Long-term evaluation of plasma electrolytic oxidation (PEO) surface-modified WE43MEO magnesium alloy screws and plates reveals safe and controlled degradation in a 18-month minipig model

Abstract

Background

Magnesium-based implants are increasingly investigated as bioabsorbable materials for temporary osteosynthesis applications due to their favorable biocompatibility and bone-like elastic modulus. However, controlling the degradation rate remains a critical barrier to clinical translation, as rapid corrosion can compromise mechanical integrity and lead to adverse effects, such as gas formation.

Methods

PEO-modified WE43MEO (WE43-PEO) and non-modified WE43MEO screws and plates were implanted in the humerus and femur of Göttinger minipigs in a non-fracture model and assessed after 18 months. Explants were analyzed using micro-computed tomography and non-decalcified histology with histomorphometric quantification of residual implant structure and peri-implant bone response.

Results

No significant differences in cortical implant volume were observed. However, PEO modified implants were surrounded by significantly higher volumes of lamellar bone, suggesting reduced remodeling and improved bone integration. Non-modified WE43MEO implants underwent complete degradation, while PEO modified implants showed partial resorption with preserved structure, indicating effective degradation control. Both implant types remained integrated without long-term complications.

Conclusion

PEO surface modification of WE43 magnesium implants supports predictable, biocompatible long-term degradation and promotes favorable bone quality without late complications, underscoring the potential of surface-engineered magnesium fixation devices for load-bearing applications.

The translational potential of this article

PEO-modified WE43MEO osteosynthesis systems may offer clinically relevant, bioabsorbable fixation with controlled degradation and improved long-term bone integration, potentially reducing implant-related complications and the need for secondary removal procedures in orthopaedic and craniomaxillofacial surgery.

Graphical abstract

1. Introduction

The most common causes of facial fractures include accidental falls, physical assaults, sports injuries, and traffic accidents. The global age-standardized incidence rate of facial fractures is 98 per 100,000. In order to achieve anatomical reduction and functionally stable fixation, the standard surgical treatment involves open reduction and titanium-based miniplate osteosynthesis.

There are numerous advantages of titanium. However, complications can still arise from long-term fixation using durable materials, which frequently necessitate the removal of fixation devices in subsequent procedures, known as revision surgeries. These complications may include infections, palpable osteosynthesis material, structural failure, soft tissue complications, unfavorable sensations, interference with skeletal growth in children or adolescents and additional surgeries [1].

The reported implant removal rate varies widely, ranging from 3% up to 64.1%, with the majority of the studies identifying infection as the primary reason for implant removal [[2], [3], [4], [5]]. Nevertheless, patient desire has been demonstrated to play a significant role in such revision surgeries. There are reports in medical literature that a substantial proportion of plates are removed at the patient's request, with studies suggesting that over 80% of cases fall into this category [6,7].

To mitigate the risks and costs associated with a second surgery for metal removal, extensive research has been conducted on bioabsorbable osteosynthesis materials [8]. In the growing skeleton of children or adolescents, and in certain non-load-bearing indications, polymer-based plates, such as those made from polylactides, have shown satisfactory results. However, strict limitations in mechanical properties, including a Young's modulus lower than that of cortical bone and insufficient creep-resistance, have been observed [9]. Thereby, restricting their use in most load bearing indications.

To address these mechanical limitations, magnesium has emerged as the most promising candidate among other bioabsorbable metals, such as zinc and iron, for use in bioabsorbable medical implants. Magnesium has attracted the most attention as a preferred bioabsorbable metal in medical research, as it has exhibited notable biocompatibility and the potential to stimulate bone growth [[2], [3], [4], [5], [6], [7],10]. Recent research has increasingly focused on overcoming the key challenge of rapid and uncontrolled degradation, which limits clinical adoption. Efforts to enhance degradation resistance—especially through advanced surface engineering—have shown promising results in regulating absorption while preserving mechanical integrity and biological performance [[10], [11], [12]]. These studies highlight the importance of optimized surface modifications, including PEO and hybrid layers, in tailoring implant behavior to enable reliable long-term function. Furthermore, recent studies have indicated the suitability of bioabsorbable magnesium for use in load-bearing indications [13,14]. Nonetheless, magnesium degradation is still associated with several challenges, including an accelerated evolution of hydrogen gas. This gas accumulation has the potential to interfere with bone healing, particularly when aggregating in sufficiently high quantities [15]. Alloying and surface modifications are commonly used to control magnesium implant degradation and hydrogen gas release, often in combination [16].

Plasma-electrolytic oxidation (PEO), also known as Micro-Arc Oxidation (MAO), has become a state-of-the-art surface modification technique for medical grade magnesium by markedly reducing degradation rates [17] and exhibiting biocompatibility and applicability in a plethora of pre-clinical animal studies [18] conducted over the past decade. Recent advancements have broadened the spectrum of surface engineering strategies beyond PEO, including metal ion implantation (e.g., Fe, Ti, Zn, Zr) and biofunctional coatings such as calcium carbonate and self-healing layers, which have shown potential to improve degradation resistance, mechanical integrity, and biological performance of magnesium alloys, too [[13], [14], [15],19,20].

While numerous short- and mid-term preclinical studies have evaluated the performance of magnesium-based implants, long-term in vivo data in large animal models that closely resemble human physiology remains limited. Existing studies have reported degradation outcomes for magnesium alloys in goats and minipigs at 9, 12, and 18 months, demonstrating variable degradation kinetics depending on alloy composition, surface treatment, and anatomical placement.

While PEO is an established surface modification strategy for biomedical magnesium alloys, robust long-term in vivo data in large-animal models remain limited. In particular, late-stage degradation behavior, residual implant fate, and long-term bone remodeling of clinically relevant WE43 fixation systems have thus been insufficiently reported.

Hence, the objective of the present study is to investigate and compare WE43-based implants (plates and screws) with and without PEO-surface modification over a period of 1.5 years (18 months) using a large animal minipig model. This approach facilitates the evaluation of orthopaedic implants, particularly those intended for cranio-maxillofacial (CMF) applications, in their original dimensions. Additionally, it enables the assessment of long-term degradation, bone formation kinetics and the biocompatibility of the alloy and surface modification. This study presents the final, long-term evaluation (18 months) biological endpoints of WE43MEO magnesium based implants with and without PEO surface modification in a large animal model. It follows and complements our previously published short-term (6 months) and mid-term (12 months) analyses [18], providing a complete timeline of the in vivo degradation and biocompatibility profile of these implants.

The primary hypothesis is that PEO surface modification improves long-term osseointegration and supports controlled degradation of the implant, thereby promoting bone formation in the surrounding tissue. Additionally, we hypothesize that the facilitated WE43 magnesium alloy (WE43MEO) —whether PEO surface-modified or non-modified — undergoes substantial or complete bioabsorption within bone tissue without inducing adverse local reactions or systemic effects, thus confirming its local and systemic biocompatibility.

2. Materials and methods

2.1. Implants

Mg-Y-RE-Zr based biomedical magnesium alloy WE43MEO (Meotec GmbH, Aachen, Germany), consisting of 1.4–4.2 wt% Yttrium, 2.5–3.5 wt% Neodymium, less than 1 wt% of Aluminum, Iron, Copper, Nickel, Manganese, Zinc, Zirconium, and magnesium as balance was chill-casted and then processed viaoptional hot extrusion, heat treatment, and subtractive manufacturing to produce implant geometries frequently used in craniomaxillofacial surgery. Semi-finished alloy blanks were processed into two implant types: 4-hole plates (32.5 x 5.5 × 2.0 mm) via 5-axis CNC milling and bone screws (Ø 2.3 × 11 mm) via hot extrusion followed by SWISS type turning. The plates featured locking designs with threaded holes, allowing angle-stable fixation without predefining the screw-plate trajectory (Fig. 1).

Fig. 1.

WE43 implants without (left) and with PEO surface modification (right), used in humerus and femur surgeries. Screws: 11 × 2.3 mm; plates: 33 × 2 mm. PEO-modified implants exhibit a dull, grey surface due to the PEO surface modification. Image adapted from previous work [21].

2.2. Plasma-electrolytic oxidation surface modification

PEO treatment was performed in a phosphate-based aqueous electrolyte (Kermasorb®, Meotec GmbH), containing proprietary phosphate salts and minor organic or inorganic additives intended to stabilize the oxide layer and promote uniform conversion layer growth. Implants served as the anode and were immersed in a temperature-controlled stainless-steel electrolyte bath, which acted as the cathode. The electrolyte was maintained at 21 °C and continuously stirred to ensure homogenous ionic distribution and minimize local overheating.

Oxidation was conducted in a bipolar pulsed galvanostatic mode using a programmable power supply (M-PEO A1, Meotec GmbH) for 15 min. Alternating anodic and cathodic current pulses of 2 to 3 A were applied at a pulse frequency of 20 Hz. The voltage was dynamically adjusted between 0 and 500 V in response to the impedance of the developing oxide layer. The initial current density was set to 10 A/m², governed by the feedback-controlled rectifier system.

No chemical or mechanical pre-treatment was applied, other than ultrasonic cleaning in ethanol and distilled water. After treatment, implants were rinsed with deionized water and air-dried at ambient conditions.

Cross-sectional analysis using scanning electron microscopy (SEM) revealed oxide layer thicknesses of 9 ± 6 μm on plates and 18 ± 6 μm on screws. The PEO surface modification exhibited a three-layered structure: An outer porous layer, an intermediate transitional zone with lower porosity, and a dense inner barrier layer adhering to the metallic substrate. Surface morphology was characterized by uniformly distributed micropores and moderate surface roughness, typical for PEO. Energy-dispersive X-ray spectroscopy (EDS) confirmed the incorporation of phosphate species and magnesium oxide as principal chemical constituents of the surface modification layer. Visually, surface modified implants have a matte, light-grey appearance, in contrast to the metallic gloss of the non-modified controls (Fig. 2).

Fig. 2.

Scanning electron microscopy (SEM) images of WE43MEO screws without (WE43) and with PEO surface modification (WE43-PEO). Macroscopically (left), both appear smooth, while higher modification (right) reveals the characteristic microscopic roughness of a PEO-modified surface (adapted from previous work [21]).

2.3. Study design

All animal experiments were performed in accordance with the German Federal Animal Welfare Act, the EU Directive 2010/63/EU on the protection of animals used for scientific purposes, and the guidelines for the care and use of laboratory animals [17]. Ethical approval was granted by the Ministry for Gesundheit und Verbraucherschutz Hamburg (Approval No. 49/16). The study included five skeletally mature miniature pigs (three females and two castrated males) with an average body weight of 44.88 kg (SD ± 8.83). To minimize hormonal influences on bone metabolism and implant degradation, only female and castrated male animals were selected, ensuring a more consistent biological response across the cohort. Implantation sites were randomized between the humerus and femur. Three animals received plates in a single location and two animals received plates in both anatomical regions. Each plate was affixed using four screws with identical surface characteristics. The primary aim was to assess overall implant integration and long-term stability of individual screws and plates.

Prior to surgery, all animals were evaluated by an experienced veterinarian to confirm health status. Upon arrival, the animals were allowed to acclimate to the housing environment. Throughout the 18-month observation period, they were monitored daily for signs of pain or behavioural changes. This study adheres to the ARRIVE guidelines [22] for the ethical design and transparent reporting of animal research.

2.4. Surgical procedure

As previously described, implants were placed in the humerus and/or femur of the minipigs [21] under standardized perioperative conditions. Animals received a single dose of antibiotic prophylaxis (Amoxicillin/clavulanic acid (8.75 mg/kg body weight; Synulox, Zoetis, Berlin, Germany). Surgical procedures were performed under inhalation anaesthesia (40% oxygen and 1.7–2% sevoflurane; Dräger, Lübeck, Germany) and strict aseptic conditions. A skin incision was made over the target bone, followed by careful mobilization of the muscle and periosteum to expose the diaphyseal region. Screw insertion was performed manually using consistent tactile feedback to ensure uniform torque and reproduce primary stability. The periosteum and surrounding soft tissues were meticulously repositioned and wounds were closed in two layers: subcutaneous tissue followed by skin closure. Postoperative care was provided according to established protocols [16], including oral antibiotic administration for six days post-surgery. Animals were monitored daily for signs of infection or discomfort. At the biological study endpoint (18 months), animals were euthanized in accordance with ethical guidelines, followed by bone and organ harvesting. Explanted specimens were fixed to preserve tissue architecture for subsequent analysis. Prior to histological processing, bone integration and implant stability were assessed via CT. Histological evaluation included microscopy and staining techniques to assess bone remodeling and local tissue response.

2.5. Histological preparation, imaging and analysis

To ensure proper fixation and dehydration, the epiphyses of the harvested bones were removed. Samples were fixed in 10% paraformaldehyde for five days to preserve tissue integrity, followed by thorough washing and dehydration through a graded alcohol series. The bones were immersed in xylene to facilitate infiltration with the embedding material. Finally, the samples were embedded in Technovit (Technovit 9100 New, Heraeus Kulzer, Hanau, Germany) to ensure optimal sectioning and subsequent histological evaluation.

For optimal visualization of the screw axis, embedded blocks were sectioned using precision-grinding. Re-embedding and further grinding allowed preparation of non-decalcified slices. Giemsa staining was performed to differentiate mineralized and non-mineralized tissue. A slice containing a part of non-degraded screw, served as a template to define the regions of interest (ROIs) for histological evaluation. Digitalization was carried out using an Axio Cam MRc5 and AxioVision software (Carl Zeiss Mikroskopie, Jena, Germany), generating mosaic images with manual correction and stitching as needed.

The primary focus of analysis was on implant degradation and osseointegration in the cortical region, where the screw interfaces with bone. Due to variability and degradation-related artifacts, the screw head regions were excluded from analysis.

To establish reference values, the opposite cortical bone (contralateral side) was used to measure original cortical thickness (see Section 2.6 Micro-CT). A non-implanted reference screw was also evaluated for comparison. A custom ROI mask was applied to each section to quantify mineralized bone and structural composition. ROIs were subdivided into cortical and marrow compartments and extended by 0.5 mm and 2 mm beyond the implant to assess peri-implant bone response.

Quantification was performed in ImageJ (ImageJ for java 8, version 1.52f, U.S. National Institutes of Health, Bethesda, Maryland, USA) using custom macro for histomorphometric analysis. Threshold-based masks, were applied to each tissue type, incorporating manual adjustments and corrections. The void area, mineralized bone (MdB), connective tissue, corrosion material and bulk material were then quantified automatically. A separate ROI was created to encompass the area above the plates to evaluate the newly formed bone. Due to the advanced corrosion states of screws and plates at 18 months Bone-Implant-Contact (BIC %) could not be reliably determined (Fig. 3, Fig. 4). Instead, osseointegration was assessed using lamellar bone quantification in the peri-implant region and qualitative evaluation of bone infiltration into degrading screw structures. To address this, an additional ROI extending 1.5 mm beyond the implant was created to assess lamellar bone formation. Lamellar bone was identified manually and marked accordingly (Supplementary Fig. 1).

Fig. 3.

Representative micro-CT slice of a non-modified WE43MEO screw 18 months post-implantation (sample 18M-1.1-PLU-Sc2), illustrating the influence of grayscale thresholding on segmentation. Left: Native grayscale image without thresholding. Middle: Threshold-based segmentation using a lower cutoff (≥ 513 mgHA/cm³), capturing both residual metallic implants, corrosion products, and mineralized bone. Right: Segmentation using a higher threshold (≥ 1296 mgHA/cm³), isolating only dense, non-degraded metallic material.

Fig. 4.

Non-decalcified histological slices compared to the corresponding micro-CT images of non-surface-modified (WE43) and surface-modified (WE43-PEO) screws. Progressive degradation is indicated by increased radiolucency of the screw. Scale bar: 1 mm. Red circle: bulk material. Yellow circle: Encased magnesium fragment. Orange arrow: Newly formed bone.

2.6. Micro computed tomography (Micro-CT)

Micro-computed tomography (micro-CT) was conducted using the VivaCT40 system (Scanco Medical AG, Brüttisellen, Switzerland) with an isotropic voxel size of 19 μm, operated at 70 kVp and 114 μA for radiological assessment. The reconstructed volumes were calibrated to mgHA/cm³ using ImageJ (Fiji), based on reference phantoms. Threshold values were defined according to a previous study [22] with 513 mgHA/cm³ used to differentiate mineralized bone from air, and 1296 mgHA/cm³ to distinguish magnesium from bone.

To define the Volume of Interest (VOI), the thickness of the cortical bone opposite of the implant site was measured, yielding an average cortical thickness of 3.24 mm. A non-resorbed reference screw was scanned, and its geometry extracted to generate a standardized VOI template for screw volume quantification. Each plate-associated scan was segmented into four regions to align with individual screw positions, allowing targeted volumetric analysis. Image segmentation and VOI placement were performed using a custom macro developed for Fiji/ImageJ.

Quantification was carried out separately for the cortical and medullary compartments, following histological stratification. To assess peri-implant bone volume, the screw VOI was radially expanded by 0.5 mm and 2.0 mm. Screw volume (SV) was quantified using the upper threshold (≥ 1,296 mgHA/cm³) and expressed as a percentage of the total volume (TV) of the VOI [SV/TV %]. Bone volume (BV) was determined using a grayscale window between 513 and 1,296 mgHA/cm³, effectively excluding low-density structures such as air and high-density remnants such as magnesium [BV/TV %].

It is important to note that the spatial resolution and density contrast of the applied micro-CT protocol limit the ability to differentiate between partially degraded magnesium, low-density corrosion products, and newly formed bone—particularly in the medullary region. As magnesium degrades, it converts into radiolucent by-products whose attenuation values often overlap with that of mineralized tissue, leading to potential misclassification or underestimation of residual material.

To visualize these segmentation challenges, representative CT cross-sections were processed in Fiji using threshold-based segmentation. Segmentation using Threshold 1 (≥ 513 mgHA/cm³) captured both mineralized bone and remaining implant structure, whereas Threshold 2 (≥ 1,296 mgHA/cm³) isolated the dense, metallic core. The intermediate grayscale range (513–1,296 mgHA/cm³) represents a diagnostic grey zone where degradation by-products and osseous tissue cannot be reliably distinguished, underscoring the necessity for complementary histological validation (Fig. 3).

The grayscale range between 513 and 1,296 mgHA/cm³ represents an ambiguity zone in which low-density degradation by-products cannot be reliably distinguished from newly formed bone, emphasizing the need for complementary histological analysis.

2.7. Statistics

All data was exported as TXT files and pre-processed in Microsoft Excel (Version 16.21; Microsoft Corporation, Redmond, WA). Statistical analyses were performed using Prism version 8.1.1 (GraphPad Software, San Diego, CA). Normality distribution of data was assessed using the Shapiro–Wilk test. Based on distribution characteristics, either the Mann–Whitney test or Welch's t-test was applied. A p-value of < 0.05 was considered statistically significant. The study design incorporated both paired and unpaired comparisons resulting in a nested data structure that may not be fully accounted for by the chosen statistical methods. While each parameter was measured once per sample (without technical replicates), this limitation was considered during interpretation of statistical outcomes. Given the limited number of animals, mixed-effects modeling or repeated-measures approaches were not applied, as such models would be statistically unstable and underpowered; this limitation was therefore addressed by a conservative interpretation of the results.

3. Results

3.1. Quantification of screw volume and screw area

Representative histological and micro-CT images of WE43 and WE43-PEO screws and plates are shown inFig. 4, Fig. 5. A comparative assessment of non-decalcified histological sections and their corresponding micro-CT scans reveal that degradation was primarily localized in the bone marrow region, evidenced by increased radiolucency and the accumulation of degradation products. In contrast, screw regions embedded in cortical bone largely retained their structural integrity, with no significant difference in material retention between non-modified (WE43) and PEO-modified (WE43-PEO) screws.

Fig. 5.

Micro-CT assessment in different cutting planes of a non-surface-modified screw after 18 months of implantation, showing the cross-section (above) and top view (below). The top view reveals loosening of the bony structure beneath the plate. During degradation, flake-shaped fragments detach, particularly in the bone marrow region. Histological staining in Giemsa. Scale bar: 2 mm.

Screw degradation was assessed using both micro-CT (volume analysis: SV/TV [%]) and histology (area analysis: Sc/Tar [%]), with separate evaluation of cortical and marrow compartments. In the cortical region, neither the SV/TV [%] nor Sc/Tar [%] differed significantly between groups (p > 0.05). However, in the bone marrow region, histological analysis revealed a significantly larger screw area for the WE43-PEO group (Sc/Tar [%]: 51.83 for WE43 vs. 91.58 for WE43-PEO; p = 0.0358, Mann–Whitney test). A similar trend was observed in volumetric micro-CT analysis (SV/TV [%]: 26.90 for WE43 vs. 35.07 for WE43-PEO), though the difference did not reach statistical significance (p = 0.3626, Welch's t-test). The discrepancy between imaging and histology likely reflects the limitations of micro-CT as method to distinguish between partially degraded implant material and newly formed bone due to overlapping grayscale values. Consequently, micro-CT predominantly captured the residual bulk metal, while histology provided higher sensitivity for identifying both intact and degrading implant regions (Fig. 6).

Fig. 6.

A: Results of screw and bone quantification by micro-CT and histology, showing no significant differences between WE43 and WE43-PEO screws, and thus indicating a comparable performance and bone response after 18 months of implantation. C = cortical, B = bone, M = bone marrow, Sc = Screw, SV = screw volume, TV = total volume, BC = MdB = mineralized bone, Tar = total area, SC = cortical screw segment, SM = intramedullary screw segment, BC = cortical bone interface, GM = intramedullary bone interface. Sc/Tar SM: p = 0.0358, Mann–Whitney test.

4.2. Radiological and histological bone quantification

Bone volume (BV) and medullary bone (MdB) were assessed within two defined VOIs/ROIs surrounding the screw, based on a screw template expanded by either 0.5 mm (BV/TV₀.₅ [%] and MdB/TAr₀.₅ [%]) or 2 mm (BV/TV₂.₀ [%] and MdB/TAr₂.₀ [%]) (Supplementary Fig. 2). No significant differences were observed between the two screw types in terms of bone volume or mineralized bone fraction in either 3D (BV/TV) or 2D (MdB/TAr) analyses (p > 0.05) (Fig. 6).

However, differences in degradation behavior became evident over time. PEO-modified implants exhibited delayed degradation after 18 months, as indicated by a lower number of active degradation sites and reduced remodeling of implant fragments into bone. These effects were more pronounced in the bone marrow region, where increased statistical variability in degradation within the thread region was observed in Sc/TAr [%] box plots. Histological analysis further confirmed a significantly higher residual area for PEO-modified screws in the bone marrow compared to non-modified screws (Mann–Whitney test, p = 0.0358), with median residual material of 91.58% versus 51.38%. While radiological and histological findings were consistent in the cortical region, only histological analysis revealed a statistically significant difference in degradation within the bone marrow, reinforcing the protective effect of PEO modification.

3.3. Lamellar bone quantification, bone overgrowth and degradation of plates

Quantification revealed a significantly higher percentage of lamellar bone per area surrounding the PEO-modified implants compared to the non-surface-modified implants (p < 0.05). In contrast, the amount of bone formed above the non-surface-modified plates was significantly greater (p < 0.005). To assess plate corrosion, both the bulk material and the corrosion material of the plates were quantified using histological analysis. No significant differences were found in the quantification of either the bulk material [mm²] or the degraded sections between the two groups (p > 0.05).

The screws appeared well integrated into surrounding bone tissue based on qualitative histological and micro-CT observations. This observation has been systematically documented in histological and micro-CT images (Fig. 7).

Fig. 7.

Results of plate degradation, lamellar bone, and newly formed bone above the plates. A significantly higher proportion of lamellar bone surrounding surface-modified implants (p = 0.0310 Welch's t-test) and a significantly greater volume of newly formed bone is observed above the non-surface-modified plates (p = 0.0124 Welch's t-test).

Qualitative histological observations indicated substantial bone infiltration within degrading screw structures, extending circumferentially in several specimens. Various stages of the bioabsorption process could be observed in both, the plates and screws (Fig. 4, Fig. 9).

Fig. 9.

Exemplary degradation and bone formation of a magnesium plate, highlighting distinct degradation stages: optically opaque (1), light pink (2) and blue (3). Bone invasion into the plates (black arrow) is minimal due to the substantial presence of bulk material. Cracks observed in the corrosion material are associated with the embedding process (white arrow). Implant: WE43, staining in Giemsa, scale bar in the overview: 1 mm.

4. Discussion

The novelty of the present study lies in providing rare long-term comparative in vivo evidence for a clinically relevant magnesium fixation system. Most previous studies focused on early or mid-term outcomes, whereas the present 18-month biological endpoint enables the assessment of late to full degradation stages, residual material behavior, and long-term tissue integration. Various studies have demonstrated that PEO treatment effectively delays the degradation of magnesium-based implants, reduces adverse effects such as hydrogen gas release, and influences both, implant remodeling and bone regeneration, by modulating bone resorption and formation processes in vitro and in vivo [17,[21], [22], [23], [24]]. The present data indicates that these positive effects become less pronounced at longer biological endpoints, such as after 18 months, and likely diminish further over extended time periods. No statistically significant differences in implant degradation between WE43 and WE43PEO screws could be detected in micro-CT volumes after 18 months. In advanced stages of implant degradation, imaging modalities such as micro-computed tomography (micro-CT) face inherent limitations. As magnesium degrades, its radiodensity increasingly approximates that of newly formed mineralized bone, making it difficult to differentiate between residual implant material and osseous tissue. This convergence in grayscale values can lead to an underestimation of the true extent of degradation. Moreover, highly fragmented implant remnants are challenging to detect and quantify both radiologically and histologically due to their small size and irregular distribution. As such, histological evaluation remains essential for accurately characterizing the bioabsorption process of magnesium and identifying the residual structure of degraded implants. The combined use of micro-CT and histological methods is therefore critical to overcome modality-specific constraints and to obtain a comprehensive understanding of the long-term degradation behavior of biomedical magnesium alloys and devices.

From a translational perspective, long-term safety and predictability of degradation are critical prerequisites for clinical adoption of bioabsorbable magnesium-based osteosynthesis systems. The present 18-month large-animal data therefore provides clinically relevant evidence that PEO surface modification enables controlled degradation without late adverse tissue reactions, addressing a key barrier to translation into orthopaedic and craniomaxillofacial practice.

In contrast, histopathological analysis provided a more detailed differentiation between actively degrading and preserved regions of the implants (screws). Once complemented by histological assessment the comparison between the two groups revealed statistically significant differences in the amount of remaining structure of the different screws in the bone marrow region (Fig. 6: SV/TV vs. Sc/TAr [%]). The present findings further support the potential advantages of WE43-PEO under long-term in vivo conditions, being particularly beneficial in surgical applications, where controlled degradation during the initial stages is imperative, yet prolonged degradation timelines are unfavorable, e.g. in orthopedics or CMF surgery. From a surgical perspective, bioabsorbable implants should ideally provide sufficient mechanical support and fully degrade after fracture healing.

Specifically, at 6 months, PEO-modified screws exhibited a significantly reduced degradation rate compared to non-surface-modified screws, along with improved bone formation and a higher quality of surrounding bone tissue. At 12 months, the degradation of both PEO and non-PEO screws progressed further, but the PEO group maintained a better preservation of the implant structure, lower degradation volume, and a greater proportion of lamellar bone in histological sections.

In the current 18-month analysis, these trends largely persisted, albeit with reduced effect sizes. Both PEO and non-PEO modified implants underwent extensive degradation, but no significant differences were observed in residual material volume within the cortical bone on either histological or radiological evaluation. However, PEO-modified implants exhibited significantly greater retention of structure in the bone marrow region (Sc/TAr in SM: 91.58% vs. 51.83%, p = 0.0358) and a significantly higher percentage of lamellar bone surrounding the implants, suggesting sustained osteoconductive benefits beyond initial stabilization phases.

These findings support the notion that PEO surface modification primarily delays the onset of rapid degradation and hydrogen gas evolution in early implantation phases (0–6 months), while preserving osseointegration and lamellar remodeling over long term. Importantly, the lack of significant late-stage complications at 18 months, such as gas cavities, inflammation, or implant loosening, reinforces the biosafety and mechanical compatibility of the presented WE43-PEO system in vivo.

Despite the absence of statistically significant differences in degradation between the two groups, as evidenced by imaging techniques and numerical quantifications, subtle qualitative disparities were discernible in the histological analysis (Fig. 8, Fig. 9).

Fig. 8.

Degradation and new bone formation of WE43-screws and plates after 18 months. The screws are notably infiltrated by mineralized bone (black arrow), while the plate is covered by a thick layer of bone (∗). Stained with Giemsa, scale bar in the overview: 1 mm.

The anatomical location of the implant influenced its degradation pattern. As observed in the quantitative histological analysis, screws within the bone marrow region exhibited more pronounced degradation, likely due to higher fluid exchange and metabolic activity in this vascularized area. In contrast, screws embedded in dense cortical bone showed comparatively preserved structure, suggesting that the local microenvironment plays a key role in modulating degradation behavior. This spatial variability in degradation is likely attributable to differential exposure to chlorine-containing agents, such as blood and interstitial fluids, which are known to accelerate magnesium degradation [21,23]. These findings suggest that PEO surface modification not only mitigates the rate of degradation but also contributes to more controlled and predictable bioabsorption kinetics over extended time periods, thereby minimizing variability and enhancing overall implant performance.

To further elucidate the underlying mechanisms driving this degradation pattern, future research should focus on the vascular architecture and canalicular network of the bone. Specifically, investigating whether regions exhibiting accelerated implant degradation correlate with areas of increased fluid exchange could provide critical insights into the local biochemical environment affecting magnesium resorption. Such analyses may inform the refinement of implant design and placement strategies to optimize long-term performance and biocompatibility.

Another important aspect to consider is the advanced degradation stage of both plates and screws after 18 months, which complicate the radiographic assessment of the degradation progress. Tomographically, distinguishing between newly formed bone within the degrading implant structure and the residual metal is challenging due to their similar grayscale values. As the implant degrades, the transformation of metallic regions into bone-like tissue leads to a loss of radiopacity, potentially causing an underestimation of the degradation extent. This limitation underscores the need for complementary histological analysis to accurately assess bioabsorption dynamics and differentiate between remaining implant material and newly integrated bone. Future studies may benefit from further advances in imaging methodology to improve discrimination between degrading magnesium, degradation products, and newly formed bone. Beyond higher-resolution micro-CT systems, advanced reconstruction and segmentation approaches, including machine learning–based algorithms trained on correlative histological data, may help to overcome current limitations related to overlapping grayscale values. Such multimodal and data-driven strategies could substantially enhance the accuracy of long-term degradation and osseointegration assessment in bioabsorbable magnesium implants.

Differences in degradation behavior between plates and screws can be attributed to variations in the surface modification thickness of the PEO layer, manufacturing processes, and anatomical positioning. Screws in this study received a thicker PEO layer and were fabricated via extrusion and Swiss turning, while plates were produced by direct milling of as-cast heat treated WE43MEO and surface-modified with a thinner oxide layer. This, combined with their placement outside the highly perfused bone marrow cavity, likely reduced fluid exposure and limited degradation kinetics. This difference likely contributed to the more pronounced degradation-delaying effect observed for PEO-modified screws compared to plates. In regions where the PEO coating may have been partially degraded or locally stripped, exposure of the bare magnesium surface could potentially accelerate degradation and alter the local chemical microenvironment.

However, histological analysis revealed no evidence of adverse bone reactions or impaired osteogenesis in these areas, suggesting that degradation still remained fully biocompatible even under such conditions.

Moreover, the higher material density and less porous microstructure of plates impeded bone infiltration, in contrast to screws, where circumferential bone ingrowth was commonly observed. These factors collectively contribute to the slower and more heterogeneous degradation of plates compared to screws.

The in vivo findings of the presented study agree with earlier in vitro investigations, which demonstrated that PEO surface modification of WE43 magnesium alloys results in delayed and more controlled degradation under simulated physiological conditions. These studies have consistently shown that PEO surface modifications reduce hydrogen gas evolution, mitigate pH fluctuations, and create a stable barrier layer during immersion in simulated body fluids [18,[24], [25], [26]]. Other research has highlighted the role of electrolyte composition and coating thickness in modulating degradation kinetics and surface morphology, which may explain the observed differences between screw and plate degradation profiles in this study [5,27]. The consistency between our long-term in vivo outcomes and these prior in vitro findings supports the utility of early-stage degradation testing in predicting the bioabsorption behavior of PEO surface-modified magnesium implants.

The lack of significant differences in bone quantity between PEO-modified and non-modified screws is consistent with the observed degradation behavior. Since both screw types exhibited similar degradation rates, the associated hydrogen gas formation is expected to be comparable, resulting in an equivalent impact on bone quantity. However, bone quality appears to be significantly enhanced after 18 months, as evidenced by the higher percentage of lamellar bone surrounding the PEO-modified samples. This may be attributed to a more controlled early degradation behavior with a smoother ion-release profile and reduced micromotion, as well as to the osteoconductive, microporous surface topography of the PEO modification [28,29]. Over time, it is expected that normal bone remodeling, in accordance with Wolff's law [30], will lead to a restructuring of this bone along the physiological stress lines, ultimately restoring its original form and orientation. This outcome is in contrast to earlier time points (6 and 12 months), where similar trends were observed, though these did not reach statistical significance [22]. This observation suggests that bone turnover may have been reduced around the surface-modified screws, or that the restoration of lamellar bone may have been preferentially promoted — potentially facilitated by the rough and microporous surface topography generated by PEO, which is known to enhance osteoblast attachment and osteoconductive remodeling.

Differences in degradation patterns between screws and plates (Fig. 8, Fig. 9) are influenced by two key factors. Firstly, the thickness of the surface modification and the fabrication process and secondly the positioning of the plates inside the situs. The plate's smooth surface and location outside of the highly perfused bone marrow region provides reduced contact with surrounding liquids and tissues in comparison to the thread-enlarged surface and position of the screws. Furthermore, in contrast to extruded and SWISS type turned screws, the plates in this study not only had a different layer thickness, but were fabricated using direct milling of as-cast the magnesium alloy, which is known to alter the degradation kinetics, even if rather towards faster degradation rates [17]. Taken into consideration the high amount of material associated with the plate design used in this study, an even faster degradation might easily be overcompensated by the increased volume of the plates, further explaining their slower degradation profile.

The observed degradation patterns have a substantial impact on the extent of bone infiltration into the material. In the core of the screws, significant bone infiltration is observed, extending through the entire circumference. In contrast, minimal bone invasion is evident in the plates (Fig. 9), likely due to the dense structure of the remaining, non-degraded bulk material of the plate, which appears to impede bone penetration as well as the less active tissue environment outside of the bone marrow.

No significant differences in bulk or degradation material were observed between WE43 and WE43-PEO plates, consistent with the results for the screws. The thinner PEO surface modification on plates likely resulted in a less pronounced effect on their degradation behavior. Previous research has demonstrated that PEO surface modification thickness significantly influences magnesium implant degradation rates, with thicker surface layers providing enhanced degradation resistance by forming a more durable barrier, while thinner layers offer less protection, leading to faster degradation. Other studies have confirmed this relationship, which highlights the importance of optimizing PEO layer thickness to balance degradation control and mechanical stability [31]. The small thickness of the PEO surface modification on the tested plates in this study might very likely explain the similarity in degradation rates between the respective WE43 and WE43-PEO groups, while visible differences were observed between WE43 and WE43-PEO screws, exhibiting higher surface layer thickness.

A significantly higher volume of newly formed bone was observed above non-surface-modified WE43 plates at 18 months, in contrast to earlier 6-month data favoring PEO-modified implants [16].

This suggests phase-dependent bone remodeling influenced by degradation stage. The more predictable bone apposition seen with WE43-PEO may be advantageous in anatomically sensitive regions. While Kong et al. reported minimal surface degradation in a goat model after 18 months, our study adds deeper histological insight across implant regions, emphasizing the importance of anatomical context. Extending the follow-up to 18 months offers essential data on late-stage degradation and osseointegration, aligning with clinical scenarios where implants remain in place long after healing. Long-term studies in both load-bearing and non-load-bearing regions have consistently demonstrated successful bone healing and biocompatibility of magnesium-based implants without systemic complications [[30], [31], [32], [33]]. Specifically, no adverse effects on liver or kidney function, nor elevated serum magnesium levels, were reported in these models [31]. Although serum magnesium was not assessed in the present 18-month study, prior data from the same animal model at 6 and 12 months support the conclusion of negligible systemic impact [17].

Magnesium alloys are generally regarded as biocompatible, yet their degradation can release ions and hydrogen gas, which may — if uncontrolled — trigger adverse local responses such as inflammation, immunotoxicity, and immune cell infiltration [32]. These effects are typically associated with rapid or early-stage corrosion. However, in the present study, histological analysis revealed no signs of chronic inflammation, fibrotic encapsulation, or abnormal immune cell presence, even in areas with advanced degradation. This indicates a favorable long-term tissue response to both PEO-modified and non-surface-modified WE43 implants.

Despite these positive outcomes, degradation behavior showed high variability depending on implant location, alloy composition, surface modification, and manufacturing method [33]. Full degradation and complete lamellar bone remodeling were not achieved by 18 months. Persistent gas release and ongoing degradation likely delayed full bio integration, particularly in the case of plates. Clinically, this suggests that implant remnants may still require removal or over-drilling if revision surgery is needed within this timeframe.

Nevertheless, degraded magnesium offers potential advantages during revision procedures: its weakened structure facilitates easy cutting, and any residual fragments are gradually resorbed without long-term risk. The present study provides the first systematic long-term comparison of surface-modified and non-modified WE43 implants, highlighting differences in degradation kinetics and osseointegration across anatomical regions. A notable observation was the progressive replacement of magnesium with bone, temporarily preserving the implant's morphology before remodeling into physiologically adapted lamellar bone.

The present study was designed as a long-term translational validation model rather than a fracture-healing or functional fixation study. The use of a non-fracture model allowed controlled assessment of late-stage degradation behavior, tissue response, and biosafety of PEO-modified versus non-surface-modified WE43 implants without confounding factors related to fracture instability, heterogeneous healing kinetics, or variable mechanical loading. We acknowledge that a fracture-healing model would be essential to evaluate functional fixation performance under clinically relevant load-bearing conditions. The present non-fracture model was intentionally selected to isolate long-term material degradation and tissue response without confounding effects of fracture instability or heterogeneous healing kinetics. Future fracture-model studies represent the necessary next translational step.

Methodological limitations in this study primarily arise from the advanced degradation stages of both plates and screws after 18 months. As magnesium undergoes degradation, its radiopacity increasingly resembles that of mineralized bone, which can occasionally impede differentiation between bone and implant during micro-CT analysis. Additionally, the precise quantification of fragmented screws in histology is more prone to inaccuracies due to the high number of small fragments and the increased difficulty of measuring them compared to a larger, intact implant area. Furthermore, this study did not include a fracture-healing model, which would provide more direct insights into the tested system's functional performance under load-bearing conditions.

Future research should focus on implant degradation in a large animal fracture model under load-bearing or -load-sharing conditions to determine whether degradation rates align with fracture healing times [34]. In addition, although the 18-month biological endpoint offers valuable insights into the terminal stages of degradation and long-term tissue response, the study design is limited by the absence of intermediate time points. As such, no conclusions can be drawn regarding the kinetics or temporal sequence of bone formation or implant resorption. A longitudinal design with multiple observation time points would be necessary to accurately characterize these full dynamic processes. As magnesium implants degrade, their stability diminishes, resulting in a progressive transfer of load to the bone [20]. The pivotal factor determining the efficacy of such implant systems is whether the diminished stability of the implant still ensures adequate mechanical support to facilitate effective bone union. In addition to surface modification, emerging manufacturing strategies such as additive manufacturing and compositionally tailored biodegradable metals may further expand the design space for patient-specific magnesium based orthopaedic implants.

5. Conclusion

PEO surface modification of WE43 magnesium implants supported controlled long-term degradation and favorable peri-implant bone remodeling, particularly in screws. After 18 months, degradation had progressed substantially in all implants; however, PEO-modified screws showed greater structural preservation in the bone marrow region and were associated with increased lamellar bone formation without adverse local tissue responses. The different degradation patterns observed for plates and screws highlight the relevance of implant geometry, surface characteristics, and local anatomical environment for long-term implant performance. The presented findings provide long-term preclinical evidence supporting the biosafety and controlled degradation profile of PEO-modified WE43MEO implants and support further evaluation in load-bearing fracture models to assess functional fixation performance.

Ethics

Ethical approval was granted by the Ministry for Gesundheit und Verbraucherschutz Hamburg (Approval No. 49/16).

Declaration of generative AI in scientific writing

No generative artificial intelligence (AI) or AI-assisted technologies were used in the preparation of this manuscript.

Funding

This study was funded by Karl Leibinger Medizintechnik GmbH & Co. KG.

Conflict of competing interest

Alexander Kopp: CEO of

Meotec GmbH

Alina Schmitz: none.

Heilwig Fischer: none.

Henri Kreiker: none.

Ana Prates Soares: none.

Katharina Schmidt-Bleek: none.

Georg Duda: none.

Henning Hanken: none.

Nadja Kröger: none.

Max Heiland: received speaker remuneration from Karl Leibinger Medizintechnik GmbH & Co. KG.

GmbH & Co. KG.

Carsten Rendenbach: received industrial funding from Karl Leibinger Medizintechnik GmbH & Co. KG.

GmbH & Co. KG for research projects.

Ralf Smeets: received industrial funding from Karl Leibinger Medizintechnik GmbH &

Co. KG for research projects.

Competing interests had no impact on the quality or reliability of the results.

Appendix A. Supplementary data

The following is the Supplementary data to this article: