- Email: [email protected]
Post-surgical mediastinitis due to carbapenem-resistant Enterobacteriaceae: Clinical, epidemiological and survival characteristics
G Model
ARTICLE IN PRESS
ANTAGE-4775; No. of Pages 5
International Journal of Antimicrobial Agents xxx (2016) xxx–xxx
Contents lists available at ScienceDirect
International Journal of Antimicrobial Agents journal homepage: http://www.elsevier.com/locate/ijantimicag
Post-surgical mediastinitis due to carbapenem-resistant Enterobacteriaceae: Clinical, epidemiological and survival characteristics夽 C.S. Abboud a,∗ , J. Monteiro b , M.E. Stryjewski c , E.C. Zandonadi a , V. Barbosa a , D. Dantas a , E.E. Sousa a , M.J. Fonseca d , D.M. Jacobs e , A.C. Pignatari d , C. Kiffer d , G.G. Rao e a
Instituto Dante Pazzanese de Cardiologia, São Paulo, Brazil Associac¸ão Fundo de Incentivo à Pesquisa (AFIP–Medicina Diagnóstica), São Paulo, Brazil c Department of Medicine, Division of Infectious Diseases, Centro de Educación Médica e Investigaciones Clínicas ‘Norberto Quirno’ (CEMIC), Buenos Aires, Argentina d Laboratório Especial de Microbiologia Clínica (LEMC), Universidade Federal de São Paulo (UNIFESP), Escola Paulista de Medicina (EPM), São Paulo, Brazil e University at Buffalo School of Pharmacy and Pharmaceutical Sciences, Buffalo, NY, USA b
a r t i c l e
i n f o
Article history: Received 12 December 2015 Accepted 27 February 2016 Keywords: Mediastinitis Post-surgical infection Carbapenem-resistant Enterobacteriaceae Polymyxin resistance
a b s t r a c t Invasive infections due to carbapenem-resistant Enterobacteriaceae (CRE), including polymyxin-resistant (PR-CRE) strains, are being increasingly reported. However, there is a lack of clinical data for several lifethreatening infections. Here we describe a cohort of patients with post-surgical mediastinitis due to CRE, including PR-CRE. This study was a retrospective cohort design at a single cardiology centre. Patients with mediastinitis due to CRE were identified and were investigated for clinically relevant variables. Infecting isolates were studied using molecular techniques. Patients infected with polymyxin-susceptible CRE (PS-CRE) strains were compared with those infected with PR-CRE strains. In total, 33 patients with CRE mediastinitis were studied, including 15 patients (45%) with PR-CRE. The majority (61%) were previously colonised. All infecting isolates carried blaKPC genes. Baseline characteristics of patients with PR-CRE mediastinitis were comparable with those with PS-CRE mediastinitis. Of the patients studied, 70% received at least one agent considered active in vitro and most patients received at least three concomitant antibiotics. Carbapenem plus polymyxin B was the most common antibiotic combination (73%). Over 90% of patients underwent surgical debridement. Overall, in-hospital mortality was 33% and tended to be higher in patients infected with PR-CRE (17% vs. 53%; P = 0.06). In conclusion, mediastinitis due to CRE, including PR-CRE, can become a significant challenge in centres with CRE and a high cardiac surgery volume. Despite complex antibiotic treatments and aggressive surgical procedures, these patients have a high mortality, particularly those infected with PR-CRE. © 2016 Published by Elsevier B.V.
1. Introduction Post-sternotomy mediastinitis is a life-threatening infection that occurs in a small proportion (0.5–2.7%) of patients undergoing cardiothoracic surgery [1–4], resulting in longer hospital stay, higher medical costs, additional surgical procedures and, finally,
夽 This study was partially presented at IDWeek 2014, 8–12 October 2014, Philadelphia, PA [abstract no. 487]. ∗ Corresponding author. Present address: Division of Infectious Diseases, Instituto Dante Pazzanese de Cardiologia, Av. Dr Dante Pazzanese, 500, Vila Mariana, São Paulo, SP 04012-909, Brazil. Tel.: +55 11 5085 6255; fax: +55 1150856255. E-mail address: [email protected] (C.S. Abboud).
increased mortality. Recent studies showed that 14–32% of patients who developed post-surgical mediastinitis died, indicative of a higher risk of morbidity and mortality compared with patients who do not experience infectious complications [1–5]. Although the main micro-organisms implicated are Gram-positive bacteria (e.g. staphylococci), Gram-negative pathogens can also be a relevant cause of mediastinitis at some centres [5]. Importantly, compared with other micro-organisms, mediastinitis associated with Gramnegative pathogens was shown to have higher rates of in-hospital death [5]. Invasive infections associated with carbapenem-resistant Enterobacteriaceae (CRE) pose a serious challenge to clinicians. Treatment with drugs such as polymyxins (polymyxin B or colistin), carbapenems or their combinations are basically supported
http://dx.doi.org/10.1016/j.ijantimicag.2016.02.015 0924-8579/© 2016 Published by Elsevier B.V.
Please cite this article in press as: Abboud CS, et al. Post-surgical mediastinitis due to carbapenem-resistant Enterobacteriaceae: Clinical, epidemiological and survival characteristics. Int J Antimicrob Agents (2016), http://dx.doi.org/10.1016/j.ijantimicag.2016.02.015
G Model ANTAGE-4775; No. of Pages 5
ARTICLE IN PRESS C.S. Abboud et al. / International Journal of Antimicrobial Agents xxx (2016) xxx–xxx
2
by observational studies [6–8]. The emergence of resistance to polymyxins further complicates the situation [9–11]. Whilst there is an increasing body of evidence on the clinical characteristics of patients with more common CRE infections (e.g. bloodstream infections) [6,7], we are faced with a lack of clinical information in the case of less frequent but otherwise severe CRE infections such as post-surgical mediastinitis. To the best of our knowledge, this is the first study describing a cohort of patients with post-surgical mediastinitis due to CRE, including patients infected with strains resistant to polymyxins. Given the lack of evidence described above and the potential occurrence of polymyxin resistance in CRE, this study intended to describe the clinical characteristics of a significant post-sternotomy mediastinitis CRE outbreak population and to compare survival outcomes among polymyxin-resistant and -susceptible isolates. 2. Materials and methods
meropenem) resistance was defined as a minimum inhibitory concentration (MIC) of ≥4 g/mL according to Clinical and Laboratory Standards Institute (CLSI) breakpoints. Resistance to polymyxins was defined by a colistin MIC ≥4 g/mL. Molecular testing was conducted at Laboratório Especial de Microbiologia Clínica (LEMC) of Universidade Federal de São Paulo (UNIFESP) (São Paulo, Brazil) and included a real-time multiplex PCR assay for blaKPC , blaGES , blaOXA-48-like , blaNDM , blaIMP and blaVIM carbapenemase genes [14]. The genetic relatedness for each group of CRE clinical strains was evaluated by pulsed-field gel electrophoresis (PFGE) using the restriction enzyme SpeI (New England Biolabs Inc., Ipswich, MA) and the results were analysed using BioNumerics v.6.6 (Applied Maths, Sint-Martens-Latem, Belgium). Multilocus sequence typing (MLST) for Klebsiella pneumoniae strains was performed as described by Diancourt et al. [15], and the Institut Pasteur MLST database was used to assign allelic numbers and sequence types (http://bigsdb.web.pasteur.fr/klebsiella/klebsiella. html; accessed 11 November 2015).
2.1. Study population and setting 2.3. Antibiotic dosing This retrospective cohort study was conducted at Instituto Dante Pazzanese de Cardiologia (São Paulo, Brazil) between December 2010 and June 2014. This is a 350-bed hospital specialising in cardiology and cardiovascular surgery that serves as a referral centre for 300 000 patients, 90% of whom are from São Paulo. Patients undergoing surgical procedures at the institution are followed by the institutional teams until the patient recovers or is clinically stable, allowing for external follow-up. For the purpose of this study, patients aged ≥18 years diagnosed with mediastinitis due to CRE were identified. Mediastinitis was defined as the presence of at least one of the following: (i) organisms cultured from mediastinal tissue or fluid obtained during a surgical operation or needle aspiration; (ii) evidence of mediastinitis seen during a surgical operation or histopathological examination; and (iii) fever (>38 ◦ C), chest pain or sternal instability, and at least one of the following: purulent discharge from mediastinal area, organisms cultured from blood or discharge from mediastinal area, or mediastinal widening on the chest radiograph [12]. Patients were considered to be infected with CRE if such isolates were obtained from either blood cultures, surgical specimens or by needle aspiration from the sternal area and/or mediastinum. For the duration of the study, all patients with mediastinitis were under the care of the same infectious diseases and cardiovascular surgical staff members. Patient charts were reviewed to capture both demographic and clinically relevant data, including co-morbidities, prior cardiovascular procedures, immunosuppressant drugs, previous infections, colonisation and prior antibiotic use. The Society of Thoracic Surgeons (STS) score, indicative of operative mortality and morbidity of adult cardiac surgery, was calculated [13]. Surgical information, such as type and duration of the procedure, blood loss, perfusion time, extracorporeal circulation and re-operation, was also collected. All relevant information regarding antibiotic treatments was reviewed. This study was approved by the Ethical Committee of Instituto Dante Pazzanese de Cardiologia.
The following intravenous [16] antimicrobial doses were used at the institution for the treatment of patients with CRE mediastinitis, according to the antimicrobial stewardship programme of the hospital: polymyxin B, 25 000 UI/kg/day; amikacin, 15 mg/kg/day; gentamicin, 5 mg/kg/day; meropenem, 1 g every 8 h; imipenem, 500 mg every 6 h; tigecycline, 100 mg initial dose followed by 50 mg every 12 h; and ciprofloxacin, 400 mg every 12 h. The doses were adjusted according to the patient’s renal function. In the case of patients with a body mass index (BMI) >40 (kg/m2 ), the dose of meropenem was increased to 6 g day (2 g every 8 h). 2.4. Statistics analysis Each subject was included only once (index infection) in the analysis. Descriptive statistics were used to describe the overall characteristics of the whole cohort. Patients infected with polymyxin-resistant CRE (PR-CRE) strains were compared with those infected with polymyxin-susceptible CRE (PS-CRE) strains using either Fisher’s exact test or 2 test as appropriate for categorical variables, and Wilcoxon rank-sum test for continuous variables. Patients were followed during their hospital stay after the diagnosis of mediastinitis was made, and charts were reviewed up to October 2014 (3 months after the end of the study period) to determine the survival status following discharge. If the information was not available in the chart, it was retrieved by telephone contact. A survival analysis to compare survival between groups was performed using the log-rank test. The effect of PR-CRE on mortality was evaluated with a Cox proportional hazard model to determine the hazard ratio (HR) and 95% confidence interval (CI). A multivariate model was performed to take into account differences between the PS-CRE and PR-CRE groups. Significance was defined as a P-value of ≤0.05 (two-tailed). IBM SPSS Statistics for Windows v.19.0 software package (IBM Corp., Armonk, NY) was utilised for the statistical analyses.
2.2. Microbiology 3. Results All Gram-negative strains recovered from patients with mediastinitis were submitted to the local microbiology laboratory for identification and determination of antimicrobial susceptibility profiles by VITEK® 2 system using AST-N238 and AST-N239 cards (bioMérieux, Marcy-l’Étoile, France). Based on the results of the routine antimicrobial susceptibility testing and phenotypic tests (VITEK® 2 system and modified Hodge test), the clinical isolates were classified as CRE. Carbapenem (imipenem and
From December 2010 to June 2014, 5933 cardiac surgeries were performed at this single institution. During this period 92 patients (1.6%) were diagnosed with mediastinitis, among which Grampositive and Gram-negative pathogens were isolated from 45% and 55% of the patients, respectively. Enterobacteriaceae were isolated in 88% of patients infected with Gram-negative pathogens, including 36 patients infected with CRE strains. Three patients were
Please cite this article in press as: Abboud CS, et al. Post-surgical mediastinitis due to carbapenem-resistant Enterobacteriaceae: Clinical, epidemiological and survival characteristics. Int J Antimicrob Agents (2016), http://dx.doi.org/10.1016/j.ijantimicag.2016.02.015
G Model
ARTICLE IN PRESS
ANTAGE-4775; No. of Pages 5
C.S. Abboud et al. / International Journal of Antimicrobial Agents xxx (2016) xxx–xxx
3
Table 1 Baseline demographic and clinical characteristics of patients with mediastinitis and deep sternal wounds due to carbapenem-resistant Enterobacteriaceae (CRE), stratified by polymyxin resistance.a Total CRE (n = 33)
PS-CRE (n = 18)
PR-CRE (n = 15)
P-value
Patient demographics Female sex Age (years) Race White Weight (kg) BMI (kg/m2 ) BMI ≥30
19 (58%) 63 (56–72) 25 (76%) 75 (64–83) 28.8 (25.3–31.6) 12 (36%)
12 (67%) 65 (59–72) 14 (78%) 74 (64–84) 28.8 (26.2–32.4) 5 (28%)
7 (47%) 56 (53.5–69.5) 11 (73%) 77 (64–83) 29.6 (24.5–31.6) 7 (47%)
0.30 0.24 1 0.95 0.60 0.31
Medical history Hypertension Diabetes Smoking history (current) COPD Peripheral vascular disease Myocardial infarction Heart failure (NYHA-IV) Chronic renal failure Dialysis Steroids within 1 month of surgery Previous sternotomy
28 (85%) 19 (58%) 8 (24%) 4 (12%) 10 (30%) 8 (24%) 2 (6%) 6 (18%) 0 1 (3%) 7 (21%)
15 (83%) 14 (78%) 2 (11%) 1 (6%) 6 (33%) 3 (17%) 2 (11%) 4 (22%) 0 0 4 (22%)
13 (87%) 5 (33%) 6 (40%) 3 (20%) 4 (27%) 5 (33%) 0 2 (13%) 0 1 (7%) 3 (20%)
0.59 0.015 0.10 0.23 0.72 0.24 0.49 0.42 1 0.46 1
Pre-operative variables Haematocrit (%) Serum creatinine (mg/dL) Pre-operative hospital stay (days) STS score Antibiotics within 30 days prior to surgery
40.5 (38–43.7) 1 (0.8–1.2) 4 (1–9) 9 (6.75–11.25) 3 (9%)
40.4 (37.2–43.7) 1 (0.8–1.2) 5 (1–9) 8 (7–11) 1 (6%)
40.5 (38.3–44.55) 0.9 (0.8–1.1) 3 (1.5–7) 10 (6.5–11.5) 2 (13%)
0.71 0.80 0.67 0.90 0.58
PS-CRE, polymyxin-susceptible CRE; PR-CRE, polymyxin-resistant CRE; BMI, body mass index; COPD, chronic obstructive pulmonary disease; NYHA, New York Heart Association; STS, Society of Thoracic Surgeons. a Values are n (%) for categorical data or median (interquartile range) for continuous variables.
excluded due to insufficient clinical or microbiological information, leaving a total of 33 patients infected with CRE. PR-CRE were detected in 15 patients (45%). Overall, the majority of patients were White (76%) and more were women (58%). The mean patient age was 63 years [interquartile range (IQR) 56–72 years]. Previous medical conditions included hypertension, diabetes, peripheral vascular diseases and myocardial infarction. More than one-third of the patients were obese (BMI ≥30). A small proportion of patients received antibiotics during the 30 days prior to the index surgery. The median hospital stay before index surgery was 4 days (IQR 1–9 days). The median STS score was 9 (IQR 6.75–11.25). Diabetes mellitus was observed in 19 patients (58%) and was more common in patients infected with PS-CRE than in those infected with PR-CRE (78% vs. 33%; P = 0.015) (Table 1). No other significant differences were detected at baseline between the two groups (i.e. PS-CRE versus PR-CRE). All patients received preoperative body washes with chlorhexidine in addition to antibiotic prophylaxis with vancomycin in combination with a second- or third-generation cephalosporin. The most common surgical procedures were coronary artery bypass grafting (70%) and/or valve replacement (36%). The median duration of surgery was 5.4 h (IQR 4.3–6 h). Nine patients (27%) required insulin within the first 24 h after surgery. In general, patients infected with PR-CRE were comparable in the surgical or post-operative variables with those infected with PS-CRE. All 33 patients with CRE had monomicrobial infections, most commonly due to K. pneumoniae (n = 20), followed by Enterobacter aerogenes (n = 10) and Enterobacter cloacae (n = 3). Clinical samples obtained for microbiological diagnosis included purulent discharge from the mediastinum (100%) and associated blood cultures (15%) and culture of tissue or bone (6%). Enterobacter spp. displayed polymyxin resistance more frequently than K. pneumoniae (77% vs. 25%; P < 0.01). PFGE analysis revealed two clusters of K. pneumoniae: cluster A comprising 19 strains (95%); and cluster B with just one strain (5%). MLST analysis from the index case of clusters A and
B identified two distinct sequence types (ST437 and ST16, respectively). eBURST analysis (http://eburst.mlst.net) revealed that both clones (ST437 and ST16) belonged to the same clonal complex (CC258). PFGE also revealed one cluster of E. aerogenes and another cluster of E. cloacae with 100% similarity for each species, respectively. Among all of the resistance genes explored, blaKPC was the only gene detected and was present in all CRE strains (n = 33). This epidemiology was consistent with the recently reported outbreak of CRE pathogens [17]. Regular surveillance cultures for detection of CRE were systematically implemented after 22 April 2013. Twenty patients (61%) were colonised with CRE before any signs or symptoms of mediastinitis were seen. These colonising CRE strains were detected either from standard surveillance cultures of the rectum and/or inguinal area (n = 12) or from other samples in patients not considered to be infected (n = 8). The emergence of PR-CRE strains was partially controlled at this site by adopting infection control policies and restricting the use of polymyxins (unpublished data). All of the patients with mediastinitis due to CRE had fever (ranging from low-grade 37.8 to 39 ◦ C), chest pain, sternal purulent discharge and wound dehiscence. Most patients (85%) had sternal instability. Although the median C-reactive protein value was high (8.7 mg/dL; reference value, <1.0 mg/dL), the white blood cell count was <10 000 cells/mm3 in ca. 40% of the patients. Surgical debridement was performed in 30 patients (91%) for a total of 40 procedures (mean 1.3 procedures per patient). Among those patients who underwent a surgical procedure, primary closure was carried out in 18 patients (60%). Three patients did not undergo surgery for treatment of their mediastinitis: one of the patients died before the surgical procedure and the two other patients with sternal stability responded well to antibiotics and hyperbaric oxygen therapy. Hyperbaric oxygen therapy was utilised for a total of 10 days as adjunct therapy in six patients infected with PS-CRE who were clinically stable and in whom primary sternal closure was not feasible. Regarding directed antibiotic treatment, 70% of patients received at least one agent considered active in vitro against the infecting micro-organism, and most patients received at least three
Please cite this article in press as: Abboud CS, et al. Post-surgical mediastinitis due to carbapenem-resistant Enterobacteriaceae: Clinical, epidemiological and survival characteristics. Int J Antimicrob Agents (2016), http://dx.doi.org/10.1016/j.ijantimicag.2016.02.015
G Model ANTAGE-4775; No. of Pages 5
ARTICLE IN PRESS C.S. Abboud et al. / International Journal of Antimicrobial Agents xxx (2016) xxx–xxx
4
Fig. 1. Kaplan–Meier curve comparing overall survival stratified by polymyxin susceptibility in patients with mediastinitis due to carbapenem-resistant Enterobacteriaceae (CRE). PS-CRE, polymyxin-susceptible CRE; PR-CRE, polymyxin-resistant CRE.
concomitant antibiotics. There were no detectable differences in the antibiotics or their combinations used between the two groups. Carbapenems (either meropenem or imipenem) or polymyxin B were administered in the majority of patients (81% and 90%, respectively). Polymyxin B in combination with a carbapenem was the most common antibiotic combination (73%). This combination was used in 40% of patients infected with PR-CRE. Aminoglycosides (amikacin or gentamicin) were also frequently used (67%). Of the 33 patients, 6 (18%) had chronic renal failure before the surgical procedure, including 4 with PS-CRE and 2 with PR-CRE. All patients maintained their basal clearance condition during antimicrobial treatment period, except for one patient in the PS-CRE group who was submitted to haemodialysis and later died. Only four patients were submitted to haemodialysis during the antimicrobial treatment period for CRE infection and all of them died, including 3 in the PR-CRE group and 1 in the PS-CRE group. The following items were evaluated and showed no difference between the PSCRE and PR-CRE groups: number of procedures; kind of closure (primary/secondary); and flap use (four PS-CRE and 2 PR-CRE). Eleven patients (33%) died and all of the surviving patients were clinically cured. Compared with patients infected with PS-CRE, in-hospital mortality was higher for patients infected with PRCRE (17% vs. 53%; P = 0.06). Compared with patients with PS-CRE, patients infected with PR-CRE who received polymyxin-containing regimens had lower cure rates and survival (27% vs. 83%; P < 0.01). Overall, patients infected with PR-CRE tended to have a lower survival than PS-CRE patients (P = 0.06) (Fig. 1). Moreover, PRCRE-infected patients tended to have a three times higher rate of mortality (HR = 3.38, 95% CI 0.87–13.11; P = 0.08). In the multivariate model adjusting for diabetes, patients with PR-CRE had a higher rate of mortality, although this was not statistically significant (HR = 2.88, 95% CI 0.69–11.99; P = 0.14).
4. Discussion To the best of our knowledge, this is the first study describing a cohort of patients with post-surgical mediastinitis due to CRE. Several results provided by this investigation have been highlighted here.
In centres with a high volume of cardiothoracic surgery and CRE outbreaks, a significant proportion of post-surgical mediastinitis can be caused by CRE. In this study, approximately one-third of post-surgical mediastinitis were due to CRE strains. Importantly, the rate of mediastinitis reported in this study (1.6%) is in agreement with the current literature [4]. The rate of prior colonisation detected among the patients was particularly high for PS-CRE and may have contributed to the development of the infection. Such colonisation was concomitant with a dramatic expansion of CRE strains (PS-CRE first and PR-CRE later) at this single centre. This observation suggests that CRE strains are likely to colonise before establishing an infection. Colonisation with KPC-producing K. pneumoniae has been suggested as a risk factor for subsequent infections [18,19], including for those strains displaying polymyxin resistance [11]. In fact, because colonisation was not systematically searched for in all of our patients, the real rates of colonisation may have been even higher than those reported here. In this scenario of an increasing number of CRE infections and narrow clonal distribution, infection control measures become crucial to limit the expansion of CRE pathogens [20]. As mentioned previously, we have reduced the rates of PR-CRE infections with measures such as surveillance screening, contact precautions, and antibiotic stewardship restricting the empirical use of polymyxins. Overall in-hospital mortality associated with CRE mediastinitis was high (33%). This rate is in agreement with current studies describing Gram-negative mediastinitis [5] and bacteraemia due to CRE pathogens [6,7]. However, in-hospital mortality for general mediastinitis within our institution is 10% (unpublished data, available under request). Thus, the general CRE mediastinitis in-hospital mortality in the present case series was higher and possibly driven by PR-CRE patients (>50% mortality). The current case series had a high prevalence of diabetes mellitus among the general patient population (58%), although PS-CRE patients had an even higher prevalence of the condition (78%). Diabetes is a highly prevalent condition in, and a risk factor for, general mediastinitis patients. A higher mortality was found in the PR-CRE population adjusted for diabetes (HR = 2.88, 95% CI 0.69–11.99), however this difference was not significant (P = 0.14). The non-significance of this higher mortality in the adjusted multivariate analysis for diabetes was probably due to the limited number of patients and the unequal distribution of the condition among both groups. Given the limited therapeutic options currently available, it is not uncommon for the continued use of polymyxins combined with several other agents for the treatment of these patients. Until more data on newer effective antimicrobial agents or optimised antimicrobial combinations effective against these PR-CRE strains are available, this question will remain. Because mediastinitis is a surgical disease, the role of surgery needs to be particularly highlighted for these difficult-to-treat infections. An aggressive surgical approach should include wide debridement of necrotic tissue, sectioning of compromised bone and cartilage, washing and sternum re-suturing [21] whenever possible. Unfortunately, negative pressure devices (vacuum-assisted closure) were not available at the site during the study period [22]. Whilst the role of hyperbaric oxygen is still to be determined, it was mainly used as an adjunctive therapy for PS-CRE-infected patients who were clinically stable and in whom primary sternal closure was not possible. This study has several limitations. First, it was retrospective and was conducted at a single centre. It is important to mention that most clinical and microbiological data were systematically collected from patients with mediastinitis at this site. Despite coming from a single centre, the number of cases is clinically relevant for a meaningful description. Second, the relatedness (e.g. by PFGE) between the clones causing the infection and those previously colonising could not be determined. As shown in a recent study,
Please cite this article in press as: Abboud CS, et al. Post-surgical mediastinitis due to carbapenem-resistant Enterobacteriaceae: Clinical, epidemiological and survival characteristics. Int J Antimicrob Agents (2016), http://dx.doi.org/10.1016/j.ijantimicag.2016.02.015
G Model ANTAGE-4775; No. of Pages 5
ARTICLE IN PRESS C.S. Abboud et al. / International Journal of Antimicrobial Agents xxx (2016) xxx–xxx
it is very likely that in a high proportion of patients the colonising clones match those causing the infection [19]. Third, the MICs for carbapenems and polymyxins were not determined by a confirmation methodology such as microdilution. Therefore, we could not explore further any potential relationship between the antibiotics used and achievable MICs. Finally, the study is limited by the small sample size, which makes it difficult to perform meaningful inferential analyses. The current findings show that patients with mediastinitis due to PR-CRE have a higher risk of overall mortality. Owing to the limited sample size, we were not able to achieve statistical significance; however, mediastinitis and PR-CRE are relatively rare and we believe that this is the largest discussion of patients with this relevant hospital-acquired infection. Despite all of these limitations, we believe that this study provides clinicians and surgeons with useful clinical and microbiological information to better understand post-surgical mediastinitis due to these difficult-to-treat pathogens. 5. Conclusion Mediastinitis due to CRE has become an emerging infection in this centre. Antibacterial treatments are usually based on carbapenems, frequently accompanied by polymyxin plus a third agent. Despite complex antibiotic treatments and aggressive surgical procedures, these patients have a high mortality, particularly those infected with PR-CRE. Centres with increasing rates of CRE and with a high cardiac surgery volume should be aware of these potentially devastating infections. Acknowledgements The authors acknowledge Celia Harumi Hiroshi (secretary) and LEE Laboratory of Epidemiology and Statistics–Instituto Dante Pazzanese de Cardiologia (São Paulo, Brazil). Funding: This work was funded by a grant [2012/12108-3] from the Fundac¸ão de Amparo à Pesquisa do Estado de São Paulo (FAPESP). Competing interests: MES is or has been a consultant to Achaogen, Astellas, Cempra, Cerexa, Furiex, Nabriva, PRI, Trius and Theravance. All other authors declare no competing interests. Ethical approval: The study was approved by the Ethical Committee of Instituto Dante Pazzanese de Cardiologia. References [1] Abboud CS, Wey SB, Baltar VT. Risk factors for mediastinitis after cardiac surgery. Ann Thorac Surg 2004;77:676–83. [2] Filsoufi F, Castillo JG, Rahmanian PB, Broumand SR, Silvay G, Carpentier A, et al. Epidemiology of deep sternal wound infection in cardiac surgery. J Cardiothorac Vasc Anesth 2009;23:488–94. [3] Wang FD, Chang CH. Risk factors of deep sternal wound infections in coronary artery bypass graft surgery. J Cardiovasc Surg (Torino) 2000;41:709–13.
5
[4] Fowler Jr VG, O’Brien SM, Muhlbaier LH, Corey GR, Ferguson TB, Peterson ED. Clinical predictors of major infections after cardiac surgery. Circulation 2005;112(9 Suppl.):I358–65. [5] Charbonneau H, Maillet JM, Faron M, Mangin O, Puymirat E, Le Besnerais P, et al. Mediastinitis due to Gram-negative bacteria is associated with increased mortality. Clin Microbiol Infect 2014;20:O197–202. [6] Daikos GL, Tsaousi S, Tzouvelekis LS, Anyfantis I, Psichogiou M, Argyropoulou A, et al. Carbapenemase-producing Klebsiella pneumoniae bloodstream infections: lowering mortality by antibiotic combination schemes and the role of carbapenems. Antimicrob Agents Chemother 2014;58: 2322–8. [7] Tumbarello M, Viale P, Viscoli C, Trecarichi EM, Tumietto F, Marchese A, et al. Predictors of mortality in bloodstream infections caused by Klebsiella pneumoniae carbapenemase-producing K. pneumoniae: importance of combination therapy. Clin Infect Dis 2012;55:943–50. [8] Tumbarello M, Trecarichi EM, De Rosa FG, Giannella M, Giacobbe DR, Bassetti M, et al. Infections caused by KPC-producing Klebsiella pneumoniae: differences in therapy and mortality in a multicentre study. J Antimicrob Chemother 2015;70:2133–43. [9] Mammina C, Bonura C, Di Bernardo F, Aleo A, Fasciana T, Sodano C, et al. Ongoing spread of colistin-resistant Klebsiella pneumoniae in different wards of an acute general hospital, Italy, June to December 2011. Euro Surveill 2012:17, pii: 20248. [10] Marchaim D, Chopra T, Pogue JM, Perez F, Hujer AM, Rudin S, et al. Outbreak of colistin-resistant, carbapenem-resistant Klebsiella pneumoniae in metropolitan Detroit, Michigan. Antimicrob Agents Chemother 2011;55:593–9. [11] Giacobbe DR, Del Bono V, Trecarichi EM, De Rosa FG, Giannella M, Bassetti M, et al. Risk factors for bloodstream infections due to colistinresistant KPC-producing Klebsiella pneumoniae: results from a multicenter case–control–control study. Clin Microbiol Infect 2015;21:e1–8. [12] 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;36:309–32. [13] Gaca JG, Sheng S, Daneshmand MA, O’Brien S, Rankin JS, Brennan JM, et al. Outcomes for endocarditis surgery in North America: a simplified risk scoring system. J Thorac Cardiovasc Surg 2011;141:98–106, e1–2. [14] Monteiro J, Widen RH, Pignatari AC, Kubasek C, Silbert S. Rapid detection of carbapenemase genes by multiplex real-time PCR. J Antimicrob Chemother 2012;67:906–9. [15] Diancourt L, Passet V, Verhoef J, Grimont PA, Brisse S. Multilocus sequence typing of Klebsiella pneumoniae nosocomial isolates. J Clin Microbiol 2005;43:4178–82. [16] Egito JG, Abboud CS, Oliveira AP, Maximo CA, Montenegro CM, Amato VL, et al. Clinical evolution of mediastinitis in patients undergoing adjuvant hyperbaric oxygen therapy after coronary artery bypass surgery. Einstein (Sao Paulo) 2013;11:345–9. [17] Abboud CS, Monteiro J, Franca JI, Pignatari AC, De Souza EE, Camargo EC, et al. A space–time model for carbapenemase-producing Klebsiella pneumoniae (KPC) cluster quantification in a high-complexity hospital. Epidemiol Infect 2015;143:2648–52. [18] Hayden MK, Lin MY, Lolans K, Weiner S, Blom D, Moore NM, et al. Prevention of colonization and infection by Klebsiella pneumoniae carbapenemaseproducing Enterobacteriaceae in long-term acute-care hospitals. Clin Infect Dis 2015;60:1153–61. [19] Ducomble T, Faucheux S, Helbig U, Kaisers UX, Konig B, Knaust A, et al. Large hospital outbreak of KPC-2-producing Klebsiella pneumoniae: investigating mortality and the impact of screening for KPC-2 with polymerase chain reaction. J Hosp Infect 2015;89:179–85. [20] Carmeli Y, Akova M, Cornaglia G, Daikos GL, Garau J, Harbarth S, et al. Controlling the spread of carbapenemase-producing Gram-negatives: therapeutic approach and infection control. Clin Microbiol Infect 2010;16: 102–11. [21] Robicsek F, Daugherty HK, Cook JW. The prevention and treatment of sternum separation following open-heart surgery. J Thorac Cardiovasc Surg 1977;73:267–8. [22] Yu AW, Rippel RA, Smock E, Jarral OA. In patients with post-sternotomy mediastinitis is vacuum-assisted closure superior to conventional therapy? Interact Cardiovasc Thorac Surg 2013;17:861–5.
Please cite this article in press as: Abboud CS, et al. Post-surgical mediastinitis due to carbapenem-resistant Enterobacteriaceae: Clinical, epidemiological and survival characteristics. Int J Antimicrob Agents (2016), http://dx.doi.org/10.1016/j.ijantimicag.2016.02.015
ARTICLE IN PRESS
ANTAGE-4775; No. of Pages 5
International Journal of Antimicrobial Agents xxx (2016) xxx–xxx
Contents lists available at ScienceDirect
International Journal of Antimicrobial Agents journal homepage: http://www.elsevier.com/locate/ijantimicag
Post-surgical mediastinitis due to carbapenem-resistant Enterobacteriaceae: Clinical, epidemiological and survival characteristics夽 C.S. Abboud a,∗ , J. Monteiro b , M.E. Stryjewski c , E.C. Zandonadi a , V. Barbosa a , D. Dantas a , E.E. Sousa a , M.J. Fonseca d , D.M. Jacobs e , A.C. Pignatari d , C. Kiffer d , G.G. Rao e a
Instituto Dante Pazzanese de Cardiologia, São Paulo, Brazil Associac¸ão Fundo de Incentivo à Pesquisa (AFIP–Medicina Diagnóstica), São Paulo, Brazil c Department of Medicine, Division of Infectious Diseases, Centro de Educación Médica e Investigaciones Clínicas ‘Norberto Quirno’ (CEMIC), Buenos Aires, Argentina d Laboratório Especial de Microbiologia Clínica (LEMC), Universidade Federal de São Paulo (UNIFESP), Escola Paulista de Medicina (EPM), São Paulo, Brazil e University at Buffalo School of Pharmacy and Pharmaceutical Sciences, Buffalo, NY, USA b
a r t i c l e
i n f o
Article history: Received 12 December 2015 Accepted 27 February 2016 Keywords: Mediastinitis Post-surgical infection Carbapenem-resistant Enterobacteriaceae Polymyxin resistance
a b s t r a c t Invasive infections due to carbapenem-resistant Enterobacteriaceae (CRE), including polymyxin-resistant (PR-CRE) strains, are being increasingly reported. However, there is a lack of clinical data for several lifethreatening infections. Here we describe a cohort of patients with post-surgical mediastinitis due to CRE, including PR-CRE. This study was a retrospective cohort design at a single cardiology centre. Patients with mediastinitis due to CRE were identified and were investigated for clinically relevant variables. Infecting isolates were studied using molecular techniques. Patients infected with polymyxin-susceptible CRE (PS-CRE) strains were compared with those infected with PR-CRE strains. In total, 33 patients with CRE mediastinitis were studied, including 15 patients (45%) with PR-CRE. The majority (61%) were previously colonised. All infecting isolates carried blaKPC genes. Baseline characteristics of patients with PR-CRE mediastinitis were comparable with those with PS-CRE mediastinitis. Of the patients studied, 70% received at least one agent considered active in vitro and most patients received at least three concomitant antibiotics. Carbapenem plus polymyxin B was the most common antibiotic combination (73%). Over 90% of patients underwent surgical debridement. Overall, in-hospital mortality was 33% and tended to be higher in patients infected with PR-CRE (17% vs. 53%; P = 0.06). In conclusion, mediastinitis due to CRE, including PR-CRE, can become a significant challenge in centres with CRE and a high cardiac surgery volume. Despite complex antibiotic treatments and aggressive surgical procedures, these patients have a high mortality, particularly those infected with PR-CRE. © 2016 Published by Elsevier B.V.
1. Introduction Post-sternotomy mediastinitis is a life-threatening infection that occurs in a small proportion (0.5–2.7%) of patients undergoing cardiothoracic surgery [1–4], resulting in longer hospital stay, higher medical costs, additional surgical procedures and, finally,
夽 This study was partially presented at IDWeek 2014, 8–12 October 2014, Philadelphia, PA [abstract no. 487]. ∗ Corresponding author. Present address: Division of Infectious Diseases, Instituto Dante Pazzanese de Cardiologia, Av. Dr Dante Pazzanese, 500, Vila Mariana, São Paulo, SP 04012-909, Brazil. Tel.: +55 11 5085 6255; fax: +55 1150856255. E-mail address: [email protected] (C.S. Abboud).
increased mortality. Recent studies showed that 14–32% of patients who developed post-surgical mediastinitis died, indicative of a higher risk of morbidity and mortality compared with patients who do not experience infectious complications [1–5]. Although the main micro-organisms implicated are Gram-positive bacteria (e.g. staphylococci), Gram-negative pathogens can also be a relevant cause of mediastinitis at some centres [5]. Importantly, compared with other micro-organisms, mediastinitis associated with Gramnegative pathogens was shown to have higher rates of in-hospital death [5]. Invasive infections associated with carbapenem-resistant Enterobacteriaceae (CRE) pose a serious challenge to clinicians. Treatment with drugs such as polymyxins (polymyxin B or colistin), carbapenems or their combinations are basically supported
http://dx.doi.org/10.1016/j.ijantimicag.2016.02.015 0924-8579/© 2016 Published by Elsevier B.V.
Please cite this article in press as: Abboud CS, et al. Post-surgical mediastinitis due to carbapenem-resistant Enterobacteriaceae: Clinical, epidemiological and survival characteristics. Int J Antimicrob Agents (2016), http://dx.doi.org/10.1016/j.ijantimicag.2016.02.015
G Model ANTAGE-4775; No. of Pages 5
ARTICLE IN PRESS C.S. Abboud et al. / International Journal of Antimicrobial Agents xxx (2016) xxx–xxx
2
by observational studies [6–8]. The emergence of resistance to polymyxins further complicates the situation [9–11]. Whilst there is an increasing body of evidence on the clinical characteristics of patients with more common CRE infections (e.g. bloodstream infections) [6,7], we are faced with a lack of clinical information in the case of less frequent but otherwise severe CRE infections such as post-surgical mediastinitis. To the best of our knowledge, this is the first study describing a cohort of patients with post-surgical mediastinitis due to CRE, including patients infected with strains resistant to polymyxins. Given the lack of evidence described above and the potential occurrence of polymyxin resistance in CRE, this study intended to describe the clinical characteristics of a significant post-sternotomy mediastinitis CRE outbreak population and to compare survival outcomes among polymyxin-resistant and -susceptible isolates. 2. Materials and methods
meropenem) resistance was defined as a minimum inhibitory concentration (MIC) of ≥4 g/mL according to Clinical and Laboratory Standards Institute (CLSI) breakpoints. Resistance to polymyxins was defined by a colistin MIC ≥4 g/mL. Molecular testing was conducted at Laboratório Especial de Microbiologia Clínica (LEMC) of Universidade Federal de São Paulo (UNIFESP) (São Paulo, Brazil) and included a real-time multiplex PCR assay for blaKPC , blaGES , blaOXA-48-like , blaNDM , blaIMP and blaVIM carbapenemase genes [14]. The genetic relatedness for each group of CRE clinical strains was evaluated by pulsed-field gel electrophoresis (PFGE) using the restriction enzyme SpeI (New England Biolabs Inc., Ipswich, MA) and the results were analysed using BioNumerics v.6.6 (Applied Maths, Sint-Martens-Latem, Belgium). Multilocus sequence typing (MLST) for Klebsiella pneumoniae strains was performed as described by Diancourt et al. [15], and the Institut Pasteur MLST database was used to assign allelic numbers and sequence types (http://bigsdb.web.pasteur.fr/klebsiella/klebsiella. html; accessed 11 November 2015).
2.1. Study population and setting 2.3. Antibiotic dosing This retrospective cohort study was conducted at Instituto Dante Pazzanese de Cardiologia (São Paulo, Brazil) between December 2010 and June 2014. This is a 350-bed hospital specialising in cardiology and cardiovascular surgery that serves as a referral centre for 300 000 patients, 90% of whom are from São Paulo. Patients undergoing surgical procedures at the institution are followed by the institutional teams until the patient recovers or is clinically stable, allowing for external follow-up. For the purpose of this study, patients aged ≥18 years diagnosed with mediastinitis due to CRE were identified. Mediastinitis was defined as the presence of at least one of the following: (i) organisms cultured from mediastinal tissue or fluid obtained during a surgical operation or needle aspiration; (ii) evidence of mediastinitis seen during a surgical operation or histopathological examination; and (iii) fever (>38 ◦ C), chest pain or sternal instability, and at least one of the following: purulent discharge from mediastinal area, organisms cultured from blood or discharge from mediastinal area, or mediastinal widening on the chest radiograph [12]. Patients were considered to be infected with CRE if such isolates were obtained from either blood cultures, surgical specimens or by needle aspiration from the sternal area and/or mediastinum. For the duration of the study, all patients with mediastinitis were under the care of the same infectious diseases and cardiovascular surgical staff members. Patient charts were reviewed to capture both demographic and clinically relevant data, including co-morbidities, prior cardiovascular procedures, immunosuppressant drugs, previous infections, colonisation and prior antibiotic use. The Society of Thoracic Surgeons (STS) score, indicative of operative mortality and morbidity of adult cardiac surgery, was calculated [13]. Surgical information, such as type and duration of the procedure, blood loss, perfusion time, extracorporeal circulation and re-operation, was also collected. All relevant information regarding antibiotic treatments was reviewed. This study was approved by the Ethical Committee of Instituto Dante Pazzanese de Cardiologia.
The following intravenous [16] antimicrobial doses were used at the institution for the treatment of patients with CRE mediastinitis, according to the antimicrobial stewardship programme of the hospital: polymyxin B, 25 000 UI/kg/day; amikacin, 15 mg/kg/day; gentamicin, 5 mg/kg/day; meropenem, 1 g every 8 h; imipenem, 500 mg every 6 h; tigecycline, 100 mg initial dose followed by 50 mg every 12 h; and ciprofloxacin, 400 mg every 12 h. The doses were adjusted according to the patient’s renal function. In the case of patients with a body mass index (BMI) >40 (kg/m2 ), the dose of meropenem was increased to 6 g day (2 g every 8 h). 2.4. Statistics analysis Each subject was included only once (index infection) in the analysis. Descriptive statistics were used to describe the overall characteristics of the whole cohort. Patients infected with polymyxin-resistant CRE (PR-CRE) strains were compared with those infected with polymyxin-susceptible CRE (PS-CRE) strains using either Fisher’s exact test or 2 test as appropriate for categorical variables, and Wilcoxon rank-sum test for continuous variables. Patients were followed during their hospital stay after the diagnosis of mediastinitis was made, and charts were reviewed up to October 2014 (3 months after the end of the study period) to determine the survival status following discharge. If the information was not available in the chart, it was retrieved by telephone contact. A survival analysis to compare survival between groups was performed using the log-rank test. The effect of PR-CRE on mortality was evaluated with a Cox proportional hazard model to determine the hazard ratio (HR) and 95% confidence interval (CI). A multivariate model was performed to take into account differences between the PS-CRE and PR-CRE groups. Significance was defined as a P-value of ≤0.05 (two-tailed). IBM SPSS Statistics for Windows v.19.0 software package (IBM Corp., Armonk, NY) was utilised for the statistical analyses.
2.2. Microbiology 3. Results All Gram-negative strains recovered from patients with mediastinitis were submitted to the local microbiology laboratory for identification and determination of antimicrobial susceptibility profiles by VITEK® 2 system using AST-N238 and AST-N239 cards (bioMérieux, Marcy-l’Étoile, France). Based on the results of the routine antimicrobial susceptibility testing and phenotypic tests (VITEK® 2 system and modified Hodge test), the clinical isolates were classified as CRE. Carbapenem (imipenem and
From December 2010 to June 2014, 5933 cardiac surgeries were performed at this single institution. During this period 92 patients (1.6%) were diagnosed with mediastinitis, among which Grampositive and Gram-negative pathogens were isolated from 45% and 55% of the patients, respectively. Enterobacteriaceae were isolated in 88% of patients infected with Gram-negative pathogens, including 36 patients infected with CRE strains. Three patients were
Please cite this article in press as: Abboud CS, et al. Post-surgical mediastinitis due to carbapenem-resistant Enterobacteriaceae: Clinical, epidemiological and survival characteristics. Int J Antimicrob Agents (2016), http://dx.doi.org/10.1016/j.ijantimicag.2016.02.015
G Model
ARTICLE IN PRESS
ANTAGE-4775; No. of Pages 5
C.S. Abboud et al. / International Journal of Antimicrobial Agents xxx (2016) xxx–xxx
3
Table 1 Baseline demographic and clinical characteristics of patients with mediastinitis and deep sternal wounds due to carbapenem-resistant Enterobacteriaceae (CRE), stratified by polymyxin resistance.a Total CRE (n = 33)
PS-CRE (n = 18)
PR-CRE (n = 15)
P-value
Patient demographics Female sex Age (years) Race White Weight (kg) BMI (kg/m2 ) BMI ≥30
19 (58%) 63 (56–72) 25 (76%) 75 (64–83) 28.8 (25.3–31.6) 12 (36%)
12 (67%) 65 (59–72) 14 (78%) 74 (64–84) 28.8 (26.2–32.4) 5 (28%)
7 (47%) 56 (53.5–69.5) 11 (73%) 77 (64–83) 29.6 (24.5–31.6) 7 (47%)
0.30 0.24 1 0.95 0.60 0.31
Medical history Hypertension Diabetes Smoking history (current) COPD Peripheral vascular disease Myocardial infarction Heart failure (NYHA-IV) Chronic renal failure Dialysis Steroids within 1 month of surgery Previous sternotomy
28 (85%) 19 (58%) 8 (24%) 4 (12%) 10 (30%) 8 (24%) 2 (6%) 6 (18%) 0 1 (3%) 7 (21%)
15 (83%) 14 (78%) 2 (11%) 1 (6%) 6 (33%) 3 (17%) 2 (11%) 4 (22%) 0 0 4 (22%)
13 (87%) 5 (33%) 6 (40%) 3 (20%) 4 (27%) 5 (33%) 0 2 (13%) 0 1 (7%) 3 (20%)
0.59 0.015 0.10 0.23 0.72 0.24 0.49 0.42 1 0.46 1
Pre-operative variables Haematocrit (%) Serum creatinine (mg/dL) Pre-operative hospital stay (days) STS score Antibiotics within 30 days prior to surgery
40.5 (38–43.7) 1 (0.8–1.2) 4 (1–9) 9 (6.75–11.25) 3 (9%)
40.4 (37.2–43.7) 1 (0.8–1.2) 5 (1–9) 8 (7–11) 1 (6%)
40.5 (38.3–44.55) 0.9 (0.8–1.1) 3 (1.5–7) 10 (6.5–11.5) 2 (13%)
0.71 0.80 0.67 0.90 0.58
PS-CRE, polymyxin-susceptible CRE; PR-CRE, polymyxin-resistant CRE; BMI, body mass index; COPD, chronic obstructive pulmonary disease; NYHA, New York Heart Association; STS, Society of Thoracic Surgeons. a Values are n (%) for categorical data or median (interquartile range) for continuous variables.
excluded due to insufficient clinical or microbiological information, leaving a total of 33 patients infected with CRE. PR-CRE were detected in 15 patients (45%). Overall, the majority of patients were White (76%) and more were women (58%). The mean patient age was 63 years [interquartile range (IQR) 56–72 years]. Previous medical conditions included hypertension, diabetes, peripheral vascular diseases and myocardial infarction. More than one-third of the patients were obese (BMI ≥30). A small proportion of patients received antibiotics during the 30 days prior to the index surgery. The median hospital stay before index surgery was 4 days (IQR 1–9 days). The median STS score was 9 (IQR 6.75–11.25). Diabetes mellitus was observed in 19 patients (58%) and was more common in patients infected with PS-CRE than in those infected with PR-CRE (78% vs. 33%; P = 0.015) (Table 1). No other significant differences were detected at baseline between the two groups (i.e. PS-CRE versus PR-CRE). All patients received preoperative body washes with chlorhexidine in addition to antibiotic prophylaxis with vancomycin in combination with a second- or third-generation cephalosporin. The most common surgical procedures were coronary artery bypass grafting (70%) and/or valve replacement (36%). The median duration of surgery was 5.4 h (IQR 4.3–6 h). Nine patients (27%) required insulin within the first 24 h after surgery. In general, patients infected with PR-CRE were comparable in the surgical or post-operative variables with those infected with PS-CRE. All 33 patients with CRE had monomicrobial infections, most commonly due to K. pneumoniae (n = 20), followed by Enterobacter aerogenes (n = 10) and Enterobacter cloacae (n = 3). Clinical samples obtained for microbiological diagnosis included purulent discharge from the mediastinum (100%) and associated blood cultures (15%) and culture of tissue or bone (6%). Enterobacter spp. displayed polymyxin resistance more frequently than K. pneumoniae (77% vs. 25%; P < 0.01). PFGE analysis revealed two clusters of K. pneumoniae: cluster A comprising 19 strains (95%); and cluster B with just one strain (5%). MLST analysis from the index case of clusters A and
B identified two distinct sequence types (ST437 and ST16, respectively). eBURST analysis (http://eburst.mlst.net) revealed that both clones (ST437 and ST16) belonged to the same clonal complex (CC258). PFGE also revealed one cluster of E. aerogenes and another cluster of E. cloacae with 100% similarity for each species, respectively. Among all of the resistance genes explored, blaKPC was the only gene detected and was present in all CRE strains (n = 33). This epidemiology was consistent with the recently reported outbreak of CRE pathogens [17]. Regular surveillance cultures for detection of CRE were systematically implemented after 22 April 2013. Twenty patients (61%) were colonised with CRE before any signs or symptoms of mediastinitis were seen. These colonising CRE strains were detected either from standard surveillance cultures of the rectum and/or inguinal area (n = 12) or from other samples in patients not considered to be infected (n = 8). The emergence of PR-CRE strains was partially controlled at this site by adopting infection control policies and restricting the use of polymyxins (unpublished data). All of the patients with mediastinitis due to CRE had fever (ranging from low-grade 37.8 to 39 ◦ C), chest pain, sternal purulent discharge and wound dehiscence. Most patients (85%) had sternal instability. Although the median C-reactive protein value was high (8.7 mg/dL; reference value, <1.0 mg/dL), the white blood cell count was <10 000 cells/mm3 in ca. 40% of the patients. Surgical debridement was performed in 30 patients (91%) for a total of 40 procedures (mean 1.3 procedures per patient). Among those patients who underwent a surgical procedure, primary closure was carried out in 18 patients (60%). Three patients did not undergo surgery for treatment of their mediastinitis: one of the patients died before the surgical procedure and the two other patients with sternal stability responded well to antibiotics and hyperbaric oxygen therapy. Hyperbaric oxygen therapy was utilised for a total of 10 days as adjunct therapy in six patients infected with PS-CRE who were clinically stable and in whom primary sternal closure was not feasible. Regarding directed antibiotic treatment, 70% of patients received at least one agent considered active in vitro against the infecting micro-organism, and most patients received at least three
Please cite this article in press as: Abboud CS, et al. Post-surgical mediastinitis due to carbapenem-resistant Enterobacteriaceae: Clinical, epidemiological and survival characteristics. Int J Antimicrob Agents (2016), http://dx.doi.org/10.1016/j.ijantimicag.2016.02.015
G Model ANTAGE-4775; No. of Pages 5
ARTICLE IN PRESS C.S. Abboud et al. / International Journal of Antimicrobial Agents xxx (2016) xxx–xxx
4
Fig. 1. Kaplan–Meier curve comparing overall survival stratified by polymyxin susceptibility in patients with mediastinitis due to carbapenem-resistant Enterobacteriaceae (CRE). PS-CRE, polymyxin-susceptible CRE; PR-CRE, polymyxin-resistant CRE.
concomitant antibiotics. There were no detectable differences in the antibiotics or their combinations used between the two groups. Carbapenems (either meropenem or imipenem) or polymyxin B were administered in the majority of patients (81% and 90%, respectively). Polymyxin B in combination with a carbapenem was the most common antibiotic combination (73%). This combination was used in 40% of patients infected with PR-CRE. Aminoglycosides (amikacin or gentamicin) were also frequently used (67%). Of the 33 patients, 6 (18%) had chronic renal failure before the surgical procedure, including 4 with PS-CRE and 2 with PR-CRE. All patients maintained their basal clearance condition during antimicrobial treatment period, except for one patient in the PS-CRE group who was submitted to haemodialysis and later died. Only four patients were submitted to haemodialysis during the antimicrobial treatment period for CRE infection and all of them died, including 3 in the PR-CRE group and 1 in the PS-CRE group. The following items were evaluated and showed no difference between the PSCRE and PR-CRE groups: number of procedures; kind of closure (primary/secondary); and flap use (four PS-CRE and 2 PR-CRE). Eleven patients (33%) died and all of the surviving patients were clinically cured. Compared with patients infected with PS-CRE, in-hospital mortality was higher for patients infected with PRCRE (17% vs. 53%; P = 0.06). Compared with patients with PS-CRE, patients infected with PR-CRE who received polymyxin-containing regimens had lower cure rates and survival (27% vs. 83%; P < 0.01). Overall, patients infected with PR-CRE tended to have a lower survival than PS-CRE patients (P = 0.06) (Fig. 1). Moreover, PRCRE-infected patients tended to have a three times higher rate of mortality (HR = 3.38, 95% CI 0.87–13.11; P = 0.08). In the multivariate model adjusting for diabetes, patients with PR-CRE had a higher rate of mortality, although this was not statistically significant (HR = 2.88, 95% CI 0.69–11.99; P = 0.14).
4. Discussion To the best of our knowledge, this is the first study describing a cohort of patients with post-surgical mediastinitis due to CRE. Several results provided by this investigation have been highlighted here.
In centres with a high volume of cardiothoracic surgery and CRE outbreaks, a significant proportion of post-surgical mediastinitis can be caused by CRE. In this study, approximately one-third of post-surgical mediastinitis were due to CRE strains. Importantly, the rate of mediastinitis reported in this study (1.6%) is in agreement with the current literature [4]. The rate of prior colonisation detected among the patients was particularly high for PS-CRE and may have contributed to the development of the infection. Such colonisation was concomitant with a dramatic expansion of CRE strains (PS-CRE first and PR-CRE later) at this single centre. This observation suggests that CRE strains are likely to colonise before establishing an infection. Colonisation with KPC-producing K. pneumoniae has been suggested as a risk factor for subsequent infections [18,19], including for those strains displaying polymyxin resistance [11]. In fact, because colonisation was not systematically searched for in all of our patients, the real rates of colonisation may have been even higher than those reported here. In this scenario of an increasing number of CRE infections and narrow clonal distribution, infection control measures become crucial to limit the expansion of CRE pathogens [20]. As mentioned previously, we have reduced the rates of PR-CRE infections with measures such as surveillance screening, contact precautions, and antibiotic stewardship restricting the empirical use of polymyxins. Overall in-hospital mortality associated with CRE mediastinitis was high (33%). This rate is in agreement with current studies describing Gram-negative mediastinitis [5] and bacteraemia due to CRE pathogens [6,7]. However, in-hospital mortality for general mediastinitis within our institution is 10% (unpublished data, available under request). Thus, the general CRE mediastinitis in-hospital mortality in the present case series was higher and possibly driven by PR-CRE patients (>50% mortality). The current case series had a high prevalence of diabetes mellitus among the general patient population (58%), although PS-CRE patients had an even higher prevalence of the condition (78%). Diabetes is a highly prevalent condition in, and a risk factor for, general mediastinitis patients. A higher mortality was found in the PR-CRE population adjusted for diabetes (HR = 2.88, 95% CI 0.69–11.99), however this difference was not significant (P = 0.14). The non-significance of this higher mortality in the adjusted multivariate analysis for diabetes was probably due to the limited number of patients and the unequal distribution of the condition among both groups. Given the limited therapeutic options currently available, it is not uncommon for the continued use of polymyxins combined with several other agents for the treatment of these patients. Until more data on newer effective antimicrobial agents or optimised antimicrobial combinations effective against these PR-CRE strains are available, this question will remain. Because mediastinitis is a surgical disease, the role of surgery needs to be particularly highlighted for these difficult-to-treat infections. An aggressive surgical approach should include wide debridement of necrotic tissue, sectioning of compromised bone and cartilage, washing and sternum re-suturing [21] whenever possible. Unfortunately, negative pressure devices (vacuum-assisted closure) were not available at the site during the study period [22]. Whilst the role of hyperbaric oxygen is still to be determined, it was mainly used as an adjunctive therapy for PS-CRE-infected patients who were clinically stable and in whom primary sternal closure was not possible. This study has several limitations. First, it was retrospective and was conducted at a single centre. It is important to mention that most clinical and microbiological data were systematically collected from patients with mediastinitis at this site. Despite coming from a single centre, the number of cases is clinically relevant for a meaningful description. Second, the relatedness (e.g. by PFGE) between the clones causing the infection and those previously colonising could not be determined. As shown in a recent study,
Please cite this article in press as: Abboud CS, et al. Post-surgical mediastinitis due to carbapenem-resistant Enterobacteriaceae: Clinical, epidemiological and survival characteristics. Int J Antimicrob Agents (2016), http://dx.doi.org/10.1016/j.ijantimicag.2016.02.015
G Model ANTAGE-4775; No. of Pages 5
ARTICLE IN PRESS C.S. Abboud et al. / International Journal of Antimicrobial Agents xxx (2016) xxx–xxx
it is very likely that in a high proportion of patients the colonising clones match those causing the infection [19]. Third, the MICs for carbapenems and polymyxins were not determined by a confirmation methodology such as microdilution. Therefore, we could not explore further any potential relationship between the antibiotics used and achievable MICs. Finally, the study is limited by the small sample size, which makes it difficult to perform meaningful inferential analyses. The current findings show that patients with mediastinitis due to PR-CRE have a higher risk of overall mortality. Owing to the limited sample size, we were not able to achieve statistical significance; however, mediastinitis and PR-CRE are relatively rare and we believe that this is the largest discussion of patients with this relevant hospital-acquired infection. Despite all of these limitations, we believe that this study provides clinicians and surgeons with useful clinical and microbiological information to better understand post-surgical mediastinitis due to these difficult-to-treat pathogens. 5. Conclusion Mediastinitis due to CRE has become an emerging infection in this centre. Antibacterial treatments are usually based on carbapenems, frequently accompanied by polymyxin plus a third agent. Despite complex antibiotic treatments and aggressive surgical procedures, these patients have a high mortality, particularly those infected with PR-CRE. Centres with increasing rates of CRE and with a high cardiac surgery volume should be aware of these potentially devastating infections. Acknowledgements The authors acknowledge Celia Harumi Hiroshi (secretary) and LEE Laboratory of Epidemiology and Statistics–Instituto Dante Pazzanese de Cardiologia (São Paulo, Brazil). Funding: This work was funded by a grant [2012/12108-3] from the Fundac¸ão de Amparo à Pesquisa do Estado de São Paulo (FAPESP). Competing interests: MES is or has been a consultant to Achaogen, Astellas, Cempra, Cerexa, Furiex, Nabriva, PRI, Trius and Theravance. All other authors declare no competing interests. Ethical approval: The study was approved by the Ethical Committee of Instituto Dante Pazzanese de Cardiologia. References [1] Abboud CS, Wey SB, Baltar VT. Risk factors for mediastinitis after cardiac surgery. Ann Thorac Surg 2004;77:676–83. [2] Filsoufi F, Castillo JG, Rahmanian PB, Broumand SR, Silvay G, Carpentier A, et al. Epidemiology of deep sternal wound infection in cardiac surgery. J Cardiothorac Vasc Anesth 2009;23:488–94. [3] Wang FD, Chang CH. Risk factors of deep sternal wound infections in coronary artery bypass graft surgery. J Cardiovasc Surg (Torino) 2000;41:709–13.
5
[4] Fowler Jr VG, O’Brien SM, Muhlbaier LH, Corey GR, Ferguson TB, Peterson ED. Clinical predictors of major infections after cardiac surgery. Circulation 2005;112(9 Suppl.):I358–65. [5] Charbonneau H, Maillet JM, Faron M, Mangin O, Puymirat E, Le Besnerais P, et al. Mediastinitis due to Gram-negative bacteria is associated with increased mortality. Clin Microbiol Infect 2014;20:O197–202. [6] Daikos GL, Tsaousi S, Tzouvelekis LS, Anyfantis I, Psichogiou M, Argyropoulou A, et al. Carbapenemase-producing Klebsiella pneumoniae bloodstream infections: lowering mortality by antibiotic combination schemes and the role of carbapenems. Antimicrob Agents Chemother 2014;58: 2322–8. [7] Tumbarello M, Viale P, Viscoli C, Trecarichi EM, Tumietto F, Marchese A, et al. Predictors of mortality in bloodstream infections caused by Klebsiella pneumoniae carbapenemase-producing K. pneumoniae: importance of combination therapy. Clin Infect Dis 2012;55:943–50. [8] Tumbarello M, Trecarichi EM, De Rosa FG, Giannella M, Giacobbe DR, Bassetti M, et al. Infections caused by KPC-producing Klebsiella pneumoniae: differences in therapy and mortality in a multicentre study. J Antimicrob Chemother 2015;70:2133–43. [9] Mammina C, Bonura C, Di Bernardo F, Aleo A, Fasciana T, Sodano C, et al. Ongoing spread of colistin-resistant Klebsiella pneumoniae in different wards of an acute general hospital, Italy, June to December 2011. Euro Surveill 2012:17, pii: 20248. [10] Marchaim D, Chopra T, Pogue JM, Perez F, Hujer AM, Rudin S, et al. Outbreak of colistin-resistant, carbapenem-resistant Klebsiella pneumoniae in metropolitan Detroit, Michigan. Antimicrob Agents Chemother 2011;55:593–9. [11] Giacobbe DR, Del Bono V, Trecarichi EM, De Rosa FG, Giannella M, Bassetti M, et al. Risk factors for bloodstream infections due to colistinresistant KPC-producing Klebsiella pneumoniae: results from a multicenter case–control–control study. Clin Microbiol Infect 2015;21:e1–8. [12] 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;36:309–32. [13] Gaca JG, Sheng S, Daneshmand MA, O’Brien S, Rankin JS, Brennan JM, et al. Outcomes for endocarditis surgery in North America: a simplified risk scoring system. J Thorac Cardiovasc Surg 2011;141:98–106, e1–2. [14] Monteiro J, Widen RH, Pignatari AC, Kubasek C, Silbert S. Rapid detection of carbapenemase genes by multiplex real-time PCR. J Antimicrob Chemother 2012;67:906–9. [15] Diancourt L, Passet V, Verhoef J, Grimont PA, Brisse S. Multilocus sequence typing of Klebsiella pneumoniae nosocomial isolates. J Clin Microbiol 2005;43:4178–82. [16] Egito JG, Abboud CS, Oliveira AP, Maximo CA, Montenegro CM, Amato VL, et al. Clinical evolution of mediastinitis in patients undergoing adjuvant hyperbaric oxygen therapy after coronary artery bypass surgery. Einstein (Sao Paulo) 2013;11:345–9. [17] Abboud CS, Monteiro J, Franca JI, Pignatari AC, De Souza EE, Camargo EC, et al. A space–time model for carbapenemase-producing Klebsiella pneumoniae (KPC) cluster quantification in a high-complexity hospital. Epidemiol Infect 2015;143:2648–52. [18] Hayden MK, Lin MY, Lolans K, Weiner S, Blom D, Moore NM, et al. Prevention of colonization and infection by Klebsiella pneumoniae carbapenemaseproducing Enterobacteriaceae in long-term acute-care hospitals. Clin Infect Dis 2015;60:1153–61. [19] Ducomble T, Faucheux S, Helbig U, Kaisers UX, Konig B, Knaust A, et al. Large hospital outbreak of KPC-2-producing Klebsiella pneumoniae: investigating mortality and the impact of screening for KPC-2 with polymerase chain reaction. J Hosp Infect 2015;89:179–85. [20] Carmeli Y, Akova M, Cornaglia G, Daikos GL, Garau J, Harbarth S, et al. Controlling the spread of carbapenemase-producing Gram-negatives: therapeutic approach and infection control. Clin Microbiol Infect 2010;16: 102–11. [21] Robicsek F, Daugherty HK, Cook JW. The prevention and treatment of sternum separation following open-heart surgery. J Thorac Cardiovasc Surg 1977;73:267–8. [22] Yu AW, Rippel RA, Smock E, Jarral OA. In patients with post-sternotomy mediastinitis is vacuum-assisted closure superior to conventional therapy? Interact Cardiovasc Thorac Surg 2013;17:861–5.
Please cite this article in press as: Abboud CS, et al. Post-surgical mediastinitis due to carbapenem-resistant Enterobacteriaceae: Clinical, epidemiological and survival characteristics. Int J Antimicrob Agents (2016), http://dx.doi.org/10.1016/j.ijantimicag.2016.02.015