The value of intraoperative gram stain in revision total knee arthroplasty

Transcription

1 Washington University School of Meicine Digital Open Access Publications The value of intraoperative gram stain in revision total knee arthroplasty Patrick M. Morgan Washington University School of Meicine in St. Louis Peter Sharkey Thomas Jefferson University Meical School Elie Ghanem Thomas Jefferson University Meical School Java Parvizi Thomas Jefferson University Meical School John C. Clohisy Washington University School of Meicine in St. Louis See next page for aitional authors Follow this an aitional works at: Part of the Meicine an Health Sciences Commons Recommene Citation Morgan, Patrick M.; Sharkey, Peter; Ghanem, Elie; Parvizi, Java; Clohisy, John C.; Burnett, R. Stephen J.; an Barrack, Robert L.,,"The value of intraoperative gram stain in revision total knee arthroplasty." The Journal of Bone an Joint Surgery.91, (2009). This Open Access Publication is brought to you for free an open access by Digital It has been accepte for inclusion in Open Access Publications by an authorize aministrator of Digital For more information, please contact

2 Authors Patrick M. Morgan, Peter Sharkey, Elie Ghanem, Java Parvizi, John C. Clohisy, R. Stephen J. Burnett, an Robert L. Barrack This open access publication is available at Digital

3 2124 COPYRIGHT Ó 2009 BY THE JOURNAL OF BONE AND JOINT SURGERY, INCORPORATED The Value of Intraoperative Gram Stain in Revision Total Knee Arthroplasty By Patrick M. Morgan, MD, Peter Sharkey, MD, Elie Ghanem, MD, Java Parvizi, MD, FRCS, John C. Clohisy, MD, R. Stephen J. Burnett, MD, FRCS(C), an Robert L. Barrack, MD Investigation performe at the Washington University School of Meicine, St. Louis, Missouri, an Thomas Jefferson University Meical School, Philaelphia, Pennsylvania Backgroun: The accurate preoperative iagnosis of infection is an essential component of ecision-making prior to revision total knee arthroplasty. When preoperative moalities use to etect infection reveal equivocal finings, the surgeon may rely on intraoperative testing. While intraoperative Gram stains are routinely performe uring revision total knee arthroplasty, their value remains unclear. Methos: We retrospectively reviewe the recors on 945 revision total knee arthroplasties performe at three university institutions to which patients were referre for total joint arthroplasty; the results of an intraoperative Gram stain were available for review in 921 cases (97.5%). Of these knees, 247 were classifie as infecte on the basis of (1) the presence of the same organism in two cultures; (2) growth, on soli meia, of an organism as well as other objective evience of infection; (3) histologic evience of acute inflammation; (4) gross purulence; an/or (5) an actively raining sinus. We reviewe the results of preoperative laboratory stuies, which inclue measurements of the erythrocyte seimentation rate, C-reactive protein values, an white bloo-cell count in 90%, 76%, an 98% of cases, respectively. Preoperative aspiration to obtain a specimen for culture an a cell count was performe routinely at one center an selectively at the other two centers, an the results were available for review in 439 (48%) of the 921 cases. Results: Intraoperative Gram staining was foun to have a sensitivity of 27% an a specificity of 99.9%. The positive an negative preictive values were 98.5% an 79%, respectively. The test accuracy was 80%. Patients with a truepositive Gram stain ha a significantly higher preoperative white bloo-cell count, C-reactive protein level, an nucleate cell count in the aspirate when compare with patients with a false-negative Gram stain (p < 0.001). In no case i the results of the intraoperative Gram stain alter treatment. Conclusions: The intraoperative Gram stain was foun to have poor sensitivity an a poor negative preictive value, an its results i not alter the treatment of any patient unergoing revision total knee arthroplasty because of a suspecte infection. These ata o not support the routine use of intraoperative Gram staining in revision total knee arthroplasty; instea, they suggest a much more limite role for this test. Level of Evience: Diagnostic Level I. See Instructions to Authors for a complete escription of levels of evience. Total knee arthroplasty is a successful an effective surgical treatment for arthritis, with reporte survivorship an patient satisfaction rates of >90% at ten to fifteen years 1,2. A knee arthroplasty that oes fail, however, poses a management ilemma. Paramount to choosing an appropriate treatment strategy is the correct ientification of the cause of failure. While noninfectious etiologies such as loosening, instability, an malalignment are responsible for the majority of total knee revisions, infection continues to be the reason for a substantial percentage of revisions an has been reporte to be Disclosure: In support of their research for or preparation of this work, one or more of the authors receive, in any one year, outsie funing or grants in excess of $10,000 from Stryker Orthopaeics, Smith an Nephew Orthopaeics, an the Orthopaeic Founation at the Rothman Institute. Neither they nor a member of their immeiate families receive payments or other benefits or a commitment or agreement to provie such benefits from a commercial entity. A commercial entity (Stryker Orthopaeics) pai or irecte in any one year, or agree to pay or irect, benefits in excess of $10,000 to a research fun, founation, ivision, center, clinical practice, or other charitable or nonprofit organization with which one or more of the authors, or a member of his or her immeiate family, is affiliate or associate. J Bone Joint Surg Am. 2009;91: oi: /jbjs.h.00853

4 2125 TABLE I Preoperative Testing Performe on Patients with an Intraoperative Gram Stain No. of Patients with Test Result of Gram Stain White Bloo-Cell Count Erythrocyte Seimentation Rate C-Reactive Protein Aspiration True-positive (67 patients) False-negative (180 patients) True-negative (673 patients) False-positive (1 patient) Total (921 patients) 903 (98%) 831 (90%) 697 (76%) 439 (48%) the single most common cause of early failure 3-5. Two-stage exchange arthroplasty is generally consiere the preferre metho of treatment of periprosthetic joint infection 6-9. A preoperative evaluation that incorrectly ientifies the presence or absence of infection can lea to either inappropriate surgical intervention or a elay in appropriate treatment. Currently, however, there is no gol stanar for the iagnosis of an infection at the site of a total knee arthroplasty. An equivocal result of a preoperative workup may, therefore, require the surgeon to epen on intraoperative tests to etermine that infection is absent 10. While the efficacy of frozensection analysis has been establishe by several authors 11-13, the value of an intraoperative Gram stain remains unclear. Although the test is routine at many centers, its sensitivity has been consistently reporte to be low an previous stuies have been hampere by small sample sizes or by the authors combining the results of both hip an knee revision surgery This stuy was esigne to etermine the sensitivity, specificity, accuracy, an positive an negative preictive values of intraoperative Gram stains in a large cohort of patients unergoing revision total knee arthroplasty. We also unertook a review of the preoperative an intraoperative finings to etermine whether there were instances in which the intraoperative Gram stain may have provie valuable information to the surgeon. In oing so, we hope to ientify the role that an intraoperative Gram stain plays in etermining the correct course of treatment. Materials an Methos The ata for this stuy were prospectively gathere at three university-affiliate institutions over a six-year perio an entere into an institutional review boar-approve, multiinstitution atabase of information on all revision total knee arthroplasties. This atabase was querie retrospectively for information relevant to the stuy question. Routine preoperative testing inclue measurement of the erythrocyte seimentation rate, C-reactive protein level, an white bloo-cell count (Table I). Preoperative aspiration was performe routinely at one center, an it was carrie out selectively at the other two institutions when infection was suspecte on the basis of clinical or raiographic finings. Aspirates were sent for aerobic an anaerobic culture an, when sufficient flui was available, for a cell count an ifferential. Aspiration results were available for 439 patients (48%). An intraoperative culture of synovial flui was performe in all cases, an aitional tissue was sent for culture in all but three cases. The results of the preoperative an intraoperative cultures an the meium on which the bacteria were ientifie (soli, or enhance broth) were ocumente. Clinical follow-up notes were reviewe for evience of infection. Knees were classifie as infecte when three of the five following criteria escribe by Leone an Hanssen 22 were met: (1) the presence of the same organism in two cultures, (2) growth of an organism on soli meia as well as other objective evience of infection such as elevate levels of inflammatory markers in the absence of systemic inflammatory isease or an elevate cell count an percentage of polymorphonuclear leukocytes in aspirate joint flui, (3) histologic evience of acute inflammation, (4) gross purulence at the time of surgery, or (5) an actively raining sinus. Threshol values for evience of infection base on results of bloo tests an synovial flui analysis were erive from the publishe literature an inclue an erythrocyte seimentation rate of >30 mm/hr, a C-reactive protein level of >10 mg/l, a synovial flui nucleate cell count of >1700 cells/ml, a white bloo-cell count of > /L, an a synovial flui leukocyte ifferential of >65% polymorphonuclear leukocytes 15,23. A normal result of a preoperative workup was efine as one in which all of its components were within these efine normal limits. Statistical Analysis The Mann-Whitney one-taile t test with Gaussian approximation, an calculations of sensitivity, specificity, accuracy, an positive an negative preictive values, were performe with use of GraphPa Prism, version 5.01 for Winows (Graph- Pa Software, San Diego, California). Descriptive analysis was performe with use of univariate statistics for the continuous variables an frequency istribution for the categorical variables. Results for the continuous variables are reporte as means an range istributions. Gaussian istribution was evaluate to etermine if there was a normal istribution of the ata. Chisquare analysis was carrie out to etermine the sensitivity, specificity, accuracy, an positive an negative preictive values. With use of one-taile Mann-Whitney statistics, the means an stanar eviations for the continuous variables, incluing the erythrocyte seimentation rate, C-reactive protein level, white bloo-cell count, cell count in the aspirate, an percentage of

5 2126 TABLE II 2 2 Table for Results of Gram Staining for Diagnosis of Infection During Revision Total Knee Arthroplasty Infection Confirme* Gram Stain Yes No Positive 67 true-positive 1 false-positive Negative 180 false-negative 673 true-negative *True-positive = infection present an Gram stain positive, falsepositive = infection absent an Gram stain positive, false-negative = infection present an Gram stain negative, an true-negative = infection absent an Gram stain negative. Sensitivity = true-positive/ (true-positive 1 false-negative) or true-positive/total with isease = 0.27 (95% confience interval, 0.22 to 0.33). Specificity = truenegative/(false-positive 1 true-negative) or true-negative/total without isease = (95% confience interval, 0.99 to 1.00). Positive preictive value = true-positive/(true-positive 1 false-positive) or truepositive/all with positive test = (95% confience interval, 0.92 to 1.00). Negative preictive value = true-negative/(false-negative 1 true-negative) or true-negative/all with negative test = 0.79 (95% confience interval, 0.76 to 0.82). polymorphonuclear leukocytes in the aspirate, were compare between the patients with a true-positive result an the ones with a false-negative result. The Mann-Whitney statistic is a nonparametric test an was use for this analysis as the ata were collecte from inepenent samples an were not normally istribute. The results of these comparisons are reporte as p values. Source of Funing No funing was receive specifically for this stuy; however, funs were receive in general support of the total joint registries that were the sources of the ata presente in the stuy. The funing sources were Smith an Nephew Orthopaeics an the Orthopaeic Founation at Rothman Institute. Funing for total joint research was also receive from Stryker Orthopaeics. Results The recors on 945 consecutive revision total knee arthroplasties performe over a six-year perio were reviewe; the results of an intraoperative Gram stain were available for review in 921 (97.5%) of these cases. Two hunre an forty-seven knees were classifie as infecte, an all ha an intraoperative Gram stain available for review. Of the 698 knees etermine not to be infecte, 674 ha a Gram stain available for review; the Gram stain was reporte to be negative in 673 of these cases an positive in one. In the positive case, all other tests (white bloo-cell count an measurements of the erythrocyte seimentation rate an level of C-reactive protein) emonstrate normal results an an intraoperative frozen section showe no acute inflammation. A revision was performe with no subsequent evience of infection, so the Gram stain was classifie as false-positive. Intraoperative Gram staining was foun to have a sensitivity of 27% an a specificity of 99.9%. The positive an negative preictive values were 98.5% an 79%, respectively (Table II). Intraoperative Gram staining i not influence treatment in any case. There were sixty-seven true-positive an 180 falsenegative Gram stains (Table III). The white bloo-cell count, C- reactive protein level, an aspirate cell count were significantly higher in the group of infecte knees with a true-positive Gram stain than they were in the group with a false-negative Gram stain (p < 0.001). The erythrocyte seimentation rate an ifferential cell count (percentage of polymorphonuclear leukocytes) in the joint flui aspirate i not iffer significantly between the two groups (p = 0.3 an p = 0.4, respectively) (Table III). The erythrocyte seimentation rate, C-reactive protein level, aspirate cell count, an percentage of polymorphonuclear leukocytes in the aspirate were all significantly higher in the infecte knees with a negative Gram stain (falsenegative cases) than they were in the uninfecte knees with a negative Gram stain (true-negative cases) (p < 0.01); this fining was consistent with the results in a large boy of literature 7,13,15,24. Of the sixty-seven patients with a true-positive Gram stain, only one ha normal results of the preoperative workup (white bloo-cell count, erythrocyte seimentation rate, an C-reactive protein level). A preoperative aspiration ha not been performe in that patient, an purulent flui was encountere. On the basis of the intraoperative appearance of the tissue, a frozen-section analysis was performe an it showe acute inflammation (>10 polymorphonuclear leuko- TABLE III Comparison of Preoperative Laboratory Results for Infecte Knees with Positive an Negative Intraoperative Gram Stains Parameter True-Positive* (N = 67) False-Negative* (N = 180) P Value White bloo-cell count ( 10 9 /L) 11.8 ± ( ) 8.6 ± ( ) <0.001 Erythrocyte seimentation rate (mm/hr) ± (4-141) 72.7 ± (1-140) 0.3 C-reactive protein (mg/l) 22.8 ± ( ) ± ( ) <0.001 Aspirate nucleate cell count (cells/ml) 60,000 ± 120,000 ( ,000) 27,600 ± 100,000 ( ,000) Aspirate polymorphonuclear leukocytes (%) ± (17-100) 85.7 ± (4-99) 0.4 *The values are given as the mean an stanar eviation with the range in parentheses.

6 2127 TABLE IV Results of Intraoperative Gram Staining for the Detection of Periprosthetic Infection as Reporte in the Literature Authors Year No. of Cases Site Sensitivity (%) Specificity (%) Athanasou et al Hip/knee Atkins et al Hip/knee Barrack et al Knee Bauer et al Hip/knee 22 an 35* 100 Chimento et al Hip/knee 0 0 Della Valle et al Hip/knee Felman et al Hip/knee Ko et al Hip/knee 0 0 Kraemer et al Hip Paney et al Hip Spangehl et al Hip *Tissue an flui, respectively. cytes per high-power fiel). A two-stage exchange arthroplasty was performe. Discussion Describe by Hans Christian Gram in 1884, the Gram stain exploits biochemical ifferences between bacterial cell walls to broaly classify many bacteria as either gram-positive or gram-negative 25. This categorization is base on a number of morphological characteristics of the bacterium, incluing the relative thickness of the bacterial peptioglycan layer an the presence or absence of an outer membrane. We are not aware of any available ata concerning the average bacterial loa seen within the tissues of an infection at the site of a total knee arthroplasty. Periprosthetic infection often occurs with a low organism buren in the synovial flui. This is influence by the formation of biofilms, which have a higher organism buren, an this may in part explain the variable sensitivity of Gram stains reporte for ifferent organisms 26,27. The results of the present stuy inicate that an intraoperative Gram stain is more likely to be positive in the setting of a more fulminant periprosthetic knee infection, a hypothesis supporte by the fact that the white bloo-cell count, C-reactive protein level, an aspirate cell count were significantly higher in cases with a true-positive Gram stain than they were in those with a falsenegative stain. A Gram stain is commonly performe on operatively retrieve specimens as a means of screening for the presence of infection 24,28, but this practice appears to be of questionable value. Previous stuies of relatively small cohorts have shown Gram stains to have poor sensitivity, which has been as low as 0% in some reports (Table IV). Inee, previous authors who reviewe a mixe cohort of hip an knee revisions questione the value of an intraoperative Gram stain an suggeste that the test shoul not be orere on a routine basis 17. The sensitivity of an intraoperative Gram stain wasalsolow(27%)inourseriesof921revisiontotalknee arthroplasties. Preoperative planning of the treatment of a faile total knee arthroplasty epens in part on the results of a preoperative workup, in which ientification of infection is important. The intraoperative use of a Gram stain ha little or no iagnostic role for the patients with positive results of the preoperative workup for infection in our series. The sensitivity of the intraoperative Gram stain also was too low to be consiere reliable for the patients with equivocal results of the preoperative workup. In aition, the Gram stain prove to be of no value for the patients with completely normal results of the preoperative workup since no true-positive cases were ientifie within this groupantherewasonefalse-positivecase. There is a potential for substantial variability in the interpretation of the Gram stain by the laboratory personnel performing the test. There is the potential for error in both the staining process an the interpretation of the slies, problems that may have contribute to the prevalence of false-negative results observe in this stuy. False-positive cases were rare in this series (only one case), but less experience technicians coul overinterpret Gram-stain results, leaing to more falsepositive finings. Specimen contamination can also result in a false-positive fining. Previous investigators have reporte that intraoperative Gram staining has a very high specificity for a positive result 17,21. Our stuy confirms this fining. Of the sixty-seven patients with a true-positive Gram stain, only one ha normal finings on the preoperative workup, which was incomplete. The patient was an eighty-year-ol man who was note to have purulent joint flui at the time of surgery; a preoperative joint aspiration ha not been performe. An intraoperative frozen-section analysis was performe, an reimplantation was aborte because of the gross purulence. Therefore, the intraoperative Gram stain i not alter the treatment of any patient.

7 2128 The stuy ha some weaknesses. First, as a result of its retrospective esign, there may have been some variability in ata collection. Secon, because it was a multi-institutional stuy, it is possible that the protocols for the workup an management iffere among the patients. In fact, this was the case with respect to aspiration of the joint. All knees scheule to unergo revision arthroplasty were aspirate routinely in one institution an only selectively in the others. This resulte in a lack of availability of aspiration ata for some patients. Also, because of ifferences in the workup of these patients, serological tests such as measurement of the C- reactive protein level were not performe for all patients. Such variability in the management of these patients presents the possibility of bias; specifically, it is possible that the relative value of iagnostic tests other than the Gram stain may have been incorrectly exaggerate or iminishe. This stuy, however, was not esigne or intene to establish the value of either measurement of the C-reactive protein level or preoperative aspiration in the iagnosis of periprosthetic infection. A complete preoperative workup for periprosthetic infection in a patient with a faile total knee arthroplasty inclues serological testing, synovial flui analysis, an raiographic imaging. We consier preoperative aspiration to be particularly useful. The accuracy of preoperative aspiration for the iagnosis of infection prior to revision of a faile total knee arthroplasty has been reporte to be very high, approaching 100% in some series 13,29 but generally ranging from 75% to 80% if the patient is not being treate with antibiotics 15. In our series, in which strict criteria were employe for the efinition of periprosthetic infection, intraoperative Gram staining playe no role in the iagnosis of infection in patients who ha ha a full preoperative workup an the selective use of intraoperative frozen-section analysis. We suggest that the practice of routinely performing a Gram stain at the time of revision total knee arthroplasty may safely be abanone. n Patrick M. Morgan, MD John C. Clohisy, MD R. Stephen J. Burnett, MD, FRCS(C) Robert L. Barrack, MD Department of Orthopaeic Surgery, Washington University School of Meicine, One Barnes-Jewish Hospital Plaza, West Pavilion, St. Louis, MO aress for R.L. Barrack: barrackr@wustl.eu Peter Sharkey, MD Elie Ghanem, MD Java Parvizi, MD, FRCS Rothman Institute of Orthopaeics, Thomas Jefferson University Meical School, 925 Chestnut Street, Philaelphia, PA References 1. Colizza WA, Insall JN, Scueri GR. The posterior stabilize total knee prosthesis. Assessment of polyethylene amage an osteolysis after a ten-year-minimum follow-up. J Bone Joint Surg Am. 1995;77: Font-Roriguez DE, Scueri GR, Insall JN. Survivorship of cemente total knee arthroplasty. Clin Orthop Relat Res. 1997;345: Sharkey PF, Hozack WJ, Rothman RH, Shastri S, Jacoby SM. Why are total knee arthroplasties failing toay? Clin Orthop Relat Res. 2002;404: Fehring TK, Oum S, Griffin WL, Mason JB, Naau M. Early failures in total knee arthroplasty. Clin Orthop Relat Res. 2001;392: Weir DJ, Moran CG, Piner IM. Kinematic conylar total knee arthroplasty. 14-year survivorship analysis of 208 consecutive cases. J Bone Joint Surg Br. 1996;78: Winsor RE, Insall JN, Urs WK, Miller DV, Brause BD. Two-stage reimplantation for the salvage of total knee arthroplasty complicate by infection. Further follow-up an refinement of inications. J Bone Joint Surg Am. 1990;72: Callaghan JJ, Salvati EA, Brause BD, Rimnac CM, Wright TM. Reimplantation for salvage of the infecte hip: rationale for the use of gentamicin-impregnate cement an beas. Hip. 1985: Walenkamp GH, Vree TB, van Rens TJ. Gentamicin-PMMA beas. Pharmacokinetic an nephrotoxicological stuy. Clin Orthop Relat Res. 1986;205: Antti-Poika I, Josefsson G, Konttinen Y, Ligren L, Santavirta S, Sanzén L. Hip arthroplasty infection. Current concepts. Acta Orthop Scan. 1990;61: Levitsky KA, Hozack WJ, Balerston RA, Rothman RH, Gluckman SJ, Maslack MM, Booth RE Jr. Evaluation of the painful prosthetic joint. Relative value of bone scan, seimentation rate, an joint aspiration. J Arthroplasty. 1991;6: Athanasou NA, Paney R, e Steiger R, Crook D, Smith PM. Diagnosis of infection by frozen section uring revision arthroplasty. J Bone Joint Surg Br. 1995;77: Felman DS, Lonner JH, Desai P, Zuckerman JD. The role of intraoperative frozen sections in revision total joint arthroplasty. J Bone Joint Surg Am. 1995;77: Bauer TW, Parvizi J, Kobayashi N, Krebs V. Diagnosis of periprosthetic infection. J Bone Joint Surg Am. 2006;88: Atkins BL, Athanasou N, Deeks JJ, Crook DW, Simpson H, Peto TE, McLary- Smith P, Berent AR. Prospective evaluation of criteria for microbiological iagnosis of prosthetic-joint infection at revision arthroplasty. The OSIRIS Collaborative Stuy Group. J Clin Microbiol. 1998;36: Barrack RL, Jennings RW, Wolfe MW, Bertot AJ. The value of preoperative aspiration before total knee revision. Clin Orthop Relat Res. 1997;345: Chimento GF, Finger S, Barrack RL. Gram stain etection of infection uring revision arthroplasty. J Bone Joint Surg Br. 1996;78: Della Valle CJ, Scher DM, Kim YH, Oxley CM, Desai P, Zuckerman JD, Di Cesare PE. The role of intraoperative Gram stain in revision total joint arthroplasty. J Arthroplasty. 1999;14: Ko PS, Ip D, Chow KP, Cheung F, Lee OB, Lam JJ. The role of intraoperative frozen section in ecision making in revision hip an knee arthroplasties in a local community hospital. J Arthroplasty. 2005;20: Kraemer WJ, Saplys R, Waell JP, Morton J. Bone scan, gallium scan, an hip aspiration in the iagnosis of infecte total hip arthroplasty. J Arthroplasty. 1993;8: Paney R, Drakoulakis E, Athanasou NA. An assessment of the histological criteria use to iagnose infection in hip revision arthroplasty tissues. J Clin Pathol. 1999;52: Spangehl MJ, Masri BA, O Connell JX, Duncan CP. Prospective analysis of preoperative an intraoperative investigations for the iagnosis of infection at the sites of two hunre an two revision total hip arthroplasties. J Bone Joint Surg Am. 1999;81: Leone JM, Hanssen AD. Management of infection at the site of a total knee arthroplasty. J Bone Joint Surg Am. 2005;87:

8 Trampuz A, Hanssen AD, Osmon DR, Manrekar J, Steckelberg JM, Patel R. Synovial flui leukocyte count an ifferential for the iagnosis of prosthetic knee infection. Am J Me. 2004;117: Spangehl MJ, Younger AS, Masri BA, Duncan CP. Diagnosis of infection following total hip arthroplasty. Instr Course Lect. 1998;47: Gram C. The ifferential staining of Schizomycetes in tissue sections an in rie preparations. Fortschritte er Meizin. 1884;2: Marker D, Seyler TM, Delanois RE, Plate JF, Mont MA. The sensitivity of Gram stains from multiple sites in patients with infecte total hip arthroplasty. Presente as a poster exhibit at the Annual Meeting of the American Acaemy of Orthopaeic Surgeons; 2007 Feb 14-18; San Diego, CA. Poster no Tunney MM, Ramage G, Patrick S, Nixon JR, Murphy PG, Gorman SP. Antimicrobial susceptibility of bacteria isolate from orthopeic implants following revision hip surgery. Antimicrob Agents Chemother. 1998;42: Moya TF, Thornhill T, Estok D. Evaluation an management of the infecte total hip an knee. Orthopeics. 2008;31:581-8; quiz Duff GP, Lachiewicz PF, Kelley SS. Aspiration of the knee joint before revision arthroplasty. Clin Orthop Relat Res. 1996;331:132-9.