first 3 months and then, 3 more months of valganciclovir. QuantiFERON-CMV was measured at the same time points of surveillance bronchoscopies. CMV infection was defined as any DNAemia detected and CMV disease with proven biopsy or antigenemia pp65 above 10 cells/300.000 neutrophils. RESULTS: Thirty-eight patients were included. On days 45, 90, and 365 days post-LTx, 60%, 72%, and 81% QuantiFERON-CMV were reactive, respectively. Eleven patients (28.9%) presented CMV-disease and 27 DNAemia/CMV infections. Reactive tests were able to predict CMV disease only at 90 days after LTx (p = .027) but failed on DNAemia/CMV infection (p = .148). Daily prophylaxis, for D+/R- patients (13.2%), remained as an independently associated factor for not achieving reactive QuantiFERON-CMV (adjusted OR .27, 95%CI .12-.60, p = .02). CONCLUSION: QuantiFERON-CMV may be another diagnostic tool to help stratify CMV-disease risk and individualized antiviral prophylaxis after LTx.

Genomic analysis and antimicrobial activity of ß-lactam/ß-lactamase inhibitors and other agents against KPC-producing Klebsiella pneumoniae clinical isolates from Brazilian hospitals.

Carbapenem-resistant Klebsiella pneumoniae (CRKP) are highly disseminated worldwide, and isolates co-resistant to other antimicrobial agents pose a threat to effective antimicrobial therapy. Therefore, evaluation of novel antimicrobial drugs is needed to identify potential treatments with better outcomes. We evaluated the in vitro activity of novel antimicrobial drugs/combinations against 97 KPC-producing Klebsiella pneumoniae isolates recovered from different hospitals in Brazil during 2021-2022. Clonality, resistance and virulence genes were detected by whole-genome sequencing. The majority of the isolates (54.6%) were classified as extensively drug resistant or multidrug resistant (44.3%); one isolate showed a pandrug resistance phenotype. The most active antimicrobial agents were meropenem-vaborbactam, cefiderocol, and ceftazidime-avibactam, with sensitivities higher than 90%; resistance to ceftazidime-avibactam was associated with KPC-33 or KPC-44 variants. Colistin and polymyxin B were active against 58.6% of the isolates. The 97 isolates were distributed into 17 different sequence types, with a predominance of ST11 (37.4%). Although high in vitro susceptibility rates were detected for meropenem-vaborbactam and cefiderocol, only ceftazidime-avibactam is currently available in Brazil. Our findings showed limited susceptibility to antimicrobial drugs employed for infection treatment of carbapenem-resistant K. pneumoniae, underscoring the urgent need for stringent policies for antimicrobial stewardship to preserve the activity of such drugs.

Genomics and Antimicrobial Susceptibility of Clinical Pseudomonas aeruginosa Isolates from Hospitals in Brazil.

Pseudomonas aeruginosa, an opportunistic pathogen causing infections in immunocompromised patients, usually shows pronounced antimicrobial resistance. In recent years, the frequency of carbapenemases in P. aeruginosa has decreased, which allows use of new beta-lactams/combinations in antimicrobial therapy. Therefore, the in vitro evaluation of these drugs in contemporary isolates is warranted. We evaluated the antimicrobial susceptibility and genomic aspects of 119 clinical P. aeruginosa isolates from 24 different hospitals in Brazil in 2021-2022. Identification was performed via MALDI-TOF-MS, and antimicrobial susceptibility was identified through broth microdilution, gradient tests, or disk diffusion. Whole-genome sequencing was carried out using NextSeq equipment. The most active drug was cefiderocol (100%), followed by ceftazidime-avibactam (94.1%), ceftolozane-tazobactam (92.4%), and imipenem-relebactam (81.5%). Imipenem susceptibility was detected in 59 isolates (49.6%), and the most active aminoglycoside was tobramycin, to which 99 (83.2%) isolates were susceptible. Seventy-one different sequence types (STs) were detected, including twelve new STs described herein. The acquired resistance genes blaCTX-M-2 and blaKPC-2 were identified in ten (8.4%) and two (1.7%) isolates, respectively. Several virulence genes (exoSTUY, toxA, aprA, lasA/B, plcH) were also identified. We found that new antimicrobials are effective against the diverse P. aeruginosa population that has been circulating in Brazilian hospitals in recent years.

Transmission of Carbapenem-Resistant Enterobacterales in an Overcrowded Emergency Department: Controlling the Spread to the Hospital.

BACKGROUND: Overcrowded emergency departments (EDs) may increase the risk of carbapenem-resistant Enterobacterales (CRE) transmission. METHODS: We conducted a quasi-experimental study divided into 2 phases (baseline and intervention) to investigate the impact of an intervention on the acquisition rate and identify risk factors for CRE colonization in an ED of a tertiary academic hospital in Brazil. In both phases, we did universal screening with rapid molecular test (blaKPC, blaNDM, blaOXA48, blaOXA23, and blaIMP) and culture. At baseline, both screening test results were not reported, and patients were put under contact precautions (CP) based on previous colonization or infection by multidrug-resistant organisms. During the intervention, all patients hospitalized in the ED were placed in empiric CP and the result of CRE screening was reported; if negative, patients were released from CP. Patients were rescreened if they stayed >7 days in the ED or were transferred to an intensive care unit. RESULTS: A total of 845 patients were included: 342 in baseline and 503 in intervention. Colonization at admission was 3.4% by culture and molecular test. Acquisition rates during ED stay dropped from 4.6% (11/241) to 1% (5/416) during intervention (P = .06). The aggregated antimicrobial use in the ED decreased from phase 1 to phase 2 (804 defined daily doses [DDD]/1000 patients to 394 DDD/1000 patients, respectively). Length of stay >2 days in the ED was a risk factor for CRE acquisition (adjusted odds ratio, 4.58 [95% confidence interval, 1.44-14.58]; P = .01). CONCLUSIONS: Early empiric CP and rapid identification of CRE-colonized patients reduce cross-transmission in ED. Nevertheless, staying >2 days in ED compromised efforts.

Crude Mortality Associated With the Empirical Use of Polymyxins in Septic Patients in a Setting of High Prevalence of Carbapenem-Resistant Gram-negative Bacteria: Retrospective Analysis of a Cohort.

BACKGROUND: Our aim in this retrospective cohort study was to assess the impact on mortality of the empirical use of polymyxin as therapy for carbapenem-resistant gram-negative bacteria (CR-GNB) in septic patients. The study was performed at a tertiary academic hospital in Brazil, from January 2018 to January 2020, the pre-coronavirus disease 2019 period. METHODS: We included 203 patients with suspected sepsis. The first doses of antibiotics were prescribed from a "sepsis antibiotic kit", which contained a selection of drugs, including polymyxin, with no preapproval policy. We developed a logistic regression model to assess risk factors associated with 14-day crude mortality. Propensity score for polymyxin was used to control biases. RESULTS: Seventy (34%) of 203 patients had infections with at least 1 multidrug-resistant organism isolated from any clinical culture. Polymyxins in monotherapy or in combination therapy were prescribed to 140 of the 203 (69%) patients. The overall 14-day mortality rate was 30%. The 14-day crude mortality was associated with age (adjusted odds ratio [aOR], 1.03; 95% confidence interval [CI], 1.01-1.05; P = .01), SOFA (sepsis-related organ failure assessment) score value (aOR, 1.2; 95% CI, 1.09-1.32; P < .001), CR-GNB infection (aOR, 3.94; 95% CI, 1.53-10.14; P = .005), and time between suspected sepsis and antibiotic administration (aOR, 0.73; 95% CI, .65-.83; P < .001). The empirical use of polymyxins was not associated with decreased crude mortality (aOR, 0.71; 95% CI, .29-1.71; P = .44). CONCLUSIONS: Empirical use of polymyxin for septic patients in a setting with high CR-GNB prevalence was not associated with decreased crude mortality.

Impact of COVID-19 on healthcare-associated infections: Antimicrobial consumption does not follow antimicrobial resistance.

BACKGROUND: This study aimed to analyze the Healthcare-Associated Infections (HAI) rates and antimicrobial consumption in Intensive Care Units (ICU) in São Paulo city during the COVID-19 pandemic and compare them with the pre-pandemic period. METHODS: This cohort included all hospitals that reported HAI rates (Central-Line-Associated Bloodstream Infection ‒ CLABSI and Ventilator-Associated Pneumonia ‒ VAP), the proportion of microorganisms that caused CLABSI, the proportion of resistant microorganisms, and antimicrobial consumption from January 2017 ‒ December 2020. Hospitals were stratified by the number of beds, Central Venous Catheter (CVC) utilization rate, Mechanical-Ventilation (MV) utilization rate, and type of funding. Statistical analyses were based on time-series plots and regression models. RESULTS: 220 ICUs were included. The authors observed an abrupt increase in CLABSI rates after the pandemic onset. High CLABSI rates during the pandemic were associated with hospital size, funding (public and non-profit private), and low CVC use (≤ 50%). An increase in VAP rates was associated with public hospitals, and high MV use (> 35%). The susceptibility profile of microorganisms did not differ from that of the pre-pandemic period. polymyxin, glycopeptides, and antifungal use increased, especially in COVID-19 ICUs. CONCLUSIONS: HAI increased during COVID-19. The microorganisms' susceptibility profile did not change with the pandemic, but the authors observed a disproportionate increase in large-spectrum antimicrobial drug use.

A surveillance program for long-term central venous access-associated infections in outpatient chemotherapy services.

OBJECTIVE: In this study, we described the first results of a surveillance system for infections associated with long-term central venous catheters (LT-CVC) in patients under outpatient chemotherapy. DESIGN: This was a multicentric, prospective study. SETTING: Outpatient chemotherapy services. PARTICIPANTS: The study included 8 referral cancer centers in the State of São Paulo. INTERVENTION: These services were invited to participate in a newly created surveillance program for patients under chemotherapy. Several meetings were convened to share previous experiences on LT-CVC infection surveillance and to define the surveillance method. Once the program was implemented, all bloodstream infection (LT-CVC BSIs), tunnel infection, and exit-site infections associated with LT-CVC were reported. Data from January to May 2021 were analyzed. The median monthly number of chemotherapy sessions per clinic was 925 (IQR, 270-5,855). We used Poisson regression to analyze the association of rates with the characteristics of the services. RESULTS: In total, 107 LT-CVC infections were reported, of which 95% were BSIs, mostly associated with totally implantable devices (76%). Infections occurred a median of 4 days after the last catheter manipulation and 116 after the LT-CVC insertion. Also, 102 microorganisms were isolated from LT-CVC BSIs; the most common pathogen was Staphylococcus epidermidis, at 22%. Moreover, 44 infections (44%) fulfilled the criteria for CVC-related LT-CVC BSI and 27 infections (27%) met the criteria for mucosal barrier injury. The 1-year cumulative LT-CVC BSI rate was 1.94 per 1,000 CVC days of use. The rates were higher in public hospitals (IRR, 6.00; P < .001) and in hospitals that already had in place surveillance for LT-CVC infections (IRR, 2.01; P < .01). CONCLUSION: Our study describes an applicable surveillance method for infections in cancer outpatients using LT-CVC.

Efficacy of ceftazidime-avibactam in solid organ transplant recipients with bloodstream infections caused by carbapenemase-producing Klebsiella pneumoniae.

We aimed to compare the efficacy of ceftazidime-avibactam (CAZ-AVI) versus the best available therapy (BAT) in solid organ transplant (SOT) recipients with bloodstream infection caused by carbapenemase-producing Klebsiella pneumoniae (CPKP-BSI). A retrospective (2016-2021) observational cohort study was performed in 14 INCREMENT-SOT centers (ClinicalTrials.gov identifier: NCT02852902; Impact of Specific Antimicrobials and MIC Values on the Outcome of Bloodstream Infections Due to ESBL- or Carbapenemase-producing Enterobacterales in Solid Organ Transplantation: an Observational Multinational Study). Outcomes were 14-day and 30-day clinical success (complete resolution of attributable manifestations, adequate source control, and negative follow-up blood cultures) and 30-day all-cause mortality. Multivariable logistic and Cox regression analyses adjusted for the propensity score to receive CAZ-AVI were constructed. Among 210 SOT recipients with CPKP-BSI, 149 received active primary therapy with CAZ-AVI (66/149) or BAT (83/149). Patients treated with CAZ-AVI had higher 14-day (80.7% vs 60.6%, P = .011) and 30-day (83.1% vs 60.6%, P = .004) clinical success and lower 30-day mortality (13.25% vs 27.3%, P = .053) than those receiving BAT. In the adjusted analysis, CAZ-AVI increased the probability of 14-day (adjusted odds ratio [aOR], 2.65; 95% confidence interval [CI], 1.03-6.84; P = .044) and 30-day clinical success (aOR, 3.14; 95% CI, 1.17-8.40; P = .023). In contrast, CAZ-AVI therapy was not independently associated with 30-day mortality. In the CAZ-AVI group, combination therapy was not associated with better outcomes. In conclusion, CAZ-AVI may be considered a first-line treatment in SOT recipients with CPKP-BSI.

Validation of the INCREMENT-SOT-CPE score in a large cohort of liver transplant recipients with carbapenem-resistant Enterobacterales infection.

BACKGROUND: Management of infections due to carbapenemase-resistant Enterobacterales (CRE) in solid organ transplant (SOT) recipients remains a difficult challenge. The INCREMENT-SOT-CPE score has been specifically developed from SOT recipients to stratify mortality risk, but an external validation is lacking. METHODS: Multicenter retrospective cohort study of liver transplant (LT) recipients colonized with CRE infection who developed infection after transplant over 7-year period. Primary endpoint was all-cause 30-day mortality from infection onset. A comparison between INCREMENT-SOT-CPE and other selected scores was performed. A two-level mixed effects logistic regression model with random effects for the center was fitted. Performance characteristics at optimal cut-point were calculated. Multivariable Cox regression analysis of risk factors for all-cause 30-day mortality was carried out. RESULTS: Overall, 250 CRE carriers developed infection after LT and were analyzed. The median age was 55 years (interquartile range [IQR]: 46-62) and 157 were males (62.8%). All-cause 30-day mortality was 35.6%. A sequential organ failure assessment (SOFA) score ≥ 11 showed a sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy of 69.7%, 76.4%, 62.0%, 82.0%, and 74.0%, respectively. An INCREMENT-SOT-CPE ≥ 11 reported a sensitivity, specificity, PPV, NPV, and accuracy of 73.0%, 62.1%, 51.6%, 80.6% and 66.0%, respectively. At multivariable analysis acute renal failure, prolonged mechanical ventilation, INCREMENT-SOT-CPE score ≥ 11 and SOFA score ≥ 11 were independently associated with all-cause 30-day mortality, while a tigecycline-based targeted regimen was found to be protective. CONCLUSIONS: Both INCREMENT-SOT-CPE ≥ 11 and SOFA ≥ 11 were identified as strong predictors of all-cause 30-day mortality in a large cohort of CRE carriers developing infection after LT.

Applying mucosal barrier injury laboratory-confirmed bloodstream infection criteria in patients with solid tumors and hematologic malignancies: A retrospective cohort study looking for the real source of infection.

We evaluated the interference of the mucosal barrier injury (MBI) laboratory-confirmed bloodstream infection (MBI-LCBI) criteria on the central-line-associated bloodstream infection (CLABSI) incidence density, and the proportion of catheter-related bloodstream infections (CRBSIs) among those classified as MBI. We detected 339 CLABSIs: 15.0% were classified as MBI-LCBIs, and among these, 19.6% were classified as CRBSIs.

Impact of COVID-19 on healthcare-associated infections: Antimicrobial consumption does not follow antimicrobial resistance

Abstract Background This study aimed to analyze the Healthcare-Associated Infections (HAI) rates and antimicrobial consumption in Intensive Care Units (ICU) in São Paulo city during the COVID-19 pandemic and compare them with the pre-pandemic period. Methods This cohort included all hospitals that reported HAI rates (Central-Line-Associated Bloodstream Infection ‒ CLABSI and Ventilator-Associated Pneumonia ‒ VAP), the proportion of microorganisms that caused CLABSI, the proportion of resistant microorganisms, and antimicrobial consumption from January 2017 ‒ December 2020. Hospitals were stratified by the number of beds, Central Venous Catheter (CVC) utilization rate, Mechanical-Ventilation (MV) utilization rate, and type of funding. Statistical analyses were based on time-series plots and regression models. Results 220 ICUs were included. The authors observed an abrupt increase in CLABSI rates after the pandemic onset. High CLABSI rates during the pandemic were associated with hospital size, funding (public and non-profit private), and low CVC use (≤ 50%). An increase in VAP rates was associated with public hospitals, and high MV use (> 35%). The susceptibility profile of microorganisms did not differ from that of the pre-pandemic period. polymyxin, glycopeptides, and antifungal use increased, especially in COVID-19 ICUs. Conclusions HAI increased during COVID-19. The microorganisms' susceptibility profile did not change with the pandemic, but the authors observed a disproportionate increase in large-spectrum antimicrobial drug use.

Brazilian perspective: antimicrobial stewardship in solid organ transplant.

BACKGROUND: The incidence of multidrug resistant organisms (MDROs) infections among solid organ transplant (SOT) patients is very high in Brazil. METHODS: This review will discuss antimicrobial use and resistance in SOT in Brazil, highlighting the main barriers and facilitators for implementation of an antimicrobial stewardship programme (ASP). RESULTS: The most common group of MDROs is carbapenem-resistant Gram-negative bacteria and vancomycin-resistant Enterococcus. Carbapenem-resistant Enterobacterales (CREs) are the most frequent MDROs and have been reported as donor-derived as well. Although ASPs are mandatory in the country, there is a lack of information regarding ASPs in SOT recipients. The main barriers for the implementation of ASPs in Brazilian hospitals are lack of electronic medical records, absence of national guidelines specific to SOT recipients, lack of recommendations on surveillance culture to evaluate colonization and transmission of donor-derived MDROs, limited availability of rapid diagnostic tests, and insufficient pharmacist and clinician time allocated to ASP activities in some SOT centers. CONCLUSIONS: The incidence of MDRO infections caused mainly by VREs and CREs is very high in the country. There is limited data regarding antimicrobial use among SOT recipients in Brazil. The absence of antimicrobial stewardship national guidelines specific to SOT recipients is one of the main barriers for the implementation of ASPs in Brazilian hospitals.

Correlating drug prescriptions with prognosis in severe COVID-19: first step towards resource management.

BACKGROUND: Optimal COVID-19 management is still undefined. In this complicated scenario, the construction of a computational model capable of extracting information from electronic medical records, correlating signs, symptoms and medical prescriptions, could improve patient management/prognosis. METHODS: The aim of this study is to investigate the correlation between drug prescriptions and outcome in patients with COVID-19. We extracted data from 3674 medical records of hospitalized patients: drug prescriptions, outcome, and demographics. The outcome evaluated was hospital outcome. We applied correlation analysis using a Logistic Regression algorithm for machine learning with Lasso and Matthews correlation coefficient. RESULTS: We found correlations between drugs and patient outcomes (death/discharged alive). Anticoagulants, used very frequently during all phases of the disease, were associated with good prognosis only after the first week of symptoms. Antibiotics very frequently prescribed, especially early, were not correlated with outcome, suggesting that bacterial infections may not be important in determining prognosis. There were no differences between age groups. CONCLUSIONS: In conclusion, we achieved an important result in the area of Artificial Intelligence, as we were able to establish a correlation between concrete variables in a real and extremely complex environment of clinical data from COVID-19. Our results are an initial and promising contribution in decision-making and real-time environments to support resource management and forecasting prognosis of patients with COVID-19.

Prediction models for carbapenem-resistant Enterobacterales carriage at liver transplantation: A multicenter retrospective study.

BACKGROUND: Carbapenem-resistant Enterobacterales (CRE) colonisation at liver transplantation (LT) increases the risk of CRE infection after LT, which impacts on recipients' survival. Colonization status usually becomes evident only near LT. Thus, predictive models can be useful to guide antibiotic prophylaxis in endemic centres. AIMS: This study aimed to identify risk factors for CRE colonisation at LT in order to build a predictive model. METHODS: Retrospective multicentre study including consecutive adult patients who underwent LT, from 2010 to 2019, at two large teaching hospitals. We excluded patients who had CRE infections 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 the waiting list, MELD and CLIF-SOFA scores, antibiotic use, intensive care unit and hospital stay, and infections. A machine learning model was trained to detect the probability of a patient being colonized with CRE at LT. RESULTS: A total of 1544 patients were analyzed, 116 (7.5%) patients were colonized by CRE at LT. The median time from CRE isolation to LT was 5 days. Use of antibiotics, hepato-renal syndrome, worst CLIF sofa score, and use of beta-lactam/beta-lactamase inhibitor increased the probability of a patient having pre-LT CRE. The proposed algorithm had a sensitivity of 66% and a specificity of 83% with a negative predictive value of 97%. CONCLUSIONS: We created a model able to predict CRE colonization at LT based on easy-to-obtain features that could guide antibiotic prophylaxis.

Predicting the outcome for COVID-19 patients by applying time series classification to electronic health records.

BACKGROUND: COVID-19 caused more than 622 thousand deaths in Brazil. The infection can be asymptomatic and cause mild symptoms, but it also can evolve into a severe disease and lead to death. It is difficult to predict which patients will develop severe disease. There are, in the literature, machine learning models capable of assisting diagnose and predicting outcomes for several diseases, but usually these models require laboratory tests and/or imaging. METHODS: We conducted a observational cohort study that evaluated vital signs and measurements from patients who were admitted to Hospital das Clínicas (São Paulo, Brazil) between March 2020 and October 2021 due to COVID-19. The data was then represented as univariate and multivariate time series, that were used to train and test machine learning models capable of predicting a patient's outcome. RESULTS: Time series-based machine learning models are capable of predicting a COVID-19 patient's outcome with up to 96% general accuracy and 81% accuracy considering only the first hospitalization day. The models can reach up to 99% sensitivity (discharge prediction) and up to 91% specificity (death prediction). CONCLUSIONS: Results indicate that time series-based machine learning models combined with easily obtainable data can predict COVID-19 outcomes and support clinical decisions. With further research, these models can potentially help doctors diagnose other diseases.

Clinical profile and mortality predictors for tuberculosis disease among patients with solid tumours.

BACKGROUND: The characteristics of Mycobacterium tuberculosis (MTb) disease are still obscure in patients with solid tumours, as well as the benefits of screening and treating latent tuberculosis infection (LTBI) in these patients. Our objective was to trace the clinical profile of these individuals and assess the mortality predictors related to tuberculosis (TB). METHODS: We reviewed the medical records of 126 patients with solid tumours malignancy and who developed TB disease between January 2009 and April 2018 at a cancer referral centre. RESULTS: The most common locations of malignancy were head and neck, with squamous cell carcinoma being the most frequent histological type, the majority (97/126) presenting locally invasive tumours (T3 or T4). A total of 120 had TB pulmonary and the microbiological diagnosis was performed in 103/126. The following variables were associated with the risk of death from TB: DPOC lung cancer, BMI, malnutrition, metastasis and ECOG 3 or 4. Previous chemotherapy treatment was a protective factor. CONCLUSIONS: Male, usage of alcohol and smoking were the most predominant patients characteristics in our sample. In the multivariate analysis, lung cancer, presence of metastasis and ECOG ≥ 3 were associated with death from TB.