Safety and efficacy of the polymer enhanced AS-ELARIS^® pedicle screw system using ultrasonically liquefied polylactide: a short-, mid-, and long-term posterolateral fusion study in sheep

Abstract

Background

Lumbar fusion surgery using pedicle screws and intervertebral cages is frequently performed to treat low back pain when conservative therapies were unsuccessful. Screw loosening, however, remains one of the most common postoperative complications. The purpose of this study was to show that the fixation with polymer-enhanced AS-ELARIS^® Pedicle Screws by ultrasonic-assisted extrusion of 0.104 mL resorbable polylactide per screw, represents a suitable alternative to current fixation techniques.

Methods

Forty adult female sheep were enrolled in the study and posterolateral fusions were performed at two non-adjacent lumbar levels (L1+L2, L4+L5), implanting eight AS-ELARIS^® Pedicle Screws per animal. After a survival period of 2 days, 8 weeks, 6, 12 or 24 months, animals were sacrificed and the instrumented vertebrae harvested for macroscopic, radiological, and histological evaluation.

Results

Despite surgical challenges in achieving optimal screw positioning, all evaluations demonstrated good biocompatibility and progressive osseointegration. Ultrasonic liquefaction-mediated polymer enhancement proved safe at all investigated timepoints, without evidence of thermal or inflammatory tissue damage.

Conclusions

The AS-ELARIS^® Pedicle Screw System may offer a promising alternative to current fixation techniques. Its biocompatibility, biodegradability, safe removal, and ease of handling make it a valuable addition to spinal fixation strategies.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13018-026-06773-9.

Introduction

Low back pain (LBP) is one of the most common diseases in the world’s human population. As much as 50–85% of the population worldwide is estimated to suffer from low back pain once in their life [1–3]. Da Silva et al. reported that 69% of patients had a recurrent episode after 12 months [4]. Risk factors for LBP are commonly associated with age (spinal degeneration), psychological health, weight, height and physical activity [3–7]. Particularly, loss of muscle strength often leads to spinal instability [5]. If conservative therapy is unsuccessful, lumbar fusion surgeries represent an option to stabilize the spine and minimize associated pain. Among the feasible methods for fusion or stabilization of adjacent vertebrae are surgeries using intervertebral cages [8] and posterolateral fusion by pedicle screws with or without augmentation.

Although spinal fixation is commonly performed, screw loosening is a postoperative complication in 12–58% of the patients [9, 10], more often reported for multi-level fusions [10]. Bredow et al. tried to predict the risk of screw loosening with the help of CT scans [9] and found a significant correlation between age and measured bone density, with lower bone density leading more often to screw loosening. Furthermore, up to half of the patients with loosened screws required revision surgeries [9, 11].

Polymethylmethacrylate (PMMA) bone cement augmentation is the most common option to enhance the pull-out strength and decrease the risk of screw loosening [12–14]. Liu et al. investigated the influence of screw shape, thread profile and the necessity of augmentation, and stated that conical screws with dual-core/dual-thread and cement augmentation are less likely subject to screw loosening than regular pedicle screws [15]. Furthermore, increased screw diameter causes higher fixation strength and stiffness [16, 17]. Another option to reduce the risk of postoperative screw loosening is the use of expandable pedicle screws. Similarly to the study results of Kiyak et al. [13], Gao et al. found that expandable screws with additional augmentation using PMMA bone cement could increase the pull-out strength [14]. While these techniques decrease the risk of screw loosening in different ways, they also come along with multiple disadvantages. PMMA bone cement, for example, is non-biodegradable and due to the exothermic reaction during polymerization [18, 19] bears the risk of thermal injury of surrounding bone and nerves and renders possible hardware removal more difficult. Also the removal of expandable screws is difficult and may lead to more bone/cement remnants on the threads which often increases the diameter of the screw hole [14]. The use of PMMA bone cement itself may result in cement leakage into adjacent structures, such as the intervertebral disc space, spinal canal, neural foramen, or paraspinal muscles. Depending on the site and extent, this could compress spinal cord or nerve roots with potentially severe neurological deficits. In addition, during its application, fat or cement embolism may occur, which can cause pulmonary embolism with clinical presentations ranging from asymptomatic to life-threatening [20].

The ultrasonic liquefaction process provides a promising alternative to the commonly used techniques. With the help of ultrasonic energy, a resorbable polylactide polymer pin (PLDLLA) in the inner core of the distally fenestrated ELARIS^® Pedicle Screw is liquefied, extends into the trabecular bone and hardens within seconds [21]. As the amount of polymer is limited to 0.184 mL in human applications, and it is immediately quenched when contacting bone, the polymer only spreads into the immediate surrounding of the screw. In comparison, PMMA bone cement has a higher flowability and infiltrates larger bone volumes in the vicinity of the screw and adjacent structures. Therefore, the ultrasonic liquefaction process is called bone enhancement in contrast to bone augmentation when using bone cement. Upon removal postoperatively, and also later in the course of the bone remodeling and polymer resorption process, the enhancement of the screw by PLDLLA should not create larger bone defects than removal of a screw without PLDLLA enhancement.

During resorption of the polymer, poly-L/DLactide hydrolyses firstly to lactic acid and subsequently degrades to carbon dioxide and H₂O. Due to its amorphous structure, it should resorb completely and without late inflammatory reactions that have been observed for semicrystalline PLLA [22]. Previous studies examined the risk of thermal injuries due to the ultrasonic liquefaction process [23, 24]. Eriksson et al. found that the threshold temperature of heat-induced bone tissue damage is 47 °C [25].

If bone tissue is exposed to temperatures above 47 °C for 5 min, it will lose its functionality, be resorbed, and eventually be replaced by fat cells. This process can lead to a reduction in bone tissue volume of up to 40% compared to pre-heating measurements.

Heidenreich et al. demonstrated that the temperature/time threshold level of 47 °C for 5 min was never reached when using the ultrasonic liquefaction technology. In fact, they reported a maximum temperature increase of 11.3 °C, which lasted for less than 1 s [23].

The purpose of this study was to test the short-, mid- and long-term safety and efficacy of the ultrasonic polymer enhancement process for the AS-ELARIS^® Pedicle Screw (AS=Animal Study adapted ELARIS^® Pedicle Screw for the anatomy of sheep). A posterolateral fusion model in sheep was designed to answer the following six questions:

1. Is the ultrasonic liquefaction technology safe regarding its thermal impact (short and transient temperature increase of the screw and/or polymer) on the surrounding tissue?

2. Is the polymer-screw combination safe regarding its biocompatibility including its impact of potential polymer migration outside the vertebral body caused by the ultrasonic process?

3. How efficient and safe is the polymer enhancement regarding its osseointegration at different timepoints?

4. Will the polymer degrade completely and be replaced by new bone during the observation period?

5. Is it possible to remove the polymer enhanced screws without causing harm to the surrounding bone?

6. Does the risk of pulmonary embolism increase using the ultrasonic liquefaction technology (results will be published elsewhere)?

The hypothesis of this study was, that the ultrasonic polymer enhancement process for the AS-ELARIS^® Pedicle Screw is biocompatible and safe throughout the entire polymer resorption process.

Materials and methods

Study design and experimental animals

All animal experiments were carried out according to the Swiss laws of animal protection and welfare and approved by the local governmental veterinary authorities (ZH013/16).

Forty-five adult, female, Swiss alpine sheep, averaging 31.9 months of age (24–38 months) and 73.9 kg body weight (65–83.4 kg) underwent surgery. Five animals had to be excluded from the group analysis for reasons not related to Test Item (TI) or Reference Item (RI) (n= 2 anesthesia complications, n= 3 surgical complications). Two posterolateral fusions were performed at two non-adjacent levels of the lumbar spine, resulting in a total of 8 pedicle screws implanted per sheep (Fig. 1). Only healthy animals with blood analyses values in normal range were acclimatized for at least 7 days under test conditions and randomly assigned to five groups:

Fig. 1.

Experimental design. Schematic illustration showing implantation of 8 AS-ELARIS^® Pedicle Screws and 4 rods per sheep

Eight sheep were included in the acute group analysis, specifically set up to detect pulmonary embolism during surgery inferred by potential migration of polymer particles or displaced bone marrow/fat cells to the lungs as well as immediate local biocompatibility (e.g. acute adverse thermal effects). These sheep either received a total of 8 Test Item screws (TI: AS-ELARIS^® Pedicle Screws with polymer enhancement) or 8 Control Item screws (CI: AS-ELARIS^® Pedicle Screw without polymer enhancement) and were sacrificed after 2 days. Results for the risk of pulmonary embolism of this group are not included in this manuscript, but will be published elsewhere (publication in preparation).

Thirty-two sheep were assigned to one of four osseointegration groups; survival was 8 weeks (n= 8), 6 months (n= 10), 12 months (n= 8) or 24 months for the long-term group (n= 6). In these sheep, one level was fused with CI and the other with TI, with the implants alternating between cranial and caudal positions.

Test and control item characterization

TI consisted of a titanium monoaxial AS-ELARIS^® Pedicle Screw adapted in its shape and size to the sheep vertebra. The screw was cannulated with three distal openings. For enhanced fixation, a poly-L-lactide-co-D, L-lactide (PLDLLA) 70:30 (trade name Resomer^® LR 706 S, Evonik Röhm GmbH, Darmstadt, Germany) polymer pin was inserted into the cannulated screw and subsequently ultrasonically melted and liquified (0.104 mL, Fig. 2). The polymer distributed through the three distal openings into the vertebral body in a cloud-like shape and hardened within 3–6 s. CI consisted of the AS-ELARIS^® Pedicle Screw alone without polymer enhancement.

Fig. 2.

AS-ELARIS^® implants and instruments: a Pedicle screw (Ø5.2×25 mm, 2.5 mm cannulation), b Pin (Ø2.4×23 mm, 0.104 mL liquefied)), c Rod (Ø5.5 mm, length 50–70 mm), d Set screw, e Handpiece tip, f Screwdriver, g Pedicle screw with pin in cannulation, h Handpiece

Surgical procedure

A standardized anesthesia protocol was used. Briefly, anesthesia was conducted using xylazine (0.08–1.0 mg/kg i.m., Xylazin^®2%, Streuli Pharma Uznach, Switzerland) and buprenorphine (0.01 mg/kg i.m., Temgesic^®, Essex Chemie AG, Lucerne, Switzerland) for sedation and pre-emptive analgesia, while ketamine (3 mg/kg, i.v., Narketan^®, Vetoquinol AG, Belp-Berne, Switzerland), midazolam (0.1 mg/kg i.v., Synthetica SA, Mendrisio, Switzerland) and propofol (0.4–0.98 mg/kg i.v., Fresenius AG, Stans, Switzerland) were used for induction. All animals’ tracheas were intubated and the lungs ventilated (to normocapnia end-tidal CO₂= 35–45 mmHg) throughout the procedure. Anesthesia was maintained with a balanced anesthetic protocol, employing the administration of isoflurane (1%–3%, Attane™ Isoflurane ad us. Vet., Provet AG, Lyssach, Schweiz) in oxygen and air (to a FiO₂ of 50–70%) via an adult F-circuit, and variable rate infusions of propofol (0.5–1 mg/kg/h) and ketamine (1.2–3 mg/kg/h). Antibiotics, tetanus serum and preemptive analgesia were administered prophylactically prior to surgery.

The anesthetized sheep were placed in sternal recumbency on the operating table. After cleaning the surgical area according to routine, the specific vertebrae were localized under fluoroscopic guidance. A midline skin incision was performed, soft tissue and fascia were incised and paraspinal muscles detached from the bone.

The pedicle screws were inserted through the pedicles of the intervertebral joints on both sides of each vertebra (Fig. 1).

The screw insertion site was prepared using a specifically designed aiming device to guide the correct entrance angle (25–30°), an awl, a probe, a tap and a feeler. The screws were inserted with a DID-4-E digital torque screwdriver. To perform the polymer enhancement process (TI screws only), the AS-ELARIS^® Pin was inserted, a specially developed sonotrode was connected to the screw and the ELARIS^® ultrasonic handpiece was used to perform the ultrasonic liquefaction process (Fig. 2). The two adjacent screws were connected with a rod (5.5 mm), set screws were used to stabilize the construct and simulate the clinical loading situation. The surgical area was closed routinely and a postoperative CT scan was performed. The health status of all animals was checked twice daily postoperatively (appetite, posture, alertness, respiration, weight bearing/gait, pain). Neurological examinations were performed in a standardized manner by a board-certified neurologist. Acute group animals were neurologically examined on postoperative day 1, while long-term animals were examined on postoperative days 1, 4, and 7 and prior to sacrifice. Neurological function was scored using a spinal cord injury score adapted from the Texas Spinal Cord Injury Score for dogs (Fig. S1). Animals not achieving the maximum score on day 1 were re-examined, and additional assessments were performed if neurological deficits persisted beyond 2 days, at the discretion of the neurologist.

Postoperatively, antibiotics and analgesics were administered for maximally 10 days. Food and water were offered ad libitum.

Computed tomography analysis of the lumbar levels

Standardized CT scans (16-slice Siemens Somatom Sensation Open CT scanner with sliding gantry, Siemens Schweiz AG) of the lumbar levels were performed under general anesthesia at different time intervals post-surgery as well as postmortem to assess implant positioning and loosening, hypertrophic ossification and spontaneous facet arthrodesis, as well as polymer visibility and distribution. Timepoints for CT scans were: immediately after surgery, 6, 12, 18 months (in-life) and postmortem (see Table 1). Implant positioning was evaluated as well as bone and screw protrusion ventrolaterally out of the vertebra or into the spinal canal.

Table 1.

Time schedule of CT examinations and fluorescence injections by group

Group	CT-scans					Fluorescence injections
	Post surgery	6 months	12 months	18 months	Post mortem	4 weeks (calcein green)	3 days prior to sacrifice (xylenol orange)	3 days prior to sacrifice (calcein green)
Acute	X				X
8 weeks	X				X	X	X
6 months	X				X			X
12 months	X	X			X			X
24 months	X	X	X	X	X			X

In vivo fluorescence labeling and fluorescence analysis

To track new bone deposition and remodeling kinetics, fluorescent dyes, which co-localize with newly deposited calcium, were injected at different timepoints post-surgery (see Table 1). Native ground sections were evaluated using a fluorescence microscope and a camera (Leica DM6000B, Leica DFC 350 FX, Leica Microsystems CMS GmbH, Mannheim, Germany) to detect fluorescent dye deposition at different timepoints after surgery (see Table 1).

Sacrifice and specimen preparation for evaluative procedures

At the planned sacrifice timepoints (2 days, 8 weeks, 6, 12, 24 months), sheep were slaughtered and the lumbar spine and its local draining lymph nodes (Lnn. lumbales aortici, Lnn. renales) were harvested. Lymph nodes were macroscopically examined and changes in size, color, consistency and any other observations were documented. Subsequently, they were fixed in buffered 4% formalin, dehydrated and embedded in paraffin. Histological sections were cut, stained with hematoxylin/eosin and microscopically examined with special attention to the presence and amount of polymer particles extra- or intracellularly.

The treated spinal area was cleansed from surrounding tissue and macroscopically examined for screw positioning/loosening, metallosis, fibrosis, signs of inflammation and ossification. A CT scan of the lumbar spine was performed with the implants in place. Thereafter, the rods were removed, and the treated vertebrae isolated from each other. Each vertebra was further cleansed and macroscopically examined for hardware failures, screw loosening and hypertrophic bone formation. Two vertebrae per sheep (2 screws per vertebra, each vertebra either CI or TI) were prepared for histology in a standardized manner: Screws meant for removal testing and thin section evaluation were removed in a standardized manner via torque-out testing (first 360° rotation and torque speed of 30° per second) at the Institute for Biomechanics ETHZ, while the other screws intended for histomorphometric and BIC analysis stayed in place.

µCT after torque removal or with screw in place (only 24 months group)

Torque removal tests were performed to assess screw removability in a standardized way to evaluate the risk of trabecular damage or pedicle fracture associated with polymer-enhancement of TI screws. Following screw removal or with screws left in place (in the 24-month group only), vertebrae were fixed in 40% ethanol and analyzed by µCT.

µCT scans were acquired using a Scanco µCT 100 (90 kVp, 200 µA, 500 projections/180°, 600 ms). Three-dimensional datasets were virtually sectioned along the screw axis, and the bone contour width along the screw trajectory was quantitatively evaluated to quantify bone tunnel dimensions. In addition, for the 24 months group, bone tissue was segmented using Gaussian filtering and a threshold of 600 mgHa/cm³, and parameters including bone volume, bone volume over total volume (BV/TV), volumetric Bone Mineral Density (BM/TV), and trabecular structure (number, thickness, separation, and maximum separation) were quantified. This approach enabled assessment of bone quantity, quality, and structural integrity around screws.

Sample processing for histology

Vertebrae were fixed in 40% ethanol, then serially dehydrated in increasing concentrations of ethanol (50–100%) followed by defatting in xylene and infiltration in liquid methylmethacrylate (MMA). Prior to the polymerization, a guiding pin was inserted into the torque-out hole to support alignment during histological sectioning. The polymerized bone blocks were then cut for ground and thin sections.

Two ground sections  (400–800 m, with screws in place) were cut in axial direction through the screw. Then, microradiographs of the ground sections were taken. One native section was used for fluorescence evaluation, and the other one was surface stained with toluidine blue to assess bone healing qualitatively and semi-quantitatively.

Thin sections were prepared at the distal openings, close to the tip of the screws, and cut perpendicular to the screw axis. All thin sections were stained using toluidine blue, von Kossa/Mc Neal and hematoxylin and eosin (HE) for the evaluation of tissue response to the ultrasonic liquefaction process.

Ground sections were captured using a Leica Z6 APOA microscope equipped with a Leica DFC 420 C camera (Glattbrugg, Switzerland) at standardized magnification. Image acquisition was performed using Imagic IMS software (Imagic Bildverarbeitung AG, Opfikon, Switzerland).

Histological evaluation

Quantitative histomorphometry

Quantitative histomorphometry was performed using a standardized pixel-detecting tool in Adobe Photoshop Elements 2019 (Adobe Systems, San Jose, CA) and the Fiji image processing package (Image J, University of Wisconsin). Pixels within the region of interest (ROI), excluding background and screw, were defined as 100%, and the relative percentages of different tissue types were quantified. Two ROIs were analyzed: one adjacent to the implant and one located more distally. A third ROI included the total area combining the two previous ROIs (Table S1). All ROIs included the proximal part of the screw as well as the distal part with the polymer extruding area. The region at the screw tip was excluded from the analysis due to the presence of bone debris resulting from drilling and screw insertion, which is associated with increased remodeling activity. In shorter survival groups (8 weeks to 6 months) both, new and old bone could be distinguished. In contrast, for longer survival times (12 and 24 months), only bone in total (new and old combined) was evaluated due to advanced remodeling, which made differentiation between new and old bone impossible. (Fig. S2)

Bone to implant contact area evaluation (BIC)

BIC was performed for all stained ground sections of the four osseointegration groups (8 weeks, 6, 12 and 24 months). Measurements of the BIC included only the proximal part of the screw up to the extrusion hole. The BIC percentage was calculated as the ratio of the implant surface in direct bone-to-implant contact to the total implant surface measured. The evaluation differentiated between cancellous and cortical bone.

Qualitative and semi-quantitative analysis of local tissue effects

Assessment of bone activity parameters, polymer distribution and interaction with bone

For TI, bone activity parameters and polymer interaction with bone were evaluated in the polymer infiltration zone and at the edge of the polymer to the surrounding bone.

The analysis for the CI implants was performed at the level of the distal holes at the zone of implant surface (See Table S1).

Assessment of biocompatibility parameters

Semi-quantitative scoring criteria in accordance with ISO 10993-6 guidelines were applied to assess the local tissue response (biocompatibility) in the peri-implant area (Table S1) using light microscopy (Leica^® DMR system equipped with a DFC320 camera; Leica Microsystems, Switzerland). For each time point, scores for TI and CI were summed, averaged, and the mean CI score was subtracted from the mean TI score to determine the severity grade of tissue reaction as minimal or no reaction (0.0–2.9), slight reaction (3.0–8.9), moderate reaction (9.0–15.0), and severe reaction (≥ 15.1).

Statistical analysis

TI and CI were compared at five time points using a significance level of p< 0.05. Histomorphometric data, bone-implant contact (BIC), µCT-based screw hole widening, and µCT evaluation of the 24 months group for BV/TV and BM/TV were analyzed using paired t-tests or Wilcoxon signed-rank tests, depending on data distribution. Normality of within-animal differences was assessed with the Shapiro–Wilk test (IBM SPSS Statistics).

Results

Surgical procedure

A total of 320 screws (160 TI, 160 CI) were implanted in 40 sheep included in the group analysis. Ultrasonic liquefaction of TI screws was successful in 159/160 implantations. In one instance, a defective sonotrode caused a small amount of polymer to remain on the device, although the majority of the material was delivered as intended. All animals recovered from surgery. Thirteen sheep showed transient mild-to-moderate postoperative neurological deficits attributed to prolonged surgery or procedural complications (e.g. hemorrhage, screw misplacement or tissue damage), but all recovered fully within 2 to 28 days of follow-up (Tables S2 and S3).

CT evaluations of lumbar spine at different timepoints

Screw positioning Post-surgical CT scans demonstrated a critical relation of screw diameter compared to the size of pedicle and vertebral body so that already minimal deviations in positioning led to slightly misplaced screws. 135/320 screw positions (42.2%) were found to be not completely within the vertebral body (tip of screw slightly out of vertebral body, screw thread leading to bone protrusion into the spinal canal or screw thread slightly visible in spinal canal, details see macroscopic evaluation at sacrifice).

Screw loosening (Table S3) In 35/40 animals, CT scans revealed no radiolucent lines or halos around the screws at any time point. In 5/40 animals, radiolucent lines of variable extent were detected around a total of 22 screws. Three animals belonged to the 6 months (n = 14 screws, TI = 7, CI = 7) and two to the 12 months group (n = 8 screws, TI = 8) (Fig. 3). Loosening predominantly affected screws implanted at caudal vertebral levels (19/22) and/or screws that were malpositioned with an incorrect entrance angle (14/22).

Fig. 3.

Example CT images for early loosened TI screws in the 12 months group, L4, incompletely inserted right screw may have led to lever effect; a CT post-surgery; b 6 months post-surgery; c at sacrifice, 12 months after surgery

Vertebral fusion evaluated at time of sacrifice No radiological fusions of the instrumented levels were detected in the acute and 8 weeks groups, although facet joint degeneration and hypertrophic ossifications were evident at 8 weeks. In the 6 months group, radiological evidence of bilateral fusion was present in 9/10 sheep (19/20 instrumented levels), with one unfused cranial level (CI). In the 12 months group, 6/8 sheep demonstrated radiological fusion (14/16 levels). In two sheep with loosened screws (TI), the caudal levels displayed facet joint lysis consistent with pseudarthrosis. In the 24 months group, all instrumented levels (12/12) showed radiological fusion.

Polymer distribution Polymer distribution could only be visualized indirectly on CT due to the intrinsic radiolucency of the material, identifiable solely by altered radiodensity in comparison to bone. Polymer extrusion out of the vertebral body (14/160, 8.7%) or into the spinal canal (3/160, 1.9%) could be localized in CT images only when associated with compression of adjacent tissues. Radiodensity of the polymer infiltration zone changed over time: Immediately postoperatively, the polymer in the infiltration zone appeared as a cloud-like area of reduced radiodensity surrounded by a radiodense rim. At 8 weeks, the polymer displayed radiodensity similar to mineralized bone, still bordered by a radiodense rim (Fig. 4a–b). At the 6 and 12 months, radiodensity of the infiltration zone decreased slightly, while the surrounding rim increased in both diameter and density. At 18 and 24 months (Fig. 4c), the infiltration zone appeared homogeneously radiodense, exceeding the density of the adjacent bone.

Fig. 4.

Representative CT images after sacrifice, a–b: L4, TI, 8 weeks group; a Screws with extrusion hole areas visible; b Polymer visible with a radiodense rim around the infiltration zone, with the polymer infiltration zone showing the same density as mineralized bone; c L4, TI, 24 months group: Homogeneous radiodense polymer infiltration zone without rim

Macroscopic evaluation of implantation sites and surrounding tissue at sacrifice

At sacrifice, only animals in the acute group showed minimal hematoma and edema at the surgical site, no inflammatory reactions were observed in the osseointegration groups. Screw loosening, assessed by manipulating the screw head after rod removal, was detected in 22/320 (6.9%) screws, consistent with CT findings. No clinical or neurological symptoms were noted in these animals during twice daily examinations and neurological examination prior to sacrifice(Tables S2 and S3).

No metallosis was observed in the acute or 8 weeks groups. Mild metallosis was detected in 17/320 screws, distributed across the 6 months (n=6), 12 months (n=2) and 24 months groups (n=9). Affected sites included set screws (n=7, TI:4, CI:3) and the rod junctions (n=10, TI:5, CI:5), and no associated adverse tissue reactions were observed. Postmortem macroscopical evaluation of screw positioning and intervertebral fusions was consistent with CT results.

Macroscopic evaluation of screw positioning confirmed the CT findings, with 72/160 TI and 63/160 CI screws slightly malpositioned. Slight ventrolateral protrusion of the screw tip occurred in 28/160 TI and 24/160 CI screws, while in 44/160 TI and 39/160 CI screws, either bone or small portions of the screw threads breached the spinal canal to varying degrees. In all cases where screw threads entered the spinal canal, the integrity of the dura mater and spinal cord was preserved. Correctly positioned screws within the vertebral body showed no polymer migration. Polymer remained confined to the vertebral body in 146/160 TI implants, whereas in 11/160 TI implants, small amounts of polymer protruded ventrolaterally out of the vertebra, and in 3/160 TI implants, into the spinal canal. In all cases, the polymer remained attached and restricted to the malpositioned screws (Table 2).

Table 2.

Macroscopic examination of the implantation sites at sacrifice. Fusion of instrumented levels is reported per sheep, not per implant

Survival time	Acute		8 weeks		6 months		12 months		24 months		All groups included in the group analysis
Group	TI (n = 32)	CI (n = 32)	TI (n = 32)	CI (n = 32)	TI (n = 40)	CI (n = 40)	TI (n = 32)	CI (n = 32)	TI (n = 24)	CI (n = 24)	TI (n = 160)	CI (n = 160)	Combined (n = 320)
Bone protrusion (ventrolaterally), without cortical breach of screw tip	1	0	3	6	0	4	2	1	3	3	9 (2.8%)	14 (4.4%)	23 (7.2%)
Screw tip slightly out of vertebra (ventrolaterally)	8	2	1	2	3	1	0	0	7	5	19 (5.9%)	10 (3.2%)	29 (9.1%)
Bone protrusion without cortical breach of screw thread (spinal canal)	3	5	10	3	7	7	4	1	2	4	26 (8.1%)	20 (6.3%)	46 (14.4%)
Screw thread slightly visible (spinal canal)	3	3	0	2	8	7	2	1	5	6	18 (5.6%)	19 (6%)	37 (11.6%)
Polymer (spinal canal)	0		0		3		0		0		3 (1.9%)
Polymer (ventrolaterally out)	6		1		2		0		2		11 (6.9%)
Sum of screw misplacements	15	10	14	13	18	19	8	3	17	18	72 (22.5%)	63 (19.7%)	135 (42.2%)
Screw loosening (macroscopically & clear halo formation in CT)	0	0	0	0	7	7	8	0	0	0	15 (4.7%)	7 (2.2%)	22 (6.9%)
Metallosis	0	0	0	0	5	1	1	1	3	6	9 (2.8%)	8 (2.5%)	17 (5.3%)
Fusion cranial level (left and right)	0	0	0	0	10	9	8	8	6	6	24 (7.5%)	23 (7.2%)	47 (14.7%)
Fusion caudal level (left and right)	0	0	0	0	10	10	4	8	6	6	20 (7%)	24 (6.8%)	44 (13.8%)

µCT evaluation of the samples after screw removal (torque test)

µCT scans were performed on 74/80 vertebrae after standardized screw removal using torque removal tests; in 6/80, standardized torque removal could not be performed due to loosened screws. In these cases, µCT evaluations could not be performed as planned and these screws were excluded of the analysis. Overall, all screws could be easily removed, with no pedicle or vertebral fractures and no widening of the screw holes in the polymer zone in any of the five TI survival groups. Only at 24 months a statistically significant difference was found between TI and CI, showing a slight distal screw hole widening for CI (p= 0.006, Table 3).

Table 3.

µCT evaluation of screw hole diameters after torque removal: Statistically significant differences (paired-samples t-tests) between TI and CI are indicated by an asterisk and were observed only in the 24 months group for the difference between proximal and distal screw hole, showing a slight distal screw hole widening for CI (p= 0.006)

Group	n	TI		CI		TI	CI
		Detail proximal Ø [mm] [mean ± STD]	Detail distal Ø [mm] [mean ± STD]	Detail proximal Ø [mm] [mean ± STD]	Detail distal Ø [mm] [mean ± STD]	Difference proximal-distal [mean ± STD]	Difference proximal-distal [mean ± STD]	p-value
Acute	8	5.20	5.13	5.29	5.24	0.07	0.05	0.825
		± 0.13	± 0.31	± 0.05	± 0.10	± 0.19	± 0.11
8 weeks	8	5.29	5.26	5.28	5.24	0.02	0.04	0.709
		± 0.11	± 0.03	± 0.06	± 0.06	± 0.11	± 0.08
6 months	7	5.75	5.23	5.71	5.13	0.51	0.58	0.833
		± 0.53	± 0.08	± 0.25	± 0.26	± 0.48	± 0.47
12 months	7	5.30	5.19	5.24	5.22	0.11	0.02	0.198
		± 0.11	± 0.14	± 0.03	± 0.04	± 0.20	± 0.04
24 months	6	5.25 ± 0.02	5.21 ± 0.03	5.24 ± 0.03	5.25 ± 0.02	0.03 ± 0.03	-0.01 ± 0.03	0.006*

In the acute group, statistical analysis was performed using a Student’s t-test for independent samples, as only TI or CI screws were implanted

Additional µCT evaluation for long-term group

Twelve vertebrae (n= 6 TI and 6 CI) of the long-term group, evaluated with screws in place, demonstrated good bone-implant contact and peri-implant bone densification. Paired t-test analysis revealed significantly higher bone volume fraction (BV/TV; p= 0.024; TI = 0.49 ± 0.07; CI = 0.40 ± 0.03) and volumetric bone mineral density (BM/TV; p = 0.02; TI = 578.98 ± 53.73 mgHA/cm³; CI = 511.11 ± 30.95 mgHA/cm³) around the extrusion holes of TI screws compared with CI screws,.

Histological evaluation

Histomorphometric measurements of ground sections

Histomorphometric measurements showed good osseointegration either along the polymer surface (TI) or directly at the metal (CI). Due to the polymer infiltration into the bone, TI showed an extended implant surface. In the 24 months group, the polymer infiltration zone was filled with new bone indicating an advanced degradation process of the polymer. Statistically significant differences were only found for the 8 weeks group using a paired T-test. Mean values for CI were significantly higher for new bone (CI: 53.44%, TI: 45.5%, p = 0.002) as well as for bone combined (CI: 71.22%, TI: 65.57%, p= 0.004) both close to the implant, whilst significantly higher values were found for TI for non-bone structures close to the implant (TI: 34.43%, CI: 28.78%, p = 0.004) as well as for new bone formation distant to the implant (TI: 17.86%, CI: 11.88%, p = 0.044) (Fig. 5).

Fig. 5.

Graph and table below show mean ± SD values (as a percentage of the total area measured) of histomorphometric measurements of old bone, new bone, both combined and non-bone structures for TI and CI at 8 weeks, 6, 12 and 24 months post-surgery (i = area close to implant, s = distant to implant, total = i and s combined). Significant differences between TI and CI are indicated with asterisks and were found for the 8 weeks group close to the implant for new bone (p = 0.002, red asterisk) and bone combined (p = 0.004, blue asterisk). Further, statistically significant differences with higher values for TI were found for the 8 weeks group for non-bone close to the implant (p = 0.004, green asterisk) and new bone distant to the implant (p = 0.044, purple asterisk), using a paired-samples t test. In the 12 months group, bone combined and non-bone values close to the implant were compared using a Wilcoxon signed-rank test for related samples; median and interquartile range (IQR) values are presented in brackets below the mean ± SD values

Bone to implant contact area (BIC)

BIC differed between cortical and cancellous bone. Cortical bone showed higher values at 8 weeks (TI, CI), decreased at 6 and 12 months, and peaked again at 24 months (highest in TI). In cancellous bone, BIC declined slightly from 8 weeks to 6 months (TI, CI) and reached maximum values at 24 months (TI). A statistically significant difference was observed only in cortical bone at the 24 months, with higher values for TI (TI: 59.8%, CI: 37.3%, p=0.02, Fig. 6).

Fig. 6.

Graph and table below show mean ± SD values for BIC measurements of cortical, cancellous, and total bone for TI and CI at 8 weeks, 6, 12, and 24 months post-surgery. Statistically significant differences between TI and CI are indicated by an asterisk and were observed only for cortical bone at 24 months, with higher values for TI (p=0.021, paired-samples t-test). At 8 weeks, total bone values were compared using a Wilcoxon signed-rank test for related samples; median and interquartile range (IQR) values are presented in brackets below the mean ± SD values

Fluorescence evaluation

At 4 weeks, calcein green was visible in both TI and CI, but more pronounced in TI at the surface of the polymer infiltration zone (Fig. 7). At 8 weeks, xylenol orange was more intense in the implant surrounding in TI, indicating a higher rate of calcium deposition (bone remodeling) (Fig. 7). At 6 and 12 months (TI, CI) and 24 months (CI), a gradually decreasing remodeling activity in the cancellous bone area resulted in only slight fluorescence. In contrast, TI at 24 months showed marked calcein green deposition in the infiltration zone, suggesting high remodeling with concurrent polymer degradation (Fig. 8).

Fig. 7.

a–d: L4 left, TI, 8 weeks group; a Microradiograph image showing bone-to-implant contact area and polymer infiltration zone (red ring); b–d Fluorescence images showing b Calcein green deposition at 4 weeks post-surgery, with dark area (red arrow corresponding to polymer infiltration zone, c Enhanced xylenol deposition at 8 weeks post-surgery particularly in the polymer infiltration zone d Merged image of both fluorescence timepoints e–h: L1 left, CI, 8 weeks group; e Microradiograph image showing bone-to-implant contact area; f–h Fluorescence images showing (f) Calcein green deposition at 4 weeks post-surgery; g Xylenol deposition at 8 weeks post-surgery; h Merged image of both fluorescence timepoints

Fig. 8.

a–b: L4 left, TI, 24 months group; c–d: L4 left, CI, 24 months group; a Microradiograph image showing good bone-to-implant contact with dense bone structure both in direct screw surrounding and within the polymer infiltration zone; b Fluorescence image showing increased calcein green deposition in the polymer infiltration zone, indicating new bone formation at 24 months post-surgery; c Microradiograph image showing good bone-to-implant contact with dense bone structure in direct screw surrounding; d Fluorescence image showing only slight calcein green deposition in screw surrounding at 24 months post-surgery

Qualitative and semi-quantitative analysis of the local tissue effects

Assessment of bone activity parameters, polymer distribution and interaction with bone

In total, 80 toluidine blue stained histological ground sections of all five timepoints were analyzed for bone activity and polymer interaction (TI, CI).

Trabecular thickness was normal for TI and CI in the acute group. At 8 weeks, in both TI and CI, enhanced bone remodeling with thickened trabeculae was observed at the tip of the screw and in direct surrounding of the screw surface. Between 8 weeks and 12 months, 18 TI implants showed a small number of slightly thinned trabeculae when completely embedded in polymer. No thinned trabeculae were observed outside or at the margin of the polymer infiltration zone (Fig. 9). At 24 months, trabecular thickness in the screw surrounding increased in both groups. In all TI samples, new bone formation and thickened trabeculae were found within the polymer infiltration zone, with only traces of polymer residuals left (Fig. 9).

Fig. 9.

Representative histologic images showing toluidine blue stained ground sections, scale bars indicating 5 mm and 2 mm; a L2 right, TI, 12 months group: Extrusion hole with polymer infiltration zones on both side of the screw (marked with red line), within the polymer zone, thinned trabeculae with isolated fragments are visible. A moderate degree of bone resorption is present within polymer zone, while normal bone structure is preserved at the margin around the infiltration zone, (BM=bone marrow). b L1 left, TI, 24 months group: Polymer infiltration zones at the distal holes, no polymer detectable. New bone formation with thickened trabeculae (dark blue) is observed

Isolated fragments of trabecular bone In the acute group, a normal amount of bone debris was present at the screw surface in both TI and CI. At later timepoints bone debris was remodeled in both groups. In TI, fragments embedded in polymer persisted up to 12 months but were absent at 24 months.

Bone resorption Within the polymer infiltration zone, bone resorption was only observed in trabeculae or bone fragments that were completely embedded in the polymer. The extent was very mild and showed no association with screw loosening. No bone resorption was detected at the direct polymer surface. A slight increase of bone resorption within the infiltration zone was noted up to 12 months, whereas at 24 months bone remodeling and new bone formation predominated.

Polymer distribution In all samples of the acute, 8 weeks, 6- and 12-months groups, the polymer remained histologically visible and was not yet fully degraded. In contrast, in all 24- month samples, only small traces of polymer residuals were detectable, which had been replaced by new bone without evidence of inflammatory tissue reaction. In 4/40 samples with slightly misplaced screws (8 weeks n= 1, 6 months n= 2, 24 months n= 1), polymer extended beyond the trabecular bone, being additionally observed either within the spinal canal (n= 1) or ventrolaterally out of the vertebral body (n= 3) while being firmly attached and spatially confined to misplaced screws. Around the polymer, no or only minimal fibrous tissue formation was noted. No inflammatory reactions were detected (Fig. 10).

Fig. 10.

Representative histologic images showing toluidine blue stained ground sections in low a, medium b and high c magnification, with scale bars indicating 1 cm, 2 mm, and 500 m, respectively; L5 right, TI, 6 months group; a Overview image showing screw and polymer (P) extending into the spinal canal (SK), (VB=vertebral body); b Detail image illustrating some new bone formation on the polymer surface and fibrous tissue without visible inflammatory cells; c High magnification image showing polymer material in spinal canal surrounded by fibrous tissue (F), no signs of inflammatory reaction are visible

Biocompatibility assessment

A total of 80 histological sections (hematoxylin-eosin, toluidine blue, and von Kossa; 240 slides) from five timepoints and two implant types (TI and CI) were evaluated. For detailed results see Table 4.

Table 4.

Results of semi-quantitative histology scoring in all groups (thin sections) according to ISO-Norm 10993-6:2016(E)

Group	Acute		8 weeks		6 months		12 months		24 months
	TI (n = 8)	CI (n = 8)	TI (n = 8)	CI (n = 8)	TI (n = 10)	CI (n = 10)	TI (n = 8)	CI (n = 8)	TI (n = 6)	CI (n = 6)
Polymorphnuclear cells	8	8	0	0	0	0	0	1	0	0
Eosinophils	1	1	0	0	0	0	0	0	0	0
Lymphocytes	7	6	3	2	4	3	3	1	0	0
Plasma cells	1	2	0	0	4	2	2	1	0	0
Macrophages	9	10	8	8	6	3	10	1	1	0
Giant cells	0	0	0	0	4	0	11	0	6	0
Necrosis	0	0	0	0	0	0	0	0	0	0
Subtotal x2	52	54	22	20	36	16	52	8	14	0
Neovascularization	0	0	4	1	6	4	12	1	1	0
Fibrosis	0	0	6	4	8	6	13	1	1	0
Fatty infiltration	0	0	4	2	0	0	3	0	0	0
Total score	52	54	36	27	50	26	80	10	16	0
Average score (Ø)	6.5	6.75	4.5	3.38	5	2.6	10	1.25	2.67	0
Ø TI – Ø CI	−0.25=0		1.1		2.4		8.75		2.67

For each time point, the sum of scores for each parameter per group (TI, CI) is illustrated, scores for inflammatory cells and necrosis are multiplied by a factor of 2 due to their greater importance. The total scores per group are averaged, and the mean CI score is subtracted from the mean TI score to determine the severity grade of tissue reaction: minimal or no reaction (0.0–2.9), slight reaction (3.0–8.9), moderate reaction (9.0–15.0), and severe reaction (≥15.1)

Acute group Inflammatory cells (polymorphonuclear cells, lymphocytes, macrophages, plasma cells) were found in low numbers equally distributed in both TI and CI, reflecting a physiological acute tissue response to surgery. No polymer fragmentation, multinucleated giant cells or necrosis were observed. In the acute group TI was considered to show no reaction (ØTI–ØCI=0) to the tissue compared to CI.

8 weeks group The tissue response consisted mainly of macrophages and lymphocytes (lower number than in the acute group) indicating a chronic remodeling response. Neither polymer fragmentation (TI) nor foreign body giant cells (TI and CI) were detected, and no necrosis (TI and CI) was observed in any sample of either group (Fig. 11a–b). Mild tissue remodeling with slight fibrosis (TI: 6/8, CI: 4/8), minimal neovascularization (TI: 4/8, CI: 1/8), and fatty infiltration (TI: 4/8, CI: 2/8) were observed, corresponding to score 1. TI demonstrated a minimally higher tissue reaction compared to CI (ØTI–ØCI=1.1).

Fig. 11.

Representative histologic images showing hematoxylin eosin stained thin sections, scale bars indicating 1 mm, 8 weeks group, a L1 left, CI: Screw hole (S) surrounded by remodeled bone debris. A physiological number of fat and inflammatory cells is present within trabecular space; b L4 left, TI: Screw hole (S) surrounded by normal trabecular bone. Polymer (P, outlined in black) is distributed within the trabecular space. A physiological number of fat and inflammatory cells is present

6 months group In 2/10 TI samples, some polymer fragmentation with multinucleated giant cell reaction was observed, associated with abundant fibrosis due to loosened screws. In these samples and in 2/10 loosened CI samples, macrophages, lymphocytes, and plasma cells, indicated a low-to-moderate chronic remodeling response (biodegradation/remodeling, screw loosening). Overall, TI showed a minimally higher tissue reaction compared to CI (ØTI–ØCI = 2.4).

12 months group TI samples exhibited higher score levels than CI. All TI samples (8/8) contained polymer degradation debris, both cell-associated and free, with macrophages and multinucleated giant cells present but no elevated other inflammatory cells, indicating normal progression of biodegradation. A thin fibrovascular layer surrounded the polymer residuals in all TI samples (Fig. 12a–b). In CI, only 1/8 samples contained a low number of lymphocytes, plasma cells, and macrophages (Fig. 12c). TI demonstrated a slightly higher tissue reaction compared to CI (ØTI–ØCI=8.75).

Fig. 12.

Representative histologic images showing hematoxylin eosin stained thin sections in low a, c and high magnification b, with scale bars indicating 1 mm a, c and 100 µm b, respectively;12 months group, a L1 right, TI: Screw hole (S) surrounded by normal trabecular bone; polymer (P) is distributed within trabecular space, adjacent to a physiological number of fat and inflammatory cells at 12 months post-surgery. b L1 right, TI: Higher magnification view showing polymer (P) within the trabecular space. A thin fibrovascular layer surrounds the polymer, along with some multinucleated foreign body giant cells (GC), indicating normal polymer biodegradation. The polymer is embedded in normal trabecular bone, with a physiological number of fat and bone marrow cells and no signs of inflammation. c L5 left, CI: Screw hole (S) surrounded by remodeled bone debris; trabecular space contains a physiological number of fat and bone marrow cells at 12 months post-surgery

24 months group A physiological amount of bone marrow and fat cells was found in both groups (Fig. 13 a and b). No increase of inflammatory cells was detected. Focal neovascularization (1/6 TI), fibrosis (1/6 TI) and traces of partially fragmented polymer residuals with mild phagocytic cell reaction (6/6 TI) were found only in TI (Fig. 13c). No necrosis was observed. Polymer residuals were almost completely replaced by new bone (Fig. 13b). Overall, TI showed a minimally higher tissue reaction as compared to CI (ØTI–ØCI = 2.67).

Fig. 13.

Representative histologic images showing hematoxylin eosin stained thin sections in medium magnification, with scale bars indicating 5 mm a, b and toluidine blue thin section in high magnification, with scales bar indicating 100 µm c 24 months, a L5 left, CI: Screw hole (S) surrounded by remodeled bone debris. A physiological number of fat and bone marrow cells is present within the trabecular space; b L5 left, TI: Screw hole (S) surrounded by remodeled bone within the polymer infiltration zone which is dense. Adjacent trabecular space shows a physiological number of fat and bone marrow cells; c L1 left, TI: Polymer within the trabecular space (P). Multinucleated foreign body giant cells (GC) are visible. Polymer is surrounded by normal trabecular bone with a physiological number of fat and bone marrow cells and no signs of inflammation

Analysis of the draining lymph nodes

79/80 draining lymph nodes were harvested and analyzed macroscopically and histologically. For one animal, the renal lymph node was not harvested because the adrenal gland was accidentally taken instead of the lymph node. No foreign material (polymer residuals) was detected in any of the examined lymph nodes.

Discussion

The results of this study provide important safety and efficacy data regarding the AS-ELARIS^® pedicle screws with ultrasonic polymer enhancement in sheep. The chosen animal model enabled detailed evaluation of osseointegration, thermal effects, polymer extrusion, screw removal, and tissue reactions.

Although the human ELARIS^® pedicle screw was adapted for sheep (AS-ELARIS^®), its diameter remained relatively large for the slim, hourglass-shaped vertebrae of the sheep [26], leaving narrow bone margins around the implant. Further diameter reduction would have hindered polymer insertion and extrusion of volumes similar to human applications. Due to this unfavorable diameter to bone size relationship, minor variations in entry angle or position caused misplacements ventrolaterally out of the vertebra or into the spinal canal (42.2% overall; TI = 22.5%, CI = 19.7%), but even in these worst-case scenarios, screw or polymer protrusion did not cause clinical symptoms, abnormal tissue reactions, or adverse effects on animal wellbeing, as long as spinal canal compromise did not produce mechanical stenosis. Histological analysis at 6 months revealed small amounts of polymer in the spinal canal in a single sample due to severe screw misplacement, without evidence of an inflammatory response and only minimal fibrous tissue formation. In humans, such misplacement of screws is expected to be unlikely due to a more favorable screw-to-bone ratio and the consequent positioning of extrusion holes away from the spinal canal and other sensitive structures. Moreover, correct screw positioning is mandatorily verified using fluoroscopy prior to polymer enhancement.

Fluorescence labeling demonstrated good osseointegration for both TI and CI. Calcium deposition appeared at 4–8 weeks, more pronounced in TI and extending into the polymer zone by 8 weeks. At 24 months, calcein green deposition indicated ongoing bone remodeling with only minimal polymer residuals remaining. Histomorphometry and BIC analyses confirmed higher early remodeling and expanded implant surfaces for TI until 12 months. By 24 months, the polymer was almost completely degraded and replaced by new bone, with no significant differences for cancellous BIC (82.15%) for TI compared to CI (73,48%, p = 0.44). Cortical BIC peaked at 8 weeks for both, TI and CI, decreased at 6–12 months, and increased again at 24 months with significantly higher values for TI when compared to CI (TI: 59.8%, CI: 37.3%, p = 0.02), consistent with cortical remodeling. These results are consistent with the histological evaluation of the screw hole widening, which occurs primarily at the cortical bone. However, according to Defino et al., cancellous bone has the highest influence on implant stability [27] and therefore, TI showed superior results compared to CI at 24 months.

Screw loosening occurred in 5/40 animals (22/320 screws), mainly at 6–12 months, without clinical symptoms in these otherwise healthy sheep. Loosening may result from screw misplacement, pedicle-screw diameter mismatch, fusion level, number of fused levels [10], insertion angle [11, 28], or inhibited neovascularization due to excessive motion [29, 30]. These previously reported factors were consistent with the loosened screws observed in this study. A limitation of this study is its transferability to humans regarding pain assessment in sheep. Since pain evaluation is more difficult in sheep [31], it is not possible to say whether humans would have experienced pain in a clinical trial involving loosened screws. Bone activity analysis revealed enhanced remodeling around the screw tip (CI and TI), in direct surrounding of the screw surface (CI) and at the expanded implant surface of the polymer infiltration zone (TI) at 8 weeks, with thickened trabeculae and no evidence of thinning or fragmentation in surrounding bone. Mild bone resorption occurred only in trabeculae fully embedded in polymer, likely due to reduced perfusion, but did not provoke screw loosening or bone voids.

In 1966, Kulkarni et al. first demonstrated the good biocompatibility of poly-lactide polymers [32]. Since then, numerous studies have investigated a variety of polymers - including polylactides, polylactide composites, polyglycolides, and polymer-ceramic blends - reporting a wide range of outcomes [33–37].

In the present study, histological evaluation confirmed that both the ultrasonic liquefaction process and the applied polymer exhibited good biocompatibility, comparable to CI at all five timepoints. Even polymer extrusion outside the vertebral body did not trigger an inflammatory response. Importantly, no acute inflammatory cells or eosinophils were detected at 8 weeks, 6, 12, or 24 months postoperatively. At 8 weeks, small numbers of lymphocytes and macrophages were present, consistent with their role in chemokine production and the initiation of tissue remodeling. The reduced amounts of fibrous tissue, fatty infiltration, and neovascularization at this stage may reflect a transient scaffold for mesenchymal cell differentiation into osteoblasts and adipocytes, thereby supporting subsequent formation of new bone, bone marrow, and vessels. Even at the onset of polymer degradation, only a slight increase in phagocytic cells was observed, without void formation.

Polymer degradation occurred concurrently with replacement by new bone. At 24 months, when only traces of polymer remained, the polymer infiltration zone was remodeled with higher bone turnover, a greater proportion of bone adjacent to the implant, and reduced non-bone tissue compared to CI. As a result, no foreign material persisted within the vertebral bodies, representing a distinct advantage over non-absorbable bone cement. These findings are consistent with the 24-month µCT results, which demonstrated a significant increase in BV/TV and BM/TV around TI compared to CI screws. Importantly, histology confirmed that despite this increase in bone density, the bone maintained its cancellous structure, and no sclerosis was detected.

In line with earlier studies [23, 24], the ultrasonic liquefaction process proved to be safe with respect to thermal impact during polymer melting, producing no histologically detectable tissue damage. This represents a clear advantage over bone cement, where local temperatures exceeding 50 °C have been reported [18, 38]. In addition, µCT analysis demonstrated that removal of the polymer enhanced screws using torque-out tests did not cause trabecular damage or pedicle fracture. The polymer broke at the screw surface close to the extrusion holes and remained firmly adhered to bone, thereby enabling revision surgery and safe screw removal without compromising the trabecular structure at any timepoint investigated. In contrast, removal of expandable or cement-augmented pedicle screws has been associated with trabecular breakage and widening of the screw hole due to bone-cement remnants within the threads [14].

Conclusion

Ultrasonic liquefaction-mediated, polymer-enhanced AS-ELARIS^® Pedicle Screws seems to be safe at all timepoints evaluated, providing mechanical stability until solid vertebral fusion and osseointegration were achieved. No bone void formation was observed following polymer resorption. This technology appears to be a valuable advancement in pedicle screw fixation in sheep, offering not only biocompatibility, biodegradability, and straightforward removal, but also reliable and user-friendly handling.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

BIC

Bone to implant contact

BM/TV

Bone mineral density/Total volume

BV/TV

Bone volume / Total volume

CI

Control item

CT

Computed tomography

HE

Hematoxylin/Eosin

i.m.

intramuscular

i.v.

intravenously

LBP

Low back pain

MMA

Methylmethacrylate

PMMA

Polymethylmethacrylate

ROI

Region of interest

TI

Test item

µCT

Micro computed tomography

Author contributions

IH managed the data summary and was critical in drafting the manuscript. AK performed histological assessment. PK designed and supervised the analgesia and anesthesia protocols for the study, provided clinical support, revised the resubmitted manuscript. BvR provided clinical advice and support, revised the manuscript critically, approved the submitted and final manuscript versions. KKL designed the in vivo study, performed surgeries, radiographic, macroscopic, histological and histomorphometric assessments, provided clinical supervision, evaluated and analyzed data, and was critical in drafting the manuscript. All authors have read and approved the final manuscript.

Funding

The pre-clinical in vivo study was funded by SpineWelding AG, Wagistrasse 6, CH-8952 Schlieren, Switzerland.

Data availability

The datasets generated and/or analyzed during the current study are not publicly available due to confidentiality requested by the study sponsor, but are archived at the corresponding author’s institution (both paper and electronic raw data as well as materials e.g. histology blocks and slides) and are available from the corresponding author on reasonable request.

Declarations

Ethics approval

All animal experiments were conducted according to the Swiss laws of animal protection and welfare and were approved by the local governmental veterinary authorities (license number: ZH013/16).

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.