- Email: [email protected]
Meticillin resistance in orthopaedic coagulase-negative staphylococcal infections
Journal of Hospital Infection 79 (2011) 248e253
Available online at www.sciencedirect.com
Journal of Hospital Infection journal homepage: www.elsevierhealth.com/journals/jhin
Meticillin resistance in orthopaedic coagulase-negative staphylococcal infectionsq I. Uçkay a, b, c, *, S. Harbarth b, c, T. Ferry b, A. Lübbeke a, S. Emonet b, P. Hoffmeyer a, D. Pittet b, c a
Orthopaedic Surgery Service, University of Geneva Hospitals and Faculty of Medicine, Geneva, Switzerland Infectious Diseases Service, University of Geneva Hospitals and Faculty of Medicine, Geneva, Switzerland c Infection Control Programme, University of Geneva Hospitals and Faculty of Medicine, Geneva, Switzerland b
a r t i c l e i n f o
s u m m a r y
Article history: Received 2 March 2011 Accepted 12 June 2011 by J.A. Child Available online 27 September 2011
Orthopaedic infections due to coagulase-negative staphylococci (CoNS) and meticillin-resistant strains may be increasing. We assessed secular trends of CoNS infections and factors associated with meticillin resistance by performing a 13-year retrospective cohort study of orthopaedic patients with CoNS infections from January 1995 to December 2007. Of 60 CoNS infections, 57 (95%) were implant-related. Median follow-up after end of treatment was 5.1 years (range: 2.4e13.8). During the study period, 44,237 orthopaedic procedures were performed, 21,299 (48%) with implants. The overall cumulative incidence of CoNS-associated infection was 0.14% and 0.28% for implant-related procedures. There were non-significant changes in the absolute number or cumulative incidence of CoNS infection (chi-squared test, P values for trend: 0.45 and 0.97, respectively). Forty-five episodes (75%) were due to meticillin-resistant strains. The proportion of meticillin resistance remained stable over time (P for trend: 0.65). Whereas few (4/15) meticillin-susceptible strains were associated with prior prophylaxis that covered the causative pathogen, 28/45 meticillin-resistant strains were associated with inadequate prophylaxis (P ¼ 0.03). The cumulative incidence of orthopaedic CoNS infection is low and stable in our institution and almost exclusively implant-related. The proportion of meticillin resistance among CoNS has remained stable over the last decade with a favourable clinical outcome. Ó 2011 The Healthcare Infection Society. Published by Elsevier Ltd. All rights reserved.
Keywords: Coagulase-negative staphylococci Implant-associated infection Meticillin resistance Orthopaedic surgery Secular trends
Introduction Coagulase-negative staphylococci (CoNS) are human skin commensals. Staphylococcus epidermidis is the most prevalent among about 30 CoNS species.1 In orthopaedic series, CoNS account for 23e44% of implant-associated infections.2,3 These infections are considered difficult to treat because of the ability of CoNS to grow in biofilms and form small-colony variants.4,5
q Data partially presented as a poster at the 27th Annual Meeting of the European Bone and Joint Infection Society, 18e20 September 2008, Barcelona, Spain, and as a short communication at the 68th Swiss Congress of Orthopaedics and Traumatology, 24e26 September 2008, Basle, Switzerland. * Corresponding author. Address: Infection Control Programme, University of Geneva Hospitals and Faculty of Medicine, 4 Rue Gabrielle Perret-Gentil, 1211 Geneva 14, Switzerland. Tel.: þ41 22 372 9834; fax: þ41 22 372 3987. E-mail address: [email protected] (I. Uçkay).
The number of orthopaedic infections due to CoNS and the proportion of meticillin-resistant strains among clinical CoNS isolates might be increasing in healthcare settings.3,6e10 According to the latest United States National Nosocomial Infections Surveillance report, 65% of clinical isolates of CoNS were resistant to meticillin in 2004.11 Skin carriage of polyclonal meticillin-resistant S. epidermidis (MRSE) has been reported among healthcare workers and orthopaedic patients.7,8,10,12,13 Mohanty and Kay reported 55% meticillin resistance among 312 CoNS isolates grown in tissue samples of total joint replacement patients. Moreover, study patients with meticillin-susceptible infections showed higher implant survival rates than those who harboured meticillin-resistant strains.1 The objectives of this retrospective, cohort study were to determine the cumulative incidence of orthopaedic infections due to CoNS at our institution and to analyse trends of meticillin resistance among infecting CoNS isolates. We examined the potential role of glycopeptides as preoperative antimicrobial prophylaxis and investigated factors associated with meticillin-resistant infection.
0195-6701/$ e see front matter Ó 2011 The Healthcare Infection Society. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.jhin.2011.06.014
I. Uçkay et al. / Journal of Hospital Infection 79 (2011) 248e253
Methods Setting The University of Geneva Hospitals (HUG), Switzerland, is a tertiary medical centre serving a population of w800,000 inhabitants with 47,000 annual admissions and 1901 acute care beds. It is the only public hospital in the canton of Geneva and comprises 61% of all acute care beds. The orthopaedic surgery service has 132 acute care beds, a dedicated infectious diseases specialist and manages two cohorts: a cohort of elective arthroplasty operations with active post-discharge surveillance for a minimum of five years since 1995; and a septic orthopaedic cohort where community-acquired and healthcare-associated infections are registered.14 The latter infections are recorded by the different dedicated infectious diseases physicians who note treatment modalities and follow-up for each case. Follow-up for fracture-fixation devices is at three months and annually thereafter. In the case of infection, patients are followed for at least two years. Orthopaedic surgery is performed under vertical laminar airflow, antibiotic prophylaxis, and with gowns and double gloving. There is no routine use of cement in arthroplasty surgery. Whenever cement is used for a particular reason, it contains tobramycin. Overall adherence to local guidelines for preoperative antibiotic prophylaxis was 95% in 2005 and in 2008 (data not shown). Data collection This retrospective, cohort study spanned a 13-year period from 1 January 1995 to December 2007. Four electronic databases of the HUG clinical microbiology laboratory and administrative coding system (details to the staphylococcal level), the Geneva Total Arthroplasty Cohort Registry, and the Septic Orthopaedic Cohort were searched to identify orthopaedic infections due to CoNS originating from earlier orthopaedic surgery.14 A suspicion of CoNS infection in any of these four databases was further explored. Sixty parameters were collected for each case. Study follow-up was closed on 31 July 2010. Definitions A strict definition for orthopaedic CoNS infection was used in order to remove any doubt of possible contamination, and only orthopaedic infections related to patients operated earlier at our institution were assessed. For instance, healthcare-associated infections of peripheral venous catheters due to CoNS were not recorded. The definition of orthopaedic CoNS infection required the presence of all of the following items: a medical report confirming the infection; the presence of CoNS in more than one intraoperative sample; a targeted antibiotic treatment prescribed by an orthopaedic surgeon or the infectious diseases specialist; and at least one clinical sign of infection, such as local heat, redness, pain, purulent discharge, fistula, or functional impairment. In addition, the patient should not have received antibiotic treatment during the previous six weeks in order to avoid a potential selection bias of meticillin-resistant CoNS strains in pretreated microbiological samples. This is due to the fact that meticillin-resistant CoNS in particular may be selected from standard antibiotic therapies, such as amoxicillineclavulanic acid or cephalosporins. In such a case, the co-pathogen, which could be S. aureus, would not be detected in intraoperative samples, and the case would be wrongly attributed to metcillin-resistant CoNS infection as if this were the only pathogen. Of note, we did not exclude patients who received antibiotic prophylaxis, only patients receiving empirical pre-admission treatment for orthopaedic infections (always spanning several days).
249
Date of infection was not defined as the day of microbiological sampling, but by the first clinical sign (e.g. cicatricial dehiscence, redness, unusual pain, etc.). Histology was not mandatory and co-pathogens were accepted in microbiological specimens only if CoNS outnumbered them by at least three-fold and the co-pathogen was only witnessed in one of all samples. We used this quantitative arbitrary threshold to ensure that the majority of pathogenicity was due to CoNS in the case of the presence of co-pathogens. For example, in an intraoperative sampling with equal amounts of CoNS and Pseudomonas aeruginosa, it is difficult to prove that CoNS contributes the same virulence as P. aeruginosa. Susceptibility to clindamycin in the presence of erythromycin resistance was interpreted as a potentially inducible resistance to clindamycin.15 CoNS strains were distinguished by antimicrobial susceptibility testing and species identification. Polyclonality of CoNS infection was defined as an infection by more than one CoNS strain according to the antibiotic susceptibility pattern and required at least three differences. Genotyping or virulence factor analysis of bacterial isolates was not performed as specimens were not stored. Microbiologically inappropriate preoperative antibiotic prophylaxis was defined as the absence of coverage of an antibiotic agent on the infecting pathogen and/or omission of antibiotic prophylaxis. Microbiological procedures Tissue specimens and swabs were taken intraoperatively and transported during and immediately after surgery to the microbiology laboratory, which is located on the fifth floor of the same building as the operating theatres. The delay between sampling and laboratory processing was around 2 h during day shifts, and several hours during night shifts. Tissues were handled under sterile conditions. Specimens were incubated for five days on in-house blood and chocolate agar culture media, CDC anaerobic agar, and in braineheart infusion liquid. To enhance specificity, only cultures grown on plates were considered and growth in enrichment broth alone was not accepted. CoNS were identified as catalase positive, slidex agglutination negative (PastorexÒ, Bio-Rad, Reinach, Switzerland) and DNAse negative (in-house). Strains were further characterized to the species level using the ID32 Staphylococcus Gallery (bioMérieux, Marcy l’Etoile, France) and/or the Vitek ID (bioMérieux) system. Susceptibility testing was performed according to the Clinical and Laboratory Standards Institute guidelines.16 Procedures remained unchanged during the entire study period. Trends of CoNS infection and meticillin resistance Each CoNS infection was classified by calendar year to assess trends in the cumulative incidence and proportion of meticillin-resistant infection. Antimicrobial resistance patterns among infecting strains in orthopaedic patients were compared with clinical CoNS isolates from other services at HUG. The antimicrobial profile of 2007 (the last full study year) was used as the reference year for comparison purposes. Statistical analyses Logistic regression was used to determine factors associated with meticillin-resistant infection. Since the number of events per variable was too small, multivariate analyses were not performed to avoid spurious findings.17 Group comparisons were performed using Pearson’s chi-squared test, Fisher’s exact text, or Wilcoxon’s rank-sum test, as appropriate. The P-value for trend of the chisquared test evaluated tendencies over time. P 0.05 (two-tailed)
250
I. Uçkay et al. / Journal of Hospital Infection 79 (2011) 248e253
defined statistical significance. STATAÔ software (9.0, STATA Corp., College Station, TX, USA) was used.
co-pathogens within one single sample (Proteus mirabilis and Enterococcus faecalis) with a majority of S. epidermidis in other samples.
Results Antibiotic prophylaxis and meticillin-resistant infection Sixty episodes of CoNS infections in 59 patients were identified (median age: 71 years; interquartile range: 59e79). Thirty-two patients were female; 16 were immune-suppressed. The median length of hospital stay between admission and original surgery was three days. Forty-seven percent of all patients were operated on within two days of admission. The median active follow-up was 5.1 years (range: 2.4e13.8). Fifty-seven (95%) infections were implant-related. These included total hip prostheses (N ¼ 24), total knee prostheses (7), femoral nails (5), dynamic hip screws (4), ankle plates (4), mid-foot plate and screws (3), tibia nails (2), femoral plate (2), clavicular plate (1), humerus plate (1), kneecap cerclage (1), olecranon cerclage (1), arthrodesis stitch (1), and lumbar spondylodesis material (1). Among the 31 arthroplasties, one was a revision for aseptic loosening. Among the non-implant-related cases, we noted a valgization osteotomy, hallux valgus surgery, and a post-traumatic abscess. Median duration between implantation and first clinical sign of infection was six months. Overall, 72% (N ¼ 43) developed infection within one year post implantation, 10% (N ¼ 6) between one and two years, and 18% (N ¼ 11) after two years. There was no death directly related to CoNS infection. During the study period, 44,237 orthopaedic procedures were performed, 21,299 (48%) with implants, of which 8263 were arthroplasty cases. The overall cumulative incidence of orthopaedic CoNS-associated infection was 0.14% [95% confidence interval (CI): 0.10e0.17%], 0.28% (0.21e0.36) for implant-related procedures, and 0.37% (0.28e0.42%) for arthroplasty surgery. Table I summarizes the clinical characteristics of the study population. Microbiology of orthopaedic CoNS infections CoNS were further identified to the species level in 45/60 (75%) infections. Pathogens identified were S. epidermidis (36/45, 80%), S. capitis (2), S. hominis (2), S. intermedius, S. lugdunensis, S. simulans, S. xylosus, and S. schleiferi. Twelve infections (20%) were probably polyclonal (same CoNS species, but very different antibiotic susceptibility results). In eight episodes (13.3%), a non-CoNS co-pathogen was identified in one of several samples: meticillin-resistant S. aureus (MRSA) (4); P. aeruginosa (2); and Klebsiella pneumonia (1). In one episode, there were two
First- and second-generation cephalosporins (cefazolin, cefuroxime) were the most frequently used agents for surgical prophylaxis (48/60; 80%), followed by ciprofloxacin (2) and amoxicillineclavulanic acid. One patient received combined prophylaxis of vancomycin and imipenem. There was no documented, systemic, antibiotic prophylaxis in eight cases. In 11 arthroplasties, tobramycin-containing cement was used in addition to systemic antibiotic prophylaxis. Forty-five infections (75%) were due to meticillin-resistant CoNS strains. None of these patients had received glycopeptide prophylaxis at time of surgery. By univariate analysis, possibly inadequate or missing antibiotic prophylaxis was associated with infection due to meticillin-resistant CoNS strains (odds ratio: 42; 95% CI: 4.0e445.7), but not the number of prior hospitalizations or delay between admission and surgery (Table I). Temporal trends of CoNS infections and antibiotic resistance There were fluctuations in the absolute number and cumulative incidence of implant-related infections with a peak among devices implanted in 2003 (P-values for trend: 0.45 and 0.97, respectively; Figure 1). The cumulative incidence was 2.6 per 1000 devices implanted during 1998e1999, reaching 5 per 1000 devices in 2003 and decreasing to 1.5 per 1000 in 2007. The corresponding incidences for the subgroup of arthroplasty infections were 5 infections per 1319 arthroplasty procedures during 1998e1999 (revisions included), 3 per 750 arthroplasties in 2003, and 4 per 728 arthroplasties in 2007, respectively. No epidemiological link could be established between CoNS infections observed during 2003, and antibiotic susceptibility patterns revealed different patterns associated with different surgeons or procedures. The proportion of meticillin resistance remained stable throughout the study period (Figure 1). The corresponding proportions among CoNS revealed no significant trend throughout the entire study period (P ¼ 0.65) (Figure 1). Table II compares the antibiotic susceptibility patterns between CoNS strains responsible for orthopaedic infections in the current cohort and those identified hospital-wide in 2007. As shown, resistance patterns for all antibiotics were quite similar, with the exception of meticillin resistance which was markedly more prevalent among strains identified from orthopaedic infections. In 2008, which was outside
Table I Comparison between meticillin-susceptible and meticillin-resistant orthopaedic coagulase-negative staphylococcus infections and parameters potentially associated with meticillin-resistant infection (University Hospitals of Geneva cohort study, 1995e2007) Associated variables Female Age >70 years Immune suppression/comorbiditiesb Emergency surgery Adequate antibiotic prophylaxis Inadequate/missing antibiotic prophylaxis Polyclonal infection Median no. of prior hospitalizations 1e2 prior hospitalization(s) Delay between admission and surgery (median)
Meticillin-susceptible (N ¼ 15) 7 6 2 6 2 4 3 2 9 3
(47%) (40%) (13%) (40%) (13%) (27%) (20%) (range: 0e4) (60%) days
Meticillin-resistant N ¼ 45 25 26 14 20 11 28 9 3 20 2
(56%) (57%) (31%) (44%) (24%) (62%) (20%) (range: 0e21) (44%) days
OR, odds ratio; CI, confidence interval. a Estimated by univariate logistic regression. b Diabetes mellitus, alcohol abuse, chronic renal failure, cancer, rheumatic polyarthritis, cirrhosis, solid organ transplant.
ORa (95% CI) 0.7 0.5 0.3 0.8 0.5 0.2 1.0 1.2 1.8 1.0
(0.2e2.7) (0.1e1.9) (0.1e1.9) (0.2e3.2) (0.1e2.7) (0.1e0.9) (0.1e5.0) (1.0e1.4) (0.5e7.5) (1.0e1.0)
P-value 0.55 0.23 0.31 0.76 0.49 0.03 1.00 0.13 0.30 0.68
I. Uçkay et al. / Journal of Hospital Infection 79 (2011) 248e253
251
Table II Antibiotic resistance among all coagulase-negative staphylococci (except Staphylococcus saprophyticus) isolates compared to the resistance pattern of coagulase-negative staphylococci responsible for orthopaedic infections (University of Geneva Hospitals) Antibiotic class
Beta-lactams Aminoglycosides Quinolones Lincosamides Others Glycopeptides
Antibiotic
Hospital-wide (N ¼ 950)
Orthopaedic (N ¼ 60)
(2007)
(1995e2007)
84% 59% 32% 47% 56%b 56% 38% 6% 0%
NA 78% 45% 47% 45%b NA 36% 9% 0%
Penicillin Meticillin Gentamycin/tobramycin Ciprofloxacin Clindamycin Erythromycin Cotrimoxazole Rifampin Vancomycin
OR (95% CI)
e 2.5 1.7 1.0 0.6 e 0.9 1.4 1.0
(1.3e5.1) (1.0e3.0) (0.6e1.7) (0.4e1.1) (0.5e1.7) (0.4e3.7)
P-valuea
e 0.01 0.04 0.95 0.10 e 0.84 0.47 1
OR, odds ratio; CI, confidence interval; NA, not available. a Two-sided chi-squared test or Fisher’s exact test, as appropriate. b Microbiological susceptibility to clindamycin in the presence of erythromycin resistance was interpreted as potential resistance to clindamycin.
the study period, meticillin resistance among all hospital-wide clinical CoNS isolates further decreased to 56% (data not shown). Discussion The overall cumulative incidence of deep orthopaedic surgical site infections due to CoNS is low at our institution (1.4 per 1000 surgical procedures) and almost exclusively implant-related. Importantly, the proportion of meticillin-resistant strains remained unchanged over more than a decade, despite a continuous increase in implant-related procedures. CoNS strains infecting orthopaedic sites were no more resistant than overall CoNS strains identified at HUG, with the exception of meticillin resistance. Few (4/15) meticillin-susceptible strains were associated with prior prophylaxis that covered the causative pathogen, but 28/45 meticillin-resistant strains were associated with inadequate prophylaxis; the difference was statistically significant (P ¼ 0.03). First- and second-generation cephalosporins are generally used in
surgery because of good intrinsic activity against staphylococci and streptococci, fewer side-effects, and low costs.18e20 However, by definition, meticillin-resistant CoNS strains, the predominant CoNS in most healthcare institutions including HUG, are resistant to all cephalosporins.4,6,21 The delay between hospital admission and surgery and the number of previous hospitalizations was not associated with meticillin-resistant infection. It remains unknown when patients become colonized with meticillin-resistant CoNS following hospital admission. In a Swedish study, most patients in an orthopaedic ward were colonized with meticillin-resistant CoNS within 14 days following admission.10 Importantly, resistance to other antibiotic classes also increased, independent of prior antibiotic exposure. In a prospective survey among surgical patients in Switzerland, only 20% of CoNS sampled from the future surgical incision site were meticillin resistant on admission, but this rose to 47% immediately after elective surgery on the same site.21 In the current study, the delay between admission and surgery did not seem to play a role.
16
4,500 3,487
14 1 12
3,523
3,873 3,552
3,243
4,000 3,500
3 2,500 8 6
2
5
13
3
2,000
No. of procedures
No. of infections
3,000 10
1,500 8
4
7
5
1,000 5
2
500 0
0 1998–1999
2000–2001
2002–2003
2004–2005
2006–2007
Figure 1. Secular trends in orthopaedic implant procedures and infections due to meticillin-susceptible (white bars) and meticillin-resistant (black bars) coagulase-negative staphylococci, University of Geneva Hospitals, 1998e2007. Infections are indicated according to the year of implantation. Follow-up closed 31 July 2010. Median follow-up: 5.1 years.
252
I. Uçkay et al. / Journal of Hospital Infection 79 (2011) 248e253
Our finding that microbiologically appropriate antibiotic prophylaxis was the single protective parameter against meticillinresistant strains in established CoNS infection, including the fact that CoNS strains were more frequently resistant to meticillin than other CoNS isolates from elsewhere in the hospital, opens the debate regarding the possible use of routine glycopeptide prophylaxis in orthopaedic surgery. Among patients colonized with MRSA, vancomycin has become the prophylaxis of choice in many institutions.22,23 Nevertheless, we do not consider that routine screening for meticillin-resistant CoNS or vancomycin prophylaxis against meticillin-resistant CoNS would be warranted at HUG for the following reasons:22,24 (a) a significant proportion of CoNS strains are still susceptible to beta-lactam antibiotics (Table II); (b) although not the case at present, vancomycin resistance among enterococci and CoNS may emerge in the future;11 (c) vancomycin needs slow infusion to avoid excessive histamine release, responsible for the so-called ‘red man’ syndrome, which renders its use more difficult in prophylaxis where accurate timing of administration is of utmost importance;25 (d) the cumulative incidence of orthopaedic infections due to meticillin-susceptible or -resistant CoNS at HUG is very low and does not show any increasing trend, contrary to some reports;3,6 (e) routine glycopeptide prophylaxis does not guarantee decreased infection rates due to CoNS, but may increase surgical site infection due to other pathogens. A review of four randomized trials comparing prophylactic teicoplanin vs prophylactic cephalosporin in settings with a high prevalence of meticillin resistance among S. epidermidis showed similar infection rates among both groups.26 This has also been confirmed in a meta-analysis of seven randomized trials in cardiac surgery, even though there have been single trials that switched to vancomycin prophylaxis in cardiac surgery.27,28 A recent systematic review and economic model of switching from non-glycopeptide to glycopeptide antibiotic prophylaxis for surgery in endemic MRSA settings failed to show increased efficacy in preventing surgical site infection due to meticillin-resistant strains. Even for MRSA, there is insufficient evidence to determine whether there is a threshold prevalence to justify a switch to routine glycopeptide prophylaxis.29 We consider that the same arguments are equally valid regarding vancomycin-containing cements for primary prophylaxis in arthroplasty surgery. Our study has some limitations. (i) Since it was retrospective and conducted in a single centre, the generalizability of our findings deserves further investigation. (ii) It includes only 60 CoNS infections, although they were identified in a large cohort of 13 years and with a median follow-up of 5.1 years. (iii) Complete case review for CoNS infection was retrospective, thus additional parameters associated with meticillin resistance acquisition may have been missed. (iv) Patients with an infection treated in another hospital may not have been detected. However, since HUG is the largest and the only public hospital in the region, we consider such a detection bias as minimal. (v) We used microbiologically documented databases that could imply an additional possible selection bias for infections where no strains grew in culture. However, taking into account the fact that we also used the HUG administration coding system, medical reports, and prospective cohort follow-up, we consider this bias as minimal. A high proportion of infections were due to meticillinresistant strains, which are unlikely to be masked by unreported exposure to antibiotics prior to admission. In the literature, culturenegative, prosthetic joint infections account for only 7% of cases and the outcome is indistinguishable from that due to identified bacterial pathogens.30 Another problem of microbiologically based, culturepositive definitions of infections is the potential selection of grown pathogens through pre-admission antibiotic therapy. In particular, CoNS may be selected through standard empirical antibiotic regimens, such as amoxicillineclavulanic acid or cephalosporins. In such
a case, the co-pathogen, e.g. S. aureus, would not be detected in intraoperative samples and the case wrongly attributed to CoNS infection only. Therefore, we established an exclusion criterion of avoiding cases with empirical antibiotic therapy within six weeks prior to admission. The cut-off of six weeks was arbitrary and only four cases were excluded from final analysis because of that criterion. This exclusion did not influence overall results. Moreover, as these cases were pretreated with antibiotics upon admission, they also showed growth of CoNS either in enrichment broth only or only in a single intraoperative sample. In other words, these four patients had several reasons to be excluded. (vi) In the literature reporting orthopaedic implant-related infections, the ideal follow-up time is unknown. The Healthcare Infection Control Practices Advisory Committee guidelines for the prevention of surgical site infections and the United States Centers for Disease Control and Prevention definitions for implant-associated infections suggest a minimal surveillance of one year for implants.23,31 Hence, we consider actively managed, individual minimal and median follow-up times of 2.4 years and 5.1 years, respectively, as an important strength of the current study. (vii) We addressed only substantial orthopaedic CoNS infection for which the patients underwent surgery. Small superficial CoNS infections during hospitalization or after discharge may have remained unrecorded. Similarly, deep infections under ongoing antibiotic treatment on admission, which could have been only due to CoNS, were excluded according to our study definitions. Despite these limitations and due to the fact that we used four different databases (Arthroplasty Cohort, Septic Orthopaedic Cohort, microbiological laboratory, and the hospital’s coding system), we are confident that we have included all severe and clinically significant CoNS infections. (viii) Our study addressed purely epidemiological issues and related perioperative antibiotic prophylaxis. Therapeutic approaches for CoNS infections were beyond the topic of this study and we recommend the readers to consult other reports.4,32 In conclusion, orthopaedic CoNS infections are rare and almost exclusively implant-related. CoNS are mostly resistant to beta-lactam antibiotics used as prophylaxis. In our institution, the cumulative incidence and the proportion of meticillin resistance has remained stable during the last decade.32 A routine change to glycopeptide prophylaxis or glycopeptide-containing cement in arthroplasty surgery is not warranted for orthopaedic procedures in our setting. Acknowledgements We are indebted to E. Huggler, Laboratory of Infectious Diseases, and C. Béant and Dr P. Tahintzi, Coding Office, for their help, to Dr A. Gayet-Ageron for her advice, and to R. Sudan for editorial assistance. Thanks also to the teams of the Orthopaedic Surgery Service and the Laboratory of Bacteriology for their support. Conflict of interest statement None declared. Funding sources None. References 1. Kloos WE, Bannerman TL. Update on clinical significance of coagulase-negative staphylococci. Clin Microbiol Rev 1994;7:117e140. 2. Segawa H, Tsukayama DT, Kyle RF, Becker DA, Gustilo RB. Infection after total knee arthroplasty. A retrospective study of the treatment of eighty-one infections. J Bone Joint Surg Am 1999;81:1434e1445. 3. Fulkerson E, Valle CJ, Wise B, Walsh M, Preston C, Di Cesare PE. Antibiotic susceptibility of bacteria infecting total joint arthroplasty sites. J Bone Joint Surg Am 2006;88:1231e1237.
I. Uçkay et al. / Journal of Hospital Infection 79 (2011) 248e253 4. Uçkay I, Pittet D, Vaudaux P, Sax H, Lew D, Waldvogel F. Foreign body infections due to Staphylococcus epidermidis. Ann Med 2009;41:109e119. 5. Adler H, Widmer A, Frei R. Emergence of a teicoplanin-resistant small colony variant of Staphylococcus epidermidis during vancomycin therapy. Eur J Clin Microbiol Infect Dis 2003;22:746e748. 6. Mohanty SS, Kay PR. Infection in total joint replacements. Why we screen MRSA when MRSE is the problem? J Bone Joint Surg Br 2004;86:266e268. 7. James PJ, Butcher IA, Gardner ER, Hamblen DL. Methicillin-resistant Staphylococcus epidermidis in infection of hip arthroplasties. J Bone Joint Surg Br 1994;76:725e727. 8. Sanzen L, Walder M. Antibiotic resistance of coagulase-negative staphylococci in an orthopaedic department. J Hosp Infect 1988;12:103e108. 9. Joshy S, Gogi N, Thomas B, Mahale A, Singh BK. Delayed onset of deep infection after total knee arthroplasty: comparison based on the infecting organism. J Orthop Surg (Hong Kong) 2007;15:154e158. 10. Thore M, Kuhn I, Lofdahl S, Burman LG. Drug-resistant coagulase-negative skin staphylococci. Evaluation of four marker systems and epidemiology in an orthopaedic ward. Epidemiol Infect 1990;105:95e105. 11. National Nosocomial Infections Surveillance (NNIS) System Report 2004. Am J Infect Control 2004;32:470e485. 12. Widerstrom M, Monsen T, Karlsson C, Wistrom J. Molecular epidemiology of methicillin-resistant coagulase-negative staphylococci in a Swedish county hospital: evidence of intra- and interhospital clonal spread. J Hosp Infect 2006;64:177e183. 13. Kitao T. Survey of methicillin-resistant coagulase-negative staphylococci isolated from the fingers of nursing students. J Infect Chemother 2003;9:30e34. 14. Lübbeke A, Stern R, Garavaglia G, Zürcher L, Hoffmeyer P. Differences in outcomes of obese women and men undergoing primary total hip arthroplasty. Arthritis Rheum 2007;57:327e334. 15. Yılmaz G, Aydın K, Iskender S, Caylan R, Köksal I. Detection and prevalence of inducible clindamycin resistance in staphylococci. J Med Microbiol 2007; 56:342e345. 16. Clinical and Laboratory Standards Institute. Performance standards for antimicrobial susceptibility testing. 17th informational supplement (standard M100-S17). Wayne, PA: CLSI; 2007. 17. Peduzzi P, Concato J, Feinstein AR, Holford TR. Importance of events per independent variable in proportional hazards regression analysis. II. Accuracy and precision of regression estimates. J Clin Epidemiol 1995;48:1503e1510. 18. Petignat C, Francioli P, Harbarth S, et al. Cefuroxime prophylaxis is effective in herniated disk surgery: a double-blind, placebo-controlled study. Spine 2008;15:1919e1924.
253
19. Trampuz A, Zimmerli W. Antimicrobial agents in orthopaedic surgery: prophylaxis and treatment. Drugs 2006;66:1089e1105. 20. Hill C, Flamant R, Mazas F, Evrard J. Prophylactic cefazolin versus placebo in total hip replacement. Report of a multicentre double-blind randomised trial. Lancet 1981;1:795e796. 21. Zingg W, Imhof A, Senn G, Demartines N, Ruef C. Rapid colonization with methicillin-resistant coagulase-negative staphylococci after surgery. World J Surg 2009;33:2058e2062. 22. Lew D, Schrenzel J, Francois P, Vaudaux P. Pathogenesis, prevention, and therapy of staphylococcal prosthetic infections. Curr Clin Top Infect Dis 2001; 21:252e270. 23. Mangram AJ, Horan TC, Pearson ML, Silver LC, Jarvis WR. Guideline for prevention of surgical site infection, 1999. Hospital Infection Control Practices Advisory Committee. Infect Control Hosp Epidemiol 1999;20: 250e278. 24. Uçkay I, Vernaz-Hegi N, Harbarth S, et al. Activity and impact on antibiotic use and costs of a dedicated infectious diseases consultant on a septic orthopaedic unit. J Infect 2009;58:205e212. 25. Classen DC, Evans RS, Pestotnik SL, Horn SD, Menlove RL, Burke JP. The timing of prophylactic administration of antibiotics and the risk of surgical-wound infection. N Engl J Med 1992;326:281e286. 26. Mini E, Nobili S, Periti P. Methicillin-resistant staphylococci in clean surgery. Is there a role for prophylaxis? Drugs 1997;54:39e52. 27. Bolon MK, Morlote M, Weber SG, Koplan B, Carmeli Y, Wright SB. Glycopeptides are no more effective than beta-lactam agents for prevention of surgical site infection after cardiac surgery: a meta-analysis. Clin Infect Dis 2004;38:1357e1363. 28. Garey KW, Lai D, Dao-Tran TK, Gentry LO, Hwang LY, Davis BR. Interrupted time series analysis of vancomycin compared to cefuroxime for surgical prophylaxis in patients undergoing cardiac surgery. Antimicrob Agents Chemother 2008;52:446e451. 29. Cranny G, Elliott R, Weatherly H, et al. A systematic review and economic model of switching from non-glycopeptide to glycopeptide antibiotic prophylaxis for surgery. Health Technol Assess 2008;12:1e168. 30. Berbari EF, Marculescu C, Sia I, et al. Culture-negative prosthetic joint infection. Clin Infect Dis 2007;45:1113e1119. 31. Garner JS, Jarvis WR, Emori TG, Horan TC, Hughes JM. CDC definitions for nosocomial infections. Am J Infect Control 1988;16:128e140. 32. Teterycz D, Ferry T, Lew D, et al. Outcome of orthopedic implant infections due to different staphylococci. Int J Infect Dis 2010;14:913e918.
Available online at www.sciencedirect.com
Journal of Hospital Infection journal homepage: www.elsevierhealth.com/journals/jhin
Meticillin resistance in orthopaedic coagulase-negative staphylococcal infectionsq I. Uçkay a, b, c, *, S. Harbarth b, c, T. Ferry b, A. Lübbeke a, S. Emonet b, P. Hoffmeyer a, D. Pittet b, c a
Orthopaedic Surgery Service, University of Geneva Hospitals and Faculty of Medicine, Geneva, Switzerland Infectious Diseases Service, University of Geneva Hospitals and Faculty of Medicine, Geneva, Switzerland c Infection Control Programme, University of Geneva Hospitals and Faculty of Medicine, Geneva, Switzerland b
a r t i c l e i n f o
s u m m a r y
Article history: Received 2 March 2011 Accepted 12 June 2011 by J.A. Child Available online 27 September 2011
Orthopaedic infections due to coagulase-negative staphylococci (CoNS) and meticillin-resistant strains may be increasing. We assessed secular trends of CoNS infections and factors associated with meticillin resistance by performing a 13-year retrospective cohort study of orthopaedic patients with CoNS infections from January 1995 to December 2007. Of 60 CoNS infections, 57 (95%) were implant-related. Median follow-up after end of treatment was 5.1 years (range: 2.4e13.8). During the study period, 44,237 orthopaedic procedures were performed, 21,299 (48%) with implants. The overall cumulative incidence of CoNS-associated infection was 0.14% and 0.28% for implant-related procedures. There were non-significant changes in the absolute number or cumulative incidence of CoNS infection (chi-squared test, P values for trend: 0.45 and 0.97, respectively). Forty-five episodes (75%) were due to meticillin-resistant strains. The proportion of meticillin resistance remained stable over time (P for trend: 0.65). Whereas few (4/15) meticillin-susceptible strains were associated with prior prophylaxis that covered the causative pathogen, 28/45 meticillin-resistant strains were associated with inadequate prophylaxis (P ¼ 0.03). The cumulative incidence of orthopaedic CoNS infection is low and stable in our institution and almost exclusively implant-related. The proportion of meticillin resistance among CoNS has remained stable over the last decade with a favourable clinical outcome. Ó 2011 The Healthcare Infection Society. Published by Elsevier Ltd. All rights reserved.
Keywords: Coagulase-negative staphylococci Implant-associated infection Meticillin resistance Orthopaedic surgery Secular trends
Introduction Coagulase-negative staphylococci (CoNS) are human skin commensals. Staphylococcus epidermidis is the most prevalent among about 30 CoNS species.1 In orthopaedic series, CoNS account for 23e44% of implant-associated infections.2,3 These infections are considered difficult to treat because of the ability of CoNS to grow in biofilms and form small-colony variants.4,5
q Data partially presented as a poster at the 27th Annual Meeting of the European Bone and Joint Infection Society, 18e20 September 2008, Barcelona, Spain, and as a short communication at the 68th Swiss Congress of Orthopaedics and Traumatology, 24e26 September 2008, Basle, Switzerland. * Corresponding author. Address: Infection Control Programme, University of Geneva Hospitals and Faculty of Medicine, 4 Rue Gabrielle Perret-Gentil, 1211 Geneva 14, Switzerland. Tel.: þ41 22 372 9834; fax: þ41 22 372 3987. E-mail address: [email protected] (I. Uçkay).
The number of orthopaedic infections due to CoNS and the proportion of meticillin-resistant strains among clinical CoNS isolates might be increasing in healthcare settings.3,6e10 According to the latest United States National Nosocomial Infections Surveillance report, 65% of clinical isolates of CoNS were resistant to meticillin in 2004.11 Skin carriage of polyclonal meticillin-resistant S. epidermidis (MRSE) has been reported among healthcare workers and orthopaedic patients.7,8,10,12,13 Mohanty and Kay reported 55% meticillin resistance among 312 CoNS isolates grown in tissue samples of total joint replacement patients. Moreover, study patients with meticillin-susceptible infections showed higher implant survival rates than those who harboured meticillin-resistant strains.1 The objectives of this retrospective, cohort study were to determine the cumulative incidence of orthopaedic infections due to CoNS at our institution and to analyse trends of meticillin resistance among infecting CoNS isolates. We examined the potential role of glycopeptides as preoperative antimicrobial prophylaxis and investigated factors associated with meticillin-resistant infection.
0195-6701/$ e see front matter Ó 2011 The Healthcare Infection Society. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.jhin.2011.06.014
I. Uçkay et al. / Journal of Hospital Infection 79 (2011) 248e253
Methods Setting The University of Geneva Hospitals (HUG), Switzerland, is a tertiary medical centre serving a population of w800,000 inhabitants with 47,000 annual admissions and 1901 acute care beds. It is the only public hospital in the canton of Geneva and comprises 61% of all acute care beds. The orthopaedic surgery service has 132 acute care beds, a dedicated infectious diseases specialist and manages two cohorts: a cohort of elective arthroplasty operations with active post-discharge surveillance for a minimum of five years since 1995; and a septic orthopaedic cohort where community-acquired and healthcare-associated infections are registered.14 The latter infections are recorded by the different dedicated infectious diseases physicians who note treatment modalities and follow-up for each case. Follow-up for fracture-fixation devices is at three months and annually thereafter. In the case of infection, patients are followed for at least two years. Orthopaedic surgery is performed under vertical laminar airflow, antibiotic prophylaxis, and with gowns and double gloving. There is no routine use of cement in arthroplasty surgery. Whenever cement is used for a particular reason, it contains tobramycin. Overall adherence to local guidelines for preoperative antibiotic prophylaxis was 95% in 2005 and in 2008 (data not shown). Data collection This retrospective, cohort study spanned a 13-year period from 1 January 1995 to December 2007. Four electronic databases of the HUG clinical microbiology laboratory and administrative coding system (details to the staphylococcal level), the Geneva Total Arthroplasty Cohort Registry, and the Septic Orthopaedic Cohort were searched to identify orthopaedic infections due to CoNS originating from earlier orthopaedic surgery.14 A suspicion of CoNS infection in any of these four databases was further explored. Sixty parameters were collected for each case. Study follow-up was closed on 31 July 2010. Definitions A strict definition for orthopaedic CoNS infection was used in order to remove any doubt of possible contamination, and only orthopaedic infections related to patients operated earlier at our institution were assessed. For instance, healthcare-associated infections of peripheral venous catheters due to CoNS were not recorded. The definition of orthopaedic CoNS infection required the presence of all of the following items: a medical report confirming the infection; the presence of CoNS in more than one intraoperative sample; a targeted antibiotic treatment prescribed by an orthopaedic surgeon or the infectious diseases specialist; and at least one clinical sign of infection, such as local heat, redness, pain, purulent discharge, fistula, or functional impairment. In addition, the patient should not have received antibiotic treatment during the previous six weeks in order to avoid a potential selection bias of meticillin-resistant CoNS strains in pretreated microbiological samples. This is due to the fact that meticillin-resistant CoNS in particular may be selected from standard antibiotic therapies, such as amoxicillineclavulanic acid or cephalosporins. In such a case, the co-pathogen, which could be S. aureus, would not be detected in intraoperative samples, and the case would be wrongly attributed to metcillin-resistant CoNS infection as if this were the only pathogen. Of note, we did not exclude patients who received antibiotic prophylaxis, only patients receiving empirical pre-admission treatment for orthopaedic infections (always spanning several days).
249
Date of infection was not defined as the day of microbiological sampling, but by the first clinical sign (e.g. cicatricial dehiscence, redness, unusual pain, etc.). Histology was not mandatory and co-pathogens were accepted in microbiological specimens only if CoNS outnumbered them by at least three-fold and the co-pathogen was only witnessed in one of all samples. We used this quantitative arbitrary threshold to ensure that the majority of pathogenicity was due to CoNS in the case of the presence of co-pathogens. For example, in an intraoperative sampling with equal amounts of CoNS and Pseudomonas aeruginosa, it is difficult to prove that CoNS contributes the same virulence as P. aeruginosa. Susceptibility to clindamycin in the presence of erythromycin resistance was interpreted as a potentially inducible resistance to clindamycin.15 CoNS strains were distinguished by antimicrobial susceptibility testing and species identification. Polyclonality of CoNS infection was defined as an infection by more than one CoNS strain according to the antibiotic susceptibility pattern and required at least three differences. Genotyping or virulence factor analysis of bacterial isolates was not performed as specimens were not stored. Microbiologically inappropriate preoperative antibiotic prophylaxis was defined as the absence of coverage of an antibiotic agent on the infecting pathogen and/or omission of antibiotic prophylaxis. Microbiological procedures Tissue specimens and swabs were taken intraoperatively and transported during and immediately after surgery to the microbiology laboratory, which is located on the fifth floor of the same building as the operating theatres. The delay between sampling and laboratory processing was around 2 h during day shifts, and several hours during night shifts. Tissues were handled under sterile conditions. Specimens were incubated for five days on in-house blood and chocolate agar culture media, CDC anaerobic agar, and in braineheart infusion liquid. To enhance specificity, only cultures grown on plates were considered and growth in enrichment broth alone was not accepted. CoNS were identified as catalase positive, slidex agglutination negative (PastorexÒ, Bio-Rad, Reinach, Switzerland) and DNAse negative (in-house). Strains were further characterized to the species level using the ID32 Staphylococcus Gallery (bioMérieux, Marcy l’Etoile, France) and/or the Vitek ID (bioMérieux) system. Susceptibility testing was performed according to the Clinical and Laboratory Standards Institute guidelines.16 Procedures remained unchanged during the entire study period. Trends of CoNS infection and meticillin resistance Each CoNS infection was classified by calendar year to assess trends in the cumulative incidence and proportion of meticillin-resistant infection. Antimicrobial resistance patterns among infecting strains in orthopaedic patients were compared with clinical CoNS isolates from other services at HUG. The antimicrobial profile of 2007 (the last full study year) was used as the reference year for comparison purposes. Statistical analyses Logistic regression was used to determine factors associated with meticillin-resistant infection. Since the number of events per variable was too small, multivariate analyses were not performed to avoid spurious findings.17 Group comparisons were performed using Pearson’s chi-squared test, Fisher’s exact text, or Wilcoxon’s rank-sum test, as appropriate. The P-value for trend of the chisquared test evaluated tendencies over time. P 0.05 (two-tailed)
250
I. Uçkay et al. / Journal of Hospital Infection 79 (2011) 248e253
defined statistical significance. STATAÔ software (9.0, STATA Corp., College Station, TX, USA) was used.
co-pathogens within one single sample (Proteus mirabilis and Enterococcus faecalis) with a majority of S. epidermidis in other samples.
Results Antibiotic prophylaxis and meticillin-resistant infection Sixty episodes of CoNS infections in 59 patients were identified (median age: 71 years; interquartile range: 59e79). Thirty-two patients were female; 16 were immune-suppressed. The median length of hospital stay between admission and original surgery was three days. Forty-seven percent of all patients were operated on within two days of admission. The median active follow-up was 5.1 years (range: 2.4e13.8). Fifty-seven (95%) infections were implant-related. These included total hip prostheses (N ¼ 24), total knee prostheses (7), femoral nails (5), dynamic hip screws (4), ankle plates (4), mid-foot plate and screws (3), tibia nails (2), femoral plate (2), clavicular plate (1), humerus plate (1), kneecap cerclage (1), olecranon cerclage (1), arthrodesis stitch (1), and lumbar spondylodesis material (1). Among the 31 arthroplasties, one was a revision for aseptic loosening. Among the non-implant-related cases, we noted a valgization osteotomy, hallux valgus surgery, and a post-traumatic abscess. Median duration between implantation and first clinical sign of infection was six months. Overall, 72% (N ¼ 43) developed infection within one year post implantation, 10% (N ¼ 6) between one and two years, and 18% (N ¼ 11) after two years. There was no death directly related to CoNS infection. During the study period, 44,237 orthopaedic procedures were performed, 21,299 (48%) with implants, of which 8263 were arthroplasty cases. The overall cumulative incidence of orthopaedic CoNS-associated infection was 0.14% [95% confidence interval (CI): 0.10e0.17%], 0.28% (0.21e0.36) for implant-related procedures, and 0.37% (0.28e0.42%) for arthroplasty surgery. Table I summarizes the clinical characteristics of the study population. Microbiology of orthopaedic CoNS infections CoNS were further identified to the species level in 45/60 (75%) infections. Pathogens identified were S. epidermidis (36/45, 80%), S. capitis (2), S. hominis (2), S. intermedius, S. lugdunensis, S. simulans, S. xylosus, and S. schleiferi. Twelve infections (20%) were probably polyclonal (same CoNS species, but very different antibiotic susceptibility results). In eight episodes (13.3%), a non-CoNS co-pathogen was identified in one of several samples: meticillin-resistant S. aureus (MRSA) (4); P. aeruginosa (2); and Klebsiella pneumonia (1). In one episode, there were two
First- and second-generation cephalosporins (cefazolin, cefuroxime) were the most frequently used agents for surgical prophylaxis (48/60; 80%), followed by ciprofloxacin (2) and amoxicillineclavulanic acid. One patient received combined prophylaxis of vancomycin and imipenem. There was no documented, systemic, antibiotic prophylaxis in eight cases. In 11 arthroplasties, tobramycin-containing cement was used in addition to systemic antibiotic prophylaxis. Forty-five infections (75%) were due to meticillin-resistant CoNS strains. None of these patients had received glycopeptide prophylaxis at time of surgery. By univariate analysis, possibly inadequate or missing antibiotic prophylaxis was associated with infection due to meticillin-resistant CoNS strains (odds ratio: 42; 95% CI: 4.0e445.7), but not the number of prior hospitalizations or delay between admission and surgery (Table I). Temporal trends of CoNS infections and antibiotic resistance There were fluctuations in the absolute number and cumulative incidence of implant-related infections with a peak among devices implanted in 2003 (P-values for trend: 0.45 and 0.97, respectively; Figure 1). The cumulative incidence was 2.6 per 1000 devices implanted during 1998e1999, reaching 5 per 1000 devices in 2003 and decreasing to 1.5 per 1000 in 2007. The corresponding incidences for the subgroup of arthroplasty infections were 5 infections per 1319 arthroplasty procedures during 1998e1999 (revisions included), 3 per 750 arthroplasties in 2003, and 4 per 728 arthroplasties in 2007, respectively. No epidemiological link could be established between CoNS infections observed during 2003, and antibiotic susceptibility patterns revealed different patterns associated with different surgeons or procedures. The proportion of meticillin resistance remained stable throughout the study period (Figure 1). The corresponding proportions among CoNS revealed no significant trend throughout the entire study period (P ¼ 0.65) (Figure 1). Table II compares the antibiotic susceptibility patterns between CoNS strains responsible for orthopaedic infections in the current cohort and those identified hospital-wide in 2007. As shown, resistance patterns for all antibiotics were quite similar, with the exception of meticillin resistance which was markedly more prevalent among strains identified from orthopaedic infections. In 2008, which was outside
Table I Comparison between meticillin-susceptible and meticillin-resistant orthopaedic coagulase-negative staphylococcus infections and parameters potentially associated with meticillin-resistant infection (University Hospitals of Geneva cohort study, 1995e2007) Associated variables Female Age >70 years Immune suppression/comorbiditiesb Emergency surgery Adequate antibiotic prophylaxis Inadequate/missing antibiotic prophylaxis Polyclonal infection Median no. of prior hospitalizations 1e2 prior hospitalization(s) Delay between admission and surgery (median)
Meticillin-susceptible (N ¼ 15) 7 6 2 6 2 4 3 2 9 3
(47%) (40%) (13%) (40%) (13%) (27%) (20%) (range: 0e4) (60%) days
Meticillin-resistant N ¼ 45 25 26 14 20 11 28 9 3 20 2
(56%) (57%) (31%) (44%) (24%) (62%) (20%) (range: 0e21) (44%) days
OR, odds ratio; CI, confidence interval. a Estimated by univariate logistic regression. b Diabetes mellitus, alcohol abuse, chronic renal failure, cancer, rheumatic polyarthritis, cirrhosis, solid organ transplant.
ORa (95% CI) 0.7 0.5 0.3 0.8 0.5 0.2 1.0 1.2 1.8 1.0
(0.2e2.7) (0.1e1.9) (0.1e1.9) (0.2e3.2) (0.1e2.7) (0.1e0.9) (0.1e5.0) (1.0e1.4) (0.5e7.5) (1.0e1.0)
P-value 0.55 0.23 0.31 0.76 0.49 0.03 1.00 0.13 0.30 0.68
I. Uçkay et al. / Journal of Hospital Infection 79 (2011) 248e253
251
Table II Antibiotic resistance among all coagulase-negative staphylococci (except Staphylococcus saprophyticus) isolates compared to the resistance pattern of coagulase-negative staphylococci responsible for orthopaedic infections (University of Geneva Hospitals) Antibiotic class
Beta-lactams Aminoglycosides Quinolones Lincosamides Others Glycopeptides
Antibiotic
Hospital-wide (N ¼ 950)
Orthopaedic (N ¼ 60)
(2007)
(1995e2007)
84% 59% 32% 47% 56%b 56% 38% 6% 0%
NA 78% 45% 47% 45%b NA 36% 9% 0%
Penicillin Meticillin Gentamycin/tobramycin Ciprofloxacin Clindamycin Erythromycin Cotrimoxazole Rifampin Vancomycin
OR (95% CI)
e 2.5 1.7 1.0 0.6 e 0.9 1.4 1.0
(1.3e5.1) (1.0e3.0) (0.6e1.7) (0.4e1.1) (0.5e1.7) (0.4e3.7)
P-valuea
e 0.01 0.04 0.95 0.10 e 0.84 0.47 1
OR, odds ratio; CI, confidence interval; NA, not available. a Two-sided chi-squared test or Fisher’s exact test, as appropriate. b Microbiological susceptibility to clindamycin in the presence of erythromycin resistance was interpreted as potential resistance to clindamycin.
the study period, meticillin resistance among all hospital-wide clinical CoNS isolates further decreased to 56% (data not shown). Discussion The overall cumulative incidence of deep orthopaedic surgical site infections due to CoNS is low at our institution (1.4 per 1000 surgical procedures) and almost exclusively implant-related. Importantly, the proportion of meticillin-resistant strains remained unchanged over more than a decade, despite a continuous increase in implant-related procedures. CoNS strains infecting orthopaedic sites were no more resistant than overall CoNS strains identified at HUG, with the exception of meticillin resistance. Few (4/15) meticillin-susceptible strains were associated with prior prophylaxis that covered the causative pathogen, but 28/45 meticillin-resistant strains were associated with inadequate prophylaxis; the difference was statistically significant (P ¼ 0.03). First- and second-generation cephalosporins are generally used in
surgery because of good intrinsic activity against staphylococci and streptococci, fewer side-effects, and low costs.18e20 However, by definition, meticillin-resistant CoNS strains, the predominant CoNS in most healthcare institutions including HUG, are resistant to all cephalosporins.4,6,21 The delay between hospital admission and surgery and the number of previous hospitalizations was not associated with meticillin-resistant infection. It remains unknown when patients become colonized with meticillin-resistant CoNS following hospital admission. In a Swedish study, most patients in an orthopaedic ward were colonized with meticillin-resistant CoNS within 14 days following admission.10 Importantly, resistance to other antibiotic classes also increased, independent of prior antibiotic exposure. In a prospective survey among surgical patients in Switzerland, only 20% of CoNS sampled from the future surgical incision site were meticillin resistant on admission, but this rose to 47% immediately after elective surgery on the same site.21 In the current study, the delay between admission and surgery did not seem to play a role.
16
4,500 3,487
14 1 12
3,523
3,873 3,552
3,243
4,000 3,500
3 2,500 8 6
2
5
13
3
2,000
No. of procedures
No. of infections
3,000 10
1,500 8
4
7
5
1,000 5
2
500 0
0 1998–1999
2000–2001
2002–2003
2004–2005
2006–2007
Figure 1. Secular trends in orthopaedic implant procedures and infections due to meticillin-susceptible (white bars) and meticillin-resistant (black bars) coagulase-negative staphylococci, University of Geneva Hospitals, 1998e2007. Infections are indicated according to the year of implantation. Follow-up closed 31 July 2010. Median follow-up: 5.1 years.
252
I. Uçkay et al. / Journal of Hospital Infection 79 (2011) 248e253
Our finding that microbiologically appropriate antibiotic prophylaxis was the single protective parameter against meticillinresistant strains in established CoNS infection, including the fact that CoNS strains were more frequently resistant to meticillin than other CoNS isolates from elsewhere in the hospital, opens the debate regarding the possible use of routine glycopeptide prophylaxis in orthopaedic surgery. Among patients colonized with MRSA, vancomycin has become the prophylaxis of choice in many institutions.22,23 Nevertheless, we do not consider that routine screening for meticillin-resistant CoNS or vancomycin prophylaxis against meticillin-resistant CoNS would be warranted at HUG for the following reasons:22,24 (a) a significant proportion of CoNS strains are still susceptible to beta-lactam antibiotics (Table II); (b) although not the case at present, vancomycin resistance among enterococci and CoNS may emerge in the future;11 (c) vancomycin needs slow infusion to avoid excessive histamine release, responsible for the so-called ‘red man’ syndrome, which renders its use more difficult in prophylaxis where accurate timing of administration is of utmost importance;25 (d) the cumulative incidence of orthopaedic infections due to meticillin-susceptible or -resistant CoNS at HUG is very low and does not show any increasing trend, contrary to some reports;3,6 (e) routine glycopeptide prophylaxis does not guarantee decreased infection rates due to CoNS, but may increase surgical site infection due to other pathogens. A review of four randomized trials comparing prophylactic teicoplanin vs prophylactic cephalosporin in settings with a high prevalence of meticillin resistance among S. epidermidis showed similar infection rates among both groups.26 This has also been confirmed in a meta-analysis of seven randomized trials in cardiac surgery, even though there have been single trials that switched to vancomycin prophylaxis in cardiac surgery.27,28 A recent systematic review and economic model of switching from non-glycopeptide to glycopeptide antibiotic prophylaxis for surgery in endemic MRSA settings failed to show increased efficacy in preventing surgical site infection due to meticillin-resistant strains. Even for MRSA, there is insufficient evidence to determine whether there is a threshold prevalence to justify a switch to routine glycopeptide prophylaxis.29 We consider that the same arguments are equally valid regarding vancomycin-containing cements for primary prophylaxis in arthroplasty surgery. Our study has some limitations. (i) Since it was retrospective and conducted in a single centre, the generalizability of our findings deserves further investigation. (ii) It includes only 60 CoNS infections, although they were identified in a large cohort of 13 years and with a median follow-up of 5.1 years. (iii) Complete case review for CoNS infection was retrospective, thus additional parameters associated with meticillin resistance acquisition may have been missed. (iv) Patients with an infection treated in another hospital may not have been detected. However, since HUG is the largest and the only public hospital in the region, we consider such a detection bias as minimal. (v) We used microbiologically documented databases that could imply an additional possible selection bias for infections where no strains grew in culture. However, taking into account the fact that we also used the HUG administration coding system, medical reports, and prospective cohort follow-up, we consider this bias as minimal. A high proportion of infections were due to meticillinresistant strains, which are unlikely to be masked by unreported exposure to antibiotics prior to admission. In the literature, culturenegative, prosthetic joint infections account for only 7% of cases and the outcome is indistinguishable from that due to identified bacterial pathogens.30 Another problem of microbiologically based, culturepositive definitions of infections is the potential selection of grown pathogens through pre-admission antibiotic therapy. In particular, CoNS may be selected through standard empirical antibiotic regimens, such as amoxicillineclavulanic acid or cephalosporins. In such
a case, the co-pathogen, e.g. S. aureus, would not be detected in intraoperative samples and the case wrongly attributed to CoNS infection only. Therefore, we established an exclusion criterion of avoiding cases with empirical antibiotic therapy within six weeks prior to admission. The cut-off of six weeks was arbitrary and only four cases were excluded from final analysis because of that criterion. This exclusion did not influence overall results. Moreover, as these cases were pretreated with antibiotics upon admission, they also showed growth of CoNS either in enrichment broth only or only in a single intraoperative sample. In other words, these four patients had several reasons to be excluded. (vi) In the literature reporting orthopaedic implant-related infections, the ideal follow-up time is unknown. The Healthcare Infection Control Practices Advisory Committee guidelines for the prevention of surgical site infections and the United States Centers for Disease Control and Prevention definitions for implant-associated infections suggest a minimal surveillance of one year for implants.23,31 Hence, we consider actively managed, individual minimal and median follow-up times of 2.4 years and 5.1 years, respectively, as an important strength of the current study. (vii) We addressed only substantial orthopaedic CoNS infection for which the patients underwent surgery. Small superficial CoNS infections during hospitalization or after discharge may have remained unrecorded. Similarly, deep infections under ongoing antibiotic treatment on admission, which could have been only due to CoNS, were excluded according to our study definitions. Despite these limitations and due to the fact that we used four different databases (Arthroplasty Cohort, Septic Orthopaedic Cohort, microbiological laboratory, and the hospital’s coding system), we are confident that we have included all severe and clinically significant CoNS infections. (viii) Our study addressed purely epidemiological issues and related perioperative antibiotic prophylaxis. Therapeutic approaches for CoNS infections were beyond the topic of this study and we recommend the readers to consult other reports.4,32 In conclusion, orthopaedic CoNS infections are rare and almost exclusively implant-related. CoNS are mostly resistant to beta-lactam antibiotics used as prophylaxis. In our institution, the cumulative incidence and the proportion of meticillin resistance has remained stable during the last decade.32 A routine change to glycopeptide prophylaxis or glycopeptide-containing cement in arthroplasty surgery is not warranted for orthopaedic procedures in our setting. Acknowledgements We are indebted to E. Huggler, Laboratory of Infectious Diseases, and C. Béant and Dr P. Tahintzi, Coding Office, for their help, to Dr A. Gayet-Ageron for her advice, and to R. Sudan for editorial assistance. Thanks also to the teams of the Orthopaedic Surgery Service and the Laboratory of Bacteriology for their support. Conflict of interest statement None declared. Funding sources None. References 1. Kloos WE, Bannerman TL. Update on clinical significance of coagulase-negative staphylococci. Clin Microbiol Rev 1994;7:117e140. 2. Segawa H, Tsukayama DT, Kyle RF, Becker DA, Gustilo RB. Infection after total knee arthroplasty. A retrospective study of the treatment of eighty-one infections. J Bone Joint Surg Am 1999;81:1434e1445. 3. Fulkerson E, Valle CJ, Wise B, Walsh M, Preston C, Di Cesare PE. Antibiotic susceptibility of bacteria infecting total joint arthroplasty sites. J Bone Joint Surg Am 2006;88:1231e1237.
I. Uçkay et al. / Journal of Hospital Infection 79 (2011) 248e253 4. Uçkay I, Pittet D, Vaudaux P, Sax H, Lew D, Waldvogel F. Foreign body infections due to Staphylococcus epidermidis. Ann Med 2009;41:109e119. 5. Adler H, Widmer A, Frei R. Emergence of a teicoplanin-resistant small colony variant of Staphylococcus epidermidis during vancomycin therapy. Eur J Clin Microbiol Infect Dis 2003;22:746e748. 6. Mohanty SS, Kay PR. Infection in total joint replacements. Why we screen MRSA when MRSE is the problem? J Bone Joint Surg Br 2004;86:266e268. 7. James PJ, Butcher IA, Gardner ER, Hamblen DL. Methicillin-resistant Staphylococcus epidermidis in infection of hip arthroplasties. J Bone Joint Surg Br 1994;76:725e727. 8. Sanzen L, Walder M. Antibiotic resistance of coagulase-negative staphylococci in an orthopaedic department. J Hosp Infect 1988;12:103e108. 9. Joshy S, Gogi N, Thomas B, Mahale A, Singh BK. Delayed onset of deep infection after total knee arthroplasty: comparison based on the infecting organism. J Orthop Surg (Hong Kong) 2007;15:154e158. 10. Thore M, Kuhn I, Lofdahl S, Burman LG. Drug-resistant coagulase-negative skin staphylococci. Evaluation of four marker systems and epidemiology in an orthopaedic ward. Epidemiol Infect 1990;105:95e105. 11. National Nosocomial Infections Surveillance (NNIS) System Report 2004. Am J Infect Control 2004;32:470e485. 12. Widerstrom M, Monsen T, Karlsson C, Wistrom J. Molecular epidemiology of methicillin-resistant coagulase-negative staphylococci in a Swedish county hospital: evidence of intra- and interhospital clonal spread. J Hosp Infect 2006;64:177e183. 13. Kitao T. Survey of methicillin-resistant coagulase-negative staphylococci isolated from the fingers of nursing students. J Infect Chemother 2003;9:30e34. 14. Lübbeke A, Stern R, Garavaglia G, Zürcher L, Hoffmeyer P. Differences in outcomes of obese women and men undergoing primary total hip arthroplasty. Arthritis Rheum 2007;57:327e334. 15. Yılmaz G, Aydın K, Iskender S, Caylan R, Köksal I. Detection and prevalence of inducible clindamycin resistance in staphylococci. J Med Microbiol 2007; 56:342e345. 16. Clinical and Laboratory Standards Institute. Performance standards for antimicrobial susceptibility testing. 17th informational supplement (standard M100-S17). Wayne, PA: CLSI; 2007. 17. Peduzzi P, Concato J, Feinstein AR, Holford TR. Importance of events per independent variable in proportional hazards regression analysis. II. Accuracy and precision of regression estimates. J Clin Epidemiol 1995;48:1503e1510. 18. Petignat C, Francioli P, Harbarth S, et al. Cefuroxime prophylaxis is effective in herniated disk surgery: a double-blind, placebo-controlled study. Spine 2008;15:1919e1924.
253
19. Trampuz A, Zimmerli W. Antimicrobial agents in orthopaedic surgery: prophylaxis and treatment. Drugs 2006;66:1089e1105. 20. Hill C, Flamant R, Mazas F, Evrard J. Prophylactic cefazolin versus placebo in total hip replacement. Report of a multicentre double-blind randomised trial. Lancet 1981;1:795e796. 21. Zingg W, Imhof A, Senn G, Demartines N, Ruef C. Rapid colonization with methicillin-resistant coagulase-negative staphylococci after surgery. World J Surg 2009;33:2058e2062. 22. Lew D, Schrenzel J, Francois P, Vaudaux P. Pathogenesis, prevention, and therapy of staphylococcal prosthetic infections. Curr Clin Top Infect Dis 2001; 21:252e270. 23. Mangram AJ, Horan TC, Pearson ML, Silver LC, Jarvis WR. Guideline for prevention of surgical site infection, 1999. Hospital Infection Control Practices Advisory Committee. Infect Control Hosp Epidemiol 1999;20: 250e278. 24. Uçkay I, Vernaz-Hegi N, Harbarth S, et al. Activity and impact on antibiotic use and costs of a dedicated infectious diseases consultant on a septic orthopaedic unit. J Infect 2009;58:205e212. 25. Classen DC, Evans RS, Pestotnik SL, Horn SD, Menlove RL, Burke JP. The timing of prophylactic administration of antibiotics and the risk of surgical-wound infection. N Engl J Med 1992;326:281e286. 26. Mini E, Nobili S, Periti P. Methicillin-resistant staphylococci in clean surgery. Is there a role for prophylaxis? Drugs 1997;54:39e52. 27. Bolon MK, Morlote M, Weber SG, Koplan B, Carmeli Y, Wright SB. Glycopeptides are no more effective than beta-lactam agents for prevention of surgical site infection after cardiac surgery: a meta-analysis. Clin Infect Dis 2004;38:1357e1363. 28. Garey KW, Lai D, Dao-Tran TK, Gentry LO, Hwang LY, Davis BR. Interrupted time series analysis of vancomycin compared to cefuroxime for surgical prophylaxis in patients undergoing cardiac surgery. Antimicrob Agents Chemother 2008;52:446e451. 29. Cranny G, Elliott R, Weatherly H, et al. A systematic review and economic model of switching from non-glycopeptide to glycopeptide antibiotic prophylaxis for surgery. Health Technol Assess 2008;12:1e168. 30. Berbari EF, Marculescu C, Sia I, et al. Culture-negative prosthetic joint infection. Clin Infect Dis 2007;45:1113e1119. 31. Garner JS, Jarvis WR, Emori TG, Horan TC, Hughes JM. CDC definitions for nosocomial infections. Am J Infect Control 1988;16:128e140. 32. Teterycz D, Ferry T, Lew D, et al. Outcome of orthopedic implant infections due to different staphylococci. Int J Infect Dis 2010;14:913e918.