1 Introduction

By 2030, an estimated 20% of Americans >65 years old or 70 million people will be at risk for degenerative arthritis or osteoarthritis (OA).¹ Over 95% of all total knee arthroplasty (TKA) in the United States are performed for OA.² Knee osteoarthritis with valgus deformity presents a challenge for surgeons performing primary TKA. Component constraint has become an option, but with particular caution since it carries its own risks, including premature aseptic loosening. Component constraint arises in the presence of a deficient soft-tissue envelope. Specifically, incompetent collateral ligaments, inability to achieve a varus-valgus balance, several neuropathic disorders, and intraoperative injury to collateral structures may lead to the use of increased constraint in TKA.³ The goal of the constrained articulation is to increase stability in the varus-valgus coronal plane and rotational stability.

Modern TKA systems use a wider post to provide increased rotational and varus/valgus constraint. Midlevel constraint (MLC) bearings limit rotation and varus/valgus lift-off but have less constraint than the constrained condylar knee (CCK) insert. MLC bearings derive from the posterior stabilized (PS) design but deviate in terms of a thicker and wider post on the polyethylene insert, in which the conformity is greater with higher level of articulation with the femoral box. It is often used to correct valgus deformity resulting from the lateral bone loss on the femur and deficiency of soft-tissue envelope. In this case, the contact between tibial eminence and femoral box can act to reduce a hyperextension inclination.⁴

However, the extent to which MLC bearings and PS bearings are compared in a similar patient population is limited throughout the literature. We attempt to compare a cohort of MLC and PS knee in regards to i) Patient Reported Outcome Measurements (PROMs), ii) preoperative and postoperative deformity, and iii) complications, including revision rate. We believe there will be more valgus knees in the MLC cohort, improved PROMs, reduced revision rate than in the PS cohort, and differential radiographic evidence, revealing the effectiveness of the constrained liner in a clinical setting for its specific needs.

2 Materials/methods

We performed a retrospective review of prospectively collected data in our institution's database to identify patients who underwent a primary TKA with a MLC implant (Exactech, Inc., Gainesville, FL) from a single manufacturer, single surgeon from 2013 to 2016, minimum two-year follow-up, and uniformity in principle diagnosis (OA). The Kellgren-Lawrence system was used to classify OA, with a grade 3 or greater necessary for use in this study. This was determined by the orthopedic surgeon and confirmed by the physician assistant. We then performed another retrospective review of patients who underwent a primary TKA with a PS implant from the same manufacturer, same surgeon from 2013 to 2016, a minimum two-year follow-up and the same principle diagnosis.

Physical evaluation includes a varus-valgus stress test of the collateral ligaments in extension, at 30° of flexion, and at 90° of flexion. The antero-posterior stability of the implant was determined using an anterior drawer test at 90° of flexion. Mid-flexion instability was determined by stability at 90° of flexion and instability in the 30–60° flexion arc. This was followed by radiographic evaluation to confirm the diagnosis. Lateral images were examined for femoral component flexion in terms of intramedullary axis and for tibial component posterior slope.⁵ Varus-valgus stress views reveal the degree of ligament laxity, which can be confirmed with a computed tomography to evaluate rotational alignment of the femoral and tibial components. Mid-flexion instability is then classified into three types: 1) over-released MCL and the posterior capsule, ii) normal MCL, and iii) over-release of MCL with a tight posterior capsule.⁶

The decision to use the MLC or PS implant was made by the orthopedic surgeon in i) the use of trial PS inserts and ii) a final intraoperative assessment of coronal stability in flexion and extension. We recorded demographics, preoperative deformity, postoperative PROMs and complications at an average follow-up of 4.0 years in the MLC cohort and 3.8 years in the PS cohort.

3 Results

A power analysis revealed 41 patients in each cohort was necessary to achieve a desired power of 0.80. 153 patients underwent primary TKA by a single surgeon who met the inclusion criteria of an eligible 180 patients. 57 patients underwent primary TKA using a MLC knee and 96 underwent primary TKA using a PS knee. Both cohorts were from the same manufacturer, timeframe (2013–2016), minimum two-year follow-up, and uniformity in degenerative arthritis as a primary diagnosis. The average age (68.91 vs. 68.40, p = 0.72), average BMI (30.88 vs. 29.14, p = 0.10), and males to females (25:32 vs. 28:69, p = 0.08) were all comparable between the MLC and PS cohorts, respectively. The average follow-up in the MLC cohort was 3.98 years and the average follow-up in the PS cohort was 3.82 years, p = 0.38.

There was a significant difference in the number of varus vs. valgus knees in the MLC cohort than the PS cohort (21:36 vs. 74:22, respectively, p < 0.001, Table 2). There was not a significant difference in degrees of postoperative alignment (MLC cohort: preoperative varus group had 4.74° of valgus postoperatively, preoperative valgus group had 5.43° of valgus postoperatively. PS cohort: preoperative varus group had 5.40° of valgus postoperatively, preoperative valgus group had 4.80° of valgus postoperatively).

Also, in terms of PROMs, the baseline data reveals that there was no significant difference in any of the PROM: 78.2 vs 76.10 for KSCSRS (p = 0.30), 110.32 vs. 109.39 for ROM (p = 0.38), and 31.43 vs. 33.10 for VAS pain, respectively (Table 3). Of note, the KSCRS-Total difference between the cohorts was 78.20 and 76.10, p = 0.30. At the respective postoperative timepoints, there was a significant difference in KSCRS-Total between the two cohorts (MLC 163.9 vs. PS 132.8; p = 0.003, Table 4). There was no significant difference in terms of ROM, anterior knee pain, or VAS Pain. The rate of revisions revealed low rates in both cohorts, but this was not significant (3.5% MLC vs. 2.10% PS, p = 0.90). There was also no significant difference in manipulation rate between the cohorts (8.80% MLC vs. 9.40 PS, p = 0.13). Of note, there was no significant difference in reason for revision, which shows a consistency in surgical approach (Table 5).

4 Discussion

Our hypotheses regarding the use of constraint, both support the existing literature and add to it as well. We found that the MLC cohort was able to successfully balance preoperative deformities to a similar extent as the PS cohort since both cohorts obtained a clinically acceptable measurement of 5° valgus postoperatively. This was confirmed radiographically, in which the deformity correction was achieved both groups appropriately, which confirms that larger preoperative deformities led to increase use of MLC. This is consistent with several long-term follow-ups that demonstrate the ability of constrained prostheses to correct extreme valgus deformity. Krackow et al. used a similarly designed studied and found similar alignment at postoperative times between two to ten years.⁷ Miyasaka et al. found 75% of the knees corrected between 2° and 7° valgus at an average follow-up of 14.1 years, which is consistent with our finding of 5.43° for the MLC cohort and 4.80° for the PS cohort at an average follow-up of 3.98 years and 3.82 years, respectively.⁸ However, these studies lack a group of similar demographics, surgical technique, principle diagnosis, uniform implant manufacturer, and similar timeframe.

In our study, we primarily explored a principle diagnoses of OA in both cohorts in order to maintain homogeneity. It also provides evidence for the use of constraint. OA can manifest as a valgus knee, which is demonstrated in this study more commonly in the MLC cohort. In addition to degenerative arthritis/extreme valgus knees, angular deformities leading to weakening of collateral ligaments, intraoperative injury to collateral structures, and inability to achieve a balanced flexion and extension gap can be reasons to use constraint in TKA.⁴ It would be useful to perform a similarly designed study with a different principle diagnoses to see if the results are maintained and more insight into the nature of MLC. This approach was taken by Fujiwara et al. which showed that patients with rheumatoid arthritis had improved KSS knee/function scores as well as radiographic femoral tibial angle at a minimum two-year follow-up after TKA with a constrained prosthesis.⁹

In terms of novel findings, the use of a control group allowed us to compare PROMs among comparable cohorts. Several studies have demonstrated the improvement of ROM and HSS score of constrained prostheses from preoperative visit to postoperative visit at follow-up greater than two years.^10–12 However, the extent to which there is a significant difference between the control cohort and constrained cohort, which we did in regards to KSCRS-Total (Table 3), is limited. This has significant implications in support of a constrained prostheses. Both Puah et al. and King et al. demonstrated no significant difference in ROM and knee function scores at a six-year follow-up and one-year follow-up, respectively.^13,14 In our study, we found that there was no significant difference in terms of ROM-Flexion, which acts a subset of the KSCRS-Total form, which reveals that the difference occurs in functional outcomes, such as walking, going up the stairs, and use of a cane.

We found a low revision rate of 3.5% in the MLC cohort. This is consistent with the existing literature. Moussa et al. found a revision rate of 3.43% in their constrained implant cohort of 817 TKA at a two-year follow-up.¹¹ Crawford et al. also found the revision rate of their constrained cohort to be 3.0% at a two-year follow-up.¹⁵ We had comparable but generally low manipulation rates compared to the literature (8.78% for MLC and 9.40% for PS) Pancio et al. found a manipulation rate of 10% for the constrained cohort at an average follow-up of 5.4 years.¹⁶ Ruel et al. reported a manipulation rate of 15.2% for their constrained cohort at five-year follow-up. The low manipulation rate may be accounted for by the use of one implant type in the entire MLC cohort.¹⁷

We addressed several limitations through the design of the study. The patient numbers were small because we had rigid inclusion criterion in order to accurately compare the two cohorts. While our follow up is limited, we hope to include a longer follow-up in subsequent studies to compare the cohorts in terms of survivorship outcomes. However, we maintained low revision rates outside of manipulation and no patients lost to follow-up. We believe the primary diagnosis of osteoarthritis as determined by the Kellgren and Lawrence system was important for maintaining similar patient populations and outcome comparisons.¹⁸ Finally, even though the study was a retrospective review, we maintained enough homogeneity to produce a worthwhile analysis.

We concluded that the MLC insert corrected varus-valgus imbalance as well as provided similar PROMs to the PS cohort. While our follow up is limited, we concluded that MLC inserts should be considered in patients with intraoperative coronal instability. While we feel that balancing the knee is imperative and that the least amount of constraint should be utilized, surgeons should consider the MLC option when needed.

Funding

There is no funding source.

Ethical approval

This article does not contain any studies with human participants or animals performed by any of the authors.

Informed consent

Informed consent was obtained from all individual participants included in the study.