1 Introduction

Metal backing for patellar components in total knee arthroplasty (TKA) was introduced in the 1980s by several manufacturers. At the time, the metal backing was felt to reduce contact stresses within the implant and, in some designs, facilitate biologic fixation. The early experience with these devices was negatively affected by several problems observed shortly after their introduction. These included delamination of the polyethylene from the metal substrate, fracture of the polyethylene, fracture of the component, and failure of fixation.^1–4 For example, Lombardi et al. reported on the occurrence of catastrophic separation of the polyethylene articular surface from the metal backing in 17 patients. The patients were in high demand, but the complete separation of the materials was believed to be secondary to design flaws in the method of bonding the materials. Peters et al. reported on 31 cases with two discrete failure modes.³ Similarly, Stulberg et al., described 16 cases where the metal backing of the components deformed and was cut through the overhanging edge of the device, resulting in catastrophic dissociation.⁴

Many of these problems were observed early in the follow-up, after implantation, and were related to problems in the bond between the metal backing and the polyethylene articular surface.⁵ In the past decade, techniques using additive manufacturing have evolved to allow the production of composite patellar implants, combining a polyethylene articular surface with a biologically active posterior surface, allowing for biologic fixation, and aiming to avoid the design problems associated with earlier attempts.

The purpose of this study was to describe the revision risks of cementless TKA that incorporated an additively manufactured metal-backed patellar component (AM-MBP) in comparison with aggregate cementless and cemented TKA benchmark groups and to identify whether any trends suggestive of device-related failure mechanisms could be seen in a large population-based national registry dataset.

2 Methods

3 Results

The primary TKA was linked to a revision procedure for 400 of 35,087 AM-MBP (1.1%), 191 of 10,755 aggregate cementless (1.8%), and 9,066 of 550,908 aggregate cemented (1.6%) group cases (Table 3). The most common diagnosis associated with revision was infection, recorded as a cause for 178 (0.5%), 68 (0.6%), and 3,770 (0.7%) revisions in the groups, respectively (Table 3). The cumulative percent all-cause revision (95 %CI) at 7-year follow-up was 1.9% (1.7%, 2.1%; 1,051 at risk) for the AM-MBP, 2.5% (2.1%, 2.8%; 1,281 at risk) for other aggregated cementless. and 2.1% (2.1%, 2.1%; 105,641 at risk) for aggregate cemented group (Fig. 1 and Supplemental Table 1).

Close

Fig. 1 Cumulative percent revision curves with 95 % confidence bands from stratified groups model.

Export to PPT

The adjusted hazard ratios (95% CI) for revision, controlling for age and sex, were 0.77 (0.64, 0.93; P = 0.007) for comparing AM-MBP with aggregate cementless, 0.92 (0.83, 1.03, P = 0.138) for comparing AM-MBP with aggregated cemented, and 1.2 (1.0, 1.4, P = 0.030) for comparing aggregate cementless to aggregate cemented.

4 Discussion

This is the first report to utilize national registry data to describe the survivorship of TKA incorporating modern cementless implants for reconstruction of the anterior compartment of the knee. Though the median follow-up for the AM-MBP group was only 22.4 months (Table 2), previous experience has demonstrated that design-related failures are typically demonstrated within this early time frame. Andersen et al. noted an 8% failure rate in 111 TKAs within two years of implantation, mostly related to fracture and/or separation of the polyethylene from the metal backing.⁵ Chan and Giori noted a 20% component fracture rate in porous tantalum buttons within five years of implantation.⁶ There are also reports of wear-related failures occurring much later, after implantation. These failures are related to the relative thinness of the polyethylene, which resulted from the need to accommodate a minimum metal thickness to achieve the required strength in these early designs.^7–9

These registry findings corroborate the observation of Grau et al., who reported on their experience with 388 TKA followed prospectively for a minimum of two years. In this study, 80 cases with the AM-MBP component were followed radiographically for more than five years. There was only one patellar component without obvious radiographic evidence of ingrowth at two years and one at five years. Neither of these patients required revision; in fact, there were no revisions in this group who had patellar problems.¹⁰ Harwin et al. reported on 261 patients at a mean follow-up of 4.5 years (range, 4 to 5) who had this AM-MBP component. They found only one patellar component that loosened (rate 0.3%) following manipulation under anesthesia six weeks post-operatively.¹¹

To address device micromotion, Sporer et al. investigated the biological fixation of the 3D-printed baseplate and patella in a radio-stereometric analysis. Small tantalum beads were implanted in 29 patients who received a cementless, highly porous tibial baseplate and patella and were followed for two years post-operatively. The mean difference in maximum total point motion between 12 and 24 months was 0.021 mm (range, −0.265 to 0.572) for the tibial implant and 0.089 mm (range, −0.337 to 0.758) for the patellar implant. The rate of tibial and patellar migration was highest over the first six postoperative weeks, with no changes in mean patellar migration occurring after six weeks. The author concluded that bony fixation appeared to occur reliably on the highly porous implant surface of the tibial baseplate and metal backing for the patellar component.¹²

The positive clinical results seen in clinical trials and in this report could in part be due to the beneficial design features of this AM-MBP over predicate designs.The AM-MBP design was developed using a titanium porous matrix with a mean pore diameter of 497 μm, a 0.80 coefficient of friction, and a mean porosity of 65% to mimic cancellous bone. The components are additively manufactured using a focused laser beam to sinter numerous layers of titanium powder to grow the implant structure layer-by-layer and adjust solid and porous areas to optimize ingrowth surfaces. Additive manufacturing also allows for a metal thickness of 9 to 11 mm and a thicker, highly crosslinked polyethylene with known low-wear properties.¹³ The design was developed to address past failures of metal-polyethylene dissociation by enhancing the bond between the two components. The architecture on the back side, combined with a direct compression molding process, is designed to minimize the potential for dissociation (Fig. 2).

Close

Fig. 2 Design of the additively manufactured metal-backed patellar component.

Export to PPT

The registry data presented in this study have important potential limitations related to the incomplete capture and lack of specificity of implant survival outcomes, the potential for residual confounding, and limited generalizability. While the AJRR is a robust registry, it only captures approximately 40% of all arthroplasty procedures in the United States (US); thus, when a TKA implanted at a registry institution is revised at a non-registry institution, the revision may not be captured. The linkage of AJRR with CMS data facilitates improved capture of subsequent revisions only among fee-for-service Medicare beneficiaries who had a revision after implementation of ICD-10 (because ICD-9 does not identify the laterality of the procedure). Part C Medicare claims data for the AJRR cases was not available at the time of this query. Also, the AJRR does not track the explantation of specific components, but rather the occurrence of any subsequent revision TKA procedure of the left or right knee of the patient. For example, when four components are implanted during the primary TKA procedure and a subsequent revision procedure is captured, it is not known which of the four components failed. While analyses controlled for age and sex, other factors that might bias the estimation of cumulative percent revision or that may confound comparisons with benchmark groups were not accounted for in these analyses. Also, these analyses were restricted to patients aged 64 years or more at the time of primary TKA and therefore are not generalizable to younger patients or to the overall US population.

In conclusion, although longer device-specific follow-up studies are needed to determine future performance, especially concerning wear, the present study is encouraging for the AM-MBP. It suggests that the durability of the construct in real-world use in the US Medicare-eligible population (age ≥65 years) was at least comparable to the average durability of other cementless constructs captured by AJRR over the recent decade. Though the AJRR does not capture or analyze revisions on a component-specific basis, these results are consistent with the absence of any clinically apparent design-related failure mechanism.

CRediT authorship contribution statement

Ormonde Mahoney: Methodology, editing. Manoshi Bhowmik-Stoker: editing. Denis Nam: editing, Software, Validation. Michael A. Mont: Conceptualization, editing. Sean B. Sequeira: Conceptualization, Methodology, Writing – original draft, editing. Michael Dunbar: Methodology, editing.

Consent

Patient consent was not necessary for this investigation.

Ethics

The authors confirm that all procedures were performed in compliance with relevant laws and institutional guidelines. This project utilized a national insurance database with de-identified patient information and therefore met criteria for an IRB exempt study.

Funding

No funding was obtained for this investigation.