ORIGINAL ARTICLE ABSTRACT

Transcription

1 VOLUME 81 No. 9 SEPTEMBER 2015

2 ORIGINAL ARTICLE Changed epidemiology of ICU acquired bloodstream infections over 12 years in an Italian teaching hospital G. B. ORSI 1, S. GIULIANO 1, C. FRANCHI 2, V. CIORBA 3, C. PROTANO 1, A. GIORDANO 1, M. ROCCO 4, M. VENDITTI 1 1Department of Public Health and Infectious Diseases, Sapienza University of Rome, Rome, Italy; 2 Department of Clinical Medicine, Sapienza University of Rome, Rome, Italy; 3 Department of Biomedical, Biotechnological and Translational Sciences (SBiBiT), University of Parma, Parma, Italy; 4 Anesthesiology and Intensive Care, Sapienza University of Rome, Rome, Italy ABSTRACT Background. We compared the etiology of 203 ICU-acquired laboratory confirmed bloodstream infections (LC- BSI) prospectively collected between January 2000-December 2007 (first period) with 83 LC-BSI recorded between January 2010-December 2012 (second period), after the diffusion in 2008 of K. pneumoniae expressing carbapenemresistance due to K. pneumoniae carbapenemases production (KPC-CR-Kp). Methods. In the general ICU of teaching hospital Umberto I in Rome, all ICU-acquired LC-BSI episodes occurring in patients admitted to ICU 48h were included. Baseline characteristics, clinical features, antimicrobial resistance and outcome were recorded. All isolated strains multidrug resistance (MDR) were evaluated according to the European Centre for Disease Control (ECDC) guidelines. Results. Overall the study included 329 isolates, 214 in and 115 in In the second period we registered a Gram-positive reduction (55.1% vs. 26.9%; P<0.01) and Gram-negative increase (40.2% vs. 69.6%; P<0.01). In staphylococci were responsible for 45.8% LC-BSI s, whereas 18.3% during Enterobacteriaceae increased dramatically (15.4% vs. 39.2%; P<0.01), especially Klebsiella spp. (5.6% vs. 31.3%; P<0.01). LC-BSI associated mortality decreased among Gram-positive (56.8% vs. 51.6%), but increased in Gramnegative (41.9% vs. 60.0%; P<0.03), especially in Enterobacteriaceae (RR 2.13; 95% CI ; P<0.01). MDR increased remarkably among Enterobacetriaceae (51.5% vs. 73.3%). The study highlighted the associated mortality for Enterobacteriaceae when comparing MDR to non-mdr microorganisms. Conclusion. ICU-acquired LC-BSI etiology shifted from Gram-positive to Gram-negative during the study period in our ICU. Also associated mortality decreased among the former, whereas it increased in the latter. Last MDR increased enormously among Enterobacteriaceae with the diffusion of KPC (75% of strains), adding significantly to associated mortality (RR 2.17; ; P<0.01). (Minerva Anestesiol 2015;81:980-8) Key words: Infection, blood - Etiology - Epidemiology - Intensive care units. ICU-acquired laboratory confirmed bloodstream infection (LC-BSI) is an important cause of morbidity and mortality in intensive care units (ICU), where it affects more than 10% of patients, determining high associated mortality >20% 1-3 and adding costs. 4 Appropriate empiric therapy is considered to Comment on p be the most important factor for patients outcome, and is associated with reduced mortality and length of stay. 3, 5, 6 For this reason it is important to know the pathogens causing infection and their antimicrobial resistance pattern to guide appropriate antimicrobial treatment. Unfortunately the epidemiology and antimicrobial profile of microorganisms responsible for bloodstream infection varies between insti- 980 MINERVA ANESTESIOLOGICA September 2015

3 ICU ACQUIRED BLOODSTREAM INFECTIONS ORSI tutions, and among ICUs within the hospitals. Also, as the rates of antimicrobial resistance in pathogens principally among Gram-negatives are increasing, every center should be familiar with its local trends in order to target a more appropriate empirical therapy. In our hospital during the years 2008 and 2009 we documented the appearance of a first Klebsiella pneumoniae clone (ST37) expressing ertapenem resistance by modification of the outer membrane permeability (Porin-ER-Kp). 7 Subsequently in 2010, we observed the disappearance of Porin-ER Kp strains and the appearance and rapid spread of a new clone of Klebsiella pneumoniae expressing carbapenem resistance due to K. pneumoniae carbapenemases production (KPC-CR-Kp), especially in high risk areas such as ICUs. 8 As the two carbapenem resistant K. pneumoniae clones were different, we decided to compare the earlier period before the diffusion of carbapenem resistance in Enterobacteriaceae ( ), to our current epidemiological situation ( ), excluding the intermediate period Therefore to understand the changing etiology and antimicrobial profile of microorganisms responsible for ICU-acquired LC-BSI in our ICU, we carried out an epidemiological study in order to compare the ICU-acquired LC-BSI etiology of the two periods (precarbapenemases period) versus (carbapenemases period). Setting Materials and methods We conducted a prospective observational study at the 13 bed general ICU of the bed University hospital Policlinico Umberto I of Rome. From January 2010 to December 2012 all episodes of ICU-acquired LC-BSI occurring in patients admitted to the ward 48h were included. Information on baseline characteristics, clinical features, antimicrobial resistance and outcome were recorded. These episodes ( ) were compared with those prospectively collected from January 2000 to December 2007 at the same institution ( ). Data regarding the first period has been previously described. 9 Definitions For the purpose of the study, only ICU-acquired LC-BSI were taken into account, diagnosed in patients at least 48 hours after ICU admission. ICU-acquired LC-BSI was defined as the isolation of one or more microorganisms from a blood culture in a patient with two or more of the following: temperature >38 C or <36 C, heart rate (HR) >90 beats/min, respiratory rate >20 breaths/min, WBC>12,000/ mm 3 or >10% immature neutrophils. Coagulase negative staphylococcus LC-BSI was defined as 2 blood cultures demonstrating the same phenotype on separate occasions within a 48 hours period. 10 When blood cultures were collected, duplicate isolates were excluded from the analysis. Source of ICU-acquired LC-BSI was determined on the basis of the isolation of the microorganisms from the presumed portal of entry and clinical evaluation. When no link with a primary source could be found, LC-BSI was considered primary in origin. Data collection During the 12 years survey an infection control team (ICT), composed by one physician specialized in intensive care, two in infectious diseases and one epidemiologist, actively participated to the surveillance. Data were collected prospectively by two physicians especially trained, using a specific database oriented software (Epi-info Version 2011, CDC). The following information was recorded: demographic characteristics (i.e. sex, age, etc.), date of admission and discharge, patient origin (i.e. emergency, operating rooms wards, other ICU), admission diagnosis (principal diagnosis leading to ICU admission), severity score (SAPS II), underlying diseases presence (diabetes mellitus, chronic renal failure, cirrhosis, chronic obstructive pulmonary disease) and final ICU outcome. Invasive procedures are associated to BSI s in ICU patients, therefore surveillance also Vol No. 9 MINERVA ANESTESIOLOGICA 981

4 ORSI ICU ACQUIRED BLOODSTREAM INFECTIONS included central venous catheter (CVC), mechanical ventilation and urinary tract catheter exposure and duration All ICU-acquired LC-BSI isolated microorganisms and their antibiotic susceptibility were screened and recorded. Microbiological methods All patient samples were taken for culture according to the general principles of specimen collection and transport. 15 The species identification and the antimicrobial susceptibility testing were performed on the isolated strains by using the VITEK system and performed a phenotypic confirmation test of carbapenemase production in Enterobacteriaceae isolates having a minimum inhibitory concentration >0.5 mcg/ ml for meropenem and/or imipenem. Carbapenemase production was confirmed by first with Hodge-Test modified and after with disc diffusion synergy test including meropenem and two carbapenemase inhibiting compounds (dipicolinic acid and boronic acid). 16 From 2000 to 2011 antimicrobial susceptibility was referred to the most recent CLSI breakpoints. Since 2012 Breakpoints were interpreted in accordance with the European Committee on Antimicrobial Susceptibility Testing (EUCAST) guidelines. 17, 18 Among microorganisms the multi-drug resistance (MDR) was defined according to the ECDC guidelines. 18 We defined early mortality as death occurring 7 days after ICU-acquired LC-BSI. 20, 21 Data Analysis Statistical analyses were carried out using SPSS software (version 14.0 for Windows, Chicago, IL, USA). The χ 2 test was used to examine differences between groups. Statistical significance was defined as a P value of less then When the observed frequencies were >5 χ2 with Yates correction was used to compare the proportions observed in the two groups. When observed frequencies were less or equal to 5 Fisher s exact test was used. The normality of quantitative data (age, SAPS II score, length of stay) was assessed by non-parametric technique (Kolmogorov-Smirnov test). When data resulted normally distributed, t-test for independent samples was used, otherwise Mann-Whitney test was applied. The univariate relationship between infection and death was tested using relative risk and its 95% confidence interval (CI 95 ). Results Between and respectively 1741 and 1165 patients, with length of stay >48 hours, were surveyed and evaluated for ICU-acquired LC-BSI in the ICU. During the first period ( ) 167 (9.6%) patients developed 203 (11.7%) ICU-acquired LC-BSI, whereas in the second period ( ) 83 (7.1%) patients developed 101 (8.7%) ICUacquired LC-BSI. Overall the study included 329 isolates. In total, 214 microorganisms isolated from 203 LC-BSI in 167 patients during the first period ( ) were compared with 115 microorganisms isolated from 98 LC-BSI in 83 patients during the second period ( ). As reported in Table I, the two patient clusters showed similar demographic and clinical characteristics, also for the SAPS II score which was 46.5±15.7 vs. 43.0±12.0; P= Table II shows the causative microorganisms of all ICU-acquired LC-BSI compared by study periods. Between the first period ( ) before KPC appearance and the second ( ) with the diffusion of KPC s in the enterobacteriaceae we registered significant Gram-positive reduction (55.1% vs. 26.9%; P<0.01), Gram-negative increase (40.2% vs. 69.6%; P<0.01) and a minimum variation of fungi (4.7% vs. 3.5%). In particular, during the first period ( ) staphylococci were responsible for 45.8% of all LC-BSI s, whereas only 18.3% during On the contrary enterobacteriaceae increased dramatically (15.4% vs. 39.2%; P<0.01), especially Klebsiella spp. (5.6% vs. 31.3%; P<0.01) and also Acinetobacter baumannii (7.5% vs. 20.0%) became more common. We also considered microorganisms according to ICU-acquired LC-BSI source, and found that primary and secondary strains were respectively 982 MINERVA ANESTESIOLOGICA September 2015

5 ICU ACQUIRED BLOODSTREAM INFECTIONS ORSI Table I. Characteristic of patients with ICU-acquired laboratory confirmed bloodstream infection in the two periods. Characteristics Period (167 patients) Period (83 patients) P Age (SD) 52.3± ± Male 108 (64.8%) 53 (63.8%) Female 59 (35.2%) 30 (36.2%) SAPS II (SD) 46.5± ± Length of stay (days) 36.2± ± ICU Crude mortality 606 (34.8%) 383 (32.9%) Admission diagnosis Medical 38 (22.7%) 18 (21.7%) Surgical 79 (47.3%) 33 (39.8%) Traumatic 50 (29.9%) 32 (38.5%) Underlying conditions Neurologic disorders 22 (13.2%) 4 (4.8%) Cardiovascular disorder 41 (24.5%) 11 (13.2%) Hypertension 24 (14.4%) 25 (30.1%) COPD 33 (19.8%) 11 (13.2%) Renal disorders 7 (4.2%) 6 (7.2%) Liver cirrhosis 5 (3.0%) 1 (1.2%) *0.67*4 Diabetes mellitus 16 (9.6%) 11 (13.2%) Invasive procedures Central venous catheter 159 (95.2%) 83 (100.0%) Mechanical ventilation 164 (98.2%) 83 (100.0%) Urinary catheter 167 (100.0%) 83 (100.0%) - P values are obtained using the χ 2 corrected test except when indicated by (*) which are obtained from the Fisher exact test. COPD: Chronic obstructive pulmonary disease. Table II. Distribution of microorganisms responsible for ICU-acquired laboratory confirmed bloodstream infection. Microorganism Period Period Total Isolates % Isolates % Isolates % MRSA 32 (14.9%) 2 (1.7%) 34 (10.3%) <0.0001* MSSA 5 (2.3%) 2 (1.7%) 7 (2.1%) 1.00* CNS-MR 56 (26.3%) 17 (14.9%) 73 (22.2%) 0.02 CNS-MS 5 (2.3%) (1.5%) - Enterococcus spp. 2 (0.9%) 3 (2.6%) 5 (1.5%) 0.35* E. faecalis 13 (6.1%) 3 (2.6%) 16 (4.9%) 0.19* E. faecium 5 (2.3%) 4 (3.5%) 9 (2.7%) 0.72* Total Gram+ 118 (55.1%) 31 (26.9%) 149 (45.3%) < A. baumannii 16 (7.5%) 23 (20.0%) 39 (11.8%) <0.001 Aeromonas spp. 1 (0.5%) - 1 (0.3%) - Citrobacter freundi - 1 (0.9%) 1 (0.3%) - Enterobacter spp. 5 (2.3%) 3 (2.6%) 8 (2.4%) 1.00* E. coli 5 (2.3%) 1 (0.9%) 6 (1.8%) 0.70* Klebsiella spp. 12 (5.6%) 36 (31.3%) 48 (14.6%) < (KPC-CR-Kp) # 31 # (27.0%) # 31 # (9.4%) # - Morganella spp. 1 (0.5%) 1 (0.9%) 2 (0.6%) 1.00* Proteus spp. 2 (0.9%) 3 (2.6%) 5 (1.5%) 0.35* Providencia spp. 2 (0.9%) (0.6%) - P. aeruginosa 29 (13.6%) 12 (10.4%) 41 (12.5%) 0.41 S. maltophilia 7 (3.3%) (2.1%) - S. marcescens 6 (2.8%) (1.8%) - All Enterobacteriaceae 33 (15.4%) 45 (39.1%) 78 (23.7%) < Total Gram- 86 (40.2%) 80 (69.6%) 166 (50.5%) < Candida spp. 10 (4.7%) 4 (3.5%) 14 (4.2%) 0.78* TOTAL ,0% ,0% ,0% MRSA: Methicillin resistant S. aureus; MSSA: Methicillin susceptible S. aureus; CNS-MR: methicillin resistant coagulase negative staphylococci; CNS-MS: Methicillin susceptible coagulase negative staphylococci; Enterobacteriaceae; # KPC-CR-Kp strains are a subgroup of the overall Klebsiella. P values are obtained using the χ 2 corrected test except when indicated by (*) which are obtained from the Fisher exact test. P Vol No. 9 MINERVA ANESTESIOLOGICA 983

6 ORSI ICU ACQUIRED BLOODSTREAM INFECTIONS 47.1% and 52.9%. The latter 31.3% (CVC), 14.3% (respiratory tract), 4.6% (surgical site) and 2.7% (UTI). Results showed that onset time (days) between ICU admission and ICU-acquired LC- BSI was higher (P<0.01) among Gram-negative (21.1±18.0) compared to Gram-positive (15.5±16.0), whereas for fungi were 18.8±22.7. The distribution of microorganisms associated mortality rate in the two periods is illustrated in Table III. Overall we observed a higher LC- BSI associated mortality in the second period (51.4% vs. 59.1%), with a modest reduction among Gram-positive (56.8% vs. 51.6%) but significant increase in Gram-negative (41.9% vs. 60.0%; P<0.03). However during the second period among Gram-negative LC-BSI associated mortality was lower in non-fermentative microorganisms as A. baumanni (75.0% vs. 65.2%) and P. aeruginosa (48.3% vs. 33.3%), whereas it increased over two-fold in enterobacteriaceae (RR 2.13; 95% CI ; P<0.01). Antimicrobial susceptibility testing showed that staphylococci expressed along time a high resistance to methicillin (>80%). In enterobacteriaceae, from the first to the second period we observed a significant increase (P<0.01) in antimicrobial resistance to a wide range of antibiotics: imipenem (6.7% vs. 68.3%), meropenem (10.0% vs. 73.2%), amikacin (10.0% vs. 65.9%), ceftazidime (37.9% vs. 82.5%), cefepime (26.7% vs. 75.6%), cefotaxime (10.0% vs. 80.5%), cyprofloxacin (34.5% vs. 82.5%), levofloxacin (33.3% vs. 78.4%) and piperacillin/tazobactam (27.2% vs. 74.3%). Only during the second period ( ) resistance to colistin resulted 28.9%. All isolated strains multidrug resistance (MDR) were evaluated according to the ECDC guidelines. 19 MDR rate and evolution along the two periods is shown in Table IV. Table III. Distribution of microorganisms associated mortality rate ( vs ). Microorganism Early Death ( 7 days) Final associated mortality P* Period Period Period Period RR (95%CI) S. aureus 1/4 (25.0%) 11/37 (29.7%) 1.00* 2/4 (50.0%) 26/37 (70.3%) 0.71; * CNS 3/17 (17.6%) 10/61 (16.4%) 1.00* 9/17 (52.9%) 31/61 (50.8%) 1.04; Enterococci 0/10 (0.0%) 2/20 (10.0%) 0.88* 5/10 (50.0%) 10/20 (50.0%) 1.00; * All Gram+ 4/31 (12.9%) 23/118 (19.5%) /31 (51.6%) 67/118 (56.8%) 0.91; A. baumannii 6/23 (26.1%) 6/16 (37.5%) /23 (65.2%) 12/16 (75.0%) 0.87; Enterobacteriaceae 12/45 (26.7%) 6/33 (18.2%) /45 (64.4%) 10/33 (30.3%) 2.13; (KPC-CR-Kp) # /31 (77.4%) P. aeruginosa 2/12 (16.7%) 7/29 (24.1%) 0.94* 4/12 (33.3%) 14/29 (48.3%) 0.69; * All Gram- 20/80 (25.0%) 19/86 (22.1%) /80 (60.0%) 36/86 (41.9%) 1.43; Candida spp. 1/4 (25.0%) 4/10 (40.0%) 1.00* 4/4 (100.0%) 7/10 (70.0%) 1.43; * TOTAL 25/115 (21.7%) 46/214 (21.5%) /115 (59.1%) 110/214 (51.4%) 1.15; #KPC-CR-Kp strains are a subgroup of the overall Klebsiella spp. P values are obtained using the χ 2 corrected test except when indicated by (*) which are obtained from the Fisher exact test. P Table IV. Evolution of multidrug resistance in microorganisms responsible for ICU-acquired LC-BSI ( vs ) Microorganism Multidrug resistance Period Period RR (95%CI) S. aureus 2/4 (50.0%) 32/37 (86.5%) 0.58; * Enterococci 3/10 (30.0%) 2/20 (10.0%) 3.00; * A. baumannii 20/23 (87.0%) 11/16 (68.8%) 1.26; Enterobacteriaceae 33/45 (73.3%) 17/33 (51.5%) 1.42; P. aeruginosa 5/12 (41.7%) 16/29 (55.2%) 0.75; * P values are obtained using the χ 2 corrected test except when indicated by (*) which are obtained from the Fisher exact test. P 984 MINERVA ANESTESIOLOGICA September 2015

7 ICU ACQUIRED BLOODSTREAM INFECTIONS ORSI Table V. Associated mortality in MDR vs. non MDR microorganisms (data cumulative of periods and ). Microorganism MDR Associated mortality Non MDR RR (95%CI) P* S. aureus 24/34 (70.6%) 4/7 (57.1%) 1.23; * Enterococci 2/5 (40.0%) 13/25 (52.0%) 0.77; * A. baumannii 22/31 (71.0%) 5/8 (62.5%) 1.13; * Enterobacteriaceae 31/50 (62.0%) 8/28 (28.6%) 2.17; (KPC-CR-Kp) # 24/31 (77.4%) 5/17 (29.4%) 2.63; * P. aeruginosa 16/21 (57.1%) 6/20 (30.0%) 1.90; #KPC-CR-Kp strains are a subgroup of the overall Klebsiella spp. P values are obtained using the χ 2 corrected test except when indicated by (*) which are obtained from the Fisher exact test. Among Gram-positive S. aeruginosa baumannii expressed always high methicillin resistance ( 50%), although in the period staphylococci isolation was much less common. On the contrary no vancomycin resistant enterococci were isolated during study periods. During the study period there was an impressive increase of MDR in enterobacteriaceae, whereas among non-fermentative bacilli MDR increased slightly in A. baumannii, but diminished in P. aeruginosa. Table V shows the associated mortality comparing MDR to non-mdr microorganisms, highlighting the growing impact of Enterobacteriaceae (RR 2.17; ; P<0.01), and particularly K. pneumoniae carbapenemases producing strains (RR 2.63; ; P<0.01). We also considered strains antimicrobial susceptibility according to the sources of LC- BSI. Results showed that strains MDR rates were respectively 70.2% (secondary to respiratory tract), 44.4% (UTI), 46.7% (surgical site), 40.6% (unknown) and 33.0% (CVC). Discussion The main purpose of surveillance on a specific disease etiology such as LC-BSI is to detect shifts in antimicrobial susceptibility of the involved bacteria in order to help improving the choice of empiric therapy. The strength of such studies on LC-BSI relies on clear and standardized clinical diagnostic criteria which make data reliable and realistic, avoiding confounding colonizing agents not directly related to clinical disease (as it may occur when ventilator associated pneumonia are considered). The decision whether to use χ2 with Yates correction rather than Fisher s exact test to compare proportions was made by arbitrarily using a cut off observed frequency of >5. Some authorities recommend other cut-offs also based on the expected frequency. In fact, we have used both tests for all comparisons and we found that these are not important differences in the results. In order to reduce bias we decided to compare the earlier period before the diffusion of carbapenem resistance in Enterobacteriaceae ( ), to our current epidemiological situation ( ), excluding the intermediate period , because as we explained in the introduction the two carbapenem resistant K. pneumoniae clones were different. Demographically there were limited differences between the two period patient populations which appeared homogeneous by age, gender and SAPS II score (Table I). Major differences were represented by an increase, in proportion, of trauma admissions and a reduction of patients with neurologic disorders as underlying conditions in the second period. This because a separate specialistic neurosurgical ICU was enlarged and the opening of new beds in it explains the shift in percentages between the two periods. Especially in a general ICU, the case-mix is a very important parameter to evaluate and compare the clusters, therefore it was encouraging to note that in and the principal diagnosis at admission as the underlying diseases had remained mostly unchanged. 22 Main finding was a dramatic change in Vol No. 9 MINERVA ANESTESIOLOGICA 985

8 ORSI ICU ACQUIRED BLOODSTREAM INFECTIONS ICU-acquired LC-BSI etiology between the two study periods, with a marked reduction of Gram-positive microorganisms and a fierce increase in isolation of Gram-negative (Table II). Among Gram-positive we observed mainly a decrease of staphylococci, particularly MRSA. In the last decade Gram-negative pathogens have emerged as dominant contributors for bloodstream infection in various institutions of developed countries. The study results appear consistent with these trends and other reports from USA, 23 Brazil, 24 Taiwan, 25 and Europe 26 also have shown a decreasing trend in the incidence of MRSA associated to BSI. Overall the reasons for this are probably multifactorial and still partially unclear. Hand hygiene reinforcement 27 and CVC management policy, which became more selective in our ICU at the beginning of the survey 2000/2001, 22 may have contributed. Also as a consequence of the MDR microorganisms diffusion, since 2011 a senior infection disease physician (MV) was directly involved in the routine ICU assistance. As this is an observational study, the reasons for MRSA decrease in our institution are still partially unclear, and further analysis from various ICU s carrying out relationship between infection control measures is needed. Our observations support the notion that S. aureus BSI is decreasing, however, although this should be interpreted cautiously since the MRSA disease burden remains high. 28, 29 During the second period a significant increase in incidence was noted for various of the most common Gram-negative organisms associated to LC-BSI, but the most remarkable increase was noted for Enterobacteriaceae particularly K. pneumoniae. This striking change of the etiology with a predominance of Gram-negative was also seen in the large European multicentre EPIC II study although considering all bacterial isolates from different body sites. 14 The second study notable finding was related to mortality. During the second period we observed a reduction associated both to all Grampositive and Gram-negative non-fermentative microorganisms (A. baumannii and P. aeruginosa), whereas we registered a fierce over two-fold increase of mortality associated to Enterobacte- riaceae. This trend was similar when considering separately only the early death (Table III). The rise in resistance for all Enterobacteriaceae but particularly K. pneumoniae may had an impact on the increased mortality rate. Gramnegative LC-BSI has been associated with high mortality and the risk is further increased if appropriate antibiotic treatment is delayed because of antimicrobial resistance. 30 Several case-control studies reported a worse outcome associated with infections due to a resistant strain Although negative outcome was not a primary objective of our study and may be influenced by several factors, undoubtedly the inversion in mortality trend between Gram-positive and Gram-negative appears influenced by MDR among Enterobacteriaceae (Table IV). As in the past many different definitions for MDR microorganisms have been used in the medical literature to characterize the different patterns of resistance found in healthcare associated bacteria, limiting the possibility to compare the surveillance data, we decided to adopt the ECDC MDR standard definitions. 19 Actually the third finding in our study was an increase of Gram-negative MDR in the second period. This is consistent with those of recent studies, which report an increase in antibiotic resistance among Gram-negative in immunocompetent and immunocompromised hosts. 26 As resistance to multiple antibiotics increases the chances for inappropriate empiric therapy, which has been shown to be an independent risk factor for adverse outcome among bacteremic patients, 6 it was not surprising a strict relationship between mortality and MDR P. aeruginosa and Enterobacteriaceae (Table V). In the latter adverse outcome was even higher when considering separately K. pneumoniae KPC producing strains (77.4%). In relation to the source of infection it was not surprising that the highest antimicrobial susceptibility rate was associated to LC-BSI secondary to respiratory tract infections. We are aware of some limitations to our study. First it was carried out in a single center, secondly the data were collected for the purposes of infection control surveillance and conse- 986 MINERVA ANESTESIOLOGICA September 2015

9 ICU ACQUIRED BLOODSTREAM INFECTIONS ORSI quently some clinical information may be limited, third mortality was reported as all-cause mortality, and not as infection related death. However, although the study did not focus on antimicrobial treatment, as in the second period ( ) many strains were MDR we can assume that also many patients could not receive an early adequate therapy before microbiological results (i.e. colistin ± tigecycline for KPC-CR-Kp) affecting the final outcome. We are also aware that in a long period study (12 years) the adoption of different breakpoints might lead to some time bias effect on antimicrobial susceptibility, but surely not on the different bacterial species isolated from blood. The study adopted a clear and standardized clinical diagnostic criteria which avoided confounding colonizing agents not directly related to clinical disease, excluding blood culture duplicates was important, as inclusion of duplicates can result in overestimation of some microorganisms. Data was collected prospectively by an infection control team using the same methodology along all the study period, and in order to improve surveillance data comparison, ECDC MDR standard definitions were adopted. 19 Conclusions The present study provided some useful insight into the LC-BSI pathogens epidemiology at our institution. Following the appearance and spread of carbapenem resistance in Enterobacteriaceae, over the last years in our ICU the LC-BSI etiology shifted from Gram-positive to Gram-negative. It also associated mortality decreased among the former, whereas it increased in the latter, especially among Enterobacteriaceae with the diffusion of KPC producing K. pneumoniae. Our results underline the importance of surveillance, particularly of MDR organisms which add significantly to mortality. Since early appropriate empirical treatment improves survival in patients affected by serious infections, specific studies are needed to identify risk factors for MDR organisms isolation in ICU patients. Key messages Over the last decade in our institution microorganisms responsible for ICUacquired LC-BSI shifted from Gram-positive to Gram-negative. During the study period also the associated mortality decreased among the former, whereas it increased in the latter; mainly as a consequence for the diffusion of enterobacteriaceae expressing carbapenem-resistance due to K. pneumoniae carbapenemases production. In the second period we observed an increase of Gram-negative MDR and adverse outcome was even higher when considering separately K. pneumoniae KPC producing strains. References 1. Sligl W, Taylor G, Brindley PG. Five years of nosocomial gram-negative bacteremia in a general intensive care unit: epidemiology, antimicrobial susceptibility patterns, and outcomes. Int J Infect Dis 2006;10: Vincent JL, Sakr Y, Sprung CL, Ranieri VM, Reinhart K, Gerlach H et al. Sepsis in european intensive care units: result of the SOAP study. Crit Care Med 2006;34: Wisplinghoff H, Bischoff T, Tallent SM, Seifert H, Wenzel RP, Edmond MB. Nosocomial bloodstream infections in US hospitals: analysis of 24,179 cases from a prospective nationwide surveillance study. Clin Infect Dis 2004;39: Al-Rawajfah OM, Hewitt JB, Stetzer F, Cheema J. Length of stay and charges associated with health care-acquired bloodstream infections. Am J Infect Control 2012;40: Kang CI, Kin SH, Park WB, Lee KD, Kim HB, Kim EC et al. Bloodstream infections caused by antibiotic-resistant gram-negative bacilli: risk factors for mortality and impact of inappropriate initial antimicrobial therapy on outcome. Antimicrob Agents Chemother 2005;49: Marchaim D, Zaidenstein R, Lazarovitch T, Karpuch Y, Ziv T, Weinberger M. Epidemiology of bacteremia episodes in a single center: increase in Gram-negative isolates, antibiotics resistance and patients age. Eur J Microbiol Infect Dis 2008;27: Orsi GB, García-Fernández A, Giordano A, Venditti C, Bencardino A, Gianfreda R et al. Risk factors and clinical significance of endemic ertapenem resistant Klebsiella pneumoniae isolates in hospitalized patients. J Hosp Infect 2011;78: Orsi GB, Bencardino A, Vena A, Carattoli A, Venditti C, Falcone M et al. Patient risk factors for outer membrane permeability and KPC producing carbapenem-resistant Klebsiella pneumoniae isolation: results of a double casecontrol study. Infection 2013;41: Orsi GB, Franchi C, Marrone R, Giordano A, Rocco M, Venditti M. Laboratory confirmed bloodstream infection etiology in an intensive care unit: eight years study. Ann Ig 2012;24: Vol No. 9 MINERVA ANESTESIOLOGICA 987

10 This document is protected by international copyright laws. No additional reproduction is authorized. It is permitted for personal use to download and save only one file and print only one copy of this Article. It is not permitted to make additional copies (either sporadically or systematically, either printed or electronic) of the Article for any purpose. It is not permitted to distribute the electronic copy of the article through online internet and/or intranet file sharing systems, electronic mailing or any other means which may allow access to the Article. The use of all or any part of the Article for any Commercial Use is not permitted. The creation of derivative works from the Article is not permitted. The production of reprints for personal or commercial use is not permitted. It is not permitted to remove, cover, overlay, obscure, block, or change any copyright notices or terms of use which the Publisher may post on the Article. It is not permitted to frame or use framing techniques to enclose any trademark, logo, or other proprietary information of the Publisher. ORSI 10. Garner JS, Jarvis WR, Emori TG, Horan TC, Hughes JM. CDC definitions for nosocomial infections. In: Olmsted R.N., editors. APIC Infection Control and Applied Epidemiology: Principles and Practice. St Louis: Mosby; p. A-1-A Maki DG, Kluger DM, Crnich CJ. The risk of bloodstream infection in adults with different intravascular devices: a systematic review of 200 published prospective studies. Mayo Clin Proc 2006;81: Mermel LA. Prevention of intravascular catheter-related infections. Ann Intern Med 2000;132: Polderman KH, Girbes AR. Central venous catheter use. Part 2: infectious complications. Intensive Care Med 2002;28: Vincent JL, Rello J, Marshall J, Silva E, Anzueto A, Martin CD et al. International study of the prevalence and outcomes of infection in intensive care units. JAMA 2009;302: Wilson ML. General principles of specimen collection and transport. Clin Infect Dis 1996;22: Cohen Stuart J, Leverstein-Van Hall MA, Dutch working Party on the detection of highly resistant microorganisms. Guideline for phenotypic screening and confirmation of carbapenemases in Enterobacteriaceae. Int J Antimicob Agents 2010;36: Clinical and Laboratory Standards Institute. Performance standards for antimicrobial susceptibility testing. CLSI document. Wayne, PA 18. EUCAST breakpoint tables for interpretation of MIC s and zone diameters. [Internet]. Available from: [cited on 2015, July 21]. 19. Magiorakos AP, Srinivasan A, Carey RB, Carmeli Y, Falagas ME, Giske CG et al. Multidrug-resistant, extensively drug-resistant and pandrug-resistant becteria: an international expert proposal for interim standard definitions for acquired resistance. Clin Microbiol Infect 2012;18: Almirante B, Rodrìguez D, Park BJ, Cuenca-Estrella M, Planes AM, Almela M et al. Epidemiology and predictors of mortality in cases of Candida bloodstream infection: results from population-based surveillance, Barcelona, Spain, from 2002 to J Clin Microbiol 2005;43: Velasco E, Portugal RD, Salluh JIF. A simple score to predict early death in adult cancer patients with bloodstream infections. J Infect 2009;59: Orsi GB, Raponi M, Franchi C, Rocco M, Mancini C, Venditti M. Surveillance and infection control in an intensive care unit. Infect Control Hosp Epidemiol 2005;26: Burton DC, Edwards JR, Horan TC, Jernigan JA, Fridkin SK. Methicillin-resistant Staphylococcus aureus central ICU ACQUIRED BLOODSTREAM INFECTIONS line associated bloodstream infections in US intensive care units, JAMA 2009;301: Marra AR, Camargo LF, Pignatari AC, Sukiennik T, Behar PR, Medeiros EA et al. Nosocomial bloodstream infections in brazilian hospitals: analysis of 2,563 cases from a prospective nationwide surveillance study. J Clin Microbiol 2011;49: Liu CY, Liao CH, Chen YC, Chang SC. Changing epidemiology of nosocomial bloodstream infections in 11 teaching hospitals in Taiwan between 1993 and J Microbiol Immunol Infect 2010;43: Gudiol C, Bodro M, Simonetti A, Tubau F, Gonzalez-Barca E, Cisnal M et al. Changing etiology, clinical features, antimicrobial resistance and outcomes of bloodstream infection in neutropenic cancer patients. Clin Microbiol Infect 2013;19: Jain R, Kralovic SM, Evans ME, Ambrose M, Simbartl LA, Obrosky DS et al. Veterans affairs initiative to prevent methicillin-resistant Staphylococcus aureus infections. N Engl J Med 2011;364: Klein E, Smith DL, Laxminarayan R. Hospitalizations and deaths caused by methicillin-resistant Staphylococcus aureus, United States, Emerg Infect Dis 2007;13: Klevens RM, Morrison MA, Nadle J, Petit S, Gersham K, Ray S et al. Invasive methicillin-resistant staphylococcus aureus infections in the United States. JAMA 2007;298: Orsi GB, Falcone M, Venditti C. Surveillance and management of multi-drug resistant microrganisms. Expert Rev Anti Infect Ther 2011;9: Borer A, Saidel-Odes L, Riesenberg K, Eskira S, Peled N, Nativ R et al. Attributable mortality rate for carbapenemresistant Klebsiella pneumoniae bacteremia. Infect Control Hosp Epidemiol 2009;30: Gasink LB, Adelstein PH, Lautenbach E, Synnestvedt M, Fishman NO. Risk factors and clinical impact of Klebsiella pneumoniae carbapenemase-producing K. pneumoniae. Infect Control Hosp Epidemiol 2009;30: Patel G, Huprikar S, Factor SH, Jenkins SG, Calfee DP. Outcomes of carbapenem-resistant Klebsiella pneumoniae infection and the impact of antimicrobial and adjunctive therapies. Infect Control Hosp Epidemiol 2008;29: Schwaber MJ, Klarfeld-Lidji S, Navon-Venezia S, Schwartz D, Leavitt A, Carmeli Y. Predictors of carbapenem-resistant Klebsiella pneumoniae acquisition among hospitalized adults and effect of acquisition on mortality. Antimicrob Agents Chemother 2008;52: Conflicts of interest. The authors certify that there is no conflict of interest with any financial organization regarding the material discussed in the manuscript. Received on April 16, Accepted for publication on November 18, Epub ahead of print on November 20, Corresponding author: GB Orsi, Dipartimento di Sanità Pubblica e Malattie Infettive, Sapienza Università di Roma, P.le Aldo Moro 5, Roma, Italy. giovanni.orsi@uniroma1.it 988 MINERVA ANESTESIOLOGICA September 2015