Abstract
Peri-prosthetic infection is amongst the most common causes of failure following total knee replacement (TKR). In the presence of established infection, thorough joint debridement and removal of all components is necessary following which new components may be implanted. This can be performed in one or two stages; two-stage revision with placement of an interim antibiotic-loaded spacer is regarded by many to be the standard procedure for eradication of peri-prosthetic joint infection.
We present our experience of a consecutive series of 50 single-stage revision TKRs for established deep infection performed between 1979 and 2010. There were 33 women and 17 men with a mean age at revision of 66.8 years (42 to 84) and a mean follow-up of 10.5 years (2 to 24). The mean time between the primary TKR and the revision procedure was 2.05 years (1 to 8).
Only one patient required a further revision for recurrent infection, representing a success rate of 98%. Nine patients required further revision for aseptic loosening, according to microbiological testing of biopsies taken at the subsequent surgery. Three other patients developed a further septic episode but none required another revision.
These results suggest that a single-stage revision can produce comparable results to a two-stage revision. Single-stage revision offers a reduction in costs as well as less morbidity and inconvenience for patients.
Cite this article: Bone Joint J 2014;96-B:759–64.
Deep infection is a devastating complication of total knee replacement (TKR), imposing a heavy burden on the patient, the surgeon and the healthcare system.1 As the number of primary procedures steadily increases, it is probable that the absolute number of infections will also rise. The incidence of infection following TKR is reported to be 1% to 2%2,3 and the management of an infected TKR can cost three to four times that of a primary procedure.4 The forms of treatment include retention of the components with antibiotic suppression,5-7 revision in one or two stages,8and salvage procedures such as arthrodesis9,10 or amputation.11
The success rate of two-stage revision has been reported to be between 72% and 100% in eradicating infection and re-establishing a functioning replacement.12-17 It is the most widely performed technique for treating an infected TKR, and was first described by Insall, Thompson and Brause in 198318 in 11 patients. The infection was reported to have been eradicated in all the patients at short-term follow-up. Goldman, Scuderi and Insall19 reported a 91% success rate in the treatment of 64 infected TKRs, at a mean follow-up of 7.5 years. However, more recently Mortazavi et al17 reported a 28% failure rate following a two-stage revision in a series of 117 patients with a mean follow-up of 3.8 years. Although there are fewer reports of the results of single-stage revision for infected TKR, success rates of between 73% and 100% have been described.20-24
We report the results of single-stage revision TKR for infection performed in our department over a 30-year period.
Patients and Methods
Between 1979 and 2010, one surgical team in our hospital performed 50 consecutive single-stage revision TKRs for infection. The remaining surgeons preferred a two-stage technique. Patients with infected TKR were selected for a single-stage revision by our team if they had a bacteriologically proven infection, an identified organism, available culture and sensitivities and intact soft-tissue cover of the knee.
All patients were followed-up on an annual basis and data were collected prospectively (Table I).
Table I
Patient characteristics and data collected and analysed
| Patient characteristics |
|---|
| Age |
| Gender |
| Initial diagnosis |
| Previous surgery |
| Time between primary and revision |
| Bacteriology |
| Radiographic assessment |
| Functional assessment |
The outcome in 18 of the patients in this cohort has been previously reported at a mean follow-up of five years.22 These patients are included in the current report with their follow-up extended. Our current group consists of 33 women and 17 men, with a mean age at revision of 66.8 years (42 to 84). No bilateral procedures were undertaken. The mean time between the primary TKR and revision procedure was 2.05 years (1 to 8). For our 14 most recent patients Oxford Knee Scores (OKS)25, 26 are also available, where 40 is the best possible score and 0 the worst.
Radiological analysis was performed using the Knee Society roentgenographic evaluation and scoring system27 in order to assess the quality of fixation of the components. A score of < 4 is regarded as probably insignificant; if the score is between 5 and 9, the patient should be followed for progression; and a score of > 10 signifies possible or impending failure regardless of symptoms. A stable radiolucent line at the cement-bone interface of < 2 mm in width is frequent and has no clinical significance.28
Surgical management
We describe our protocol, which has remained unchanged from that originally outlined by Goksan and Freeman.22
Prior to definitive surgery, any infecting organisms are identified by needle aspiration and culture, or arthroscopic biopsy if the aspiration was unsuccessful. No antibiotics are commenced at this point. At operation, having approached the knee through the old incision, swabs and tissue samples are taken from the joint. All components and cement are removed. Any interface material is taken for both bacteriological and histological examination. The joint is then debrided, excising all tissue of doubtful viability and washed out with copious quantities of normal saline; it is then packed with povidone-iodine soaked swabs. The wound edges are approximated with sutures and a temporary compressive dressing applied. Appropriate antibiotics, according to the sensitivities established pre-operatively, are administered intravenously and the tourniquet is deflated for 30 minutes.
At this point, the entire operating team re-scrub and put on new gowns and gloves, the patient is re-draped and a new set of instruments is made available. The tourniquet is re-inflated, the knee is re-opened and the packs removed. The joint is copiously irrigated again with normal saline and culture swabs are once more taken from bone surfaces. The new components are introduced with antibiotic-impregnated cement, including supplementation with appropriate antibiotics, according to previous cultures and sensitivities. No cement is used around the stems, but the intramedullary canals are dusted with appropriate antibiotic powder. Two suction drains are placed into the joint and the wound is closed in a standard fashion with layers of interrupted sutures and, since the mid-1990s, with staples for the skin. Wool and crepe bandaging is applied. The patient is mobilised fully weight-bearing according to their ability.
The antibiotics, at first given intravenously, are chosen following microbiological advice according to the sensitivities obtained from pre-operative cultures. This regime is changed, if necessary, when the results of the intra-operative culture and sensitivities are available. Generally, at about two weeks post-operatively, the antibiotics are changed to an oral regime, which continues for three months.
Statistical analysis
The STATA statistical package for Windows version 12.0 (STATA Corp, College Station, Texas) was used for all analyses. A paired t-test was performed to compare the change in OKS between the pre- and post-operative phase. We fitted multivariable logistic regression models to investigate potential differences between the patients who had died and those who were still alive at the time of final analysis. Length of mean follow-up was calculated according to the method of Schemper & Smith.29 A Kaplan–Meier curve with a 95% confidence band (derived using a robust standard error) was computed to estimate the cumulative incidence of re-revision. P-values < 0.05 were considered statistically significant.
Results
The microbiological results at revision are shown in Fig. 1. In eight patients, infection was due to a single organism other than a methicillin sensitive staphylococcal species, and there was one case each of methicillin-resistant staphylococcus aureus (MRSA) and methicillin-resistant staphylococcus epidermidis (MRSE).
Fig. 1
Pie chart showing absolute and relative frequencies of micro-organisms confirmed at revision surgery (S. Epidermis - staphylococcus epidermidis; S. Aureus - staphylococcus aureus; MRSA - methicillin-resistant staphylococcus aureus; MRSE - methicillin-resistant staphylococcus epidermidis; pts - patients).
A total of 29 patients died during the period of the study without evidence of recurrent infection and from conditions unrelated to their TKR. The mean follow-up in this subset of patients was 10.3 years (2 to 22). The mean follow-up for the whole group was 10.5 years (2 to 24), at which time (last follow-up) 46 patients (92%) were asymptomatic with well-functioning knees and no evidence of infection. Four patients (8%) had a further septic episode following the single-stage revision and these are described later in the paper.
Multivariable logistic regression revealed that an underlying diagnosis of rheumatoid arthritis was associated with a greater than twenty times higher odds of death than in patients with osteoarthritis (Odds ratio (OR) = 23.75, 95% confidence interval (CI): 1.15 to 488.56, p = 0.04). Additionally, patients who were older at the time of revision had higher odds of death (OR for one year increase in age from the mean age of 66.8 years = 1.20, 95% CI 1.06 to 1.37, p = 0.004). The other variables, including previous orthopaedic procedures, time between primary and revision surgery, gender and mixed bacterial growth were not associated with death (Table II).
Table II
Association of patient variables with death (CI confidence interval)
| Variable | Odds ratio | 95% CI |
|---|---|---|
| Gender | 1.35 | 0.14 to 12.85 |
| Time between primary and revision | 0.61 | 0.20 to 1.82 |
| Previous procedures | 4.46 | 0.53 to 37.75 |
| Mixed bacterial growth | 1.67 | 0.18 to 15.56 |
A total of ten patients (20%) required a further procedure for pain and loosening of their revision implant. The further revisions were undertaken at a mean 9.4 years (1 to 21) post-operatively. Figure 2 shows the cumulative incidence of further revision for any reason during follow-up. The cumulative five- and ten-year probability of further revision for any reason was 4.1% (95% CI 1.1 to 15.5) and 9.6% (95% CI 3.7 to 23.9) respectively.
Fig. 2
Graph showing the cumulative incidence of re-revision for any reason; vertical bars on the curve indicate censoring, shaded area represents 95% confidence band.
Biopsies were taken in all patients for microbiological examination at further revision. In one patient, the same micro-organisms were grown as had been identified in the initial procedure, which was a mixed growth of Staphylococcus epidermidis and Streptococcus viridans. This patient had a history of an infected high tibial osteotomy prior to the TKR, which had not been disclosed to the surgeon. This was considered to be the only true re-infection within our cohort. She ultimately underwent an arthrodesis.
No other knees which required further revision had any micro-organisms isolated and were considered aseptically loose; these patients all made an uncomplicated recovery. Thus, the rate of success with re- revision for confirmed further infection as the endpoint is 98% (49 of 50 patients).
A paired t-test showed that the mean OKS in the 14 most recent patients (Fig. 3) increased by a factor of 2.4 from 14.5 (6 to 25) pre-operatively to 34.5 (26 to 38) one year after surgery. This represents a mean absolute improvement of 20.0 points (95% CI: 17.8 to 22.2, p < 0.001).
Fig. 3
Pre- and post-operative Oxford Knee Score: transverse line, median; box, interquartile range (Q1-Q3); end of the whiskers, most extreme value that lies within Q3 or Q1 +/- 1.5 times the IQR.
We were able to review the most recent radiographs of eight of the remaining 11 living patients who had not required any further surgery to their TKR. These radiographs were taken at a mean follow-up of 15.25 years (8 to 24). In all patients the implants were well fixed, with no radiolucent lines or signs of progressive lysis (Fig. 4).
Figs. 4a - 4f
Radiographs of a 73-year-old man who underwent a single-stage revision for an infected TKR two years following implantation of the primary prosthesis. Pre-operative a) anteroposterior and b) lateral radiographs showing areaanteroposteriors of lysis. Immediate post-revision c) anteroposterior and d) lateral views; and e) anteroposterior and f) lateral views at follow-up of 11 years.
A total of four patients (8%) had a further infection, of which one has already been described above as a confirmed re-infection that required further revision; there were thus three other patients with a subsequent infection that did not require further revision. These are regarded as ‘new’ infective episodes rather than re-infection. The first was in a 51-year- old woman with rheumatoid arthritis, who had recurrent infections in her contralateral knee, which had been arthrodesed following an infected primary TKR. At seven months following her revision for infection, she developed an acute infection within the revision TKR, which was managed successfully with drainage and antibiotics for three months. She was asymptomatic 22 years following the revision surgery.
The second was a 62-year-old woman with rheumatoid arthritis, who developed an infection six years following the revision. This occurred in conjunction with an infected varicose ulcer on the ipsilateral ankle. The infection resolved with antibiotics and the TKR remained asymptomatic ten years post-operatively.
The final patient was a morbidly obese 43-year-old woman who developed an extra-articular ipsilateral abscess ten months following revision surgery. The abscess was explored and found to be within the supra-patellar fat layer and not communicating with the TKR. It was drained and she recovered well. The synovium of the knee was also biopsied and did not grow any organisms. She was asymptomatic four years post-operatively.
Discussion
Our series shows a 98% success rate judged by one further revision for recurrent infection, in a cohort of 50 patients with a mean follow-up of 10.5 years. Single-stage revision in infected TKR, although technically demanding, can produce excellent results when specific criteria and meticulous surgical technique are followed. There remains an ongoing debate between those who advocate single-stage revision20-24 and those who favour two-stage procedure.12-18
Two-stage revision involves an initial stage with removal of the components and thorough debridement, with the introduction of an antibiotic loaded spacer. This is followed by a second stage procedure with systemic antibiotic treatment during the interval. The disadvantages of this approach include impaired mobility, joint stiffness and pain during the interval between stages. It also involves two major procedures with the associated morbidity, discomfort, cost and prolonged stay in hospital.20
There is much less information in the literature regarding the use of single-stage revision TKR (Table III) but success rates are comparable with those published for two-stage revision.12-17 A successful single-stage revision confers a shorter recovery period, less limitation of the patient’s activities, and avoids the potential morbidity from a second procedure as well as allowing financial savings for the healthcare system.
Table III
Single stage revision published studies
| Author | No. of patients | Follow-up (years) | Success rate |
|---|---|---|---|
| Von Foerster et al21 | 104 | 5 to 15 | 73% |
| Buechel23 | 22 | Mean 10.2 (1.4 to 19.6) | 90.9% |
| Parkinson et al20 | 12 | Mean 2 | 100% |
| Singer et al24 | 63 | Mean 3 (2 to 5.8) | 95% |
Initial reports on single-stage revision of infected TKRs from the Endoklinik, Hamburg, Germany, described a success rate in eradication of any infection of only 73%21 while Buechel23 reported a 90.9% success rate in 22 patients at a mean follow-up of 10.2 years. More recently, Parkinson et al20 reported a success rate of 100% at a mean follow-up of two years in 12 patients. Singer et al24 reported a success rate of 95% at a mean follow-up of three years in 63 patients, although MRSA and MRSE infections were excluded.
In our series, we have identified only one true recurrent infection confirmed with positive microbiological results identified at the further revision procedure, representing a 98% success rate. If we were to include the three new infections, none of which resulted in further revision, the rate of success remains at 92%, at a mean follow-up of 10.5 years, with all bacterial species included.
We believe that extensive debridement, identification of the infecting organisms and suitable antibiotic treatment, as directed in discussion with a microbiologist, are the key factors in the successful outcome of this procedure. Single-stage revision for the management of an infected TKR should gain further acceptance and we agree with a recent review that advocates its use in selected cases.8
1 Lavernia C , LeeDJ, HernandezVH. The increasing financial burden of knee revision surgery in the United States. Clin Orthop Relat Res2006;446:221–226. Google Scholar
2 Bengston S , KnutsonK, LidgrenL. Treatment of infected total knee arthroplasty. Clin Orthop Relat Res1989;245:173–178. Google Scholar
3 Blom AW , BrownJ, TaylorAH, et al.Infection after total knee arthroplasty. J Bone Joint Surg [Br]2004;86-B:688–691. Google Scholar
4 Herbert CK , WilliamsRE, LevyRS, BarrackRL. Cost of treating an infected total knee replacement. Clin Orthop Relat Res1994;331:140–145. Google Scholar
5 Garvin KL , CorderoGX. Infected total knee arthroplasty: diagnosis and treatment. Instr Course Lect2008;57:305–315. Google Scholar
6 Trebse R , PisotV, TrampuzA. Treatment of infected retained implants. J Bone Joint Surg [Br]2005;87-B:249–256. Google Scholar
7 Nickinson RS , BoardTN, GambhirAK, PorterML, KayPR. The microbiology of the infected knee arthroplasty. Int Orthop2010;34:505–510. Google Scholar
8 Vanhegan IS , Morgan-JonesR, BarrettDS, HaddadFS. Developing a strategy to treat established infection in total knee replacement: a review of the latest evidence and clinical practice. J Bone Joint Surg [Br]2012;94-B:875–881. Google Scholar
9 Yeoh D , GoddardR, MacnamaraP, et al.A comparison of two techniques for knee arthrodesis: the custom made intramedullary Mayday nail versus a monoaxial external fixator. Knee2008;15:263–267. Google Scholar
10 Bargiotas K , WohlrabD, SeweckeJJ, et al.Arthrodesis of the knee with a long intramedullary nail following the failure of a total knee arthoplasty as the result of infection. J Bone Joint Surg [Am]2006;88-A:553–558. Google Scholar
11 Sierra RJ , TrousdaleRT, PagnanoMW. Above-the-knee amputation after a total knee replacement: prevalence, etiology, and functional outcome. J Bone Joint Surg [Am]2003;85-A:1000–1004. Google Scholar
12 Haddad FS , MasriBA, CampbellD, et al.The PROSTALAC functional spacer in two-stage revision for infected knee replacements: prosthesis of antibiotic-loaded acrylic cement. J Bone Joint Surg [Br]2000;82-B:807–812. Google Scholar
13 Laffer RR , GraberP, OchsnerPE, ZimmerliW. Outcome of prosthetic knee-associated infection: evaluation of 40 consecutive episodes at a single centre. Clin Microbiol Infect2006;12:433–439. Google Scholar
14 Leone JM , HanssenAD. Management of infection at the site of a total knee arthroplasty. J Bone Joint Surg [Am]2005;87-A:2335–2348. Google Scholar
15 Pitto RP , CastelliCC, FerrariR, MunroJ. Pre-formed articulating knee spacer in two-stage revision for the infected total knee arthroplasty. Int Orthop2005;29:305–308. Google Scholar
16 Zimmerli W , TrampuzA, OchsnerPE. Prosthetic-joint infections. N Engl J Med2004;351:1645–1654. Google Scholar
17 Mortazavi SM , VegariD, HoA, ZmistowskiB, ParviziJ. Two-stage exchange arthroplasty for infected total knee arthroplasty: predictors of failure. Clin Orthop Relat Res2011;469:3049–3054. Google Scholar
18 Insall JN , ThompsonFM, BrauseBD. Two-stage reimplantation for the salvage of infected total knee arthroplasty. J Bone Joint Surg [Am]1983;65-A:1087–1098. Google Scholar
19 Goldman RT , ScuderiGR, InsallJN. 2-stage reimplantation for infected total knee replacement. Clin Orthop Relat Res1996;331:118–124. Google Scholar
20 Parkinson RW , KayPR, RawalA. A case for one-stage revision in infected total knee arthroplasty?Knee2011;18:1–4. Google Scholar
21 von Foerster G , KlüberD, KäblerU. Mid- article long-term results after treatment of 118 cases of periprosthetic infections after knee joint replacement using one-stage exchange surgery. Orthopade1991;20:244–252 (article in German). Google Scholar
22 Goksan SB , FreemanMA. One-stage reimplantation for infected total knee arthroplasty. J Bone Joint Surg [Br]1992;74-B:78–82. Google Scholar
23 Buechel FF . The infected total knee arthroplasty: just when you thought it was over. J Arthroplasty2004;19(Suppl):51–55. Google Scholar
24 Singer J , MerzA, FrommeltL, FinkB. High rate of infection control with one-stage revision of septic knee prostheses excluding MRSA and MRSE. Clin Orthop Relat Res2012;470:1461–1471. Google Scholar
25 Dawson J , FitzpatrickR, MurrayD, CarrA. Questionnaire on the perceptions of patients about total knee replacement. J Bone Joint Surg [Br]1998;80-B:63–69. Google Scholar
26 Murray DW , FitzpatrickR, RogersK, et al.The use of the Oxford hip and knee scores,. J Bone Joint Surg (Br)2007:89(8)1010–1014. Google Scholar
27 Ewald FC . The Knee Society total knee arthroplasty roentgenographic evaluation and scoring system. Clin Orthop Relat Res1989;248:9–12. Google Scholar
28 Schneider R , HoodRW, RanawatCS. Radiographic evaluation of knee arthroplasty. Orthop Clin North (Am)1982;13:225–244. Google Scholar
29 Schemper M , SmithTL. A note on quantifying follow-up in studies of failure time. Control Clin Trials . 1996;17:343–346. Google Scholar
No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article.
This article was primary edited by D. Rowley and first proof edited by G. Scott.
