Search

Introduction and Objectives: there is insufficient data regarding bacterial infections in patients with cirrhosis to support recommendations for empiric antibiotic treatments, particularly in Latin America. This study aimed to evaluate bacterial infection's clinical impact and microbiological characteristics, intending to serve as a platform to revise current practices. Materials and Methods: multicenter prospective cohort study of patients with cirrhosis and bacterial infections from Argentina and Uruguay. Patient and infection-related information were collected, focusing on microbiology, antibiotic susceptibility patterns, and outcomes. Results: 472 patients were included. Spontaneous bacterial infections and urinary tract infections (UTIs) were registered in 187 (39.6%) and 116 (24.6%) patients, respectively, representing the most common infections. Of the 256 culture-positive infections, 103 (40.2%) were caused by multidrug-resistant organisms (reaching 50% for UTI), and 181 (70.7%) received adequate initial antibiotic treatment. The coverage of cefepime and ceftriaxone was over 70% for the empirical treatment of community-acquired spontaneous infections, but ceftazidime´s coverage was only 40%. For all UTI cases and for healthcare-associated or nosocomial spontaneous bacterial infections, the lower-spectrum antibiotics that covered at least 70% of the isolations were imipenem and meropenem. During hospitalization, a second bacterial infection was diagnosed in 9.8% of patients, 23.9% required at least one organ support, and 19.5% died. Conclusions: short-term mortality of bacterial infections in patients with cirrhosis is very high, and a high percentage were caused by multidrug-resistant organisms, particularly in UTIs. The information provided might serve to adapt recommendations, particularly related to empirical antibiotic treatment in Argentina and Uruguay. The study was registered in Clinical Trials (NCT03919032).

Introduction and Objectives: Selecting an empiric antibiotic treatment in patients with cirrhosis and spontaneous bacterial infections is challenging. It is of paramount importance to have local epidemiological data to maximize pathogen coverage while minimizing the unnecessary use of broad-spectrum antibiotics. This study aimed to describe the patterns of antibiotic resistance of spontaneous bacterial infections according to the site of acquisition. Materials and Methods: Analysis of the multicenter prospective cohort study of cirrhotic patients with bacterial infections in Argentina and Uruguay (NCT03919032). Only culture-positive spontaneous infections were included in this study: spontaneous bacterial peritonitis (SBP), spontaneous bacterial empyema (SBE), and spontaneous bacteremia (SB). We estimated the proportion of infections that were sensitive to various antibiotics according to where the infection was acquired: community-acquired (CA), healthcare-associated (HCA), or nosocomial (NOS). Approximately 80% coverage is advisable for empiric treatments in stable patients and 90% for critically-ill patients. Results: The main cohort included 472 patients, of whom 97 presented culture-positive spontaneous infections and were included: with 57 (59%) SBP, 34 (35%) SB, and 4 (6%) SBE. Regarding the site of acquisition, 43% were CA, 36% NOS, and 21% HCA. Gram-positive and negative bacteria were found in 53% and 47% of the infections. The most frequent isolations were Streptococcus spp (26%), E coli (20), K pneumonia (15%), S Aureus (10%), E. faecium (6%) and E. faecalis (4%). Multidrug-resistant organisms (MDROs) were isolated in 35% of the patients. As shown in the table, cefepime and ceftriaxone offer the most rational coverage for CA and HCA infections, and imipenem or meropenem for NOS infections. However, in critically-ill patients, broader-spectrum antibiotics are needed to achieve a coverage closer to 90% (table). Conclusions: We present, for the first time in our region, evidence-based recommendations for the empirical treatment of spontaneous bacterial infections. Prior colonization and/or infections by MDROs might refine even more the antibiotic selection and should be explored.

Background: Antibiotic de-escalation is considered a safe strategy that reduces costs and the risk of multi-drug resistant infections. However, its prevalence and associated factors in real-life practice were not reported in patients with cirrhosis in Latin-America. Aims: To estimate the prevalence of antibiotic de-escalation in patients with cirrhosis in real life-practice, and to explore its associated factors. Methods: We performed an analysis of the multicenter prospective cohort study of cirrhotic patients with bacterial infections throughout Argentina and Uruguay (clinicatrials.gov NCT03919032). Patients who died in the first 72 hs from the diagnosis of the infection were excluded. In accordance with guidelines, de-escalation was defined as changing the initially antimicrobials to a narrower spectrum regimen, or suspending one or more of the empirical antibiotics, according to culture results or to other clinical reasons, either in patients with culture-positive or culture-negative bacterial infections. We used inverse probability weighting (IPW) of having a culture-positive infection to estimate its causal effect on de-escalation. Results: We included 450 patients. Most frequent infections were SBP (30.4%), and urinary tract infection (12.9%). Overall, 243 (54%) infections were culture-positive, and 207 (46%) culture-negative. De-escalation was reported in 85 patients (18.9%: 95% CI 15%-22%) at a mean of 3.3 ± 2.4 days from treatment initiation and was more frequent in culture-positive than culture-negative infections (28.4% vs 7.7%, p< 0.001). The table shows the crude analyses of variables associated with de-escalation. Culture-positive infection was strongly and independently associated with de-escalation (ORIPW 6.08; 95% CI: 2.90-12.70; p < 0.001). Conclusions: Antibiotic de-escalation was reported in one-fifth of in-patients with cirrhosis. Given that having a culture-positive infection had a strong effect on de-escalation, efforts should be made to increase the likelihood of obtaining adequate culture samples in a timely manner.

Background: Predicting short-term mortality in patients with cirrhosis and bacterial infections is challenging. Aims: To compare the performance of various scores in predicting in-hospital mortality in this population. Methods: We performed an analysis of the multicenter prospective cohort study of patients with cirrhosis with bacterial infections throughout Argentina and Uruguay (clinicatrials.gov.NCT03919032). Patients were classified according to the CLIF criteria as having ACLF or mere acute decompensation (AD). We evaluated the performance of scores of liver disease and infection severity in predicting in-hospital mortality. MELD, MELD-Na, and Quick SOFA (qSOFA) were computed in all patients. CLIF-AD was only computed in patients without ACLF, and CLIF-ACLF only in patients with ACLF. We plotted ROC curves and estimated their area under the curve (AUROC). Results: We included 472 patients: 66% male, mean age 57 ± 12 years. Most frequent infections: SBP (30%) and urinary tract infection (25%). Overall, 332 (70%) patients had acute decompensation, and 140 (30%) ACLF. In-hospital mortality rate was 19%: 41% in patients with ACLF vs 10% in patients with AD (p