Search

Clostridioides difficile infection (CDI) is a disease with high morbidity and mortality rates. The objective of this study was to describe CDI epidemiology and patient characteristics over a 5-year period in Switzerland and assess risk factors for mortality, recurrence and severe CDI.We retrospectively included all consecutive CDI cases having occurred in adult patients hospitalized in two tertiary centers: the Lausanne University Hospital (1000 beds) and the University Hospital of Zurich (900 beds), between 2014 and 2018. Suspected cases of CDI were identified from the microbiology laboratory database on the basis of a positive test and confirmed by records review.During first CDI episodes, the median age was 67 years and the median Charlson comorbidity index (CCI) score was 5. All in all, 299 out of 826 patients (36.2%) had severe infection based on the Infectious Diseases Society of America criteria. In the multivariable analysis, CCI was associated with increased risk of mortality. None of the factors recorded on admission were significantly associated with increased risk of recurrence. In the multivariable analysis, male sex and CCI were associated with severity, while immunosuppression was associated with less severe presentation.If we did not identify any criteria on admission that could be predictive of recurrences, this could be explained the retrospective nature of the study. A higher comorbidity index is a key driver for severe CDI and mortality. Reporting of CDI is not mandatory in Switzerland; structuration of CDI reporting should be a short-term priority.

International audience; BACKGROUND: Cohorting carbapenemase-producing Enterobacteriaceae (CPE) carriers during hospitalization limits in-hospital spreading.OBJECTIVE: To identify risk factors for CPE acquisition among contacts of an index patient in non-cohorted populations.METHODS: A multicentre retrospective matched case-control study was conducted in five hospitals. Each contact patient (case) who acquired Klebsiella pneumoniae (KP)-OXA-48 from an index patient was compared to three contact (controls) with the same index patients matched with hospitalization in the same unit and similar exposure times.RESULTS: Fifty-one secondary cases and 131 controls were included. By univariate analysis, exposure time (odds ratio: 1.06; 95% confidence interval: 1.02-1.1; P = 0.006), concomitant infection at admission (3.23; 1.42-7.35; P = 0.005), antimicrobial therapy within the last month before hospitalization (2.88; 1.34-6.2; P = 0.007), antimicrobial therapy during the exposure time (5.36; 2.28-12.6; P

Resume Les recommandations pour la prevention de l’infection avant un acte invasif, principalement la pose et l’entretien des acces vasculaires et la preparation cutanee avant chirurgie, ont evolue ces dernieres annees. L’absence d’utilite de la detersion sur peau propre avant antisepsie cutanee avait ete affirmee pour la preparation avant chirurgie en 2013 ; sur la base de nouvelles donnees de la litterature, elle a ete reaffirmee et etendue a tous les actes invasifs en 2016. La France rejoint ainsi la grande majorite des recommandations d’autres pays. L’autre nouveaute est le choix preferentiel de la chlorhexidine alcoolique plutot que la polividone iodee alcoolique pour l’antisepsie de la peau saine pour la pose des catheters vasculaires, laissant le choix de l’un ou l’autre gamme pour l’antisepsie avant chirurgie.

Summary Background Since 1990, several studies have focused on safety and patient satisfaction in connection with day surgery. However, to date, no meta-analysis has investigated the overall prevalence of surgical site infections (SSI). Aim To estimate the overall prevalence of SSI following day surgery, regardless of the type of surgery. Method A systematic review and a meta-analysis of the prevalence of SSI following day surgery, regardless of the type of surgery, was conducted, seeking all studies before June 2016. A pooled random effects model using the DerSimonian and Laird approach was used to estimate overall prevalence. A double arcsine transformation was used to stabilize the variance of proportions. After performing a sensitivity analysis to validate the robustness of the method, univariate and multi-variate meta-regressions were used to test the effect of date of publication, country of study, study population, type of specialty, contamination class, time of postoperative patient visit after day surgery, and duration of hospital care. Findings Ninety articles, both observational and randomized, were analysed. The estimated overall prevalence of SSI among patients who underwent day surgery was 1.36% (95% confidence interval 1.1–1.6), with a Bayesian probability between 1 and 2% of 96.5%. The date of publication was associated with the prevalence of SSI (coefficient −0.001, P = 0.04), and the specialty (digestive vs non-digestive surgery) tended to be associated with the prevalence of SSI (coefficient 0.03, P = 0.064). Conclusion The meta-analysis showed a low prevalence of SSI following day surgery, regardless of the surgical procedure.

The increasing bacterial resistance to antibiotics has become a major public health concern bringing the threat of therapeutic impasses. In this context, control of the spread of highly-resistant bacteria emerging antibiotics (BHRe), such as glycopeptide-resistant enterococci (VRE) and Enterobacteriaceae producing carbapenemases (CPE), is based on a dual strategy of reducing the prescription of antibiotics to limit the pressure selection and preventing the spread from carriers. Prevention strategy is based on three different levels such as standard precautions for all patients with a particular focus on the management of excreta, and additional precautions for BHRe carriers. What makes it difficult is that carriage is usually completely asymptomatic, enterobacteria and enterococci are normal commensal of gut microbiota. Explosive dissemination of Enterobacteriaceae producing extended spectrum beta-lactamases in hospital and community heralds the emergence of CPE whose import by patients with a history of hospitalization in abroad may be the main source of spread in France.

Controlling the spread of multi- or extensively drug-resistant bacteria (MDR or XDR) includes a dual strategy for reducing antibiotic prescriptions and preventing their spread from patient carriers. Standard precautions are applicable to all health professionals caring for any patients; additional barrier precautions (isolation) are recommended for patients carrying transmissible infectious diseases or MDR bacteria in sporadic or epidemic situations. Moreover, additional precautions may be required for populations at particular risk of infection or colonization by emerging XDR (eXDR), defined in our country as carbapenemase-producing Enterobacteriaceae and vancomycin-resistant enterococci. Our ability to detect and identify eXDR carriers early and ensure their follow-up, through effective communication between all those involved, is a significant challenge for controlling their spread. Thus, the French High Committee for Public Health has updated and standardized all French existing recommendations concerning the prevention of the cross-transmission of these bacteria, and these recommendations are summarized in this review. The recommendations are based on scientific and operational knowledge up to 2013. Different preventive strategies are recommended for patients found to be carrying eXDR and those who are considered to be at risk of having eXDR because of a history of contact. The local context, the experience of the infection control team, the different times at which detection of eXDR takes place (during admission, hospitalization, etc.) and the epidemiological situation (sporadic cases, clusters, outbreaks, widespread epidemic) must be included in risk assessments that in turn inform the control measures that should be applied in each clinical circumstance.

International audience; OBJECTIVE: Two randomized controlled trials (RCTs) showed the non-inferiority of fidaxomicin compared with vancomycin for Clostridium difficile infection (CDI) treatment and its superiority regarding recurrence rate. The aim of this study was to evaluate fidaxomicin's efficacy in clinical practice.METHODS: This single-center prospective cohort study included hospitalized patients treated with fidaxomicin for CDI. Demographic, clinical and biological data were collected. Primary outcome was efficacy of fidaxomicin (clinical cure, recurrence and global cure) at 10 weeks. Secondary outcome was efficacy among different subgroups.RESULTS: Ninety-nine patients were included: 42 severe CDI, 16 complicated CDI and 41 recurrent CDI. Rates of clinical cure, recurrence and global cure were 87, 15 and 59%, respectively. Subgroup analysis showed a higher recurrence rate for patients with recurrent CDI compared with first episode (8 vs. 26%; p = 0.04). Binary toxin was associated with severe/complicated CDI (80 vs. 50%; p

Résumé La dernière décennie a été marquée par l’incidence croissante des infections émergentes avec, pour principale conséquence, la nécessité de faire évoluer notre pratique et notre système de santé pour s’y adapter. Cette revue rassemble les principales crises auxquelles nous avons été confrontés ces dix dernières années avec une emphase particulière sur Ebola et les conséquences sur notre politique de prise en charge.

« Gestion preoperatoire du risque infectieux : strategie de decolonisation preoperatoire du portage de Staphylococcus aureus » (mise a jour 2013) D. Lepelletier a,j,∗, P. Saliou b,j, A. Lefebvre c,j, J.-C. Lucet d,j, B. Grandbastien e,j, F. Bruyere f, J.-P. Stahl g, O. Keita-Perse h,j, P. Berthelot i,j, S. Aho c,j, for the workgroup SF2H a Service de bacteriologie et d’hygiene hospitaliere, CHU de Nantes, 44000 Nantes, France b Service d’hygiene hospitaliere, CHU de Brest, 29000 Brest, France c Service d’epidemiologie et d’hygiene hospitaliere, CHU de Dijon, 21000 Dijon, France d Unite hospitaliere de lutte contre l’infection nosocomiale, hopital Bichat-Claude-Bernard, AP–HP, 75018 Paris, France e Service de gestion du risque infectieux, des vigilances et d’infectiologie, CHRU de Lille, 59000 Lille, France f Service d’urologie, hopital Bretonneau, CHRU de Tours, 37000 Tours, France g Service des maladies infectieuses et tropicales, CHU de Grenoble, 38000 Grenoble, France h Service d’hygiene hospitaliere, hopital Princesse Grâce, 44000 MC98012, France i Unite d’hygiene inter-hospitaliere, service des maladies infectieuses, laboratoire de microbiologie, CHU de Saint-Etienne, 42000 Saint-Etienne, France j Societe francaise d’hygiene hospitaliere, France

Resume Objectifs La methode de detection de synergie entre les disques de cefepime (FEP) et d’amoxicilline–acide clavulanique a ete comparee avec la methode des quatre disques combines Mastdiscs™ ID AmpC sLSE pour depister le portage digestif des enterobacteries productrices de s-lactamase a spectre etendu (EsLSE) dans des ecouvillons rectaux. Methodes Deux cent quarante-neuf ecouvillons rectaux ont ete directement ensemences sur une gelose de Mueller-Hinton et analyses selon les deux methodes. Resultats Parmi les 249 ecouvillons rectaux analyses, 41 (16 %) et 208 (84 %) etaient positifs et negatifs a EsLSE, respectivement. Douze (29 %) et 20 (49 %) ecouvillons positifs a EsLSE ont ete detectes des 24 heures avec les methodes de synergie et Mastdiscs™, respectivement (p = 0,013). Cent cinquante-huit (76 %) et 183 (88 %) des 208 ecouvillons ont ete definitivement rendus negatifs a 24 heures avec les methodes de synergie et Mastdiscs™, respectivement (p Conclusions La methode Mastdiscs™ est un test phenotypique rapide et simple qui a permis, d’une part, de detecter des la 24e heure un plus grand nombre d’EsLSE et, d’autre part, de preciser plus rapidement l’absence de portage d’EsLSE.

In the present study, two pre-analytic processes for mass spectrometric bacterial identification were compared: the time-consuming reference method, chemical extraction, and the direct smear technique directly using cultured colonies without any further preparation. These pre-analytic processes were compared in the identification of a total of 238 strains of anaerobic bacteria representing 34 species. The results showed that 218/238 strains were identified following chemical extraction, 185 identifications (77.7%) were secured to both genus and species [log(score) > 2.0] whereas 33 identifications (14%) were secured to genus only [log(score) between 1.7 and 2.0]. Following direct smear, 207/238 anaerobic bacteria were identified, 158 identifications (66.4%) were secured to both genus and species [log(score) > 2.0] whereas 49 identifications were secured to genus only [log(score) between 1.7 and 2.0]. Twenty strains were not identified [log(score) < 1.7] by MALDI-TOF MS following chemical extraction whereas 31 strains were not identified with the direct smear technique. Although direct smear led to a significant decrease of the log(score) values for the Clostridium genus and the Gram positive anaerobic bacteria (GPAC) group (p < 0.0001, Wilcoxon test), identification to both species and genus were not changed. However these differences were not statistically significant (p = 0.1, Chi square). Therefore, MALDI-TOF MS identification following the direct smear technique appears to both non-inferior to the reference method and relevant for anaerobic bacteria identification.

Summary Aim To determine the number of deaths associated with nosocomial infection (NI) occurring in French hospitals, and to evaluate the role of hospital mortality reviews in assessing the preventability of NI-associated death. Methods The study considered 13,537 consecutive deaths occurring in 14 French hospitals in 2007 and 2008. An expert commission including attending physicians and nurses determined the likelihood that deaths could be attributed to NI, and the preventability of NI and death. Findings Medical records of the 2355 eligible patients with a McCabe score of 0 or 1 who died more than 48 h after admission were reviewed. Among them, 33% had at least one NI. Death was attributable to NI in 182 patients, and was considered preventable in 35 cases. Among these, 10 deaths were unexpected. Conclusion Extrapolating these figures nationally, approximately 3500 [95% confidence interval (CI) 2605–4036] deaths attributable to NI occur in France annually. Among these, approximately 1300 NIs (95% CI 357–2196) and 800 deaths (95% CI 51–1481) can be considered preventable. Hospital mortality review commissions can help to improve the quality of health care by identifying circumstances associated with NI contributing to death, and targeting specific preventative measures. Such hospital commissions should involve all healthcare personnel.

International audience; PURPOSE:Our goal is to assess the prevalence of questioning about the appropriateness of initiating or maintaining life-sustaining treatments (LST) in French-speaking paediatric intensive care units (PICUs) and to evaluate time utilisation related to decision-making processes (DMP).METHODS:18-month, multicentre, prospective, descriptive, observational study in 15 French-speaking PICUs.RESULTS:Among the 5,602 children admitted, 410 died (7.3%), including 175 after forgoing LST (42.7% of deaths). LST was questioned in 308 children (5.5%) with a prevalence of 13.3 per 100 patient-days. More than 30% of children survived despite the appropriateness of LST being questioned (23% despite a decision to forgo treatment). Median caregiver time spent on making and presenting the decisions was 11 h per child.CONCLUSIONS:In this study, on any given day in each 10-bed PICU, there was more than one child for whom a DMP was underway. Of children, 23% survived despite a decision to forgo LST being made, which underlines the need to elaborate a care plan for these children. Also, DMP represented a large amount of staff time that is undervalued but necessary to ensure optimal palliative practice in PICU.

Resume Objectif Evaluer la qualite des prescriptions de fluoroquinolones, d’aminosides et de l’amoxicilline-acide clavulanique au sein des secteurs de medecine. Methodes Recueil un jour donne des prescriptions antibiotiques ciblees sur ces molecules. Comparaison aux recommandations locales et aux manuels de references. Evaluation des antibiotherapies sur le choix de la molecule, les conditions d’administration (posologies, voies et rythmes d’administration, durees de traitement), la reevaluation des traitements a 48–72 heures, l’adaptation posologique des aminosides selon les dosages seriques. Resultats Soixante-trois patients ont ete inclus et 67 antibiotiques « cibles » prescrits. La prevalence des patients sous antibiotiques etait de 24,4 %, celle sous antibiotiques « cibles » de 14,6. Le choix de la molecule etait conforme pour 67 % des prescriptions. Les posologies administrees etaient adequates dans 94 % des cas et les rythmes d’administration dans 97 % des cas. Le relais per os des que possible a ete realise pour la moitie des patients. Les durees de traitement ont ete respectees pour 94 % des prescriptions. La reevaluation des antibiotherapies a 48–72 heures a ete realisee dans 66 % des cas. L’adaptation posologique des aminosides, quand elle etait necessaire, a ete realisee une fois sur trois. Pour l’ensemble des criteres de qualite mesures, le niveau de conformite globale des prescriptions etait de 44 %. Conclusion Differentes actions pourraient ameliorer la qualite de l’antibiotherapie : diffusion large de protocoles, reevaluation systematique des traitements a 48–72 heures, formation des jeunes prescripteurs, renforcement du role des correspondants medicaux en antibiotherapie et des infectiologues, evaluation periodique de l’antibiotherapie.

Resume Propos Clostridium difficile est un bacille a Gram positif anaerobie sporule responsable de 15 a 25 % des cas de diarrhees postantibiotiques et de plus de 95 % des cas de colites pseudomembraneuses (CPM). Cet article resume les principales connaissances des infections a C. difficile et leur evolution recente. Actualites et points forts Depuis 2003, des epidemies de diarrhees severes a C. difficile ont ete decrites au Canada et aux Etats-Unis. Ces epidemies sont liees a l’emergence et a la dissemination d’un clone particulier de C. difficile appele 027 en reference a son profil par PCR-ribotypage ou North American pulsotype 1 (NAP1) en reference a son profil en electrophorese en champ pulse. Ce clone possede la particularite de produire in vitro des quantites plus importantes de toxines A et B et d’etre positif pour une troisieme toxine appelee toxine binaire. Ce clone a ete diffuse en Grande-Bretagne, en Belgique, aux Pays-Bas et plus recemment en France ou il a ete responsable d’epidemies, principalement dans la region Nord–Pas-de-Calais. Perspective et projets La surveillance systematique de l’incidence des infections a C. difficile dans les etablissements de sante devrait permettre de mieux apprehender l’evolution en France de cette pathologie.

Resume Objectifs Nous relatons les difficultes de la prise en charge d’une epidemie d’infection a Clostridium difficile de type 027 dans un etablissement de long et moyen sejour. Seize patients ont eu une ICD027, soit une incidence de 2,9/10 000 journees d’hospitalisation. 20 cas d’ICD027 ont ete identifies, incidence d’infections evaluee a 3,7/10 000 journees d’hospitalisation. Recommandations L’application des recommandations nationales a ete difficile a mettre en œuvre et a engendre un surcout majeur pour l’etablissement. Les recidives, meme apres un traitement bien conduit, etaient evaluees a 31 %. Ce type d’evenement necessite des reorganisations lourdes et la perennisation de ces mesures, dans un contexte ou Cd 027 est susceptible de s’installer, est tres « destabilisant ».

Central venous catheter (CVC)-related infections (CRIs) are a key target for infection control in intensive care units (ICUs). The aim of this study was to describe temporal trends of CRI incidence in a network of volunteer ICUs in Northern France. During a 4 month surveillance period each year, all CVCs in place for more than 48h were prospectively followed until removal or patient discharge. Standard clinical and microbiological criteria were used to define colonization and CRI. The standardized incidence ratio (SIR) was estimated by dividing the number of observed CRIs by the number of expected CRIs, which was computed using a logistic regression model including risk factors for CRI. CRI incidence and SIR were fed back to ICUs as a benchmark at the end of each period. From 2001 to 2005, 135 ICUs participated for at least one surveillance period. Overall, 11 703 CVC in 9182 patients (122 495 CVC-days) were included. CRI incidence was 2.8 per 1000 CVC-days. Among 35 ICUs that participated for three or more consecutive periods, CRI incidence decreased significantly by 58.6%. SIR also decreased significantly from the first to the third surveillance period in these ICUs. These results suggest that surveillance programmes have a significant impact on CRI risk in ICUs and remain an important strategy for combating nosocomial infections in these settings.

Objective: To determine the incidence of different modes of death in French paediatric intensive care units and to compare patients' characteristics, including a severity of illness score (Paediatric Risk of Mortality: PRISM score) and prior health status (Paediatric Overall Performance Category scale), according to the mode of death. Design: A 4-month prospective cohort study. Setting: Nine French multidisciplinary paediatric intensive care units. Patients: All patients who died in PICUs. except premature babies. Main results: Among 712 admissions, 13% patients died. Brain death was declared in 20%. failure of cardiopulmonary resuscitation occurred in 26%, do-not-resuscitate status was identified in 27%,and withdrawal of supportive therapy was noted in 27%, The PRISM score and the baseline Paediatric Overall Performance Category were not different between the four groups. Brain-dead patients were older than those in whom a do-not-resuscitate order and withdrawal of therapy were made (median age 81 vs 7 and 4 months). Conclusions: Decisions to limit or to withdraw supportive care were made for a majority of patients dying in French paediatric intensive care units. Chronic health evaluation and severity of illness index are not sufficient to describe dead-patient populations.

International audience; Preoperative hair removal has been used to prevent surgical site infections (SSIs) or to prevent hair from interfering with the incision site. We aimed to update the meta-analysis of published randomized controlled trials about hair removal for the prevention of SSIs, and conduct network meta-analyses to combine direct and indirect evidence and to compare chemical depilation with clipping. The PubMed, ScienceDirect and Cochrane databases were searched for randomized controlled trials analysing different hair removal techniques and no hair removal in similar groups. Paired and network meta-analyses were conducted. Two readers independently assessed the study limitations for each selected article according to the Grading of Recommendations Assessment, Development and Evaluation (GRADE) method. Nineteen studies met the inclusion criteria. No study compared clipping with chemical depilation. Network meta-analyses with shaving as the reference showed significantly fewer SSIs with clipping, chemical depilation, or no depilation [relative risk 0.55, 95% confidence interval 0.38-0.79; 0.60, 0.36-0.97; and 0.56, 0.34-0.96, respectively]. No significant difference was observed between the absence of depilation and chemical depilation or clipping (1.05, 0.55-2.00; 0.97, 0.51-1.82, respectively] or between chemical depilation and clipping (1.09, 0.59-2.01). This meta-analysis of 19 randomized controlled trials confirmed the absence of any benefit of depilation to prevent surgical site infection, and the higher risk of surgical site infection when shaving is used for depilation. Chemical depilation and clipping were compared for the first time. The risk of SSI seems to be similar with both methods.

Resume Objectif Decrire l'investigation d'une epidemie d'infections nosocomiales neonatales a Pseudomonas aeruginosa (PA). Methodes Dans le service de neonatalogie du groupe hospitalier Sud-Reunion, entre janvier 2003 et octobre 2005, nous avons denombre les infections a PA, mesure retrospectivement leur incidence et realise un audit des pratiques puis une enquete cas–temoins pour en rechercher les facteurs de risque. Resultats Sur 1432 sejours, 40 nouveau-nes (âge gestationnel median : 29 semaines, poids median : 1195 g) ont ete infectes, soit un taux d'attaque de 2,8 %. Entre janvier 2003 et janvier 2004, l'incidence de ces infections etait inferieure a 2 pour 1000 jours d'hospitalisation. Au dernier trimestre 2004, un pic d'incidence a 5,6 infections pour 1000 jours d'hospitalisation et la presence de PA dans 100 % des prelevements oculaires authentifiaient une epidemie, mais ce n'est que 3 mois plus tard et apres 3 nouveaux deces de grands prematures que des mesures de controle et un audit des pratiques permirent d'arreter la transmission. La letalite globale a ete de 25 %, et de 71 % dans les formes graves (septicemies ou pneumopathies). Le profil evolutif de l'epidemie etait en faveur d'une source hydrique commune. Deux facteurs de risque ont ete identifies : la ventilation mecanique au-dela de 4 jours (OR 3,3 ; IC 95 % : 1,3–8,4) et la grande prematurite (OR 2,7; IC 95 % : 1,0–7,7). Conclusion Ces resultats soulignent l'interet d'une etroite collaboration entre neonatologistes et hygienistes afin d'ameliorer les pratiques et la surveillance.

To evaluate the epidemiology of hospital admissions for varicella in children, a 1-year prospective multicentre study was done in Northern France in the pre-varicella vaccine era. The 405 children aged

International audience; We evaluated retrospectively the yield of stool culture (SC) depending on the length of hospitalization, and we characterized the patients missed by the 3-day rejection rule. SC detects bacterial enteric pathogens (Campylobacter spp., Salmonella enterica, Yersinia spp., Shigella spp.). During this 5-year study period, 13,039 SCs were requested, and 376 were positive (2.9%). The yield of SC dropped from 11.7% before 3 days of hospitalization to 0.7% after 3 days in children and 4.3% to 0.3% in adults. Finally, only 13 clinically relevant cases (0.2% of SC prescribed after 3 days) were undiagnosed by strict application of the 3-day rule. In conclusion, rejection of SC prescribed after 3 days of hospitalization allows to reduce workload by 37.8% for children and 65.7% for adults, representing a cost of €12,500 ($16,250) per year in our hospital.

Resume Objectif : Nous avons voulu identifier les facteurs de risque favorisant le portage de souches de H. influenzae productrices de β-lactamase et de pneumocoques de sensibilite diminuee a la penicilline (PSDP). Methode : En 1999, 2565 enfants tires au sort au sein de 66 creches collectives des Alpes-Maritimes, du Doubs et du Nord, ont ete inclus dans l’etude. Pour chacun d’eux, un prelevement nasopharynge a ete realise pour l’etude du portage, et deux questionnaires ont ete remplis, l’un par l’investigateur a partir des carnets de sante et prescriptions du trimestre precedent, l’autre par les parents. Resultats : 44,9 % des enfants sont porteurs de souches de H. influenzae dont 48,1 % productrices de β-lactamase et 49,2 % des enfants sont porteurs de pneumocoques dont 64,7 % de PSDP. En plus des variations liees aux differences departementales et saisonnieres, il est montre un âge ≥ 18 mois est associe a un portage accru de souches de H. influenzae et un âge Commentaires : Cette etude confirme l’usage important des antibiotiques dans cette population d’enfants et souligne l’association etroite entre les antecedents d’hospitalisation et/ou d’antibiotherapie, et le portage de souches PSDP ; la force de l’association augmente avec le nombre d’antibiotiques et le caractere recent de la prescription.

The increasing bacterial resistance to antibiotics has become a major public health concern bringing the threat of therapeutic impasses. In this context, control of the spread of highly-resistant bacteria emerging antibiotics (BHRe), such as glycopeptide-resistant enterococci (VRE) and Enterobacteriaceae producing carbapenemases (CPE), is based on a dual strategy of reducing the prescription of antibiotics to limit the pressure selection and preventing the spread from carriers. Prevention strategy is based on three different levels such as standard precautions for all patients with a particular focus on the management of excreta, and additional precautions for BHRe carriers. What makes it difficult is that carriage is usually completely asymptomatic, enterobacteria and enterococci are normal commensal of gut microbiota. Explosive dissemination of Enterobacteriaceae producing extended spectrum beta-lactamases in hospital and community heralds the emergence of CPE whose import by patients with a history of hospitalization in abroad may be the main source of spread in France.

Resume Les diarrhees aigues du nourrisson sont responsables de 10 % des hospitalisations a cet âge, avec de tres importantes variations du taux d'hospitalisation selon les centres. Objectifs. – Evaluer le traitement hospitalier des diarrhees aigues du nourrisson, et notamment l’impact de la presence permanente d’un pediatre et d’une structure d’hospitalisation de courte duree sur les modalites de rehydratation et les taux d’hospitalisation. Methodes. – L’etude prospective, multicentrique, a inclus durant deux periodes epidemiques d’infection a rotavirus 1 187 nourrissons, d’âge moyen 10 ± 6 mois, amenes a l’hopital pour une diarrhee aigue avec ou sans deshydratation. La premiere phase a concerne 12 centres hospitaliers volontaires. Dans la deuxieme phase, les modalites de rehydratation et d’hospitalisation ont ete comparees dans six centres hospitaliers (parmi les 12) ou les enfants etaient accueillis dans le service d’urgences ou de pediatrie sans la presence permanente d’un pediatre, et dans une unite d’urgences pediatriques ayant un pediatre en permanence sur place et dix lits d’hospitalisation de courte duree. Resultats. – Dans la premiere phase, 32 % des nourrissons etaient deshydrates et 6 % etaient choques a l’admission. Le taux d’hospitalisation etait de 81 %. Seuls 19 % des enfants deshydrates etaient rehydrates exclusivement par voie orale tandis que 19 % des normohydrates etaient perfuses. Dans la deuxieme phase, le taux global d’hospitalisation etait moindre dans l’unite d’urgences pediatriques : 42 % (soit 28 % d’hospitalisation de courte duree aux urgences + 14 % d’hospitalisation en pediatrie) contre 77 % dans les six centres hospitaliers. Dans l’unite d’urgences pediatriques, par comparaison aux six centres hospitaliers, le nombre de journees d’hospitalisation etait reduit de 65 %, le recours a la voie veineuse de 58 %, et le recours a des examens sanguins de 66 %. Conclusions. – Cette etude montre l’interet de la presence d’un pediatre et de lits d’hospitalisation de courte duree dans la prise en charge de cette affection frequente et couteuse.

Resume Objectif – Evaluer les moyens consacres a la lutte contre l'infection nosocomiale, et les politiques de maitrise de la diffusion des bacteries multiresistantes. Materiel et methodes – Des recommandations pour la maitrise de la diffusion des bacteries multi-resistantes ont ete elaborees. En les utilisant comme referentiel, une enquete multicentrique dans la region Nord–Pas-de-Calais a ete realisee. 62 etablissements (21 631 lits), ont ete inclus dans l'etude. Resultats – Six Unites d'Hygiene ont ete identifiees; 17 etablissements (27%) beneficiaient des prestations d'une unite operationnelle en hygiene directement ou par convention. Il a ete denombre 1 equivalent temps plein (ETP) medecin / 4 081 lits de court sejour, 1 ETP d'infirmiere hygieniste / 600 lits de court sejour. Un seul etablissement avait le ratio recommande de praticien hygieniste, et 12 celui d'infirmiere hygieniste. Le deficit etait de 21,7 ETP (medecins) et de 19 ETP (infirmieres hygienistes). 81% des etablissements declaraient l'existence de referents en hygiene. La proportion de chambres seules etait de 58%, celle des points d'eau equipes pour le lavage des mains au savon doux de 45% et de 38% pour le lavage anti-septique. La signalisation des patients porteurs de bacteries multiresistantes etait utilisee correctement dans 12,5%. Dans 55% des etablissements, aucune unite n'avait la totalite du materiel necessaire pour un isolement technique. 77% d'entre elles disaient modifier l'organisation des soins en presence de patients porteurs de bacteries multiresistantes. Des protocoles ecrits (lavage des mains, procedures d'isolement) etaient accessibles dans 71% des unites.

Resume Une etude d'incidence de Pseudomonas aeruginosa a ete menee en 1996 par 25 centres hospitaliers (13913 lits) de la region Nord-Pas-de-Calais. L'enquete s'est deroulee sur deux periodes trimestrielles durant lesquelles ces etablissements ont realise 230 018 entrees. L'incidence des patients infectes ou colonises par P. aeruginosa s'etablit a 9,4/1 000 admissions et rapportee a l'annee a 31,2/100 lits. Les incidences d'isolement calculees par discipline montrent des taux eleves pour les services de reanimation : 425,8 P. aeruginosa par an et pour 100 lits. Trois serotypes (O 11 , O 6 , O 4 ) representent plus de 40 % du recrutement. Les pourcentages de sensibilite aux antibiotiques (3 056 souches) sont de 66,5 % pour la ticarcilline, 84,7 % pour la ceftazidime, 82,2 % pour l'imipeneme, 73,3 % pour l'amikacine et 61,2 % pour la ciprofloxacine. Cette enquete confirme l'importance de P. aeruginosa dans les hopitaux de la region mais montre de tres grandes differences quant a la situation locale de chaque etablissement tant en termes d'incidence qu'en termes de niveau de resistance.

La diarrhee aigue est une affection frequente et couteuse. La deshydratation et les deces qu'elle entraine peuvent etre prevenus par l'utilisation des solutions de rehydratation orale (SRO). Objectifs. - Evaluer le traitement ambulatoire des diarrhees aigues du nourrisson, et notamment la prescription des SRO. Methodes. - L'etude prospective multicentrique a inclus, en janvier 1996, 326 nourrissons, d'âge moyen 10 ± 6 mois, examines a l'hopital pour une diarrhee aigue avec ou sans deshydratation. Resultats. - Avant l'admission, 81% des enfants avaient ete examines par au moins un medecin, et 89 % de de ces medecins avaient redige une prescription. Le medecin prescripteur etait un medecin generaliste dans 81 % des cas et un pediatre dans 15% des cas. Cette prescription comportait dans 35% des cas une SRO. La frequence de prescription des SRO n'etait pas differente selon l'âge de l'enfant. Les pediatres prescrivaient plus souvent une SRO que les generalistes (58% vs 29%; p < 0,001). Le taux d'echec de la prise de SRO apres une premiere prescription medicale etait de 25% (deux non-observances par les parents, dix refus de la SRO par l'enfant et huit arrets pour vomissements parmi 81 enfants). Un lait de regime (essentiellement un lait sans lactose) etait prescrit chez 46% des enfants, et l'observance etait de 82%. Au moins un medicament etait prescrit chez 94% des enfants, avec en moyenne 2,6 medicaments par enfant; un antibiotique etait prescrit dans 33 % des cas. Les parents emmenaient leur enfant a l'hopital sans consultation prealable (18%), sur leur initiative mais apres avoir consulte au moins un medecin au prealable (52 %), ou sur demande immediate du medecin apres examen de l'enfant (30%). Trente-trois pour cent etaient deshydrates; un enfant decedait et deux enfants presentaient des sequelles. Conclusions. - L'utilisation des SRO est insuffisante. La frequence et la morbidite de l'affection justifient des efforts d'information aupres des medecins et du public.

Resume L’evaluation du recrutement et de l’activite des services de reanimation pediatrique est devenue une necessite. Objectifs. — Il s'agit de decrire l’activite diagnostique et therapeutique par le systeme Omega, de rechercher une relation entre celte activite et les caracteristiques des patients, notamment la gravite appreciee par le score PRISM (Pediatric Risk of Mortality), et d’analyser la frequence et la nature des sejours en reanimation pouvant paraitre inappropries. Methodes. — Les auteurs ont mene une etude prospective dans neuf services de reanimation pediatrique francais, de decembre 1993 a avril 1994, en excluant les nouveau-nes prematures. Resultats. — Sept cent douze sejours ont ete retenus. Vingt-cinq pour cent des patients etaient des nouveau-nes, 30% des nourrissons de moins de 1 an et 45% avaient plus de 1 an. Vingt-deux pour cent avaient une affection chirurgicale, 10% etaient immunodeprimes et 45% des enfants (nouveau-nes exclus) avaient une affection chronique. La mediane de la duree de sejour etait de 4 jours (extremes: 1–155). L’activite moyenne en points Omega etait: Omega total = 92 ± 124, Omega/jour = 14 ± 9. Soixante-quatre pour cent des enfants etaient ventiles et 37% etaient ventiles plus de 2 jours: 42% avaient un catheter veineux central et 23% un catheter arteriel peripherique, La charge en soins etait plus importante chez les enfants decedes, les malades chirurgicaux, les nouveau-nes et les immunodeprimes. Le PRISM moyen etait de 9 ± 9. Le taux de mortalite en reanimation etait de 13%, Il existait une correlation entre d’une part la probabilite de deces mesuree par le PRISM et, d’autre part, l’Omega/jour et l’Omega total. La proportion de sejours paraissant inappropries etait de 7,6% et consommait 1.0% de l’activite Omega. Conclusions. — Cette etude etablit une base de references d’activite pour les services de reanimation pediatrique polyvalente. Elle demontre la relation entre gravite et charge en soins, et le faible taux de sejours inappropries par comparaison aux etudes nord-americaines ou hollandaises.

There are still many uncertainties regarding the benefit of public reporting on indicators of healthcare-associated infections. Publishing the results of performance indicators in healthcare facilities seems to stimulate quality insurance activity, but for the moment, without any demonstrated impact on the rates of healthcare-associated infections. Furthermore, it is not proved that public reporting of these results is beneficial. The French Society for Hospital Hygiene, in this synthesis, considered the advantages and drawbacks of public reporting of healthcare-associated infections. Beyond the difficult choice of indicators, public reporting of such indicators requires taking many precautions listed in this article, including a broad dialogue between the various stakeholders involved, a test period, and mandatory control of the produced indicator validity by health authorities.

Background:Patients with end-stage renal disease (ESRD) and latently infected with Mycobacterium tuberculosis (LTBI) are at higher risk to develop tuberculosis (TB) than healthy subjects. Interferon-gamma release assays (IGRAs) were reported to be more sensitive than tuberculin skin tests for the detection of infected individuals in dialysis patients.Methods:On 143 dialysis patients prospectively enrolled, we compared the results from the QuantiFERON®-TB Gold assay (QFT), to those of an IGRA in response to in vitro stimulation of circulating mononuclear cells with the mycobacterial latency antigen Heparin-Binding Haemagglutinin purified from Mycobacterium bovis BCG (native HBHA, nHBHA).Results:Seven patients had a past history of active TB and 1 had an undetermined result with both IGRAs. Among the other 135 patients, 94 had concordant results with the QFT and nHBHA-IGRA, 40.0% being negative and therefore not latently infected, and 29.6% being positive and thus LTBI. Discrepant results between these tests were found for 36 patients positive only with the nHBHA-IGRA and 5 only with the QFT.Conclusions:The nHBHA-IGRA is more sensitive than the QFT for the detection of LTBI dialysis patients, and follow-up of the patients will allow us to define the clinical significance of discrepant results between the nHBHA-IGRA and the QFT. © 2013 Dessein et al., SCOPUS: ar.j, info:eu-repo/semantics/published

The double-disk synergy test was compared to the Mastdiscs™ ID AmpC and ESßL method for detection of ESßL production in rectal swab.Two hundred and forty-nine rectal swabs were directly inoculated onto Mueller-Hinton plates and analyzed according to both methods.A total of 41 (16%) and 208 (84%) were positive and negative for ESßL, respectively. Twelve (29%) and 20 (49%) of the 41 rectal swabs positive for ESßL were detected after 24h of incubation with the double-disk synergy test and the Mastdiscs™ method, respectively (P=0.013). One hundred fifty-eight (76%) et 183 (88%) of the 208 rectal swabs were detected negative for ESßL after 24h of incubation with the double-disk synergy test and the Mastdiscs™ method, respectively (P0.001). Finally, 79 (32%) and 46 (18%) rectal swabs respectively inoculated according to the double-disk synergy test and the Mastdiscs™ method were inconclusive after 24h of incubation. The better performance of the Mastdiscs™ method was due to an easier detection of cephalosporinase producing bacteria.The Mastdiscs™ method is a simple phenotypic method that detects more easily ESßL and non-ESßL producing bacteria in rectal swab.

Resume L'hepatite B (HB) est la premiere maladie professionnelle en milieu hospitalier. Une enquete, par auto-questionnaire, realisee en mars 1990 au CHU de Tours, a estime la couverture vaccinale des personnels exposes a 73,8 %, precisant le statut vaccinal par categorie professionnelle. Elle a egalement precise la prevalence de contacts directs accidentels recents avec le sang : 8,7 %. Enfin, la prevalence des antecedents d'HB dans cette population a ete de 5,0 %.

Introduction L’utilisation des dispositifs intra-vasculaires (DIV) donne lieu a l’observation d’infections locales, generales et de bacteriemies responsables d’une importante morbi-mortalite en neonatalogie [1] , [2] . La pose de DIV concerne un nombre croissant d’enfants dans le bloc operatoire pediatrique. Dans le cadre general d’une evaluation de la dimension organisationnelle en anesthesie pediatrique en termes d’hygiene et de securite, cette enquete est centree sur les pratiques individuelles lors de la pose de catheters veineux centraux (CVC) et de PICC line avec comme objectifs principaux : de mesurer les ecarts avec les recommandations existantes et de re-sensibiliser les professionnels concernes a l’identification du risque infectieux lie a la pose de DIV et aux mesures recommandees de prevention. L’analyse des ecarts nous permet de definir des axes d’amelioration pour lesquels des actions pourront etre menees. Materiel et methodes Il s’agit d’une approche prospective, par observation directe de 50 poses de DIV au bloc operatoire pediatrique sur 10 semaines. Population etudiee : toutes les categories professionnelles consentantes et concernees par la pose de DIV. L’anonymat des personnes auditees etait garanti. Le referentiel est constitue par les protocoles valides de l’etablissement et les recommandations recentes [3] . L’analyse item par item a ete completee par la definition de conformites combinant certains items entre eux. La conformite « association produits » s’interesse a la coherence de gamme du couple « produit detersion/produit antisepsie (ATS) » utilise, associee au respect de la classe d’âge comme decrit dans les recommandations sur l’ATS en pediatrie [3] . Le score attendu est de 100 %. Resultats Au total 33 CVC (66 %) et 17 PICC (34 %) ont ete poses. L’observance de l’hygiene des mains (HDM) avant la preparation cutanee etait de 72 %. Le taux de conformite de « l’association produits » etait de 84 %. Le taux de conformite globale de la pratique d’une detersion suivie d’une desinfection large du site d’insertion etait de 56 %. Le type de pansement etait conforme dans 42 % des poses. Dans plus d’un tiers des cas, le pansement n’integre pas le premier raccord du catheter. Une seule observation (2 %) presente la tracabilite des prescriptions pour le suivi apres pose. Discussion Certaines recommandations importantes en hygiene sont integrees dans les pratiques quotidiennes des equipes du bloc pediatrique. Sur la base de l’analyse des ecarts et de leurs causes nous avons degage plusieurs orientations pour la poursuite de l’amelioration des pratiques notamment par le rappel des recommandations existantes et de la reglementation sur l’HDM, la tracabilite des soins et les etapes a respecter pour l’ATS cutanee ; l’adaptation des protocoles a la specificite pediatrique notamment le choix de l’ATS selon la classe d’âge et la coherence de gamme du couple « produit detersion/produit ATS » ; l’analyse du management en particulier sur la qualite des pansements, l’utilisation d’outils d’aide a l’observance des mesures de prevention comme la Check-list de pose d’un CVC ou autre dispositif vasculaire et la mise en œuvre de la tracabilite des soins.

In 2006 and 2007, a large outbreak of Clostridium difficile infections (CDIs) with PCR-ribotype 027 was identified in northern France. Overall, 38 healthcare facilities notified 529 CDIs over a 22-month period, including 281 laboratory-confirmed CDI 027 and 248 non-confirmed CDI 027 cases (incidence rate per 10,000 elective bed days: 1.63, range: 0.07 to 7.94). The cases occurred mainly in long-term care hospital facilities and nursing homes, near the border between France and Belgium. An active surveillance and prevention campaign was launched at the first epidemic peak including hygiene precautions for healthcare professionals, which supported healthcare facilities to improve care organisation. The outbreak was controlled at the end of 2007, but sporadic cases were identified until the end of 2009. A bundle of appropriate control measures may halt the spread of such outbreaks, provided that substantial human resources and financial support are available.

Introduction: The multiple organ dysfunction syndrome (MODS) is a major cause of death for patients admitted to pediatric intensive care units (PICU). The Pediatric Logistic Organ Dysfunction (PELOD) score has been validated in order to describe and quantify the severity of organ dysfunction (OD). There are several physiological differences between neonates and older children. The objective of the study was to determine whether there are differences in incidence of ODs and mortality rate between full-term neonates (age

Resume Six enquetes de prevalence concernant l'antibiotherapie ont ete realisees entre 1978 et 1990 au CHU de Tours. Ces etudes concernaient l'ensemble des patients hospitalises. En moyenne d'une annee a l'autre, 30 % des patients hospitalises ont recu au moins un antibiotique. Chez un tiers de ceux-ci, les antibiotiques etaient prescrits a titre prophylactique. En 1990, la prevalence de l'antibiotherapie etait de 35 % de la prescription globale d'antibiotiques. Les Penicillines G et V predominaient nettement en 1978 (25 %). La prevalence de leur utilisation a considerablement diminue pour ne plus representer en 1990 que 5 % des prescriptions d'antibiotiques. En 1990, l'amoxicilline et l'ampicilline representaient 11 %, et l'association amoxicilline-acide clavulanique, apparue en 1984, 16 % des prescriptions. Les fluoroquinolones, apparues en 1985, occupent en 1990 la troisieme place des molecules prescrites (9 %). L'utilisation des cephalosporines etait en legere augmentation, essentiellement due aux cephalosporines de troisieme generation. Certaines molecules, comme le metronidazole et les aminosides, ont vu leur place considerablement diminuer depuis 1985. Malgre les limites d'interpretation d'ordre methodologique, les resultats de ces enquetes permettent d'etablir des tendances concernant l'utilisation des antibiotiques, et peuvent contribuer a une politique de bonne utilisation des antibiotiques.

Resume Une etude a ete realisee dans le but d'evaluer la situation de la consommation d'antibiotiques de 1987 a 1990 au CHR de Tours. L'etude concernait la situation de l'antibiotherapie dans le budget des medicaments et la place relative de chaque classe d'antibiotiques. Les donnees ont ete agregees en grammes, francs et en journees-traitement. En 1990, les antibiotiques representaient 18,8 % du budget medicament. Cette part allait croissante depuis 1987, avant une legere diminution en 1990. Les fluoroquinolones, en augmentation constante de 1987 a 1989 (+150 %) et les cephalosporines de troisieme generation predominaient en 1990, avec une part respective de 14,7 et 14,4 % du budget total des antibiotiques. La part des aminosides ne cesse de decroitre (8,79% du budget antibiotique en 1990). La prescription des cephalosporines de troisieme generation parenterales reste relativement stable ; la place des formes orales se developpe rapidement et merite d'etre suivie (22,4 % des prescriptions du groupe en 1990). Les aminopenicillines restent tres utilisees, et la part de l'association amoxicilline-acide clavulanique n'a cesse de croitre de 1987 a 1990 (+180 %), au detriment des autres aminopenicillines, dans l'analyse en francs. Les glycopeptides prennent une place de plus en plus importante (10 % du budget antibiotique en 1990). La consommation des macrolides est stable. Parmi les autres molecules ou familles representant des postes budgetaires plus modestes, on notait une diminution reguliere de la consommation en imidazoles depuis 1987. Ces resultats de tendance encouragent l'usage de guides de prescription ou de protocoles evalues et l'etude longitudinale de consommation de quelques familles ou molecules, dans le but de mieux approcher cette consommation dans le cadre d'une politique de bonne utilisation des antibiotiques.

Background Adequate glycopeptide use is necessary to optimize efficacy and minimize adverse events and selection of drug resistant microorganisms. Methods We performed a multicenter prospective observational study of glycopeptides indications based on modified HICPAC criteria, and quality of prescription based on French consensus recommendations. We assessed the adequacy of indications, loading dose, maintenance dose, administration frequency, serum monitoring and its consequences. Results We evaluated 117 curative adult prescriptions from nine hospitals. Glycopeptide indications were adequate in 71% with either a clinical (55%) or microbiological justification (27%). The indication was more frequently adequate for vancomycin than for teicoplanin (77% vs. 60%, p = 0.04). De-escalation therapy was performed in 48% of all documented infections and 78% of documented infections with betalactam susceptible strains. We observed high rates of correct maintenance dose (90%), administration frequency (88%), and serum monitoring (74%). Loading doses (43%), adequate serum monitoring timing (43%), correction of dose because of inadequate levels (32%) were less adequate. Overall, only three (3%) treatments were adequate for all evaluated items. Conclusion Glycopeptide indications were adequate. However, the quality of prescription can be improved and should focus on loading doses and serum monitoring.

Clostridium difficile is an anaerobic gram positive, spore-forming bacterium which is responsible for 15-25% of antibiotic-associated diarrhea and for more than 95% of pseudomembranous colitis (PMC). This paper will review the main knowledge on C. difficile-associated infections and their recent evolution.Since 2003, outbreaks of severe C. difficile-associated diarrhea (CDAD) have been increasingly reported in Canada and the United States. This trend is assumed to be associated with the rapid emergence and spread of a specific clone of C. difficile belonging to PCR-ribotype 027 or North American Pulsotype 1, pulsotype (NAP1). This clone is characterized by the overproduction of toxins A and B and is positive for a third toxin named binary toxin. This clone has spread in UK, in Belgium, in the Netherlands, and, more recently, in France where it has been responsible for large outbreaks mainly in northern France.A systematic reporting of C. difficile incidence by health facilities should enable a better assessment of this pathology in France.

Simple clinical markers have poor sensitivity; specificity and predictive value in both infants and adults when predicting the success of weaning from mechanical ventilation. Recently, multi-parametric indices, such as the CROP (Compliance-Respiratory Rate-Oxygenation-Pressure) and the RSB (Rapid-Shallow-Breathing) have been used in adults and subsequently in children. The aim of this study was to test the value of the pediatric CROP and RSB (CROPp, RSBp) and the accuracy of a simplified pediatric CROP (CROPpS) that does not require an arterial blood gas sample.This prospective study was conducted in a pediatric ICU which does not admit neonates. All infants were intubated and ventilated at the time of entry. Spontaneous tidal volume and maximal negative inspiratory pressure, that are required to assess and calculate the indices, were measured using a Fleish pneumotachograph and a unidirectional valve. The other parameters were recorded or calculated. A maximum 4 hour-duration trial of spontaneous ventilation was then performed. Weaning failure was defined as the requirement of re-ventilation within 48 hours of extubation. The discriminant power of CROPp and RSBp was determined by calculating the area under the receiver operating characteristic (ROC) curve. The best cut-off value of the CROPpS was determined by chi2 optimisation.39 children (20 males) were included in the trial. They had a median age of 3.2 years and a median duration of mechanical ventilation of 1.3 days. 89.7% of children were successfully weaned of mechanical ventilation. Sensitivity of CROP, and RSB, was 97% and 94%, specificity was 0% and 0%, positive predictive value was 89% and 89%, and negative predictive value was 0% and 0% respectively; the area under the ROC curve was 0.57 and 0.74. The CROP,S was found to be as accurate as the CROP, index using the same cut-off value. Comparison of the 2 groups (success, failure) revealed a significant difference in duration of ventilation (longer in the failure group).Even though they correctly classified 87% and 85% of patients respectively, the CROPp and RSBp are not good predictors of weaning from mechanical ventilation as the area under the ROC curve is less than 0.80. Other indices need to be evaluated.

To investigate an outbreak of Pseudomonas aeruginosa (PA) hospital-acquired infections in neonates.Incidences were assessed retrospectively in the neonatal care units of the Groupe Hospitalier Sud-Réunion, from January 2003 to September 2005. Environmental survey, audit of health care workers and case-control study were performed to reinforce staff training and to determine risk factors.Of 1432 neonates, 40 were infected (median gestational age: 29 weeks, median birth weight: 1195 g), accounting for an attack rate of 2.8%. Between January 2003 and January 2004, incidence rates were less than 2 infections per 1000 hospitalisation days. In the last trimester of year 2004, the incidence rose to 5.6 infections per 1000 hospitalisation days and PA was found in all ocular swabs, leading to diagnose an epidemic. However, it was only 3 months later, after 3 new deaths of very preterm neonates, that the implementation of control measures and an audit of health care practices focused on water utilisation ruled out the outbreak. The overall fatality rate was 25%, and of 71% in severe diseases (septicemia or pneumonia). The epidemic pattern argued for a common unique source. Two risk factors were identified by logistic regression: exposure to mechanical ventilation beyond 4 days (OR 3.3; CI 95%: 1.3-8.4) and very preterm birth (OR 2.7; CI 95%: 1.0-7.7).Our findings highlight the need for a close collaboration between neonatologists and hygienists to improve health care practices and surveillance.