p - CCLIN Est
Transcription
p - CCLIN Est
Prévention des infections liées aux CVC en réanimation Journée scientifique du CClin-Est Dijon, 14 mars 2013 Dr Rémi BRUYERE Chef de clinique - Assistant Service de Réanimation Médicale - CHU de Dijon SAD 13,1% Figure 2 Répartition des patients selon l’l’exposition aux dispositifs invasifs SAD + CVC 10,5% SAD + INT 11,9% SAD + CVC + INT 49,2% CVC + INT 2,4% Les cathéters, c’est fréquent ? Aucun 10,3% INT 1,1% Aucun INT CVC veineu 10,3% = cathéter 1,1% INT = intubation SAD = sonde urinaire à demeure INT = intubation CVC = cathéter veineux central Tableau 17 Exposition des patients aux dispositifs invasifs Exposition des patients aux dispositifs Dispositifinvasifs invasif Patients exposés n n’ Intubation 25 648 16 548 Cathéter veineux central 25 562 16 181 Sonde urinaire 24 931 21 690 n : nombre de patients dont on sait s’ils sont exposés ou non n’ : nombre de patients exposés au risque Patients exposés n n’ Intubation 25 648 16 548 Cathéter veineux central 25 562 16 181 Tableau 18 Sonde urinaire 24 931 21 690 Exposition aux dispositifs invasifs n : nombre de patients dont on saitExposition s’ils sont en joursexposés ou non n moy. (± ds) Durée de séjour 25 685 11,6 (14,1) n’ : nombre de patients exposés au risque Durée d’intubation 16 548 10,9 (15,1) Dispositif invasif CVC 1,6% SAD + CVC + INT 49,2% SAD = sonde urinaire à demeure Tableau 17 SAD 13,1% CVC 1,6% Durée de cathétérisme Durée de sondage 16 180 21 690 12,2 11,3 (13,1) (13,4) % 64,5 63,3 87,0 % 64,5 63,3 87,0 min. 3 1 1 1 P25 4 3 5 4 méd. 7 6 8 7 P75 13 13 15 13 Données REA-Raisin 2010 max 3 3 2 2 Réintubations Epidémiologie Parmi les 16 424 patients (sur les 16 548 patients intubés) pour lesquels l’information a été recueillie : - 14 230 (86,7 %) n’ont jamais subi de réintubation, - 2 192 (13,3 %) ont subi 1 ou plusieurs réintubations. Le délai moyen de 1ère réintubation est de 10,9 jours +/- 10 (médiane à 8 j). ème - ILC = 3 cause d’IN en réanimation Cathétérisme veineux central - ILC concerne à 10 % des veineux centraux Un patient peut être porteur5 de plus d’un cathéter lors de cathéters son séjour. On observe 19 773 CVC parmi les 16 181 patients avec un cathétérisme soit un ratio de 1,22 CVC / patient : - patients avec 1 CVC : 84,3 %, - patients avec 2 CVC : 11,4 %, - patients avec 3 CVC et + : 4,3 %. - Variable selon le germe concerné Tableau Tableau 20 Répartition et durée de maintien des CVC selon le site de pose Site de pose Sous-clavier Jugulaire interne Fémoral Autre Total Tableau 21 Devenir du cathéter veineux central Répartition des CVC % n 9 185 47,1 6 441 33,1 3 801 19,5 61 0,3 19 488 100,0 n 9 184 6 441 3 801 61 19 772 Durée de maintien du CVC moy. (± ds) 11,5 (9,3) 9,3 (7,8) 8,8 (6,9) 8,7 (7,7) 10,2 (8,5) méd. 9 7 7 6 8 Données REA-Raisin 2010 PREVENTION Facteurs de risque - Prévention Liés au cathéter Liés aux patients et aux soins • Matériau • Site de pose • Technique de pose • Expérience de l’opérateur • Nombre de voies • Durée de cathétérisme • Asepsie lors de la pose et des soins • Pansements • Echange sur guide • Personnel • Protocoles • Surveillance Asepsie lors de la pose du cathéter En Urgence vs Asepsie Chirurgicale Pose en urgence (n=167) Asepsie chirurgicale (n=178) p 67 (2-100) 70 (2-100) 0,3 41 34 0,24 123 (74) 135 (76) 0,5 Colonisation, % 7,2 2,3 0,04 Bactériémie, n 67 (2-100) 1 0,5 0,08 Durée de cathétérisme, j (extr) Multi-lumière (%) Voie sous-clavière, n (%) Densité d’incidence / 1000 j de CVC 0,02 Raad et al, ICHE 1994;15:231-36 Solution antiseptique Groupe Chlorhexidine Colonisation Groupe Povidone Iodine ILC BLC Mimoz et al. CCM 1996 hexidine gluconate in preventing catheter-related is likely to be clinically important. bloodstream infection in patients who are catheterized Two additional issues should be considered regardfor an average of longer than 10 days or the potential ing the use of chlorhexidine gluconate for catheter-site additional benefit of chlorhexidine gluconate when an care. Hypersensitivity reactions have been reported with antimicrobial catheter is used (40, 41). use of central venous catheters impregnated by chlorNathorn Chaiyakunapruk, PharmD,and PhD;with Daviduse L. Veenstra, PharmD, PhD; Benjamin A. Lipsky, MD; and Sanjayaccumulation Saint, MD, MPHof studies Considering the progressive hexidine-silver sulfadiazine of chlorhexiover the past decade supporting the efficacy of chlordine gluconate for bathing (34 –38). No hypersensitivity Purpose: Bloodstream infections related to use of catheters, parData Synthesis: Eight studies involving a total of 4143 cath hexidine gluconate in reducing catheter colonization, reactions were reported by the studies in our meta-analticularly central-line catheters, are an important cause of patient ters met the inclusion criteria. All studies were conducted in Chlorhexidine Compared with Povidone-Iodine Solution for Vascular Catheter–Site Care: A Meta-Analysis morbidity, mortality, and increased health care costs. This study evaluated the efficacy of skin disinfection with chlorhexidine gluFigure compared 2. Analysis of catheter-related bloodstream infection in conate with povidone-iodine solution in preventing povidone-iodine solutionsinfection. for care of vascular catheter sites. catheter-related bloodstream Data Sources: Multiple computerized databases (1966 to 2001), reference lists of identified articles, and queries of principal investigators and antiseptic manufacturers. Study Selection: Randomized, controlled trials comparing chlorhexidine gluconate with povidone-iodine solutions for cathetersite care. Data Extraction: Using a standardized form, two reviewers abstracted data on study design, patient population, intervention, and incidence of catheter-related bloodstream infection from all included studies. I ntravascular catheters are commonly used in caring for hospitalized patients but can lead to serious infectious complications (1). Catheter-related bloodstream infection is associated with increased morbidity, mortality, length of hospitalization, and medical costs (2– 6). Use hospital setting, and various catheter types were used. The sum mary risk ratio for catheter-related bloodstream infection was 0.4 studies comparing chlorhexidine gluconate and (95% CI, 0.28 to 0.88) in patients whose catheter sites we disinfected with chlorhexidine gluconate instead of povidon iodine. Among patients with a central vascular catheter, chlorhex dine gluconate reduced the risk for catheter-related bloodstrea infection by 49% (risk ratio, 0.51 [CI, 0.27 to 0.97]). Conclusions: These results suggest that incidence of bloo stream infections is significantly reduced in patients with centr vascular lines who receive chlorhexidine gluconate versus pov done-iodine for insertion-site skin disinfection. Use of chlorhex dine gluconate is a simple and effective means of reducing va cular catheter–related infections. Ann Intern Med. 2002;136:792-801. For author affiliations, see end of text. www.annals. We sought to aid clinical decision making by eva uating the effectiveness of chlorhexidine gluconate ve sus povidone-iodine as a skin disinfectant for cathete site care. We performed a meta-analysis of all availab Ann Intern Med 2002;136:792-801 published and unpublished studies comparing chlo The diamond indicates the summary risk ratio and 95% CI. Studies are ordered chronologically. The size of squares is proportional to the reciprocal of Alcoholic povidone-iodine to prevent central venous catheter colonization: A randomized unit-crossover study* Jean-Jacques Parienti, MD, DTM&H; Damien du Cheyron, MD; Michel Ramakers, MD; Brigitte Malbruny, MD; Roland Leclercq, MD; Xavier Le Coutour, MD; Pierre Charbonneau, MD; for Members of the NACRE Study Group considered for the main analys parison of the 57 pairs of CVCs Objective: To compare effectiveness in preventing central ve- significantly lower in the alcoholic povidone-iodine solution protocol Aqueous Alcoholic tients risk, in both protocols is pres nous catheter colonization and infection of two protocols of than in the aqueous povidone-iodine solution protocol (relative Povidone-Iodine Povidone-Iodine cutaneous antisepsis using povidone-iodine solution in combina- 0.38; 95% confidence interval, 0.22– 0.65, p < .001), and Table so was3.the Characteristics (n ! 117) (n ! 106) tion with ethanol or water. incidence of catheter-related infection (relative risk, 0.34; 95% conThirty-one pairs and one pair Design: Randomized trial. fidence interval, 0.13– 0.91, p < .04). Catheter-related bacteremia respectively, to negative and posi No. of days catheter in place, mean (SD) 9.0 (4.4) 8.7 (4.8) Setting: Medical intensive Internal care department in a university hospital. were similar in both protocols. After adjusting for other risk factors, jugular vein 9.2 (3.3) 7.6 (3.6) colonization. Twenty-one and fo Patients: Consecutive patients requiring Subclavian veincentral venous catheter in time to central venous 8.7 (4.6)catheter colonization 9.3 was (5.1)significantly werelonger only positive in the aqueo two similar 11-bed units fromFemoral Januaryvein 1, 2001, to January 1, 2002. in the alcoholic solution ratio, 0.3; 95% confidence 9.5 (4.7) (adjusted hazards8.3 (5.2) and alcoholic PVP-I protocols, Site of central venous catheter Interventions: Alcoholic povidone-iodine solution protocol was interval, 0.2– 0.6, p < .001). Based on a subgroup of 114 patients (57 The number of conflicti vein the study began. in each protocol), 24 analysis (20.5) of 57 pairs of 28 central (26.4) venoustively. randomly assigned to one Internal of two jugular units when catheters was statistically significant (chi-s Subclavian vein 58 (49.6) 59 (55.7) Every 3 months the alcoholic protocol was switched from one unit matched for age, duration, and site of insertion found similar results Femoral vein 35 (29.9) 19 (17.9) p" to the other. Depending on the unit and the time the patient was regarding the superiority of alcoholic povidone-iodine 11.5, solution in .002), and the use of a Number of lumina PVP-I was associated with a redu admitted, catheters were inserted and cared for with 10% aque- preventing central 1 10 venous (8.5) catheter colonization 7 (6.6)and infection. for CVC colonization in alcoho ous povidone-iodine solution Conclusions:37The(31.6) use of alcoholic povidone-iodine 2 or 5% povidone-iodine solution 70% 34 (32.1) for skin disinfecparedinfecwith aqueous PVP-I (od ethanol-based combination.3 tion reduced the 42incidence (35.9) of catheter colonization 42 (39.6) and related 23 (21.7) disinfection Measurements and Main4Results: Rates of catheter colonization, tion compared28with(24) aqueous 10% povidone-iodine 0.2; 95%inCI, 0.1– 0.6). Use of catheter for administration of antibiotics catheter-related bacteremia, and catheter-related infection were an adult intensive care unit. (Crit Care Med 2004; 32:708–713) Fifty-five pairs and no pair y Yes 223 catheters were included in an (59.8) 55 central (51.9) venous catheter; alcohol-based antiseptic; compared in the two protocols; KEY WORDS: 70 negative and positive CVC-relate 47 (40.2) 51 (48.1) prevention intent-to-treat analysis. TheNo incidence of catheter colonization was nosocomial infection; tion, respectively. Ten and two pa Location of insertion and care Unit A 57 (48.7) 63 (59.4) positive in the aqueous PVP-I a Unit B 60 (51.3) 43 (40.6) holic PVP-I protocols, respectivel PVP-I in a osocomial infection remains a causes of central venous catheter (CVC)- Viatris, Merignac, France) with 5%the number of conflicting pairs No catheter characteristic between protocolscutaneous (p # .05). Results expressed inaqueous numbersolution (Beta70% are ethanol-based frequent complication in critrelated differed infection (3). Efficient tistically significant (chi-square (percent). ically ill patients, representing antisepsis before CVC insertion and follow- dine alcoolique, Viatris) for the prevention of " .05), and the use of alcoholic P a major source of morbidity, ing hubline manipulations is therefore es- catheter colonization and catheter-related infection. Alcoholic PVP-I, a recentlyassociated developed with a reduced risk f Table 223 (4). catheters inserted during the study mortality, and medical costs (1,2.2).Outcome Cutane- of the sential Aqueous 2% chlorhexidine so-period disinfectant, was approved for clinical use infection in related compared with ous microorganisms are predominant lution has significantly reduced nosocomial this setting. The study was approved by the PVP-I (odds ratio, 0.2; 95% CI, 0 Aqueous catheter-related Alcoholic infections (5) compared local ethics committee andp was conducted in Povidone-Iodine Povidone-Iodine Relative Risk CVCs were removed more fr with 10% aqueous polyvidone iodine (PVP- two similar 11-bed adult medical Parienti intensiveet al. CCM 2004 (n ! 106) (n ! 117) (95% CI) Valuea because they were not needed wh Table 1. Characteristics of the 223 catheters inserted during the study period N Characteristics (n ! 117) (n ! 106) Thirty-one pairs and one pair yielded, respectively, to negative and positive CVC colonization. Twenty-one and four pairs were only positive in the aqueous PVP-I and alcoholic PVP-I protocols, respectively. The number of conflicting pairs was statistically significant (chi-square ! 11.5, p " .002), and the use of alcoholic PVP-I was associated with a reduced risk for CVC colonization in alcoholic compared with aqueous PVP-I (odds ratio, 0.2; 95% CI, 0.1– 0.6). Fifty-five pairs and no pair yielded to negative and positive CVC-related infection, respectively. Ten and two pairs were positive in the aqueous PVP-I and alcoholic PVP-I protocols, respectively. Again, the number of conflicting pairs was stasignificantly lower in the alcoholic povidone-iodine solution protocol tistically significant (chi-square ! 4.1, p in the aqueous povidone-iodine solution protocol (relative risk, "than .05), and the use of alcoholic PVP-I was associated with a reduced risk for CVC0.38; 95% confidence interval, 0.22– 0.65, p < .001), and so was the related infection compared with aqueous incidence of catheter-related infection (relative risk, 0.34; 95% conPVP-I (odds ratio, 0.2; 95% CI, 0.0 – 0.9). fidence interval, p < .04). Catheter-related bacteremia CVCs were removed0.13– more0.91, frequently because they wereinnot needed when alcowere similar both protocols. After adjusting for other risk factors, holic was used (chi-square ! 5.0,colonization was significantly longer timePVP-I to central venous catheter p " .05). Alcoholic povidone-iodine to prevent central venous catheter colonization: A randomized unit-crossover study* No. of days catheter in place, mean (SD) Internal jugular vein Subclavian vein Femoral vein Site of central venous catheter Internal jugular vein Subclavian vein Femoral vein Number of lumina 1 2 3 4 Use of catheter for administration of antibiotics Yes No Location of insertion and care Unit A Unit B 9.0 (4.4) 9.2 (3.3) 8.7 (4.6) 9.5 (4.7) 8.7 (4.8) 7.6 (3.6) 9.3 (5.1) 8.3 (5.2) 24 (20.5) 58 (49.6) 35 (29.9) 28 (26.4) 59 (55.7) 19 (17.9) 10 (8.5) 7 (6.6) Jean-Jacques Parienti, MD, DTM&H; du Cheyron, MD; Michel Ramakers, MD; 37 (31.6) Damien 34 (32.1) 42 (35.9) 42 (39.6) Brigitte Malbruny, MD; Roland Leclercq, MD; 23Xavier 28 (24) (21.7) Le Coutour, MD; Pierre Charbonneau, MD; for Members of the NACRE Study Group 70 (59.8) 55 (51.9) 47 (40.2) 51 (48.1) 57 (48.7) 60 (51.3) 63 (59.4) 43 (40.6) No catheter characteristic differed between protocols (p # .05). Results are expressed in number Objective: To compare effectiveness in preventing central ve(percent). nous catheter colonization and infection of two protocols of cutaneous using povidone-iodine solution in combinaTable 2. Outcome antisepsis of the 223 catheters inserted during the study period tion with ethanol or water. Alcoholic Aqueous Design: RandomizedPovidone-Iodine trial. Povidone-Iodine Relative Risk p (n !care 106) department (n ! 117)in a university (95% CI) hospital. Valuea Setting: Medical intensive Patients: Consecutive patients central venous catheter in Catheter-tip colonization 14 (13.2) requiring 41 (35.0) 0.38 (0.22–0.65) ".001 Gram 5 (4.7) (18.8) to January 1, 2002. .002 two positive similar 11-bed units from January 1,222001, in the alcoholic solution (adjusted hazards ratio, 0.3; 95% confidence Staphylococcus epidermidis 5 (4.7) 15 (12.8) .04 Interventions: interval, 0.2– 0.6, p < .001). Based on a subgroup of 114 patients (57 Staphylococcus aureus Alcoholic0 povidone-iodine 3 (2.6) solution protocol was DISCUSSION Enterococcus 0 3 (2.6) randomly assigned to one of two units when the study began. in each protocol), analysis of 57 pairs of central venous catheters Corynebacterium 0 1 (0.9) Gram negative 9 (7.7) (16) switched from one unit .11 Every 3 months the alcoholic protocol 19 was matched foralcoholic age, duration, and ofsite of insertion found similar results The use of PVP-I instead Escherichia coli 0 7 (6.0) .02 aqueous PVP-I for insertion site disinfectoKlebsiella the other. Depending on4 (3.4) the unit and the time the patient was regarding the superiority of alcoholic povidone-iodine solution in pneumoniae 5 (4.3) tion and catheter care procedures markProteus mirabilis (3.4) (2.6) admitted, catheters were 4inserted and3 cared for with 10% aque- edly preventing central colonization and infection. reduced the risk of venous catheter catheter coloniPseudomonas aeruginosa 1 (0.9) 3 (2.6) ous povidone-iodine solution or 5% povidone-iodine solution 70% zationConclusions: use of in alcoholic povidone-iodine for skin disinfecEnterobacter cloacae 0 1 (0.9) and related The infections this Catheter-related infection 5 (4.7) 16 (13.7) 0.34 (0.13–0.91) .04 study. The benefitthe wasincidence most apparent as ethanol-based combination. tion reduced of catheter colonization and related infecBacteremia 1 (0.9) 4 (3.4) regards Gram-positive bacteria. The use S. Measurements epidermidis 0 Results: Rates 2 (1.7)of catheter colonization, and Main tion compared with aqueous 10% povidone-iodine disinfection in of alcoholic PVP-I also appeared to be S. aureus 0 1 (0.9) catheter-related bacteremia, and catheter-related infection were an adult intensive care unit. (Crit Care Med 2004; 32:708–713) independently associated with longer colK. pneumoniae 1 (0.9) 1 (0.9) Clinical infectionin the two protocols; 4 (3.8) 223 catheters 12 (10.3) were included in .09 catheter survival. To our : alcohol-based antiseptic; central venous catheter; compared an onization-free KEY WORDS knowledge, only infection; one previous study (10) Figure 1. Kaplan-Meier estimates of time to catheter-tip colonization according to alcoholic povidonenosocomial prevention intent-to-treat analysis. The incidence of catheter colonization was CI, confidence interval. reported a significant CVC iodinereduction or aqueousofpovidone-iodine protocols. Time to catheter-tip colonization was significantly colonization associated with use ofpovidone-iodine protocol (p ! .005 by the log-rank test). The time to longer in thethe alcoholic iodine-based disinfectant alcoholcolonization in remained statistically longer in the alcoholic povidone-iodine protocol (adjusted hazards iodine) water-based 0.3;vs. 95% confidence interval, 0.2– 0.6; p ! .001) in the Cox model after adjusting for site of onization-free catheter survival was sig- burning sensation in one conscious pa- based (tincture of ratio, Viatris, Merignac, France) with 5% PVP-I a infection remains a causes of central venous insertion vs. subclavian and femoral vs. subclavian), age, andin immunodepression status. (PVP-I)catheter solutions (CVC)(9% vs.(jugular 14% respecnificantly higher osocomial during the alcoholic tient, but ablation of the catheter because 70% aqueous solution (Betafrequentlog-rank complication crit- problems related never infection (3). tively, Efficient cutaneous p " .001), but this study wasethanol-based obPVP-I protocol (chi-square test, of in tolerance occurred effect of the disinfecdine alcoolique, Viatris) for the prevention of 7.8; p ! .005). ically ill patients, representing during the study. In the aqueous antisepsis beforePVP-I CVC servational insertion and andthefollowin the In the Cox multivariable analysis, al- protocol, one death was a consequence of tant’s choice disappeared It is noteworthy that alcoholic PVP-I tained release ofiniodine and reduces skin catheter colonization and catheter-related a major source of morbidity, ing hubline manipulations is therefore escoholic PVP-I significantly and indepen- bloodstream CVC-related bacteremia multivariate analysis. reduced thefection. risk of catheter colonization irritationdeveloped (17). Alcoholic PVP-I, a recently mortality, and costscol(1, 2).(methicillin-resistant Cutane- sentialStaphylococcus (4). Aqueous 2%The chlorhexidine so-of catheterpathogens, overall incidence coldently prolonged the medical time to catheter by Gram-positive which is an for Although not in statistically significant, disinfectant, was approved clinical use our increasing study (29.0cause per 1000 onization (adjusted hazards ratio, aureus endocarditis). ous microorganisms are 0.3; predominant lution has significantlyonization reducedinnosocomial of CVC-related blood- duration of catheter insertion was shorter this setting. The study was approved by the very similar that prevalence 95% CI, 0.2– 0.6; p " .001) compared Matched Case-Control Subgroup catheter-days) wasstream infectionto(12): rates of and number of subclavian sites higher in catheter-related infections (5) compared For example, a rate local ethics committee and conducted in protocol than in the with assignment to aqueous PVP-I. Analysis. Matching was possible for 114 previously reported. 28% during 1986 –1989 and 36% during was the alcoholic PVP-I with 10% aqueousin-polyvidone iodine (PVPof 31 per 1000 catheter-days with PVP-I Adverse Events and Mortality Rate. CVCs in the alcoholic PVP-I protocol two been similar 11-bed adult medical 1992–1997 have reported for Staphaqueous intensive PVP-I protocol. Moreover, paa Fisher’s exact test. Only p values " .20 are reported. Results are expressed in number (percent). N T tio cat lat wit iod adu p " .0 provid adequ trolled Third, and ab formed both p manip and th opmen Howev cathet extralu Parienti et al. CCM 2004 Mimoz O. et al. Arch Intern Med 2007;167:2066-2072 Mimoz O. et al. Arch Intern Med 2007;167:2066-2072 Site de pose Sous-Clavière > Jugulaire Richet et al. J Clin Microbiol 1990;28:2520-2525 Medical 99 (68) 108 (75) 93 (69) 101 (74) 78 (67) 81 (77) Complications of Femoral and Subclavian Venous Catheterization in Critically Ill Patients 41 (31) 35 (26) 38 (33)were 26 (23) complication in bivariate analysis COMMENT 101 (75) 105 (77) site and 84 (72) insertion at the femoral center 82In(78) this prospective, randomized, con2.0 (1.2) 2.0 (1.2) 1.8 (1.2) 2.0 (1.2) and 2 in the subclavian group (1.5%; (1 center had a lower risk and 1 center cealed multicenter study in critically ill Time between intensive care unit admission 1 (0-3) 1 (0-3) 1 (0-3) 1 (0-3) 1 (0-3) 1 (0-3) P = .07 by log-rank test). The inci- had a higher risk). In a multivariate lo- patients, we found that catheterizaand insertion, median (interquartile range), d unlessinfectious otherwise specified. SAPS II indicates Acute Physiologic Score II ;(HosmerODIN, Organ Dysfunction and/or *P..20 for femoral vs subclavian gisticSimplified regression model tion of the femoral vein was associdence catheterization densities groups, of major comInfection. 2 †P = .03. Lemeshow x ated with a significantly higher risk of = noncalculable; likeliplications were 4.5 per 1000 femoral ‡P = .13. Merrer et al. JAMA 2001 A Randomized Controlled Trial 2 catheter-days and 1.2 per 1000 subcla- hood ratio test x 1 =23.8), the only risk overall complications compared with factorContext for thrombotic complications was catheterization of the subclavian vein. vian catheter-days. Significant Jacques Merrer, MD factors Whether venous catheterization at the femoral site is associated with an Table 2. Characteristics Associated With Central Venous Catheter Insertion* at the femoral sitecompared (OR, 14.42; catheterization increased the associated with Bernard occurrence of anMDinfec- insertion De Jonghe, increased risk of complications with that atFemoral the subclavian site is debated. Patients With To compare mechanical, infectious, and thrombotic of fem- infection and Patients With Catheter Ultrasonographic All Patients CI, 3.33-62.57; P,.001). risk of complications catheter-related tious complication inGolliot, bivariate analy- 95% Objective Franck MS Culture (n = 270) Examination (n = 223) (n = 289) oral and subclavian venous catheterization. Overall Reduction in Complicathrombosis, whereas the rate of mesis included 3 risk factors: insertion at Jean-Yves Lefrant, MD Design and Setting Concealed,Femoral randomized controlled clinical trial conducted beFemoral Subclaviantions. Femoral Subclavian Subclavian Brigitte Raffy, MD The estimated absolute risk rechanical complications did not differ befemoral site, high Organ Dysfunction tween at 8 intensiveGroup care units (ICUs) in France. Group Group GroupDecember 1997 Groupand July 2000 Group Eric Barre, MD Characteristics = 145) (n = 144) duction (n = 134) (n = 136) (n = adult 116) patients (ntween = 107) associated with subclavian groups. and/or Infection score at(nadmission, and Patients Two hundred eighty-nine receiving athe first 2 central venous To our knowlLumens, No. of patients catheter. Jean-Philippe Rigaud, MD catheterization rather than femoral edge, this is the first randomized study insertion at 2 centers; and 2 protective 1 26 28 24 28 22 25 Interventions Patients were randomly assigned to undergo central venous cathDominique Casciani, MD catheterization was (95% providing direct comparison of 3 types factors: use of the catheter for sys2 69 66 65 60 33%site 57CI,or23%eterization at the femoral (n = 145) subclavian45site (n = 144). Benoı̂t Misset, MD 43%)Main for complications and 6% (95% of complications associated with subtemic administration of 50 antibiotics and 3 50 45 all 48 37 37 Outcome Measures Rate and severity of mechanical, infectious, and thrombotic Christophe Bosquet, MD Catheter inserted by insertion a senior physician, No. (%) 77 (53) 84 (58) 68 (51)† 83 (61)† 59 (51)‡ 64 (60)‡ complications, compared by catheterization site in 289, 270, and 223and patients, respectively. for major complicaclavian femoral catheterization. at 2 centers. In a Cox model CI, 0.2%-12%) 2 (86) Use of povidone-iodine antisepsis, No. (%) Hervé Outin, 125 122 (85) 115 (86) 115 (85) 101 (87) 89 (83) MD Femoral catheterization was associated incidence rate of catheter–related overall Consequently, 3 patients (95%with a higher Central venous com(likelihood ratio test x 2 = 22.1), inser- tions.Results Use of catheter for administration of, No. (%)Christian Brun-Buisson, MD infectious complications (19.8% vs 4.5%; P,.001; incidence density of 20 vs 3.7 per 1000 tion at the femoral site (HR, wouldand need be treated plications ill patients are usuAntibiotics 89 (61) 4.83; 95% 92 (64) CI, 2-4) 80 (62) 87 (64)toinfectious 65complications (56) us- (clinical 69 (64)sepsis within catheter-days) of major orcritically without bloodGérard Nitenberg, MD stream 4.4% vs 1.5%;femoral P=.07; incidence of 4.5 vs 1.2 per 1000 catheterBlood products CI, 1.96-11.93; P,.001) 30increased (21) 26 (18) ing subclavian 28 (21)infection, 24 (18) 23 (20) 18 (17) the rather than cath-density ally classified as mechanical, infectious, days), as well as of overall thrombotic complications (21.5% vs 1.9%; P,.001) and comfor the French Catheter Study Group Parenteral nutrition 88 (61) 91 (63) 83 (62) 89 (65) 71 (61) 70 (65) risk of infection, whereas eterization to prevent 1 (6% complication thrombotic. In our study, the 4 maplete thrombosis of the vessel vs 0%; P=.01); ratesand of overall and major mechanical in Intensive Careuse of the Prophylactic anticoagulation, No. (%) 123 (85) 123 (85) 116 (87) 120 (88) between101 97 (91) complications were similar the(87) 2 groups 18.8 %; P=.74 and 1.4% vs catheter for systemic antibiotic therapy of catheterization, and 16 patients (95%(17.3% jorvsmechanical complications (2.8%) obDuration of catheter placement, mean (SD), d 9.3 (6.2)§ 11.0 (6.3)§ 9.4 (6.2)\ 11.1 (6.3)\ Risk factors 9.6 (6.3)¶ 11.3 (6.4)¶ 2.8%; P=.44, respectively). for mechanical complications were duration of in(HR, 0.41; 95% CI, 0.18-0.93; P = .03) CI,- 8-411) to prevent 1 major compli- 2.8served with the subclavian approach ENTRAL VENOUS CATHETER sertion 1.03-1.08 per additional minute; Time between catheter removal and NA NA NA (odds ratio [OR], NA1.05; 95% confidence 2.6 (1.5) interval [CI], (1.7) ization is often necessarycation. to P,.001); insertion in 2 of the centers (OR, 4.52; 95% were CI, 1.81-11.23; P=.001); and inultrasonographic decreased study, mean (SD), d risk of infectious the compneumothoraces necessitating chest sertion during the night (OR, 2.06; 95% CI, 1.04-4.08; P=.03). The only factor associated critically patients otherwise noted.ill NA indicateshosnot applicable. *P..20 for femoral vs subclavian catheterization groups, unlesstreat plications. Microorganisms recovered †P = .09. pitalized in intensive care with infectious complications was femoral catheterization (hazard ratio [HR], 4.83; 95% ‡P = .18. (ICUs).are However, this proce- CI, 1.96-11.93; P,.001); antibiotic administration via the catheter decreased risk of infecfrom catheter units cultures summa§P = .01. complications (HR, 0.41; 95% CI, 0.18-0.93; P=.03). Femoral catheterization was dure can lead to serious and some\P = .009. Tabletious 3. Catheter-Related Infectious Complications rized in TABLE 4. ¶P = .02. times life-threatening complications, the only risk factor for thrombotic complications (OR, 14.42; 95% CI, 3.33-62.57; P,.001). Femoral Subclavian Thrombotic Complications. the whether mechanical,Of infectious, or Conclusion Femoral venous catheterization is associated with a greater risk of inGroup, No. Group, No. fectious and thrombotic complications than subclavian catheterization in ICU patients. choice of insertion 223 patients who The underwent ul- site 704 JAMA, August 8, 2001—Vol 286, (77.2%) No. 6thrombotic. (Reprinted) ©2001 American Medical Association. All rights reserved. (n = 134) * Code Classification (n = 136) P Value† JAMA. 2001;286:700-707 www.jama.com can influence the incidence and type of trasonographic examination for detecNA Sterile 100 127 NA such complications. 3,4 subclavian cath-3 Reported(,1000 rates ofcolony-forming catheter- from 10% to 50% with 7 tion of a catheter-related thrombosis, 116 he- 1 of cases. Major femoral or retroperitoneal Contamination NA thrombosis range from 6.6% to eterization.8,9 These differences may be matoma is theand most107 frequent major me- related units/mL and no clinical sepsis) were in the femoral group were 5-7 and related to use of various diagnostic ap25% with catheterization chanical complication of femoral ve-by Downloaded from jama.ama-assn.org onfemoral April 6, 2012 colony-forming 2 guest Colonization ($1000 units/mL 19 3 in the subclavian group. Among the 66 nous catheterization, occurring in up and no clinical sepsis) Author Affiliations and Study Group Members are (e-mail: jmerrer@chi-poissy-st-germain.fr). patients in whom ultrasonography was pneumo- listed at the end of this article. to 1.3% of cases,1,2 whereas for the Critically Ill Patient Section Editor: Clinical sepsis without bloodstream Caring infection 4 1 Author and Reprints: Jacques Deborah J. Cook, MD, Consulting Editor, JAMA. is the most frequent major com- 3 Corresponding .07 ,.001 not performed, thorax 41 (65%) died before Merrer, MD, Service de Réanimation Médicale, Advisory Board: David Bihari, MD; Christian BrunClinical sepsis with bloodstream infection 2 1 plication of subclavian venous cath- 4 Hôpital de Poissy/St Germain-en-Laye, 10 Rue Buisson, MD; Timothy Evans, MD; John Heffner, MD; catheter removaleterization, or ultrasonographic ex-to 2.3% d u C h a m p - G a i l l a r d , 7 8 3 0 3 P o i s s y , F r a n c e Norman Paradis, MD. occurring in 1.5% 5 Unable to discriminate between codes 2 and 3 2 1 amination, 10 were discharged from the Codes 2, 3, 4, and 5 were collectively considered infectious complications. 700 JAMA, August 8, 2001—Vol 286, No. 6*(Reprinted) ©2001 Americancatheter-related Medical Association. All rights reserved. Codes 3 and 4 were hospital before examination was perconsidered major catheter-related infectious complications. NA indicates not applicable. Surgical 46 (34) 36 (25) major catheter-related infectious comMechanical ventilation, No. (%) 112group (77) 109 (76) plications in the femoral (4.4%) ODIN score, mean (SD) 2.0 (1.3) 2.0 (1.2) 19 20 C Site de pose « En raison du risque accru d’ILC en territoire cave inférieur, il est recommandé d’insérer les cathéters en territoire cave supérieur. » 5ème conférence de consensus – Prévention des infections nosocomiales en réanimation – SFAR - SRLF O’Grady NP et al. Guidelines for the prevention of intravascular catheter-related infections. Clin Infect Dis 2011 Durée de cathétérisme Richet et al. J Clin Microbiol 1990;28:2520-2525 Mode de pose « A l’aveugle » ? Echographie ? Research Results OpenaAccess longitudinal axis provided clear image of both walls of the vessel (theinternal actual vein puncture using either the longitudinal or Baselineultrasound-guided characteristics of the study population are presented eal-time catheterisation of the jugular the transverse axis of the 2D image was left to the discretion 1. There were no significant differences between the vein:intwo aTable prospective comparison with the landmark technique in of the operator). Also, using this approach, a single-wall puncgroups of patients in gender ratio, age, BMI, or side of critical care patients ture can be made by observing the point at which the needle catheterisation or in the presence of risk factors for difficult 1 2 3 first indents the anterior4,wall of the IJV. A short stabbing venous cannulation such as Labropoulos prior catheterisation, limited sites, Alexandros Dimitrios Karakitsos , Nicolaos , Eric De Groot P Patrianakos 5, Johnprevious 1, George 6, Dimosthenis 7, at this point will tend to puncture the motionA of the needle for access attempts, difficulties during catheterisaGregorios Kouraklis Poularas Samonis Tsoutsos anterior wall without opposing it to the posterior wall, thereby tion, previous mechanical complication, known vascular abnorManousos M Konstadoulakis8 and Andreas Karabinis1 avoiding a double-wall puncture (Figure 2). Single-wall puncmality, untreated coagulopathy, skeletal deformity, and tures were achieved in all cases using ultrasound guidance. cannulation during cardiac arrest (Table 1). 1Department of Intensive Care, General State Hospital of Athens, 154 Mesogeion Avenue, 11527 Athens, Greece of Vascular Surgery, University of Medicine and Dentistry of New Jersey, The University Hospital-150 Bergen Street Newark, NJ 07103 USA 3Academic Medical Center, Department of Vascular Medicine, University of Amsterdam Tafelbergweg 51 .1105 BD Amsterdam, The Netherlands 4Department of Cardiology, University Hospital of Heraklion, PO Box 1352 Stavrakia, Heraklion, Crete, Greece Table 1 52nd Department of Propedeutic Surgery, University of Athens School of Medicine, Laiko General Hospital, 17 Agiou Thoma street-11527 Athens, Greece Characteristics of the total study population 6Department of Internal Medicine and Infectious Diseases, University of Crete, P. O. Box 2203, 71003 Heraklion, Greece 7'J. Ioannovic' Burn Center, General State Hospital of Athens, 154 Mesogeion Avenue, 11527 Athens, Greece 81st Department Characteristics Ultrasound group (n = Hospital,114 450) Landmark of Propedeutic Surgery, University of Athens School of Medicine, Hipokrateion University Vasilis Sofias Avenue 11527 group (n = 450) Athens, Greece 2Division Age (years)a 58.3 ± 10.3 59 ± 9.5 Gender (male/female ratio)a 0.56 ± 0.4 0.6 ± 0.4 Side of catheterisation (left/right) 222/228 218/232 Limited sites for access attempts 51 (11.3%) 55 (12.2%) 44 (9.7%) 40 (8.8%) 18 (4%) 20 (4.4%) Corresponding author: Dimitrios Karakitsos, echolabicu@gmail.com Received: 23 May 2006 Revisions requested: 15 Jun 2006 Revisions received: 8 Sep 2006 Accepted: 10 Nov 2006 Published: 17 Nov 2006 Critical Care 2006, 10:R162 (doi:10.1186/cc5101) mass index (kg/m2)a 24.1 ± 5.3 23.7 ± 5.9 This article Body is online at: http://ccforum.com/content/10/6/R162 © 2006 Karakitsos et al.; licensee BioMed Central Ltd. Prior catheterisation 85 (18.8%) 76 (16.8%) This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. See relatedPrevious commentary by Bodenham, http://ccforum.com/content/10/6/175 difficulties during catheterisation Previous mechanical complication Abstract Known vascular abnormality Introduction Central venous cannulation is crucial in the Untreated coagulopathy management of the critical care patient. This study was designed Skeletal to evaluate whether real-time ultrasound-guided deformity cannulation of the internal jugular vein is superior to the standard Cannulation during cardiac arrest landmark method. Values are presented as mean ± standard Methods In this randomised study, 450 critical care deviation. patients who underwent real-time ultrasound-guided cannulation of the internal jugular vein were prospectively compared with 450 critical care patients Page 4 of 8in whom the landmark technique was used. Randomisation was performed by means of a computer(page number not for citation purposes) a 4 (0.8%) 3 (0.6%) coagulopathy, skeletal deformity, and cannulation during cardiac 25 (5.5%) 24 (5.3%) arrest between the two groups of patients. Furthermore, the physicians 15 who performed the procedures had comparable (3.3%) 13 (2.8%) experience in the placement of central venous catheters (p = 31 (6.8%) 35 (7.7%) non-significant). Cannulation of the internal jugular vein was achieved in all patients by using ultrasound and in 425 of the patients (94.4%) by using the landmark technique (p < 0.001). Average access time (skin to vein) and number of attempts were significantly reduced in the ultrasound group of patients compared with the landmark group (p < 0.001). In the landmark Crit Care group, puncture of the carotid artery occurred in 10.6% of 2006:10;R162 Research Open Access medial to theultrasound-guided common carotid artery in 53catheterisation (12.6%) cases and of the internal jugular eal-time directly medial to the artery in 17 (3.7%) cases. Discussion vein: a prospective comparison with the landmark technique in The use of CVCs may be associated with adverse effects that critical care Results using thepatients landmark technique are in sharp contrast to are both hazardous to patients and expensive to treat [18]. 1, Nicolaos Labropoulos2, Eric De Groot3, Alexandros P Patrianakos4, Dimitrios Karakitsos those obtained by the ultrasound method and are presented in Mechanical complications are reported to occur in 5% to 19% 5, John Poularas1, George Samonis6, Dimosthenis A Tsoutsos7, Gregorios Kouraklis Table 2. Average access time and number of attempts were of patients, infectious complications in 5% to 26%, and throm8 1 Manousos M Konstadoulakis and Andreas compared Karabinis with the both significantly reduced using ultrasound botic complications in 2% to 26% [19,20]. These complicalandmark technique (p < 0.001) (Table 2). The success rate tions increase in association with several characteristics, 1Department of Intensive Care, General State Hospital 154 Mesogeion Avenue, 11527 Athens, Greece patient anatomy (for example, morbid obesity, was significantly lower and the rateofofAthens, mechanical complicaincluding 2Division of Vascular Surgery, University of Medicine and Dentistry of New Jersey, The University Hospital-150 Bergen Street Newark, NJ 07103 USA Medical Center, Department of Vascular Medicine, University of Amsterdam Tafelbergweg 51 .1105 BD Amsterdam, The Netherlands 4Department of Cardiology, University Hospital of Heraklion, PO Box 1352 Stavrakia, Heraklion, Crete, Greece Table 2 52nd Department of Propedeutic Surgery, University of Athens School of Medicine, Laiko General Hospital, 17 Agiou Thoma street-11527 Athens, Greece 6Department of Internal Medicine Diseases, University of Crete, P. landmark O. Box 2203, group 71003 Heraklion, Greece Outcome measures inand theInfectious ultrasound group versus the of patients 7'J. Ioannovic' Burn Center, General State Hospital of Athens, 154 Mesogeion Avenue, 11527 Athens, Greece 81st Department of Propedeutic Surgery, University of Athens School of Medicine, Hipokrateion University Hospital,114 Vasilis Sofias Avenue 11527 Outcome measures Ultrasound group (n = 450) Landmark Athens, Greece 3Academic Access time (seconds) Corresponding author: Dimitrios Karakitsos, echolabicu@gmail.com 17.1 ± 16.5 (11.5 to 41.4)a 44 ± 95.4 (33.2 to 77.5) a 10 Nov 2006 Published: 17 Nov 2006 Received: 23 May 2006 Revisions requested: 15 Jun 2006 Revisions received: 8 Sep450 2006(100%) Accepted: Success rate Critical Care 2006, 10:R162 (doi:10.1186/cc5101) Carotid puncture 5 (1.1%)a This article is online at: http://ccforum.com/content/10/6/R162 © 2006 Karakitsos et al.; licensee BioMed Central Ltd. Haematoma 2 (0.4%)a This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), a which permits unrestricted use, distribution, and reproduction in any medium, provided the original is properly cited. Haemothorax 0 (0%)work SeePneumothorax related commentary by Bodenham, http://ccforum.com/content/10/6/175 Average number of attempts Abstract CVC-BSI group (n = 450) 425 (94.4%) 48 (10.6%) 38 (8.4%) 8 (1.7%) 0 (0%)a 11 (2.4%) 1.1 ± 0.6 (1.1 to 1.9)a 2.6 ± 2.9 (1.5 to 6.3) 47 (10.4%)a 72 (16%) Introduction Central venous cannulation is crucial in the coagulopathy, skeletal deformity, and cannulation during cardiac aComparison management of the critical care measures patient. This study the wasultrasound arrest between the the twolandmark groups ofgroup patients. Furthermore, of the outcome between group and of patients (p <the 0.001). Access time and average designed real-timeasultrasound-guided physicians who confidence performed the procedures had rate, comparable numbertoofevaluate attemptswhether are expressed mean ± standard deviation (95% interval). Success carotid puncture, haematoma, cannulation of the internal jugular veinand is superior to theare standard in the placement central venous catheters (pof=their group. CVC-BSI, central haemothorax, pneumothorax, CVC-BSI expressedexperience as the absolute number ofofpatients and percentage landmark method. non-significant). Cannulation of the internal jugular vein was venous catheter-associated blood stream infection. achieved in all patients by using ultrasound and in 425 of the Methods In this randomised study, 450 critical care patients patients (94.4%) by using the landmark technique (p < 0.001). who underwent real-time ultrasound-guided cannulation of the Average access time (skin to vein) and number of attempts were Page 5 of 8 internal jugular vein were prospectively compared with 450 significantly reduced in the ultrasound group of patients critical care patients in whom the landmark technique was used. compared with the landmark group (p < 0.001). In the landmark (page number not for citation purposes) Crit Care 2006:10;R162 Randomisation was performed by means of a computergroup, puncture of the carotid artery occurred in 10.6% of p < 0,001 Pansements Quel type de pansement ? Pansements transparents semi-perméables Car permettent un examen quotidien du site d’insertion Randomized Controlled Trial of Chlorhexidine Dressing and Highly Adhesive Dressing for Preventing Catheter-related Infections in Critically Ill Adults Jean-François Timsit1,2, Olivier Mimoz3, Bruno Mourvillier4, Bertrand Souweine5, Maı̈té Garrouste-Orgeas6, Serge Alfandari7, Gaétan Plantefeve8, Régis Bronchard9, Gilles Troche10, Remy Gauzit11, Marion Antona12, Emmanuel Canet13, Julien Bohe14, Alain Lepape14, Aurélien Vesin1, Xavier Arrault15, Carole Schwebel2, 17, Stéphane Ruckly1, Caroline Tournegros2, and Jean-Christophe Lucet18 Christophe Adrie16, Jean-RalphTimsit, Zahar Mimoz, Mourvillier, et al.: CHG-Gel Dressing for Catheter-related Infection Prevention 1 12 2 Université Grenoble 1 (Joseph Fourier), U823 “Outcome of Cancers and Critical Illness,” Albert Bonniot Institute, La Tronche, France; Université Grenoble 1, Medical ICU, Albert Michallon Hospital, Grenoble, France; 3Service d’Anesthésie Réanimation, Centre Hospitalier Universitaire, Université de Poitiers et Inserm U1070, Poitiers, France; 4Medical ICU, 9Surgical ICU, and 15Drug Delivery Department, Bichat-Claude Bernard University Hospital, Paris, France; 5Université Clermont-Ferrand, Medical ICU, Gabriel Montpied Hospital, Clermont-Ferrand, France; 6MedicalSurgical ICU, Saint Joseph Hospital Network, Paris, France; 7Intensive Care and Infectious Diseases Unit, General Hospital, Tourcoing, France; 8 Medical Surgical ICU, General Hospital, Argenteuil, France; 10Medical Surgical Intensive Care Unit, General Hospital, Versailles, France; 11 Surgical Intensive Care Unit, Hotel-Dieu University Hospital, Paris, France; 12General ICU, Raymond Poincaré Hospital, University of Versailles SQY, Garches, France; 13Université Paris VII, Medical ICU, Saint Louis Hospital, Paris, France; 14Université Claude-Bernard Lyon 1, Medical-Surgical ICU, Centre Hospitalier Lyon Sud, Lyon, France; 16Physiology Department, Cochin Hospital, Paris, France; 17Micro-Biology and Hygiene, Necker Hospital, Paris, France; and 18Infection Control Unit, Bichat-Claude Bernard University Hospital, Assistance-Publique Hôpitaux de Paris, and Université Paris Diderot, Sorbonne Paris Cité, France Rationale: Most vascular catheter-related infections (CRIs) occur extraluminally in patients in the intensive care unit (ICU). Chlorhexidineimpregnated and strongly adherent dressings may decrease catheter colonization and CRI rates. Objectives: To determine if chlorhexidine-impregnated and strongly adherent dressings decrease catheter colonization and CRI rates. Methods: In a 2:1:1 assessor-masked randomized trial in patients with vascular catheters inserted for an expected duration of 48 hours or more in 12 French ICUs, we compared chlorhexidine dressings, highly adhesive dressings, and standard dressings from May 2010 to July 2011. Coprimary endpoints were major CRI with or without catheterrelated bloodstream infection (CR-BSI) with chlorhexidine versus nonchlorhexidine dressings and catheter colonization rate with highly adhesive nonchlorhexidine versus standard nonchlorhexidine dressings. Catheter-colonization, CR-BSIs, and skin reactions were secondary endpoints. Measurements and Main Results: A total of 1,879 patients (4,163 catheters and 34,339 catheter-days) were evaluated. With chlorhexidine dressings, the major-CRI rate was 67% lower (0.7 per 1,000 vs. 2.1 per 1,000 catheter-days; hazard ratio [HR], 0.328; 95% confidence interval AT A GLANCE COMMENTARY Figure 1. Flow chart of the study. C BSI ¼ catheter-related bloodstream Scientific Knowledge on the Subject fection; CRI ¼ catheter-related infectio Chlorhexidine-impregnated sponges decrease catheter-related ITT ¼ intention to treat. infections (CRI) in the intensive care unit (ICU) but make impossible the continuous inspection of insertion site. Dressing disruption is frequent in the ICU and a major risk factor of CRI. What This Study Adds to the Field Chlorhexidine-impregnated gel dressings decrease by 60% the risk of CRI in the ICU. This second large, multicenter randomized control trial confirmed the benefits of chlorhexidine dressings. A highly adhesive nonchlorhexidine transparent dressing decreased dressing disruption but increased cutaneous and positive catheter tip culture. different between these two groups (Table 3; see Figure E changes per catheter-days was significantly lower in the highly J Respir Crit Care Meddifferent 2012;186:1272-8 Also, CLA-BSI was between groups (H per catheter-day; ,0.0006) 0.0001;Am [CI], 0.174–0.619;PP ¼ and the CR-BSI rate 60%rate lower (0.5not per adhesive (0.33 [0.20–0.50] (Received in original form June 9, 2012; accepted in final group form September 17, 2012) Transport to operating room with catheter in place, n (%) No Once Twice More than twice Transport out of ICU with catheter in place, n (%) No Once Twice More than twice Number of dressing changes per catheter, median (IQR) 1,2Catheter removal for suspected 3 infection, n (%) Data for arterial catheters only 7 8 9 Arterial catheter, n (%) 13 14 14 Femoral 3,436 587 94 46 (82.5) (14.1) (2.3) (1.1) 877 148 24 18 (82.2) (13.9) (2.2) (1.7) 812 140 25 11 (82.2) (14.2) (2.5) (1.1) 1,747 299 45 17 (82.9) (14.2) (2.1) (0.8) 2,638 1,109 294 122 2 563 (63.4) (26.6) (7.1) (2.9) (1–4) (13.5) 675 272 86 34 3 155 (63.3) (25.5) (8.1) (3.2) (1–5) 5 (14.5) 632 284 45 27 2 130 (64) (28.7) (4.6) (2.7) (1–4) (13.2) 1,331 (63.1) 553 (26.2) 163 (7.7) 61 (2.9) 2 (1–4) 6 278 (13.2) 12 1,128 (53.5) 393 (34.8) 73518(65.2) 473 (47.9) 980 (46.5) Randomized Controlled Trial of Chlorhexidine Dressing and Highly Adhesive Dressing for Preventing Catheter-related Infections in Critically Ill Adults Jean-François Timsit , Olivier Mimoz , Bruno Mourvillier4, Bertrand Souweine , Maı̈té Garrouste-Orgeas , Serge Alfandari , Gaétan Plantefeve , Régis Bronchard , Gilles Troche10, Remy Gauzit11, Marion Antona , 2,201 (52.9) 558 (52.3) 515 (52.1) 1, Xavier Arrault 2, 773 (35.1) 207 (37.1) 15, Carole 173Schwebel (33.6) Emmanuel Canet , Julien Bohe , Alain Lepape , Aurélien Vesin Radial 1,428 (64.9) 351 (62.9) 342 (66.4) 17, Stéphane Ruckly1, Caroline 2, and Jean-Christophe Christophe Adrie16, Jean-Ralph Zahar Tournegros Lucet Data for CVCs only 1 All CVCs, n (%) 1,962 (47.1) 509 (47.7) Jugular CVCs 728 (37.1) 180Bonniot (35.4) 3732Université (38.1) Université Grenoble 1 (Joseph Fourier), U823 “Outcome of Cancers and Critical Illness,” Albert Institute,175La(37) Tronche, France; Subclavian CVCs 567 (28.9) 152 (29.9) 140 (29.6) 275 (28.1) 3 Timsit, Mourvillier, Hospital, et al.: CHG-Gel Dressing for Catheter-related Infection Réanimation, Prevention 1277 Grenoble 1, Medical ICU, Mimoz, Albert Michallon Grenoble, France; Service d’Anesthésie Centre Hospitalier Universitaire, Femoral CVCs 667 (34) 177 (34.8) 158 (33.4) 332 (33.9) 4 9 Guidewire exchange, n (%) France; Medical ICU, Surgical 76 (3.9) 23 (4.5) 20 (4.2) Bichat-Claude33Bernard (3.4) Université de Poitiers et Inserm U1070, Poitiers, ICU, and 15Drug Delivery Department, 5 catheters, n (%) 5 (0.3) (0.2) 2 (0.4) 2 (0.2) 3. Tunneled HAZARD RATIOS IN THE INTENTION-TO-TREAT ANALYSIS University Hospital,TABLE Paris, France; Université Clermont-Ferrand, Medical ICU, Gabriel Montpied 1Hospital, Clermont-Ferrand, France; 6MedicalVenous catheter lumens, n (%) 7 Surgical ICU, Saint Joseph Hospital Network, Paris, France; Intensive Care and17Infectious Diseases Unit, General Hospital, Tourcoing, One (0.9) 6 (1.2) 3 (0.6)Standard 8Adhesive (0.8) Nonchlorhexidine vs. Chlorhexidine Dressings vs. HighlyFrance; Dressings 8 Two Hospital, Argenteuil, France; 10Medical Surgical 201 (10.2) (9.8) General Hospital, 42 (8.9) Versailles, France; 109 (11.1) Medical Surgical ICU, General Intensive Care Unit, (941 patients/2,055 catheters vs.50938 (476 patients/1,067 catheters vs. 465 Three 1,458 375 (73.7) 358 (75.7) 725 (74) 11 12 (74.3) patients/2,108 catheters) patients/988 Surgical IntensiveVariable Care Unit, Hotel-Dieu ICU, Raymond University of catheters) Versailles Greater than three University Hospital, Paris, France; 286General (14.6) 78 (15.3) Poincaré Hospital, 70 (14.8) 138 (14.1) 13 14 Use of lipids,Paris n (%) VII, Medical ICU, Saint Louis Hospital, 938 (47.8) 247 (48.5) Claude-Bernard 219 (46.3) 472 (48.2) ICU, SQY, Garches, France; Université Paris, France; Université Lyon 1, Medical-Surgical Catheter colonization of heparin, n (%) 615 (31.3) 159 146 (30.9) 310 (31.6) 16 17 (31.2) Centre Hospitalier Lyon Sud,Use Lyon, Physiology Department, Cochin Hospital, Paris, France;201 Micro-Biology and Hygiene, Necker Paris, Incidence (n perFrance; 1,000 9.6Hospital, vs. 12.5 Red-blood-cell packcatheter-days) transfused, n (%) 766 10.9 (39) vs. 4.3 (39.5) 182 (38.5) 383 (39.1) 18 Hazard ratio 0.412 (0.306–0.556), P , 0.0001 1.651 (1.208–2.256), P ¼ 0.0016 France; and Infection Control Unit, Bichat-Claude Bernard University Hospital, Assistance-Publique Hôpitaux de Paris, and Université Paris Diderot, of abbreviations: CVC ¼ central venous catheter; ICU ¼ intensive care unit; IQR ¼ interquartile range. bloodstream infection Sorbonne Paris Cité,Catheter-related France Definition Incidence (n per 1,000 catheter-days) 1.3 vs. 0.5 1.3 vs. 1.3 Hazard ratio 0.402 (0.186–0.868), P ¼ 0.02 1.215 (0.470–3.142), P ¼ 0.689 Skin Colonization catheters (67.5%; P , 0.001). Median (IQR) rate of Count-Tact Major catheter-related infections was significantly lower in the chlorhexidine Incidence (n per 1,000 catheter-days) 0.7 2.3 versus vs. 1.9 the Rationale: Most vascular catheter-related infections (CRIs) occur extra-of 2,965 cathe-2.1 vs.positivity Count-Tact cultures were performed at removal nonchlorhexidine group and significantly higher in the highly adHazard ratio 0.328 (0.174–0.619), P ¼ 0.0006 1.052 (0.517–2.142), P ¼ 0.888 ters and were negative in 918 (31%) cases. Bacterial growth was luminally in patients in the intensive care unit (ICU). Chlorhexidine- AT A GLANCE COMMENTARY hesive nonchlorhexidine versus the standard group (23 [1–101] vs. common in patients with colonization (89%), major-CRI impregnated and stronglymore adherent dressings may decrease catheter 10 [0–100]; P ¼ 0.010) Table E3). (87.5%), or CR-BSI (87.5%) than in patients with noncolonized Scientific Knowledge on (see theand Subject colonization and CRI rates. Adverse Events decreased major-CRI CR-BSI rates in our study similarly to Objectives: To determine if chlorhexidine-impregnated and strongly the sponge (13). The sponges rate of severe contact dermatitis was comChlorhexidine-impregnated decrease catheter-related No systemic adverse reaction to chlorhexidine occurred. Severe parable with that reported with chlorhexidine adherent dressingscontact decrease catheter colonization and CRI rates. (CRI) in the intensive care unit (ICU) sponges. but make dermatitis requiring permanent discontinuation ofinfections the In our previous study (13), we standard Methods: In a 2:1:1 assessor-masked randomized trial in patients with study dressing occurred in 22 chlorhexidine-group patients impossible (1.1 the continuous inspection of used insertion site.transparent Dressing Tegaderm dressings, of which two-thirds were replaced earlier than vascular catheters per inserted for an expected duration of 48 hours or 100 catheters), four highly adhesive-group patients (0.5disruption per is frequent in the ICU and a major risk factor scheduled, because of soiling or detachment. Spontaneous dressmore in 12 French ICUs, we compared chlorhexidine dressings, 100 catheters), and one standard-group patienthighly (0.1 per 100 cathof CRI. ing detachment was associated with catheter colonization and inadhesive dressings, and (P standard from May 2010 Julygroups, P , eters) ¼ 0.0005dressings for comparison among theto three fection (28), suggesting that a highly adhesive transparent dressing 0.0001 forwere comparison between chlorhexidine and nonchlorhex2011. Coprimary endpoints major CRI with or without catheter(Tegaderm HP Transparent Film Dressing) might decrease cathidine dressings, P ¼ 0.17with for chlorhexidine comparison between and This related bloodstream infection (CR-BSI) versusstandardWhat Adds and to the Field rates. However, skin and catheter eterStudy colonization major-CRI highly adhesive nonchlorhexidine dressings). nonchlorhexidine dressings and catheter colonization rate with highly colonization rates at catheter removal were higher with TegaChlorhexidine-impregnated gel dressings decrease by 60% Contact dermatitis usually occurred for a single adhesive nonchlorhexidine versus standard nonchlorhexidine dress-catheter per derm HP, despite a significant decrease in early dressing changes. the risk of CRI in the ICU. This second large, multicenter patient andCR-BSIs, selectively patients withsecondmultiple organ failings. Catheter-colonization, andaffected skin reactions were The control hydrophilic acrylate component in Tegaderm HP may have randomized trial confirmed the benefits of chlorhexiure, subcutaneous edema, and fragile skin. No systemic adverse ary endpoints. resulted in skin toxicity or dermabrasion during dressing changes, reactions to chlorhexidine occurred. dine dressings. A highly adhesive nonchlorhexidine transMeasurements and Main Results: A total of 1,879 patients (4,163 cathincreasing the risk of colonization from the pilosebaceous units. The rate of abnormal International Contact Dermatitis Research parent dressing decreased dressing disruption but increased eters and 34,339 catheter-days) were evaluated. With chlorhexidine Our study is the first to evaluate chlorhexidine-gel dressings Group scores was significantly higher with chlorhexidine (2.3%) cutaneous and positive catheterand tipthe culture. dressings, the major-CRI rate was 67% lower (0.7 per 1,000 vs. 2.1 per for major-CRI prevention second large randomized conthan without chlorhexidine (1%; P , 0.0001). Abnormal scores 1,000 catheter-days; hazard ratio [HR], 0.328; 95% confidence interval trolled trial (after [13]) showing that chlorhexidine dressings dewere significantly more common with the highly adhesive dresscrease major-CRI CR-BSI rates in ICUs. Nine out Figure 2. Cumulative of (A) major catheter-related (CRI) with chlorhexidine-gel (CHG) and dressings and nonantiseptic dressings, and (B) of the 12 ing (1.4%) than with the risk standard dressing (0.7%; Pinfections ¼ 0.0039). J Respir Crit Care Med 2012;186:1272-8 never usedAm chlorhexidine dressings. In addition, we obcatheter colonization with highly adhesive nonchlorhexidine dressings versusICUs standard dressings. Pansements Rôle primordial pour la prévention des ILC Périodicité de réfection des pansements reste discutée… Recommandations françaises : tous les 3 jours Recommandations américaines : tous les 7 jours sauf si souillure 5ème conférence de consensus – Prévention des infections nosocomiales en réanimation – SFAR - SRLF O’Grady NP et al. Guidelines for the prevention of intravascular catheter-related infections. Clin Infect Dis 2011 Maki et al. Am J Med 1981;70,739-744 3 groupes : - Groupe contrôle : rien - PI2 : Iodophore - PNB : Polymyxine, Néomycine et Bacitracine Place des pommades antibiotiques Pas moins d’infections … … et au contraire : + d’infections à Levures ! Nombre de lumières Nombre de lumières : méta-analyse Cathéter et héparine Si CVC imprégnés d’héparine : moindre risque de thrombose ? Hoar PF. Heparin bonding reduces thrombogenicity of pulmonary-artery catheters. N Engl J Med 1981;305:993-995 Daniel HW. Heparin in the prevention of infusion phlebitis. A double-blind controlled study. JAMA 1973;226:1317-21 Prévention de l’infection de cathéter par l’héparine Randolph AG. Benefit of heparin in central venous and pulmonary artery catheters. A meta-analysis of randomized controlled trials. Chest 1998;113:165-171 Cathéters imprégnés … Cathéter imprégnés d’antiseptiques Imprégnés (n=227) Non imprégnés (n=215) Nb jours-cathéters 6 +/- 3,4 6 +/- 2,8 Colonisation, n (%) 28 (13,5) 47 (24) Taux/1000 j 22,5 40 Bactériémie, n (%) 2 (1) 9 (4,6) Taux/1000 j 1,6 7,6 Nouveau site 1,4 4,6 Après échange/ guide 2 6 p 0,005 0,03 Maki et al. Ann Intern Med 1997;127:257-266 Méta-analyse Colonisations Bactériémies Veenstra et al. JAMA 1999;281:261-267 VOLUME 340 J A N U A R Y 7, 1999 NUMBER 1 A C O M PA R I S O N O F T WO A N T I M I C R O B I A L - I M P R EG N AT E D C E N T R A L V E N O U S C AT H ET E R S A COMPARISON OF TWO ANTIMICROBIAL-IMPREGNATED CENTRAL VENOUS CATHETERS an independent effect. The limit for entering or removRABIH O. DAROUICHE, M.D., ISSAM I. RAAD, M.D., STEPHEN O. HEARD, M.D., JOHN I. THORNBY, PH.D., bles in the logistic-regression models was a P value of TABLE 1. CHARACTERISTICS OF THE PATIENTS O C. WENKER , M.D., ANDREA GABRIELLIAND , M.D., JOHANNES BERG, M.D., NANCY KHARDORI, M.D., HEND HANNA, M.D., LIVIER ess. All computations were performed with SAS/STAT ANTIMICROBIAL-IMPREGNATED CATHETERS.* 12 An independent monitoring RAY board HACHEM , M.D., of RICHARD L. HARRIS, M.D., AND GLEN MAYHALL, M.D., FOR THE CATHETER STUDY GROUP* composed exinfectious diseases reviewed and helped interpret the MINOCYCLINE– CHLORHEXIDINE–SILVER of the study. An interim analysis of the data was not per- ABSTRACT I RIFAMPIN SULFADIAZINE associated with the use of central CNFECTION ATHETERS CATHETERS (N=356) (N=382) venous catheters can result in serious medical Background The use of centralCHARACTERISTIC venous catheters RESULTS complications and expensive care.1 In prospecimpregnated with either minocycline and rifampin No. of patients 350 370 eristics of Patients and Catheters or chlorhexidine and silver sulfadiazine reduces the tive, clinical trials, the use of cenMale sex (%) 59 randomized63 rates of catheter colonization and catheter-related tral venous catheters impregnated with either mital of 865 study catheters (414 impregnated Median age (yr) 56 56 bloodstream infection as compared with the use of 2 or chlorhexidine and silver nocycline and rifampin inocycline and rifampin and 451 impregnatUnderlying disease (%) unimpregnated catheters. We compared the rates Cancer 28 3 was associated 26 sulfadiazine with reduced rates of h chlorhexidine and silver sulfadiazine) were of catheter colonization and catheter-related bloodCardiopulmonary disease 32 colonization 34and catheter-related bloodcatheter d into 817 patients. Complete data could bewith these stream infection associated two disorder kinds of Neurologic 16 19 stream 24 infection, as compared with unimpregnated Other 21 ed for 738 catheters (85 percent): 356 imantiinfective catheters. 4 and studies in animals5 catheters. In vitro studies Patients in intensive care unit (%) 66 67 We conducted Methods ted with minocycline and rifampin and 382a prospective, randomRisk factors for infection (%) have suggested that catheters impregnated with miized clinical in 12 university-affiliated hospitals. nated with chlorhexidine andtrial silver sulfadiaHyperalimentation 16 and rifampin 16 can resist infection more efnocycline High-riskThe adult patients127 in whom central venous nserted in 698 patients. remaining Immunosuppressive therapy fectively22than catheters 20impregnated with chlorhexicatheters were expected to remain Bone in place for three marrow transplantation 6 4 rs (58 impregnated with minocycline and ridine and5 silver sulfadiazine, but the clinical efficacy or more days were randomly assigned to undergo Neutropenia 3 and 69 impregnated with chlorhexidine and insertion of polyurethane, triple-lumen catheters imMechanical ventilation 60 65 of these two types of antiinfective catheters has not ulfadiazine, with similar patient catheter catheter been compared 53 54 We compared catheters impregnated with and either minocycline Other and intravascular rifampin (on directly. Urinaryor catheter 85 86 eristics) were not both cultured (84 were removed the luminal and external surfaces) chlorhexpregnated with minocycline and rifampin with those Receiving systemic antibiotics (%) 89 90 idine coordinators, and silver sulfadiazine t notification of study 19 were (on only the external impregnated with chlorhexidine and silver sulfadiaInsertion (%) surface). After their andsite subcutanecontaminated during removal, and removal, 24 were the tips zine in terms of the rates Subclavian vein 54 53 of colonization of catheters ous segments of the were catheters cultured by both ilable for other reasons) and therefore ex- wereJugular vein 38 36 and bloodstream infection. the roll-plate and the sonication methods. Peripheral11 Femoral vein 8 from further analysis. The two groups of blood cultures were obtained if clinically indicated. Duration of placement (days) METHODS rs that could be evaluated with Results were Of 865similar catheters inserted, 738 (85 percent) Mean 8.4 8.2 to characteristicsproduced of patients and catheters Patients 6 7 culture results that could Median be evaluated. The Range 1–55 1–36 1). clinical characteristics of the patients and the risk facThe trial was conducted between December 1995 and July Reason removal (%) tors for infection were similar in the for two groups. 1997 in 12 university-affiliated hospitals. The study was approved Catheter no longer needed by the appropriate 67 69 review boards. Hospitalized adults ation of Catheters Catheters impregnated with minocycline institutional and rifamSuspected catheter infection 14 13 pin were 1/3 as likely to be colonized as catheters ty-seven of 382 catheters impregnated with Occluded catheter im3 1 with chlorhexidine Other sulfadiazine 16 17 xidine and silver pregnated sulfadiazine (22.8 percent) and silver (28 of 356 catheters [7.9 percent] vs. 87 of 382 [22.8 Antiseptiques ou Antibiotiques ? Minocycline Rifampicine Chlorhexidine Sulfadiazine (n=356) (n=382) Durée séjour en réa (jours) 8,4 8,2 Colonisation, n (%) 28 (7,9) 87 (22,8 RR=2,9 Bactériémie, n (%) 1 (0,3) 13 (3,4) RR=12,5 1 (0,3) 11 (6,4) p=0,01 survenue > 7 j Incidence BLC/1000 0,3 The jNe w E n g l a nd Jo u r n a l o f Me d ic i ne 4,1 Minocycline–rifampin catheters 1.0 Proportion of Catheters! without Infection p<0,001 0.9 0.8 0.7 Chlorhexidine–silver sulfadiazine catheters 0.6 0.5 0.4 p=0,001 0.3 0.2 0.1 0.0 0 5 10 15 20 25 30 Duration of Catheterization (Days) NO. AT RISK Darouiche et al. NEJM 1999 Cathéters imprégnés … L’utilisation de cathéters imprégnés d’agents antiinfectieux diminue l’incidence des ILC. En raison de leur coût et du risque potentiel de sélection de BMR ou de levures, il n’est pas recommandé de les utiliser en 1ère intention. Ils peuvent trouver leur indication dans les unités où l’incidence des ILC reste élevée malgré la mise en place de mesures préventives. 5ème conférence de consensus – Prévention des infections nosocomiales en réanimation – SFAR - SRLF O’Grady NP et al. Guidelines for the prevention of intravascular catheter-related infections. Clin Infect Dis 2011 Entretien des lignes veineuses Le changement des lignes de perfusion peut n’être effectué que tous les 3 à 4 jours. En revanche, les tubulures doivent être changées après chaque transfusion sanguine ou quotidiennement lors de perfusions d’émulsions lipidiques. 5ème conférence de consensus – Prévention des infections nosocomiales en réanimation – SFAR - SRLF O’Grady NP et al. Guidelines for the prevention of intravascular catheter-related infections. Clin Infect Dis 2011 Autres mesures préventives... En l’absence d’efficacité, certaines mesures ne sont pas recommandées : antibioprophylaxie à l’insertion, pommade antibiotique, filtres antibactériens, boîtiers protecteurs, changement systématique du cathéter à intervalle régulier. 5ème conférence de consensus – Prévention des infections nosocomiales en réanimation – SFAR - SRLF O’Grady NP et al. Guidelines for the prevention of intravascular catheter-related infections. Clin Infect Dis 2011 Education ACADEMIA AND CLINIC Education of Physicians-in-Training Can Decrease the Risk for Vascular Catheter Infection Robert J. Sherertz, MD; E. Wesley Ely, MD, MPH; Debi M. Westbrook, RN; Kate S. Gledhill, RN; Stephen A. Streed, MS; Betty Kiger, RN; Lenora Flynn, MT; Stewart Hayes, RRT; Sallie Strong, RN; Julia Cruz, MD; David L. Bowton, MD; Todd Hulgan, MD; and Edward F. Haponik, MD V s. The second year’s supBackground: Procedure instruction for physicians-inascular catheter infection is a substantial cause training is usually nonstandardized. The authors observed $12 000. Almost all physiof morbidity and death in hospitalized patients. that during insertion of central venous catheters (CVCs), It has been estimated that 50 000 to 100 000 bloodwere fellows; most other few physicians used full-size sterile drapes (an intervention stream infections related to vascular devices occur ad salaries equivalent to proven to reduce the risk for CVC-related infection). yearly in the United States; 90% of these infections yearly salary plus benefits Objective: To improve standardization of infection conoriginate from central venous catheters (CVCs) (1). cost for trol thepractices participating and techniques during invasive procedures. The attributable mortality rate for CVC-related each day of faculty time Design: Nonrandomized pre–post observational trial. bloodstream infections ranges from 14% to 28% For the purposes of this (2– 6). The attributable cost of such infections has Setting: Six intensive care units and one step-down unit ider costsatof faculty prepbeen estimated to be as high as $29 000 per episode Wake Forest University Baptist Medical Center, Winrtunity. Assuming (4). Various interventions, including skin preparaston-Salem,that Norththe Carolina. tions withParticipants: two facultyThird-year per tion with chlorhexidine (7), use of vascular catheters medical students and physicians Figure. Effect of a procedure course on the risk for primary bloodwith anti-infective coatings (8, 9), and use of maxirse days, completing the total cost for postgraduate their first year. stream infection (white bars) and catheter-related infection (striped barrier during catheter insertion, ately $9600 for 1 year A and bars) inon sixinfection intensivecontrol care units and onemum step-down unit.precautions The course Intervention: 1-day course pracAnn to Intern Med offered twice; participants were medicalhave students andshown physicians combeen reduce risk2000;132:641-48 for catheter-related e full-size drapes givenwas ticessterile and procedures in June 1996 and June 1997. Programme multi-facette process, yielding an “exact” zero that our inter- balances (22–24). an 81% reduction in unadjusted CLABSI The appropriateness of the distribution rates at 19-month postimplementation vention would not affect. In other units, zeroes might be occasional, could be brought about by was confirmed using a Vuong test and visual and the control group, after undertaking our intervention (along with smaller and small- inspection of the data. Models were compared the intervention, achieved a 69% reducer positive rates of infection), and are thus not using the Akaike information criterion and tion 12-month postimplementation. Both considered structurally produced (21). The zero- Bayesian information criterion goodness-of-fit groups reduced CLABSIs to approximately inflated Poisson model uses two parts: it specifies statistics. All analyses used Stata 10.0 (Stata 0.8 per 1,000 line days by the end of the the conditional probability of an exact zero using Corp., College Station, TX). study. A trend of increasing numbers of a logit model to predict exact zeros and uses a Jill A. Marsteller, PhD, MPP; J. Bryan Sexton, PhD; Yea-Jen Hsu, PhD, MHA; Chun-Ju Hsiao, PhD, MHS; consistently performed infection prevenPoisson model to specify the other zeroes and the Christinecontinuous G. Holzmueller, BLA; Peter J. Pronovost, MD, PhD, FCCM; David A. Thompson, DNSc, MS, RN tion behaviors over time was also found in positive portion of the distribution. RESULTS We included ICU structural variables of system, both intervention groups (Fig. 2). Ninety-five percent (35/37) of ICUs bed size, and number of CLABSIs in year 2006 In multivariate Objectives: To determine the causal effects of an intervention the intervention and control groups (p = .28), respectively. By October analysis, model 1 demto predict exact zeros. The intervention effect from the Adventist Health System (East) a significantly lower adjusted Lavage mains proven effective in pre-post studies in reducing central line-asso- to December 2007,des the infection rate declined toonstrates 1.33 in the intervenwas assessed in the nonexact zero part where and 59% (10/17) Adven-groupincidence ciated bloodstream infections in the intensive care unit. we participated tion group compared to 2.16 from in the control (adjusted incidence rate ratio for CLABSI in the Asepsie chirurgicale infections in the postimplementation tist-Health not interval Design:regressed We conducted a multicenter, phased, cluster-randomrate ratio(West) 0.19; p (Fig. = .003;1). 95%Those confidence 0.06–0.57). group The intervention postimplementation periodtrial (with an offset for linewere days)randomized on baselineinto two intervention group sustained rates <1/1,000 central line days at 19 ized controlled in which hospitals Chlorhexidine participating in the East system were compared to the control group (adjusted infections and group groupstarted assignment. The2007 offset forthe con- months (an 81% reduction). The control group also reduced infection groups. The intervention in March and extremely small orsite recently opened; rea- incidence rate ratio 0.19; p = .003, 95% -rates Eviter fémoral number of central for theSeptemtrol group the started in October 2007;line the days studyadjusted period ended to <1/1,000 central line days (a 69% reduction) at 12 months. sons for nonparticipation were not col- confidence interval 0.06–0.57; Table 3). risk of infection facedgroups by each ber 2008. Baseline data for both areICU. fromBecause 2006. a - Retrait Conclusions: This study demonstrated a causal relationship cathéters inutiles lectedbetween from the system.intervention The overalland the Thisreduced corresponds regression offset the in two Setting:Poisson Forty-five intensivewith carethe units fromconverts 35 hospitals theWest multifaceted central to a reduction of 81% into the log of the count divided by the log participation rate was 83% (45/54). CLABSI rates in the intervention group Adventist count healthcare systems. line-associated bloodstream infections. Bothingroups decreased Interventions: A multifaceted intervention involving evidence- infection rates after implementation and sustained these results based practices to prevent central line-associated bloodstream over time, replicating the results found in previous, pre-post stud2. Comprehensive Central line-associated bloodstream infectiontorateies group comparisons infectionsTable and the Unit–based Safety Program of this multifaceted intervention and providing further evidence improve safety, teamwork, and communication. that most central line-associated bloodstream infections are preIntervention Group (Crit (n = Care 23) Med 2012; 40:2933–2939) Control Group (n = 22) Measurements and Results: We measured central line-associventable. KEY WORDS: catheter-related infections; evidence-based pracated bloodstream infections per 1,000 central line days and reported CLABSI Rate CLABSI rate quarterly rates. Baseline average central line-associated blood- tice; intensive care units; prevention and control; quality improvestream infections per 1,000 central line days was 4.48 and 2.71, for ment; randomized controlled trial Median Incidence A multicenter, phased, cluster-randomized controlled trial to reduce central line-associated bloodstream infections in intensive care units* N Median (Interquartile Incidence % of (Interquartile Rate % of Range) Mean (SD) Rate Ratio Reductiona Range) Mean (SD) Ratio Reductiona early 250,000 healthcare- life-saving medical care (1). Furthermore, known as the Comprehensive Unit–based associated infections 25% of patients 4.48 contracting a central Safety Program (CUSP) (13). Together Baseline (Jan-Dec 2006) occur 2.56 (0.74–5.87) (7.03) 1.00 1.78 (0.00–3.80) 2.71 (3.01) 1.00 b annually in patients with 0.00 line-associated bloodstream these 5interventions reduced the overall (0.00–1.79) 4.71 (17.31) infection 1.05 —— Mar 2007 central lines placed First Q (Apr-Jun 2007) to deliver 0.00 (0.00–2.02) 1.12 (2.34) −80 rate by 66% in — a cohort of ICUs — (CLABSI) in the intensive care unit0.25 (ICU) CLABSI Second Q (Jul-Sep 2007) 0.00 (0.00–2.59) (4.25) 16 — die, totaling 31,0001.83 deaths annually 0.41 in the (7). Nevertheless, this— cohort study, with 0.00 (0.00–1.50) 1.33 (2.77) review0.30 −11 0.00 (0.00–2.89) 2.16 (3.93) 0.79 −21 Third Q (Oct-Dec 2007)c United States (2). A recent estino concurrent control group, was not able *See also p. 3083. Fourth Q (Jan-Mar 2008) 0.00 (0.00–0.00) 0.96 (2.78) 0.21 −8 0.00 (0.00–0.00) 0.56 (1.35) 0.21 −59 mated an added annual cost of $9 billion to establish a causal relationship between From the Department of Health Policy and Fifth Q (Apr-Jun 2008) 0.00 (0.00–0.00) 0.88 (2.19) 0.20 −2 0.00 (0.00–0.00) 0.52 (1.46) 0.19 −2 Management (JAM, JBS, Y-JH, C-JH, PJP), Johns to the U.S. healthcare system (3). the intervention and the reduced CLABSI Sixth Q (Jul-Sep 2008) 0.85 (1.39) 0.19 −1 0.00 (0.00–0.00) 0.83 (2.30) 0.31 12 Hopkins Bloomberg School of Public Health, Baltimore, 0.00 (0.00–1.94) Previous quality improvement studies rate.−81 The rational next scientific step was CLABSI rate reduction from −69 MD; and Department of Anesthesiology and Critical suggest that these infections are largely to test a causal relationship between this Care Medicine baseline (JAM, JBS, CGH, PJP, DAT), Johns to sixth quarter preventable (4–8). However, these studies multifaceted intervention and reduced Hopkins School of Medicine, Baltimore, MD. The work was performed at Adventist Health and were based infection; on nonrandomized withnot collected. CLABSI rates in a randomized controlled CLABSI, central line-associated bloodstream Q, quarter;trials —, data the Adventist Health System. Crit Care Med reductions 2012;40:2933-39 historical or contemporaneous controls. trial (RCT) to evaluate theto magnitude of (percentage a Calculated as differences of CLABSI rate or incidence rate ratio between two consecutive quarters relative baseline values are Supported by a grant from the Robert Wood Au total : quelle méthode de prévention ? • Lavage des mains • Asepsie : chirurgicale • Antisepsie cutanée : chlorhexidine ou PVI-alcool • Site : Sous-clavier > Jugulaire interne > Fémoral • Nombre de voie : multi-lumière = mono-lumière • Matériau : Polyuréthane ou élastomères de silicone • Cathéters imprégnés : pas en 1ère intention • Pansement : tous les 3 jours • Entretien des lignes : tous les 3-4 jours • Retrait des cathéters inutiles • Education 5ème conférence de consensus – Prévention des infections nosocomiales en réanimation – SFAR - SRLF O’Grady NP et al. Guidelines for the prevention of intravascular catheter-related infections. Clin Infect Dis 2011