A. Mizrahi

Hospital cost of Clostridium difficile infection including the contribution of recurrences in French acute-care hospitals

BACKGROUND The impact of Clostridium difficile infection (CDI) on healthcare costs is significant due to the extra costs of associated inpatient care. However, the specific contribution of recurrences has rarely been studied. AIM The aim of this study was to estimate the hospital costs of CDI and the fraction attributable to recurrences in French acute-care hospitals. METHODS A retrospective study was performed for 2011 on a sample of 12 large acute-care hospitals. CDI costs were estimated from both hospital and public insurance perspectives. For each stay, CDI additional costs were estimated by comparison to controls without CDI extracted from the national DRG (diagnosis-related group) database and matched on DRG, age and sex. When CDI was the primary diagnosis, the full cost of stay was used. FINDINGS A total of 1067 bacteriological cases of CDI were identified corresponding to 979 stays involving 906 different patients. Recurrence(s) were identified in 118 (12%) of these stays with 51.7% of them having occurred within the same stay as the index episode. Their mean length of stay was 63.8 days compared to 25.1 days for stays with an index case only. The mean extra cost per stay with CDI was estimated at €9,575 (median: €7,514). The extra cost of CDI in public acute-care hospitals was extrapolated to €163.1 million at the national level, of which 12.5% was attributable to recurrences. CONCLUSION The economic burden of CDI is substantial and directly impacts healthcare systems in France.

Prospective evaluation of the Alere i Influenza A&B nucleic acid amplification versus Xpert Flu/RSV

The rapid and accurate detection of influenza virus in respiratory specimens is required for optimal management of patients with acute respiratory infections. Because of the variability of the symptoms and the numerous other causes of influenza-like illness, the diagnosis of influenza cannot be made on the basis of clinical criteria alone. Thus, rapid influenza diagnostic tests have been developed such as the Alere i Influenza A&B isothermal nucleic acid assay. We prospectively evaluated the performance of the Alere i Influenza A&B assay in comparison with our routine Xpert Flu/RSV assay. Positive samples were subtyped according to the protocol from the National Influenza Center (Paris, France). A total of 96 respiratory nasal swab samples were analyzed: with both methods, 38 were positive and 56 were negative. Samples were prospectively collected from January 20 to April 8, 2015, from patient (86 adult and 10 pediatric patients) presenting with an influenza-like illness through the French influenza season. In comparison with the Xpert Flu/RSV assay, the overall sensitivity and specificity of the Alere i Influenza A&B assay were 95% and 100%, respectively. Our results indicate that the Alere i Influenza A&B assay has a good overall analytical performance and a high degree of concordance with the PCR-based Xpert Flu/RSV assay. The Alere i Influenza A&B isothermal nucleic acid amplification test is a powerful tool for influenza detection due to its high sensitivity and specificity as well as its ability to generate results within 15min.

Carriage of ESBL-producing Enterobacteriaceae in French hospitals: the PORTABLSE study

BACKGROUND Currently, contact precautions are recommended for patients colonized or infected with extended-spectrum beta-lactamase-producing Enterobacteriaceae (ESBL-PE). Recent studies have challenged this strategy. This study aimed to assess the rate of ESBL-PE faecal carriage among hospitalized patients according to type of hospital ward, and to identify risk factors associated with carriage. METHODS A point prevalence study was conducted in five different types of hospital ward [medical, surgical, intensive care unit (ICU), after care and rehabilitation, and geriatric] in eight French hospitals. All patients included in the study provided a fresh stool sample. RESULTS In total, 554 patients were included in the study, with a median age of 73 years (range 60-82 years). The overall faecal carriage rate of ESBL-PE was 17.7%. The most frequently encountered species among ESBL-PE was Escherichia coli (71.4%), followed by Klebsiella pneumoniae (14.3%). Risk factors associated with ESBL-PE faecal carriage on univariate analysis were: living in the Paris region (P<0.01) and hospitalization on a geriatric ward (P<0.01). Interestingly, the cumulative duration of hospital stay before screening was not associated with a significantly higher prevalence of ESBL-PE carriage, regardless of ward type. The ESBL-PE colonization rate was much higher for patients hospitalized on geriatric wards (28.1%) and ICUs (21.7%) compared with those for patients hospitalized on surgical wards (14.8%), medical wards (12.8%) or aftercare and rehabilitation (11.2%). CONCLUSION The overall prevalence of ESBL-PE faecal carriage was 17.7%, with only 21% of patients identified previously as carriers. The delay between admission and screening was not associated with an increase in ESBL-PE faecal carriage.

Iconography : Implementation of Alere i Influenza A & B point of care test for the diagnosis of influenza in an emergency department

Study objective This study aimed to evaluate the impact of implementing rapid point‐of‐care testing (POCT) with the Alere i Influenza A & B in an emergency department (ED) during an influenza epidemic. Methods Direct nasal swabs were prospectively collected following the physical examination of patients aged > 18 years who presented to the ED of a tertiary hospital in France with influenza‐like illness (ILI) symptoms (N = 301) between February 1st and March 31st, 2016, which coincided with an influenza epidemic. Laboratory‐based testing (standard of care) was used to obtain a diagnosis in February 2016 (pre‐POCT cohort) and positive results were confirmed using polymerase chain reaction. The primary endpoint was patient time in the ED. Results A total of 169 and 132 patients participated in the pre‐POCT phase and POCT phase respectively. A significantly higher proportion of patients received a positive diagnosis in the POCT cohort compared with the pre‐POCT cohort (31% versus 5.3%, P < 0.01). Mean time spent in the ED and hospitalization rate were significantly lower in the POCT cohort (6.06 h versus 4.15 h, P = 0.03, and 44.4% versus 9.7%, P = 0.02, respectively). Despite similar rates in the prescription of antibiotics and antiviral therapies, the proportion of patients who were referred for additional tests was significantly lower in the POCT cohort (78.1% versus 62.1%, P = 0.003, and 80.5% versus 63.6%, P = 0.01, respectively). Conclusions The Alere i Influenza A & B POCT reduced the length of stay in ED, the hospitalization rates, and the number of additional diagnostic tests compared with standard of care testing.

Clinical impact of rapid bacterial identification by MALDI-TOF MS combined with the bêta-LACTA™ test on early antibiotic adaptation by an antimicrobial stewardship team in bloodstream infections

Abstract Background: Bloodstream infections (BSI) can potentially be life-threatening infections and are associated with a high crude mortality, moreover with an inappropriate first-line antibiotic therapy. Bacterial resistance is more and more frequently observed. New strategies of BSI management are urgently needed. Materials and methods: During an 18-months period, we prospectively evaluated the clinical impact of rapid bacterial identification by MALDI-TOF MS technology combined with an antimicrobial stewardship team (AST) intervention. Furthermore, during an 8-months period, we combined this strategy with the rapid detection of third-generation cephalosporin (3GC) resistance by the Bêta-LACTA™ test (BLT) directly on blood cultures. We then evaluated the theoretical impact of BLT on antibiotic therapy adaptation and establishment of infection control measures. Results: A total of 335 blood cultures were enrolled during the study. MALDI-TOF MS gave accurate identification for 301 blood cultures (89,8%) and led to early antibiotic therapy adaptation for 73 episodes (21.8%). BLT was performed on 141 blood cultures, revealing 28 3GC-resistant bacteria (19.9%). Twenty-one patients (75%) received a non-adapted first-line treatment. The antibiotic therapy adaptation was delayed by 28.1 hours and the establishment of infection control measures by 35 hours with antimicrobial susceptibility testing, compared to the theoretical adaptation with BLT result. Conclusions: These tools can be included in a strategy of bloodstream infections management for a better patient care, optimizing and saving the use of antibiotics, notably carbapenems as well as diminishing the spread of multi-drug resistant bacteria by applying rapidly infection control measures.

Optimization of the β LACTA test for the detection of extended-spectrum-β-lactamase-producing bacteria directly in urine samples

Abstract The β LACTA™ test (BLT) is a chromogenic test detecting resistance to third-generation cephalosporins on bacterial colonies. Some studies have shown that this test can be used directly in urine samples. The aim of this study was to determine the optimal conditions of use of this test in order to detect the ESBL-producing bacteria directly in urine samples. During a 4-months period, a total of 365 consecutive urine samples were tested with the BLT using the recommendation of the manufacturer. We isolated 56 ESBL-producing bacteria and 17 AmpC-overproducing bacteria. ESBL- and/or AmpC β-lactamase-producing bacteria isolates were systematically characterized by disc diffusion antibiotic susceptibility testing interpreted according to the guidelines of EUCAST. The sensitivity and the specificity for 3GC-resistance detection, regardless the mechanism of resistance, were, respectively, 60.3% and 100%, whereas for ESBL detection, it was, respectively, 75.4% and 99.7%. We applied then modification of the initial protocol considering urines with a bacteriuria >1000/μL, a reading time at 30 min and considering any change of the initial colour as positive. The overall sensitivity was 81% and the sensitivity for ESBL-detection raised to 95.7%.

Pseudo-Outbreak of Oxa-23-Mediated Carbapenem-Resistant Acinetobacter baumannii in Urinary Tract Infections Caused by an Automated Urine Analyzer

To the Editor—Acinetobacter baumannii is a gram-negative coccobacilli, isolated from clinical and environmental sources, that is mainly responsible for hospital-acquired pneumonia, catheter-related bacteriemia, and urinary tract infections. This emerging nosocomial pathogen has the ability to acquire various antibiotic resistance mechanisms, facilitating the emergence of multidrug-resistant strains that are able to persist in the hospital environment. OXA-23 is the most widespread b-lactamase with carbapenemase activity found in A. baumannii and has been frequently associated with outbreaks. However, urinary tract infections caused by OXA-23-producing A. baumannii remain rare. Here, we report a pseudo-outbreak of A. baumannii urinary tract infection involving 3 patients that was caused by an automated urine analyzer. On October 13, 2013, the infection control unit at our institution (Groupe Hospitalier Paris Saint Joseph, Paris, France) was notified of 3 imipenemresistant A. baumannii isolated from urine samples of 3 patients. All isolates exhibited an identical phenotype with respect to antibiotic susceptibility. The French national reference center for detection of antimicrobial resistance (Besançon, France) confirmed the presence of the OXA-23 enzyme in 3 genetically related strains. The first patient was hospitalized in the otolaryngology ward and had a urinary catheter. The second patient has been admitted to the emergency department for a fever resulting from an alveolarinterstitial pneumonia. The third patient, notable for a diagnosis of spina bifida, experienced urinary retention. He was hospitalized for a laminectomy and had a urine sample with a positive culture of a susceptible A. baumannii obtained on October 8, 5 days before the pseudo outbreak. Review of the previous 3 months of microbiology records did not find any other imipenem-resistant A. baumannii strains. A thorough ward-based investigation revealed no epidemiological link to suggest cross-infection between the patients. During this period, there was no change in staff, microbiological technique, or culture medium. The identity label on each sample was checked to rule out potential human error. Therefore, the investigation focused on the microbiology laboratory, especially on the flow cytometry–based UF-500i instrument (bioMérieux) used for analysis of urine samples. This analyzer uses a single needle for sample aspiration that is washed by a rinsing system, which is designed to prevent sample-to-sample carryover in the bacterial count. Depending on the quantity of bacteria detected in the sample, an adequate number of rinsing cycles with buffer are performed. Finally, we found that the 3 samples were analyzed successively by the UF-500i. The UF-500i detected 2,461, 1,421, and 10 colony-forming units (CFU)/mL for the first, second, and third patient urine samples, respectively. Bacteria count was 10 CFU/mL for the first sample, 10 CFU/mL for the second, and 10 CFU/mL for the third. Occurrence of 3 consecutive samples harboring an imipenem-resistant A. baumannii suggested a contamination due to the analyzer needle. The urine samples were cultured again, which led to identical colony counts. New urine samples were requested to obtain control samples for each patient, and the first patient was the only one found to be infected with A. baumannii. He was affected by a catheter-related urinary tract infection due to an imipenem-resistant A. baumannii strain 3 days before. No evidence of contamination of the following samples was found at the time. Unexpectedly, attempts to reproduce the contamination were unsuccessful. Occurrence of 3 patients with urinary tract infections due to an OXA-23-producing A. baumannii on the same day from 3 independent care units is an improbable event. In such a situation, a pseudo-outbreak should be suspected. Among 20 reported pseudo-outbreaks, Weinstein and Stamm identified 7 that originated within the microbiology laboratory. Imipenem-resistant A. baumannii have already been reported in pseudo-outbreaks, but the origin was a false susceptibility testing by a rapid automated system. Pseudo-outbreaks involving the same automated urine analyzer were only reported with Pseudomonas aeruginosa. To the best of our knowledge, this is the first description of a pseudo-outbreak of A. baumannii urinary tract infection caused by an automated urine analyzer. This pseudo-outbreak was due to inadequate decontamination of a sampling needle. The UF-500i analyzer does not use a disinfectant to sterilize the sampling needle between the samples but just rinses the needle with the washing buffer. A. baumannii creates biofilms on medical devices, such as urinary catheters. A part of the preexisting biofilm could have been dropped in the urine sample and then drawn by the automated needle, leading to the partial failure of the washing process. Such an event could explain why attempts to reproduce the contamination were unsuccessful. We focused on these cases because a multidrug-resistant strain was involved, but contamination involving susceptible bacteria may occur without being detected. Therefore, we recommend that urine samples be plated before being analyzed by the UF500i. If not, regular sterility control must be applied when bacterial cultures are performed after treatment of the sample by the UF-500i. This pseudo-outbreak did not lead to antimicrobial therapy. Rapid identification of a pseudo-outbreak is particularly important to prevent the implementation of time-consuming and costly measures, such as patient isolation, microbiological analysis to detect secondary transmission, and inappropriate antimicrobial therapy. Infection control units as well as microbiologists should suspect a pseudo-outbreak if a cluster of

Clinical diagnostic and therapeutic aspects of 221 consecutive anorectal Chlamydia trachomatis and Neisseria gonorrhoeae sexually transmitted infections among men who have sex with men

OBJECTIVES Proctitis caused by Chlamydia trachomatis (CT) and Neisseria gonorrhoeae (NG) are known as sexually transmitted infections (STI). This study describes their clinical, diagnostic and therapeutic aspects. METHODS Between 01/2013-03/2015, all MSM consulting for proctitis at proctology Institute-Saint-Josephs Hospital, Paris, were included. Demographic, past-medical history, STI status and medical treatment were collected. Detection of CT/NG was performed by Transcription-Mediated Amplification (TMA) and antimicrobial susceptibilities for Ng by agar diffusion method. RESULTS On 441 rectal samples collected, 221 (50.1%) were positive: 109 Ct (49.3%), 70 Ng (31.7%), 42 positive for both etiologies (19%). Among Ng infections, no resistance was detected to azithromycin and ceftriaxone. However, 84 strains (43.2%) were resistant to fluoroquinolones. More than one episode was diagnosed for 10 (5.1%) and 12 (6.2%) patients with CT and NG infections respectively. Anal abscesses were found for 27 (13.9%) patients, and 14 (7.2%) of them underwent surgery for anal fistula. CONCLUSIONS The prevalence of CT/NG anorectal infections described is high on symptomatic patients, and a significant level of abscess was reported. These results confirm the interest of the association of recommended antibiotics excluding quinolones. Prospective studies would be relevant on complicated forms of anorectal infections.

Prospective evaluation of rapid antimicrobial susceptibility testing by disk diffusion on Mueller-Hinton rapid-SIR directly on blood cultures

BACKGROUND With the worldwide spread of antibiotic resistance, delivering antibiotic susceptibility test (AST) results in a timely manner represents a major challenge. In cases of sepsis, rapid AST may facilitate early optimization of empiric antibiotic therapy. Disc diffusion is a well-standardized AST method, however 16 to 24 h are required to achieve an overall AST profile according to antimicrobial societies. METHODS In this prospective pilot study, we evaluated the performance of Mueller-Hinton-Rapid-SIR (MHR-SIR) agar after 6-8 h of incubation in comparison with standard MH agar after 16 h of incubation directly on positive blood cultures caused by Enterobacteriaceae and Staphylococcus aureus from routine clinical microbiology. A total of 133 positive blood samples including 110 Enterobacteriaceae (83%) and 23 Staphylococcus aureus (17%) were tested in parallel by two direct AST methods, each using EUCAST breakpoints. For each combination bacterium and antibiotic, we compared the categorical agreement and the correlation between the diameters obtained by MHR-SIR and by standard MH. RESULTS Our results showed 97.7% categorical agreement for Enterobacteriaceae, with 1.4% minor errors, 0.4% major errors and 0.5% very major errors. For S. aureus, we observed 97.8% categorical agreement, 1.9% minor errors, 0.3% major errors and no very major errors. CONCLUSION Our results showed excellent categorical agreement and correlations between diameters for MHR-SIR and standard MH methods. MHRSIR can predict the result of overall AST profile within 6-8 h with reliable results. AST is obtained on the same day the blood culture becomes positive, with a very moderate cost.

Infective endocarditis: Clinical presentation, etiology, and early predictors of in-hospital case fatality.

OBJECTIVE We aimed to assess the clinical presentation, microbial etiology and outcome of patients presenting with infective endocarditis (IE). PATIENTS AND METHODS We conducted a four-year retrospective study including all patients presenting with IE. RESULTS We included 121 patients in the study. The median age was 74.8years. Most patients had native valve IE (57%). Staphylococcus aureus accounted for 24.8% of all IE. Surgery was indicated for 70 patients (57.9%) but actually performed in only 55 (44.7%). Factors associated with surgery were younger age (P=0.002) and prosthetic valve IE (P=0.001). Risk factors associated with in-hospital mortality were diabetes mellitus (OR=3.17), chronic renal insufficiency (OR=6.62), and surgical indication (OR=3.49). Mortality of patients who underwent surgery was one sixth of that of patients with surgical indication who did not have the surgery (P<0.001).