JAC Advance Access published online on July 17, 2008
Journal of Antimicrobial Chemotherapy, doi:10.1093/jac/dkn289
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Original research |
Prediction models to identify hospitalized patients at risk of being colonized or infected with multidrug-resistant Acinetobacter baumannii calcoaceticus complex
1 Department of Infectious Diseases, Università Cattolica Sacro Cuore, 00168 Rome, Italy 2 Department of Microbiology, Università Cattolica Sacro Cuore, 00168 Rome, Italy 3 Hospital Head Unit, Policlinico Agostino Gemelli, 00168 Rome, Italy 4 Anaesthesiology and Intensive Care Medicine, Università Cattolica Sacro Cuore, 00168 Rome, Italy
* Corresponding author. Tel: +39-06-30155527; Fax: +39-06-3054519; E-mail: etacconelli{at}rm.unicatt.it
Received 6 March 2008; returned 8 April 2008; revised 18 June 2008; accepted 20 June 2008
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionFundingTransparency declarationsReferences |
|---|
Background: The multidrug-resistant (MDR) Acinetobacter baumannii calcoaceticus complex (Abc) has emerged as an important cause of nosocomial infections. The aims of the study were to evaluate risk factors for MDR-Abc in intensive care units (ICUs) as well as in medical and surgical wards, to define the likelihood ratios (LRs) of risk factors and to determine if risk factors differ depending on whether colonization or infections are considered.
Methods: Two prospective matched case–control studies were performed. MDR-Abc was defined as a strain resistant to four or more classes of antibiotics. The two case groups included patients with MDR-Abc infections or colonization. Controls were selected among patients not harbouring Abc. Matching criteria were the number of days from admission to MDR-Abc isolation among cases and the duration of hospitalization among controls.
Results: Overall, 514 patients were included in the study. One hundred and thirty-seven patients were infected and 120 colonized. A Charlson score >3 and previous methicillin-resistant Staphylococcus aureus isolation and β-lactam use were independent risk factors for colonization and infection. Bedridden status and previous ICU admission were associated with colonization, while the presence of a central venous catheter and surgery were related to infection. The analysis of LRs showed an association between the presence of more than two risk factors and colonization or infection. The highest predicting value was observed for the presence of more than two risk factors and colonization in patients with no history of ICU admission.
Conclusions: This study provides novel information that can be used to identify interventions for different stages of the spread of MDR-Abc.
Key Words: nosocomial infections , antibiotic resistance , likelihood ratio
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionFundingTransparency declarationsReferences |
|---|
Among nosocomial pathogens, in recent years, genospecies within the Acinetobacter baumannii calcoaceticus complex (Abc) have been increasingly implicated in severe nosocomial infections, including bloodstream infections and ventilator-associated pneumonia.1–3 Treatment of these infections can be particularly difficult because of the frequency of drug resistance to multiple antibiotics, including aminoglycosides, quinolones and carbapenems.1,4,5 Ampicillin–sulbactam and colistin are frequently the only drugs available to treat multidrug-resistant Abc (MDR-Abc) infections.6,7 A systematic review has provided evidence that the infections are associated with a mortality rate in excess of 24% and an additional hospital stay of 14 days.8
Previous research failed to explain how the bacteria reach and colonize patients, and how they avoid the host's defence mechanisms to create infection.9 Previous faecal colonization seems to increase the risk of infection.10 It is still not clear which body site should be cultured to detect carriage. Marchaim et al.11 demonstrated that the sensitivity of culturing body sites is low, even if multiple sites were cultured.
Many risk factors were found to be associated with MDR-Abc infection and colonization: admission to intensive care units (ICU),7 intra-tracheal devices,12 mechanical ventilation,13 central venous catheters (CVCs),14 enteral feeding,15 severe underlying conditions,14 colonization of health workers’ hands and skin16 and antibiotic treatment.7,12,13,17–19 However, studies have produced different results due to the lack of methodological standards. Previous studies mainly reported retrospective small series of patients. Data were usually gathered in the context of an outbreak in an ICU population without differentiating between infected and colonized patients. Control groups were not always selected independently of their exposure status. A systematic review, although limited by significant heterogeneity among studies, reported that the acquisition and spread of the microorganism appeared to be related mainly to deficiencies in infection control measures and to the use of carbapenems and third-generation cephalosporins.20 Falagas et al.20 also suggested that the development of innovative strategies was needed in order to limit Abc hospital spread.
Therefore, we performed a prospective study to better understand the epidemiology of MDR-Abc. The objectives of the study were to evaluate risk factors for MDR-Abc acquisition in ICU as well as in the medical and surgical wards, to define the likelihood ratios (LRs) of risk factors to be transferred in different epidemiological settings and to determine if risk factors differ depending on whether colonization or infection is considered.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionFundingTransparency declarationsReferences |
|---|
Study site, subjects and design
The study was performed at the Agostino Gemelli Hospital, Catholic University of Medicine in Rome, Italy, a 1850 bed tertiary care teaching hospital with 
60 000 discharged patients per year. It includes a transplantation service, haematology/oncology units, many surgical wards and an 18 bed general ICU that admits surgical, medical and trauma patients.
From January 2003 to November 2005, all clinical cultures yielding Abc processed by the microbiology laboratory at our hospital were identified through a daily review of the microbiology computerized databases. All adult patients (
18 years of age) with an isolation of MDR-Abc were included in the study. Patients were enrolled only once, during the first recovery of MDR-Abc. Infected or colonized patients were all isolated and measures were taken to prevent contact with the multiresistant strain according to local and international protocols.21
To assess potential risk factors associated with MDR-Abc colonization and infection, two prospective matched (ratio 1:1) case–control studies were performed. In the first case–control study, the case was defined as a patient with MDR-Abc infection. Centers for Diseases Control (CDC) criteria were used to define nosocomial infections.22 In the second case–control study, the case was defined as a patient with MDR-Abc colonization. In the absence of supporting clinical and laboratory signs of infection and in the absence of specific therapy, patients with a positive culture for MDR-Abc were considered to be colonized.2 Since screening samples for MDR-Abc were not routinely performed in our institution during the study period, patients with positive screening samples were excluded from the analysis (nine patients). Because of the lack of a standard definition for multidrug resistance in the literature, we defined multidrug resistance as resistance to at least four of the following antibiotic classes commonly prescribed for the treatment of Gram-negative bacterial infections: carbapenems, antipseudomonal cephalosporins, β-lactam/β-lactamase inhibitor combinations, quinolones and aminoglycosides.2
The matching criteria chosen for our case–control study were based on the need to control for spatial-temporal factors such as hospital ward, length of hospital stay and calendar time. Therefore, cases and controls were matched according to the following criteria: length of hospital stay at the time of matching, hospital location and month of hospitalization. Controls were selected from a list of patients admitted for a long time period (at the time of matching) as the infection-free days for cases and in whom MDR-Abc were not isolated.
All patients’ medical records were reviewed. A standardized survey was administered by an infectious disease fellow at the study enrolment and included patient demographics; date of hospital admission, hospital service, presence of a CVC; previous hospitalization within 6 months; ambulatory status; history of neoplastic disease or organ transplantation; ICU admission; and surgical procedures within 30 days. A composite score of co-morbidity was derived using the Charlson index score.23 A value of APACHE III score was reported for patients admitted to ICU.24 Antibiotics administered during a 30 day period prior to the study enrolment and for at least 48 h were recorded. For the risk factor analysis, antibiotic exposure was analysed by classes and included penicillins, glycopeptides, cephalosporins, aminoglycosides, quinolones and carbapenems. β-Lactams included penicillins, cephalosporins and carbapenems. Microbiology records were reviewed to identify patients harbouring methicillin-resistant Staphylococcus aureus (MRSA) and vancomycin-resistant enterococci within 6 months of enrolment.
Bacterial strains and susceptibility assays. Identification was determined using standard methods.3 Antimicrobial susceptibility was evaluated with the VITEK 2 system (bioMérieux, Inc., Hazelwood, MO, USA), the PhoenixTM system (Becton–Dickinson Microbiology Systems, Sparks, MD, USA) and the Epsilometer tests (Etest, AB Biodisk, Solna, Sweden). Results were interpreted according to the CLSI criteria.25
Molecular typing. The clonal relationship among MDR-Abc isolates was determined by a repetitive extragenic palindromic (REP)-PCR, as described elsewhere.26 Isolates were considered to be clonally related when band patterns differed by less than three bands.
Quantitative variables were tested for distribution and compared using the Kruskal–Wallis test. Differences in group proportions were assessed using the 
2 test and Fisher's exact test. Potential risk factors for MDR-Abc colonization and infection were analysed by univariate analysis. Only those variables with an association of P < 0.05 were included in the multivariate analysis by conditional (fixed effects) regression model. Goodness-of-fit of the model was calculated by the McFadden R2. LRs were calculated. LRs are alternative statistics for summarizing diagnostic accuracy and are usually more clinically useful than sensitivity and specificity.27 An LR >5 is considered to provide a significant association between the presence of the variable and the disease.27 Positive and negative predictive values and positive LR were calculated using established formulae.27,28 Statistical analyses were performed using the software program Intercooled Stata (Stata Statistical Software, release 8.0, College Station, TX, USA).
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionFundingTransparency declarationsReferences |
|---|
Patient characteristics
During the study period, 276 patients became colonized or infected with MDR-Abc. Nineteen patients were excluded from the analysis since it was not possible to define the state of infection or colonization. Among the remaining 257 cases, 120 patients were colonized and 137 infected. Cases were matched to 257 controls. Demographic and epidemiological characteristics of the 514 patients included are described in Table 1. Case patients with infections were almost equally distributed between ICU (55%) and non-ICU hospital services (23% medical and 22% surgical wards). Colonization with MDR-Abc was detected equally in ICU and medical wards (38% each).
|
The mean age (±SD) of the population was 63 ± 15 years, and the majority of the patients (56%) were male. One-hundred and nine case patients (42%) had been hospitalized within the previous 6 months. The mean length of hospitalization (±SD) prior to the recovery of MDR-Abc was 23 ± 21 days (range, 0–92) in colonized patients and 27 ± 25 days (range, 0–150) in infected patients.
Twenty-three cases (14 colonized and 9 infected) were identified within 48 h of hospital admission. Sixty-five percent of these patients had been exposed to antibiotic therapy in the previous 30 days, 61% had been hospitalized in the previous 6 months, 52% were bedridden and 35% had been transferred from other hospitals. All the 23 patients diagnosed upon hospital admission demonstrated at least one risk factor for healthcare-associated infections, according to the 2007 CDC definitions.29
In colonized patients, MDR-Abc was recovered from the following culture sites: skin/mucosa (68 cases), sputum (49) and urine (3). In infected patients, MDR-Abc was isolated from the following culture sites: blood and/or CVC (48 cases), sputum/bronchoalveolar (45), wound (35), liquor (5), urine (3) and conjunctiva (1). The distribution of sources of infections according to the isolation ward is detailed in Table 2.
|
One hundred and ninety-nine (77%) patients had carbapenem-resistant MDR-Abc. One hundred and twenty-five (49%) had strains that were susceptible to colistin only. Nine strains (3%) retained sensitivity to colistin and ampicillin–sulbactam only. There were no statistically significant differences in susceptibility patterns between colonizing and infecting strains.
Risk factors for MDR-Abc colonization
Univariate analysis. Patients with MDR-Abc colonization were more frequently bedridden with CVC and had severe underlying conditions at the time of isolation. Cases were also more likely to have had surgical procedures, ICU admission, exposure to antibiotics and isolation of MRSA compared with controls (Table 1). Among individual classes of antibiotics, penicillins, cephalosporins, quinolones, glycopeptides, aminoglycosides and carbapenems exposure was significantly associated with MDR-Abc colonization (Table 3).
|
Multivariate analysis. Conditional regression analysis identified a bedridden status, a Charlson score >3, previous MRSA isolation, ICU admission and β-lactam use as independent risk factors for colonization (Table 4). The McFadden R2 was 0.70. The clinical and epidemiological variables included in the model could not explain every episode of MDR-Abc, and there was an estimated rate of MDR-Abc colonization of 2.5% (3/120) in the absence of these characteristics. The five risk factors for MDR-Abc colonization were modelled. Overall, patients who had 0, 1, 2 or >2 risk factors had 8%, 26%, 39% and 91% probabilities of being colonized with MDR-Abc, respectively (Table 5).
|
|
Likelihood ratios. The presence of more than two risk factors was associated with a positive LR of 4.9 (Table 5). In order to help to adapt the results of this study to clinical practice, we distinguished patients according to the history of ICU admission within 30 days, since this variable might be easily obtained from the patient's chart or history. Table 6 shows the pre-test probability of MDR-Abc, positive- and negative-predicted values, and positive LRs, according to the number of risk factors and the history of ICU admission. The presence of more than two risk factors was associated with significant LRs in both groups (>4). The strongest association between risk factors and MDR-Abc colonization was observed in patients with no previous ICU admission and more than two risk factors (LR = 33).
|
To define the clinical application of these results in different epidemiological settings, a post-test analysis was done for pre-test probabilities (i.e. prevalence of the disease in the absence of risk factors) of 0.1 and 0.15. The probabilities of MDR-Abc colonization in these hospitals would be 5% and 47%, respectively, in patients with more than two risk factors.
Risk factors for MDR-A. baumannii infection
Univariate analysis. Patients with MDR-Abc infections were more likely to have severe underlying conditions, to be bedridden and organ recipients compared with controls. Cases also had surgical procedures, CVC, admission to ICU and isolation of MRSA prior to being included in the study (Table 1). Sixteen infected patients (10%) had a history of colonization by MDR-Abc within 6 months of the beginning of the study compared with only one patient among the controls (P = 0.001). Cases also had greater antibiotic exposure within 30 days of inclusion in the study (Table 3). Several classes of antibiotics were significantly associated with MDR-Abc infections, including penicillins, cephalosporins, glycopeptides, aminoglycosides and carbapenems. All patients with previous colonization with MDR-Abc had at least one other risk factor for infections.
Multivariate analysis. After adjusting for previous colonization with MDR-Abc, conditional regression analysis identified the presence of CVC, a Charlson score >3, MRSA isolation, β-lactam use and surgery as independent risk factors for MDR-Abc infections (Table 4). The McFadden R2 was 0.65. The clinical and epidemiological variables included in the model could explain all episodes of MDR-Abc except one. The estimated rate of MDR-Abc infection was 1% in the absence of these characteristics.
We modelled the five risk factors for MDR-Abc infection. Overall, patients who had 0, 1, 2 or >2 risk factors had 2%, 20%, 54% and 84% probabilities of being infected with MDR-Abc, respectively (Table 7).
|
Likelihood ratios. The presence of more than two risk factors was associated with a positive LR of 7.
We then analysed patients without a history of MDR-Abc colonization. Table 8 shows the pre-test probability, positive- and negative-predicted values and positive LRs according to the number of risk factors. The presence of more than two risk factors was associated with an LR of 5.3.
|
Microbiological analysis
In order to assess the epidemiological link among the isolates, all strains were analysed by REP-PCR.26 All MDR-Abc carbapenem-resistant isolates were included in a unique profile. The remaining MDR carbapenem-susceptible isolates showed other non-correlated profiles.
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionFundingTransparency declarationsReferences |
|---|
In the past few decades, Abc has emerged as an important nosocomial pathogen responsible for significant morbidity and mortality. The risk factors associated with MDR-Abc infections or colonization have raised controversy. In recent publications, methodological issues specific to risk factors were raised.30–32 The majority of previous studies selected controls among patients with antibiotic-susceptible Abc and failed to distinguish between infection and colonization.
We evaluated the accuracy of several variables in hospitalized patients in diagnosing MDR-Abc colonization or infection. Using a multivariate model, we developed a five variable model. A high Charlson score, β-lactam usage and MRSA isolation within 30 days were independently associated with infection and colonization. Bedridden status and previous ICU admission predicted colonization, while the presence of CVC and previous surgery predicted infection.
The association between antibiotic usage and antibiotic-resistant bacteria is still under debate.33,34 A multicentre prospective cohort study documented that the highest risk of being colonized with antibiotic-resistant bacteria is associated with the usage of carbapenems.35 Receipt of a quinolone was an independent risk factor for endemic MDR-Abc infection. The selection pressure caused by the indiscriminate use of quinolones was responsible for the persistence of MDR clones over at least 5 years.36 The use of third-generation cephalosporins, carbapenems and broad-spectrum penicillins was associated with an increased risk of nosocomial pneumonia due to MDR-Abc.5,13,19 In our study, receipt of β-lactam antibiotics was a significant variable for both colonization and infection. Prolonged antibiotic treatment with these antibiotics may encourage the selection and outgrowth of Abc strains after colonization of patients. The exposure to β-lactams might provide a selective advantage for MDR-colonizing strains competing with susceptible clones.
Previous admission into an ICU with a high density of patients infected with MDR-Abc and being bedridden with severe underlying conditions were associated with both colonization and infection, suggesting the important role of cross-transmission in the acquisition of MDR-Abc.
In our study, intravascular catheter and recent surgery were the risk factors for MDR-Abc infections. The presence of CVC might reflect the severity of underlying disease or may well be a surrogate for healthcare workers' contacts. Previous studies have reported that Abc adheres well to epithelial cells.37,38 It is possible that the adherence properties of Abc favour colonization of the skin followed by colonization of wounds during catheterization and surgery.
The analysis of LRs suggests that the presence of at least three risk factors among those selected by the conditional regression analysis has the highest value to rule a diagnosis of MDR-Abc colonization or infection. In clinical practice, the LR can be multiplied to generate an LR for an individual patient. For example, in our hospital, a bedridden patient with no history of ICU admission and previous MRSA and β-lactam use has an LR of 33 to rule a diagnosis of MDR-Abc colonization. Thus, in a different epidemiological setting with a pre-test probability of only 0.1, this group of variables increases the post-test probability of being colonized to 0.8. To simplify, the calculation nomogram is available.39
Our results might have important clinical implications in differentiating risk factors between colonized and infected patients. First, our data suggest that colonization with MDR-Abc in non-ICU patients is related to a bedridden status, previous MRSA colonization, therapy with β-lactams and severe co-morbidities. A further study might use these variables to perform a cost-effectiveness analysis of targeted screening on the population admitted to medical or surgical wards in hospital with epidemic/endemic MDR-Abc. Second, in patients not previously colonized with MDR-Abc, severe co-morbidities, recent β-lactam usage, MRSA isolation, surgery and presence of CVC were associated with an LR of 5. This might be important in the selection of empirical therapy for infections in hospitals with epidemic/endemic MDR-Abc. If we consider a pre-test probability of 0.9, the presence of these four risk factors, in patients with no history of colonization, would increase the risk of MDR-Abc infections up to 0.98.
The results of the molecular analysis suggest that the dissemination of MDR-Abc complex isolates has occurred through the spread of a unique clone that included all MDR-carbapenem-resistant isolates and different unrelated strains of MDR carbapenem-susceptible isolates. An endemic–epidemic distribution of MDR-carbapenem-resistant isolates was observed.
There are some limits to our analysis. First, the study is a case–control study. It is possible that significant confounders such as hospitalization nearby a colonized patient were not addressed. The history of colonization with MRSA, for example, might be a surrogate of multiple factors such as previous antibiotics usage and hospitalization. Second, the study was performed in a single centre and the predictive model should be tested in a population different from that used to develop it. This limitation, however, is overcome by the calculation of the LR that allows physicians to translate the results to other epidemiological settings. Another limitation of this study might be that not all colonized patients may have been included because active surveillance was not performed. Colonized patients without clinical manifestations could have been erroneously considered as controls. However, this might have led to the underestimation of some risk factors. We cannot rule out cross-transmission among patients through the hands of healthcare workers. We could hypothesize that cross-transmission might have occurred, despite the implementation of stringent infection-control measures that had been put into place in our hospital after the first cases of MDR-Abc.
Our study adds substantially to the literature on the MDR-Abc epidemiology. Results might be used to plan targeted screening in non-ICU populations at risk of being colonized in epidemic–endemic settings. Moreover, since β-lactam therapy is a risk factor for the acquisition of MDR-Abc, as once again demonstrated by our study, physicians should sharpen their clinical judgement when choosing antibiotic therapy and strongly reinforce compliance with basic control measures for nosocomial infections. Our results could have implications for the design of future studies on control programmes for the spread of multiresistant Acinetobacter in hospital.
|
Funding |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionFundingTransparency declarationsReferences |
|---|
No funding has been received.
|
Transparency declarations |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionFundingTransparency declarationsReferences |
|---|
None to declare.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionFundingTransparency declarationsReferences |
|---|
1 . Gales AC, Jones RN, Forward KR, et al. Emerging importance of multidrug-resistant Acinetobacter species and Stenotrophomonas maltophilia as pathogens in seriously ill patients: geographic patterns, epidemiological features and trends in the SENTRY Antimicrobial Surveillance Program (1997–99). Clin Infect Dis (2001) 32(Suppl_2):104–13.[CrossRef]
2 . Allen DM, Hartman BJ. Acinetobacter species. In: Principle and Practice of Infectious Diseases—Mandell GL, Bennett JE, Dolin R, eds. (2005) 6th edn. Churchill, Livingstone: Elsevier. 2632–6.
3 . Bergogne-Berezin E, Towner KJ. Acinetobacter spp. as nosocomial pathogens: microbiological, clinical, and epidemiological features. Clin Microbiol Rev (1996) 9:148–65.[Web of Science][Medline]
4 . Go ES, Urban C, Burns J, et al. Clinical and molecular epidemiology of Acinetobacter infections sensitive only to polymyxin B and sulbactam. Lancet North Am Ed (1994) 344:1329–32.
5
.
Corbella X, Montenero A, Pujol M, et al. Emergence and rapid spread of carbapenem resistance during a large and sustained hospital outbreak of multiresistant Acinetobacter baumannii. J Clin Microbiol (2000) 38:4086–95.
6 . Levin AS, Barone AA, Penco J, et al. Intravenous colistin as therapy for nosocomial infections caused by multidrug-resistant Pseudomonas aeruginosa and Acinetobacter baumannii. Clin Infect Dis (1999) 28:1008–11.[Web of Science][Medline]
7 . Smolyakov R, Borer A, Riesenberg K, et al. Nosocomial multi-drug resistant Acinetobacter baumannii bloodstream infection: risk factors and outcome with ampicillin-sulbactam treatment. J Hosp Infect (2003) 54:32–8.[CrossRef][Web of Science][Medline]
8 . Falagas ME, Bliziotis IA, Siempos II. Attributable mortality of Acinetobacter baumannii infections in critically ill patients: a systematic review of matched cohort and case–control studies. Crit Care (2006) 10:R48.[CrossRef][Medline]
9 . Urban C, Segal-Maurer S, Rahal JJ. Considerations in control and treatment of nosocomial infections due to multidrug-resistant Acinetobacter baumannii. Clin Infect Dis (2003) 36:1268–74.[CrossRef][Web of Science][Medline]
10 . Corbella X, Pujol M, Ayats J, et al. Relevance of digestive tract colonisation in the epidemiology of nosocomial infections due to multiresistant Acinetobacter baumannii. Clin Infect Dis (1996) 23:329–34.[Web of Science][Medline]
11
.
Marchaim D, Navon-Venezia S, Schwartz D, et al. Surveillance cultures and duration of carriage of multidrug-resistant Acinetobacter baumannii. J Clin Microbiol (2007) 45:1551–5.
12 . Mahgoub S, Ahmed J, Glatt AE. Underlying characteristics of patients harboring highly resistant Acinetobacter baumannii. Am J Infect Control (2002) 30:386–90.[CrossRef][Web of Science][Medline]
13 . Abbo A, Navon-Venezia S, Hammer-Muntz O, et al. Multidrug-resistant Acinetobacter baumannii. Emerg Infect Dis (2005) 11:22–9.[Web of Science][Medline]
14 . Wong TH, Tan BH, Ling ML, et al. Multi-resistant Acinetobacter baumannii on a burns unit-clinical risk factors and prognosis. Burns (2002) 28:349–57.[CrossRef][Web of Science][Medline]
15 . Kaul R, Burt JA, Cork L, et al. Investigation of a multiyear multiple critical care unit outbreak due to relatively drug-sensitive Acinetobacter baumannii: risk factors and attributable mortality. J Infect Dis (1996) 174:1279–87.[Web of Science][Medline]
16 . Herruzo R, de la Cruz J, Fernandez-Acenero MJ, et al. Two consecutive outbreaks of Acinetobacter baumannii 1-a in a burn intensive care unit for adults. Burns (2004) 30:419–23.[CrossRef][Web of Science][Medline]
17 . Husni RN, Goldstein LS, Arroliga AC, et al. Risk factors for an outbreak of multi-drug-resistant Acinetobacter nosocomial pneumonia among intubated patients. Chest (1999) 115:1378–82.[CrossRef][Web of Science][Medline]
18 . Manikal VM, Landman D, Saurina G, et al. Endemic carbapenem-resistant Acinetobacter baumannii species in Brooklyn, New York: citywide prevalence, interinstitutional spread, and relation to antibiotic usage. Clin Infect Dis (2000) 31:101–6.[CrossRef][Web of Science][Medline]
19 . Mulin B. Risk factors for nosocomial colonisation with multiresistant Acinetobacter baumannii. Eur J Clin Microbiol Infect Dis (1995) 14:569–76.[CrossRef][Web of Science][Medline]
20 . Falagas ME, Kopterides P. Risk factors for the isolation of multi-drug-resistant Acinetobacter baumannii and Pseudomonas aeruginosa: a systematic review of the literature. J Hosp Infect (2006) 64:7–15.[CrossRef][Web of Science][Medline]
21 . Harris AD, McGregor JC, Furuno JP. What infection control interventions should be undertaken to control multidrug-resistant Gram-negative bacteria? Clin Infect Dis (2006) 43(Suppl 2):S57–61.[CrossRef][Web of Science][Medline]
22 . Garner JS, Jarvis WR, Emori TG, et al. CDC definitions for nosocomial infections, 1988. Am J Infect Control (1988) 16:128–40.[CrossRef][Web of Science][Medline]
23 . Charlson ME, Pompei P, Ales KL, et al. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis (1987) 40:373–83.[CrossRef][Web of Science][Medline]
24 . Knaus WA, Wagner DP, Draper EA, et al. The APACHE III Prognostic System. Risk prediction of hospital mortality for critically ill hospitalised adults. Chest (1991) 100:1619–36.[CrossRef][Web of Science][Medline]
25 . Clinical and Laboratory Standards Institute. Performance Standards for Antimicrobial Susceptibility Testing: Seventeenth Informational Supplement M100-S17 (2007) Wayne, PA, USA: CLSI.
26 . Snelling AM, Gerner-Smidt P, Hawkey P, et al. Validation of use of whole-cell repetitive extragenic palindromic sequence-based PCR (REP-PCR) for typing strains belonging to the Acinetobacter calcoaceticus–Acinetobacter baumannii complex and application of the method to the investigation of a hospital outbreak. J Clin Microbiol (1996) 34:1193–202.[Abstract]
27
.
Deeks JJ, Altman DG. Diagnostic tests 4: likelihood ratios. BMJ (2004) 329:168–9.
28
.
Altman DG, Bland JM. Diagnostic tests. 1: Sensitivity and specificity. BMJ (1994) 308:1552.
29 . Siegel JD, Rhinehart E, Jackson M, et al. Guidelines for isolation precautions: preventing transmission of infectious agents in healthcare settings. Am J Infect Control (2007) 35(Suppl 2):S65–164.[CrossRef][Web of Science][Medline]
30 . Harris AD, Samore MH, Lipsitch M, et al. Control-group selection importance in studies of antimicrobial resistance: examples applied to Pseudomonas aeruginosa, Enterococci, and Escherichia coli. Clin Infect Dis (2002) 34:1558–63.[CrossRef][Web of Science][Medline]
31 . Harris AD, Karchmer TB, Carmeli Y, et al. Methodological principles of case–control studies that analyzed risk factors for antibiotic resistance: a systematic review. Clin Infect Dis (2001) 32:1055–61.[CrossRef][Web of Science][Medline]
32 . Paterson DL. Looking for risk factors for the acquisition of antibiotic resistance: a 21st-century approach. Clin Infect Dis (2002) 34:1564–7.[CrossRef][Web of Science][Medline]
33
.
Tacconelli E, De Angelis G, Cataldo MA, et al. Does antibiotic exposure increase the risk of methicillin-resistant Staphylococcus aureus (MRSA) isolation? A systematic review and meta-analysis. J Antimicrob Chemother (2008) 61:26–38.
34 . Hyle EP, Lipworth AD, Zaoutis TE, et al. Risk factors for increasing multidrug resistance among extended-spectrum β-lactamase-producing Escherichia coli and Klebsiella species. Clin Infect Dis (2005) 40:1317–24.[CrossRef][Web of Science][Medline]
35 . Tacconelli E, De Angelis G, Cataldo A, et al. The link between use of antibiotics and acquisition of methicillin-resistant Staphylococcus aureus (MRSA), vancomycin-resistant enterococci (VRE), and ciprofloxacin-resistant Pseudomonas aeruginosa (CR-PA) in a cohort of 864 patients. In: Abstracts of the Forty-seventh Interscience Conference on Antimicrobial Agents and Chemotherapy, Chicago, IL, 2007. Washington, DC, USA: American Society for Microbiology. Abstract C2-2037.
36
.
Villers D, Espaze E, Coste-Burel M, et al. Nosocomial Acinetobacter baumannii infections: microbiological and clinical epidemiology. Ann Intern Med (1998) 129:182–9.
37 . Koeleman JG, Parlevliet GA, Dijkshoorn L, et al. Nosocomial outbreak of multi-resistant Acinetobacter baumannii on a surgical ward: epidemiology and risk factors for acquisition. J Hosp Infect (1997) 37:113–23.[CrossRef][Web of Science][Medline]
38
.
Rosenberg M, Perry A, Bayer EA, et al. Adherence of Acinetobacter calcoaceticus RAG-1 to human epithelial cells and to hexadecane. Infect Immun (1981) 33:29–33.
39 . Fagan TJ. Nomogram for Bayes theory. N Engl J Med (1975) 293:257.[Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||