- Email: [email protected]
The impact of carbapenemase-producing Enterobacteriaceae colonization on infection risk after liver transplantation: a prospective observational cohort study
Clinical Microbiology and Infection xxx (xxxx) xxx
Contents lists available at ScienceDirect
Clinical Microbiology and Infection journal homepage: www.clinicalmicrobiologyandinfection.com
Original article
The impact of carbapenemase-producing Enterobacteriaceae colonization on infection risk after liver transplantation: a prospective observational cohort study M. Giannella 1, *, M. Bartoletti 1, C. Campoli 1, M. Rinaldi 1, S. Coladonato 1, R. Pascale 1, S. Tedeschi 1, S. Ambretti 2, F. Cristini 1, F. Tumietto 1, A. Siniscalchi 3, V. Bertuzzo 4, M.C. Morelli 5, M. Cescon 4, A.D. Pinna 4, R. Lewis 1, P. Viale 1 1)
Infectious Diseases Unit, Department of Medical and Surgical Sciences, Policlinico Sant'Orsola Malpighi, University of Bologna, Bologna, Italy Operative Unit of Clinical Microbiology, Policlinico Sant'Orsola Malpighi, University of Bologna, Bologna, Italy 3) Anaesthesia Unit, Department of Medical and Surgical Sciences, Policlinico Sant'Orsola Malpighi, University of Bologna, Bologna, Italy 4) Liver and Multiorgan Transplant Unit, Department of Medical and Surgical Sciences, Policlinico Sant'Orsola Malpighi, University of Bologna, Bologna, Italy 5) Internal Medicine Unit for the Treatment of Severe Organ Failure, Department of Medical and Surgical Sciences, Policlinico Sant'Orsola Malpighi, University of Bologna, Bologna, Italy 2)
a r t i c l e i n f o
a b s t r a c t
Article history: Received 3 February 2019 Received in revised form 29 March 2019 Accepted 12 April 2019 Available online xxx
Objective: To investigate the impact of colonization with carbapenemase-producing Enterobacteriaceae (CPE) on the CPE infection risk after liver transplantation (LT). Methods: Prospective cohort study of all adult patients undergoing LT at our centre over an 8-year period (2010e2017). Individuals were screened for CPE colonization by rectal swabs at inclusion onto the waiting list, immediately before LT and weekly after LT until hospital discharge. Asymptomatic carriers did not receive decolonization, anti-CPE prophylaxis or pre-emptive antibiotic therapy. Participants were followed up for 1 year after LT. Results: We analysed 553 individuals who underwent a first LT, 38 were colonized with CPE at LT and 104 acquired colonization after LT. CPE colonization rates at LT and acquired after LT increased significantly over the study period: incidence rate ratios (IRR) 1.21 (95% CI 1.05e1.39) and 1.17 (95% CI 1.07e1.27), respectively. Overall, 57 patients developed CPE infection within a median of 31 (interquartile range 11 e115) days after LT, with an incidence of 3.05 cases per 10 000 LT-recipient-days and a non-significant increase over the study period (IRR 1.11, 95% CI 0.98e1.26). In multivariable analysis, CPE colonization at LT (hazard ratio (HR) 18.50, 95% CI 6.76e50.54) and CPE colonization acquired after LT (HR 16.89, 95% CI 6.95e41.00) were the strongest risk factors for CPE infection, along with combined transplant (HR 2.60, 95% CI 1.20e5.59), higher Model for End-Stage Liver Disease at the time of LT (HR 1.03, 95% CI 1.00e1.07), prolonged mechanical ventilation (HR 2.63, 95% CI 1.48e4.67), re-intervention (HR 2.16, 95% CI 1.21 e3.84) and rejection (HR 2.81, 95% CI 1.52e5.21). Conclusions: CPE colonization at LT or acquired after LT were the strongest predictors of CPE infection. Prevention strategies focused on LT candidates and recipients colonized with CPE should be investigated. M. Giannella, Clin Microbiol Infect 2019;▪:1 © 2019 European Society of Clinical Microbiology and Infectious Diseases. Published by Elsevier Ltd. All rights reserved.
Editor: A. Kalil Keywords: Carbapenem-resistant Carbapenemase-producing Enterobacteriaceae Colonization Infection risk Liver transplantation
Introduction
* Corresponding author. M. Giannella, Infectious Disease Unit, Policlinico Sant'Orsola-Malpighi, Via Massarenti 11, 40138, Bologna, Italy. E-mail address: [email protected] (M. Giannella).
Solid organ transplant (SOT) recipients are at higher risk than non-SOT recipients for colonization and/or infection with carbapenemase-producing Enterobacteriaceae (CPE) [1]. Higher rates of CPE acquisition occur in lung and liver transplant recipients compared with other types of SOT, with liver transplantation (LT)
https://doi.org/10.1016/j.cmi.2019.04.014 1198-743X/© 2019 European Society of Clinical Microbiology and Infectious Diseases. Published by Elsevier Ltd. All rights reserved.
Please cite this article as: Giannella M et al., The impact of carbapenemase-producing Enterobacteriaceae colonization on infection risk after liver transplantation: a prospective observational cohort study, Clinical Microbiology and Infection, https://doi.org/10.1016/j.cmi.2019.04.014
2
M. Giannella et al. / Clinical Microbiology and Infection xxx (xxxx) xxx
accounting for most episodes [2]. Infection with CPE after SOT is associated with a mortality risk up to ten times higher than that with carbapenem-susceptible Enterobacteriaceae [3]. The knowledge of risk factors for CPE infection is key to designing preventive strategies. In LT recipients, CPE infection is usually preceded by CPE rectal carriage. Indeed, CPE colonization has been described as an important predictor of CPE infection [4e6]. However, colonization alone does not seem to be sufficient for developing infection [5]. In addition, data concerning the risk of CPE infection according to the moment of rectal carriage acquisition, before or after LT, are controversial [4e6]. We performed a large prospective cohort study of individuals undergoing LT to investigate risk factors for CPE infection. The main goal of this study was to define the impact of CPE colonization status before or after LT. Material and methods Study design and participants We performed a prospective observational cohort study of all consecutive adults (18 years) undergoing LT at our hospital, from June 2010 to December 2017. For the purpose of this study, patients were followed up for 12 months after LT by hospital and outpatient clinical assessments. Only the first episode of CPE infection was considered for the analysis. The study was approved by our Institutional Review Board, and patient informed consent was obtained. Setting Sant'Orsola-Malpighi Hospital is a 1420-bed tertiary-care University Hospital in Bologna (northern Italy) with an average of 72 000 admissions per year. The hospital has an active LT programme, performing an average of 90 procedures per year. During the study period, all LT candidates were screened with rectal swab (RS) for CPE carriage at the time of inclusion in the waiting list, in case of hospital (re-)admission, and at the time of transplant. RS was repeated once weekly after LT until hospital discharge. RS was repeated with the same frequency also in the case of hospital re-admission. Isolation and contact precautions were activated for CPE carriers, but no pharmacological interventions, including selective intestinal decontamination, CPEtargeted surgical prophylaxis, or antimicrobial treatment, were administered to asymptomatic patients. Standard perioperative prophylaxis consisted of short-course ampicillin-sulbactam. All suspected infections were managed in collaboration with a dedicated team of infectious disease specialists. Additional information on the transplant management at the study centre is available in the Supplementary material (Data S1). Data collection and definitions Variables were collected on standardized case report forms, which have been collected electronically since 2014. The end-point variable, CPE infection, was assessed according to CDC criteria [7], from LT to 12 months after LT. Exposure variables were evaluated from LT to the first episode of CPE infection, death or 12 months after LT. The main exposure variable, CPE colonization, was defined as the isolation of CPE from RS in the absence of symptoms and signs of infection. Every sample other than blood cultures or sterile fluids (including urine, respiratory samples, surgical site and other skin cultures) positive for CPE in asymptomatic patients was considered as colonization. Individuals found to be colonized with CPE by
cultures performed at the inclusion in waiting list, during the stay in waiting list, or at the time of LT were considered as CPE carriers at LT, and those found to be colonized afterwards were considered as having acquired CPE carriage after LT. The other exposure variables included demographic data (age and sex), and co-morbidities according to Charlson index, aetiology of liver disease, Model for End-Stage Liver Disease (MELD) at inclusion in the waiting list, and at LT. Pre-transplant variables (within 90 days) included: intensive care unit admission, grade 3 and/or refractory ascites [8], hepatorenal syndrome [9], gastrointestinal bleeding, hepatic encephalopathy [10], any infection, infection with multidrug-resistant bacteria, candidaemia and Clostridium difficile infection. Graft variables included donor age and cold ischaemia time. Intraoperative variables were bleeding with the need for 40 units of cellular blood products transfusion, and type of biliary anastomosis. Post-LT complications included reintervention, acute kidney injury according to RIFLE criteria [11], renal replacement therapy, prolonged (48 hours) mechanical ventilation, primary graft non-function, biopsy-proven rejection, re-transplantation, histological recurrence of hepatitis C virus infection and cytomegalovirus infection defined as >100 000 copies/mL. Microbiology Surveillance RS were collected using E-swab (Copan, Brescia, Italy), inoculated onto a chromogenic medium (Kima Meus, Arzegrande, Italy) containing a carbapenem antibiotic as selective agent and incubated at 35e37 C for 18e24 hours. Clinical isolates were processed following standard procedures. Species identification and routine antimicrobial susceptibility testing were performed rieux, using Vitek-MS MALDI-TOF and VITEK2 systems (bioMe Marcy l'Etoile, France). Results for MIC were interpreted according to EUCAST breakpoints. For all Enterobacterales isolates with MIC values higher than the EUCAST cut-off for at least one of the tested carbapenems (ertapenem, imipenem, meropenem), carbapenemase production was evaluated. Different confirmation methods have been used from 2010 to 2017, including modified Hodge-test, disc diffusion synergy test with inhibitors (Rosco Diagnostika, Taastrup, Denmark), lateral flow immunoassays (Coris BioConcept, Gembloux, Belgium), the Neo-Rapid carb sCPEen test (Rosco Diagnostika) [12] and Xpert Carba-R molecular method (Cepheid, Sunnyvale, CA, USA) [13]. In addition, carbapenem MICs were rieux). confirmed by E-test (Biome Statistical analysis Incidence rates ±95% CI for CPE colonization at LT were calculated and plotted per 10 000 patient-days using only new episodes observed during the study period by the overall number of patientdays computed from the day of inclusion in waiting list to LT. To calculate the incidence of CPE colonization and infection after LT, we divided the number of new episodes by the overall number of patient-days in the study cohort computed from the day of transplantation until day 365 or patient death. Poisson regression was used to analyse trends in the incidence of pre- and post-transplant CPE colonization as well as CPE infection per year adjusting for patient-days of exposure and reported as incidence rate ratio (IRR) (±95% CI) over the 8-year period. The impact of pre-transplant MELD score on the rate of CPE colonization was analysed using classification and regression tree analysis. Martingale residuals of a Cox model were used to calculate c2 values for all possible MELD score cut points, resulting in significantly different rates of pre-transplant CPE colonization.
Please cite this article as: Giannella M et al., The impact of carbapenemase-producing Enterobacteriaceae colonization on infection risk after liver transplantation: a prospective observational cohort study, Clinical Microbiology and Infection, https://doi.org/10.1016/j.cmi.2019.04.014
M. Giannella et al. / Clinical Microbiology and Infection xxx (xxxx) xxx Table 1 Characteristics of study cohort
(a)
5.6 ± 1.9 6 (1.1) 274 (49.5) 127 (23) 20 (3.6) 237 (42.9) 27 (4.9) 16 (11e23) 19 (12e26) 179 (36e450) 266 (48.1) 114 (20.6) 93 (16.8) 31 (5.6) 68 (12.3) 120 (21.7) 34 (6.1) 1 (0.2) 7 (1.3)
Rate (per 10 000)
52.8 ± 10.7 394 (71.2)
12 10 8 6 4 2 0 1
(b) 24
38 (6.9) 109 (19.7) 553 (100) 63 (46e75) 7 (6e7.3)
2
3
4
Year
5
6
7
8
IRR 1.16 (95% CI 1.01–1.23), p 0.03
22 20 18 16 14 12 10 8 6 4 2 0 1
2
3
4
5
6
7
8
7
8
Year 93 (16.8) 129 (23.3) 12 (2.2) 33 (6) 143 (25.9) 68 (12.3) 102 (18.4) 89 (16.1) 26 (4.7) 71 (12.8) 42 (7.6) 21 (3.8) 17 (11e31) 57 (10.3)
Abbreviations: CMV, cytomegalovirus; CPE, carbapenemase-producing Enterobacteriaceae; HCV, hepatitis C virus; HIV, human immunodeficiency virus; IQR, interquartile range; LT, liver transplant; MELD, Model for End-stage Liver Disease; PGNF, primary graft non-function; SD, standard deviation. a Some patients presented more than one underlying liver disease. b Episodes needing 40 units of cellular blood products transfusion.
Patients who developed a CPE infection within 12 months after LT were compared with those who did not develop CPE infection during that period. Categorical variables were compared using Pearson c2 or Fisher's exact test when appropriate. Continuous variables were compared using Student's t or ManneWhitney U test according to their distribution. All the variables with a p value 0.1 at the univariate analysis were entered into a multivariable Cox regression model after verifying for proportional hazards (Fig. S1) and non-collinearity (see Supplementary material: Data S1, Table S1 and Fig. S2). Patients were considered from the day of LT up to 12 months, or before in case of CPE infection or death. We also examined the effects of these covariates in a multivariable Fine and Gray model considering death before CPE
(c) 16
IRR 1.11 (95% CI 0.99–1.25), p 0.09
14 Rate (per 10 000)
Demographic data Age (years), mean ± SD Sex, male, n (%) Co-morbidities Charlson index, mean ± SD HIV co-infection, n (%) Underlying liver disease, n (%)a Viral disease Alcohol Fulminant liver failure Hepatocellular carcinoma Combined transplant MELD at inclusion in waiting list, median (IQR) MELD at LT, median (IQR) Time on waiting list (days), median (IQR) Complications within 90 days before LT, n (%) Grade 3 and/or refractory ascites Encephalopathy Hepato-renal syndrome Gastrointestinal bleeding Admission to intensive care unit Any infection Infection with multidrug-resistant bacteria and/or candidaemia Clostridium difficile infection CPE infection pre-LT Rectal swab positive for CPE, n (%) Colonization at LT Colonization acquired after LT Graft characteristics Deceased donor, n (%) Donor age (years), median (IQR) Cold ischaemia time (hours), median (IQR) Intraoperative variables, n (%) Choledochojejunostomy Bleedingb Post-operative complications, n (%) CMV infection (>100 000 copies/mL) HCV relapse Acute kidney injury Renal replacement therapy Re-intervention Mechanical ventilation >48 hours PGNF Rejection Biopsy proven rejection Re-transplantation Outcome Length of hospital stay after LT (days), median (IQR) All-cause 1-year mortality, n (%)
IRR 1.21 (95% CI 1.05–1.40); p 0.008
14
Rate (per 10 000)
Total (n ¼ 553)
16
3
12 10 8 6 4 2 0 1
2
3
4
5
6
Year Fig. 1. Trends in carbapenemase-producing Enterobacteriaceae (CPE) colonization and infections over the study period. (a) Pre-transplant CPE colonization; (b) posttransplant CPE colonization; (c) post-transplant CPE infection. Points show the estimated incidence per year with 95% CI indicated by shading. Changes in the incidence rate per year were analysed using Poisson regression to derive the incidence rate ratio (IRR).
infection as a competing risk. All analyses were performed with STATA 13 (Stata Corp, College Station, TX, USA).
Results During the study period, 553 individuals underwent LT. The general characteristics of the study cohort are shown in Table 1. Overall, CPE colonization was detected in 147 individuals (26.6%). In all cases, the first sample showing CPE carriage was RS,
Please cite this article as: Giannella M et al., The impact of carbapenemase-producing Enterobacteriaceae colonization on infection risk after liver transplantation: a prospective observational cohort study, Clinical Microbiology and Infection, https://doi.org/10.1016/j.cmi.2019.04.014
4
M. Giannella et al. / Clinical Microbiology and Infection xxx (xxxx) xxx
(a)
(b)
6–33
349
Col. RHR 13
Pre-transplant CPE colonization
N 6–19
.42
MELD score
20–33
170
18
2.91
MELD score
1.00
MELD 6–19
0.75
MELD 20–33 MELD >33
0.50
Log rank p<0.0001
0.25
0.00 34–47
22
6 33.74
0
100
200
300
400
Days on transplant list Number at risk MELD 6–19 = 349 MELD 20–33 = 170 MELD >33 = 34
279 54 6
217 36 4
171 26 1
135 15 1
Fig. 2. Impact of pre-transplant Model for End-Stage Liver Disease (MELD) score at time of patient listing and rates of carbapenemase-producing Enterobacteriaceae (CPE) colonization. (a) Martingale residuals of a Cox regression model for newly-acquired CPE colonization were used to calculate c2 values for all possible MELD score cut-points using classification and regression tree (CART) analysis resulting in significantly different rates of pre-transplant CRE colonization. (b) KaplaneMeier failure curves for pre-transplant CRE colonization at CART-derived MELD score cut-offs.
CPE INFECTION 406 No CPE carriage
553 LT recipients
6 (1.5%) 48 (IQR 20–362) days 14 (36.8%)
38 CPE carriage at LT
13 (IQR 3–21.5) days
109 CPE carriage aer LT
56 (IQR 14–129) days
37 (33.9%)
Fig. 3. Carbapenemase-producing Enterobacteriaceae (CRE) infection rate and median time to CRE infection after liver transplant (LT) according with the colonization status before and after LT.
in some individuals CPE was subsequently isolated also from other sites, mainly urine. Among CPE-colonized individuals, 38 (25.8%) were found to be colonized at LT. The median time of carriage detection was 12 (interquartile range (IQR) 0.75e40) days (maximum time 309 days) before LT. The incidence of CPE colonization at LT was 1.88 cases per 10 000 candidate-days. The incidence rate of CPE colonization at LT increased significantly over the study period (IRR 1.21, 95% CI 1.05e1.39, p 0.008) (Fig. 1a). The risk of CPE colonization varied depending on the patient MELD score at the time of listing, classification and regression tree analysis showed that the risk of being colonized with CPE at LT increased significantly in individuals with a MELD score at listing >19 (Fig. 2). However, there was not a significant change in the number of individuals with MELD scores >19 over the study period (IRR 1.01, 95% CI 0.95e1.06, p 0.79). One hundred nine individuals (19.7%) acquired CPE colonization within a median of 15 (IQR 6e34) days after LT. The overall incidence of CPE colonization after LT was 5.84 cases per 10 000 LT recipient-days with a significant increase per year (IRR 1.17, 95% CI
1.07e1.27, p < 0.001) (Fig. 1b). Although a relationship between CPE colonization post-transplant and MELD at the time of transplant was detected (OR 1.02, 95% CI 1.00e1.04; p 0.03) no cut-off for high risk versus low risk was identified by classification and regression tree analysis. Overall, 57 (10.3%) individuals developed one or more episodes of CPE infection after LT. The infection rate according with the colonization status is shown in Fig. 3. The median time to the first CPE infection episode was 31 (IQR 11e115) days after transplant, varying from 13 (IQR 3e21.5) days in individuals colonized with CPE before LT, to 56 (IQR 14e129) days in those who acquired CPE colonization after LT. In this latter group, the median time from colonization acquisition to infection development was 20 (IQR 4e72) days. The overall incidence of CPE infection was 3.05 per 10 000 LT recipient-days, without a significant increase over the study period (IRR 1.11, 95% 0.98e1.26, p 0.08) (Fig 1c). Most individuals who developed CPE infection were bacteraemic (84.2%), the infection types/sources were lower respiratory tract infection (n ¼ 19, 33.3%), surgical site including superficial wound (n ¼ 1), biliary (n ¼ 5) and intra-abdominal tract (n ¼ 11), infections (n ¼ 17, 29.8%), undefined (n ¼ 14, 24.5%), central venous catheter (n ¼ 4, 7%) and urinary tract infection (n ¼ 3, 5.2%). Severe sepsis and septic shock were present at infection onset in 33.3% and 24.6% of cases, respectively. The mechanism of carbapenem resistance was determined in 32 out of 57 patients with infection showing Klebsiella pneumoniae carbapenemase production in all cases. A comparison of patients with and without CPE infection is shown in Table 2. Factors associated with CPE infection included alcoholic liver disease, median MELD at listing, grade 3 and/or refractory ascites within 90 days before LT, previous CPE infection, CPE carriage at LT, median MELD at LT, combined transplant, prolonged cold ischaemia time of the graft, CPE carriage acquisition after LT, cytomegalovirus infection, acute kidney injury, renal replacement therapy, re-intervention, prolonged mechanical ventilation, primary graft non-function, rejection and re-
Please cite this article as: Giannella M et al., The impact of carbapenemase-producing Enterobacteriaceae colonization on infection risk after liver transplantation: a prospective observational cohort study, Clinical Microbiology and Infection, https://doi.org/10.1016/j.cmi.2019.04.014
M. Giannella et al. / Clinical Microbiology and Infection xxx (xxxx) xxx
5
Table 2 Comparison of patients with and without CPE infection within 1 year after liver transplantation
Demographic data Age (years), mean ± SD Sex, male, n (%) Co-morbidities Charlson index, mean ± SD HIV co-infection, n (%) Underlying liver disease, n (%)a Viral disease Alcohol Fulminant liver failure Hepatocellular carcinoma Combined transplant MELD at inclusion in waiting list, median (IQR) MELD at LT, median (IQR) Complications within 90 days before LT, n (%) Grade 3 and/or refractory ascites Encephalopathy Hepato-renal syndrome Gastrointestinal bleeding Admission to intensive care unit Any infection Infection with multi-drug-resistant bacteria and/or candidaemia Clostridium difficile infection CRE infection pre-LT Rectal swab positive for CPE, n (%) Colonization at LT Colonization acquired after LT Graft characteristics Donor age (years), median (IQR) Cold ischaemia time (hours), median (IQR) Intraoperative variables, n (%) Choledochojejunostomy Bleedingb Post-operative complications, n (%) CMV infection (>100 000 copies/mL) HCV relapse Acute kidney injury Renal replacement therapy Re-intervention Mechanical ventilation >48 hours PGNF Rejection Biopsy proven rejection Re-transplantation Outcome Length of hospital stay after LT (days), median (IQR) All-cause 1-year mortality, n (%)
Patients with CPE infection, (n ¼ 57)
Patients without CPE infection (n ¼ 496)
p
53.6 ± 10.5 44 (77.2)
52.7 ± 10.8 350 (70.6)
0.53 0.35
5.7 ± 1.8 0
5.6 ± 1.9 6 (1.2)
0.65 1
28 (49.1) 20 (35.1) 0 19 (33.3) 8 (14) 20 (14e25) 24 (15e30)
246 (49.6) 107 (21.6) 20 (4) 218 (44) 19 (3.8) 15 (10e22) 18 (12e26)
1 0.03 0.15 0.16 0.004 0.001 0.002
40 (70.2) 14 (24.6) 13 (22.8) 5 (8.8) 11 (19.3) 17 (29.8) 7 (12.3) 0 3 (5.3)
226 (45.6) 100 (20.2) 80 (16.1) 26 (5.2) 57 (11.5) 103 (20.8) 27 (5.4) 1 (0.2) 4 (0.8)
<0.001 0.5 0.3 0.35 0.13 0.13 0.07 1 0.03
14 (24.6) 37 (64.9)
24 (4.8) 72 (14.5)
<0.001 <0.001
67 (46e77) 7 (6e7.3)
63 (46e75) 6.5 (6e7.2)
0.35 0.04
9 (15.8) 17 (29.8)
84 (16.9) 112 (22.9)
0.85 0.25
4 (7) 5 (8.8) 29 (50.9) 21 (36.8) 31 (54.4) 23 (40.4) 10 (17.5) 18 (31.6) 11 (19.3) 10 (17.5)
8 (1.6) 28 (5.6) 114 (23) 47 (9.5) 71 (14.3) 66 (13.3) 16 (3.2) 53 (10.7) 31 (6.3) 11 (2.2)
0.03 0.37 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 0.002 <0.001
46 (26e104) 24 (42.1)
16 (11e26) 33 (6.7)
<0.001 <0.001
Abbreviations: CMV, cytomegalovirus; CPE, carbapenemase-producing Enterobacteriaceae; HCV, hepatitis C virus; HIV, human immunodeficiency virus; IQR, interquartile range; LT, liver transplant; MELD, Model for End-stage Liver Disease; PGNF, primary graft non-function; SD, standard deviation. a Some patients presented more than one underlying liver disease. b Episodes needing 40 units of cellular blood products transfusion.
transplantation. The all-cause 1-year mortality rates in individuals with and without CPE infection were 42.1% and 6.7% (p < 0.001), respectively. In addition, the median length of hospital stay after LT was significantly longer among individuals who developed CPE infection (46 versus 16 days, p < 0.001).
At multivariate Cox regression and competing risk analyses, the independent risk factors for CPE infection within 1 year after LT were: MELD at LT, combined transplant, colonization at LT, colonization acquired after LT, prolonged mechanical ventilation, reintervention and rejection (Table 3, Fig. 4).
Table 3 Multivariate analysis of risk factors for CPE infection within 1 year after liver transplantation Variable
HR (95% CI)
p
SHR (95% CI)
p
MELD score at transplant (per 1 point increase) Combined transplant CPE colonization at LT CPE colonization acquired after LT Prolonged mechanical ventilation Re-intervention Rejection
1.03 (1.00e1.07) 2.60 (1.20e5.59) 18.50 (6.76e50.54) 16.89 (6.95e41.00) 2.63 (1.48e4.67) 2.16 (1.21e3.84) 2.81 (1.52e5.21)
0.03 0.02 <0.001 <0.001 0.001 0.009 0.001
1.03 (1.00e1.08) 2.74 (1.08e6.96) 18.12 (6.98e47.01) 17.60 (7.36e42.08) 2.39 (1.32e4.35) 2.16 (1.18e3.94) 2.94 (1.59e5.45)
0.05 0.03 <0.001 <0.001 0.004 0.012 0.001
Abbreviations: CPE, carbapenemase-producing Enterobacteriaceae; HR, hazard ratio; LT, liver transplantation; MELD, Model for End-stage Liver Disease; SHR, sub-hazard ratio.
Please cite this article as: Giannella M et al., The impact of carbapenemase-producing Enterobacteriaceae colonization on infection risk after liver transplantation: a prospective observational cohort study, Clinical Microbiology and Infection, https://doi.org/10.1016/j.cmi.2019.04.014
6
M. Giannella et al. / Clinical Microbiology and Infection xxx (xxxx) xxx
Competing-risks regression .2
Colonization pre-OLT (SHR 18.12, 95% CI 6.98–47.01, p<0.0001) Cumulative Incidence
Colonization post-OLT(SHR 17.60, 95% CI 7.36–42.08, p<0.0001) .15
.1
.05
No colonization (reference) 0 0
100
200
300
400
Days after OLT Fig. 4. Comparison of estimated cumulative incidence of carbapenemase-producing Enterobacteriaceae (CPE) infection based on patient CPE colonization status. A competing risk multivariable regression model accounting for Model for End-Stage Liver Disease (MELD) at liver transplant (LT), combined transplant, prolonged mechanical ventilation, reintervention and rejection was used to estimate the 1-year cumulative incidence of CPE infection in patients depending on the time(s) of CPE colonization relative to LT.
Discussion In this large cohort of liver transplant recipients, we have confirmed the key role of CPE colonization, either at LT or acquired after LT, as the pre-eminent risk factor for developing CPE infection after transplant. Factors associated with a complicated clinical course, both before and after transplant, along with combined transplant increased the rate of CPE infection among colonized individuals. The key role of rectal colonization in developing CPE infection has been confirmed in both non-SOT and SOT patients [6,14]. However, the impact of colonization according to the moment of carriage acquisition, before or after transplant, is a matter of debate. In our previous study, including 237 LT recipients, we found that only colonization acquired after LT was associated with an augmented rate of CPE infection, and that the impact of colonization was significant in the presence of a complicated post-operative course [5]. Subsequent larger studies from other centres observed that both colonization at LT and that acquired after LT were associated with increased risk of CPE infection after transplant [4,6]. In particular, Freire et al. analysed 386 LT recipients: 68 were found to be colonized at transplant and 104 acquired colonization after LT; overall, 59 patients developed CPE infection after LT [6]. At multivariate analysis, colonization at LT was the strongest risk factor for CPE infection followed by CPE colonization acquired after LT, combined transplant, re-operation, renal replacement therapy and MELD >32 at LT. Our current study confirmed these findings. We found a similar hazard and sub-hazard ratio (accounting for the potential competing risk of dying within 1-year after LT) for both colonization at LT and that acquired after LT, underlying the key role of colonization whenever it occurs. Along with colonization, we confirmed the increased risk of CPE infection among individuals undergoing combined transplant, probably reflecting the higher complexity of such patients pre-operatively, intraoperatively and after LT. The other variables associated with CPE infection risk also indicate a more severe patient condition at LT (higher MELD values) and a complicated course after LT (prolonged mechanical ventilation, re-intervention and rejection). Among preventive measures for CPE infection after SOT, several strategies targeted to the CPE carrier have been proposed, including denying transplant, selective decontamination with non-
absorbable drugs, targeted antibiotic prophylaxis, pre-emptive treatment and faecal microbiota transplantation [15,16]. None of these procedures was adopted in our centre during the study period, because data concerning the efficacy, safety and feasibility of such interventions in the setting of LT are still unproven. Our study has some limitations. Being a monocentric study, our data may vary from those found in other centres. However, as discussed above, our data are similar to those reported by colleagues from other countries with high endemic levels of CPE [6]. Incidence of CPE colonization per 10 000 candidate days may have been overestimated as we considered only patients who were definitively transplanted. Cohort studies of LT candidates to establish the impact of CPE colonization on long-term patient outcome should be pursued. Our surveillance screening consisted only of RS, we did not evaluate routinely other sites of colonization that could have been linked to an augmented risk of CPE infection [14]. Finally, the type of CPE was not available for all colonized and/or infected patients, but according to national and local previous data [2,17], we can affirm that the main mechanism (>90%) of carbapenem resistance in our study population was Klebsiella pneumoniae carbapenemase production. To conclude, along with infection control activities, interventions targeted to the colonized LT candidate and/or recipient seem to be necessary for reducing CPE infection incidence after LT, mainly in individuals with higher MELD values at LT, undergoing combined transplant or with a complicated post-operative course. Further studies aimed at best identifying those colonized individuals at higher risk of infection, in whom preventive strategies could be maximally effective overcoming potential harms, are needed.
Transparency declaration The authors state no conflict of interest related to the content of the present study.
Financial support No external funding was received for the present study.
Please cite this article as: Giannella M et al., The impact of carbapenemase-producing Enterobacteriaceae colonization on infection risk after liver transplantation: a prospective observational cohort study, Clinical Microbiology and Infection, https://doi.org/10.1016/j.cmi.2019.04.014
M. Giannella et al. / Clinical Microbiology and Infection xxx (xxxx) xxx
Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi.org/10.1016/j.cmi.2019.04.014.
[8]
[9]
References [1] 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;29:1099e106. [2] Errico G, Gagliotti C, Monaco M, Masiero L, Gaibani P, Ambretti S, et al. Colonization and infection due to carbapenemase-producing Enterobacteriaceae in liver and lung transplant recipients and donor-derived transmission: a prospective cohort study conducted in Italy. Clin Microbiol Infect 2019;25: 203e9. Epub 2018/05/26. [3] Lanini S, Costa AN, Puro V, Procaccio F, Grossi PA, Vespasianoi F, et al. Incidence of carbapenem-resistant Gram negatives in Italian transplant recipients: a nationwide surveillance study. PLoS One 2015;10:e0123706. Epup 2015/04/04. [4] Mazza E, Prosperi M, Panzeri MF, Limuti R, Nichelatti M, De Gasperi A. Carbapenem-resistant Klebsiella pneumoniae infections early after liver transplantation: a single-center experience. Transplant Proc 2017;49:677e81. Epub 2017/05/02. [5] Giannella M, Bartoletti M, Morelli MC, Tedeschi S, Cristini F, Tumietto F, et al. Risk factors for infection with carbapenem-resistant Klebsiella pneumoniae after liver transplantation: the importance of pre- and posttransplant colonization. Am J Transplant 2015;15:1708e15. Epub 2015/03/11. [6] Freire MP, Oshiro IC, Pierrotti LC, Bonazzi PR, de Oliveira LM, Song AT, et al. Carbapenem-resistant Enterobacteriaceae acquired before liver transplantation: impact on recipient outcomes. Transplantation 2017;101:811e20. Epub 2016/12/24. [7] Horan TC, Andrus M, Dudeck MA. CDC/NHSN surveillance definition of health care-associated infection and criteria for specific types of infections in
[10] [11] [12] [13]
[14]
[15]
[16]
[17]
7
the acute care setting. Am J Infect Control 2008;36:309e32. Epub 2008/06/ 10. Moore KP, Wong F, Gines P, Bernardi M, Ochs A, Salerno F, et al. The management of ascites in cirrhosis: report on the consensus conference of the international ascites club. Hepatology 2003;38:258e66. Epub 2003/06/28. Angeli P, Gines P, Wong F, Bernardi M, Boyer TD, Gerbes A, et al. Diagnosis and management of acute kidney injury in patients with cirrhosis: revised consensus recommendations of the International Club of Ascites. J Hepatol 2015;62:968e74. Epub 2015/02/02. Dharel N, Bajaj JS. Definition and nomenclature of hepatic encephalopathy. J Clin Exp Hepatol 2015;5:S37e41. Epub 2015/06/05. Chen YC, Chen TH. The RIFLE criteria and renal prognosis in acute kidney injury. Kidney Int 2008;74:1492. author reply 1492e3, Epub 2008/11/15. Tamma PD, Simner PJ. Phenotypic detection of carbapenemase-producing organisms from clinical isolates. J Clin Microbiol 2018;56. Epub 2018/08/31. Traczewski MM, Carretto E, Canton R, Moore NM. Multicenter evaluation of the Xpert Carba-r assay for detection of carbapenemase genes in Gramnegative isolates. J Clin Microbiol 2018;56. Epub 2018/06/01. Giannella M, Trecarichi EM, De Rosa FG, Del Bono V, Bassetti M, Lewis RE, et al. Risk factors for carbapenem-resistant Klebsiella pneumoniae bloodstream infection among rectal carriers: a prospective observational multicentre study. Clin Microb Infect 2014;20:1357e62. Epub 2014/07/02. Aguado JM, Silva JT, Fernandez-Ruiz M, Cordero E, Fortun J, Gudiol C, et al. Management of multidrug resistant Gram-negative bacilli infections in solid organ transplant recipients: SET/GESITRA-SEIMC/REIPI recommendations. Transplant Rev (Orlando) 2018;32:36e57. Epub 2017/08/16. Tacconelli E, Mazzaferri F, Marie de Smet A, Bragantini D, Eggimann P, Huttner BD, et al. ESCMID-EUCIC clinical guidelines on decolonisation of multidrug-resistant Gram-negative bacteria carriers. Clin Microbiol Infect 2019 Jan 29. https://doi.org/10.1016/j.cmi.2019.01.005. pii: S1198-743X(19) 30025-4. [Epub ahead of print]. Gaibani P, Ambretti S, Berlingeri A, Gelsomino F, Bielli A, Landini MP, et al. Rapid increase of carbapenemase-producing Klebsiella pneumoniae strains in a large Italian hospital: surveillance period 1 Marche30 September 2010. Euro Surveill 2011;16.
Please cite this article as: Giannella M et al., The impact of carbapenemase-producing Enterobacteriaceae colonization on infection risk after liver transplantation: a prospective observational cohort study, Clinical Microbiology and Infection, https://doi.org/10.1016/j.cmi.2019.04.014
Contents lists available at ScienceDirect
Clinical Microbiology and Infection journal homepage: www.clinicalmicrobiologyandinfection.com
Original article
The impact of carbapenemase-producing Enterobacteriaceae colonization on infection risk after liver transplantation: a prospective observational cohort study M. Giannella 1, *, M. Bartoletti 1, C. Campoli 1, M. Rinaldi 1, S. Coladonato 1, R. Pascale 1, S. Tedeschi 1, S. Ambretti 2, F. Cristini 1, F. Tumietto 1, A. Siniscalchi 3, V. Bertuzzo 4, M.C. Morelli 5, M. Cescon 4, A.D. Pinna 4, R. Lewis 1, P. Viale 1 1)
Infectious Diseases Unit, Department of Medical and Surgical Sciences, Policlinico Sant'Orsola Malpighi, University of Bologna, Bologna, Italy Operative Unit of Clinical Microbiology, Policlinico Sant'Orsola Malpighi, University of Bologna, Bologna, Italy 3) Anaesthesia Unit, Department of Medical and Surgical Sciences, Policlinico Sant'Orsola Malpighi, University of Bologna, Bologna, Italy 4) Liver and Multiorgan Transplant Unit, Department of Medical and Surgical Sciences, Policlinico Sant'Orsola Malpighi, University of Bologna, Bologna, Italy 5) Internal Medicine Unit for the Treatment of Severe Organ Failure, Department of Medical and Surgical Sciences, Policlinico Sant'Orsola Malpighi, University of Bologna, Bologna, Italy 2)
a r t i c l e i n f o
a b s t r a c t
Article history: Received 3 February 2019 Received in revised form 29 March 2019 Accepted 12 April 2019 Available online xxx
Objective: To investigate the impact of colonization with carbapenemase-producing Enterobacteriaceae (CPE) on the CPE infection risk after liver transplantation (LT). Methods: Prospective cohort study of all adult patients undergoing LT at our centre over an 8-year period (2010e2017). Individuals were screened for CPE colonization by rectal swabs at inclusion onto the waiting list, immediately before LT and weekly after LT until hospital discharge. Asymptomatic carriers did not receive decolonization, anti-CPE prophylaxis or pre-emptive antibiotic therapy. Participants were followed up for 1 year after LT. Results: We analysed 553 individuals who underwent a first LT, 38 were colonized with CPE at LT and 104 acquired colonization after LT. CPE colonization rates at LT and acquired after LT increased significantly over the study period: incidence rate ratios (IRR) 1.21 (95% CI 1.05e1.39) and 1.17 (95% CI 1.07e1.27), respectively. Overall, 57 patients developed CPE infection within a median of 31 (interquartile range 11 e115) days after LT, with an incidence of 3.05 cases per 10 000 LT-recipient-days and a non-significant increase over the study period (IRR 1.11, 95% CI 0.98e1.26). In multivariable analysis, CPE colonization at LT (hazard ratio (HR) 18.50, 95% CI 6.76e50.54) and CPE colonization acquired after LT (HR 16.89, 95% CI 6.95e41.00) were the strongest risk factors for CPE infection, along with combined transplant (HR 2.60, 95% CI 1.20e5.59), higher Model for End-Stage Liver Disease at the time of LT (HR 1.03, 95% CI 1.00e1.07), prolonged mechanical ventilation (HR 2.63, 95% CI 1.48e4.67), re-intervention (HR 2.16, 95% CI 1.21 e3.84) and rejection (HR 2.81, 95% CI 1.52e5.21). Conclusions: CPE colonization at LT or acquired after LT were the strongest predictors of CPE infection. Prevention strategies focused on LT candidates and recipients colonized with CPE should be investigated. M. Giannella, Clin Microbiol Infect 2019;▪:1 © 2019 European Society of Clinical Microbiology and Infectious Diseases. Published by Elsevier Ltd. All rights reserved.
Editor: A. Kalil Keywords: Carbapenem-resistant Carbapenemase-producing Enterobacteriaceae Colonization Infection risk Liver transplantation
Introduction
* Corresponding author. M. Giannella, Infectious Disease Unit, Policlinico Sant'Orsola-Malpighi, Via Massarenti 11, 40138, Bologna, Italy. E-mail address: [email protected] (M. Giannella).
Solid organ transplant (SOT) recipients are at higher risk than non-SOT recipients for colonization and/or infection with carbapenemase-producing Enterobacteriaceae (CPE) [1]. Higher rates of CPE acquisition occur in lung and liver transplant recipients compared with other types of SOT, with liver transplantation (LT)
https://doi.org/10.1016/j.cmi.2019.04.014 1198-743X/© 2019 European Society of Clinical Microbiology and Infectious Diseases. Published by Elsevier Ltd. All rights reserved.
Please cite this article as: Giannella M et al., The impact of carbapenemase-producing Enterobacteriaceae colonization on infection risk after liver transplantation: a prospective observational cohort study, Clinical Microbiology and Infection, https://doi.org/10.1016/j.cmi.2019.04.014
2
M. Giannella et al. / Clinical Microbiology and Infection xxx (xxxx) xxx
accounting for most episodes [2]. Infection with CPE after SOT is associated with a mortality risk up to ten times higher than that with carbapenem-susceptible Enterobacteriaceae [3]. The knowledge of risk factors for CPE infection is key to designing preventive strategies. In LT recipients, CPE infection is usually preceded by CPE rectal carriage. Indeed, CPE colonization has been described as an important predictor of CPE infection [4e6]. However, colonization alone does not seem to be sufficient for developing infection [5]. In addition, data concerning the risk of CPE infection according to the moment of rectal carriage acquisition, before or after LT, are controversial [4e6]. We performed a large prospective cohort study of individuals undergoing LT to investigate risk factors for CPE infection. The main goal of this study was to define the impact of CPE colonization status before or after LT. Material and methods Study design and participants We performed a prospective observational cohort study of all consecutive adults (18 years) undergoing LT at our hospital, from June 2010 to December 2017. For the purpose of this study, patients were followed up for 12 months after LT by hospital and outpatient clinical assessments. Only the first episode of CPE infection was considered for the analysis. The study was approved by our Institutional Review Board, and patient informed consent was obtained. Setting Sant'Orsola-Malpighi Hospital is a 1420-bed tertiary-care University Hospital in Bologna (northern Italy) with an average of 72 000 admissions per year. The hospital has an active LT programme, performing an average of 90 procedures per year. During the study period, all LT candidates were screened with rectal swab (RS) for CPE carriage at the time of inclusion in the waiting list, in case of hospital (re-)admission, and at the time of transplant. RS was repeated once weekly after LT until hospital discharge. RS was repeated with the same frequency also in the case of hospital re-admission. Isolation and contact precautions were activated for CPE carriers, but no pharmacological interventions, including selective intestinal decontamination, CPEtargeted surgical prophylaxis, or antimicrobial treatment, were administered to asymptomatic patients. Standard perioperative prophylaxis consisted of short-course ampicillin-sulbactam. All suspected infections were managed in collaboration with a dedicated team of infectious disease specialists. Additional information on the transplant management at the study centre is available in the Supplementary material (Data S1). Data collection and definitions Variables were collected on standardized case report forms, which have been collected electronically since 2014. The end-point variable, CPE infection, was assessed according to CDC criteria [7], from LT to 12 months after LT. Exposure variables were evaluated from LT to the first episode of CPE infection, death or 12 months after LT. The main exposure variable, CPE colonization, was defined as the isolation of CPE from RS in the absence of symptoms and signs of infection. Every sample other than blood cultures or sterile fluids (including urine, respiratory samples, surgical site and other skin cultures) positive for CPE in asymptomatic patients was considered as colonization. Individuals found to be colonized with CPE by
cultures performed at the inclusion in waiting list, during the stay in waiting list, or at the time of LT were considered as CPE carriers at LT, and those found to be colonized afterwards were considered as having acquired CPE carriage after LT. The other exposure variables included demographic data (age and sex), and co-morbidities according to Charlson index, aetiology of liver disease, Model for End-Stage Liver Disease (MELD) at inclusion in the waiting list, and at LT. Pre-transplant variables (within 90 days) included: intensive care unit admission, grade 3 and/or refractory ascites [8], hepatorenal syndrome [9], gastrointestinal bleeding, hepatic encephalopathy [10], any infection, infection with multidrug-resistant bacteria, candidaemia and Clostridium difficile infection. Graft variables included donor age and cold ischaemia time. Intraoperative variables were bleeding with the need for 40 units of cellular blood products transfusion, and type of biliary anastomosis. Post-LT complications included reintervention, acute kidney injury according to RIFLE criteria [11], renal replacement therapy, prolonged (48 hours) mechanical ventilation, primary graft non-function, biopsy-proven rejection, re-transplantation, histological recurrence of hepatitis C virus infection and cytomegalovirus infection defined as >100 000 copies/mL. Microbiology Surveillance RS were collected using E-swab (Copan, Brescia, Italy), inoculated onto a chromogenic medium (Kima Meus, Arzegrande, Italy) containing a carbapenem antibiotic as selective agent and incubated at 35e37 C for 18e24 hours. Clinical isolates were processed following standard procedures. Species identification and routine antimicrobial susceptibility testing were performed rieux, using Vitek-MS MALDI-TOF and VITEK2 systems (bioMe Marcy l'Etoile, France). Results for MIC were interpreted according to EUCAST breakpoints. For all Enterobacterales isolates with MIC values higher than the EUCAST cut-off for at least one of the tested carbapenems (ertapenem, imipenem, meropenem), carbapenemase production was evaluated. Different confirmation methods have been used from 2010 to 2017, including modified Hodge-test, disc diffusion synergy test with inhibitors (Rosco Diagnostika, Taastrup, Denmark), lateral flow immunoassays (Coris BioConcept, Gembloux, Belgium), the Neo-Rapid carb sCPEen test (Rosco Diagnostika) [12] and Xpert Carba-R molecular method (Cepheid, Sunnyvale, CA, USA) [13]. In addition, carbapenem MICs were rieux). confirmed by E-test (Biome Statistical analysis Incidence rates ±95% CI for CPE colonization at LT were calculated and plotted per 10 000 patient-days using only new episodes observed during the study period by the overall number of patientdays computed from the day of inclusion in waiting list to LT. To calculate the incidence of CPE colonization and infection after LT, we divided the number of new episodes by the overall number of patient-days in the study cohort computed from the day of transplantation until day 365 or patient death. Poisson regression was used to analyse trends in the incidence of pre- and post-transplant CPE colonization as well as CPE infection per year adjusting for patient-days of exposure and reported as incidence rate ratio (IRR) (±95% CI) over the 8-year period. The impact of pre-transplant MELD score on the rate of CPE colonization was analysed using classification and regression tree analysis. Martingale residuals of a Cox model were used to calculate c2 values for all possible MELD score cut points, resulting in significantly different rates of pre-transplant CPE colonization.
Please cite this article as: Giannella M et al., The impact of carbapenemase-producing Enterobacteriaceae colonization on infection risk after liver transplantation: a prospective observational cohort study, Clinical Microbiology and Infection, https://doi.org/10.1016/j.cmi.2019.04.014
M. Giannella et al. / Clinical Microbiology and Infection xxx (xxxx) xxx Table 1 Characteristics of study cohort
(a)
5.6 ± 1.9 6 (1.1) 274 (49.5) 127 (23) 20 (3.6) 237 (42.9) 27 (4.9) 16 (11e23) 19 (12e26) 179 (36e450) 266 (48.1) 114 (20.6) 93 (16.8) 31 (5.6) 68 (12.3) 120 (21.7) 34 (6.1) 1 (0.2) 7 (1.3)
Rate (per 10 000)
52.8 ± 10.7 394 (71.2)
12 10 8 6 4 2 0 1
(b) 24
38 (6.9) 109 (19.7) 553 (100) 63 (46e75) 7 (6e7.3)
2
3
4
Year
5
6
7
8
IRR 1.16 (95% CI 1.01–1.23), p 0.03
22 20 18 16 14 12 10 8 6 4 2 0 1
2
3
4
5
6
7
8
7
8
Year 93 (16.8) 129 (23.3) 12 (2.2) 33 (6) 143 (25.9) 68 (12.3) 102 (18.4) 89 (16.1) 26 (4.7) 71 (12.8) 42 (7.6) 21 (3.8) 17 (11e31) 57 (10.3)
Abbreviations: CMV, cytomegalovirus; CPE, carbapenemase-producing Enterobacteriaceae; HCV, hepatitis C virus; HIV, human immunodeficiency virus; IQR, interquartile range; LT, liver transplant; MELD, Model for End-stage Liver Disease; PGNF, primary graft non-function; SD, standard deviation. a Some patients presented more than one underlying liver disease. b Episodes needing 40 units of cellular blood products transfusion.
Patients who developed a CPE infection within 12 months after LT were compared with those who did not develop CPE infection during that period. Categorical variables were compared using Pearson c2 or Fisher's exact test when appropriate. Continuous variables were compared using Student's t or ManneWhitney U test according to their distribution. All the variables with a p value 0.1 at the univariate analysis were entered into a multivariable Cox regression model after verifying for proportional hazards (Fig. S1) and non-collinearity (see Supplementary material: Data S1, Table S1 and Fig. S2). Patients were considered from the day of LT up to 12 months, or before in case of CPE infection or death. We also examined the effects of these covariates in a multivariable Fine and Gray model considering death before CPE
(c) 16
IRR 1.11 (95% CI 0.99–1.25), p 0.09
14 Rate (per 10 000)
Demographic data Age (years), mean ± SD Sex, male, n (%) Co-morbidities Charlson index, mean ± SD HIV co-infection, n (%) Underlying liver disease, n (%)a Viral disease Alcohol Fulminant liver failure Hepatocellular carcinoma Combined transplant MELD at inclusion in waiting list, median (IQR) MELD at LT, median (IQR) Time on waiting list (days), median (IQR) Complications within 90 days before LT, n (%) Grade 3 and/or refractory ascites Encephalopathy Hepato-renal syndrome Gastrointestinal bleeding Admission to intensive care unit Any infection Infection with multidrug-resistant bacteria and/or candidaemia Clostridium difficile infection CPE infection pre-LT Rectal swab positive for CPE, n (%) Colonization at LT Colonization acquired after LT Graft characteristics Deceased donor, n (%) Donor age (years), median (IQR) Cold ischaemia time (hours), median (IQR) Intraoperative variables, n (%) Choledochojejunostomy Bleedingb Post-operative complications, n (%) CMV infection (>100 000 copies/mL) HCV relapse Acute kidney injury Renal replacement therapy Re-intervention Mechanical ventilation >48 hours PGNF Rejection Biopsy proven rejection Re-transplantation Outcome Length of hospital stay after LT (days), median (IQR) All-cause 1-year mortality, n (%)
IRR 1.21 (95% CI 1.05–1.40); p 0.008
14
Rate (per 10 000)
Total (n ¼ 553)
16
3
12 10 8 6 4 2 0 1
2
3
4
5
6
Year Fig. 1. Trends in carbapenemase-producing Enterobacteriaceae (CPE) colonization and infections over the study period. (a) Pre-transplant CPE colonization; (b) posttransplant CPE colonization; (c) post-transplant CPE infection. Points show the estimated incidence per year with 95% CI indicated by shading. Changes in the incidence rate per year were analysed using Poisson regression to derive the incidence rate ratio (IRR).
infection as a competing risk. All analyses were performed with STATA 13 (Stata Corp, College Station, TX, USA).
Results During the study period, 553 individuals underwent LT. The general characteristics of the study cohort are shown in Table 1. Overall, CPE colonization was detected in 147 individuals (26.6%). In all cases, the first sample showing CPE carriage was RS,
Please cite this article as: Giannella M et al., The impact of carbapenemase-producing Enterobacteriaceae colonization on infection risk after liver transplantation: a prospective observational cohort study, Clinical Microbiology and Infection, https://doi.org/10.1016/j.cmi.2019.04.014
4
M. Giannella et al. / Clinical Microbiology and Infection xxx (xxxx) xxx
(a)
(b)
6–33
349
Col. RHR 13
Pre-transplant CPE colonization
N 6–19
.42
MELD score
20–33
170
18
2.91
MELD score
1.00
MELD 6–19
0.75
MELD 20–33 MELD >33
0.50
Log rank p<0.0001
0.25
0.00 34–47
22
6 33.74
0
100
200
300
400
Days on transplant list Number at risk MELD 6–19 = 349 MELD 20–33 = 170 MELD >33 = 34
279 54 6
217 36 4
171 26 1
135 15 1
Fig. 2. Impact of pre-transplant Model for End-Stage Liver Disease (MELD) score at time of patient listing and rates of carbapenemase-producing Enterobacteriaceae (CPE) colonization. (a) Martingale residuals of a Cox regression model for newly-acquired CPE colonization were used to calculate c2 values for all possible MELD score cut-points using classification and regression tree (CART) analysis resulting in significantly different rates of pre-transplant CRE colonization. (b) KaplaneMeier failure curves for pre-transplant CRE colonization at CART-derived MELD score cut-offs.
CPE INFECTION 406 No CPE carriage
553 LT recipients
6 (1.5%) 48 (IQR 20–362) days 14 (36.8%)
38 CPE carriage at LT
13 (IQR 3–21.5) days
109 CPE carriage aer LT
56 (IQR 14–129) days
37 (33.9%)
Fig. 3. Carbapenemase-producing Enterobacteriaceae (CRE) infection rate and median time to CRE infection after liver transplant (LT) according with the colonization status before and after LT.
in some individuals CPE was subsequently isolated also from other sites, mainly urine. Among CPE-colonized individuals, 38 (25.8%) were found to be colonized at LT. The median time of carriage detection was 12 (interquartile range (IQR) 0.75e40) days (maximum time 309 days) before LT. The incidence of CPE colonization at LT was 1.88 cases per 10 000 candidate-days. The incidence rate of CPE colonization at LT increased significantly over the study period (IRR 1.21, 95% CI 1.05e1.39, p 0.008) (Fig. 1a). The risk of CPE colonization varied depending on the patient MELD score at the time of listing, classification and regression tree analysis showed that the risk of being colonized with CPE at LT increased significantly in individuals with a MELD score at listing >19 (Fig. 2). However, there was not a significant change in the number of individuals with MELD scores >19 over the study period (IRR 1.01, 95% CI 0.95e1.06, p 0.79). One hundred nine individuals (19.7%) acquired CPE colonization within a median of 15 (IQR 6e34) days after LT. The overall incidence of CPE colonization after LT was 5.84 cases per 10 000 LT recipient-days with a significant increase per year (IRR 1.17, 95% CI
1.07e1.27, p < 0.001) (Fig. 1b). Although a relationship between CPE colonization post-transplant and MELD at the time of transplant was detected (OR 1.02, 95% CI 1.00e1.04; p 0.03) no cut-off for high risk versus low risk was identified by classification and regression tree analysis. Overall, 57 (10.3%) individuals developed one or more episodes of CPE infection after LT. The infection rate according with the colonization status is shown in Fig. 3. The median time to the first CPE infection episode was 31 (IQR 11e115) days after transplant, varying from 13 (IQR 3e21.5) days in individuals colonized with CPE before LT, to 56 (IQR 14e129) days in those who acquired CPE colonization after LT. In this latter group, the median time from colonization acquisition to infection development was 20 (IQR 4e72) days. The overall incidence of CPE infection was 3.05 per 10 000 LT recipient-days, without a significant increase over the study period (IRR 1.11, 95% 0.98e1.26, p 0.08) (Fig 1c). Most individuals who developed CPE infection were bacteraemic (84.2%), the infection types/sources were lower respiratory tract infection (n ¼ 19, 33.3%), surgical site including superficial wound (n ¼ 1), biliary (n ¼ 5) and intra-abdominal tract (n ¼ 11), infections (n ¼ 17, 29.8%), undefined (n ¼ 14, 24.5%), central venous catheter (n ¼ 4, 7%) and urinary tract infection (n ¼ 3, 5.2%). Severe sepsis and septic shock were present at infection onset in 33.3% and 24.6% of cases, respectively. The mechanism of carbapenem resistance was determined in 32 out of 57 patients with infection showing Klebsiella pneumoniae carbapenemase production in all cases. A comparison of patients with and without CPE infection is shown in Table 2. Factors associated with CPE infection included alcoholic liver disease, median MELD at listing, grade 3 and/or refractory ascites within 90 days before LT, previous CPE infection, CPE carriage at LT, median MELD at LT, combined transplant, prolonged cold ischaemia time of the graft, CPE carriage acquisition after LT, cytomegalovirus infection, acute kidney injury, renal replacement therapy, re-intervention, prolonged mechanical ventilation, primary graft non-function, rejection and re-
Please cite this article as: Giannella M et al., The impact of carbapenemase-producing Enterobacteriaceae colonization on infection risk after liver transplantation: a prospective observational cohort study, Clinical Microbiology and Infection, https://doi.org/10.1016/j.cmi.2019.04.014
M. Giannella et al. / Clinical Microbiology and Infection xxx (xxxx) xxx
5
Table 2 Comparison of patients with and without CPE infection within 1 year after liver transplantation
Demographic data Age (years), mean ± SD Sex, male, n (%) Co-morbidities Charlson index, mean ± SD HIV co-infection, n (%) Underlying liver disease, n (%)a Viral disease Alcohol Fulminant liver failure Hepatocellular carcinoma Combined transplant MELD at inclusion in waiting list, median (IQR) MELD at LT, median (IQR) Complications within 90 days before LT, n (%) Grade 3 and/or refractory ascites Encephalopathy Hepato-renal syndrome Gastrointestinal bleeding Admission to intensive care unit Any infection Infection with multi-drug-resistant bacteria and/or candidaemia Clostridium difficile infection CRE infection pre-LT Rectal swab positive for CPE, n (%) Colonization at LT Colonization acquired after LT Graft characteristics Donor age (years), median (IQR) Cold ischaemia time (hours), median (IQR) Intraoperative variables, n (%) Choledochojejunostomy Bleedingb Post-operative complications, n (%) CMV infection (>100 000 copies/mL) HCV relapse Acute kidney injury Renal replacement therapy Re-intervention Mechanical ventilation >48 hours PGNF Rejection Biopsy proven rejection Re-transplantation Outcome Length of hospital stay after LT (days), median (IQR) All-cause 1-year mortality, n (%)
Patients with CPE infection, (n ¼ 57)
Patients without CPE infection (n ¼ 496)
p
53.6 ± 10.5 44 (77.2)
52.7 ± 10.8 350 (70.6)
0.53 0.35
5.7 ± 1.8 0
5.6 ± 1.9 6 (1.2)
0.65 1
28 (49.1) 20 (35.1) 0 19 (33.3) 8 (14) 20 (14e25) 24 (15e30)
246 (49.6) 107 (21.6) 20 (4) 218 (44) 19 (3.8) 15 (10e22) 18 (12e26)
1 0.03 0.15 0.16 0.004 0.001 0.002
40 (70.2) 14 (24.6) 13 (22.8) 5 (8.8) 11 (19.3) 17 (29.8) 7 (12.3) 0 3 (5.3)
226 (45.6) 100 (20.2) 80 (16.1) 26 (5.2) 57 (11.5) 103 (20.8) 27 (5.4) 1 (0.2) 4 (0.8)
<0.001 0.5 0.3 0.35 0.13 0.13 0.07 1 0.03
14 (24.6) 37 (64.9)
24 (4.8) 72 (14.5)
<0.001 <0.001
67 (46e77) 7 (6e7.3)
63 (46e75) 6.5 (6e7.2)
0.35 0.04
9 (15.8) 17 (29.8)
84 (16.9) 112 (22.9)
0.85 0.25
4 (7) 5 (8.8) 29 (50.9) 21 (36.8) 31 (54.4) 23 (40.4) 10 (17.5) 18 (31.6) 11 (19.3) 10 (17.5)
8 (1.6) 28 (5.6) 114 (23) 47 (9.5) 71 (14.3) 66 (13.3) 16 (3.2) 53 (10.7) 31 (6.3) 11 (2.2)
0.03 0.37 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 0.002 <0.001
46 (26e104) 24 (42.1)
16 (11e26) 33 (6.7)
<0.001 <0.001
Abbreviations: CMV, cytomegalovirus; CPE, carbapenemase-producing Enterobacteriaceae; HCV, hepatitis C virus; HIV, human immunodeficiency virus; IQR, interquartile range; LT, liver transplant; MELD, Model for End-stage Liver Disease; PGNF, primary graft non-function; SD, standard deviation. a Some patients presented more than one underlying liver disease. b Episodes needing 40 units of cellular blood products transfusion.
transplantation. The all-cause 1-year mortality rates in individuals with and without CPE infection were 42.1% and 6.7% (p < 0.001), respectively. In addition, the median length of hospital stay after LT was significantly longer among individuals who developed CPE infection (46 versus 16 days, p < 0.001).
At multivariate Cox regression and competing risk analyses, the independent risk factors for CPE infection within 1 year after LT were: MELD at LT, combined transplant, colonization at LT, colonization acquired after LT, prolonged mechanical ventilation, reintervention and rejection (Table 3, Fig. 4).
Table 3 Multivariate analysis of risk factors for CPE infection within 1 year after liver transplantation Variable
HR (95% CI)
p
SHR (95% CI)
p
MELD score at transplant (per 1 point increase) Combined transplant CPE colonization at LT CPE colonization acquired after LT Prolonged mechanical ventilation Re-intervention Rejection
1.03 (1.00e1.07) 2.60 (1.20e5.59) 18.50 (6.76e50.54) 16.89 (6.95e41.00) 2.63 (1.48e4.67) 2.16 (1.21e3.84) 2.81 (1.52e5.21)
0.03 0.02 <0.001 <0.001 0.001 0.009 0.001
1.03 (1.00e1.08) 2.74 (1.08e6.96) 18.12 (6.98e47.01) 17.60 (7.36e42.08) 2.39 (1.32e4.35) 2.16 (1.18e3.94) 2.94 (1.59e5.45)
0.05 0.03 <0.001 <0.001 0.004 0.012 0.001
Abbreviations: CPE, carbapenemase-producing Enterobacteriaceae; HR, hazard ratio; LT, liver transplantation; MELD, Model for End-stage Liver Disease; SHR, sub-hazard ratio.
Please cite this article as: Giannella M et al., The impact of carbapenemase-producing Enterobacteriaceae colonization on infection risk after liver transplantation: a prospective observational cohort study, Clinical Microbiology and Infection, https://doi.org/10.1016/j.cmi.2019.04.014
6
M. Giannella et al. / Clinical Microbiology and Infection xxx (xxxx) xxx
Competing-risks regression .2
Colonization pre-OLT (SHR 18.12, 95% CI 6.98–47.01, p<0.0001) Cumulative Incidence
Colonization post-OLT(SHR 17.60, 95% CI 7.36–42.08, p<0.0001) .15
.1
.05
No colonization (reference) 0 0
100
200
300
400
Days after OLT Fig. 4. Comparison of estimated cumulative incidence of carbapenemase-producing Enterobacteriaceae (CPE) infection based on patient CPE colonization status. A competing risk multivariable regression model accounting for Model for End-Stage Liver Disease (MELD) at liver transplant (LT), combined transplant, prolonged mechanical ventilation, reintervention and rejection was used to estimate the 1-year cumulative incidence of CPE infection in patients depending on the time(s) of CPE colonization relative to LT.
Discussion In this large cohort of liver transplant recipients, we have confirmed the key role of CPE colonization, either at LT or acquired after LT, as the pre-eminent risk factor for developing CPE infection after transplant. Factors associated with a complicated clinical course, both before and after transplant, along with combined transplant increased the rate of CPE infection among colonized individuals. The key role of rectal colonization in developing CPE infection has been confirmed in both non-SOT and SOT patients [6,14]. However, the impact of colonization according to the moment of carriage acquisition, before or after transplant, is a matter of debate. In our previous study, including 237 LT recipients, we found that only colonization acquired after LT was associated with an augmented rate of CPE infection, and that the impact of colonization was significant in the presence of a complicated post-operative course [5]. Subsequent larger studies from other centres observed that both colonization at LT and that acquired after LT were associated with increased risk of CPE infection after transplant [4,6]. In particular, Freire et al. analysed 386 LT recipients: 68 were found to be colonized at transplant and 104 acquired colonization after LT; overall, 59 patients developed CPE infection after LT [6]. At multivariate analysis, colonization at LT was the strongest risk factor for CPE infection followed by CPE colonization acquired after LT, combined transplant, re-operation, renal replacement therapy and MELD >32 at LT. Our current study confirmed these findings. We found a similar hazard and sub-hazard ratio (accounting for the potential competing risk of dying within 1-year after LT) for both colonization at LT and that acquired after LT, underlying the key role of colonization whenever it occurs. Along with colonization, we confirmed the increased risk of CPE infection among individuals undergoing combined transplant, probably reflecting the higher complexity of such patients pre-operatively, intraoperatively and after LT. The other variables associated with CPE infection risk also indicate a more severe patient condition at LT (higher MELD values) and a complicated course after LT (prolonged mechanical ventilation, re-intervention and rejection). Among preventive measures for CPE infection after SOT, several strategies targeted to the CPE carrier have been proposed, including denying transplant, selective decontamination with non-
absorbable drugs, targeted antibiotic prophylaxis, pre-emptive treatment and faecal microbiota transplantation [15,16]. None of these procedures was adopted in our centre during the study period, because data concerning the efficacy, safety and feasibility of such interventions in the setting of LT are still unproven. Our study has some limitations. Being a monocentric study, our data may vary from those found in other centres. However, as discussed above, our data are similar to those reported by colleagues from other countries with high endemic levels of CPE [6]. Incidence of CPE colonization per 10 000 candidate days may have been overestimated as we considered only patients who were definitively transplanted. Cohort studies of LT candidates to establish the impact of CPE colonization on long-term patient outcome should be pursued. Our surveillance screening consisted only of RS, we did not evaluate routinely other sites of colonization that could have been linked to an augmented risk of CPE infection [14]. Finally, the type of CPE was not available for all colonized and/or infected patients, but according to national and local previous data [2,17], we can affirm that the main mechanism (>90%) of carbapenem resistance in our study population was Klebsiella pneumoniae carbapenemase production. To conclude, along with infection control activities, interventions targeted to the colonized LT candidate and/or recipient seem to be necessary for reducing CPE infection incidence after LT, mainly in individuals with higher MELD values at LT, undergoing combined transplant or with a complicated post-operative course. Further studies aimed at best identifying those colonized individuals at higher risk of infection, in whom preventive strategies could be maximally effective overcoming potential harms, are needed.
Transparency declaration The authors state no conflict of interest related to the content of the present study.
Financial support No external funding was received for the present study.
Please cite this article as: Giannella M et al., The impact of carbapenemase-producing Enterobacteriaceae colonization on infection risk after liver transplantation: a prospective observational cohort study, Clinical Microbiology and Infection, https://doi.org/10.1016/j.cmi.2019.04.014
M. Giannella et al. / Clinical Microbiology and Infection xxx (xxxx) xxx
Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi.org/10.1016/j.cmi.2019.04.014.
[8]
[9]
References [1] 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;29:1099e106. [2] Errico G, Gagliotti C, Monaco M, Masiero L, Gaibani P, Ambretti S, et al. Colonization and infection due to carbapenemase-producing Enterobacteriaceae in liver and lung transplant recipients and donor-derived transmission: a prospective cohort study conducted in Italy. Clin Microbiol Infect 2019;25: 203e9. Epub 2018/05/26. [3] Lanini S, Costa AN, Puro V, Procaccio F, Grossi PA, Vespasianoi F, et al. Incidence of carbapenem-resistant Gram negatives in Italian transplant recipients: a nationwide surveillance study. PLoS One 2015;10:e0123706. Epup 2015/04/04. [4] Mazza E, Prosperi M, Panzeri MF, Limuti R, Nichelatti M, De Gasperi A. Carbapenem-resistant Klebsiella pneumoniae infections early after liver transplantation: a single-center experience. Transplant Proc 2017;49:677e81. Epub 2017/05/02. [5] Giannella M, Bartoletti M, Morelli MC, Tedeschi S, Cristini F, Tumietto F, et al. Risk factors for infection with carbapenem-resistant Klebsiella pneumoniae after liver transplantation: the importance of pre- and posttransplant colonization. Am J Transplant 2015;15:1708e15. Epub 2015/03/11. [6] Freire MP, Oshiro IC, Pierrotti LC, Bonazzi PR, de Oliveira LM, Song AT, et al. Carbapenem-resistant Enterobacteriaceae acquired before liver transplantation: impact on recipient outcomes. Transplantation 2017;101:811e20. Epub 2016/12/24. [7] Horan TC, Andrus M, Dudeck MA. CDC/NHSN surveillance definition of health care-associated infection and criteria for specific types of infections in
[10] [11] [12] [13]
[14]
[15]
[16]
[17]
7
the acute care setting. Am J Infect Control 2008;36:309e32. Epub 2008/06/ 10. Moore KP, Wong F, Gines P, Bernardi M, Ochs A, Salerno F, et al. The management of ascites in cirrhosis: report on the consensus conference of the international ascites club. Hepatology 2003;38:258e66. Epub 2003/06/28. Angeli P, Gines P, Wong F, Bernardi M, Boyer TD, Gerbes A, et al. Diagnosis and management of acute kidney injury in patients with cirrhosis: revised consensus recommendations of the International Club of Ascites. J Hepatol 2015;62:968e74. Epub 2015/02/02. Dharel N, Bajaj JS. Definition and nomenclature of hepatic encephalopathy. J Clin Exp Hepatol 2015;5:S37e41. Epub 2015/06/05. Chen YC, Chen TH. The RIFLE criteria and renal prognosis in acute kidney injury. Kidney Int 2008;74:1492. author reply 1492e3, Epub 2008/11/15. Tamma PD, Simner PJ. Phenotypic detection of carbapenemase-producing organisms from clinical isolates. J Clin Microbiol 2018;56. Epub 2018/08/31. Traczewski MM, Carretto E, Canton R, Moore NM. Multicenter evaluation of the Xpert Carba-r assay for detection of carbapenemase genes in Gramnegative isolates. J Clin Microbiol 2018;56. Epub 2018/06/01. Giannella M, Trecarichi EM, De Rosa FG, Del Bono V, Bassetti M, Lewis RE, et al. Risk factors for carbapenem-resistant Klebsiella pneumoniae bloodstream infection among rectal carriers: a prospective observational multicentre study. Clin Microb Infect 2014;20:1357e62. Epub 2014/07/02. Aguado JM, Silva JT, Fernandez-Ruiz M, Cordero E, Fortun J, Gudiol C, et al. Management of multidrug resistant Gram-negative bacilli infections in solid organ transplant recipients: SET/GESITRA-SEIMC/REIPI recommendations. Transplant Rev (Orlando) 2018;32:36e57. Epub 2017/08/16. Tacconelli E, Mazzaferri F, Marie de Smet A, Bragantini D, Eggimann P, Huttner BD, et al. ESCMID-EUCIC clinical guidelines on decolonisation of multidrug-resistant Gram-negative bacteria carriers. Clin Microbiol Infect 2019 Jan 29. https://doi.org/10.1016/j.cmi.2019.01.005. pii: S1198-743X(19) 30025-4. [Epub ahead of print]. Gaibani P, Ambretti S, Berlingeri A, Gelsomino F, Bielli A, Landini MP, et al. Rapid increase of carbapenemase-producing Klebsiella pneumoniae strains in a large Italian hospital: surveillance period 1 Marche30 September 2010. Euro Surveill 2011;16.
Please cite this article as: Giannella M et al., The impact of carbapenemase-producing Enterobacteriaceae colonization on infection risk after liver transplantation: a prospective observational cohort study, Clinical Microbiology and Infection, https://doi.org/10.1016/j.cmi.2019.04.014