Healthcare-associated Gram-negative bloodstream infections: antibiotic resistance and predictors of mortality

Accepted Manuscript Healthcare-associated Gram-negative Bloodstream Infections: Antibiotic Resistance and Predictors of Mortality Önder Ergönül, Mehtap Aydin, Alpay Azap, Seniha Başaran, Süda Tekin, Şafak Kaya, Serda Gülsün, Gülşen Yörük, Ebru Kurşun, Ayşegül Yeşilkaya, Funda Şimşek, Emel Yılmaz, Hüseyin Bilgin, Çiğdem Hatipoğlu, Hatice Cabadak, Yasemin Tezer, Turhan Togan, İlkay Karaoğlan, Asuman İnan, Aynur Engin, Hikmet Eda Alışkan, Serap Şimşek Yavuz, Şebnem Erdinç, Lütfiye Mulazimoglu, Özlem Azap, Füsun Can, Halis Akalın, Funda Timurkaynak PII:

S0195-6701(16)30328-0

DOI:

10.1016/j.jhin.2016.08.012

Reference:

YJHIN 4893

To appear in:

Journal of Hospital Infection

Received Date: 31 January 2016 Accepted Date: 2 August 2016

Please cite this article as: Ergönül Ö, Aydin M, Azap A, Başaran S, Tekin S, Kaya Ş, Gülsün S, Yörük G, Kurşun E, Yeşilkaya A, Şimşek F, Yılmaz E, Bilgin H, Hatipoğlu Ç, Cabadak H, Tezer Y, Togan T, Karaoğlan İ, İnan A, Engin A, Alışkan HE, Yavuz SŞ, Erdinç Ş, Mulazimoglu L, Azap Ö, Can F, Akalın H, Timurkaynak F, Turkish Society of Clinical Microbiology and Infectious Diseases, Health care related infections study group, Turkey, Healthcare-associated Gram-negative Bloodstream Infections: Antibiotic Resistance and Predictors of Mortality, Journal of Hospital Infection (2016), doi: 10.1016/ j.jhin.2016.08.012. 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.

ACCEPTED MANUSCRIPT Healthcare-associated Gram-negative Bloodstream Infections: Antibiotic Resistance and Predictors of Mortality

1

2

3

4

1

5

5

Önder Ergönül , Mehtap Aydin , Alpay Azap , Seniha Başaran , Süda Tekin , Şafak Kaya , Serda Gülsün , Gülşen 6

7

8

9

10

11

12

13

13

14

3

11

15

RI PT

Yörük , Ebru Kurşun , Ayşegül Yeşilkaya , Funda Şimşek , Emel Yılmaz , Hüseyin Bilgin , Çiğdem Hatipoğlu , 16

17

Hatice Cabadak , Yasemin Tezer , Turhan Togan , İlkay Karaoğlan , Asuman İnan , Aynur Engin , Hikmet 7

10

7

1

Eda Alışkan , Serap Şimşek Yavuz , Şebnem Erdinç , Lütfiye Mulazimoglu , Özlem Azap , Füsun Can , Halis 10

2

Akalın , Funda Timurkaynak , Turkish Society of Clinical Microbiology and Infectious Diseases, Health care re-

M AN U

SC

lated infections study group, Turkey.

Koç Üniversitesi Tıp Fakültesi

2

Başkent Üniversitesi Tıp Fakültesi, İstanbul Hastanesi

3

Ankara Üniversitesi Tıp Fakültesi

4

İstanbul Üniversitesi İstanbul Tıp Fakültesi

5

Diyarbakır Eğitim ve Araştırma Hastanesi

6

İstanbul Eğitim ve Araştırma Hastanesi

7

Başkent Üniversitesi Tıp Fakültesi, Adana Hastanesi

8

Başkent Üniversitesi Tıp Fakültesi, Ankara Hastanesi

9

Okmeydanı Eğitim ve Araştırma Hastanesi

EP

TE D

1

Uludağ Üniversitesi Tıp Fakültesi

11

Marmara Üniversitesi Tıp Fakültesi

12

Ankara Eğitim ve Araştırma Hastanesi

13

Ankara Yüksek İhtisas Hastanesi

14

Başkent Üniversitesi Tıp Fakültesi, Konya Hastanesi

15

Gaziantep Üniversitesi Tıp Fakültesi

16

Haydarpaşa Numune Eğitim ve Araştırma Hastanesi

17

Cumhuriyet Üniversitesi Tıp Fakültesi

1

AC C

10

ACCEPTED MANUSCRIPT

Abstract We described the prevalence of antibiotic resistance and predictors of fatality for healthcareassociated (HA) Gram-negative bloodstream infections (GN-BSI). In total 831 cases of HA GN-BSI from

RI PT

17 intensive care units in different centres in Turkey were included; the all-cause fatality rate was 44%. Carbapenem resistance in Klebsiella pneumoniae was 38%, and the colistin resistance rate was 6%. Multivariate analysis showed that age >70 (OR:2, CI: 1.22-3.51), having a central venous catheter

SC

(OR:2.1, CI:1.09-4.07), ventilator associated pneumonia (OR:1.9, CI:1.1-3.16), carbapenem resistance (OR:1.8, CI:1.11-2.95), and APACHE II score (OR:1.1, CI:1.07-1.13) were significantly associated with

M AN U

fatality.

Introduction

Healthcare associated infections (HAI) are a major problem, and are associated with signifi-

TE D

cant morbidity and mortality. Amongst HAI, bloodstream infections (BSI) are associated with the highest mortality rates; amongst BSI, mortality is greater in patients with Gram-negative BSI (GN-BSI) than in those with Gram-positive BSI.1 Multidrug-resistant GN-BSI are associated with a particularly

EP

high mortality rate; treatment options for these infections can be extremely limited, and they are becoming an increasing important cause of HAI globally.2 Of particular concern is the rapid emer-

AC C

gence and spread of carbapenem-resistant Gram-negative bacteria, because these bacteria are frequently resistant to many other antibiotic classes.3,4 Carbapenem resistance rates of 87% among Acinetobacter spp. and 34% among Pseudomonas spp. were reported by national nosocomial infections surveillance network of Turkey (UHESA) in 2013,5 but no clinical information was provided in this report. In our multicentre national study, we collected clinical and microbiological data in order to provide a clinically useful picture of the state of HA GN-BSI in Turkey.

Methods

2

ACCEPTED MANUSCRIPT Study population This was a retrospective multicentre study that included 17 tertiary care centres from seven different geographic regions of Turkey in 2013. All the centres actively applied standard infection control and prevention practices, and routinely reported their HAIs to Public Health Institute of Tur-

RI PT

key. The most commonly used empiric antibiotics in these centres were piperacillin-tazobactam and carbapenems (68% of all treatments); in all centres, antibiotic treatment was reviewed and adjusted where appropriate, based on positive culture results.

SC

All patients with HA GN-BSI in these centres during 2013 were included in the study. A cases was defined as any patient with ≥1 positive blood culture for GN bacteria presenting after 48 hours

M AN U

of hospitalization between 01/01/2013 and 31/12/2013. The diagnosis of HAI was based on the Centers for Disease Control and Prevention (CDC) criteria .6 Each centre was requested to submit the following information: demographic data; the source of BSI; causative microorganisms and their antimicrobial susceptibilities; choice of empiric antibiotic therapy; details of any operations in the pre-

TE D

ceding one month; details of intensive care; time to first positive blood culture after admission; mean CRP and/or procalcitonin concentrations and leukocyte counts; mean days for starting appropriate antibiotic; fatality (any cause); and APACHE II (Acute Physiology and Chronic Health Evaluation)

EP

scores. Data were collected using a structured electronic form; each case was validated by cross-

AC C

checking with the routine monthly reports of the infection control teams at each centre.

Microbiologic studies

Only clinically significant (as defined according to CDC criteria) blood culture isolates were included in the study. Identification of the organisms was conducted using whichever automated system was routinely used at each centre (VITEK 2, Biomerieux, Marcy l'Etoile, France; Phoenix, BD, Franklin Lakes, NJ, USA). Antibiotic susceptibilities was tested by disc diffusion or minimum inhibitory concentration tests according to the Clinical and Laboratory Standards Institute (CLSI) guideline and resistance was defined according to CLSI criteria.7 Antibiotics that were tested included carbapenems

3

ACCEPTED MANUSCRIPT (ertapenem, meropenem, and imipenem); third and fourth generation cephalosporins (ceftriaxone, ceftazidime and cefepime); fluoroquinolones (ciprofloxacin); aminoglycosides (amikacin, gentamicin); and and polymixin (colistin).

RI PT

Statistical analysis In statistical analysis, Student’s t-test was used for continous variables and Chi square test was used for the categorical variables. A multivariate analysis was performed for the predictors of all-cause

SC

fatality. STATA (USA, version 11) was used, and statistical significance was set as <0.05. Institutional Review Board of Koç University approved the study with the reference number of

M AN U

2015.294.IRB2.099.

Results

In total 831 cases of HA-GNB were identified, of which 371 (44%) were fatal. In univariate analysis,

TE D

the mean age and mean leukocyte count of fatal cases were higher. Female gender, ventilatorassociated pneumonia, central venous catheter use, antibiotic resistance, and being in intensive care unit were all more common among patients who died (Table 1). Amongst antibiotic resistances, re-

EP

sistances to carbapenems, fluoroquinolones, third generation cephalosporins and aminoglycosides were more common among fatal cases (Table 1).

AC C

Acinetobacter baumannii (256 cases; 31%), Klebsiella pneumoniae (221; 27%) and 202 Escherichia coli (202; 24%), were much the commonest species causing HA GN-BSI. Other species included Pseudomonas aeruginosa (74; 9%), Enterobacter cloacae (32; 4%), Stenotrophomonas maltophilia (15; 1.8%), Serratia marcescens (12; 1.5%), Proteus mirabilis (7; 1%), Burkholderia cepacia (5; 1%), Morganella morganii (2; 0.3%), Edwardsiella tarda (1; 0.1%), Pasteurella multocida (1; 0.1%). The highest antibiotic resistance rate was found in A. baumannii (Table 2). 94% of isolates were carbapenemresistant, 94% were fluoroquinolone-resistant , 97% were resistant to third generation cephalospor-

4

ACCEPTED MANUSCRIPT ins, 73% to aminoglycosides and 6% to colistin. The highest rate of all-cause fatality was detected among the patients with A. baumannii bacteremia (62% vs 38%, p<0.001)

In multivariate analysis, age >70, gender, resistance to antibiotics (aminoglycosides, fluoroquin-

RI PT

olones, 3rd generation cephalosporins, carbapenems, colistin), ventilator associated pneumonia and central catheter related infections were included as the independent variables. Age >70 (OR:2, CI: 1.22-3.51, p=0.006), having central venous catheter (OR:2.1, CI:1.09-4.07, p=0.025), ventilator asso-

SC

ciated pneumonia (OR:1.9, CI:1.1-3.16, p=0.02), carbapenem resistance (OR:1.8, CI:1.11-2.95,

p=0.016), and APACHE II score (OR:1.1, CI:1.07-1.13, p<0.001) were found to be significantly associ-

M AN U

ated with all cause fatality (Table 3). The logistic regression model predicted the fatality with the sensitivity of 82% (Area under ROC was 82%).

Discussion

TE D

We present a high case fatality rate (CFR) of 44% among the patients that was reported HA GN-BSI. However, the rate is comparable to that reported by Kaye et al (49%) in United States.8 Our study group included all the patients in 2013 from 17 centres of different geographic regions in Turkey, and

EP

we believe is therefore representative of the national picture. The majority of primary diagnoses of our patients on admission to our units were solid organ tumours followed by hematologic malignan-

AC C

cies, neurologic, pulmonary, cardiovascular diseases, trauma, and renal diseases, a picture that is also similar to that described by Kaye et al.8 In another recent study, coronary artery disease, immunosuppression, and inadequate empiric antimicrobial therapy were found to be independently associated with increased mortality.9

In our study, age >70 was found to predict fatality. The mortality rate in the elderly undergoing invasive procedures, who have more severe acute illness or who have multiple comorbidities may be even higher.8 In our univariate analysis the rate of fatality among female gender was found to be

5

ACCEPTED MANUSCRIPT higher (Table 1), but in multivariate analysis no significant effect was detected. Likewise, previous studies have not shown an association between fatality and gender in HA BSI.9

We found ventilator associated pneumonia and central venous catheter (CVC) use were more com-

RI PT

mon among patients who died, as has been reported previously.10 These are risk factors that should be amenable to infection control intervention. We also found that carbapenem resistance was independently associated with fatality. Previous studies from the USA and Israel reported similar find-

SC

ings.11,12 In an attempt to combat antibiotic resistance, a nationwide antibiotic restriction policy has been in place since 2003 in Turkey.13 However, there are still many opportunities to antibiotic stew-

M AN U

ardship strategies in Turkey. In previous studies, delay in starting appropriate antimicrobial therapy was found to be associated with poor outcome,14 but in our study time to start appropriate antibiotic therapy was not a significant risk factor for mortality. Relatedly, the time to first positive blood cul-

TE D

ture after admission was not correlated with fatality.

We restricted our study to GN-BSI because of its increasing importance. The limited options for antibiotic treatment in carbapenem-resistant A. baumannii are a particularly serious problem in our

EP

health care facilities. The CFR of A. baumannii BSI infection was 62% in our study, compared with 34% in a USA study.15 Rates of carbapenem resistance among Acinetobacter spp. collected from inva-

AC C

sive samples in recent studies has ranged from 0% (Finland and Norway) to 90.6% (Greece). Similar variations have been seen with respect to fluoroquinolone resistance, ranging from 0 % in Norway to 95.0 % in Greece, and aminoglycoside resistance, ranging from 0 % in Finland to 92.1% in Croatia. Perhaps unsurprisingly, resistance rates in our study were similar to those in Greece .16 We also found a higher rate of colistin resistance (6%) in A. bacteremia than in previous studies from other countries. Colisitin resistance rates of between 0.9% and 3.3% were reported by the SENTRY Antimicrobial Surveillance from the USA, Europe, Latin America and the Asia-Pacific region.17 However, our

6

ACCEPTED MANUSCRIPT results accord well with the colistin resistance rate in A. baumannii (6%) in the 2014 report of the National Nosocomial Infections Surveillance Network of Turkey (5.8%).5

In our study, K. pneumoniae was the second commonest cause of HA GN-BSI, and was also associat-

RI PT

ed with a high crude mortality rate of 44%. Of the 231 isolates, 40% were resistant to carbapenems, 60% to fluoroquinolones, 72% to 3rd generation cephalosporins and 6% to colistin. A high rate of carbapenem resistance has been reported in Greece (59.4%); we also found a lower rate of fluoro-

SC

quinolone resistance , but the rate of resistance to cephalosporins was similar.16 The aminoglycoside resistance rate in our study was 24.5%, which was similar to a European report; the rate of colistiin

M AN U

resistance in our study at 6% was lower than the rate of 8.8% in the European study.16 Although carbapenem resistance in E .coli was lower at 6.4%, this rate is much higher than the rate of <0.1% reported in European countries.16 Empiric use of piperacillin-tazobactam or carbapenem in our hospitals was very high (68%). Whilst this may be seen as justified in a country with high antibiotic re-

TE D

sistance rates, our observation that empiric antibiotic therapy did not affect survival, suggests that it would be safe to use less broad-spectrum antibiotics in at least some patients.

EP

Conclusion

In multivariate analysis for prediction of fatality, ventilator and central venous catheter associated

AC C

infections were found to be two important areas to be improved in terms of infection control. Carbapenem resistance as a significant independent risk factor for mortality all antibiotic prescribers, infection control practitioners and public health services on the urgent need to control these bacteria. The first steps in controlling the emergence and spread of carbapenen-resistant Gram-negative bacteria are within all healthcare workers’ jurisdiction, namely good infection prevention and control and effective antibiotic stewardship.

References

7

ACCEPTED MANUSCRIPT

6

7

8

9

10

11

12

13

14

8

RI PT

5

SC

4

M AN U

4

TE D

3

EP

2

Vincent JL. Nosocomial infections in adult intensive-care units. Lancet 2003;361:2068-2077. Guh AY, Bulens SN, Mu Y, et al. Epidemiology of carbapenem-resistant Enterobacteriaceae in 7 US communities, 2012-2013. JAMA 2015;314:14791487. Canton R, Akova M, Carmeli Y, et al. Rapid evolution and spread of carbapenemases among Enterobacteriaceae in Europe. Clin Microbiol Infect 2012;18:413-431. Gray JW, Mahida N. How do you solve a problem like multidrug-resistant Gramnegative bacteria? J Hosp Infect 2016;92:1-2. National hospital infections surveillance network, 2013 report. Ankara: Public Health Institute of Turkey 2015. Horan TC, Andrus M, Dudeck MA. CDC/NHSN surveillance definition of health care-associated infection and criteria for specific types of infections in the acute care setting. Am J Infect Control 2008;36:309-332. Performance standards for antimicrobial susceptibility testing: Twentieth informational supplement. CLSI document M100-S20. Wayne: Clinical and Laboratory Standards Institute 2010. Kaye KS, Marchaim D, Chen TY, et al. Effect of nosocomial bloodstream infections on mortality, length of stay, and hospital costs in older adults. J Am Geriatr Soc 2014;62:306-311. Sligl WI, Dragan T, Smith SW. Nosocomial gram-negative bacteremia in intensive care: Epidemiology, antimicrobial susceptibilities, and outcomes. Int J Infect Dis 2015;37:129-134. Martin-Loeches I, Torres A, Rinaudo M, et al. Resistance patterns and outcomes in intensive care unit (ICU)-acquired pneumonia. Validation of European Centre for Disease Prevention and Control (ecdc) and the Centers for Disease Control and Prevention (CDC) classification of multidrug resistant organisms. J Infect 2015;70:213-222. Patel G, Huprikar S, Factor SH, Jenkins SG, Calfee DP. Outcomes of carbapenemresistant Klebsiella pneumoniae infection and the impact of antimicrobial and adjunctive therapies. Infect Control Hosp Epidemiol 2008;29:1099-1106. Hussein K, Raz-Pasteur A, Finkelstein R, et al. Impact of carbapenem resistance on the outcome of patients' hospital-acquired bacteraemia caused by Klebsiella pneumoniae. J Hosp Infect 2013;83:307-313. Altunsoy A, Aypak C, Azap A, Ergonul O, Balik I. The impact of a nationwide antibiotic restriction program on antibiotic usage and resistance against nosocomial pathogens in turkey. Int J Med Sci 2011;8:339-344. Ferrer R, Martin-Loeches I, Phillips G, et al. Empiric antibiotic treatment reduces mortality in severe sepsis and septic shock from the first hour: Results from a guideline-based performance improvement program. Crit Care Med 2014;42:1749-1755. 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;39:309-317.

AC C

1

ACCEPTED MANUSCRIPT 15

AC C

EP

TE D

M AN U

SC

RI PT

16

European Centre for Disease Prevention and Control. Antimicrobial resistance surveillance in Europe 2013. Annual Report of the European Antimicrobial Resistance Surveillance Network (EARS-Net) Stockholm: ECDC 2014. Cai Y, Chai D, Wang R, Liang B, Bai N. Colistin resistance of Acinetobacter baumannii: Clinical reports, mechanisms and antimicrobial strategies. J Antimicrob Chemother 2012;67:1607-1615.

9

ACCEPTED MANUSCRIPT Table 1. Univariate analysis of risk factors for mortality in patients with healthcareassociated Gram-negative bloodstream infection Patients who died n=371

Patients who survived n=460

P value

Mean age (sd)

60 (19)

54 (18)

<0.001

Female gender

185 (50)

192 (42)

0.019

Primary diagnosis

RI PT

Variable

Solid organ tumors

72 (20)

Hematological malignancies

48 (13)

Neurological diseases

53 (14)

61 (14)

Trauma

32 (9)

36 (8)

Cardiovascular diseases

41 (11)

37 (8)

56 (15)

41 (9)

7 (2)

22 (5)

287 (77)

213 (46)

<0.001

132 (32)

172 (39)

0.567

Antibiotic use within last three months

228 (68)

277 (65)

0.5

Primary bacteremia

216 (59)

266 (58)

0.902

Ventilator-associated pneumonia

155 (43)

100 (22)

<0.001

Central venous catheter

307 (84)

262 (58)

<0.001

Time to first positive blood culture after admission (mean days)

20

23

0.091

Mean CRP level

132

126

0.409

Mean leukocyte count

13117

11317

0.04

Mean procalcitonin level

11

10

0.796

Mean days for starting appropriate antibiotic

2

2

0.94

Carbapenem resistance

236 (62)

143 (37)

<0.001

Fluoroquinolone resistance

187 (78)

172 (62)

<0.001

Third generation cephalosporin

238 (89)

211 (71)

<0.001

Renal diseases On intensive care unit

99 (22)

SC

AC C

EP

TE D

Operation within one month

M AN U

Pulmonary diseases

79 (17)

Antibiotic resistance

ACCEPTED MANUSCRIPT Aminoglycoside resistance

115 (49)

88 (33)

<0.001

Table 2. Antibiotic Resistance Rates in healthcare-associated Gram-negative bloodstream infections Carbapenem n (%)

Fluoroquinolon Third generation Aminoglycosides Colistin e Cephalosporins

Acinetobacter baumannii

239 (94)

240 (94)

247 (97)

187 (73)

15 (6)

Klebsiella pneumoniae

88 (40)

130 (60)

159 (72)

56 (25)

14 (6)

E.coli

13 (6.4)

128 (63)

143 (71)

47 (23)

0

36 (49)

37 (51)

19 (26)

1 (1)

16 (53)

5 (16)

0

5 (16)

6 (19)

M AN U

Enterobacter cloacae

SC

Pseudomonas aeruginosa 32 (43)

RI PT

Bacteria

Table 3. Multivariate analysis of risk factors for mortality in patients with healthcareassociated Gram-negative bloodstream infection Odds Ratio

Confidence interval

p

2

1.22-3.51

0.006

2.1

1.09-4.07

0.025

Ventilator associated pneumonia

1.9

1.1-3.16

0.02

Carbapenem resistance

1.8

1.11-2.95

0.016

1.1

1.07-1.13

<0.001

TE D

Variable Age >70

AC C

EP

Central venous catheter

APACHE II score