INTRODUCTION
Postoperative infection complicates 1 to 4% of primary arthroplasties and an even higher proportion in revision surgery.1 The costs of treating infected total knee arthroplasties are 3 to 4 times higher than primary arthroplasties, and twice as high as aseptic revision arthroplasties.2 Prophylactic administration of systemic antibiotics reduces infection rates. According to the Norwegian Arthroplasty Register, the revision rate for infections in patients not receiving any prophylactic antibiotics is 12 times higher than those receiving both antibiotic-containing cement and systemic antibiotics.3 The efficacy of prophylaxis depends on the susceptibility of bacteria to the antibiotics. Outcome is worse in the presence of multi-resistant pathogens.4 We aimed to classify the spectrum and antibiotic susceptibility of bacteria isolated from revision hip and knee arthroplasty specimens, and recommend appropriate empiric perioperative antibiotics before definitive culture results are available.
MATERIALS AND METHODS
From January 1999 to August 2006, specimens from revision hip and knee arthroplasties (with or without suspected infection) were routinely collected for identifying possible organisms and their susceptibility patterns. About 550 primary and 80 revision hip and knee arthroplasty procedures were performed in our hospital a year. During the study period, 147 patients undergoing revision arthroplasty had positive specimens that yielded 248 micro-organisms (from 195 tissue specimens, 43 fluid specimens, and 10 swabs). 140 isolates were from hips and 108 from knees. To avoid multiple listings from a single infection, subsequent sets of cultures taken from patients with repeated reoperations were excluded from analysis whenever the same organisms were isolated.
All tissue and fluid specimens were processed within a class-II biohazard cabinet. They were inoculated and incubated for 7 days at 37°C in CO2 conditions. All swab specimens were inoculated and incubated for 48 hours in CO2 conditions. Anaerobic cultures of specimens were incubated for 48 hours and Gram-stained.
All isolates were identified using the Microscan Walkaway (Dade), Vitex 2 (Biomerieux), API (Biomerieux), or Rapid ID (Remel). Most sensitivity tests were performed by the Microscan or Vitex 2, the rest by disc diffusion methods based on the Clinical Laboratory Standards Institute guidelines. Non-standardised sensitivity tests were performed on some genus groups; related guidelines with comments were used for susceptibility reporting on such cases.
RESULTS
Most of the isolates were Gram-positive; their distribution was similar in hip and knee specimens (Table 1). Of the 248 micro-organisms isolated, Staphylococcus was the most common genus encountered (131, 53%), in which coagulase-negative Staphylococcus accounted for 60% and Staphylococcus aureus 40%. In all 24% of isolates were Gram-negative. 88% of Gram-negative organisms were identified within 48 hours of inoculation and 94% of Grampositive organisms within 96 hours (Table 2).
Only 114 (46%) of the 248 isolates were susceptible to cephalothin (Table 3). 123 (50%) isolates were tested or predicted to be resistant to cephalothin; 11 isolates were not tested for cephalothin susceptibility. Six of the 53 S aureus isolates were methicillin-resistant. Only 35% of the coagulasenegative staphylococcal isolates were susceptible to cephalothin. All methicillin-sensitive S aureus isolates were susceptible to both cephalothin and cefazolin. All Gram-positive organisms were susceptible to vancomycin, but 2 methicillin-resistant S aureus isolates expressed vancomycin heteroresistance. 43% of Enterobacteriaceae were susceptible to cephalothin and cefazolin and all Pseudomonas aeruginosa isolates are considered intrinsically resistant to these agents. Bacteria capable of producing inducible class-1 betalactamases were considered resistant to the first, second, and third generations of cephalosporins, regardless of the in vitro results. These were known as the ESCAHPM organisms, which refer to Enterobacter, Serratia, Citrobacter freundii, Aeromonas spp, Hafnia alvei, Providencia spp, and Morganella morganii. Gentamicin and tobramycin had excellent activity against Gram-negative organisms. Combining coliforms and P aeruginosa together, 93% were susceptible to gentamicin and 96% to ciprofloxacin. Gentamicin had in vitro activity against 65% of staphylococcal species. Gram-negative organisms are intrinsically resistant to vancomycin. Of the 248 isolates, 96% would have been susceptible to the combination of vancomycin and gentamicin.
DISCUSSION
To prevent infections, a surveillance system should be established in each institution to monitor rates of procedure-specific infections, changing ecology of resistant pathogens, and susceptibility patterns to antimicrobials. Guidelines based on local microbiological epidemiology should be developed.
Peri-operative antibiotics are given to treat an existing infection and/or prevent a new infection. In our study, all isolates whether from the arthroplasty or the operative field (e.g. skin contaminants) were considered relevant to the choice of empiric antibiotics before and after revision arthroplasty.
In a study on organisms from deep infections, a similar prevalence of staphylococcal infections and high rates of first-generation cephalosporin resistance among coagulase-negative staphylococci were reported.5 Other studies have reported variable cephalosporin sensitivities of coagulase-negative staphylococci.6-8 In our study, Gram-negative organisms were more commonly present in operative cultures.
A first-generation cephalosporin alone can no longer cover the organisms likely to be encountered at revision arthroplasty. Empiric prophylactic antibiotics should include vancomycin to cover Gram-positive organisms and gentamicin to cover most Gramnegative bacteria. Prophylactic antibiotics should not be continued for >24 hours (a time course equivalent to 4 doses of many first-generation cephalosporins) if the risk of infection is low. If infection cannot be excluded, vancomycin and gentamicin should be commenced preoperatively and continued postoperatively for 96 and 48 hours, respectively, unless culture or histology results suggest otherwise.
Because few Gram-negative organisms were isolated after 48 hours of incubation, intravenous gentamicin was recommended as a single daily dose on the day of surgery and for 2 more days to minimise the risk of renal or vestibular/auditory side-effects. If renal function is normal, the single daily dose of gentamicin is 5 mg/kg in patients aged <60 years and 4 mg/kg in those aged >60 years.9 Nephrotoxicity is rare even in the elderly if gentamicin is given for <4 days.10 If a Gram-negative organism is isolated from operative cultures, an alternative with a more acceptable side-effect profile can be selected for longterm use.
As 94% of Gram-positive isolates were detected within 96 hours of incubation; routine cessation of vancomycin is recommended if no such organism is detected. It is imperative that vancomycin be used appropriately to reduce the emergence of vancomycinresistant strains.11 To avoid infusion-related toxicity ('red man syndrome'), it must be given over 1 to 2 hours. Careful coordination is needed to ensure all the recommended dosage is delivered prior to tourniquet application or skin incision. The initial dosage is 25 mg/kg up to 1 g every 12 hours.9 A predose level should be checked on the postoperative day 2 and adjusted if necessary to achieve a target blood concentration of 10 to 20 mg/l.9,12
As both vancomycin and gentamicin need to be dose- and/or dose interval-adjusted in renal failure patients, infectious disease expertise should be sought. In our hospital, the third-generation cephalosporins do not adequately cover the Gramnegative organisms likely to be encountered. A fourthgeneration cephalosporin (e.g. cefepime) should be used.
© 2008 Western Pacific Orthopaedic Association Provided by ProQuest LLC. All Rights Reserved.
Source: Journal of Orthopaedic Surgery
