Use of antibiotic spacers for knee endoprosthesis infections treatment

Abstracts

OBJETIVO: O objetivo do estudo é avaliar a utilização dos espaçadores de cimento acrílico com antibiótico no tratamento das infecções em endopróteses não convencionais de joelho. MÉTODO: Desde de 2004 foram tratados sete pacientes (seis pacientes operados no nosso serviço e um paciente que havia sido submetido a cirurgia primária do tumor em outro serviço) com infecção peri-endoprótese não convencional de joelho. Todos pacientes foram submetidos a retirada da endoprótese e reconstrução com espaçador com cimento acrílico com antibiótico. Todos os pacientes foram monitorados clínica e laboratorialmente quanto ao controle da evolução, sendo considerados aptos para a revisão e recolocação de endoprótese após 06 (seis) meses sem sinais infecciosos RESULTADOS: Notamos um discreto predomínio do do processo infeccioso nas próteses realizadas na tíbia proximal em comparação com o fêmur distal (57,1% x 42,9%). O seguimento médio dos pacientes foi 68,2 meses. Durante o seguimento, um paciente faleceu devido a doença de base. Dos sete pacientes, 6 foram considerados curados e um persistiu com sinais e sintomas de infecção. CONCLUSÃO: Os resultados obtidos até o momento tem motivado a continuidade deste método de tratamento.

Prótese; Neoplasias; Infecção; Cimentos para ossos; Osteossarcoma

ORIGINAL ARTICLE

Use of antibiotic spacers for knee endoprosthesis infections treatment

Dan Carai Maia Viola; Nivaldo Souza Cardozo Filho; Rodrigo Tristão Nunes; Fabíola Andrea Carvalho Godoy; Marcelo de Toledo Petrilli; Marcos Korukian; Reynaldo Jesus-Garcia

Department of Orthopaedics and Traumatology, Federal University of São Paulo – Paulista Medical School - UNIFESP/EPM- São Paulo, SP, Brazil

Correspondences to

ABSTRACT

OBJCTIVE: The aim of this study is to evaluate the use of cement spacers impregnated with antibiotics for the treatment of infections in the nonconventional endoprostheses of the knee.

METHODOLOGY: We have treated seven patients since 2004 (of which six were submitted to surgery in our service and one patient had been submitted to a primary tumor surgery in another removal service) with deep infection in knee tumor prosthesis. All patients were submitted to endoprosthesis removal and reconstructed with antibiotic cement spacer. All patients were monitored both clinically and by lab tests as for monitoring the evolution, being considered able for reviews after 6 (six) months without infections signs.

RESULTS: We have noted a small predominance of infectious processes on the prosthesis inserted on proximal tibia as compared with distal femur (57.1% x 42.9%). The mean follow-up time of patients was 68.2 months. During the follow up, one patient died as a result of the root disease. Six patients out of seven were regarded as cured and one persisted with infection signs and symptoms.

CONCLUSION: The results obtained up to date have motivated us to continue using this method of treatment.

Keywords: Prosthesis. Neoplasms. Infection. Bone cements. Osteosarcoma.

INTRODUCTION

Nonconventional knee endoprostheses are largely employed in reconstructions after bone tumor resections around the knee. This kind of implant presents the same complication rates as the conventional knee prostheses.

The name 'nonconventional endoprosthesis' is due to the fact that the osteotomy for its placement is more extensive than in the cases of conventional prostheses. In conventional knee prostheses, osteotomy reaches the distal metaphyses of the femur and proximal metaphyses of the tibia, being usually performed with the aid of metaphyseal templates. On the other hand, in nonconventional endoprostheses, there is no rule of thumb for the osteotomy location. Bone resection varies according to the size of the tumor and the oncologic margin of the lesion.

The bone affected by tumor, in the large majority of times, is submitted to a diaphyseal osteotomy. This kind of diaphyseal resection gives rise to a large bone failure, requiring the use of an implant enabling bone length and its perfect joint alignment reestablishment. Nonconventional endoprostheses were introduced, thus, as a method for reconstructing bone failures resulting from the resection of musculoskeletal tumors.^1-4

The major acute complications of this kind of surgery are infection, dehiscence, collections formation, such as hematoma or seroma and deep venous thrombosis.^3,5 At late stages, this kind of procedure may present other complications, such as aseptic loosening of the implant, a broken component of the endoprosthesis, peri-prosthetic infection, and septic loosening of the implant. In addition to those, local recurrence of tumor may occur, with or without implant loosening.^4,6

Peri-prosthetic infections are usually symptomatic. The classical signs of local heat, redness, and pain may be present. Fistulization of collections is a common event in these cases, with serous, blood red cells or purulent leakage. In some cases of late infection we didn't find at first any classical inflammatory signs. The patient may evolve with pain at prosthesis site and, at the X-ray evaluation, signs of components loosening can be visualized or not.⁷ These cases constitute a greater challenge for diagnosis.

The key laboratory changes are the leukogram modification (with an infectious pattern), increased hemosedimentation speed and increased serum reactive C protein. These tests, in addition to support the diagnosis, serve as parameters of the therapeutic response during patient's follow-up. On X-ray tests, we can find endoprosthesis loosening (lysis around the cement) and deep peri-prosthetic collections.

The classical treatment for peri-prosthetic infections relies on therapy with antibiotics driven by results of cultures and antibiograms and by the removal of the infected endoprosthesis. In cases of infection around nonconventional endoprostheses, the bone failure found after the implant removal requires a reconstruction method.

Many authors recommended amputation as a definitive method for fixing infections around knee endoprostheses.⁸ Reconstruction techniques are based on joint arthrodesis after bone transportation with the method by Ilizarov. Another technique fosters transportation with the use of linear fixators and subsequent arthrodesis. Arthrodesis in situ may also be provided, causing a dramatic shortening the limb. In this case, there is no bone transportation, and the limb becomes quite shorter.

Following the same treatment approach as in peri-prosthetic infections for knee total arthroplasty, the use of acrylic cement spacers with antibiotics was suggested. In this treatment approach, the spacer provides, in addition to its mechanical function, a therapeutic antibacterial function, with the local release of antibiotic agents.

Since the bone failure found after UCE removal is quite large, using a cement spacer only is not enough for providing an appropriate mechanical stabilization. With the purpose of improving mechanical stabilization, the technique using a cement spacer with a stiff nail inside it was developed. Thus, the cement with antibiotics makes contact with bone and muscular bed, and the nail at the assembly core allows a better functional stabilization.

Once the spacer system is assembled, the patient is clinically treated (for a variable period of time) with appropriate antibiotics (based on culture and antibiogram results of materials collected at the peri-prosthetic region), and, after total remission of signs and symptoms, and upon the stabilization of laboratory parameters, the spacer can be removed and replaced by a new NCE.^9,10

The objective of this study is to present a method for treating deep infections in patients submitted to tumoral resection and reconstruction with nonconventional knee endoprosthesis.

MATERIALS AND METHODS

Between January 1st, 1990 and December 31st, 2006, 283 surgeries for inserting an nonconventional knee endoprosthesis after tumoral resection (patients with tumors on distal femur and proximal tibia) have been carried out in our service.

During follow-up, 24 (twenty-four patients – 8.48%) patients evolved with deep infection of the endoprosthesis site. Up to 2004, the patients with deep peri-prosthetic infection were treated by removing the endoprosthesis and with external fixation (for elongation or arthrodesis) or with amputation of the limb. From 2004 on, we started to operate the patients by removing the endoprosthesis followed by reconstruction with acrylic cement spacer with antibiotics. The case series of this study, thus, consists of 6 (six) patients with peri-prosthetic infection operated from 2004 on, added by a patient submitted to tumoral resection surgery and reconstruction with nonconventional endoprosthesis whose surgery was performed in another hospital, being subsequently followed up in our outpatient service. When this patient was referred to our outpatient facility, he already presented with deep infection, making use of oral antibiotics for 3 months, and he had already been submitted to surgical debridement on the endoprosthesis site.

The studied patients, as approved by UNIFESP/EPM's Committee of Ethics, consisted of 2 males (28,6%) and 5 females (71.4%). The mean age when the infection was diagnosed was 35.9 years (range: 11-68 years). Concerning the baseline disease, 6 (six) patients had osteosarcoma (3 metastatic and 3 non-metastatic), and 1 patient presented giant cells tumor (GCT) as a baseline pathology. The six patients affected by osteosarcoma were submitted to a neo-adjuvant and adjuvant chemotherapy approach (according to the protocol in force at the moment of diagnosis). The patient with GCT had a tumor on distal femur with extra-bony lesion, showing resection with wide margins and reconstruction with nonconventional knee prosthesis. This patient had 2 episodes of local recurrence, and he was submitted to resection of the recurrences around the prosthesis. No patient received radiotherapy as adjuvant treatment.

All patients received therapy with antibiotics (oral and endovenous) when diagnosed with infection. During follow-up, all patients were submitted to at least one surgical debridement surgery but not removing prosthesis components. All patients were followed up through leukograms, hemosedimentation speed and reactive C protein tests. All patients received therapy with specific antibiotics after culture tests and antibiograms collected intraoperatively at debridement procedure.

All patients with persistent infection after one surgical debridement event and receiving antibiotic therapy for at least 6 weeks (specific antibiotic therapy) were regarded as eligible for endoprosthesis removal and limb reconstruction with modeled spacers with nail and cement with antibiotics. In our case series, all patients with infection after 2004 were regarded as eligible to treatment.

The seven patients were submitted to a new surgical procedure through the same port used for tumor resection. In cases where fistulas were present, fistulectomy was carried out with resection of the whole scar (ellipse resection). By performing the endoprosthesis approach, new culture and antibiogram samples were collected in order to guide the postoperative antibiotic therapy approach. The endoprostheses were removed (all the components) and all the cement previously used was removed. We performed an extensive debridement of the necrotic tissues around the endoprosthesis. At the bone approach, a careful curettage of the femoral and tibial medullar was performed. After the debridement, the area was cleaned with saline solution 0.9% (at least 15 liters were used in each procedure).

For reconstruction, femoral nails were used wrapped with cement with antibiotics. In all patients, the Simplex^® brand acrylic cement with antibiotics (Tobramycin) was used.

In order to determine infection control, we employed clinical and laboratory parameters. Clinically, we saw total healing of the surgical wound, absence of fistulas and leakages, and absence of flogistic signs. Concerning laboratory analysis, we saw the reduction of Hemosedimentation speed, reactive C protein, and leukogram levels to normal parameters. We regarded a patient as able to endoprosthesis replacement when the clinical and laboratory parameters remained at normal levels for 6 (six) months (Figures 1 and 2).

RESULTS

We found a subtle prevalence of the incidence of infectious process on prosthesis implanted in patients with proximal tibial tumors when compared to the baseline location of the tumor on the distal femur (57.1% of the cases showed a baseline tumor on proximal tibia, while 42.9% on the distal femur).

Concerning the endoprosthesis model employed, in 6 patients an nonconventional custom-made prosthesis was used, and, in one case, this was an extensible-type endoprosthesis. This was the only patient in the case series presenting with a partial knee prosthesis (only the femoral component). The size of the endoprosthesis body was, in average, 142.9 mm.

Concerning the anatomical-pathological report, all patients were submitted to resections with wide margins and presented free margins.

The patients presented the first signs and symptoms corresponding to infectious processes within a period that ranged from 3 to 127 months after the endoprosthesis insertion surgery (mean: 32.1 months, and median: 24.8 months (Figure 3)). The signs and symptoms found in our study were the following: pain (n=6, 85.7% of the patients), active fistula (n=5, 71.4% of the patients), local heat (n=3, 42.9% of the patients), redness (n=3, 42.9% of the patients), palpable liquid collection (n=2, 28.6% of the patients), skin lesion and prosthesis exposure (n=2, 28.6% of the patients) and diffuse edema on the limb (n=1, 14.3% of the patients). (Figure 4)

All patients were submitted to at least one surgical procedure for debridement and surgical cleaning previously to the indication of endoprosthesis removal. One patient (the one diagnosed with giant cell tumor) was submitted to two surgical cleaning procedures previously to the indication of endoprosthesis removal.

The time elapsed between clinical diagnosis of infection and endoprosthesis removal ranged from 2.3 months and 88.8 months (mean: 35.2 months, and median: 12.1 months). The time elapsed between the 1st surgery and the endoprosthesis removal ranged from 13.3 months to 129.3 months (mean: 67; median: 69.6 months).

Acrylic cement spacers with antibiotics may be placed for blocking knee joint (single nail) or with two nails (assembling 2 separate spacers, one for the femur and another for the tibia), allowing a pseudoarticulation between spacers (double nail). In the case series presented here, 2 (two) (28.6%) double-nail spacers, and 5 (five) (71.4%) single-nail spacers were inserted. All spacers were assembled with Simplex^® cement with Tobramycin.

For the screening of etiologic agents, cultures were collected intraoperativelly from the endoprostheses beds. The identified etiologic agents were the following: Staphylococcus aureus in 4 cases (57.1%), Acinetobacter in 1 case (14.3%), Streptococcus sp. in 1 case (14.3%), and Enterobacter in 1 case (14.3%). All patients received postoperative therapy with specific antibiotics according to the culture and the antibiogram. Thus, 4 patients (57.1%) were treated with ciprofloxacin, 2 patients (28.6%) were treated with vancomycin, and 1 patient (14.3%) was treated with sultamycin. Antibiotic therapy time ranged from 2 to 33 months (mean: 5.5 months) after endoprosthesis removal.

The most prevalent objective clinical sign is the presence of active fistula. Additionally to the fistula, classical inflammatory signs were assessed (heat, redness, edema and pain), as well as laboratory parameters. (Figure 5). Of the 7 patients assessed, 6 (85.7% of the cases) evolved with clinical and laboratory parameters improvement, while one patient didn't present improvement and evolved with persistent infectious process.

For the 6 patients responding well to the recommended treatment, the mean time for fistulas closing was 4 weeks (3-24 weeks). Inflammatory signs have also improved, returning to normal levels within 4 weeks (3-24 weeks), while the mean time for HSS and reactive C protein normalization was 20 weeks (12-100 weeks). (Figure 5)

The mean follow-up time is 68.2 months (13.9 to 120.9 months). One patient was submitted to spacer removal and endoprosthesis replacement. Currently, this patient shows no signs of infection on the new endoprosthesis. Concerning the other 6 patients, one is waiting for endoprosthesis replacement, one died as a result of the baseline disease, and the remaining 4 did not accept to be submitted to a new surgical procedure at risk of perpetuating the infectious process. These patients remained with the spacers, and no major complication was found so far.

Of the 7 (seven) patients submitted to spacer insertion, 6 (six) evolved with shortening of the affected limb, which ranged from 1.5 to 8 cm (mean: 3.2 cm).

Taking only the patients who survived throughout follow-up into account, we have a cure rate of 83.3% with the described method.

Due to the small sample, we couldn't design an appropriate multifactorial analysis to determine which variables present a stronger impact for peri-endoprosthetic infection treatment.

DISCUSSION

Infection is one of the most feared complications occurring in orthopaedic surgeries. Tumor resection surgery and the reconstruction with NCE present a risk of infection ranging from 5% to 35%. Such high incidence occurs mainly because of oncologic patients' characteristics, which, many times, have diseases leading to immunosupression. Furthermore, chemotherapy and radiotherapy play an immunosuppressant role (both local and systemic), increasing the risks of infection.

On the other hand, tumor resection surgeries requiring reconstruction with NCE are major procedures, showing a lengthy surgical time and leading to a higher risk of infections.

Since the 1970's, many authors have tried to determine which would be the best method for peri-prosthetic infection treatment. The first treatment method described for treating infections in knee total arthroplasties was the plain surgical cleaning, recommended to any type of infection around implant materials.¹¹ Brodersen et al.¹² were the first to describe, in the 1970's, knee arthrodesis as a method for treating arthroplasty failures. In their series of 45 patients, 40 presented infection as the root cause of implant failure, being submitted to knee arthrodesis.

Phillips et al.¹³ have also proposed knee arthrodesis after an implant removal, but using the Hoffman's external fixator. With this technique, they were able to achieve a faster union and less shortening compared to internal techniques.

Insall¹⁴ was the first author to write a review article about the peri-prosthetic infections theme. The author considered surgical cleaning as imperative, as well as the implant removal and endovenous antibiotic therapy. Insall was also the first to suggest the review in two steps. After the surgery, the patient would be on endovenous antibiotic therapy for six weeks, and only after that period a new surgery for inserting the new prosthesis would be indicated.

Poss et al.¹⁵, in the 1980's, assessed 4240 patients in an attempt to determine which would be the most important groups at risk of infection around the knee. In their text, the authors discuss that the pathologies leading to immunosuppression are the key risk factor for infection around knee total arthroplasty, followed by patients with previous review procedures. Bliss and McBride¹⁶ emphasized, at the time, the importance of the full prosthesis coverage with myocutaneous flaps.

In the 1980's, Borden and Gearen¹⁷ suggested a treatment protocol for knee total peri-prosthesis infection. The methods described by the authors consisted of primary review of the prosthesis, aggressive and isolated surgical cleaning, and of the review in two steps.

Wilde and Ruth¹⁸ pioneered the description of a review method in two steps with the use of an acrylic cement spacer with antibiotics. In that initial study, the authors removed the prosthesis and kept the patient under endovenous antibiotic therapy for 4.2 weeks, and removing the spacer and replacing arthroplasty. Morrey et al.¹⁹, in 1989, recommended that, in acute infections, aggressive debridement and surgical cleaning should be made, which could cure the infection in up to 80% of the cases.

Wirganowicz et al.³ assessed the causes of nonconventional endoprosthesis failures. The authors assessed 278 prostheses, finding an infection rate of 13%. More recently, Gosheger et al.²⁰ assessed the effects of a silver layer on the endoprosthesis as a way to reduce the rate of infections. In an experimental study, the authors concluded that silver could reduce the risk of peri-prosthetic infection from 47% (control group) to 7%.

Since literature does not describe which would be the best method for treating deep infections around nonconventional knee endoprostheses, we found support on treatment techniques for infections around knee total arthroplasties. Our primary concern was maintaining a conservative surgery on the limb. The indication of surgical cleaning previously to the endoprosthesis removal is described, which can lead to resolution of the infectious process without requiring a more aggressive procedure.

CONCLUSION

Based on our findings, we couldn't determine the key positive predictive factors for a better evolution in cases of infection around nonconventional endoprostheses. However, we conclude that the method employed shows good results with great chances of curing the infectious process.

We also conclude that the major complication resulting from the method is dysmetria of the affected limb.

REFERENCES

• 1. Marcove RC, Khafagy MM.Total femur and knee replacement using a metallic prosthesis. Clin Bull. 1974;4:69-71.
• 2. Harrington KD, Johnston JO, Kaufer HN, Luck JV Jr, Moore TM, Limb salvage and prosthetic joint reconstruction for low-grade and selected high-grade sarcomas of bone after wide resection and replacement by autoclaved [corrected] autogeneic grafts. Clin Orthop Relat Res. 1986;(211):180-214. Erratum in: Clin Orthop Relat Res. 1987;(216):312.
• 3. Wirganowicz PZ, Eckardt JJ, Dorey FJ, Eilber FR, Kabo JM. Etiology and results of tumor endoprosthesis revision surgery in 64 patients. Clin Orthop Relat Res. 1999;(358):64-74.
• 4. Fabroni RH, Castagno A, Aguilera AL, Steverlynck AM, Zeballos J. Long-term results of limb salvage with the Fabroni custom made endoprosthesis. Clin Orthop Relat Res. 1999;(358):41-52.
• 5. Kawai A, Muschler GP, Lane JM, Otis JC, Healey JH. Prosthetic knee replacement after resection of a malignant tumor of the distal part of the femur: Medium to long-term results. J Bone Joint Surg Am. 1998;80: 636-47.
• 6. Rech L, Schittic H, Plotz W, Trager J, Gradinger R, Burgkart R, Reeg S Custom made total knee replacement in pacients with primary or secondary bone tumor. Acta Chir Orthop Traumatol Cech. 1964 ;61:92-6.
• 7. Jeys LM, Grimer RJ, Carter SR, Tillman RM. Periprosthetic infection in pacients treated for an orthopedic oncological condition . J Bone Joint Surg Am. 2005;87:842-9.
• 8. vonStein T, Gollwitzer H, Kruis C, Buhren V. Arthrodesis after total knee replacement considering septic loosening as example. Orthopade. 2006;35:946, 948-52, 954-5.
• 9. Flin MN, Griffin Am, Bell RS, Ferguson PL. Two stage revision of infected uncemented lower extremity tumor endoprostheses J Arthroplasty. 2007;22:859-65.
• 10. Lombardi AV Jr, Berend KR, Adams JB, Kames JM. Articulating antibiotic spacers :the standard of care for an infected total knee arthroplasty. Orthopedics. 2007;30:782, 786-7.
• 11. Petty W, Bryan RS, Coventry MB, Peterson LF. Infection after total knee arthroplasty. Orthop Clin North Am. 1975;6:1005-14.
• 12. Brodersen MP, Fitzgerald RH Jr, Peterson LF, Coventry MB, Bryan RS. Arthrodesis of the knee following failed total knee arthroplasty. J Bone Joint Surg Am. 1979;61:181-5.
• 13. Phillips HT, Mears DC. Knee fusion with external skeletal fixation after an infected hinge prosthesis: a case report. Clin Orthop Relat Res. 1980;(151):147-52.
• 14. Insall JN, Thompson FM, Brause BD. Two-stage reimplantation for the salvage of infected total knee arthroplasty. J Bone Joint Surg Am. 1983;65:1087-98.
• 15. Poss R, Thornhill TS, Ewald FC, Thomas WH, Batte NJ, Sledge CB. Factors influencing the incidence and outcome of infection following total joint arthroplasty. Clin Orthop Relat Res. 1984;(182):117-26.
• 16. Bliss DG, McBride GG. Infected total knee arthroplasties. Clin Orthop Relat Res. 1985;(199):207-14.
• 17. Borden LS, Gearen PF. Infected total knee arthroplasty. A protocol for management. J Arthroplasty. 1987;2:27-36.
• 18. Wilde AH, Ruth JT. Two-stage reimplantation in infected total knee arthroplasty. Clin Orthop Relat Res. 1988;(236):23-35.
• 19. Morrey BF, Westholm F, Schoifet S, Rand JA, Bryan RS. Long-term results of various treatment options for infected total knee arthroplasty. Clin Orthop Relat Res. 1989;(248):120-8.
• 20. Gosheger G, Hardes J, Ahrens H, Streitburger A, Buerger H, Erren M, Gunsel A, Kemper FH, Winkelmann W, Von Eiff C. Silver-coated megaendoprostheses in a rabbit model--an analysis of the infection rate and toxicological side effects. Biomaterials. 2004;25:5547-56.

Publication Dates

• Publication in this collection
  20 July 2009
• Date of issue
  2009

History

• Accepted
  13 Oct 2008
• Received
  06 May 2009