Introduction

Anterior cruciate ligament reconstruction (ACLR) is a widely performed surgery aiming to restore or improve the knee joint function. Postoperative infection after ACLR is a rare complication, with its frequency estimated at about 0,5 % (0.14%-1.7%), yet with potentially devastating consequences for the knee and the extremity [1]. This procedure is an intra- and extra-articular operation, therefore infection can involve the native joint cavity, graft canals and implants fixing the graft in the femur and tibia, which can lead to the synovitis and local osteomyelitis of the surrounding bone. The most common causative pathogens are Staphylococcus aureus, but other pathogens, like coagulase-negative Staphylococcus, Cutibacterium acnes, Peptostreptococcus, Enterobacter, Enterococcus, Pseudomonas aeruginosa, Escherichia coli, Klebsiella and Methicillin-resistant Staphylococcus aureus (MRSA), can be cultured. Treatment attitudes differ regarding the advancement of the joint infection according to Gächter classification [2]. Published options include for Gächter grades 1 and 2 infections: arthroscopic lavage and graft retention; debridement with a graft and hardware removal followed by early ACL revision reconstruction; even open arthrotomy with the removal of hardware and the curettage of tunnels [1,3,4].
In Gächter stage 3 – complete arthroscopic or open synovectomy is recommended, and in grade 4 infections with osteolysis – an arthrotomy with open debridement. The decision upon graft retention is based on clinical findings: salvaging stable graft or removing in the case of insufficient graft, tunnel misplacement, or persistent infection. After removing the ACL graft, debridement – the curettage of the femoral and tibial graft tunnels should be performed [1,3,4].
The chronic, resistant to initial treatment post-reconstruction knee infection with the involvement of the proximal tibia or distal femur is a severe complication of the ACLR with limited information in literature [5-8]. All reports are case series, mostly retrospective, with case numbers ranging from 4 to 24 (Tab. 1). In the retrospective study by Zalavras et al., 5 patients were treated for persistent infection after ACLR; 3 of 5 had tunnel osteomyelitis and had previously 1 to 3 irrigations and debridement that failed [5]. The treatment protocol consisted of an arthrotomy, complete synovectomy, the removal of implants and graft, the curettage of the femoral and tibial tunnels; patients had a total of 2-4 procedures. O’Neill et al., published one case of proximal tibia osteomyelitis 6 years after ACLR, treated successfully with arthrotomy, and the debridement of the tibial tunnel without graft removal [6]. Numazaki et al., published one case with late infection after ACLR, eventually diagnosed as osteomyelitis of the proximal tibia, which did not heal after the removal of metal implants and was healed after the removal of the artificial graft portion and non-absorbable suture from the bone tunnel [7]. Schub et al., published an observational study of 4 patients with septic arthritis after ACLR – initially treated with arthroscopic lavage, open wound debridement and partial synovectomy. In 3 out of 4 cases the ACL graft was left and 3 out of 4 patients required hardware removal. All cases required, on average, 3.75 procedures for infection eradication [8].
The Authors assume that in chronic, recurrent infection after ACLR with the involvement of the proximal tibia or distal femur and failure of arthroscopic lavage and screw removal, complete revision surgery is necessary with the thorough debridement of all involved compartments – joint and bone canals with subsequent local antibiotic delivery into the graft canal and infected foci in the bone [9]. The basis of this concept is the Silo technique – first described for calcaneal osteomyelitis [10]. A potential local antibiotic carrier to fill the graft canal are the resorbable collagen gentamicin sponge, antibiotic impregnated PMMA beads containing gentamicin (miniseptopal), which must be removed if a reimplantation of ACL graft is planned. We only found 2 papers regarding the use of the collagen gentamicin sponge or PMMA beads containing gentamicin (miniseptopal) [2,11]. Petersen et al., assessed stage adopted protocol in 13 cases with infection after ACLR [2]. In the paper, the authors propose the use of a local antibiotic carrier in the tibial tunnel: a collagen gentamicin sponge or miniseptopal – after the removal of the ACL graft, but from the case description it is not clear if a local carrier was used. In the prospective study by Schultz et al., 24 patients with infection after ACLR were treated on the basis of the infection grade by Gächter [11]. Arthroscopic treatment was performed for infection grades 1 and 2 and arthrotomy for grades 3 and 4. In all cases after debridement the joint space was filled with gentamicin containing PMMA beads (Septopal), in grade 1-2 through an arthroscope, in grade 3-4 by arthrotomy. The Authors used 60 to 180 (mean 78) beads containing 7.5 mg of gentamicin-sulfate each and left the knee open with only a sterile dressing. Usually after 7-10 days PMMA bead chains were removed together with another debridement and secondary closure followed. Graft salvage was possible in 7 out of 24 cases – the decision to conduct graft retention was based on clinical findings. At a follow-up time of 11 to 142 months there was no recurrence of joint or bone infection in all cases, with a mean of 2.2 operations per case required. Bioactive glass without antibiotics has similar properties. In the study by Ferrando and al., antibiotic loaded CSBS showed a significantly better influence on infection treatment than bioactive glass without antibiotics [12].
The concept of local antibiotic application (gentamicin and vancomycin) in chronic osteomyelitis delivered together with a calcium sulphate bone substitute (CSBS) proved to be an effective technique especially in chronic localized tibial osteomyelitis [13]. CSBS is widely used as an antibiotic delivery system in a variety of infections: in diabetic food infection – deposited into bone or soft tissue, in revision surgeries due to breast implant infection, infected hernias and even in alveolar cleft correction surgery [14]. To our knowledge CSBS has not yet been used as a tibial canal filler in infection after ACLR. The aim of this study is the analysis of the results of revision surgery consisting of graft canal debridement and the application of local antibiotics – gentamicin and vancomycin delivered from CSBS, performed due to chronic, recurrent infection after ACLR with the involvement of the bone surrounding the graft canal.

Materials and methods

We present 3 cases (3 knees) – 2 women and one man, aged 28-66 years, suffering from a chronic, recurrent knee infection after ACLR. The ACL graft was made from a semitendinosus and gracilis muscle tendon, tibial fixation with a titanium interference screw and femoral fixation with an endobutton or a suture-anchor. All patients were ineffectively treated in primary surgery centers. Preoperatively all knees were assessed by means of physical examination for sinus tract, effusion, synovial hypertrophy as well as range of motion and stability (Fig. 1A). Implant presence and positioning, signs of osteomyelitis and other pathologies were assessed on radiographs. Magnetic Resonance (MR) showed ACL graft presence, the condition and its healing into the bone, the expansion of inflammatory infiltration around the implants and in graft canals and the communication of an extra-articular infection cluster with the joint cavity. The systemic inflammatory reaction was assessed with the C-reactive protein (CRP) level; causative pathogens were identified in cultures from the sinus tract – if present and from the joint fluid. In 1 case a bone scan with Indium labeled white blood cells (LeucoScan) was performed. In cases No. 2 and No. 3 – preoperatively, the MR revealed ACL graft continuity and limited osteomyelitis affecting the bone around the screw and anchor (Fig. 1B, Fig. 2A-C). Patient No. 3 was treated preoperatively with empiric antibiotic therapy. Table 2 presents more detailed material.
Patient No. 1: a 28 y.o. female had ACLR with a simultaneous lateral meniscectomy and one year later revision arthroscopy and the resection of cyclop deformity. Infection was diagnosed 2 years after ACLR. In total, patient No. 1 underwent 8 surgeries after primary ACLR i.e.: arthroscopic cyclop resection, suprapatellar recess drainage, arthroscopic synovectomy, arthrotomy and synovectomy, ACL graft remnants resection and secondary ACLR with patellar tendon (bone to bone) graft usage, knee drainage, arthroscopic synovectomy with ACL graft preservation and arthroscopic synovectomy with ACL graft removal. Summarizing – the patient encountered two ACLRs and seven revision surgeries. During the last surgery, the pathogens identified were E. Coli and E. Faecium sensitive to gentamicin and vancomycin. Patient No. 2: a 44 y.o. male who had ACLR with simultaneous LCL reconstruction. Infection was diagnosed 6 months after surgery. During the next 3 years the patient had 4 surgeries: 2 arthroscopic lavages and 2 debridement surgeries with interference screw removal from the tibia; moreover, he even had maggot-therapy and bioresonance therapy(sic!). During the last surgery, the pathogens identified were Methicillin Resistant Staphylococcus Aureus (MRSA), sensitive to gentamicin and vancomycin. Patient No. 3: a 66 y.o. female underwent ACLR with partial medial meniscectomy. Infection was diagnosed 7 years after ACLR and was treated unsuccessfully with empiric systemic antibiotic therapy. The patient had revision surgery, but no pathogens were identified pre- or intraoperatively.
Strategy and surgical technique. All surgeries were performed with the patient in the supine position with tourniquet by one surgeon (I.B). All metallic implants were localized under fluoroscopy, and then all implants i.e., the interference screw, endobuttons and non-resorbable sutures were removed (case No. 2 and No. 3) and the meticulous debridement of bone cavities was performed. The position of the surgical curette used for the debridement of the ACL graft canal was controlled under fluoroscopy to assess how deep the canal penetrates within the proximal tibia or if it penetrates the joint space. To assess the possible connection of the fistula with the knee joint cavity, a sterile saline solution was injected into the joint and observed if any leak is visible in the bone canal (case
No. 2). In 2 knees the ACL graft was healed (case No. 2 and
No. 3), in one case it had been previously removed (case No. 1). When the fistula in the ACL canal was communicating with the joint space, open synovectomy with the debridement of the knee joint from the anterior incision with a medial parapatellar capsulotomy was performed. Three tissue biopsies from the synovium and bone were taken for culture – each with separate instruments and for histological examination. In all cases the ACL graft canal was debrided: both the tibial and femoral canal in full length in case No. 1 (after the removal of the ACL graft), and only partially tibial canal in the distal part – in case No. 2 and No. 3 – with a healed graft. Inflammatory foci in the cancellous bone in the distal femur and in the proximal tibia seen in preoperative MR were debrided using a separate approach – outside of the articular surface and filled with antibiotic beads (Fig. 3A-C). After debridement and lavage with a sterile saline solution, 10 cc of calcium sulfate powder (Stimulan, Biocomposite Ltd., UK) were mixed with an addition of 240 milligrams gentamicin and 1 g vancomycin and then formed to 25 cc of beads. We used middle size beads (approximately 4,8-millimeter in diameter) for filling bone cavities – in the tibia and femur and into the ACL graft canal. The entrance to the ACL graft canal from the side of the joint space was sealed with a cylindrical block formed from a compressed bovine collagen sponge 10 × 10 × 0.5 cm containing 130 mg gentamicin (Collatamp® G, Serb) to prevent the extrusion of calcium sulfate material to the joint cavity. The entrance to the ACL graft canal in the tibia was closed from the outside with a cylinder formed from calcium sulfate paste mixed with the same antibiotics as the beads. After achieving the mechanical stability of all calcium sulfate beads, the surgical wound was closed. The study was performed following the Declaration of Helsinki principles and informed consent was obtained from each patient prior to inclusion in the study. The study was approved by responsible authorities of the Carolina Medical Center in Warsaw. Ethics committee approval – not applicable.

Results

The results of revision were assessed on the basis of a physical examination – knee stability, the presence of inflammation signs, the evolution of the CRP level after surgery and the resorption of the bone substitute on radiographs (Fig. 4). Follow-up time ranged from 20 to 26 months. All patients underwent a complete resolution of inflammation, CRP normalization, and have no complications related to CSBS like drainage from the wound or extrusion of the antibiotic carrier to the joint cavity or soft tissue – seen macroscopically, or on radiographs. Radiographs after 6 weeks showed the partial resorption of CSBS beads in the bone of about 50%. On radiographs 3-10 months after revision, the complete resorption of CSBS and no change of joint space or advancement of preoperative arthrosis were observed. After a follow-up time of 20 to 26 months, there was no deterioration in knee range motion and stability in the physical examination – compared to the pre-revision status. Knee stability was acceptable for all patients and none of the patients was willing to undergo ACL reimplantation. In 2 cases there was no need to take CSBGS out. However, we observed complications in case No. 1 – in our assessment unrelated to surgery or CSBS beads themselves. 4 months after surgery and complete operative wound healing, patient No. 1 fell down on the operated leg and unfortunately suffered injury to the skin on the antero-medial part of the proximal tibia, which resulted in a traumatic wound without exposure of the underlying bone. There was no injury to the knee joint itself, no signs of joint effusion or deterioration of knee stability. The patient was able to walk, cycle 80 km and work as an orthopedic department resident. The initially traumatic wound became chronic with the dimensions of 10 × 14 cm and was treated with wet dressings and negative pressure wound therapy. 2 skin graft procedures were performed with the intention to heal the wound, but withuout success. A biopsy performed on the edge of the wound revealed neither evident pathology or specific infiltration. 28 months after the implantation of Stimulan, the remnants of CSBS were removed in another department. The culture from the removed tissue from the bone canal revealed Methicilin Sensible Staph. Aureus (MSSA) sensible to gentamicin and histology revealed abundant neutrophilic infiltration, fibrous tissue and bone trabeculae. The wound diminished for a short time but later went back to its previous condition with no tendency to heal. The patient is now under investigation with a suspicion of pyodermia gangrenosum before an attempt to close the chronic wound with a pedicled flap from the gastrocnemius muscle.

Discussion

Our study regards the most difficult to treat – chronic, relapsing infection, previously treated without success either with suppressive antibiotic therapy or multiple revisions but without the local application of a bone graft substitute with antibiotics – as local antimicrobial therapy and dead space management in the graft canal. We focused on two aspects of chronic infection complicating ACLR, with the involvement of the bone of the proximal tibia or even distal femur – around the graft tunnel: firstly – the strategy of surgical procedure and secondly – the application of CSBS as local antibiotic delivery in the place of the bone infection, taking care to prevent extrusion of the CSBS to the joint cavity. We assessed the effectiveness and safety of the local application of CSBS with the addition of gentamicin and vancomycin in all potential infected bone foci and graft canal in proximal tibia or distal femur – if present in the apreoperative MR. We assumed that an antibiotic loaded bone substitute is not only local antibiotic delivery, but also plays the role of a “dead space management” agent after graft canal debridement. We were aware of bone substitute penetration to the joint cavity, which may lead to joint irritation and secondary to joint destruction. We also assumed that the revision should maintain knee stability existing before surgery and should address all anatomical compartments in which infection persisted.
The purpose of the study was to assess the early results of a new surgical technique for treating chronic infection complicating ACLR, with the involvement of the bone of the proximal tibia or even distal femur – around the graft tunnel. A lack of similar reports in literature makes it difficult to compare our results with other studies. The decision to use CSBS in our infected cases after ACLR was based on the assumption that what is crucial for bone infection treatment is the achievement of a high local concentration of antibiotics – against most common pathogens. Absorbable CSBS seemed to be better than nonabsorbable antibiotic loaded PMMA due to obvious advantages i.e., no need of removal. The application of CSBS is related to possible local and systemic side effects. Wound drainage is probably the most common local side effect, estimated between 3.2-51%, but it should not be associated with infection treatment failure. Pellet size is not connected with its occurrence [15]. It is hard to directly add heterotopic ossification formation to calcium sulfate side effects for not much is completely known about the etiology of its origin, possibly when it is left in skeletal muscles it may act as an ossification center [16].
In our 3 cases we only used small amounts of CSBS and have not observed wound leakage or heterotopic ossification. One of the most catastrophic side effects of CSBS use in the graft canal can be its extrusion into the joint cavity and the mechanical as well as chemical irritation of synovium and the destruction of the joint cartilage. We used our technique of sealing the tunnel of the ACL graft after debridement – from the side of the joint cavity using a “pressed” collagen sponge with gentamicin applied during arthrotomy into the entrance of the tunnel from the joint cavity side. The same technique – the application of a solid CSBS block – was used on the opposite side of the graft canal in order to prevent leakage of CSBS beads into soft tissue and wound space. We did not note any extrusion of the antibiotic carrier to the joint cavity or soft tissue – neither macroscopically nor on radiographs. We did not monitor the calcium blood level among our patients due to the small amount of CSBS used – not exceeding 5-10cc. Also, neither heterotopic ossification nor systemic toxic effects related to gentamicin and vancomycin were observed. Together with CSBS, local doses of 240 mg gentamicin and 1000 mg – per 10 cc of CSBS powder were applied. In our cases, we were able to achieve infection control with only one surgery using absorbable CSBS with antibiotics, without the deposition of any solid material into joint space and without the obvious need of antibiotic carrier removal. In mentioned studies by Petersen and Schulz with the application of a collagen gentamicin sponge or gentamicin containing PMMA beads, 2,2 to 3,4 procedures per patient were necessary for infection control [2,11]. The removal of CSBS performed 28 months after its implantation in the graft tunnel in case No. 1 in our assessment is neither related to the surgery nor the application of CSBS beads themselves. The removal of CSBS does not improve the healing of chronic wounds suspected for pyodermia gangrenosum.
The limitation of our study is the small number of patients and relatively short follow-up (20 to 26 months). Our study showed that the careful preoperative assessment of location of potential infection after ACLRs using radiographs, MRs and even bone scans with Indium labeled white blood cells is useful to adequately address the involved compartment and avoid suboptimal surgical debridement around the infected knee – both the joint cavity if infected, the proximal tibia and distal femur – using local antibiotic delivery similar to the Silo technique in calcaneal osteomyelitis.

Conclusion

The calcium sulfate bone substitute with an addition of gentamicin and vancomycin applied according to the technique described above – in the graft canal and inflammatory foci in the surrounding cancellous bone proved to be effective in the therapy of chronic infection complicating ACLR. Our study showed that this is a relatively safe procedure and unrelated to wound leakage or the extrusion of beads to the joint cavity.

References

1. Stucken C, Garras DN, Shaner JL, Cohen SB: Infections in anterior cruciate ligament reconstruction. Sports health. 2013;5(6):553-557. doi:10.1177/1941738113489099.
2. Petersen W, Herbort M, Höynck E, Zantop T, Mayr H: Stage-adapted treatment of infection after reconstruction of the anterior cruciate ligament. Operative Orthopadie und Traumatologie. 2014;26(1):63-74. doi:10.1007/s00064-013-0262-3.
3. Indelli PF, Dillingham M, Fanton G, Schurman DJ: Septic arthritis in postoperative anterior cruciate ligament reconstruction. Clinical orthopaedics and related research. 2002;(398):182-188. doi:10.1097/00003086-200205000-00026.
4. Burks RT, Friederichs MG, Fink B, Luker MG, West HS, Greis PE: Treatment of postoperative anterior cruciate ligament infections with graft removal and early reimplantation. The American journal of sports medicine. 2003;31(3):414-418. doi:10.1177/03635465030310031501.
5. Zalavras CG, Patzakis MJ, Tibone J, Weisman N, Holtom P: Treatment of persistent infection after anterior cruciate ligament surgery. Clinical orthopaedics and related research. 2005;439:52-55. doi:10.1097/01.blo.0000181499.49740.e5.
6. O’Neill BJ, Molloy AP, McCarthy T: Osteomyelitis of the tibia following anterior cruciate ligament reconstruction. International journal of surgery case reports. 2013;4(2):143-145. doi:10.1016/j.ijscr.2012.10.020.
7. Numazaki H, Kobayashi H, Yoshida K, Hakozaki M, Konno SI: Prolonged infection at the tibial bone tunnel after anterior cruciate ligament reconstruction. Fukushima journal of medical science. 2017;63(2):121-125. doi:10.5387/fms.2017-02.
8. Schub DL, Schmitz LM, Sakamoto FA, Winalski CS, Parker RD: Long-term outcomes of postoperative septic arthritis after anterior cruciate ligament reconstruction. The American journal of sports medicine. 2012;40(12):2764-2770. doi:10.1177/0363546512461903.
9. Babiak I, Pędzisz P, Janowicz J, Małdyk P: Antimicrobial biomaterials in the prevention and local treatment of infection in orthopedics. Polish Orthopaedics and Traumatology 2018; 83(6) 221-228 DOI: 10.31139/chnriop.2018.83.6.44.
10. Drampalos E, Mohammad HR, Kosmidis C, Balal M, Wong J, Pillai A: Single stage treatment of diabetic calcaneal osteomyelitis with an absorbable gentamicin-loaded calcium sulphate/hydroxyapatite biocomposite: The Silo technique. Foot (Edinburgh, Scotland). 2018;34:40-44. doi:10.1016/j.foot.2017.11.011.
11. Schulz AP, Götze S, Schmidt HGK, Jürgens C, Faschingbauer M: Septic arthritis of the knee after anterior cruciate ligament surgery: a stage-adapted treatment regimen. The American journal of sports medicine. 2007;35(7):1064-1069. doi:10.1177/0363546507299744.
12. Ferrando A, Part J, Baeza J: Treatment of Cavitary Bone Defects in Chronic Osteomyelitis: Biogactive glass S53P4 vs. Calcium Sulphate Antibiotic Beads. Journal of bone and joint infection. 2017;2(4):194-201. doi:10.7150/jbji.20404.
13. Zhou CH, Ren Y, Ali A et al.: Single-stage treatment of chronic localized tibial osteomyelitis with local debridement and antibiotic-loaded calcium sulfate implantation: a retrospective study of 42 patients. Journal of orthopaedic surgery and research. 2020;15(1):201. doi:10.1186/s13018-020-01721-7.
14. Qin CH, Zhou CH, Song HJ et al.: Infected bone resection plus adjuvant antibiotic-impregnated calcium sulfate versus infected bone resection alone in the treatment of diabetic forefoot osteomyelitis. BMC Musculoskeletal Disorders. 2019;20(1):1-8. doi:10.1186/s12891-019-2635-8.
15. McPherson, MD FACS E, Dipane, BA M, Sherif, MD S: Dissolvable Antibiotic Beads in Treatment of Periprosthetic Joint Infection and Revision Arthroplasty – The Use of Synthetic Pure Calcium Sulfate (Stimulan®) Impregnated with Vancomycin & Tobramycin. Reconstructive Review. 2013;3(1). doi:10.15438/rr.v3i1.27.
16. Raina DB, Gupta A, Petersen MM et al.: Muscle as an osteoinductive niche for local bone formation with the use of a biphasic calcium sulphate/hydroxyapatite biomaterial. Bone & joint research. 2016;5(10):500-511. doi:10.1302/2046-3758.510.BJR-2016-0133.R1.