Elution kinetics and antimicrobial efficacy of high-dose gram-negative antibiotic-loaded acrylic cement: an in vitro and in vivo study Authors: Jian Wei, Zhiyang He, Jiayue Xie, Kai Tong Published: 2026 ABSTRACT AbstractBackgroundHigh-dose antibiotic-loaded acrylic cement is routinely employed for the management of skeletal infections when implant fixation or local structural support is required. However, no established consensus or clinical guidelines currently exist regarding the selection of antibiotics in acrylic cement, particularly those targeting Gram-negative bacteria. This study aimed to systematically evaluate, through both in vitro and animal experiments, the efficacy of high-dose antibiotic-loaded acrylic cement in controlling periprosthetic skeletal infections caused by Gram-negative bacteria following orthopedic implant placement.MethodsHigh doses (4 g) of tobramycin, meropenem, piperacillin, ceftazidime, ciprofloxacin, and aztreonam were individually incorporated into acrylic cement (40 g). The elution profiles (n = 6 per group), antimicrobial efficacy (n = 5 per test), and compressive mechanical properties (n = 8 per group) were characterized in vitro over 28 days. The composite demonstrating the most favorable overall performance was selected for subsequent in vivo evaluation in a rat model of periprosthetic joint infection induced by Pseudomonas aeruginosa (P. aeruginosa) (n = 15 per group).ResultsThe incorporation of antibiotics resulted in a reduction in the mechanical strength of all tested cement composites compared with plain cement; however, all formulations except piperacillin-containing cement maintained compressive strengths above 70 MPa. After 28 days of elution, compressive strength increased in all groups except piperacillin‑cement; ciprofloxacin‑cement exhibited the highest post‑elution compressive strength among the antibiotic‑loaded cements (99.57 ± 2.64 MPa). Tobramycin-cement demonstrated the highest cumulative antibiotic release over the 28-day period (7.867 ± 0.295%). The other five antibiotics were released predominantly within the first 7 days, with peak release occurring on the first day, particularly for piperacillin. Through in vitro antibacterial assays, meropenem-loaded cement exhibited the strongest antimicrobial activity against the tested strains, while tobramycin- and ciprofloxacin-loaded cements also maintained antibacterial efficacy throughout the 28-day testing period. Meropenem was selected for in vivo testing based on its overall superior in vitro performance. In a rat model of periprosthetic joint infection induced by P. aeruginosa, animals treated with meropenem-loaded cement showed significantly lower bacterial loads in all periarticular tissues compared with antibiotic‑free cement (P < 0.01) and less severe bone destruction. The in vivo study compared meropenem‑cement only with antibiotic‑free cement, not with other antibiotic‑loaded formulations. Microbial cultures of the retrieved meropenem-cement specimens revealed no bacterial growth.ConclusionsBoth in vitro and in vivo findings indicate that meropenem-loaded acrylic cement offers distinct advantages in terms of mechanical properties, elution characteristics, and sustained antimicrobial activity against Gram-negative bacteria. Meropenem-cement may therefore represent a promising candidate for local antimicrobial delivery in the management of Gram-negative skeletal infections, warranting further clinical investigation. However, the potential selection pressure for carbapenem-resistant organisms must be carefully managed in any clinical translation of this approach.Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1186/s12866-026-05322-y.Keywords: Skeletal infection, Gram-negative bacteria, Mechanical properties, Elution, Antimicrobial activities Introduction Infections of the skeletal system—including periprosthetic joint infection (PJI), osteomyelitis, infected bone defects, and postoperative implant-related infections—represent common yet challenging clinical conditions in orthopedic practice. These infections not only cause significant impairment of joint function but also impose a considerable healthcare burden [1]. A major factor complicating treatment and leading to recurrence is the propensity of bacteria to form biofilms on implant surfaces. Within these biofilms, microorganisms are shielded from both antimicrobial agents and host immune defenses. As a result, bacteria in the biofilm state may exhibit a 10- to 1000-fold reduction in antimicrobial susceptibility compared to their planktonic counterparts [2]. Antibiotic-loaded acrylic bone cement represents a cornerstone in the management of implant‑related skeletal infections. Its clinical applications are extensive and well-established, encompassing the use of antibiotic-impregnated cement beads in chronic osteomyelitis, the filling of infected bone defects with antibiotic cement, and the deployment of antibiotic-loaded cement spacers during revision surgery for PJI. These formulations serve two critical functions: acting as a sustained-release drug delivery system to achieve high local antibiotic concentrations, and providing structural support as temporary or permanent spacers in bone defects. While Gram-positive bacteria (GPB), notably Staphylococcus aureus and Staphylococcus epidermidis, have traditionally been the predominant pathogens in skeletal infections, recent global epidemiological data indicate a substantial increase in Gram-negative bacterial (GNB) involvement over the past decade. Currently, GNB account for approximately 6% to 48% of all PJI cases [3–9], the eradication rate for GNB infections ranges merely from 27% to 77%, significantly lower than that observed for GPB infections, highlighting the particular therapeutic challenges posed by these organisms [10, 11]. In the management of GPB infections, vancomycin is widely regarded as the first-line antibiotic by the majority of clinicians and researchers [12, 13]. In contrast, no clear consensus exists regarding the selection or dosage of antibiotics to be incorporated into acrylic cement for the treatment of GNB infections, nor are there established guidelines or clinical consensus documents to inform practice. Previous in vitro studies have commonly employed antibiotic loadings of at least 4 g per 40 g of cement powder to achieve high local concentrations in experimental models [14, 15]. It is recognized that commercially available antibiotic-loaded cements may contain lower doses (e.g., 0.5–1 g per 40 g), but higher doses are often used in therapeutic spacers to ensure adequate elution [14, 15]. This study comprehensively evaluated four prevalent GNB strains—Escherichia coli (E. coli), Pseudomonas aeruginosa (P. aeruginosa), Klebsiella pneumoniae (K. pneumoniae), and Proteus vulgaris (P. vulgaris)—along with six antibiotics with recognized activity against Gram-negative pathogens: meropenem (MEM), ceftazidime (CAZ), piperacillin (PIP), ciprofloxacin (CIP), aztreonam (ATM), and tobramycin (TOB) [12, 16–19]. Through systematic assessment of mechanical properties, elution kinetics, and antimicrobial efficacy both in vitro and in vivo, this work aims to identify the most suitable antibiotic for incorporation into acrylic cement to manage Gram-negative skeletal infections. Unlike previous studies that evaluated fewer antibiotics or lacked in vivo validation, the present study provides a comprehensive, head-to-head comparison of six anti-Gram-negative antibiotics using identical experimental conditions and further substantiates the in vitro findings with a clinically relevant animal model of periprosthetic joint infection. Materials and methods Materials and animals Commercially available acrylic bone cement (VERSABOND®, Smith & Nephew, UK) was used throughout this study. Each package consisted of 40 g of radiopaque polymer powder and a 20 mL ampule of liquid monomer. Analytical-grade antibiotics, including MEM, CAZ, PIP, CIP, ATM, and TOB, were sourced from Aladdin (China). Reference-standard antibiotics were procured from the National Institutes for Food and Drug Control (China). Thirty specific pathogen-free (SPF) male Wistar rats were included in the present study. The animals, aged 10 weeks with a mean body weight of 254 ± 7 g, were purchased from the Center for Disease Control and Prevention (Hubei, China). Antibiotic-loaded acrylic cement specimen preparation Under atmospheric conditions, A high-dose (4 g) powdered antibiotic formulation was manually mixed with 40 g of cement powder (10 wt%) using manual mixing at a frequency of three stirs per second for 3 min. Subsequently, the liquid monomer was introduced into the antibiotic–cement powder mixture and manually stirred (three stirs per second for 60 s) under controlled ambient temperature (23 ± 1 °C) and relative humidity (50 ± 10%), in accordance with the manufacturer’s guidelines. Once the mixture reached the dough stage, it was compacted into stainless steel molds and allowed to polymerize for 60 min. This process yielded uniform cylindrical specimens in two distinct sizes: one measuring 12.00 ± 0.05 mm in height and 6.00 ± 0.05 mm in diameter, and the other 3.00 ± 0.05 mm in height and 6.00 ± 0.05 mm in diameter (refer to Supplementary Fig. 1A–D). All specimens were sterilized using ethylene oxide, and strict aseptic conditions were maintained throughout the pre-experimental handling phase. Mechanical properties Prior to testing, all specimens (n = 8 per group) were conditioned at 23 ± 1 °C for 24 h in accordance with ISO 5833 standards. Compressive strength was evaluated under quasi-static conditions using an Instron 5969 universal testing system (Canton, MA, USA). Cylindrical specimens (Ø 6 mm, H 12 mm) were tested at a constant crosshead displacement rate of 20 mm/min, corresponding to an approximate strain rate of 2.0 × 10³ s−¹, with the specimens positioned between rigid stainless steel compression platens (as illustrated in Fig. 1A). Compressive strength evaluation of antibiotic-impregnated acrylic cement. A Experimental setup for in vitro compression testing performed using an Instron 5969 universal testing system (Instron Corp., MA, USA), showing representative cement specimens before and after mechanical failure. B Quantitative comparison of compressive strength (mean ± SD) before and after 28-day elution in phosphate-buffered saline. Specimens include antibiotic-free control (CON) and cement loaded with tobramycin (TOB), meropenem (MEM), piperacillin (PIP), ceftazidime (CAZ), ciprofloxacin (CIP), or aztreonam (ATM). n = 8. Statistical significance was determined by one-way ANOVA with Tukey’s post-hoc test for multiple comparisons and paired t-test for within-group comparisons. ρP < 0.01 (Control group vs. all antibiotic-loaded groups, both before and after elution), **P < 0.01 (Comparison within each group: before elution vs. after elution), &P < 0.05 (Comparison between different antibiotic-loaded groups before elution; specifically, MEM and CIP vs. PIP or TOB), ##P < 0.01 (Comparison between different antibiotic-loaded groups after elution; e.g., CIP vs. other antibiotic-loaded groups) Antibiotic elution To evaluate the elution profile of antibiotics from the cement matrix, cylindrical specimens (12 mm in height × 6 mm in diameter; n = 6 per group) were individually immersed in 5 mL of phosphate-buffered saline (1× PBS) and incubated at 37 °C under constant agitation at 60 rpm [20]. The elution medium was completely replaced at predetermined intervals (days 1, 3, 5, 7, 14, 21, and 28). At each time point, specimens were carefully removed, rinsed gently with PBS, and transferred to fresh PBS to maintain sink conditions throughout the study. The eluate collected at each interval was stored in cryotubes at − 80 °C pending further analysis. The volume of 5 mL PBS was chosen to maintain approximate sink conditions based on the predicted maximum cumulative release of each antibiotic over the sampling interval and their aqueous solubilities. For all antibiotics tested, the concentration in the eluate at any time point was estimated to remain below 10% of their saturation solubility, thereby fulfilling sink condition requirements [15, 21–23]. The cumulative release over 28 days was subsequently determined. Quantification of antibiotic concentrations in all eluate samples was performed using high-performance liquid chromatography-mass spectrometry (HPLC–MS; Thermo TSQ Quantis, USA). HPLC-MS analysis was performed using a Thermo TSQ Quantis system equipped with an electrospray ionization source operating in positive ion mode. Chromatographic separation was achieved on a Hypersil GOLD column (100 × 2.1 mm, 3 μm) with a mobile phase consisting of water with 1% formic acid (A) and acetonitrile (B) at a flow rate of 0.2 mL/min. Total run time was 6 min. The specific SRM transitions and retention times were as follows: TOB: m/z 163.21 → 468.33, retention time 1.01 min; MEM: m/z 141.15 → 384.18, 3.62 min; PIP: m/z 143.13 → 518.26, 4.37 min; CAZ: m/z 468.15 → 547.22, 3.61 min; CIP: m/z 288.28 → 332.23, 3.87 min; ATM: m/z 96.03 → 434.14, 3.74 min. The lower limit of quantification was 0.1 µg/mL for all analytes. Calibration curves were constructed over the range of 0.1–100 µg/mL and exhibited excellent linearity (R² > 0.99). Intra- and inter-day precision (coefficients of variation) were < 10%. No internal standard was used. Assessment of antimicrobial activity in antibiotic-loaded cements The antimicrobial assays in this study were designed for comparative evaluation of different antibiotic-cement formulations under identical conditions, rather than for determination of clinical susceptibility breakpoints. These four strains (E. coli, P. aeruginosa, K. pneumoniae, and P. vulgaris) were selected because they represent the most clinically prevalent and challenging Gram-negative pathogens in orthopedic infections, as highlighted in the introduction [3–9]. The antimicrobial efficacy of antibiotic-impregnated acrylic cement specimens (6 mm in diameter × 3 mm in height; see Supplementary Fig. 1C-D) against GNB was evaluated using the Kirby-Bauer disk diffusion method [24–26]. The tested bacterial strains included E. coli (ATCC 25922), P. aeruginosa (ATCC 27853), K. pneumoniae (CMCC 46117), and P. vulgaris (CMCC 49027). All cement disks were sterilized with ethylene oxide following the ISO 11135:1994 standard for medical device sterilization. For the assay, Mueller-Hinton agar plates were uniformly inoculated with a bacterial suspension adjusted to a concentration of 10⁸ CFU/mL. A single sterilized antibiotic-cement disk was then aseptically placed onto the surface of each inoculated plate. Following incubation at 37 °C for 24 h, the diameter of the resulting inhibition zones was measured using a precision electronic vernier caliper with an accuracy of 0.01 mm. Bioactivity assessment of eluents via modified microdilution method The residual antimicrobial activity of antibiotics released from the cement specimens was quantified using a modified microdilution bioassay [27, 28]. In this procedure, 10 µL of each eluate sample was dispensed into individual wells of a 96-well culture plate, followed by the addition of 100 µL of a bacterial suspension standardized to 10⁵ CFU/mL. The plates were subsequently incubated at 37 °C for 24 h. Following incubation, bacterial growth in the test wells was assessed visually and compared against a drug-free growth control well. It is important to note that this procedure results in a 1:10 dilution of the eluted antibiotic concentration in the test well; therefore, the effective concentration to which bacteria were exposed was one-tenth of the concentration measured by HPLC-MS [27, 28]. Bioassay of eluents via modified disk diffusion method The antimicrobial activity of the eluted antibiotics was assessed using a modified agar disk diffusion assay [29]. Blank paper disks (PDM Diagnostic Disks, 6 mm diameter × 1 mm thickness, drug-free; AB Biodisk) were impregnated with 20 µL of each elution sample and aseptically placed onto the surface of Mueller-Hinton agar plates. These plates had previously been uniformly inoculated with 100 µL of a standardized bacterial suspension (10⁶ CFU/mL) of the test strains, including E. coli, P. aeruginosa, K. pneumoniae, and P. vulgaris. Following a 24-hour incubation period at 37 °C, the antibacterial activity was evaluated by measuring the diameter of the inhibition zones surrounding the disks with a precision electronic vernier caliper. Surgical procedure All animal procedures were approved by the Animal Care and Use Committee of Wuhan University (Approval No. AF339) and were performed in accordance with the ARRIVE guidelines. Following the induction of anesthesia with 2.5% inhalational isoflurane administered through a nose cone, rats received preoperative analgesia via subcutaneous injection of buprenorphine (0.1 mg/kg) and parecoxib (2 mg/kg). Upon achieving deep anesthesia, the right hind limb was shaved and aseptically prepared. A surgical approach was made to expose the knee joint, and a 1.3-mm bore was drilled into the intramedullary canal of the tibia. A prosthesis fabricated from orthopedic-grade titanium alloy (1.5 mm diameter × 5 mm length) was manually inserted in an anterograde manner, positioned such that the screw head protruded 1 mm into the joint space (Supplementary Fig. 2A). After capsular closure, the joint cavity was inoculated with 20 µL of a P. aeruginosa (ATCC 27853) suspension at a concentration of 1.5 × 10⁶ CFU/mL. Postoperative analgesia was maintained for 72 h using once‑daily subcutaneous buprenorphine (0.1 mg/kg; standard formulation, not extended‑release). This regimen was approved by the Institutional Animal Care and Use Committee (Approval No. AF339) at the time of study conduct. Parecoxib (2 mg/kg) was administered as a single preoperative dose and was not continued postoperatively. Establishment of infection was confirmed in all animals prior to revision surgery by the presence of clinical signs (abscess formation, see Supplementary Fig. 2C-D) and radiographic evidence of periprosthetic osteolysis (Supplementary Fig. 2B). Although quantitative cultures were not performed at the time of debridement to avoid compromising the surgical field, the consistent recovery of high CFU counts from the control group at the study endpoint (Fig. 7) indirectly validates the successful induction of infection in all animals included in the analysis. Radiographic examination at four weeks post-inoculation confirmed periprosthetic osteolysis and intra-articular abscess formation (Supplementary Fig. 2B-D). The original prosthesis was then explanted, followed by thorough debridement and implantation of a cement spacer (Supplementary Fig. 2E-F). Randomization was performed using a computer‑generated random number sequence (www.randomization.com). After debridement, rats were allocated to groups by an investigator not involved in subsequent outcome assessments. The order of surgeries alternated between groups to minimize time‑of‑day confounding, and cages were randomly positioned on the rack. No formal assessment of cage location effects was performed. The thirty animals were thus allocated into two treatment groups : (1) Control group, receiving antibiotic-free cement spacers (n = 15); and (2) MEM-Spacer group, receiving MEM-loaded cement spacers (10% w/w, n = 15). A total of 30 rats were used. Following revision surgery and spacer implantation, animals were randomly allocated for endpoint analyses: six rats per group (n = 6) were designated for micro computed tomographic (micro-CT) and histological evaluation; the remaining nine rats per group (n = 9) were used for microbiological analysis of bone, soft tissue, and explanted spacers. Microbiological Assessment Across Treatment Groups. A Quantitative culture results from bone specimens in the experimental groups. B Bacterial load analysis in peri-articular soft tissues from all treatment cohorts. C Microbiological evaluation of retrieved cement spacers from each study group. D Comprehensive bacterial burden assessment integrating all sampled biological materials; CON: antibiotic-free cement spacer; MEM: meropenem-loaded cement spacer. **P < 0.01 (Compared with control group). n = 9 per group Animal ethics and euthanasia All animal procedures were approved by the Animal Care and Use Committee of Wuhan University (Approval No. AF339, approved on May 6, 2020). The experimental work described in this study was conducted between May 2021 and June 2022. The approval remained valid throughout this period via annual renewal extensions in accordance with institutional policies. All procedures complied with the ARRIVE guidelines. At the end of the experimental period (six weeks post-revision surgery), animals were euthanized humanely for subsequent tissue collection and analysis. To minimize animal suffering, all rats were first anesthetized via intraperitoneal injection with a mixture of xylazine (10 mg/kg) and ketamine (100 mg/kg). Once a surgical plane of anesthesia was confirmed by the absence of pedal reflex, an overdose of sodium pentobarbital (150 mg/kg) was administered intraperitoneally to ensure euthanasia. Death was verified by the cessation of breathing and heartbeat. This method is consistent with the recommendations of the AVMA Guidelines for the Euthanasia of Animals (2020 edition) and was approved by our Institutional Animal Care and Use Committee (IACUC). The approval was valid for the duration of the study under the institutional policies, and all procedures were carried out within the approved timeframe. Radiographic evaluation Radiographic assessment was performed with a Bruker Xtreme BI imaging system (Germany). All images were acquired using a digital detector plate with the following standardized parameters: filter setting of 0.4 mm, tube voltage of 45 kVp, exposure time of 1.2 s, binning mode of 1 × 1 pixels, field of view of 10 cm, and an aperture set at f/2. Radiographic evaluation was performed by an investigator blinded to group allocation. Micro-CT imaging and data analysis At the six-week endpoint, following euthanasia, the right tibiae were harvested, fixed in 10% neutral buffered formalin, and subjected to ex vivo micro-CT scanning. Bone tissue responses, including resorptive changes and features of osteomyelitis, were evaluated by micro-computed tomography using a SkyScan 1276 system (Bruker, Germany). Scanning was conducted under the following parameters: X‑ray tube voltage 45 kV, current 200 µA, exposure time 794 ms per projection, total acquisition time 17 min, and an isotropic voxel size of 17.49 μm. The acquired projection data were reconstructed, and peri‑implant bone architecture was evaluated using CT‑viewer software (v2.0.4.6, Materialise, Belgium). Three‑dimensional visualization and analysis were subsequently performed with CT‑An software (Bruker, Germany). Following system calibration, a cubic volume of interest (VOI) measuring 370.282 mm³ (X: 9.55 mm; Y: 9.55 mm; Z: 4.06 mm) was defined within the tibial metaphyseal region. Bone volume (BV, mm³) inside the VOI was quantified as a measure of residual bone mass. Bone volume quantification was performed by an investigator blinded to group allocation. Microbiological analysis At the experimental endpoint (six weeks post-spacer implantation), periarticular soft tissues, bone samples, and the implanted cement spacers were aseptically harvested using sterile surgical instruments. Harvested tissues (bone and soft tissue) were placed in 8 mL of phosphate-buffered saline (PBS) and then mechanically homogenized using a high-throughput tissue grinder (JXFSTPRP, China) operated at 70 Hz for 10 min. A 100 µL aliquot of the resulting homogenate supernatant was then plated onto Mueller-Hinton agar and incubated at 37 °C for 24 h. Retrieved cement spacers were individually immersed in 2 mL of PBS containing 0.3% Tween 20 and subjected to ultrasonic agitation to dislodge adherent bacterial biofilms. Subsequently, 100 µL of the sonicated fluid was cultured following the same protocol used for tissue supernatants. Bacterial colonization was quantitatively assessed by enumerating viable colonies using standard plate counting methodology. All CFU counts were performed by an investigator blinded to group allocation. Histopathological evaluation Following a 56-day decalcification process in 0.5 M EDTA for bone specimens, all samples (bone and joint capsule) underwent sequential dehydration, were embedded in paraffin wax, and sectioned into 4 μm slices. Tissue sections were stained with hematoxylin and eosin (H&E) for morphological assessment. Imaging was performed using an H550S photomicrography system (Nikon, Japan), and all histopathological evaluations were conducted by an experienced observer who was blinded to the experimental group assignments. Statistical analysis An a priori sample size estimation was performed using G-Power software (v3.1.9.2, Germany). For the primary in vivo outcome (bacterial load), based on pilot data, an effect size (Cohen’s d) of 1.8 was assumed. With α = 0.05 and power (1 – β) = 0.80, a minimum of 6 animals per group was required for histological evaluation; 9 animals per group were allocated for microbiological CFU analysis to ensure robust statistical power. For in vitro mechanical testing, a minimum of 6 specimens per group was estimated to detect a 10% difference in compressive strength (α = 0.05, power = 0.80); 8 specimens per group were used. For elution and microbiological assays, group sizes of 5–6 were employed, consistent with previous studies [22, 27, 28]. All statistical computations were carried out in SPSS (v22.0, SPSS Inc., USA). The normality of data distribution was verified using the Shapiro-Wilk test, and homogeneity of variances was confirmed with Levene’s test. For normally distributed data, intergroup comparisons were performed using independent samples t‑tests or one‑way ANOVA with Tukey’s post‑hoc test; intragroup comparisons (pre‑ vs. post‑elution) were performed using paired t‑tests. Non-parametric alternatives, including the Mann-Whitney U test for between-group comparison, were applied to datasets violating normality assumptions (e.g., CFU data). A two-tailed P-value below 0.05 was considered statistically significant. Statistical analyses were performed by an independent analyst blinded to group identity. Results Mechanical properties The incorporation of antibiotics significantly altered the compressive strength of the cement compared to the antibiotic-free control (Fig. 1B, P < 0.01). Despite this reduction, prior to elution, all formulations satisfied the ISO 5833 minimum requirement of 70 MPa. Prior to elution, MEM- and CIP-loaded specimens demonstrated superior mechanical strength relative to other antibiotic-modified cements. One-way ANOVA with Tukey’s post-hoc test revealed that prior to elution, MEM- and CIP-cement had significantly higher compressive strength than PIP- and TOB-cement (P < 0.05), while no significant differences were observed among the other antibiotic-loaded groups. Following a 28-day elution period, compressive strength increased in all groups except the PIP-loaded cement. Among the eluted specimens, CIP-cement exhibited the highest retained strength. Post-elution analysis showed that CIP-cement was significantly stronger than all other antibiotic-loaded cements (P < 0.01), and MEM-cement retained significantly higher strength than PIP- and TOB-cement (P < 0.05). Elution characteristics TOB-loaded cement displayed the most favorable elution profile, with the highest cumulative antibiotic release over the 28-day period. The remaining five antibiotic cements showed a biphasic release pattern, characterized by an initial burst within the first 24 h—particularly pronounced for PIP—followed by sustained, gradual release throughout the observation period (Fig. 2A–H). The cumulative release percentages of the incorporated antibiotics were quantified as follows: TOB 7.867 ± 0.295%, MEM 1.729 ± 0.053%, PIP 2.748 ± 0.182%, CAZ 2.393 ± 0.101%, CIP 0.769 ± 0.030%, and ATM 1.975 ± 0.106%. To interpret the biological relevance of these elution concentrations, Table 1 summarizes the minimum inhibitory concentrations (MICs) of each antibiotic against the four test strains, compiled from CLSI guidelines and published literature [30–35]. These values represent the standard susceptibility breakpoints for planktonic bacteria. Cumulative release and elution kinetics of antibiotics from bone cement over 28 days. A-D Daily elution profiles of the six tested antibiotics from cement specimens. E–H Cumulative antibiotic release over the study period. Abbreviations: TOB, tobramycin; MEM, meropenem; PIP, piperacillin; CAZ, ceftazidime; CIP, ciprofloxacin; ATM, aztreonam. HPLC-MS analysis was performed as described in the Methods section. Data represent mean ± SD (n = 6). **P < 0.01 indicates statistically significant differences compared to all other antibiotic elution groups Table 1MICs of tested antibiotics against standard GNB strains (µg/mL)AntibioticE. coli ATCC 25,922 (µg/mL)P. aeruginosa ATCC 27,853 (µg/mL)K. pneumoniae CMCC 46,117 (µg/mL)P. vulgaris CMCC 49,027 (µg/mL)Tobramycin0.5-10.5-10.5-10.5-2Meropenem0.03–0.060.25-10.03–0.060.06–0.12Piperacillin2–44–84–82–4Ceftazidime0.12–0.251–20.12–0.250.12–0.5Ciprofloxacin0.008–0.030.25-10.03–0.120.03–0.12Aztreonam0.06–0.122–40.06–0.120.03–0.06MIC values are presented as ranges based on CLSI M100 (2023) and EUCAST breakpoints MICs of tested antibiotics against standard GNB strains (µg/mL) MIC values are presented as ranges based on CLSI M100 (2023) and EUCAST breakpoints Values represent minimum inhibitory concentration (MIC) ranges compiled from CLSI/EUCAST quality‑control reference data and published literature [36, 37]. These are MIC values, not clinical susceptibility breakpoints. Antibacterial properties of antibiotic-loaded cement All antibiotic-impregnated cement specimens maintained antibacterial activity following their incorporation into the acrylic matrix. In the 24-hour Kirby-Bauer disk diffusion assay, MEM-loaded cement exhibited the strongest inhibitory effect against all four Gram-negative strains. Against K. pneumoniae, the zones of inhibition produced by aztreonam-cement were not statistically different from those of MEM-cement (Fig. 3A–D). It should be noted that the size of the inhibition zone is influenced not only by antimicrobial potency but also by the differential diffusion rates of each antibiotic through the agar medium. Antibacterial Activity of Antibiotic-Loaded Cement Specimens Assessed by Kirby-Bauer Disk Diffusion Assay. A Inhibition zones of antibiotic-cement formulations against E. coli after 24-hour incubation. B Antibacterial efficacy against P. aeruginosa, represented by zone diameters. C Growth inhibition of K. pneumoniae by various antibiotic-cement composites. D Antimicrobial activity against P. vulgaris measured by clear zone formation. Abbreviations: TOB, tobramycin; MEM, meropenem; PIP, piperacillin; CAZ, ceftazidime; CIP, ciprofloxacin; ATM, aztreonam. Biological replicates: n = 5 per group. **P < 0.01 (Compared with the other tested antibiotic-cement group). N.S: No statistical significance Antibacterial durability assessed by microdilution bioassay In the modified microdilution bioassay, bacterial growth inhibition was indicated by clear and transparent wells. Against E. coli (Fig. 4A), eluates from MEM-, CAZ-, CIP-, and ATM-loaded cements completely inhibited bacterial growth for the entire 28-day period. Eluates from TOB-cement showed complete inhibition until day 5, after which slight turbidity appeared, indicating partial loss of activity. PIP-cement eluates inhibited growth only on day 1. For P. aeruginosa (Fig. 4B), MEM-, and CIP-cement eluates exhibited complete inhibition for 28 days. TOB-cement eluates showed inhibition up to day 5, with slight turbidity appearing on day 7. CAZ-cement eluates showed inhibition up to day 3, after which growth was observed. PIP-, and ATM-cement eluates lost activity after day 1. Against K. pneumoniae (Fig. 4C), MEM-, and ATM-cement eluates exhibited complete inhibition for 28 days. TOB-cement eluates showed inhibition up to day 3, with slight turbidity thereafter. PIP-cement eluates lost activity after day 1, CAZ-, and CIP-cement eluates exhibited complete inhibition for 21 days. In the case of P. vulgaris (Fig. 4D), Only MEM-cement eluates maintained complete inhibition for 28 days. TOB- and CIP-cement eluates showed inhibition up to day 5, followed by growth. CAZ-, and PIP-cement eluates lost activity within 3 days. Antibacterial Activity of Eluents from Antibiotic-Loaded Cement Over 28 Days. Broth microdilution assays were performed to evaluate the antimicrobial efficacy of eluents collected from antibiotic-impregnated acrylic cement specimens over a 28-day period. Bacterial growth is indicated by turbidity (cloudiness) in the culture wells. The panels illustrate the inhibitory activity of eluents against: A E. coli, B P. aeruginosa, C K. pneumoniae, D P. vulgaris. Abbreviations: TOB, tobramycin; MEM, meropenem; PIP, piperacillin; CAZ, ceftazidime; CIP, ciprofloxacin; ATM, aztreonam. Technical replicates: n = 5 Antibacterial durability of eluents in Kirby-Bauer assay Disk diffusion analysis of cement eluents revealed sustained antibacterial activity against E. coli over 28 days for MEM-, TOB-, CIP-, and ATM-loaded specimens, with MEM demonstrating the largest inhibition zones on day 1 (Fig. 5A–B). Against P. aeruginosa, eluents from MEM-, CIP-, and TOB-cement maintained inhibitory activity throughout the study period. MEM again showed the strongest initial antibacterial effect (Fig. 5C–D). For K. pneumoniae, sustained 28-day activity was observed with MEM-, CAZ-, TOB-, CIP-, and ATM-loaded cement. On the first day, both MEM and CIP eluents exhibited the most pronounced antibacterial effects (Fig. 5E–F). Against P. vulgaris, only MEM-, CIP-, and TOB-cement eluents maintained full activity over 28 days, with MEM showing superior efficacy at the 24-hour time point (Fig. 5G–H). Antibacterial activity of eluents from antibiotic-loaded bone cement specimens evaluated by agar disk diffusion assay over 28 days. A, B Inhibition zones of eluents against E. coli. C, D Inhibition zones of eluents against P. aeruginosa. E, F Inhibition zones of eluents against K. pneumoniae. G, H Inhibition zones of eluents against P. vulgaris. Abbreviations: TOB, tobramycin; MEM, meropenem; PIP, piperacillin; CAZ, ceftazidime; CIP, ciprofloxacin; ATM, aztreonam. All groups: n = 5. **P < 0.01 (Compared with other antibiotic elution group). N.S: No statistical significance Radiological evaluation Following six weeks of implantation, all cement spacers remained positioned within the tibial metaphysis. Radiographic analysis revealed varying degrees of periprosthetic bone resorption across all groups; however, these destructive changes were markedly more severe in the control group compared to the MEM-loaded cement group (Fig. 6A–F). Quantitative micro-CT analysis demonstrated a statistically significant increase in proximal tibial bone volume in the MEM group compared with the control group (mean difference = 20.69 mm³, 95% CI: 13.15–28.24 mm³; t10 = 6.11, P < 0.01, independent t-test; Fig. 6G, n = 6 per group). Radiological Assessment of Periprosthetic Bone in a Rat PJI Model following Debridement and Cement Spacer Implantation. A Representative radiographic image of the control group at 6 weeks after revision surgery. B, C Cross-sectional CT visualization and three-dimensional osseous reconstruction of the proximal tibia in a control animal at the 6-week endpoint. D Follow-up radiograph of the control group at 6 weeks post-revision. E, F Corresponding CT scan and rendered bone architecture of the proximal tibia in a rat treated with MEM-loaded cement at 6 weeks. G Quantitative analysis of residual bone volume in the proximal tibial region. CON: antibiotic-free cement spacer; MEM: meropenem-impregnated cement spacer. **P < 0.01 (Compared with control group). n = 6 per group Bacterial load quantification Bacterial culture results (Fig. 7A-D, n = 9 per group) showed that MEM-loaded cement significantly reduced bacterial counts in all sampled compartments compared to antibiotic-free cement. For bone tissue, the Hodges-Lehmann estimate of the median CFU reduction was 4540 CFU (95% CI, 2950–6620; Mann-Whitney U = 0.000, Z = -3.576, P < 0.01). For soft tissue, the median reduction was 5620 CFU (95% CI, 3430–7110; Mann-Whitney U = 0.000, Z = -3.576, P < 0.01). For retrieved cement spacers, the median reduction was 6200 CFU (95% CI, 5800–8300; Mann-Whitney U = 0.000, Z = -3.821, P < 0.01), with no bacterial growth detected on any MEM-loaded cement specimen. For total bacterial burden (bone + soft tissue + cement), the median reduction was 16,680 CFU (95% CI, 13830–21420; Mann-Whitney U = 0.000, Z = − 3.576, P < 0.01). Histopathological assessment Histological examination of knee joint sections revealed extensive inflammatory cell infiltration in both synovial and capsular tissues of control animals. In contrast, specimens from the MEM treatment group exhibited substantially attenuated inflammatory responses (Fig. 8, n = 6). Histopathological evaluation of periarticular tissues. Representative histopathological sections of the knee joint and synovial capsule from each experimental group are shown at 20× and 200× magnification. Scale bars represent 500 μm (for 20×) and 50 μm (for 200×). Tissue samples were compared between the control group (CON), which received antibiotic-free cement, and the treatment group (MEM) implanted with meropenem-loaded cement (n = 6 per group) Discussion In the present study, we systematically evaluated the mechanical, elution, and antimicrobial properties of six high‑dose antibiotic‑loaded acrylic cement formulations targeting Gram‑negative bacteria. Our key findings are threefold. First, all formulations except piperacillin‑cement met the ISO 5833 compressive strength standard both before and after 28 days of elution. Second, tobramycin-loaded cement exhibited the highest cumulative antibiotic release (7.867 ± 0.295%), whereas meropenem-loaded cement demonstrated the most potent and sustained antibacterial activity against all four tested Gram‑negative strains over 28 days, despite a more modest cumulative release (1.729 ± 0.053%). Third, in a rat model of P. aeruginosa periprosthetic joint infection, meropenem‑loaded cement significantly reduced bacterial loads in periarticular tissues, prevented bacterial colonization on the spacer, and attenuated bone destruction compared with antibiotic‑free cement. Skeletal system infection represents one of the most severe complications following orthopedic implant surgery, contributing to considerable patient morbidity and imposing substantial healthcare costs. The tenacity of these infections is largely attributable to biofilm formation on implant surfaces, rendering them notoriously refractory to systemic antibiotic therapy. GNB have emerged as prominent pathogens in such settings; they constitute the third most common causative agents in early PJI and account for nearly 30% of late-acute PJI cases [13, 38]. The clinical management of Gram-negative infections is particularly challenging due to the inherent virulence of these organisms, their escalating antimicrobial resistance profiles, and the frequent presence of patient comorbidities. Notably, GNB and GPB employ fundamentally distinct virulence mechanisms and host–microbial interactions, leading to markedly different infection phenotypes. Hsieh et al. reported that patients with Gram-negative PJI were generally older (mean age 68 vs. 59 years) and experienced earlier onset of infection following joint replacement (median prosthesis survival 74 days vs. 109 days) [39]. In their clinical series, P. aeruginosa was the most frequently isolated pathogen, followed by E. coli and K. pneumoniae. Multiple studies have associated Gram-negative PJI with higher treatment failure rates compared to Gram-positive infections [40]. The rising prevalence of drug-resistant Gram-negative strains has further heightened clinical concern. Martinez-Pastor et al. documented that among 132 patients with prosthetic joint infections following total knee arthroplasty, 7 cases (5.3%) were caused by extended-spectrum β-lactamase-producing Enterobacteriaceae, including E. coli and K. pneumoniae [41]. Because the animal experiment compared only MEM‑cement with antibiotic‑free cement, no in vivo ranking of the six antibiotic formulations can be inferred; the in vivo data confirm the efficacy of the selected formulation rather than its superiority over the other five candidates. Several investigators have characterized the physicochemical and antimicrobial properties of various antibiotic-loaded bone cement formulations targeting Gram-negative pathogens. Gasparini et al. evaluated the elution profiles of MEM, gentamicin (GM), and TOB when incorporated at 4 g per 40 g of cement. Their findings indicated that both MEM and TOB exhibited significantly greater cumulative drug release compared to GM [22]. While Gasparini et al. similarly reported superior elution of MEM and TOB compared to GM, the present study extends these findings by directly comparing a broader panel of anti-Gram-negative antibiotics and correlating their elution profiles with sustained antibacterial activity over 28 days. In a study by Gandomkarzadeh et al., the mechanical performance of cement containing 10 wt% CIP was assessed. The reported compressive and bending strengths fell below the ISO-specified thresholds (70 MPa and 50 MPa, respectively) at both initial testing and after four weeks [42]. In contrast to the findings of Gandomkarzadeh et al., who reported that 10 wt% CIP cement fell below ISO standards, our CIP-loaded cement maintained compressive strength > 70 MPa both before and after elution. This discrepancy might be attributed to differences in the specific cement brand (VERSABOND® vs. others) or mixing techniques used. Samuel et al. examined the release kinetics and antibacterial efficacy of cement loaded with 10 wt% MEM. Over a 34-day period, the cumulative antibiotic release accounted for 2.18% of the initial loading. The formulation completely inhibited growth of P. aeruginosa, E. coli, and K. pneumoniae for 7 days, and sustained antibacterial activity was observed for up to three weeks [43]. High-dose antibiotic-loaded bone cement represents a valuable option for constructing therapeutic spacers or beads following surgical debridement in skeletal infections. The present experimental findings demonstrate that the incorporation of high doses of any of the six tested anti-Gram-negative antibiotics into acrylic cement yielded compressive strengths exceeding 70 MPa. Notably, with the exception of PIP, all formulations exhibited a further increase in compressive strength after 28 days of elution. TOB-loaded cement displayed sustained release over the entire 28-day period, achieving the highest cumulative antibiotic release. In contrast, the other five antibiotic cements exhibited a biphasic elution profile, characterized by a substantial initial burst within the first 24 h, predominant release within the first week, and a subsequent phase of slow, minimal release. All six antibiotic-cement composites exhibited significant antibacterial activity on the first day, confirming the thermal stability of these antibiotics during cement polymerization. Among them, MEM-cement provided the most potent and durable inhibition against all four tested Gram-negative strains throughout the 28-day study. The superior and sustained antimicrobial performance of MEM-loaded cement observed in this study can be attributed to several factors. First, MEM, as a carbapenem, possesses exceptional stability against a wide range of β-lactamases, including extended-spectrum β-lactamases (ESBLs) often produced by the tested organisms. Second, it has high affinity for multiple penicillin-binding proteins (PBPs), leading to potent bactericidal activity. Third, its mechanism of action is time-dependent; the sustained, albeit low-level, elution from the cement may be sufficient to maintain drug concentrations above the MIC for a significant portion of the dosing interval, effectively suppressing regrowth of susceptible organisms [44–46]. This contrasts with concentration-dependent killers like TOB or CIP, which require high peak concentrations for optimal effect. Additionally, MEM may exhibit better thermal stability during PMMA polymerization compared to other β-lactams like PIP, which showed rapid loss of activity. In a rat model of post-arthroplasty skeletal infection, treatment with MEM-loaded cement resulted in significantly lower bacterial burdens in peri-prosthetic tissues—with no viable bacteria recovered from the cement itself—along with markedly reduced bone destruction and diminished inflammatory infiltration on histopathological evaluation compared to the control group. It is important to interpret the elution data in the context of each antibiotic’s potency. For instance, while the cumulative release of CIP was the lowest among all tested antibiotics (0.77%), its extremely low MIC against Gram-negative pathogens (typically ≤ 0.5 µg/mL) means that even these modest eluted concentrations were sufficient to maintain antibacterial activity for 28 days in the bioassays. This highlights that the clinical relevance of an elution profile is intrinsically linked to the antibiotic’s antimicrobial potency. The observed increase in compressive strength after 28 days of elution for most antibiotic-loaded groups may be attributed to continued polymerization (post-curing) of the PMMA matrix in an aqueous environment, or to water absorption leading to plasticization and filling of microvoids created by antibiotic leaching [47, 48]. From a clinical standpoint, this investigation provides essential evidence to inform the formulation of antibiotic-loaded acrylic cement for managing Gram-negative skeletal infections following orthopedic implantation. The findings indicate that a high elution volume does not invariably correlate with optimal antimicrobial efficacy, as antibacterial performance is influenced by multiple factors including antibiotic spectrum and drug half-life. The interpretation of elution kinetics in relation to antimicrobial activity must consider both the intrinsic potency of each antibiotic and the methodological dilution factor inherent in the bioassay. As shown in Table 1, the MICs of MEM against the tested Gram-negative strains are exceptionally low (e.g., 0.03–0.06 µg/mL against E. coli). Despite its relatively modest cumulative release (1.73% over 28 days), the concentrations of MEM eluted from the cement remained above these MICs throughout the study period. Even after accounting for the 1:10 dilution in the microdilution bioassay, the effective concentrations in the test wells (e.g., approximately 0.5 µg/mL at day 28 for E. coli) still exceeded the MIC, which explains its sustained antibacterial activity observed in Figs. 4 and 5. Conversely, PIP exhibited a higher cumulative release (2.75%), but its higher MICs (2–8 µg/mL against the tested strains) and rapid decline in elution after the initial burst resulted in an early loss of activity. This comparison underscores that the clinical relevance of an elution profile is not merely a function of the total quantity of drug released, but is intrinsically linked to the antibiotic’s antimicrobial potency and the sustained achievement of effective concentrations at the site of action [46]. Following their incorporation into bone cement, the majority of the six tested antibiotics sustained antimicrobial activity over 28 days. All antibiotic-cement composites met or exceeded the ISO minimum requirement for compressive strength, ensuring mechanical suitability for clinical use. Among the formulations evaluated, MEM-loaded cement demonstrated consistently superior antibacterial efficacy against all four Gram-negative pathogens—E. coli, P. aeruginosa, K. pneumoniae, and P. vulgaris—throughout the 28-day testing period, distinguishing it as the most promising candidate for further clinical evaluation. Nevertheless, this study has several limitations that should be considered. First, the investigation utilized a single acrylic bone cement formulation (VERSABOND), and different commercial brands with varying compositions could potentially yield different mechanical and elution characteristics. Second, while antibiotics such as GM and TOB are commonly employed in bone cement, this study focused exclusively on single-antibiotic formulations at a fixed dosage to establish a baseline understanding of each drug’s individual performance. The potential synergistic effects of dual or multi-antibiotic combinations within cement were not explored. However, the data presented here provide a critical foundation for future studies investigating combination therapies (e.g., MEM plus vancomycin or TOB) that may offer broader coverage and enhanced efficacy against mixed infections. Third, all cement samples in this study were hand-mixed under atmospheric conditions, despite the more frequent clinical use of vacuum mixing. This approach was intentionally selected because vacuum mixing has been shown to reduce cement porosity and consequently diminish antibiotic elution. Furthermore, hand-mixing remains more representative of standard practice in many primary care hospitals [49]. Fourth, the antibiotic powder was thoroughly blended with cement powder prior to monomer introduction, rather than employing the alternative technique of suspending antibiotic powder directly in the liquid monomer. It should be noted, however, that previous research indicates no significant difference in cumulative antibiotic release between these two mixing methods [50]. Fifth, while the in vivo rat model provides a holistic assessment including biofilm formation on the implant, our in vitro antibacterial assays (disk diffusion and microdilution) were performed on planktonic bacteria. It is well-established that bacteria in biofilms exhibit significantly reduced susceptibility to antibiotics. Therefore, the zone of inhibition and MIC results may overestimate the efficacy of these cements against sessile bacteria within a mature biofilm. Future studies should incorporate in vitro biofilm models, such as the Calgary Biofilm Device or flow cell systems, to directly assess the anti-biofilm activity of these antibiotic-loaded cements. Sixth, from a microbiological standpoint, the clinical application of MEM-loaded cement raises legitimate concerns about the potential for selecting carbapenem-resistant organisms. The prolonged release of low concentrations of antibiotics, particularly powerful agents like carbapenems, can create a selective pressure gradient within the joint space, potentially fostering the emergence of resistant mutants. Therefore, we strongly emphasize that this therapeutic approach should be reserved for confirmed or high-risk Gram-negative infections, used in conjunction with systemic antibiotics, and applied only after aggressive surgical debridement to minimize the bacterial burden. The benefit of eradicating an otherwise difficult-to-treat infection must be carefully weighed against the potential ecological downside, a consideration that warrants further investigation in long-term clinical studies. Seventh, while the MEM-loaded cement resulted in a statistically significant reduction in bacterial load compared to the control group (P < 0.01), the magnitude of this reduction was modest (approximately 0.8-log in some tissues) and did not achieve complete bacterial eradication in all samples. This underscores the challenge of treating established biofilm infections and suggests that in clinical practice, such local antibiotic delivery should be viewed as an adjunct to, rather than a replacement for, thorough surgical debridement and effective systemic antibiotic therapy. The primary benefit observed in this model may be the prevention of heavy recolonization and the mitigation of periprosthetic bone destruction, rather than absolute sterilization of the joint. Eighth, the postoperative analgesia protocol used once‑daily standard buprenorphine (0.1 mg/kg) without continued parecoxib. Current published recommendations for rat postoperative analgesia generally recommend buprenorphine administration every 6–12 h for up to 72 h. The once‑daily regimen was mandated by our institutional IACUC protocol in place at the time of study approval. Although all animals were closely monitored and none required rescue analgesia or early euthanasia, we acknowledge that more frequent dosing or use of an extended‑release buprenorphine formulation would have better aligned with current best practices. This should be addressed in future studies. Finally, the animal experiment was designed specifically to evaluate the in vivo antimicrobial performance of MEM-loaded cement. The study did not include other antibiotic-cement formulations or examine potential combination therapies with systemic antibiotics. Future investigations are planned to address these aspects. In summary, the present findings demonstrate that MEM-loaded cement exhibits superior sustained antimicrobial activity compared to all other tested antibiotic formulations, whereas TOB-cement achieved the highest cumulative antibiotic release over 28 days. It should be emphasized that local antibiotic delivery via bone cement alone may be insufficient to fully eradicate established infections following debridement; rather, it should be employed as part of a comprehensive therapeutic strategy that includes systemic antimicrobial therapy. The use of a carbapenem in local antibiotic delivery carries inherent antimicrobial stewardship risks that should not be underestimated. While the high local concentrations achieved may be effective, the prolonged sub‑inhibitory elution tail could create a selection gradient favoring carbapenem‑resistant subpopulations. Therefore, this strategy should be reserved for carefully selected cases with confirmed or high‑risk Gram‑negative infection, always combined with thorough surgical debridement and appropriate systemic antibiotic therapy. Further validation through well-designed clinical studies is warranted to confirm these experimental observations and establish evidence-based treatment protocols. The in vivo findings support the potential of MEM‑cement as an adjunctive local therapy, but do not demonstrate its superiority over other antibiotic‑loaded cements in vivo. Supplementary Information Supplementary Material 1: Supplementary Fig. 1 In Vitro Preparation of Antibiotic-Loaded Acrylic Cement Specimens. (A–B) Fabrication and dimensional verification of cylindrical specimens (12.00 ± 0.05 mm in height × 6.00 ± 0.05 mm in diameter) intended for compressive strength evaluation and elution testing. (C–D) Preparation and metrological characterization of disk-shaped specimens (3.00 ± 0.05 mm in height × 6.00 ± 0.05 mm in diameter) designed for antimicrobial activity assessment. Supplementary Fig. 2. Establishment of periprosthetic joint infection (PJI) and subsequent cement spacer implantation in a rat model. (A) Radiographic confirmation of prosthesis placement immediately following initial surgery. (B) Radiograph obtained 3 weeks post-implantation, showing periprosthetic osteolytic changes. (C‑D) Macroscopic views of the knee joint 3 weeks after surgery, revealing subcutaneous and intra-articular purulence with associated ulceration. (E‑F) Implantation of an antibiotic-loaded cement spacer following surgical debridement and prosthesis explanation. Acknowledgements We thank Mr Chen Wang from the microbiology laboratory of Wuhan University of China for test organisms: E. coli ATCC 25922, P. aeruginosa ATCC 27853, K. pneumoniae CMCC 46117 and P. vulgaris CMCC 49027. Authors’ contributions Jian Wei designed and performed the research, analyzed the data, prepared the manuscript and revised the paper. Zhiyang He, Jiayue Xie and Kai Tong performed the research and analyzed the data. Funding This work was supported by grants from the National Natural Science Foundation of China (No. 82260179), and Natural Science Foundation of Guangxi Zhuang Autonomous Region, China (No. 2023GXNSFAA026040). Data availability All relevant data analyzed during the current study are within the paper. Declarations Ethics approval and consent to participate The animal experiments were designed and conducted in compliance with the Animal Research Reporting In Vivo Experiments (ARRIVE) guidelines and were approved by the Institutional Animal Care and Use Committee (IACUC). All animal procedures were approved by the Animal Care and Use Committee of Wuhan University (Approval No. AF339) and were performed in accordance with the ARRIVE guidelines. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Supplementary Materials Supplementary Material 1: Supplementary Fig. 1 In Vitro Preparation of Antibiotic-Loaded Acrylic Cement Specimens. (A–B) Fabrication and dimensional verification of cylindrical specimens (12.00 ± 0.05 mm in height × 6.00 ± 0.05 mm in diameter) intended for compressive strength evaluation and elution testing. (C–D) Preparation and metrological characterization of disk-shaped specimens (3.00 ± 0.05 mm in height × 6.00 ± 0.05 mm in diameter) designed for antimicrobial activity assessment. Supplementary Fig. 2. Establishment of periprosthetic joint infection (PJI) and subsequent cement spacer implantation in a rat model. (A) Radiographic confirmation of prosthesis placement immediately following initial surgery. (B) Radiograph obtained 3 weeks post-implantation, showing periprosthetic osteolytic changes. (C‑D) Macroscopic views of the knee joint 3 weeks after surgery, revealing subcutaneous and intra-articular purulence with associated ulceration. (E‑F) Implantation of an antibiotic-loaded cement spacer following surgical debridement and prosthesis explanation. Data Availability Statement All relevant data analyzed during the current study are within the paper.