Yu Sun, Heike Helmholz, Regine Willumeit-R?mer

a Division of Metallic Biomaterials, Institute of Materials Research, Helmholtz-Zentrum Geesthacht, 21502 Geesthacht, Germany

b Department of Orthopedics, First Hospital of China Medical University, Shenyang 110001, China

Received 23 June 2020; received in revised form 13 August 2020; accepted 27 September 2020

Available online 9 November 2020

Abstract Background: Degradable magnesium implants are promising for clinical fracture treatment, providing less stress-shielding mechanical support and superior bone-strengthening benefit to traditional materials. The quality of preclinical research is essential for developing Mg implants; however, there are considerable variations in the model selection and study design in published studies, posing challenges for safe and effective clinical translation of lab discoveries.

Keywords: Magnesium; Implant; Fracture; Animal model; Study design; Biomaterial.

1. Background

Bone fracture is one of the most commonly seen diseases in clinics. In European Union countries, there were 3.5 million fragility fractures reported in 2010, with a health care cost of about 37 billion euros, and the cost is expected to increase by 25% till 2025 [1]. Current mainstream treatment with internal fixatio devices has achieved great success in the past several decades and continues evolving with precise preoperative design and individualized 3D-printed prostheses,ensuring accurate fracture reduction and early rehabilitation.However, nowadays, the incidence rate of postoperative delayed union or non-union is still reported to be more than 5%among patients. The notable risk factors of the postoperative complications include the stress-shielding effect resulted from implant materials like titanium(Ti)or stainless steel(SS)with high elastic modulus, as well as secondary surgeries for implant removal [1,2].

The development and clinical application of novel biodegradable materials, such as magnesium (Mg) and its alloys, may compensate for the above-mentioned disadvantages of Ti and SS. Degradable Mg implants are designed to provide necessary mechanical stability, supporting both anatomical healing and functional reconstruction with purely bony tissue.The recent years witnessed in vitro and in vivo studies reporting novel Mg materials with optimized biological properties as potential candidates for clinical application, promoting bone formation, and regulating immune and inflammator responses[3-5].The clinical use of an Mg-based osteosynthesis implant, the MAGNEZIXcompression screw, firs reported in 2013, was a milestone for the translation of biodegradable metallic implants [6]. A further clinical study showed its encouraging results after more than one-year follow-up,supporting successful bone healing in patients undergoing foot/ankle surgeries [7], and other clinical case series also reported successful treatment of tibia spine avulsion fractures and displaced femoral neck fractures using Mg-based implants[8,9].

However, the above Mg implants are indicated in surgeries for small to medium-sized bones, not mid-shaft repair of long weight-bearing bones. While the most severe complications and urgent needs for innovative treatment, reside in orthopedic trauma of long weight-bearing bones, such as femur and tibia shaft fractures [10,11]. For Mg-based implants, there are concerns that the materials may degrade too fast and lose mechanical stability before fracture healing, suggested by early reports of failed clinical applications [12]. With more efforts put into the production and processing technologies of Mg alloys, novel Mg-based materials are developed with tailored degradation and mechanical properties [5,6,13].

The clinical application of novel medical devices requires rigorous investigations and detailed study of the potential accompanying reactions, and preclinical animal studies conducted under the 3R principles (Replacement, Reduction and Refinement are indispensable. The challenge for developing biodegradable materials lies not only in the different characteristics between in vitro and in vivo conditions, but also in the heterogeneities among different animal models [3,14].Previous in vivo studies were primarily planned to investigate Mg-based implants’ behavior in non-fracture models,for the general study of biodegradation and biocompatibility[15-18]. Different models involving several animal species and various methods for inducing fractures have been reported, and different types of Mg materials and implant designs have been investigated; however, there are currently no guidelines regarding the proper planning of preclinical studies on Mg-based biodegradable implants in fracture models.The purpose of this study is to conduct a systematic literature review on Mg implants in animal models of fracture healing,focusing on their study design, selection of Mg materials,and evaluation techniques, to provide practical guidance for the future preclinical in vivo research.

2. Materials and methods

2.1. Search strategy

A literature search for systematic reviews on preclinical studies of Mg implants and fracture healing was firs performed, indicating no previous related publications on this topic. Then the PubMed and Embase online databases were searched to identify articles published from 1960 to the end of December 2019, for in vivo studies of Mg implants in animal models of bone fracture, using a combination of the following keywords: magnesium and fracture. The whole literature research followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines [19], and a fl w chart was used to document the screening and selection process (Fig. 1).

2.2. Search criteria

The literature search included randomized and nonrandomized in vivo studies, evaluating Mg implants in animal models of fracture healing using radiological, biomechanical,histological, and other techniques. There were no restrictions on the anatomical site of the fracture induction or comorbidities such as osteoporosis or diabetes. The inclusion criteria were: (1) original in vivo preclinical studies; (2) using animal models of bone fracture and Mg-based implants; (3) reporting postoperative information such as radiological and histological images of fractured bones, or related evidence based on other evaluation methods. The exclusion criteria were: (1) review paper, commentaries, abstract publication and letters to the editor; (2) in vitro studies without in vivo evaluation; (3)animal studies reporting evaluation of Mg implants in only non-fracture models. There were no restrictions on the language of the publications.

2.3. Selection of studies

The authors performed the selection process using the two online databases in duplicates, screening the titles and abstracts within the result lists, and articles were selected based on the inclusion and exclusion criteria. The articles identifie as potentially meeting the above criteria were further reviewed for the study design and methodology, and full texts of the relevant studies were obtained. The reference lists of these full texts were also read and evaluated, to check for other potentially eligible studies.

Fig. 1. Flow diagram of literature retrieval and screening process.

2.4. Data extraction

All full texts within the selected results were read carefully by the author, and standardized forms were used for data extraction. The following data were extracted from the full texts of selected eligible articles: authors and country, publication year, animal species, gender, number of animals, anatomical position of fracture, methods for induction of fracture, observation period, Mg and control materials, implant design,methods of evaluation and major research conclusions.

2.5. Reporting quality and analysis

The methodological quality of the studies was assessed,focusing on the report of the following items: randomization,blindness, and drop-outs. The author identified recorded and assessed the quality of each study using the following criteria:1)when all three items were reported,the study was classifie as having a low risk of bias; 2) when one of the three items was lacking, the study was classifie as having a moderate potential risk of bias; 3) there would be a high potential risk of bias if the study lack two or more of these items.Due to the heterogeneities in animal models and evaluation techniques,the authors did not conduct a quantitative meta-analysis, and only a qualitative comprehensive review was performed.

2.6. Ethics

This study is a systematic review that does not require ethical approval.

3. Results

3.1. Literature search results and study characteristics

The preliminary literature search identifie 681 and 1420 studies in PubMed and Embase, respectively, and a total of 1597 records were left after removing duplicates. The authors reviewed each title and abstract, and then removed an additional 1486 records. Among 111 potentially eligible publications,90 were further removed for not satisfying the inclusion or meeting the exclusion criteria after careful evaluation of study design and published results. Among the remaining 21 publications [4,14,20-38], two articles reported results from different aspects of one animal research, and were considered as a single study [4,31]. Finally, 20 studies were included for the systematic review (Fig. 1 and Table. 1), all meeting the inclusion criteria as preclinical in vivo studies of Mg-based implants using animal models of fracture healing.

The 20 included studies were published from 2014 to 2019(Table 1). Half of the projects (10 publications) were conducted by research groups (institutes of corresponding authors) in Asia, of which 40% (8 publications) were published by authors from China; European research institutescontributed 35% of the included studies, Germany and Spain each accounted for 3 and 2 publications, respectively; All studies in America were published by research teams from the United States, accounting for 15% (4 publications) of all included researches (Fig. 2).

Table 1 Selection of animal models in included studies.

Fig. 2. Number of publications per country.

3.2. Methodological quality assessment

All studies provided basic information regarding ethical approval, animal housing, experimental design, surgical procedures, and outcome evaluations, but methodological information was generally inadequate. Only 10% (2 studies) reported blinding during radiological evaluation. Of the included studies, 60% reported randomization of lab animals, and 75%reported whether there were drop-outs due to surgery failure or other reasons (Fig. 3). Half of the included studies were classifie as having a moderate potential risk of bias [20,23,27,30,33-38], and the other 50% were classifie as having a high potential risk of bias [4,14,21,22,24-26,28,29,31,32].

Fig. 3. Methodological evaluation of included studies (reporting of randomization, blinding and drop-out).

3.3. Selection of animal models

Rats and rabbits were the most commonly used animal species. Of all studies included, 35% selected rats and 30%selected rabbits for research, 15% selected miniature pigs,and another 10% selected beagle dogs as model species, only 1 study used mice and 1 study goat (Table 1). A specifi tendency for gender selection was observed for studies with rats (5 out of 7 selected female rats). Statistical methods for sample size calculations were not mentioned in the included studies. Researches with rats and rabbits tend to use more animals than those selecting dogs and pigs, but the number of animals varied considerably within the same species.

The same situation also exists in the settings of observation time. The overall follow-up period ranged between 3 weeks and 13 months for rat models of femoral fractures, between 4 weeks and 16 weeks for the rabbit model of forelimb fractures, and between 12 weeks and 24 weeks in rabbit models of hind limb fractures. The follow-up period for studies involving dogs and pigs varies from 4 to 12 weeks and 8 to 38.6 weeks,respectively.One study using the goat femoral condyle fracture model had the longest postoperative observation time of 18 months. Generally, more information on the degradation process of Mg-based materials could be obtained with a more extended follow-up period, but some long-term studies did not provide detailed results such as volume changes of implants with quantificatio [20,38].

All studies using rodents (mice and rats) selected femur as the anatomical location for inducing fractures. For rabbits,2 studies selected hind limbs (femur and tibia) and 4 studies selected forelimb (ulna and radius). Tibia was selected in 2 studies using rabbits and dogs, femoral condyle was selected in 2 studies using rabbits and goats, and cranial-facial bones were selected in 4 studies using dogs and miniature pigs.Most researchers adopted osteotomy for fracture induction (70%),and there were also other methods, including the use of bone cutter,bone clamp,or three-point bending devices.The details were summarized in Table 1.

3.4. Mg-based materials and types of implant

Both pure Mg and Mg-alloys have been studied in animal models of fracture healing. Among the 20 studies included,30%evaluated pure Mg,20%evaluated Mg-Al-Zn-Mn alloy,another 20% evaluated Mg-Y-RE-Zr alloy, and 10% evaluated Mg-Nd-Zn-Zr alloy. The Mg-Ag, Mg-Ca-Zn, Mg-Ca-Mn-Zr, and Mg-Y-Zn-Zr-Ca alloy were evaluated in 1 study, respectively. For the setting of control groups, both polymers such as poly-L-lactic acid (PLLA), polylactide-coglycolic acid(PLGA),and metallic materials(SS and Ti)were selected for comparison. Of the included studies, 40% investigated the influenc of surface coating on biological response or implant degradation. Related details were summarized in Table 2.

Among all studies, the implant design of the plate/screw system was only applied in large non-rodent animals, including rabbits, dogs, and pigs. All studies using small rodents(mice and rats) selected intramedullary fixatio methods, with implants in the form of pins, nails, bars, and rods. Of the included studies, 10% selected plates and strips of AZ31 alloy in rabbit models without rigid fixation and 15% studies selected screws for fracture fixatio in animal models of rabbits,dogs, and goats. The more complex external fixatio devices were not selected by the included studies.The mass of the implant was available in one study [23], and most studies (95%)provided the size of the implants (information summarized in Table 2).

Fig. 4. Number of studies grouped by combining fracture sites and fixatio methods.

As the methods for stabilization(fixatio devices)influenc the healing status and related biological events [39-41], the authors further analyzed the fixatio methods within included studies. Intramedullary fixatio without locking mechanism,which suffers both axial and rotational instability, is considered “fl xible fixation” and the screw/plate system is considered as “rigid fixation” Moreover, the selection of anatomical sites for inducing fractures is also a factor influencin the healing process. Long-weight bearing bones such as femur and tibia were considered “unstable anatomical sites”,and the cranial-facial bones, radius and ulna of rabbits were considered “stable anatomical sites”. Table 2 summarized the details of implant devices used in the included publications,and Fig. 4 showed the analysis of the selection of anatomical sites and fixatio methods among studies.

In addition to anatomical sites and design of fixatio devices, the in vivo degradation process of Mg-based implants could also influenc the stability and biological healing process of fractured bones. Of all the included studies, 60%(12 articles) selected to induce fractures at unstable anatomical sites [14,20-27, 30,34,38], and 5 of the 12 studies conducted CT analysis, which could provide radiological details of implant degradation [14,22,23,25,30]. None of the above research reported post-operative re-fractures due to the degradation of Mg implants.

3.5. Evaluation techniques and research conclusions

Of the included studies, the most commonly used methods for evaluating bone healing and implant status were radiology and histology. All included studies adopted radiology, and 95% performed both radiological and histological examinations (Table 2). X-ray plain film were used by 65% of studies. In vivo and ex vivo micro-CT were adopted by more re-searchers in 75% of all studies, involving qualitative (visibility of fracture line and implant morphology) and quantitative(bone mineral density, bone/tissue volume, and implant volume) evaluations during the healing process. The application of micro-CT for analyzing implant degradation was reported in 40% of all studies [4,14,22,25,30,31,35-37]. Histological methods included conventional hematoxylin-eosin staining for general information on tissue inflammatio and special staining such as Von-Kossa for evaluation of bone mineralization. Biomechanical analyses were less popular than radiological and histological evaluations. Three or four-point bending tests were only reported in 25% of all included studies[4,24,25,33],and finit element analysis was conducted in one study for comparison between different types of plate/screw combinations, as well as a simulation of four-point bending tests [27]. Blood examinations were reported by 25% of all research [22,28,32,33,38], and 15% reported the measurement of serum Mg ion levels [22,28,32]. Only one study reported the measurement of biomarkers of bone metabolism in the peri-implant callus tissue [38]. Technical details such as voxel resolution of micro-CT as well as methods for histological staining and mechanical tests were summarized in Table 2.

Table 2 Mg materials and evaluation methods in included studies.

Table 2 (continued)

As to research conclusions, except one study focusing on the modeling of the degradation process [26], most studies generally supported the application of Mg-based materials as potential candidates for future orthopedic implants. Improved bone healing could be expected by the application of both pure Mg and Mg alloys, when used independently or in combination with Ti and SS implants [24,27]. In two studies, the surface coating on Mg-based implants did not influenc the in vivo degradation compared to their non-coated counterparts[23,37].

4. Discussion

The publication of systematic reviews have generally been accepted and emphasized in clinical studies of novel diagnostic and therapeutic methods; however, it is still not a common practice for animal research [42,43]. Considering the limited number of publications and significan challenges in translating animal research conclusions into clinical application, a systematic review of previous in vivo studies could be informative for planning future translational research of Mg implants. The authors followed a standard structured process to retrieve and select eligible studies based on inclusion/exclusion criteria, evaluated the methodological quality,and extracted data, aiming at performing an evidence-based summary rather than a classical, narrative review based on expert opinion.

Biodegradable orthopedic implants have been used for decades, while the majority of current degradable devices were made of polymeric materials with disadvantages, such as sizeable volumetric geometry or low mechanical strength,as well as incomplete degradation [44]. In recent years,degradable Mg-based metal implants have become a research hotspot for their mechanical properties and biological bonestrengthening effects comparing with polymers. Furthermore,the successful introduction of MAGNEZIXin clinical practice has especially inspired more exploration of Mg and its alloys for bone repair, with efforts to extend the treatment indications for fractures of long weight-bearing bones [8,45].More future publications could be expected from Europe and China (Fig. 2), as researchers there contributed 85% of current included studies, and both regions have approved clinical trials of Mg-based orthopedic implants.

Well-designed, thorough, and rigorous preclinical investigations of degradable Mg implants are necessary for effica y and safety. The risk of implant degradation before fracture healing and related mechanical failure is still a significan concern,and the formation of resorption cysts was reported by clinical follow-up in patients undergoing fixatio of scaphoid fractures using Mg implants [46]. Though Mg implants have already entered clinical application and novel Mg-based materials are widely investigated for bone repair, there were still relatively few publications of preclinical studies using fracture models, compared with researches involving non-fracture models. Of the 14 included studies that did not investigate pure Mg, seven types of different Mg-based alloys were evaluated, indicating that more comprehensive preclinical investigations are necessary for each candidate material. Besides,the included studies focused on presenting methods and results from the biomedical aspect, while details in material production and implant preparation were generally poorly reported (Table 2). It should be realized that for biodegradable alloys, different material processing and post-treatment such as casting and extrusion may bring about significan differences in biological performance; thus an adequate reporting of the materials production and implant preparation is necessary for a comprehensive evaluation,comparison,and analysis of Mg-based implants.

In vitro studies provide enormous information for both Mg degradation and the underlying biological mechanisms. These investigations could be achieved by simulating in vivo physiological or pathophysiological environment, via modulation of parameters such as pH and pressure of oxygen, with or without using a series of cell lines or tissue culture techniques.However, investigations using animal disease models cannot be replaced by cellular or tissue level research. For example,the evaluation of implant degradation, the formation of gas cavity, local and systematic distribution of corrosion products, their effect on re-vascularization, and potentially negative inflammator response, are essential for examining the long-term performance and safety issues for biodegradable Mg and its alloys. Currently, the whole set of the above analysis and results could only be available by using preclinical animal models.

Besides, by selecting appropriate animal models, researchers can simulate the damaging impact of common clinical comorbidities on fracture healing, such as diabetes, traumatic or postoperative infection, and osteoporosis. During the literature search process,the authors did not set restrictions on the comorbidities in animal models and found two studies using rat models of ovariectomy-induced osteoporosis [24,25].The authors recommend that researchers may consider selecting complex fracture healing models with comorbidities,under the premise of facility conditions and animal ethics.The type of study design will help explore the potential clinical application of magnesium-based materials in patients with similar clinical conditions.

Conducting animal research could be far more time consuming than in vitro experiments, and the translational value and quality may still be problematic in many cases. Despite the application of numerous radiological, histological, mechanical, and other techniques for evaluation, many animal studies tend to have minor value and are redundant in the translational perspective, because of poor study design or inappropriate model selection. It was estimated that the preclinical research on bone morphogenetic proteins led to more than 500 publications, with the usage of at least 17,000 lab animals; however, the majority of the studies were with poor design,and variations in species and dosage yielded only misleading or inconclusive results, and clinical applications were reported with side effects [47].

To improve the model selection, methodological quality, and translational effica y for the future development of biodegradable Mg implants, tight cooperation between material scientists, biologists, statisticians, and clinicians is of importance. This systematic review revealed a relatively high risk of bias and low methodological quality among included in vivo studies, which might be improved by collaborations of the researchers mentioned above from different backgrounds.However, it should also be considered that there are inherent methodological difficultie for surgery-related studies in terms of randomization, blinding, and allocation concealment, compared with studies of novel non-surgical therapies.

Of the included studies,rats and rabbits were the most popular animal species, accounting for 65% of all projects. Mice have advantages of relatively lower cost, less space for husbandry, shorter healing period, and more genetically modifi cation applications than other larger animal species; however,only 1 included study selected mice as the research model.Possibly the tiny size and skeletal structures posed technical difficultie both for surgery and preparation of fixatio devices. Similar technical difficult also exists in included studies using rats, as only intramedullary fixatio devices were selected.

More implant designs were available for studies using large non-rodent species, with broader application field involving more anatomical sites, not only extremities but also cranialfacial bones. Sheep and goats are recognized as superior animal models for preclinical evaluation of orthopedic implants,especially in developing fixatio devices for long-weight bearing bones, due to the similarity of anatomical morphology between their tibia and human bones. However, only 1 included study selected goats and induced femoral condyle (but not bone shaft) fractures as the disease model. Moreover,it should be a common practice to use at least two laboratory species, such as rodents and non-rodents, to assess a novel fracture-healing treatment before clinical application,to compensate for the biological gap between animal species.Nevertheless, most current in vivo studies only involve one species, and huge variations existed in the setting of postoperative follow-up time points for radiological and other evaluations.

The authors also evaluated the selection of anatomical sites to induce fractures and methods of bone fixation Mg-based non-locking intramedullary fixatio devices adopted by all studies using mice and rats, could only provide limited fix ation without satisfying rotational or axial stability required in clinical fracture treatment. Further analysis revealed that a less reasonable combination of “fl xible fixation and “unstable anatomical sites” did exist in many studies, especially those using small animal species (shown in Fig. 4), which may provide relatively less translational prospects for clinical applications.

Fig. 5. General aspects of research planning.

The degradation of magnesium-based materials may lead to mechanical failure of fixatio devices, causing severe complications such as re-fractures in clinical applications. From the perspective of animal welfare, degradation-related implant failure, or breakage before fracture union, should be avoided as much as possible. The risk of such severe complications might be reduced during the experimental design stage of preclinical studies.

As to the methods for inducing fractures, the osteotomy should be the best method with reproducible results, adopted by most studies (Table 1). To better mimic the clinical situation and improve the translation effica y, the authors recommend using relatively rigid fixatio methods on unstable anatomical sites and inducing fractures using reproducible osteotomy tools.

Effective fracture treatment should focus not only on the reconstruction of continuity of bone tissues, but also restoration of bone quality and its metabolic function. The planning of preclinical evaluation should also consider not only structural but also mechanical and metabolic analysis. For structural evaluation, micro-CT has several advantages over traditional histological techniques for analyzing small animal bone specimens,and has been established as a non-destructive analysis method in bone research, with high resolution allowing quantitative measurement of bone microstructure and implant degradation, without complicated preparation of tissue samples [48]. For biomechanical evaluation, three or four-point bending tests were widely adopted in preclinical studies of bone healing and osteoporosis, and serum biomarkers such as alkaline phosphatase (ALP), osteocalcin (OCN), procollagen type 1 amino-terminal propeptide (P1NP), and C-telopeptide(CTX) have been demonstrated to be essential indicators reflectin the state of bone metabolism in both preclinical and clinical studies [49-53]. In vivo and ex vivo micro-CT were adopted in the majority of included studies,while biomechanical tests and serum biochemistry were relatively less reported(Table 2).

Based on the results of current literature research, several aspects should be considered when planning in vivo experiments, as listed in Fig. 5, not only for ensuring the rational use of lab animals, but also for attaining reproducible and translational results. To obtain reliable data regarding the effica y and safety of Mg-based implants, researchers should carefully consider the available small and large animal species and their characteristics under the 3R principles, as well as the anatomical location and surgical techniques for inducing fractures, which directly influenc the implant design and fix ation methods.

Moreover, effective animal models should simulate the intended clinical situation, and appropriate methods must be chosen for the assessment of bone healing and implant degradation. The authors recommend using in vivo or ex vivo micro-CT in combination with histology, biomechanical evaluation, and serum biomarker of bone metabolism as primary methods. Besides, it should be realized that, from a practical and ethical point of view,it could be challenging to ensure animal experiments with high methodological levels, especially in studies involving complex surgical procedures. However, a thorough review of the listed significan aspects during study design will still support investigators to form a detailed research plan with a considerable value of clinical significance

5. Conclusions

Current evidence from preclinical experiments suggested that the application of Mg implants is beneficia to promote fracture healing. However, considerable heterogeneity existed among studies, regarding animal species, implant preparation,surgery, and evaluation techniques. The literature research also revealed the limitations in methodological quality and risk of bias among in vivo surgical studies,to be improved via detailed planning and reporting of randomization, blindness,and drop-out. The finding indicated that there is still a lack of a standardized reference model to develop Mg-based implant materials for fracture treatment. Nevertheless, the technical details extracted from published articles, may function as building blocks for comprehensive study design and standardized protocol.This evidence-based systematic review may provide useful information on the selection of clinically relevant animal models, design of implants, and evaluation techniques, for planning and conducting preclinical research with a translational perspective.