Contribution of tap water to patient colonisation with Pseudomonas aeruginosa in a medical intensive care unit

Journal of Hospital Infection (2007) 67, 72e78

www.elsevierhealth.com/journals/jhin

Contribution of tap water to patient colonisation with Pseudomonas aeruginosa in a medical intensive care unit ´ras b, A. Boyer c, A.-M. Rogues a,b,*, H. Boulestreau a,b, A. Lashe ´bear d, J.-P. Gachie a,b D. Gruson c, C. Merle b, Y. Castaing c, C.M. Be a

Unite´ INSERM 657, Universite´ Victor Segalen Bordeaux 2, Bordeaux, France Service d’Hygie`ne Hospitalie`re Groupe Hospitalier Pellegrin, Bordeaux, France c Service de Re´animation Me´dicale Groupe Hospitalier Pellegrin, Bordeaux, France d Service de Bacte´riologie Groupe Hospitalier Pellegrin, Bordeaux, France b

Received 9 October 2006; accepted 8 June 2007 Available online 28 August 2007

KEYWORDS Pseudomonas aeruginosa; Critical care; Epidemiology; Tap water samples; Cross-infection; Nosocomial infection

Summary This study examined tap water as a source of Pseudomonas aeruginosa in a medical intensive care setting. We prospectively screened specimens of patients, tap water and hands of healthcare workers (HCWs) over a six-month period in a 16-bed medical intensive care unit. Molecular relatedness of P. aeruginosa strains was investigated by pulsed-field gel electrophoresis. A total of 657 tap water samples were collected from 39 faucets and 127 hands of HCWs were sampled. P. aeruginosa was found in 11.4% of 484 tap water samples taken from patients’ rooms and in 5.3% of 189 other tap water samples (P < 0.01). P. aeruginosa was isolated from 38 patients. Typing of 73 non-replicate isolates (water samples, hands of HCWs and patients) revealed 32 major DNA patterns. Eleven (52.4%) of the 21 faucets were contaminated with a patient strain, found before isolation from tap water in the corresponding room in nine cases, or from the neighbouring room in two cases. Among seven P. aeruginosa strains isolated from HCW hands, the genotype obtained was the same as that from the last patient they had touched in six cases, and in the seventh with the last tap water sample used. More than half of P. aeruginosa carriage in patients was acquired via tap water or cross-transmission. Carriage of P. aeruginosa by patients was both the source and the consequence of

* Corresponding author. Address: Service d’Hygie `ne Hospitalie `re, Groupe Hospitalier Pellegrin, Place Amelie Raba Le ´on, 33076 Bordeaux, France. Tel.: þ33 5 5679 5553; fax: þ33 5 5679 4997. E-mail address: [email protected] 0195-6701/$ - see front matter ª 2007 The Hospital Infection Society. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.jhin.2007.06.019

P. aeruginosa and faucets in intensive care

73

tap water colonisation. These results emphasise the need for studies on how to control tap water contamination. ª 2007 The Hospital Infection Society. Published by Elsevier Ltd. All rights reserved.

Introduction Pseudomonas aeruginosa is associated with nosocomial infection, especially among mechanically ventilated or immunocompromised patients in intensive care units (ICUs).1e5 Within hospital settings, several P. aeruginosa outbreaks have been linked to contaminated tap water.6e8 Most hospital infections are endemic, and although several studies report the major reservoir of P. aeruginosa as ICU patients’ endogenous flora, other studies have shown patient-to-patient spread via hands of healthcare workers (HCWs) or via fomites. The role of tap water as a source or reservoir for P. aeruginosa remains controversial.9e13 Knowledge of the relative significance of different transmission routes is important for the design of optimal infection control strategies for ICUs. The aim of this study was to further elucidate the importance of tap water as a propagating source for Pseudomonas aeruginosa in a medical ICU setting.

Methods Design and study population We monitored patients, tap water and hands of HCWs for contamination with P. aeruginosa over a six-month period in a 16-bed medical ICU at a large teaching hospital (27 weeks between May and November 2003). This ICU mainly admitted patients with respiratory failure and w80% of them received ventilator treatment. The nurse:patient ratio was 1:4. Each single room had its own washbasin with a conventional water tap and the same water source, from the central piping system, supplied all 39 washbasins or sinks of the ICU. In July (week 11 of the study), due to patient colonisation with P. aeruginosa probably caused directly or indirectly by transmission of strains from tap water, we attempted to eliminate the organism with twice monthly chlorine disinfection. An aqueous solution (4.5%) of sodium hypochlorite (diluted household bleach) was injected into taps with a 60 mL syringe for 15 min. Furthermore, aerators were removed every two weeks, immersed and brushed in a detergent-disinfectant solution. Once

the disinfection programme had been instituted, samples were obtained prior to the disinfection procedure. Hand disinfection with an alcohol-based solution was required between patient contacts. Only bottled water was used for enteral nutrition and to administer drugs through gastric tubes. Bottled water is not sterile but analyses performed every year on bottles used for immunocompromised patients in another unit were always satisfactory. Sterile water was used for mouth care. An outbreak of imipenem-resistant P. aeruginosa involving patients undergoing bronchoscopy was detected in our hospital. This was related to a defective flexible bronchoscope used in another ICU, which remained contaminated with an epidemic strain after manual reprocessing. This bronchoscope was removed on week 8 of the study.

Surveillance culture All patients admitted to the ICU during the study period were included. Surveillance (throat and rectal swabs, sputum and urine) specimens were screened for P. aeruginosa. Samples were collected on admission and then weekly on pre-defined days until discharge from the ICU or death. Other specimens were collected as clinically indicated. Culture and identification of P. aeruginosa were performed using standard techniques with non-selective plates. All clinical isolates of P. aeruginosa were stored throughout the study period. Demographic and epidemiological data (admission and discharge dates, room number) were collected prospectively. Environmental specimens were taken by the infection control team every week from tap water in patients’ rooms and every three weeks from other taps, and samples examined for the presence of P. aeruginosa. Cold water samples were taken with an aerator in place. Taps were opened and the first 250 ml flush of water collected immediately into a sterile flask with sodium thiosulphate. The aerator was swabbed and the swab broken into the water samples. Hands of HCWs were cultured for P. aeruginosa three times a month using a modification of the ‘glove juice method’, without prior warning. Only nursing staff working directly with patients were sampled. The predominant hand of the worker

74 was inserted into a sterile polyethylene bag containing 50 mL of sampling solution [3% (v/v) Tween 80, 0.1% (w/v) sodium thiosulphate, 0.1% L-histidine (w/v) and 0.3% (w/v) lecithin]. His/her hand was massaged by an infection control practitioner through the wall of the bag for 15 to 30 s, and samples delivered to the microbiology laboratory within 1 h for processing.14 HCWs were asked for the last faucet contact and the name of the last patient contact before sampling. HCWs’ hands and tap water samples were processed by membrane filtration. A 100 ml volume was filtered through a 0.45 mm pore size membrane filter (Millipore Microfil, France). Swabs and filters were cultured onto cetrimide agar plates (BioRad, Marnes-la-Coquette, France) at 37  C and examined for growth after 24 and 48 h. Any colonies that grew on cetrimide agar were identified using the API20 NE identification system (bioMe ´rieux, Marcy l’Etoile, France). Serotyping was performed from morphologically distinguishable colonies by slide agglutination using commercial antisera (BioRad, Marnes-la-Coquette, France).

Genotyping The first isolate of each serotype and non-serotypable isolates found in tap water or in a patient was selected for genotyping analysis. Two patients’ strains isolated were not available (week 23). P. aeruginosa strains were genotyped by pulsed-field gel electrophoresis (PFGE) with Spe1 as described by Talon et al. and compared using the Genepath System, group 3 reagent kit (Bio-Rad Laboratories, Hercules, CA, USA), including the GenePath control strain. Gels were stained and photographed and DNA patterns compared and interpreted according to criteria previously described.15,16 Analysis of PFGE profiles was made with the software Fingerprinting II (Bio-Rad, France).

Definitions Cross-transmission was considered possible when indistinguishable isolates were found in more than two patients hospitalised during overlapping periods. P. aeruginosa was considered as originating (directly or indirectly) from the tap water when the isolate(s) from a patient was (were) of the same PFGE genotype as one previous isolate recovered from the room where the patient was hospitalised. Exogenous colonisation was defined as colonisation by a strain of P. aeruginosa with a pulsotype previously isolated from another patient, a HCW’s hand or tap water.

A.-M. Rogues et al.

Statistical analysis The Chi-square test was used to compare observed with expected proportions (Epi Info, version 6.0; CDC, Atlanta, GA, USA). The significance was set at P < 0.05.

Results A total of 657 cold water samples were obtained from taps and 127 HCW hand samples. P. aeruginosa was found in 55 (11.4%) of 484 tap water samples taken from patients’ rooms and in 10 (5.3%) of 189 other tap water samples (P < 0.01). Only three taps (18.8%) from patients’ rooms were consistently negative as opposed to 17 (81%) of the others. Seven of the 39 taps checked were positive only once during the study. P. aeruginosa was found in 34 out of 180 (18.8%) samples before and in 22 of 288 (7.6%) after disinfection was implemented (P < 0.01). P. aeruginosa was identified in seven of 127 (5.5%) HCW hand samples; four were sampled whilst the HCW was in a patient’s room. The percentage of positive hand specimens was 14% (3/ 21) when the last faucet contact was with positive tap water, and 3.8% (4/106) when the last contact was with negative tap water (P < 0.01). When the last patient contact was with a positive patient, P. aeruginosa was identified in 12% (7/59) of the hand samples and never occurred when the last contact was a negative patient (0/68). During the study period, among the 415 patients admitted to ICU, only 153 stayed more than 72 h; 38 of them were P. aeruginosa carriers. Six were colonised at the beginning of the study. After implementation of tap disinfection, the number of P. aeruginosa carriers decreased: 21 in the 10 weeks before implementation, with 10 probably acquired (incidence of 10 per 1000 patient-days) and 15 in the 17 weeks after implementation, with five probably acquired (incidence of 3.5 per 1000 patient-days). Isolates from 36 patients, 30 tap water samples and hands of seven HCWs were available for molecular typing. A total of 32 pulsotypes as determined by PFGE were detected among the 73 strains studied: nine pulsotypes were detected in both clinical and tap water strains, 14 from patient isolates only and nine from tap water samples only. A unique pulsotype was found for 17 strains. Clinical strains showed 21 pulsotypes; six of them were isolated from more than two patients. One pulsotype related to a bronchoscope was detected in three patients. Eighteen patients

P. aeruginosa and faucets in intensive care

75

harboured strains which were genotypically identical to those recovered from one of the ICU faucets during the study. Chronological epidemiological analysis and PFGE results shown in Figure 1 suggested transmission from tap water to patient in seven cases of the 15 strains identified 72 h after patient’s admission to the ICU. Six patients harboured a pulsotype undetected in water but found in at least one other patient during the same stay suggesting cross-transmission. Faucets in the ICU rooms had been contaminated with a patient strain in 11 (52.4%) of the 21 positive tap water samples. The same pulsotype was found in a patient before isolation from tap water in the corresponding rooms in nine cases or in the neighbouring rooms in two cases. Among the seven P. aeruginosa isolated from hands of HCWs, the same pulsotype was obtained with the last patient with whom they were in contact in six cases, and for the remaining case with the last tap water used. Two pulsotypes were present in the ICU throughout the study period. One of these was found in eight isolates (Pulsotype 7), first identified from a patient, and subsequently in five tap water samples

and in two other patients. The other one (Pulsotype 18), found in eight isolates, was initially related to a bronchoscope in three cases and was then identified in one hand, two patients and in three tap water samples.

Discussion Pseudomonas aeruginosa is a common cause of hospital-acquired infection in ICUs. In contrast to outbreak situations in which various sources of P. aeruginosa have been identified, it is more difficult to trace the reservoirs of the pathogen in endemic situations. We monitored patients, tap water and HCWs’ hands for contamination with P. aeruginosa in a medical ICU that mainly admitted patients with respiratory failure. We showed that P. aeruginosa was found most often in tap water samples in patients’ rooms than in other tap water of the unit. Half of the environmental isolates of P. aeruginosa derived from colonised patients and did not stem from a central source in the supply mains. Carriage of P. aeruginosa by patients was both the source

Disinfection

21

21

1

26 7

2 7

4

7

4

4

3

7 7 18

4 4

18

5 18

4

Patient room no.

6 25

4

7

25

8

26 7

9 21 10 1 11

21 20

20 18

18

12 1

1 18

1

13 18 14

18

15 16

24

1

24

24

2

3

4

5

6

7

8

9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 Week

Figure 1 Chronology of Pseudomonas aeruginosa genotypes (N ¼ 9) isolated from patients (numbers in circles) and from tap water (numbers in stars) in different patient rooms (horizontal lines represented patient’s individual stay).

76 and the consequence of tap water colonisation. Both water-related and non-water-related strains appeared to have spread in half of the instances. In several studies using molecular typing in a non-epidemic ICU setting, P. aeruginosa was reported to originate from the patient’s endogenous flora.9e11,17e19 In a Dutch ICU, Bonten et al. prospectively investigated patients’ colonisation and infection with P. aeruginosa. They concluded that respiratory tract colonisation was exogenous in origin in only 8% of cases.19 In a French ICU, 21 of 26 cases of acquired lung colonisation with P. aeruginosa resulted from endogenous colonisation and the remaining five from cross-transmission.10 However, results of the studies comparing genotypes of endemic P. aeruginosa strains isolated from patients and tap water outlets varied according to the environmental sampling technique.13 This could explain why our results were similar to those studies which used a sensitive technique for sampling water with an aerator in place and swabbing of the tap.20,21 More recently, similar results were observed in a medical ICU where 50% of patient isolates were genotypically identical to strains isolated previously from water of the same or adjacent rooms.22 We may have underestimated the number of exogenous sources as we did not perform extensive microbiological sampling of the environment other than tap water. In particular, we did not analyse the bottled water used in the unit, and outbreaks due to various exogenous equipment or fibroscopes have been described.23 Our microbiological analysis could also have underestimated exogenous colonisation because, firstly, we selected isolates for genotyping with serotyping, and secondly, we only performed genotyping from one colony of each positive culture. In a study with more than 1600 isolates of P. aeruginosa over a period of three years, Valle `s et al. found that more than 60% of tap water samples in his ICU were contaminated by P. aeruginosa and overall 83% of patient strains were classified as exogenous. They subcultured at least four colonies representative of the different morphological types of P. aeruginosa present on each culture plate.24 The results of molecular typing should always be interpreted with traditional epidemiology data. In the nosocomial setting, an overlapping period of hospitalisation in the same ward or in the same room, or other common epidemiological features (for example, a common exogenous exposure such as ventilator equipment or bronchoscopy) must be considered. Reuter et al. reported that in their ICU, 35% of all cases of acquired colonisation with P. aeruginosa originated from contaminated tap

A.-M. Rogues et al. water and that retrograde contamination of faucets by patients occurred in 15% of cases. In this study, the corresponding tap water was examined retrospectively after isolation of P. aeruginosa from a patient hospitalised in a surgical department; P. aeruginosa was found in 58% of tap water samples taken from patients’ rooms.20 In a threeyear prospective study, P. aeruginosa was isolated from the rooms’ tap water in 62.4% of samples. In addition, more than 90% of tap water samples had pulsotypes which were frequently isolated from the patient’s stomach (59%), so requiring a change in infection control strategy using mineral water only through the gastric tube.24 In our study contamination of taps by a patient strain occurred in half of positive tap water samples. However, the exact modes of retrograde contamination have not been explored. Colonisation of tap water may be initiated by splashing the faucet when water is drawn (e.g. during handwashing) or when something is poured into the basin.25 Following the implementation of a disinfection procedure, the proportion of tap water isolates containing P. aeruginosa and the number of patients colonised with P. aeruginosa decreased. However, our study design did not allow us to conclude that this was a direct effect of disinfection since many others factors could have influenced patient acquisition of P. aeruginosa. There were other possible environmental reservoirs for the bacterium in this medical ICU, especially during the first part of the study (e.g. bronchoscope or other equipment). In some outbreaks, regular cleaning and chlorine disinfection of sinks was shown to reduce the number of hospital infections due to P. aeruginosa.26,27 Nevertheless, as a routine measure, the efficacy of regular disinfection to control retrograde contamination has not been proven and recent reports note how difficult it is to reduce colonisation in tap water, especially with electronic faucets.12,20,27 The incidence rate of 6% hands colonised by P. aeruginosa was similar to that in two previous studies.14,24 Our results suggest that hand contamination seems to be more frequently due to inadequate hand disinfection after patient contact than to direct contact with contaminated tap water. However, in hospital settings where sinks may be a reservoir of nosocomial pathogens, there is all the more reason to promote compliance with standard precautions and alcohol hand disinfection in order to limit the spread of P. aeruginosa.27e30 Petignat et al. suggested that the observed decrease in incidence of cases of P. aeruginosa in an intensive care unit (from 59 patients per 1000 admissions in 1998 to 26 patients per 1000 admissions in 2000) was related to implementation

P. aeruginosa and faucets in intensive care of infection control measures, confirming that P. aeruginosa strains had been of exogenous origin in a substantial proportion of patients during the pre-intervention period.31 In conclusion, the dynamics of colonisation with P. aeruginosa in ICUs are complex and may differ considerably between units. Carriage by patients could be both the source and consequence of faucet contamination. These results emphasise the need for adequate and effective hand disinfection after patient care and after hand washing, but also for studying how to control tap water retrograde contamination. Further studies are needed since sampling technique and microbiological analysis of tap water could underestimate exogenous colonisation.

77

10.

11.

12.

13.

14.

Conflict of interest statement None. Funding sources This study was supported by the Conseil Re ´gional de d’ Aquitaine, France.

References 1. Fagon JY, Chastre J, Hance AJ, Montravears P, Novara A, Gilbert C. Nosocomial pneumonia in ventilated patients: a cohort study evaluating attributable mortality and hospital stay. Am J Med 1993;94:281e288. 2. Thuong M, Arvaniti K, Ruimy R, et al. Epidemiology of Pseudomonas aeruginosa and risk factors for carriage acquisition in intensive care unit. J Hosp Infect 2003;53:274e282. 3. Bukholm G, Tannaes T, Britt Bye Kjelsberg A, SmithErichsen N. An outbreak of multidrug-resistant Pseudomonas aeruginosa associated with increased risk of patient death in an intensive care unit. Infect Control Hosp Epidemiol 2002; 23:441e446. 4. Osmon S, Ward S, Fraser VJ, Kollef MH. Hospital mortality for patients with bacteremia due to Staphylococcus aureus or Pseudomonas aeruginosa. Chest 2004;125:607e616. 5. Aloush V, Navon-Venezia S, Seigman-Igra Y, Cabili S, Carmeli Y. Multidrug-resistant Pseudomonas aeruginosa: risk factors and clinical impact. Antimicrob Agents Chemother 2006;50:43e48. 6. Kolmos HJ, Thuesen B, Nielsen SV, Lohmann M, Kristffersen K, Rosdahl VT. Outbreak of infection in a burns unit due to Pseudomonas aeruginosa originating from contaminated tubing used for irrigation of patients. J Hosp Infect 1993;24:11e21. 7. Ferroni A, Nguyen L, Pron B, Quesne G, Brusset MC, Berche P. Outbreak of nosocomial urinary tract infections due to Pseudomonas aeruginosa in a paediatric surgical unit associated with tap-water contamination. J Hosp Infect 1998;39:301e307. 8. Bert F, Maubec E, Bruneau B, Berry P, Lambert-Zechovsky N. Multi-resistant Pseudomonas aeruginosa outbreak associated tap water in neurosurgery intensive care unit. J Hosp Infect 1998;39:53e62. 9. Bertrand X, Thouverez M, Talon D, et al. Endemicity, molecular diversity and colonisation routes of Pseudomonas

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

aeruginosa in intensive care units. Intensive Care Med 2001;27:1263e1268. Berthelot P, Grattard F, Mahul P, et al. Prospective study of nosocomial colonization and infection due to Pseudomonas aeruginosa in ventilated patients. Intensive Care Med 2001; 27:503e512. Speijer H, Savelkoul PHM, Bonten MJ, Stobberingh EE, Tjhie JHT. Application of different genotyping methods for Pseudomonas aeruginosa in a setting of endemicity in an intensive care unit. J Clin Microbiol 1999;37:3654e3661. Trautmann M, Michalsky T, Wiedeck H, Radosavljevic V, Ruhnke M. Tap water colonization with Pseudomonas aeruginosa in a surgical intensive care unit and relation to Pseudomonas infections of ICU patients. Infect Control Hosp Epidemiol 2001;22:49e52. Trautmann M, Lepper PM, Haller M. Ecology of Pseudomonas aeruginosa in the intensive care unit and the evolving role of water outlets as a reservoir of the organism. Am J Infect Control 2005;33:S41eS49. Foca M, Jakob K, Whittier S, et al. Endemic Pseudomonas aeruginosa infection in a neonatal intensive care unit. N Engl J Med 2000;343:696e700. Talon D, Cailleux V, Thouverez M, Michel-Briand Y. Discriminatory power and usefulness of pulsed field gel electrophoresis in epidemiological studies of Pseudomonas aeruginosa. J Hosp Infect 1995;30:125e131. Tenover F, Arbeit RD, Coering RV. How to select and interpret molecular strain typing methods for epidemiological studies of bacterial infections. A review for health care epidemiologists. Infect Control Hosp Epidemiol 1997;18:426e439. Bergmans D, Bonten M, van Tiel F, et al. Cross-colonisation with Pseudomonas aeruginosa of patients in an intensive care unit. Thorax 1998;53:1053e1058. Kropec A, Huebner J, Riffel M, et al. Exogenous or endogenous reservoirs of nosocomial Pseudomonas aeruginosa and Staphylococcus aureus infections in a surgical intensive care unit. Intensive Care Med 1993;19:161e165. Bonten MJM, Bergmans DCJJ, Speijer H, Stobberingh EE. Characteristics of polyclonal endemicity of Pseudomonas aeruginosa colonization in intensive care units. Am J Respir Crit Care Med 1999;160:1212e1219. Reuter S, Sigge A, Wiedeck H, Trautmann M. Analysis of transmission pathways of Pseudomonas aeruginosa between patients and tap water outlets. Crit Care Med 2002;30(10): 2222e2228. Blanc DS, Nahimana I, Petignat C, Wenger A, Bille J, Francioli P. Faucets as a reservoir of endemic Pseudomonas aeruginosa colonization/infection in intensive care units. Intensive Care Med 2004;30:1964e1968. Trautmann M, Bauer C, Schumann C, et al. Common RAPD pattern of Pseudomonas aeruginosa from patients and tap water in a medical intensive care unit. Int J Hyg Environ Health 2006;209:325e333. Blanc DS, Parret T, Janin B, Raselli P, Francioli P. Nosocomial infections and pseudoinfections from contaminated bronchoscopes: two-year follow up using molecular markers. Infect Control Hosp Epidemiol 1997;18:134e136. Valle ´s J, Mariscal D, Corte ´s P, et al. Patterns of colonization by Pseudomonas aeruginosa in intubated patients: a 3-years prospective study of 1607 isolates using pulsed-field gel electrophoresis with implications for prevention of ventilator-associated pneumonia. Intensive Care Med 2004;30: 1768e1775. Pitten FA, Panzig B, Schro ¨der G, Tietze K, Kramer A. Transmission of a multiresistant Pseudomonas aeruginosa strain at a German University Hospital. J Hosp Infect 2001;47: 125e130.

78 26. Hargreaves J, Shireley L, Hansen S, et al. Bacterial contamination associated with electronic faucets: a new risk for healthcare facilities. Infect Control Hosp Epidemiol 2001; 22:202e205. 27. Merrer J, Girou E, Ducellier D, et al. Should electronic faucets be used in intensive care and haematology units? Intensive Care Med 2005;31:1715e1718. 28. Widmer AF, Wenzel RP, Trilla A, Bale MJ, Jones RN, Doebbeling BN. Outbreak of Pseudomonas aeruginosa infections in a surgical intensive care unit: probable transmission via hands of a health care worker. Clin Infect Dis 1993;16: 372e376.

A.-M. Rogues et al. 29. Moolenaar RL, Crutcher JM, San Joaquin VH, et al. A prolonged outbreak of Pseudomonas aeruginosa in a neonatal intensive care unit: did staff fingernails play a role in disease transmission? Infect Control Hosp Epidemiol 2000;21:80e85. 30. Anaissie AJ, Penzag SR, Dignani C. The hospital water supply as a source of nosocomial infections. A plea for action. Arch Intern Med 2002;162:1483e1492. 31. Petignat C, Francioli P, Nahimana I, et al. Exogenous sources of Pseudomonas aeruginosa in intensive care unit patients: implementation of infection control measures and follow-up with molecular typing. Infect Control Hosp Epidemiol 2006;27:953e957.