Current therapeutic interventions combating biofilm-related infections in orthopaedics

Abstract

Aims

Biofilm-related infection is a major complication that occurs in orthopaedic surgery. Various treatments are available but efficacy to eradicate infections varies significantly. A systematic review was performed to evaluate therapeutic interventions combating biofilm-related infections on in vivo animal models.

Methods

Literature research was performed on PubMed and Embase databases. Keywords used for search criteria were “bone AND biofilm”. Information on the species of the animal model, bacterial strain, evaluation of biofilm and bone infection, complications, key findings on observations, prevention, and treatment of biofilm were extracted.

Results

A total of 43 studies were included. Animal models used included fracture-related infections (ten studies), periprosthetic joint infections (five studies), spinal infections (three studies), other implant-associated infections, and osteomyelitis. The most common bacteria were Staphylococcus species. Biofilm was most often observed with scanning electron microscopy. The natural history of biofilm revealed that the process of bacteria attachment, proliferation, maturation, and dispersal would take 14 days. For systemic mono-antibiotic therapy, only two of six studies using vancomycin reported significant biofilm reduction, and none reported eradication. Ten studies showed that combined systemic and topical antibiotics are needed to achieve higher biofilm reduction or eradication, and the effect is decreased with delayed treatment. Overall, 13 studies showed promising therapeutic potential with surface coating and antibiotic loading techniques.

Conclusion

Combined topical and systemic application of antimicrobial agents effectively reduces biofilm at early stages. Future studies with sustained release of antimicrobial and biofilm-dispersing agents tailored to specific pathogens are warranted to achieve biofilm eradication.

Cite this article: Bone Joint Res 2022;11(10):700–714.

Article focus

• In this review, we analyzed current animal models, interventions, and outcome measures of biofilm-related bone infections in vivo.

• We elucidated the research gap in effective methods and therapies for the eradication of biofilm-related bone infections, and provided insight for future experimental designs targeting biofilm in animal studies.

Key messages

• Recent findings on the dynamic evolution from biofilm formation, and maturation to distant repopulations, will facilitate the design of targeted therapy.

• Given the limited effects of systemic mono-antibiotics, a combined local and systemic therapy is recommended to achieve effective biofilm reduction.

• For established biofilms, a tailored release with matrix-dispersing, antimicrobial, and bone-forming agents loaded on biodegradable hydrogels would potentially eradicate infection and facilitate bone regeneration.

Strengths and limitations

• The current study provides a comprehensive overview of animal model development and biofilm characterizations in different orthopaedic scenarios, current therapeutic interventions, and effectiveness for biofilm-related bone infection research.

• Due to the heterogeneity of the interventions and outcome measures, only qualitative analysis was performed.

Introduction

Device-related infections (DRIs) are a major concern in orthopaedic surgery. Despite tremendous efforts to reduce the risk, these events still occur. The rate of periprosthetic joint infection (PJI) is present in 1% to 2% of all cases. ¹ The increasing number of open fractures in aged patients also adds to the risk of fracture-related infection (FRI), ^2,3 which occurs in 4% to 52% of Gustilo Anderson grade III ⁴ fractures with concomitant soft-tissue damage. ⁵ Based on different diagnostic techniques and heterogeneous patient populations, the risk for implant-associated spinal infection ranged from 0.5% to 10%. ⁶ One of the critical issues in the clinical treatment of DRI is biofilm formation, which leads to antibiotic tolerance, infection recurrence, ^1,7 and poor clinical outcomes. Biofilm-related infections often cause prolonged disability, recurrent hospital admissions, and even patient mortality, triggering a huge care burden. The estimated cost per patient reaches USD $17,000 to $150,000. ⁸

Biofilm is defined as clusters of microorganisms that are adhered to biological or non-biological surfaces, often encased in an outer polymer layer. ⁹ Mediated by quorum-sensing signalling systems, bacteria cells undergo orchestrated biofilm formation, maturation, release of virulent factors, and dispersal in a population-based manner. ^10,11 In a mature biofilm, the non-growing bacteria cells and high-density extracellular polymeric substances (EPSs) cause antibiotic tolerance and persistent infection. ¹² In clinical settings, the success rate of debridement, antibiotics, and implant retention (DAIR) for acute PJI reaches 92.3%, but only 44.4% for late-onset infections. ¹³ Comparatively, the success rate of DAIR is 86% to 100% for acute FRI, but only 67% for late-onset FRI. ¹⁴ The major difficulty is biofilm formation, which often requires debridement, implant exchange/cement spacer for PJI, ¹⁵ and external fixation for FRI. ³ Despite two-stage strategies, reinfection rates reach 8.4% and 16.2% for hip and knee arthroplasties, respectively. ¹⁶ Given the progressive nature of biofilm-related infection and low success rate of treatment at late stages, one study has suggested that the best possible treatment is to inhibit bacteria attachment and prevent biofilm maturation at the beginning. ¹⁷

Therefore, research and development of biomaterials targeting biofilm-related infection have been of great interest in the recent decade. Various techniques to prevent DRI, including titanium and copper alloy implants, antimicrobial surface coatings, antimicrobial agent(s) loaded scaffold, or hydrogel have been reported. ^18-20 However, emerging cases of antibiotic-resistant strains, polymicrobial infections, and biofilms pose new challenges. ^21,22 To our knowledge, numerous studies have reported successful eradication of biofilm in vitro, however there is a lack of validation on clinically relevant animal models before translation. ^23-25 The purpose of this study was to summarize relevant animal models, bacterial strains, evaluation of biofilm and bone infection, complications, key findings, observations, prevention and treatment of biofilm to provide information for development of novel treatments, and future clinical translation.

Methods

Search strategy

PubMed and Embase (date last accessed 1 February 2022) were searched. Keywords used for search criteria were “bone AND biofilm”.

The inclusion criteria were: 1) preclinical studies; 2) use of animal models; and 3) study on biofilm-related DRI. The exclusion criteria were: 1) lack of analysis or evaluation of biofilm ex vivo; 2) in vitro study; 3) review article; 4) abstract or conference paper; and 5) non-English article.

Selection of studies was based on the evaluation of biofilm. ⁹ Eligible studies must include the confirmation of in vivo biofilm formation by visualization or imaging techniques, and the quantification of biofilm mass and colony-forming unit (CFU) load. Two independent reviewers (JL, RMYW) screened all titles and abstracts, and performed the selection from the search results based on inclusion and exclusion criteria. Each article was reviewed with any disagreement resolved by discussion until consensus was met.

Data extraction and analysis

For eligible studies, the two reviewers extracted information on: 1) species and strains of the animal; 2) bacteria strain; 3) animal model characteristics; 4) evaluation of biofilm and bone infection; 5) complications; and 6) key findings on observational study, prevention of biofilm formation, and treatment of biofilm. Due to the large heterogeneity in animal models and methodology, a qualitative review was performed.

Results

Characteristics of the papers

A total of 1,681 and 2,122 studies were identified from Embase and PubMed, respectively. After removal of duplicated entries, 2,071 records remained. Each title and abstract were reviewed, and 1,969 records were excluded based on criteria. Upon detailed review of the full text, an additional 59 studies were excluded: in vitro study of antibiotics, bone substitutes, biomaterials, and other agents (n = 39); lack of explicit biofilm evaluation (n = 18); and studies on the dental bone (n = 2). Finally, a total of 43 studies published from 1998 to 2021 (Table I and Table II) were included in our systematic review (Figure 1). ^11,26-67

Table I.

Summary of the study characteristics.

Study	Animal	Model features	Bacteria strain and inoculation methods	Clinical scenarios	Intervention	Systemic antibiotics	Treatment duration	Biofilm evaluation
Li et al 39	SD rats	Osteomyelitis (femur Intramedullary nail with infection)	MRSA pre-cultured on the implant	Implant associated infection	Magnesium implant	None	None	CV staining FESEM RTPCR: icaA, agr RNAIII expression
Hazer et al 53	14 New Zealand rabbit	Pedicel screw related infection in the lumbar spine	Planktonic MRSA at 106	Implant associated spinal infection	Plyethylene glycol grafted, Polypropylene-base silver nanoparticles	None	None	CFU counting SEM
Yao et al 67	50 SD rat	Titanium rod inserted into femur	Planktonic S. aureus ATCC25923 at 106	Implant associated infection	enoxacin -loaded mesoporous silica nanoparticles	None	None	SEM live and dead staining CFU counting
Li et al 54	56 New Zealand white rabbits	Intramedullary nail inserted in tibia	Planktonic MRSA at 106	Implant associated infection	Mg-Cu with 0.25% Cu implant	None	None	CFU plating CLSM FESEM CV staining RT-PCR: biofilm-associated genes including atlE, clfA,
Min et al 40	Adult SD rats	Implant inserted at tibia	Planktonic S. aureus Xen 29. at 105	Implant associated infection	Implant coating with sequential release of gentamicin and BMP-2	None	None	SEM CFU counting
Gracia et al 66	36 Wistar male rats	Metal inserted into femur	S. aureus strain ATCC 29213 biofilm at 109 pre-cultured on implant	Implant associated infection	None	21 day systemic antibiotic treatment using cefuroxime, vancomycin, or tobramycin	None	CFU counting in blood, bone tissue, and implant
Tran et al 41	SD rats	Plate and screw fixation at femur	Planktonic MRSA and MRSE at 105	Implant associated infection	Selenium nanoparticle coating	None	None	CFU counting CLSM (in vivo)
Ashbaugh et al 26	C57BL/6 mice	Intramedullary K-wire fixation at femur	Planktonic S. aureus Xen36 at 103	Implant associated infection	Electrospinning of PLGA and PCL with Vancomycin, Rifampin, linezolid and daptomycin.	None	None	SEM bioluminescent monitoring of S. aureus. CFU counting
Carli et al 27	20 C57BL/6 mice	Tibia-implant insertion (arthrotomy)	Planktonic S. aureus Xen36	PJI	Vancomycin-loaded polymethylmethacrylate spacers	None	None	SEM
Inzana et al 33	13 to 15 weeks BALB/cJ mice	6-hole polyether ether ketone plate with titanium coating at femur	S. aureus Xen36 pre-cultured on a Collagen sheet at 104	FRI	Rifampin and Vancomycin laden calcium phosphate scaffolds (CPS)	None	None	SEM CFU counting
van der Horst et al 42	85 Female SD rats	A titanium wire wrapped high-density polyethylene and chrome cobalt at femur	Planktonic S. aureus 25923 at 106	Implant associated infection	None	Systemic ceftriaxone for 5 or 10 days	Tobramycin, gentamycin, rifampin, vancomycin	CV staining CFU counting
Greimel et al 48	61 Male Wistar rats	Intravenous catheter implanted into femur	Planktonic S. aureus 29213 at 108	Implant associated infection	None	Rifampin (Rif); flucloxacillin (Flu); flu+ Rif; moxifloxacin (Mox); Mox+ Rif	2 weeks antibiotics treatment starts from day 7	CFU counting
Zhu et al 43	60 Adult SD rats	A titanium rod inserted into the tibia	S. aureus 25,923 biofilm cultured on implant at 20 h	Implant associated infection	Human β-defensin 3	None	None	Live and dead bacterial viability assay assessed by CLSM
Jensen et al 59	25 Pigs	A k-wire inserted at the proximal tibia	Planktonic S. aureus S54F9 spa type t1333 at 104	Implant associated infection	None	None	Single dosage in 2X 160, to160,000 times MIC Gentamycin injected locally	Immunohistochemical staining CFU counting
Boles et al 28	C57BL/6 mice	K-wire inoculated with bacterial inserted into femur	Planktonic S. aureus (UAMS-1) and E. Coli (ATCC 25922) at 104	Implant associated infection	Chitosan sponge loaded with amikacin and vancomycin	None	None	CFU counting
Singh et al 60	5 Yorkshire pig	Surgical implantation of a titanium rods in spinous process	Planktonic MRSA solution at 106	Implant associated spinal infection	Negative pressure wound therapy with antiseptic instillation	None	None	Tissue culture SEM
Kandemir et al 52	26 Rats	A silicone drain inserted in the medullary canal	Planktonic Pseudomonas aeruginosa solution at 108	Implant associated infection	None	Subcutaneous injection of Ceftazidime and Clarithromycin	None	CFU counting SEM
Inzana et al 34	Female BALB/cJ mice	Transverse osteotomy at femur fixed with polyether ether ketone (PEEK) plate	S. aureus Xen36 pre-cultured on implant overnight	FRI	Polymethyl methacrylate spacer loaded with vancomycin	Vancomycin subcutaneous injection	None	SEM CFU counting Gram staining
Fang et al 49	60 Wistar rats	18 G needle inserted into the femur bone marrow cavity	Planktonic S. aureus (MSSA BCRC10451)	Implant associated infection	Nano-particle-induced hyperthermia therapy combined with vancomycin therapy	Systemic vancomycin for 40 days	Vancomycin injection into the cavity for 40 days	SEM CFU counting Cango red staining Gram staining (biofilm in bone)
Tomizawa et al 35	125 BALB/c rats	Stainless-steel implant inserted at the tibia	S. aureus (UAMS-1) pre-cultured on implant overnight	Implant associated infection	None	cefazolin, gentamycin, and vancomycin with or without rifampin for 14 days started from 0, 3, 7 days	None	SEM CFU counting
Lindsay et al 29	58 Male C57BL/6 j mice	gauge needle fixation at femur with osteotomy	Planktonic S. aureus 29,213	FRI	None	Oral antibiotics (cephalexin) for 14 days	Metalloporphyrin Antioxidant (MnTE-2-PyP) injection start one-day post-surgery	CFU counts
Jørgensen et al 30	65 Female C57B6 mice	A pin inserted transcortically in the tibia metaphysis	S. aureus Xen29 and S. aureus Xen 31 pre-cultured on implant	Implant associated infection	None	Systemic injection of Vancomycin (110 to 180 mg/kg) for 14 days	None	SEM Live/dead staining; CLSM CFU counting bioluminescent monitoring
Shiono et al 65	18 BALB/c mice	A hole drilled at distal femur with or without an implant	Planktonic Propionibacterium acnes at 106	Implant associated infection	None	None	None	SEM imaging Fluorescent bacterial probe detection
Thompson et al 31	Male C57BL/6	K-wire inserted into the femur	Planktonic Pseudomonas aeruginosa Xen 41, E. coli Xen 14, E. coli ATCC25922, E. coli ATCC K12 at 105	PJI	Bispecific antibody MEDI3902 targeting Biofilm related antigens PcrV and Psl	None	None	bioluminescent imaging SEM
Nishitani et al 11	C57BL/6 and BALB/c mice	L shaped implant inserted to the tibia	S. aureus (UAMS-1, Xen 40, agr gene mutant UAMS-1Δagr) pre-cultured on implant	Implant associated infection	None	None	None	SEM bioluminescent imaging CFU counting and PCR for bacteria load quantification
Gahukamble et al 55	New Zealand white rabbits	Intramedullary nail fixation	Planktonic Propionibacterium acnes LED2 (from a clinical isolate) and Staphylococcus lugdunensis at 107	Implant associated infection	None	None	None	bacteria molecular analysis SEM (bacteria in medullary canal) CFU counting
Lovati et al 50	24 Wistar rats	Osteotomy at the femur fixed with plate and screw	Planktonic MRSE GOI1153754-03-14 at 103 105 108	FRI	None	None	None	CFU counting Gram staining SEM
Johansen et al 61	7 Yorkshire-Landrace-cross pig	Femur artery injection	Planktonic S. aureus S54F9, S. aureus NCTC-8325 to 4, S. aureus UAMS-1 at 104	Hematogenous osteomyelitis	None	None	None	CFU counting IHC staining for S. aureus Fluorescent in situ hybridization
Zhang et al 56	24 New Zealand rabbits	Plate and screw fixation with osteotomy	Staphylococcus. aureus 25,923 biofilm pre-cultured on implant	FRI	None	None	None	SEM of the plate CLSM
Cahill et al 32	C57BL/6 j mice	Tibia fracture fixed with pin	Planktonic MRSA USA 300 at 106	FRI	None	Systemic injection of vancomycin and rifampin for 3 days started on day 0	local injection rifampin in hydrogel	CFU counting Gram staining
Marston et al 44	SD rat	Intramedullary pin within distal femur	Planktonic S. aureus at 104	Implant associated infection	None	Systemic ceftriaxone treatment for 4 weeks	tobramycin or doxycycline powder placed on pin and soft tissue	CFU counting SEM
Wei et al 51	32 Wistar rat	A screw inserted at the knee	Planktonic MRSA (ATCC BAA-1026) at 108	PJI model	None	Vancomycin with IP injection for 14 days	intra-articular injection for 14 days	CFU counting SEM
Lovati et al 38	32 CD 1 mice and NOD/ShiLtJ mice	Gauge inserted into femur	Planktonic S. aureus ATCC 25923 at 105	PJI in diabetic patients	None	None	None	SEM CFU counting
Hu et al 47	12 SD rat	Intramedullary Neil inserted into femur	MSSA ATCC 25923, Escherichia coli ATCC 25922 pre-cultured on the implant	Fracture related infection	TaON-Ag Nanocomposite coated titanium	None	None	SEM imaging
Stewart et al 64	9 Sheep	Tibia osteotomy fixed with a titanium plate	Planktonic S. aureus at 106	Fracture related infection	Vancomycin modified AEEA-AEEA-APTS-Ti surface	None	None	SEM imaging CFU counting Live and dead staining
Tomizawa et al 37	49 BALB/c female mice	L shaped rod inserted into the tibia	Planktonic S. aureus USAA300 LAC and S. epidermidis RP62A at 105	Implant related infection	None	None	None	CFU counting SEM Gram staining
Schaer et al 63	16 Sheep	Compression plate fixation after tibia osteotomy	Planktonic S. aureus ATCC 25923 at 106, 108,10 11	Fracture related infection	hydrophobic polycation N, N- dodecyl,methyl-PEI (PEI 1⁄4 polyethylenimine) coated surface	None	None	SEM CFU counting
Gordon et al 57	14 Rabbit	Pedicel screw and titanium plate	Planktonic Community-acquired MRSA strain SAP231 at 104,105,106	Implant associated spinal infection (IASI)	None	None	None	SEM bioluminescent imaging CFU counting
Hadden et al 46	8 Sprague-Dawley rats	cemented hemiarthroplasty	Planktonic Staphylococcus aureus Xen36 at 108	PJI related infection	None	None	None	Field emission scanning electron microscopy In vivo luminescent imaging Tissue culture
Windolf et al 36	Balb/c mice	Osteotomy at the femur with plate fixation	Planktonic Staphylococcus aureus ATCC 29213	FRI	None	None	None	SEM CFU counting Giemsa staining
Wu et al 45	10 Sprague Dawley rats	A hole in tibia	Planktonic MRSA ASyycG over-expression, and MRSA ATCC29213	Osteomyelitis	None	None	None	647-labelled dextran conjugate, and SYTO9 labeling; CLSM SEM Gram staining FISH
Hovis et al 58	18 New Zealand rabbits	Plate and screw fixation at the tibia	Planktonic MRSA at 108	Implant associated infection	None	None	125 mg vancomycin powder applied direct to the implant	CFU counting SEM imaging
Blirup-Plum et al 62	9 pigs	A K-wire insert into tibia	Planktonic S. aureus S54F9 at 104	Implant associated infection	Injectable ceramic bone graft substitute loaded with gentamycin and debridement surgery	Intramuscular injection of gentamycin	None	Immunohistochemistry staining for biofilm CFU counting

CFU, colony-forming unit; CLSM, confocal laser scanning microscope; CV staining, crystal violet staining; ELISA, enzyme-linked immunosorbent assay; FESEM, field emission scanning electron microscopy; FISH, fluorescence in situ hybridization; FRI, fracture-related infection; IHC, immunohistochemistry; LC/MS, liquid chromatography-mass spectrometry; MRSA, methicillin-resistant Staphylococcus aureus; PEI, polyethyleneimine; PJI, periprosthetic joint infection; PLGA, poly(lactic-co-glycolic acid); PMMA, polymethyl methacrylate; ROS, reactive oxygen species.

Table II.

Key findings of each study.

Study	Key findings	Other assessments conducted
Li et al 39	1) Mg was highly effective against MRSA-induced osteomyelitis and improved the peri-implant bone formation. 2) The antibiofilm effects of Mg were achieved by reducing bacterial icaA and agr RNAIII transcription levels.	X-ray99mTc radioactivity emission CT Micro-CT CFU plating
Hazer et al 53	1) PP-g-PEG-A g grafted pedicle screw showed antimicrobial effect and inhibit biofilm formation. 2) Complications: 4 of 14 rats had pus formation localized in the fascia plane.	Radiograph H&E staining of muscle tissue
Yao et al 67	The current study provides a novel biomaterial in preventing Staphylococcus aureus related implantation infections and bone loss.	Micro-CT H&E and TRAP staining Drug release study In vitro antimicrobial study
Li et al 54	The Mg-Cu alloy showed antibacterial ability demonstrated by microbiological and biofilm formation assays with reduced expression of biofilm, virulence, and antibiotic-resistant genes.	Radiograph Bone histology: H&E and Giemsa staining
Min et al 40	The rapid release of antibiotics and sustained release of BMP-2 successfully eradicated the biofilm and accelerated bone tissue formation.	Bioluminescent monitoring Micro-CT Pull-out tensile test Histology: H&E, Masson’s trichome, TRAP, and gram staining
Gracia et al 66	Cefuroxime significantly reduced the bacteria load on bone and K-wire, which was consistent with the antimicrobial effects of 48 hours biofilm in vitro.	Serum antibodies against S. aureus
Tran et al 41	The nanoparticles coating strongly inhibited biofilm formation on the implant and reduced the number of CFU in the surrounding tissue.	In vitro characterization of nanoparticles In vitro antimicrobial properties Biocompatibility test
Ashbaugh et al 26	The polymeric coating can be applied to deliver various antibiotics to prevent biofilm-associated orthopaedic infection by varying the PLGA versus PCL ratios.	Radiograph Micro-CT Sanderson’s and acid fuchsin counterstaining
Carli et al 27	The antimicrobial effects of PMMA spacers fail to eradicate periprosthetic joint infection in the clinically representative mouse model.	Radiograph Serum amyloid A
Inzana et al 2015 33	Co-delivery of rifamycin and vancomycin from 3D-printed CPS significantly reduced the bacteria burden but cannot fully eradicate the biofilm on implant.	Bioluminescent imaging Radiograph and micro-CT
van der Horst et al 42	1) 5 days daily injection significantly reduced CFU but cannot eradicate. 2) 10 days of systemic ceftriaxone and local gentamicin showed complete clearance.	None
Greimel et al 48	1) Only mono rifampin can significantly reduce the biofilm on the implant but not the bone and soft tissue. 2) Mox plus rif or Flu plus rifamycin showed significant reduction in CFU in bone, soft-tissue and biofilm, and Mox + rif showed eradication of biofilm on implant, but not on bone tissue.	None
Zhu et al 43	β-defensin 3 inhibits the bacterial growth by regulating inflammatory and immune response in the MRSA-induced implant biofilm infection. Complications: 3 rats died with swelling and white purulent secretion on the wound.	ELISA test: IL-10, TNF-α, IL-1α, and interferon-γ IHC staining of NF-κB and TLR-4
Jensen et al 59	1,600 times of MIC is required to prevent the bacteria attachment, indicating that susceptibility in intro may not reflect in vivo susceptibility.	Bone H&E staining
Boles et al 28	Chitosan loaded vancomycin and amikacin (5 mg/ ml) showed higher percentage of clearance rate, which can be further augmented by double the serum concentration.	None
Singh et al 60	NPWTi therapy is associated with decreased bacterial load and biofilm formation compared to wet-to-dry wound dressing.	None
Kandemir et al 52	Clarithromycin enhanced the activity of concomitantly used bactericidal agents by destroying the biofilm formation.	None
Inzana et al 2015 34	PMMA-loaded vancomycin only showed significant effect of decreasing the bacterial burden and osteolysis when combined with systemic antibiotics in a revision model.	Bioluminescence imaging Histology: ABH/Orange G, TRAP Micro-CT
Fang et al 49	1) Systemic application of vancomycin did not eradicate the biofilm infection. 2) Magnet nanoparticles combined with local administration of vancomycin enhance the eradication of bacteria in the biofilm-based colony.	H&E staining Micro-CT
Tomizawa et al 35	Combination with rifampin is recommended to inhibit implant associated osteomyelitis, due to the limited effects of monotherapy, especially cefazolin.	None
Lindsay et al 29	The S. aureus biofilm is redox-sensitive and ROS scavenger treatment improved the efficacy of antibiotic treatment on bacteria clearance.	Crystal violet staining Nitroblue tetrazolium assays
Jørgensen et al 30	1) This model is suitable for testing antimicrobial agent treatment as both biofilm and CFU can be assessed. 2) 14 days vancomycin injection was unable to eradicate biofilm infection. 3) Key research gap: despite clear dosage, the serum vancomycin levels cannot be monitored.	None
Shiono et al 65	The presence of an implant is essential for the development of delayed surgical site infection model.	Myeloperoxidase activity Bone histology Genetic confirmation of C. acnes by PCR
Thompson et al 31	1) In vitro biofilm-producing activity was associated with the in vivo gram-negative bone infection characterized by bacteria infection, biofilm formation reactive bone changes and inflammatory cells infiltration. 2) Biospecific antibody-targeting Pseudomonas aeruginosa virulence factors reduced the bacteria burden in vivo.	Radiograph imaging Bone histology PET imaging Flow cytometry
Nishitani et al 11	1) This study showed the S. aureus attachment, proliferation and maturation from day 0 to day 7. 2) Biofilm dispersal was achieved by S. aureus migration in an agrin-dependent way, as presented with empty lacunae and retention of few culture-negative, RNA-positive residual bacteria.	None
Gahukamble et al 55	1) C. acnes and S. lugdunensis infection model caused different clinical presentations, including low-grade infection in C. acnes and acute infection in S. lugdunensis. 2) This model is relevant to the periprosthetic joint infection and nonunion fracture fixation, which may be reported as aseptic failure.	H&E staining Modified Brown and Brenn staining
Lovati et al 50	1) The severity of osteomyelitis signs and nonunion rate was dosage-dependent. 2) This study provides a relevant preclinical model for subclinical infections in orthopaedic trauma.	WBC count Micro-CT H&E staining CV staining
Johansen et al 61	Bacteria embedded in the opaque biofilm matrix was demonstrated by FISH in a haematogenously spread osteomyelitis model.	Blood culture for bacteremia Micro-CT H&E staining
Zhang et al 56	This study provides a novel rabbit model of infection following internal fixation with biofilm formation.	Radiograph and micro-CT H&E staining
Cahill et al 32	Local application of high-dosage rifampin-loaded hydrogel reduced the bacteria load, which was further enhanced when combined with systemic rifampin application. Complications: two rats died within 24 hours postoperatively with no specified course.	Radiograph and micro-CT H&E staining Immunohistochemistry for IL-1β, p-p65, Sox9, Runx2
Marston et al 44	Local tobramycin showed more significant CFU reduction than doxycycline in the synovium, supporting the current evidence of local application of antibiotics	Monitoring the serum antibiotics concentrations by LC/MS Bone histology
Wei et al 51	IA injection is superior to systemic injection, whereas the combined treatment can eradicate the infection in the two-week course after revision surgery.	Serological analysis of vancomycin Radiograph and micro-CT H&E staining
Lovati et al 38	Diabetic mice challenged with a single inoculum of S. aureus displayed severe osteomyelitis changes and biofilm formation on implant. Complications: two mice had severe signs of infection including joint abscess and fistulae.	Serum white blood cell and CRP Micro-CT analysis H&E staining Gram staining
Hu et al 47	TaON-Ag nanocomposite coated titanium inhibited pathogen adhesion and biofilm formation in both S. aureus and Escherichia coli in vivo.	H&E staining Radiograph In vitro antibacterial assay
Stewart et al 64	Vancomycin-derivatized plate surfaces inhibited implant colonization with S. aureus and supported bone healing in an infected large animal model.	Calcified tissue staining Radiograph Micro-CT
Tomizawa et al 37	Biofilm‐producing S. epidermidis RP62A does not cause prominent osteolysis, reactive bone formation, but persists in biofilm, stimulates a low-grade proinflammatory environment, and inhibits osseous integration.	Micro-CT: bone healing H&E staining TRAP staining Proinflammatory cytokines transcriptome in tibia
Schaer et al 63	The presence of a N, Ndodecyl,methyl PEI coating on the surface of a metal implant was effective in eliminating the clinical signs of infection, preventing biofilm formation and support bone healing.	Von Kossa staining Safranin O staining Micro-CT
Gordon et al 57	This rabbit model could serve a valuable preclinical model of IASI to study the pathogenesis and novel diagnostic and therapeutic methods, which allows for real-time monitoring of bacteria burden and inflammation.	FDG-PET/CT imaging Micro-CT analysis
Hadden et al 46	The study showed a new, high-fidelity model of in vivo PJI using cemented hip hemiarthroplasty in rats. Complications: One rat with prosthesis dislocation and one with implant loosening.	Gait analysis MRI Micro-CT
Windolf et al 36	Implant-associated localized osteitis in murine femur fracture by biofilm-forming S. aureus was established, with increased leucocyte count and IL-6 levels	Lavage: cell sorting by flow cytometry, IL-6 expression H&E staining
Wu et al 45	The overexpression of ASyycG leads to a reduction in biofilm formation and bacterial pathogenicity in vivo.	qPCR: inflammatory markers and biofilm-related genes H&E staining Micro-CT
Hovis et al 58	Vancomycin spreading at the infection site successfully prevents infection of the bone and implant in all cases.	H&E staining Serum vancomycin measurement
Blirup-Plum et al 62	The injectable ceramic bone graft substitute loaded with gentamycin cannot be used as a standalone alternative to extensive debridement or be used without the addition of systemic antibiotics.	CT scanning FISH Bone histology: H&E and Masson-trichome staining Gentamycin concentrations

ABH, alcian blue/haematoxylin; BMP-2, bone morphogenetic protein 2; CFU, colony-forming unit; CPS, calcium phosphate scaffold; CV, crystal violet; ELISA, enzyme-linked immunosorbent assay; FDG, fluorodeoxyglucose; FISH, fluorescence in situ hybridization; H&E, haematoxylin and eosin; IA, intra-articular; IASI, implant-associated spinal infection; IHC, immunohistochemistry; IL, interleukin; K-wire, Kirschner-wire; LC/MS, liquid chromatography-mass spectrometry; MIC, minimum inhibitory concentration; MRSA, methicillin-resistant Staphylococcus aureus; NF-κB, nuclear factor kappa B; NPWTi, negative pressure wound therapy with instillation; PCL, poly-є-caprolactone; PCR, polymerase chain reaction; PEI, polyethyleneimine; PET, positron emission tomography; PJI, periprosthetic joint infection; PLGA, poly(lactic-co-glycolic acid); PMMA, polymethyl methacrylate; ROS, reactive oxygen species; TLR-4, toll-like receptor 4; TNF-α, tumour necrosis factor-alpha; TRAP, tartrate-resistant acid phosphatase.

Fig. 1.

Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) flowchart of study selection.

Animals

Mice were used in 15 studies including eight with C57BL6, ^11,26-32 eight with BALC/c, ^{11,30,33-37,65} and CD 1 mice and NOD/ShiLtJ mice in one study. ³⁸ Rats were used in 16 studies, including ten using Sprague-Dawley rats, ^39-47,67 five using Wistar rats, ^48-51,66 and one with no specified strain. ⁵² The other models were developed with New Zealand rabbits in six studies, ^53-58 pigs in four studies, ^59-62 and sheep in two studies. ^63,64

Bacteria strains

Staphylococcus aureus was used in 25 studies, methicillin-resistant S. aureus (MRSA) in ten studies, and methicillin-resistant Staphylococcus epidermidis (MRSE) in two studies. Other staphylococcus strains used included Staphylococcus lugdunensis ⁵⁵ and S. epidermidis RP62A. ³⁷ Bioluminescently engineered strains including S. aureus Xen 29, Xen 31, and Xen36 were applied to allow continuous monitoring of bacteria load. ^{11,26,30,31,33,34,40,46} An additional gram-positive species was Propionibacterium acnes, ^55,65 which is now reported as Cutibacterium acnes. Gram-negative strains included Pseudomonas aeruginosa, ^31,52 and various strains of Escherichia coli, ^28,31,47 among which one study developed a polymicrobial infection model with S. aureus and E. coli. ²⁸

Bacteria isolated from a patient with PJI were used in one study, and strain characterization based on polymerase chain reaction (PCR) was performed to identify the C. acnes strain LED2 and S. lugdunensis 010729. ⁵⁵ One study reported genetic confirmation of C. acnes isolated from an implant, showing that the bacteria strain could survive for over six months. ⁶⁵ Additionally, one study used agr gene mutant S. aureus (UAMS-1)Δ agr, ¹¹ and showed that biofilm formation was agr-dependent. Another study found that MRSA with antisense yycG overexpression had reduced biofilm formation and pathogenicity in a rat osteomyelitis model. ⁴⁵

Animal models

Animal models were developed to mimic FRI in ten studies, ^{29,32-34,36,47,50,56,63,64} PJI in five studies, ^{31,38,46,51,68} implant-associated spinal infection model (IASI) in three studies, ^53,57,60 and other implant-associated bone infections. The healing in FRI was monitored by radiograph analysis, ^32,33,56,64 micro-CT, ^{32-34,50,56,63,64} and bone histology. ^{32,36,47,50,56} The anatomical location of infection included the femur in 20 studies, tibia in 15 studies, spine in three studies, ^53,57,60 and both the knee and hip joint in five studies. ^{31,38,51,68,46} A total of 41 studies used implants, including intramedullary nails, screws, or pins in 14 studies, ^{26,28,29,32,35,39,40,44,47,49,54,55,59,62} plates and screw fixation, ^{33,34,36,41,50,56,63,64} joint prostheses, ^{27,31,46,51,68} pedicle screws or rods, ^53,57,60 transcortical pins, ³⁰ and other implants. ^{11,37,38,42,43,48,52,65-67} Among the studies, 33 used planktonic bacteria solution for inoculation in bone tissue or implant, and ten used precultured bacteria on the implant, three of which showing biofilm formation on the implant or collagen sheet. ^33,43,56 In another two studies, infections were performed by injection of S. aureus into the femoral artery or a hole in the tibia. ^45,61

Evaluation of biofilm and bone infection

A total of 36 studies performed quantification of CFU load on tissue or implant. Biofilm was visualized by scanning electron microscopy (SEM) in 31 studies, confocal laser scanning microscopy (CLSM) in six studies, ^{30,41,43,45,54,56} and light microscope with crystal violet staining in another six studies. ^{29,39,42,50,54,55} Other imaging techniques included radioactive emission CT (ECT) for quantifying bacterial load through measuring the concentrated ^99mTc radioactivity, ³⁹ FDG-PET/CT imaging, ⁵⁷ and fluorescent microscope for monitoring bioluminescent-engineered bacteria strains. Radiological assessments to monitor bone lysis included radiograph analysis in ten studies, ^{26,27,31-33,39,51,53,56,64} micro-CT in 18, and MRI in one. ⁴⁶ Staining methods for bacteria identification included Giemsa stain ^36,54 and gram stain. ^{32,34,37,38,40,45,49,50} Commonly used histological assessments were haematoxylin and eosin (H&E) to observe inflammation in 21 studies, and tartrate-resistant acid phosphatase (TRAP) staining to evaluate osteoclastogenesis in three studies. ^34,37,40 Serum and local inflammatory markers include white blood cell count, ^38,50 CRP, ³⁸ amyloid, ²⁷ interleukin-10 (IL-10), tumour necrosis factor-α, IL-1α, and IL-1β. ^32,37,43 Local inflammation was also observed with positron emission tomography (PET) imaging and flow cytometry cell sorting. ^31,36

Complications

Complications were documented in five studies. ^{32,38,43,46,53} One study reported one dislocation and one loose implant among eight animals with hip PJI infection. ⁴⁶ Zhu et al ⁴³ documented that three of 60 rats with implant-associated tibia infection died from swelling and white purulent secretion at the wound. Hazer et al ⁵³ showed that four of 14 rats with IASI had pus formation localized in the fascia plane. Cahill et al ³² reported two mice with FRI dead within 24 hours post-surgery, with no cause specified. Lovati et al ³⁸ showed that two out of 32 animals with diabetic PJI demonstrated severe signs of infection, including joint abscess and fistulae.

Basic observational study findings

A total of 12 studies were observational without interventions. Two studies showed atypical implant-associated infection at a lower grade and lack of osteolysis caused by C. acnes and S. epidermidis RP62A. ^37,55 In the study by Nishitani et al, ¹¹ the natural history of biofilm maturation was revealed by SEM, including an initial attachment on day 1, robust proliferation on day 3, maturation with increased matrix on day 7, and dispersal after day 14 (Figure 2). The authors also showed that bacteria migration from biofilm was agr-dependent. Johansen et al ⁶¹ developed a haematogenous osteomyelitis model by injecting S. aureus into the femoral artery and found biofilm formation in the bone tissue.

Fig. 2.

A summary of the animal models, biofilm development, and therapeutic effects against Staphylococcus aureus (SA) and methicillin-resistant S. aureus (MRSA). CPS, calcium phosphate scaffold; FRI, fracture-related infection; Gen, gentamicin; IASI, implant-associated spinal infection model; OM, osteomyelitis; PJI, periprosthetic joint infection; PMMA, polymethyl methacrylate; Rif, rifampin; Van, vancomycin.

Prevention effects on biofilm formation

A total of 26 papers studied the preventive effects of various techniques on biofilm formation (Figure 2). Antibiotic treatment for durations of three, ³² five, ⁴² ten, ⁴² 14, ^29,30,35 21, ⁶⁶ 28, ⁴⁴ and 40 days ⁴⁹ were used in different studies for the prevention of biofilm formation. Accordingly, serum concentrations of vancomycin or tobramycin were measured in three studies, ^44,51,58 including one by liquid chromatography-mass spectrometry (LC-MS). ⁴⁴ Only one study quantified the concentration of gentamycin in bone tissue. ⁶² Biomaterials and antibiotic-loaded carriers were applied in 16 studies, of which only one study applied systemic antibiotic treatment as a control group. ³⁴

Tomizawa et al ³⁵ found that gentamycin and vancomycin could reduce the bacteria load of S. aureus when starting on day 0, which was enhanced in combination with rifampin. Cahill et al ³² showed that a higher dosage of topical rifampin reduced MRSA biofilm load, which was enhanced with systemic rifampin application. By loading amikacin and vancomycin on a chitosan sponge, one study found synergistic bactericidal effects against polymicrobial biofilm infections caused by S. aureus and E. coli. ²⁸ Jensen et al ⁵⁹ showed that a single dose of gentamycin, at least 1,600 × minimum inhibition concentration (MIC) value, was required to prevent S. aureus attachment to bone implants. van der Horst et al ⁴² showed that topical aminoglycoside and systemic ceftriaxone could eradicate S. aureus biofilm. Hovis et al ⁵⁸ showed successful infection prevention in all nine rabbits with implant-associated MRSA infection by the local spreading of vancomycin. One study showed more significant CFU reduction of S. aureus effects of topical tobramycin than doxycycline. ⁴⁴ Clarithromycin was found to enhance the bactericidal ability of ceftazidime in reducing the biofilm caused by P. aeruginosa. ⁵² Among the ten studies on the effects of systemic mono-antibiotic therapy, only two of six using vancomycin showed significant biofilm reduction, ^{30,32,35,49,51,66} and one of two studies using tobramycin showed significant biofilm reduction. ⁶⁶ None of these studies reported eradication. Two studies showed a significant reduction effect with ceftriaxone on implant, but one found no effects on CFU on bone. ^42,44 For the three studies on topical mono-antibiotic therapy, including vancomycin, tobramycin, and doxycycline, only one showed eradication, ⁵⁸ while the other two studies showed no significant reduction. ^42,44

Inzana et al ³³ showed that 3D-printed rifampin and vancomycin-laden calcium phosphate scaffold (CPS) reduced culture-positive rate to 50%. An electrospun composite coating composed of poly (lactic-co-glycolic acid) PLGA nanofibres and poly-є-caprolactone (PCL) was shown to deliver multiple antibiotics, including vancomycin, rifampin, linezolid, and daptomycin to eradicate MRSA biofilm in vivo. ²⁶ Stewart et al ⁶⁴ first showed the vancomycin-modified aminoethoxyethoxyacetate (AEEA)-AEEA-aminopropyltriethoxysilane (APTS)-Ti surface technique in a commercially available titanium plate, and found significant biofilm inhibition effects in a sheep model with FRI. Another study also showed that N, N-dodecyl, methyl-PEI coating prevented biofilm formation, and supported healing in a sheep FRI model. ⁶³ Li et al ³⁹ found that magnesium implant reduced biofilm formation through downregulation of the transcription levels of icaA, agr ribonucleic acid (RNA) III, and other virulence and antibiotics-associated genes. ⁵⁴ Min et al ⁴⁰ showed that sequential release of gentamicin and BMP-2 eradicated biofilm and promoted bone healing. Additionally, silver and selenium nanoparticles were shown to inhibit biofilm on pedicle screws and femur implants. ^41,53

For the specific pathogens, antibody MEDI3902 targeting P. aeruginosa PcrV and Psl exopolysaccharide were found to decrease biofilm by ten-fold load. ³¹ A novel antibiotic agent, human β-defensin, was shown to reduce MRSA biofilm in vivo by regulating inflammation and immune responses. ⁴³ Singh et al ⁶⁰ showed that negative-pressure wound therapy decreased CFU and biofilm formation. Fang et al ⁴⁹ showed that magnet nanoparticles-induced hyperthermia enhanced the biofilm eradication rate compared with systemic vancomycin treatment. Lindsay et al ²⁹ showed that reactive oxygen species (ROS) scavenger treatment improved oral antibiotic treatment on bacteria clearance.

Treatment effects on established biofilm

Six articles showed that treatment effects on established biofilms started at three to seven days after planktonic inoculation or revision surgery (Figure 2). ^{27,34,35,48,51,62} Notably, Tomizawa et al ³⁵ found that gentamycin reduced biofilm load on day 3, and none of the antibiotics groups showed any effects when starting on day 7. Inzana et al ³⁴ showed that vancomycin loaded on polymethyl methacrylate (PMMA) spacers only showed positive effects when combined with systemic vancomycin in a revision model for FRI. Similarly, Carli et al ²⁷ showed that vancomycin-loaded PMMA failed to decrease bacterial load. However, it prevented biofilm formation on the implant, reduced the inflammatory response, and preserved the tibial bone in a PJI revision model. An injectable composite of ceramic bone graft substitute loaded with gentamycin was ineffective in reducing S. aureus biofilm formation without extensive debridement and systemic antibiotic treatment in a pig osteomyelitis model. ⁶² Greimel et al ⁴⁸ tested different combinations of moxifloxacin, flucloxacillin, and rifampin, and found that moxifloxacin combined with rifampin was most effective in reducing CFU of S. aureus and biofilm in vivo. Wei et al ⁵¹ showed that intra-articular injection of vancomycin showed superior antibiofilm efficacy than systemic application in a PJI model caused by MRSA.

Discussion

Device-related infection is a major complication that occurs in orthopaedic surgery. The major obstacle in clinical treatment is the eradication of bacteria biofilm. To address this, research and development of new techniques for prevention and treatment, including implant surface modification, ^69,70 bone allografts, ⁷¹ controlled release of antibiotics, ⁷² and novel bactericidal agents have shown promising potential. ⁵³ In this review, we summarized the current preclinical evidence in animal studies, the advancement of recent therapeutic interventions, and the clinical translational value of the in vivo findings.

The principles of developing a biofilm-related bone infection animal model include the presence of a foreign implant, ⁶⁵ and biofilm that was cultured in advance. ⁴³ Large animals, including rabbits or sheep, are suggested to have better translational potential due to size, and a phylogenetically closer immune response to humans. ⁷³ Lovati et al ⁵⁰ identified dosage-dependent osteolysis and histological changes with increased bacterial load. In their study, low-grade infection and impaired healing caused by low-dosage CFU of S. epidermidis matched the observations in subclinical infections without prominent infection signs. Nishitani et al ¹¹ showed the natural history of biofilm attachment, proliferation, maturation, and dispersal. This explains the decrease of therapeutic effects when antibiotic treatment was started at a later period post-infection, and the high recurrence rate of implant retention. ¹¹ In another study by Windolf et al, ³⁶ local but not systemic elevation of leucocyte and IL-6 was observed with the formation of biofilm in the FRI model, where the unfavourable environment created by local inflammation and necrosis was found to be associated with poor healing. Among the 42 studies with implants, 15 used actual prosthesis or fracture devices customized to the animal size, including 3D-printed knee joint prostheses and spacers, ^27,68 cemented hip hemiarthroplasty implant, ⁴⁶ and commercially available plate fixation. ⁶⁴ Functional implants that provide better tolerance to surgery and mechanical stability enhance clinical relevance, because they enable functional outcome analyses like gait analysis for the PJI models and load-bearing in the FRI models.

Senneville et al ⁷⁴ showed a significantly higher rate of coagulase-negative strains from bone biopsy samples than from swab culture. Unlike the classic infection induced by S. lugdunensis, asymptomatic biofilm infection caused by C. acnes underscores the importance of long-term observation and the possibility of latent C. acnes infection in aseptic prosthetic loosening. ⁵⁵ Asymptomatic biofilm formation and low-grade inflammatory cytokines without prominent bone lysis were also observed in S. epidermidis infections. ³⁷ Meanwhile, a model of gram-negative PJI, which accounts for 3% to 6% of total cases, has been developed. ³¹ In the gram-negative (GN)-PJI model developed by Thompson et al, ³¹ the different infection rates among different strains revealed a chronic inflammatory response in P. aeruginosa infection compared to E. coli infection. These species-specific pathological patterns were suggested to cause different clinical outcomes. ⁵⁵ Given the high comorbidities in elderly patients, further investigation with osteoporotic or diabetic models is needed to elucidate disease-specific pathology of biofilm-related infections. ³⁸

Various imaging tools have been applied to observe biofilm ex vivo, but each has its advantages and limitations. Crystal violet staining is the most feasible method for biofilm observation and quantification, ^{29,39,42,50,54,55} but the morphological observation lacks detail. Other techniques include peptide nucleic acid fluorescence in situ hybridization (FISH) to detect biofilm formation in bone tissue, ^45,61,62 immunohistochemical staining, ^59,61,62 and fluorescent bacteria detection probe. ⁶⁵ Positively stained bacteria cells and biofilm can be visualized under the light microscope, but magnification is limited. The most common method with SEM can provide images with a high range of magnifications and complex shapes, ^11,75 but the potential limitations are a long sample preparation time and lack of vertical resolution. Another frequently used technique is the CLSM that can visualize 3D images at single-cell resolution, offer discrimination of bacteria and biofilm polysaccharide matrix, distinguish live or dead cells, and provide quantification by the integrated optical density. ^{41,43,45,56,64} However, the biofilm structure and properties may interfere with the fluorescence probe. Other imaging tools, such as radioactive CT or fluorescent microscope of bioluminescent bacterial strains, allow for non-invasive real-time monitoring of the infection development and inflammation in vivo, but are not specific to biofilm formation. ^11,30,46,57 For future studies, we recommend using SEM or CLSM, combined with CV staining and CFU counting, for better visualization and quantification.

The major difference between prevention and intervention is the decreased therapeutic effect of antibiotics against mature biofilms due to antibiotic tolerance and protection from the host’s immune response. When planktonic bacteria were inoculated, systemic or topical antibiotics, ⁴² coating techniques with antibiotic or nanoparticles, ^26,47,53 and hydrogel delivery of antibiotics ²⁸ showed significant biofilm reduction. Eradication of biofilm was observed when topical and systemic antibiotics were applied together. ^32,42,48 Therapies applied to prevent biofilm adhesion from overnight precultured bacteria also showed preventative effects ^33,35,39,66 as biofilm maturation often occurs three days after attachment. Despite numerous preventative methods proposed, only three studies had intervention on mature biofilms, ^35,48,62 and another three studies had intervention on established biofilm in revision surgeries. ^27,34,51 More often, established biofilm was found to be resistant to combined topical treatment with vancomycin and rifampin when the treatment was started after biofilm maturation. ³⁵ Greimel et al ⁴⁸ found that systemic moxifloxacin and rifampin treatment, starting on day 7 post-infection, eradicated the biofilm on implant but failed to clear the bacteria in the knee joint and bone tissue. In cases of revision surgery, Wei et al ⁵¹ showed that systemic and intra-articular injection of vancomycin eradicated the infection. Other studies found a reduction, but not eradication, of biofilm treated with vancomycin loaded on PMMA, ²⁷ even when combined with systemic application. ³⁴ We would encourage future investigations on the treatment effects of novel EPS dispersing agents in conjunction with antimicrobials as an alternative for current methods.

Systemic antibiotic treatment is the conventional treatment method for biofilm infections, albeit with limited effectiveness. ^30,35,42 For methicillin-sensitive S. aureus (MSSA) biofilm, first-line antibiotic cefazolin showed no effects of biofilm reduction even when it was started at day 0, and eradication was only achieved through a combination of topical aminoglycoside and systemic ceftriaxone. ⁴² For MRSA and MRSE infections, vancomycin remains the primary course of treatment. ^43,51 However, treatment failure by three or 40 days of systemic vancomycin therapy for biofilm reduction caused by S. aureus or MRSA was noted. ^32,49 Three studies suggested that the lack of sufficient local concentrations is the cause of the failure, while none measured tissue concentration. ^30,35,42 Furthermore, some strains even showed resistance to single vancomycin treatment in vivo at increased MIC to 4 ug/ml. ⁷⁶ Rifampin has been recognized as a cornerstone for treating biofilm in PJI, yet the monotherapy of rifampin is no longer recommended due to antibiotic tolerance. ⁴⁸ The effects of systemic rifampin have been tested in three studies. ^32,35,48 These studies suggested that systemic rifampin combined with high-dosage topical rifampin or moxifloxacin enhanced the effects against MRSA or MRSE. ^32,35,48 Quinolones including moxifloxacin can be a favourable option for combination with rifampin, due to the advantages of higher bioavailability and broad-spectrum bactericidal activity. ⁴⁸

Given that a sufficient local concentration of antibiotics is the prerequisite for successful bacteria eradication, topical application of antibiotics has been widely studied. ^{32,42,44,51,59} Topical spreading of vancomycin successfully prevented biofilm infection of the bone in all nine rabbits with implant-associated infection caused by MRSA. Similarly, topical tobramycin powder spreading showed more significant biofilm reduction compared to four weeks of systemic ceftriaxone, suggesting similar prophylactic effects of topical application compared to systemic antibiotics. ⁴⁴ However, current clinical evidence supporting antibiotic powder in orthopaedic trauma and infection is sparse. ⁷⁷ Increased MRSA biofilm reduction or eradication was achieved by intra-articular injection of vancomycin combined with systemic application, ⁵¹ or high-dosage topical rifampin combined with systemic rifampin. ³² The major advantage of this application method was a higher concentration at the target site. However, only one study in our review evaluated antibiotics concentration in bone tissue. ⁶² Importantly, vancomycin has nephrotoxicity, and rifampin has toxicity on osteoblasts and antagonistic effects with gentamicin, ^33,42,51 warranting more attention for application. Despite some evidence reporting positive results of topical antibiotic treatment, direct injection is flawed by its drug distribution and leakage problems, which warrants drug encapsulation and delivery with biomedical carriers. ⁴²

Biocompatibility, degradability, and sustained drug release are the fundamental characteristics of successful biomaterials targeting bone infection. ⁷⁸ As a clinically used carrier, PMMA loaded with vancomycin shows bacterial load reduction but cannot eradicate the biofilm. ^27,34 However, the primary concern is the elution from the depot to the surrounding tissue, which may fail to reach the optimal concentration. ³⁴ The non-degradable nature and low compatibility of rifampin with PMMA also limit the application of this combination. ⁷⁹ Alternative carriers with good biocompatibility and biodegradability have been studied. ^28,33,62 Inzana et al ³³ successfully incorporated rifampin with vancomycin-loaded calcium phosphate scaffold that is not feasible in PMMA. Compared to PMMA, their findings confirmed the advantage of co-delivery on CFU reduction on bone and implant, but the observation of persistent biofilm indicated limited effects of this combination. ³³ Chitosan loaded with amikacin and vancomycin showed complete clearance against polymicrobial infection caused by S. aureus and P. aeruginosa. ²⁸

Surface modification or coating with nanoparticles or hydrogels represents another direction of intervention. Ashbaugh et al ²⁶ developed a polymeric nanofibre coating with tunable combinatorial antibiotic delivery that prevented biofilm formation in vivo. Metal particles including Ag coating techniques have also been shown to reduce biofilm, but concerns include the durability of antibiotic activity and low cytotoxicity of Ag. ^47,53 Other proposed therapies including selenium nanoparticles or magnet nanoparticle-induced hyperthermia therapy show biofilm reduction effects at pilot stages. ^41,49 Notably, mesoporous silica nanoparticles loaded with enoxacin inhibited osteoclast activation, thereby mitigating bone loss and reducing biofilm formation. ⁶⁷ Biodegradable magnesium-based implants have shown promising potential in preventing biofilm attachment. ^39,54 The major advantage is that local degradation of Mg will reduce bacteria adhesion and biofilm formation, and stimulate bone formation. Interestingly, one study that applied a dual therapy by sequentially delivering gentamicin and BMP-2 showed biofilm eradication effects and promoted healing. ⁴⁰ These findings inspire novel therapies bridging the gap between degeneration and regeneration profiles of single treatments, highlighting the promising potential of a layered release strategy with antibiotic and bone-forming agents to achieve pathogen clearance and subsequent bone regeneration.

The strength of the current review is that we have presented existing knowledge gaps of research, novel findings regarding biofilm development within bone tissue, and the benefits and pitfalls of new treatment options. For this purpose, we had stringent inclusion and exclusion criteria to allow for an in-depth and accurate assessment of the treatment against biofilm. However, our study is limited by a lack of meta-analysis due to the large heterogeneity of bacteria species and strains, animal models, and treatment protocols. Another limitation is that many studies that did not explicitly evaluate biofilm were excluded. Additionally, newly identified intestinal MRSA carried by neutrophils and intracellular S. aureus in phagocytes are also suggested to be associated with PJI and recurrent infections on FRI models. ^80,81 Further studies are required to confirm if there is any interaction between these intracellular bacteria and biofilm pathology. With improved detection methods and growing use of permanent implants, similar challenges for biofilm eradication have been encountered in other devices including dental implants, catheters, and shunts. ^82,83 Treatment success rates for device-related biofilm infections range from 32% to 70%. ^84,85 Despite different pathologies of infections in different scenarios, surface modification techniques including anti-adhesive and antibacterial coating may serve as promising directions for future implant design in orthopaedics. However, these therapies require substantial modifications for orthopaedic applications, due to the disparities of the pathogens, wound environment, and the function of the implant.

In this review, we summarized the development of preclinical models on biofilm-related bone infections, in vivo characterization of biofilm, and advances in therapeutic interventions and outcomes. A comprehensive understanding of the matrix complex components, homeostasis, and molecular pathways in the local biofilm environment is the prerequisite for developing novel therapies. Given that the biofilm natural history and biological behaviour varies greatly between each pathogen and animal model, experimental designs should match the particular clinical scenario. Despite robust knowledge of the molecular pathways of biofilm formation in vitro, studies looking into the pathology of genetically modified strains are warranted to elucidate functional roles of the target genes in vivo. ⁴⁵

Current antibiotic treatment regimens show significant biofilm reduction effects when applied at early stages, and these effects can be further enhanced by a combination of rifampin with topical vancomycin that is already used in clinical application. ⁸⁶ However, only three studies that used systemic moxifloxacin plus topical rifampin, topical vancomycin powder spreading, ⁵⁸ or systemic plus topical vancomycin showed eradication. ⁴⁸ Given the difficulty of achieving eradication, surgical debridement is still the keystone in infection control. For clinical translation, data from animal studies will illustrate the mechanisms underlying persistent and recurrent infections. The findings from the current review revealed novel targets for future study and aid clinical decision-making on the optimal timing and combination of therapies. Emerging biodegradable and biocompatible biomaterials that act as drug carriers showed promising therapeutic potential when applied locally. Further validations of these novel treatments designated to individual clinical scenarios will allow for continuous eradication rates and safety improvement.

So far, there is still a major knowledge gap on the combination, dosage, delivery, and pharmacokinetics of current interventions. To address this, novel therapies should be tested in a valid animal model under a clear design for prevention or treatment for biofilm-related infections in orthopaedics. Evaluating the pharmacokinetics of antibiotics is necessary to determine the optimal dosage and concentration and avoid systemic toxicity. Measurements of antibiotic concentrations at tissue levels will justify the claimed drug release profile and further confirm the superiorities of topical application. The development of novel antimicrobial materials or agents is warranted to improve therapeutic effects and mitigate the tolerance of conventional antibiotics. For biofilm prevention, surface coating techniques of implants with nanoparticles or hydrogels showed the most promising potential, as the integration of antimicrobial properties with functional devices can facilitate recovery and improve clinical outcomes. We recommend a tailored release with EPS-dispersing, antimicrobial, and bone-forming agents loaded on biodegradable hydrogels for established biofilms. These therapies can be applied adjunctly with surgical debridement to promote subsequent healing and bone regeneration. With the ageing population, future directions, including the susceptibility, pathology, and treatments for biofilm-related bone infection with concurrent osteoporosis will provide therapeutic targets and improve outcomes in elderly patients.