Title: Bloodstream infections in cancer patients. Risk factors associated with mortality. International Journal of Infectious Diseases

Transcription

1 Accepted Manuscript Title: Bloodstream infections in cancer patients. Risk factors associated with mortality. Authors: Beda Islas-Muñoz, Patricia Volkow-Fernández, Cynthia Ibanes-Gutiérrez, Alberto Villamar-Ramírez, Diana Vilar-Compte, Patricia Cornejo-Juárez PII: S (18)30081-X DOI: Reference: IJID 3209 To appear in: International Journal of Infectious Diseases Received date: Revised date: Accepted date: Please cite this article as: Islas-Muñoz Beda, Volkow-Fernández Patricia, Ibanes-Gutiérrez Cynthia, Villamar-Ramírez Alberto, Vilar-Compte Diana, Cornejo-Juárez Patricia.Bloodstream infections in cancer patients.risk factors associated with mortality.international Journal of Infectious Diseases This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

2 Title: Bloodstream infections in cancer patients. Risk factors associated with mortality. Authors: Beda Islas-Muñoz, Patricia Volkow-Fernández, Cynthia Ibanes-Gutiérrez, Alberto Villamar-Ramírez, Diana Vilar-Compte and Patricia Cornejo-Juárez*. Infectious Diseases Department, Instituto Nacional de Cancerología (INCan), Av. San Fernando No. 22, Col. Sección XVI, Del. Tlalpan, Mexico City, Mexico. Phone: (+52) , ext *Corresponding author. PCJ: BIM: PVF: CIG: AVR: DVC: Word count of main text: Highlights 1

3 Bloodstream infections (BSI) are cause of severe complications in cancer patients Secondary BSI and central-related BSI were the most common in solid tumors Abstract Primary BSI and mucosal barrier injury BSI were described in hematological patients Mortality at 30-day was increased with multidrug resistant Gram-negative bacteria Inappropriate antimicrobial treatment in the first 24 h was related with mortality Objective: The aim of this study was to evaluate the clinical characteristics and risk factors associated with mortality in cancer patients with bloodstream infections (BSI), analyzing multidrug resistant bacteria (MDRB). Methods: We conducted a prospective observational study at a cancer referral center from August 2016 to July 2017, which included all BSI. Results: 4,220 patients were tested with blood cultures; 496 were included. Mean age was 48 years. In 299 patients with solid tumors, secondary BSI and Central Line-Associated BSI (CLABSI) were the most common (55.9% and 31.8%, respectively). In 197 hematologic patients, primary and mucosal barrier injury (MBI) BSI, were the main type (38.6%). Gram-negative were the most frequent bacteria 2

4 (72.8%), with E. coli occupying the first place (n=210, 42.3%), 48% were Extended- Spectrum Beta-Lactamase (ESBL) producers, and 1.8% were resistant to carbapenems. Mortality at day 30, was 22%, but reached 70% when patients did not receive an appropriate antimicrobial treatment. Multivariate analysis showed that progression or relapse of the oncologic disease, inappropriate antimicrobial treatment, and having resistant bacteria were independently associated with 30- day mortality. Conclusions: Emergence of MDR bacteria is an important healthcare problem worldwide. Patients with BSI, particularly those patients with MDR bacteria have a higher mortality risk. Keywords: Bloodstream infection; Multidrug resistant bacteria; mortality; cancer Introduction BloodStream Infections (BSI) are a major cause of life-threatening complications in patients with cancer, due to the causing delays in chemotherapy, longer hospital stay, suboptimal treatment, higher mortality rate, and increased healthcare costs [Montassier et al 2013, Marin et al 2014, Schelenz et al 2013). The prevalence of BSI ranges from 11% to 38% and overall mortality reaches 40% (Montassier et al 2013). The extensive emergence of MultiDrug-Resistant (MDR) bacteria has increased the burden of morbidity and mortality among cancer patients with BSI (Marin et al 2014). In recent years, a shift from Gram-positive to Gram-negative organisms has been documented in the etiology of BSI; however, these data vary depending on the 3

5 geographic area. Gram-negative bacteria are most frequent in the U.S. and Latin America, while Gram-positive bacteria have been reported as the prevalent cause in Europe (Marin et al 2014). The aim of this study was to evaluate the clinical epidemiological characteristics and risk factors associated with mortality in cancer patients with BSI, with special emphasis on MDR bacteria. Methods We conducted a prospective observational study at the Instituto Nacional de Cancerología (INCan), a 135-bed cancer referral center in Mexico City, Mexico. From August 2016 to July 2017, all positive blood cultures detected at the Microbiology Laboratory were recorded, and all episodes confirmed as BSI were included in the study. The latter was approved by the INCan Institutional Review Board (016/055/INI and CEI/1126/16). BloodStream Infection (BSI) definitions: BSI was defined as a laboratory-confirmed isolation of at least one Gram-negative or Gram-positive bacterium from blood samples classified as follows: A) Central- Line Associated BSI (CLABSI): If time- to-positivity between blood cultures taken from the catheter was 2 or more h apart from that of the blood culture taken from peripheral puncture, with signs of systemic infection (fever, chills, and/or hypotension) and no apparent source of infection, except for the catheter, and/or the catheter-tip culture being positive for the same organisms (when the catheter was removed), and/or signs and symptoms of catheter entry-site infection with the same strain isolated from the blood (Schelenz et al 2013); B) MBI related 4

6 bloodstream infection: If the patient had neutropenia (neutrophils <500 cells/mm 3 at least in two occasions or <100 cells/mm 3 at least on one), or an allogeneic hematopoietic stem-cell transplantation or grade 3-4 gastrointestinal graft-versushost disease or diarrhea 1 L or more in 24 h documented within 7 days prior to or on the day of collection of positive blood culture (See et al 2013); C) Secondary BSI was diagnosed when there was another source of infection, such that the primary site of infection may have seeded the bloodstream secondarily. Secondary bacteraemias were classified either as related to a Urinary Tract Infection (UTI), pneumonia, skin and soft tissue, cholangitis, abdominal or other source, and D) Primary BSI occurred when no underlying infection was diagnosed despite intense a clinical and radiological workout. Persistent bacteremia was defined as the continuous presence of bacteria in the bloodstream 48 h after administration of appropriate treatment. Polymicrobial bacteraemia, when two or more microorganisms were isolated from blood cultures. Contamination: Coagulase-Negative Staphylococcus (CoNS) and other skin flora isolated from a single blood culture in a patient without any clinical evidence of infection. Appropriate antimicrobial treatment: An antibiotic with activity against the bacteria isolated initiated within the first 48 h of the blood culture sample and duration of treatment of at least 72 h. De-escalation: When microbiological results were available and empirical antibiotics were stopped or reduced in number and/or narrowed in antimicrobial spectrum. 5

7 The following information was recorded: age; gender; comorbidities; cancer types (classified as solid or hematological malignancy); current status of cancer (complete or partial remission, recent diagnosis, relapse, or progression); presence of long-term central venous catheter; treatment with chemotherapy during the last month; radiotherapy; surgery, hospitalization, and previous antimicrobials in the last 3 months; Charlson Index, Intensive Care Unit (ICU) admission; Sequential Organ Failure Assessment score (SOFA); length of ICU stay, and mechanical ventilation. Bacterial identification was performed by Mass Spectrometry Especially Matrix- Assisted Laser Desorption and Ionization - Time of Flight - Mass Spectrometry (MALDI-TOF-MS; Microflex, USA). Antimicrobial susceptibility testing was performed by means of BD Automated Phoenix TM (USA) and the Kirby-Bauer disk diffusion technique in the case of resistant strains (Clinical Laboratory Standards Institute, CLSI). The following MDR bacteria were evaluated: Methicillin-resistant Staphylococcus aureus (MRSA); Vancomycin-resistant Enterococcus faecium (VRE); Extended- Spectrum Beta-Lactamases (ESBL) Escherichia coli, Klebsiella spp., or Enterobacter spp.; Pseudomonas aeruginosa, and Acinetobacter spp., resistant to third-generation cephalosporins and carbapenems. Other Gram-negative bacteria were considered MDR if they were resistant at least to three active antimicrobial groups, such as fluoroquinolones, third-generation cephalosporins, aminoglycosides, and/or carbapenems (Magiorakos et al 2012). Type and days of antimicrobial treatment were recorded. 6

8 Early BSI mortality rate was defined as death within 48 h from the day blood cultures were obtained. Overall case-mortality rate was defined as death by any cause within 30 days of onset. Patients confidentiality was preserved, all data were anonymous, and patients cannot be identified. Statistical analysis The Student t test or the Mann-Whitney U test was used to compare continuous variables according to a normal or non-normal data distribution. The Chi-square or Fisher exact test was utilized to compare categorical variables. Variables with p values of <0.5 in the univariate analysis were included in the multivariate analysis. A logistic regression model was performed for risk factors associated with mortality. Odds Ratios (OR) with 95% Confidence Intervals (95% CI) were calculated. Rates of Overall Survival (OS) were estimated by means of the Kaplan- Meier analysis. P values of 0.05 were considered statistically significant. Data was analyzed employing Epi-Info (ver. 7) and STATA (ver. 12) software. Results During the study period, 4,220 patients were tested with blood cultures: 585 (13.9%) were positive; 22 were considered contaminated, and 67 were classified as with persistent bacteremia; the current analysis included 496 patients (11.7%). General characteristics 7

9 These are shown in table 1. Mean age was 48.1 ± 17.1 years, 299 (60.3%) had solid tumors, and 197 (39.7%) had hematologic malignancies. Two hundred ninetynine patients were women (60.3%); there were more women among patients with solid tumors (n = 187, 62.5%) compared with hematologic patients (n = 95, 48.2%; p = 0.001). Sixty four (13%) patients were admitted in the ICU (no differences were found between patients with solid vs. hematologic cancer, 11% vs. 15.6%, respectively (p = 0.126). Two hundred fifty seven patients (51.8%) had progression or relapse of the oncologic disease, and 238 (48%) had received chemotherapy during the last month. Secondary BSI was the most frequent type of bacteremia (n = 227, 45.8%), mainly due to a UTI (n = 93, 40.8%); 18 patients (19.2%) had nephrostomies. CLABSI was the second most frequent BSI (31.5%). In patients with solid tumors, secondary bacteraemias and CLABSI were the most common BSI (55.9% and 31.8%, respectively). In hematologic patients, primary and MBI bacteremia comprised the most frequent type (38.6%); 93 (47%) had severe neutropenia. Pathogens The distribution of isolated microorganisms is depicted in Figure 1. Gram-negative bacteria were the most common isolated microorganisms (n = 361, 72.8%), while Escherichia coli was the most frequent isolate (n = 210, 42.3%), followed by Klebsiella spp. (n = 45, 9.1%) and Pseudomonas aeruginosa (n = 29, 5.8%). 8

10 On comparing Enterobacteriaceae ESBL producers (n = 123, 38.1%) vs. non- ESBL (n = 148, 54.6%), patients with ESBL less frequently had neutropenia (p = 0.001), received fewer appropriate antimicrobials (p = 0.03), and were not deescalated as much as non-esbl (p <0.001). There was no relation with previous recent hospitalization, chemotherapy, or antibiotics (Table 2), Comparison was performed between P. aeruginosa and Acinetobacter spp. MDR (n = 9, 20.9%) vs. non-mdr (n = 34, 79.1%). Nosocomial bacteremia was more frequent for MDR vs. non-mdr (66.7% vs. 8.8% respectively, p = 0.04). Eight patients had been hospitalized in the ICU; only two had an isolated MDR (p = 1). Gram-positive bacteria were found in 135 isolates (27.2%). The most common pathogens were coagulase-negative Staphylococci (n = 58, 11.7%) and Staphylococcus aureus (n = 50, 10.1%). Methicillin-Resistant S.aureus (MRSA) was isolated in only four patients. Enterococcus spp. was recovered from 22 patients (4.4%); seven of these were E. faecium, only two of which were Vancomycin-Resistant Enterococcus (VRE). There were no clinical differences between MRSA and VRE vs. Methicillin-Susceptible S. aureus (MSSA) and Vancomycin-Susceptible Enterococcus (VSE) (Table 2). Outcome Forty-one patients (8.3%) had persistent bacteremia, 53.6% of these with CLABSI. No endocarditis was diagnosed. Thirty six patients (7.3%) died during the first 48 h of the BSI diagnosis, in 33 (91.7%) a Gram-negative bacterium was isolated, and only three (8.3%) had a Gram-positive (p = 0.002). 9

11 At 30-day, 109 patients (22%) had died: 73 with a Gram-negative (67%), no differences in mortality were documented between ESBL (18.7%) vs. non-esbl (21%, p = 0.625). Mortality for MDR P. aeruginosa or Acinetobacter spp. was 66.7% (6/9), and for non-mdr, mortality was 14.7% (5/34) (p = 0.004). Thirty-six patients (26.7%) with a Gram-positive bacteria died during the same period of time; without differences between MRSA/EVR 33.3% (2/6) vs. MSSA/VSE 9.3% (4/43) (p = 0.151). Fifty patients (10.1%) received an inadequate empirical antimicrobial treatment, 35 (70%) died during the first month. Of 446 patients who received appropriate antibiotics, mortality was 74 patients (16.6%) during the same time period (p <0.001). Analyzing patients who received empirical treatment with carbapenems vs. piperacillin/tazobactam, there was no difference in 30-day mortality (p = 0.216). Multivariate analysis for 30-day mortality revealed that progression or relapse of the oncologic disease (OR, 1.9; 95% CI, ; p = 0.008), inappropriate antimicrobial treatment (OR, 10.9; 95% CI, ; p = <0.001), and having resistant bacteria, MDR or ESBL, (OR, 1.59; 95% CI, ; p = 0.05) were independently associated. Discussion The prevalence of BSI in patients with cancer described in this series was 11.7%, lower compared with other reports from different geographical areas, which ranged between 11.8 and 33.3% (Obeng-Nkrumah et al 2015). Secondary BSI and CLABSI were more prevalent in patients with solid tumors, although primary and MBI-BSI were more frequent in patients with hematological 10

12 malignancies. These findings are consistent with previous reports (Marin et al 2014). In studies performed worldwide throughout the 1960s and 1970s, Gram-negative bacilli were the most frequent causative agents of BSI in patients with cancer and neutropenia. During the next two decades, there was a shift to Gram-positive cocci (30-70%) (Gudiol et al 2013, Ruhnke et al 2014, Fortún et al 2011, Trecarichi et al 2015). This substitution was not observed in our hospital, where Gram-negative bacteria have predominated in the last three decades. In the current study, we describe a prevalence of 72.8%, probably related to the non-use of antimicrobial prophylaxis in neutropenic patients, particularly fluoroquinolones. Escherichia coli was the bacterium more frequently isolated (42.3%). This is in contrast with that reported in other countries, where P. aeruginosa is the predominant bacterium isolated from blood in neutropenic patients with cancer (Marin et al 2014, Miedema et al 2013). Even in series in which E. coli is also the main Gram-negative bacteria isolated, the proportion is lower than that in our study (11% to 18%) (Montassier et al 2013, Schelenez et al 2013). Enterobacteriaceae-ESBL-producing strains were identified in 38%; one half of all E. coli isolated were ESBL. In previous reports from our hospital, we have described E. coli as the main bacterium isolated (40% to 73%), with the predominant ESBL identified CTX-M15 (Cornejo-Juárez et al 2012, Montes et al 2014, Hernaiz-Leonardo et al 2017). Other studies have reported ESBL- Enterobacteriaceae between 0 and 42%; the wide variation in these results is directly related with stewardship programs, antibiotic prescription policies, and the country-of-report (Montassier et al 2013, Gudiol et al 2013, Miedema et al 2013). 11

13 The high prevalence of ESBL-harboring Enterobacteriaceae has resulted in an increased use of carbapenems, and has been related with the emergence of strains resistant to these antibiotics (Peratla et al 2012). We found four carbapenem-resistant strains (0.8%), similar to what has been reported at other hospitals (0.8%-3.3%) (Satlin et al 2016). We did not isolate Klebsiella-producing carbapenems, in contrast with that reported in series including patients with hematological neoplasms, where this reaches up to 38% (Trecarichi et al 2015). At present, Gram-positive bacteria comprise a group that is an important cause of MDR-BSI in some countries. We only found 8% of MRSA (4/50 isolates), and 28.6% of VRE (2/7 Enterococcus isolates), compared with what we had found in previous years, when the number of these strains was much higher. Since 2012, we intensified infection-control policies, including a strict pre-authorization prescription program of broad- spectrum antibiotics, hand hygiene, and new hospital facilities, with less crowding and more isolation rooms that, overall, have resulted in a decrease in the number of resistant strains. The proportion of appropriate empirical antimicrobial therapy exerts an important impact on the survival of patients with BSI (Koupetori et al 2014, Retamar et al 2012, Kang et al 2005). In this study, we found statistically higher mortality in those who did not receive an appropriate antimicrobial scheme in the first 48 h (50%), in comparison with those who received it (19%). Appropriate selection of empirical antimicrobial treatment depends on some epidemiological factors, such as having received broad-spectrum antibiotics, or prior hospitalization, particularly in high-risk areas such as the ICU (Koupetori et al 2014). Currently, it is recognized that the initial treatment with broad-spectrum 12

14 antimicrobials covers the pathogens most probably involved in the infection and improves outcomes; however, it could lead to the emergence of MDR pathogens (Koupetori et al 2014). One of the strategies recommended to avoid this MDR increase is timely antimicrobial de-escalation, which includes either switching to a narrower-spectrum antimicrobial when the susceptibility test is available, or reducing the number of antibiotics used (Koupetori et al 2014). The 30-day mortality rate was 22%, similar to other reports in cancer patients with BSI (13% to 36%) (Gudiol et al 2013, Satlin et al 2016, Velasco et al 2006). However, mortality was significantly increased when the BSI was secondary to Gram-negative resistant bacteria: for ESBL-Enterobacteriaceae, this was 30%, and for MDR P. aeruginosa and Acinetobacter, it reached 67%. We did not find any difference in mortality in patients who initiated an empirical treatment with carbapenems vs. Piperacillin/tazobactam. It is important to point out, as there is an ongoing controversy between initial use of these two antibiotics and outcomes (Ofer-Friedman et al 2015). This study has several strengths. All data, clinical and microbiological, was prospectively collected, which probably decreased some of the bias from other reports. The study was conducted in a tertiary-care cancer referral center with a captive population that allows close monitoring and evaluation of outcomes. Our results can be extrapolated to other hospitals with similar characteristics, mainly in other Latin American countries which population characteristics and economic development are similar to Mexico. On the other hand, as it was only conducted at a single institution, it may not reflect a national epidemiology, nor from other cancer centers. 13

15 In summary, the emergence of MDR bacteria is an important healthcare problem at our Institution, as it is in many other countries worldwide. This is of particular concern among patients with cancer. Thirty-day mortality was higher in patients with Gram-negative MDR bacteria, closely related with inappropriate antimicrobial treatment in the first 24 hours. We highly recommend initiating broad-spectrum antibiotics based on local susceptibility patterns and, as soon as the susceptibility is available, the de-escalation of antimicrobials, narrowing the spectrum and/or the number. Conflict of interest: All authors report no conflicts of interest. Financial Support: This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. Acknowledgments: We thank the staff of the Microbiology Laboratory 14

16 References 1. Montassier E, Batard E, Gastinne T, Potel G, de La Cochetière MF. Recent changes in bacteremia in patients with cancer: a systematic review of epidemiology and antibiotic resistance. Eur J Clin Microbiol Infect Dis 2013; 32: Marín M, Gudiol C, Ardanuy C, García-Vidal C, Calvo M, Arnan M, Carratalà J. Bloodstream infections in neutropenic patients with cancer: differences between patients with haematological malignancies and solid tumours. J Infect 2014; 69: Schelenz S, Nwaka D, Hunter PR. Longitudinal surveillance of bacteraemia in haematology and oncology patients at a UK cancer centre and the impact of Ciprofloxacin use on antimicrobial resistance. J Antimicrob Chemother 2013; 68: See I, Iwamoto M, Allen-Bridson K, Horan T, Magill SS, Thompson ND. Mucosal barrier injury laboratory-confirmed blood infection: results from a field test of a new national healthcare safety network definition. Infect Hosp Control Epidemiol 2013; 34: Magiorakos AP, Srinivasan A, Carey RB, Carmeli Y, Falagas ME, Giske CE, et al. Multidrug-resistant, extensively drug-resistant and pandrug-resistant bacteria: an international expert proposal for interim standard definitions for acquired resistance. Clin Microbiol Infect. 2012; 18: Obeng-Nkrumah N, Labi AK, Acquah ME, Donkor ES. Bloodstream infections in patients with malignancies: implications for antibiotic treatment in a Ghanaian tertiary setting. BMC Res Notes 2015; 8:

17 7. Gudiol C, Bodro M, Simonetti A, Tubau F, González-Barca E, Cisnal M, et al. Changing aetiology, clinical features, antimicrobial resistance, and outcomes of bloodstream infection in neutropenic cancer patients. Clin Microbiol Infect 2013; 19: Ruhnke M, Arnold R, Gastmeier P. Infection control issues in patients with haematological malignancies in the era of multidrug-resistant bacteria. Lancet Oncol 2014; 15: e606-e Fortún J, Sanz MÁ, Madero L, López J, de la Torre J, Jarque I, Vallejo C. Update on bacteraemia in oncology and hematology. Enferm Infecc Microbiol Clin 2011; 29 (Suppl 4): Trecarichi EM, Pagano L, Candoni A, Pastore D, Cattaneo C, Fanci R, et al; HeMABIS Registry SEIFEM Group, Italy. Current epidemiology and antimicrobial resistance data for bacterial bloodstream infections in patients with hematologic malignancies: an Italian multicentre prospective survey. Clin Microbiol Infect 2015; 21(4): Miedema KG, Winter RH, Ammann RA, Droz S, Spanjaard L, de Bont ES, et al. Bacteria causing bacteremia in pediatric cancer patients presenting with febrile neutropenia- species distribution and susceptibility patterns. Support Care Cancer 2013; 21: Cornejo-Juárez P, Pérez-Jiménez C, Silva-Sánchez J, Velázquez-Acosta C, González-Lara F, Reyna-Flores F, Sánchez-Pérez A, Volkow-Fernández P. Molecular analysis and risk factors for Escherichia coli producing extendedspectrum β-lactamase bloodstream infection in hematological malignancies. PLOS ONE. 2012;7:e

18 13. Montes CV, Vilar-Compte D, Velázquez C, Golzarri MF, Cornejo-Juárez P, Larson EL. Risk factors for extended spectrum β-lactamase-producing Escherichia coli versus susceptible E. coli in surgical site infections among cancer patients in Mexico. Surg Infect (Larchmt) 2014; 15: Hernaiz-Leonardo JC, Golzarri MF, Cornejo-Juárez P, Volkow P, Velázquez C, Ostrosky-Frid M, Vilar-Compte D. Microbiology of surgical site infections in patients with cancer: a 7-year review. Am J Infect Control 2017; 45: Peralta G, Lamelo M, Álvarez-García P, Velasco M, Delgado A, Horcajada JP, et al; SEMI- BLEE STUDY GROUP. Impact of empirical treatment in extendedspectrum beta-lactamase-producing Escherichia coli and Klebsiella spp. bacteremia. A multicentric cohort study. BMC Infect Dis 2012; 12: Satlin MJ, Cohen N, Ma KC, Gedrimaite Z, Soave R, Askin G, et al. Bacteremia due to carbapenem-resistant Enterobacteriaceae in neutropenic patients with hematologic malignancies. J Infect 2016; 73: Koupetori M, Retsas T, Antonakos N, Vlachogiannis G, Perdios I, Nathanail C, et al. Hellenic Sepsis Study Group. Bloodstream infections and sepsis in Greece: over-time change of epidemiology and impact of de-escalation on final outcome. BMC Infect Dis 2014; 14: Retamar P, Portillo MM, López-Prieto MD, Rodríguez-López F, de Cueto M, García MV, et al; SAEI/SAMPAC Bacteremia Group. Impact of inadequate empirical therapy on the mortality of patients with bloodstream infections: a propensity score-based analysis. Antimicrob Agents Chemother 2012; 56:

19 19. Kang CI, Kim 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: Velasco E, Byington R, Martins CA, Schirmer M, Dias LM, Gonçalves VM. Comparative study of clinical characteristics of neutropenic and non-neutropenic adult cancer patients with bloodstream infections. Eur J Clin Microbiol Infect Dis 2006; 25: Ofer-Friedman H, Shefler C, Sharma S, Tirosh A, Tal-Jasper R, Kandipalli D, et al. Carbapenems versus Piperacillin-Tazobactam for bloodstream infections of non-urinary source caused by Extended-spectrum beta-lactamase-producing Enterobacteriaceae. Infect Control Hosp Epidemiol 2015; 36:

20 Figure legend Figure 1. Microorganisms isolated from 496 cancer patients with bloodstream infections. Footnote: CNS: Coagulase Negative Staphylococcus; GN-NF: Gram-Negative Non-Fermenters. 19

21 Table 1. Clinical and demographic characteristics of the cancer patients with BloodStream Infection (BSI) Characteristic N (%) Total (n = 496) Age (years)* 48.1 ± 17.1 Gender-Female 299 (60.3) Solid tumors Hematologic malignancy Stage Recent diagnosis Progression Relapse Partial remission Complete remission 299 (60.3) 197 (39.7) 169 (34.1) 201 (40.5) 56 (11.2) 51 (10.3) 19 (3.8) Chemotherapy within last 30 days 356 (71.8) Radiotherapy within 6 months 87 (17.5) Surgery within 3 months 113 (22.8) Hospitalization 295 (59.5) Previous antibiotics 256 (51.6) Central venous catheter 266 (53.6) Charlson score (50.5) >2 245 (49.5) Hospitalization for bacteremia treatment 335 (75.3) 20

22 Nosocomial bacteremia 109 (22) Gram-positive 135 (27.2) Gram-negative 361 (72.8) Polymicrobial bacteremia 44 (8.9) Persistent bacteremia 41 (8.3) Appropriate antimicrobial treatment 448 (90.3) Source of bacteremia Primary Mucosa-barrier injury CLABSI & Secondary Urinary tract infection Abdominal source Skin and soft tissue Pneumonia Other 75 (15.1) 38 (7.7) 156 (31.5) 227 (45.8) 93 (41) 82 (36.1) 34 (15) 17 (0.5) 1 (0.5) *Mean ± standard deviation. & CLABSI: Central Line-Associated Bloodstream Infection; Secondary sources percentage was obtained considering secondary BSI as the whole. 21

23 Table 2. Clinical and microbiological characteristics of patients with bloodstream infection (BSI) with main resistant bacteria Characteristic N (%) Solid tumor Escherichia coli, Klebsiella spp. (n = 123, (n = 148, Pseudomonas aeruginosa, Staphylococcus aureus or and Enterobacter spp. (n =271) Acinetobacter spp. (n = 43) Enterococcus faecium (n = 49) ESBL* Non-ESBL P MDR & Non- P MRSA MSSA** P (n = 9, MDR (n = or VRE or EVS && 45.4%) 54.6%) 20.9%) 34, (n = 6, (n = 43, 79.1%) 12.2%) 87.8%) 81 (65.9) 80 (54.1) (66.7) 23 (67.6) 1 3 (50) 28 (65.1) Hematologic malignancy 42 (34.1) 68 (45.9) 3 (33.3) 11 (32.4) 3 (50) 15 (34.9) Chemotherapy 60 (48.8) 64 (43.2) (33.3) 17 (50) (50) 17 (39.5) Hospitalization 76 (61.8) 94 (63.5) (66.7) 19 (55.9) (50) 24 (55.8) 1 Antibiotics 73 (59.3) 75 (50.7) (66.7) 17 (50) (66.7) 16 (37.2) Nosocomial bacteremia 32 (26) 26 (17.6) (66.7) 3 (8.8) (16.7) 9 (20.1) 1 Neutropenia 29 (23.6) 62 (41.9) (22.2) 7 (20.5) 1 1 (16.7) 8 (18.6) 1 22

24 Required hospitalization 105 (85.4) 131 (88.5) (100) 27 (79.4) (100) 41 (95.3) NA Appropriate antimicrobial treatment 106 (86.2) 139 (93.9) (44.4) 32 (94.1) (100) 40 (93) NA De-escalation of 34 (27.6) 82 (55.4) < (22.2) 18 (52.9) (50) 28 (65) antimicrobials Early mortality (48 h) 13 (10.6) 12 (8.1) (44.4) 2 (5.9) (2.4) 1 30-day mortality 37 (30.1) 35 (23.6) (66.7) 5 (14.7) (33.3) 4 (9.3) *ESBL: Extended-Spectrum Beta-Lactamase. & MDR: MultiDrug-Resistant; MRSA: Methicillin-Resistant S. aureus; VRE: Vancomycin-Resistant Enterococci; **MSSA: Methicillin-Susceptible S. aureus; && VSE: Vancomycin- Susceptible Enterococci; Within 1 month; Within 3 months. 23

25 Table 3. Multivariate analysis of factors associated with 30-day mortality in cancer patients with BloodStream Infection (BSI) 30-day mortality Characteristic - Age 60 years Alive (n = 105 (27.1) Death (n = Univariate Multivariate N (%) 387, 78%) 109, 22%) OR (95% CI) P OR (95% CI) P Age <60 years 282 (72.9) 76 (69.7) 1.16 ( ) Solid tumor Hematological Recent diagnosis or remission* Progression/re-lapse 230 (59.4) 157 (40.6) 199 (51.4) 188 (48.6) Gram-positive 99 (25.6) Gram-negative 288 (74.4) Monomycrobial 352 (91) 33 (30.3) 69 (63.3) 40 (36.7) 40 (36.7) 1.18 ( ) 0.84 ( ) ( ) 1.82 ( ) ( ) (63.3) 36 (33) 0.69 ( ) (67) 0.75 ( ) 100 (91.7) 0.9 (0.37-2) Polymicrobial 35 (9) 9 (8.3) 0.95 ( ) 24

26 Appropriate treatment 372 (96.1) 74 (67.9) 11.7 ( ) < Innappropiate treatment 15 (3.9) 35 (32.1) 10.9 ( ) Non-MDR 245 (63.3) 55 (50.5) 1.69 ( ) MDR and/or ESBL 142 (36.7) 54 (49.5) 1.59 ( ) Neutrophils > (72.4) 80 (73.4) 0.94 ( ) Neutrophils (27.6) 29 (26.6) 1.18 ( ) *Oncologic disease. <