Clinical and Biomechanical Characteristics of Posterior-Stabilized Polyethylene Post Fractures in Total Knee Arthroplasty: A Retrieval Analysis

Abstract

Background:

Fractures of the polyethylene post are a rare but known complication after posterior-stabilized (PS) total knee arthroplasty (TKA). We evaluated the polyethylene and patient characteristics for 33 primary PS polyethylene components that were revised with fractured posts.

Methods:

We identified 33 PS inserts revised between 2015 and 2022. Patient characteristics collected included age at index TKA, sex, body mass index, length of implantation (LOI), and patient-reported details on events surrounding the post fracture. Implant characteristics recorded were manufacturer, cross-linking properties (highly cross-linked polyethylene [XLPE] versus ultra-high molecular weight polyethylene [UHMWPE]), wear characteristics based on subjective scoring of the articular surfaces and scanning electron microscopy (SEM) of fracture surfaces. Mean age at index surgery was 55 years (range, 35 to 69), mean body mass index was 29.5 (range, 18.5 to 37.2), and mean LOI was 10.0 (range, 4 to 26).

Results:

Total surface damage scores were significantly higher in the UHMWPE group versus the XLPE group (57.3 versus 44.2, P = .003). SEM demonstrated fracture initiation at the posterior edge of the post in 10 of 13 cases. UHMWPE fracture surfaces posts had more tufted, irregularly clamshell features, while XLPE posts had more precise clamshell marking and a diamond pattern in the region of acute, final fracture.

Conclusion:

Characteristics of PS post fracture differed between XLPE and UHMWPE implants, with fractures occurring in the XLPE with less generalized surface damage, after a shorter LOI, and with SEM evaluation indicative of a more brittle fracture pattern.

Fractures of the polyethylene post are a rare but known complication after posterior-stabilized (PS) total knee arthroplasty (TKA). Although the reported rate of post fractures after primary PS TKA is typically reported to be <1% [1,2], this problem necessitates return to the operating room for revision surgery, causing morbidity for the patient and potential for subsequent complications. Thus, understanding the factors that cause post fractures could minimize the risk of this complication after TKA.

Prior literature on PS polyethylene post fractures predominately consists of case reports of fractures among specific designs. Authors pointed out specific material characteristics such as highly cross-linked polyethylene (XLPE) [2] or in vivo oxidation of inserts that could have contributed to the fracturing of polyethylene posts. Other authors suggested that post fractures occur due to subtle malposition of TKA components, which can lead to cam-post impingement or flexion instability, both of which place added stress on the polyethylene post [3,4]. It is understood that post fractures can occur as an isolated event [2] or as a late sequelae of severe post wear over time [5]. Also, material type may play a role as XLPE components have shown a higher susceptibility to fracture [6,7] when compared to ultra-high molecular weight polyethylene (UHMWPE) components.

In this study, we examined a large collection of 33 PS polyethylene retrieved tibial components from primary TKAs that had failed due to a fractured post. Our goal was to examine the roles that implant and patient characteristics played in these fractures. We specifically examined the differences in patients who have highly cross-linked (XLPE) versus UHMWPE inserts. Identifying material and clinical characteristics of post fractures will help surgeons and implant designers decrease the incidence of this complication after TKA, and thus increase implant longevity.

Methods

Utilizing our institutional review board approved retrieval system, we identified 33 PS polyethylene inserts with fractured posts that had been removed at revision surgery in 30 patients (3 bilateral procedures) conducted between January 1, 2015, and December 31 2022. Patient characteristics collected for each insert included age at index TKA, sex, body mass index, length of implantation (LOI), and patient-reported details on the events surrounding the post fracture. Among the 30 patients from whom the 33 fractured post components had been retrieved, the mean age at index surgery was 55 years (range, 35 to 69), the mean body mass index was 29.5 (range, 18.5 to 37.2), and the mean LOI prior to revision for post fracture was 10 years (range, 4 to 27). The majority of the post fractures occurred in men (81.8%, 27 of 33).

We also recorded patients’ University of California Los Angeles (UCLA) activity scores prior to the post fractures occurring and at most recent follow-up, as well as the Knee Injury and Osteoarthritis Outcome Score for Joint Replacement scores at most recent follow-up. If patient had magnetic resonance imaging (MRI) scans performed pre-revision surgery, alignment of the tibial tray and femoral component were measured using multiacquisition variable-resonance image combination MRI sequencing, read by a trained radiologist. Rotation, recorded in degrees, was defined as either internal or external. The femoral component was measured with respect to the femoral shaft, and the tibial tray with respect to the tibial tubercle. Tibial tray alignment, reported in degrees varus or valgus, and tibial slope were measured from immediate post-operative short leg radiographs.

There were 13 patients who had XLPE implants and 20 who had UHMWPE implants. Implant characteristics recorded were manufacturer, material (highly cross-linked or conventional), and damage characteristics based on Hood scoring [8] and scanning electron microscopy (SEM). We evaluated damage characteristics of each polyethylene component using the Hood grading system [8], which has been used for multiple prior retrieval studies [9–13]. Briefly, the implant articular surfaces (inferior and superior regions) were divided into 12 zones, and each zone was assigned a damage score on a scale of 0 to 3 (indicating no, mild, moderate, and severe damage). The articular surface of the implant contained zones 0 to 7 and the backside consisted of regions 8 to 11 (Fig. 1). As the posts had been completely disassociated by the fracture, the post surfaces were excluded from analysis. Damage modes assessed included scratching, pitting, burnishing, abrasion, surface deformation, delamination, and embedded debris.

Fig. 1.

Hood damage grading regions for posterior stabilized (PS) knees with post fracture. The post region was excluded from analysis. (A) Articular surface regions (0 to 7) and (B) Backside regions (8-11); each region was graded from 0 to 3 for severity/area seen with each damage mode independently. Scores summed for 1 aggregate damage score per implant. This example is based on a right knee. Zones were switched according to implant laterality.

Scoring was done by 2 independent observers blinded to all clinical data. After grading implants that showed embedded debris and delamination (9 implants total, all UHMWPE) for those specific damage modes, all implants were sputter coated in gold or palladium coating using a sputter coater (Denton II, Moorestown, New Jersey) to better visualize surface features for grading and later SEM imaging. Implants were sputter coated for 40 seconds at 60 mA current supply, and in cases of larger inserts, each half (medial and lateral) was coated separately to ensure even coverage. After sputter coating, the implants were scored for the remaining 5 damage types. Each damage mode was assigned a separate score in each of the 12 regions. The scores were averaged and aggregated to report a total score per implant and each respective damage mode. Damage scores were also reported regionally for the articular surface and backside of the implant.

We selected 13 inserts (7 UHMWPE, 6 XLPE) out of the cohort for SEM imaging (Carl Zeiss Gemini 300 FE-SEM, White Plains, New York) to visualize the fracture surface and identify distinct fracture progression features. We chose those inserts that had not sustained damage to the fracture surfaces after the posts had broken. All images were acquired using accelerating voltage between 5 to 10 keV using the secondary electron detector. Aperture size used was 120 μm, and beam size was auto selected by the imaging software depending on the magnification settings. Images at high and low magnification were acquired, and fracture progression features unique to each material type were compared qualitatively.

Data Analyses

Damage scores were normally distributed as established by Shapiro-Wilk tests. Descriptive data were reported as means and standard deviations. Student’s t-tests were used to assess implant damage between material groups, damage type versus material type, and patient demographics within material type. Linear regressions and correlation matrices were used to assess the relationships between patient demographic data and damage scores. Significance was deemed at P < .05. Analyses were performed using Prism V.9 (GraphPad, San Diego, California USA).

Results

Among our cohort, 76% (25 of 33 total inserts) were from a single manufacturer (Smith & Nephew, Memphis, Tennessee). The remaining insert manufacturers were Zimmer (Warsaw, Indiana) 12.1% (4 of 33), Exactech (Gainesville, Florida) 6.1% (2 of 33), Biomet (Warsaw, Indiana) 3.0% (1 of 33), and Stryker (Mahwah, New Jersey) 3.0% (1 of 33).

On average across the entire cohort, the tibial tray was found to be in 1.2 (±1.6°) of varus with 4.6 (±2.6°) tibial slope. Tibial slope did not differ between groups when assessed for material type; however, tibial tray alignment was lower for the XLPE cohort when compared to the UHMWPE components (0.3 varus ±1.3° versus 1.7 varus ±1.6°, P = .02) (Fig. 2). For the entire cohort, total damage score based on Hood scoring did not correlate with tibial tray alignment or tibial slope. Also, MRI rotational alignment did not correlate with total damage score (Fig. 3).

Fig. 2.

Radiographic alignment for the tibial tray and the tibial slope of each knee. Damage score was reported on the y axis with alignment along the x axis.

Fig. 3.

Magnetic resonance imaging (MRI) rotational alignment for both the femoral component and the tibial tray. Damage scores were reported on the y axis with rotation along the x axis. There were 25 of 33 implants for damage scoring that also had an MRI pre-revision surgery.

The LOI prior to revision for post fracture was significantly shorter in the XLPE group versus the UHMWPE group (mean 7.3 years [range, 4 to 11] versus mean 12 years [range, 7 to 27] P =.002). The total damage score was significantly higher in the UHMWPE group versus the XLPE group (mean overall damage score: 57.3 ± 11.2 versus 44.2 ± 9.7, P =.002, Fig. 4). The XLPE group also had less damage (a lower damage score) on the backside when compared to the UHMWPE (11.13 ± 4.4 versus 19.7 ± 5.4, P < .001). No differences among the 7 damage modes were found between the 2 material groups.

Fig. 4.

Damage scores for each implant sorted by material type. The highly crosslinked polyethylene (XLPE) inserts had less damage overall when compared to ultra-high molecular weight polyethylene (UHMWPE) implants. Bars correspond to the means and standard deviations for each group.

The mean overall Knee Injury and Osteoarthritis Outcome Score for Joint Replacement score improved from mean 56 points (range, 31 to 76) after breaking the post (immediately prior to the revision for post fracture) to a mean of 74 points (range, 42 to 100) 1 year after revision surgery. UCLA activity scores were available for 24 patients both prior to post fracture and after revision for post fracture. The mean UCLA activity score prior to post fracture was 6.7 ± 2.2. After revision for post fracture, the mean UCLA activity score was 6.4 ± 2.1.

SEM of the fracture surfaces showed that crack initiation occurred on the posterior side in 10 of the 13 components examined, with the fractures occurring usually around the midpoint of the post height. In most cases where initiation could be determined, the immediate area at initiation had a brittle appearance (Fig. 5A), with outward radiating scars (Fig. 5B). The fracture surface of XLPE posts had more precise clamshell marks (Fig. 6A), while UHMWPE posts had a more tufted, irregularly distributed clamshell appearance (Fig. 6B). The XLPE fracture surfaces also demonstrated a diamond pattern towards the anterior side of the post, consistent with final acute fracture (Fig. 6C). The UHMWPE posts did not exhibit this diamond pattern. The apex of the diamond markings points towards the fracture initiation origin on the posterior edge of the post and indicates acute breakage towards the anterior fracture surface. As the fracture progressed through the post, the fracture plane often changed resulting in multidirectional fracture planes. These kinds of plane shifts along both vertical and horizontal directions were observed in all 13 devices under SEM.

Fig. 5.

Scanning electron micrograph showing fracture initiation point on the posterior face of the post. (A) Low magnification image showing relatively low deformation at initiation point (white rectangle). (B) High magnification showing region within the rectangle from (A). Note the outward radiating scars from the point of initiation at the edge.

Fig. 6.

Scanning electron micrograph showing (A) highly crosslinked polyethylene (XLPE) fracture progression along the fracture surface. Note the distinct clamshell markings showing fracture progression (B) ultra-high molecular weight polyethylene (UHWMPE) fracture progression along the fracture surface. Note the tufted, irregular markings showing fracture progression. (C) Diamond markings near the anterior edge of the fractured post, with the apex of the diamond pointing towards fracture initiation point on the posterior side (not pictured). These diamond markings were only seen in XLPE posts and were absent in UHMWPE posts.

Discussion

In this series of 33 PS TKA cases, the fractured posts predominantly occurred in younger men TKA patients who had high activity levels. The characteristics of the fracture differed between XLPE and UHMWPE implants, with fractures occurring in the XLPE group displaying less generalized damage on the tibiofemoral articular surfaces and after a shorter LOI. The fractures did not appear to be related to positioning based on MRI measurements.

Prior reports of post fractures in the literature are limited, with most studies reporting an incidence of post fractures below 1% [1]. Diamond et al. reported on a series of 5 cases of post fractures, which had occurred in a series of 955 consecutive XLPE knees from 1 company. They noted that in over 2,000 prior TKAs performed using UHMWPE from the same company with the same surgical technique, no post fractures had occurred [2]. Our study demonstrated post fractures occurring in both XLPE and UHMWPE knees, though the specific characteristics and SEM findings differed based on material type.

The XLPE has a highly crystalline structure and a higher degree of crosslinking, which imparts improved wear resistance as compared to UHMWPE [14]. The well-defined clamshell markings seen on the fracture surfaces of the XLPE material (Fig. 6A) are likely due to the highly organized crystallinity. The mechanical properties, including fracture toughness, of polyethylene are determined by not just crosslinking density and degree of crosslinking, but also by the molecular weight between crosslinks. Pruitt et al. reported a series of case studies illustrating the reduction in fatigue and fracture resistance properties that can occur by increasing the degree of crosslinking [15]. Consistent with the improved wear resistance in XLPE, we found a lower prevalence of surface damage in XLPE implants as compared to UHMWPE implants.

Among our cohort, 76% (25 of 33 total inserts) were from a single manufacturer (Smith & Nephew, Memphis, Tennessee). However, it cannot be concluded that this implant design has a higher incidence of post fracture. We do not have information on the total number of cases from which any of the 33 fractures occurred. A previously published study has indicated different manufacturers utilize different methods for modulating crosslinking and crosslinking density, as well as for postirradiation processing, which could lead to changes in oxidation [16]. However, direct correlations between individual post-processing methods, crosslinking density, and proclivity for fracture at the post have not been studied yet.

This study has several limitations. This is a single institution series with a small number of cases. However, due to the rarity of this complication, this appears to be the largest series of patients reported who had this complication. Also, we were unable to account for all factors that may have led to post fractures including differences in cam-post geometry, loading conditions differences among the patients, and variations in other material factors. Since specifics of the crosslinking techniques employed, the degree of crosslinking, and other manufacturing specifications for each implant are not available, we were unable to determine the extent to which these factors influenced fracture progression and fracture mechanics. Some notable differences were seen in the dimensions of the post across designs, however, due to the limited sample size and uneven distribution across designs, we are not able to quantitatively compare if design differences had a bearing on the fracture progression.

In summary, PS post fractures in our retrieval registry occurred predominantly in active, younger men. The characteristics of the fracture were different between XLPE versus UHMWPE implants, with fractures occurring in the XLPE displaying less generalized surface damage, after a shorter LOI, and with SEM evaluation indicative of an acute fracture pattern. Neither material appeared superior in terms of fatigue resistance to post fractures.