Patient specific implants versus conventional implants in primary total knee arthroplasty: No significant difference in patient reported outcomes at 5 years

Abstract

Background

Patient specific implants (PSI) represent a novel innovation aimed to improve patient satisfaction and function after total knee arthroplasty (TKA); however, longitudinal patient reported outcome measures (PROMs) for PSI are not well described. We sought to primarily evaluate PROMs of patients undergoing TKA with either PSI or off-the-shelf (OTS) implants at mid-term follow-up.

Methods

A retrospective review was performed on a prospectively collected cohort of 43 primary, cruciate-retaining TKAs performed with PSI (n = 23) and OTS implants (n = 20) by a single surgeon. Patient demographics, operative characteristics, range of motion (ROM) return, reoperations, and outcomes [Patient-Reported Outcomes Measurement Information System (PROMIS) T-score, Knee Injury and Osteoarthritis outcome score (KOOS), and Knee Society Score-Function (KSS-F)] were compared. Mean follow-up was 5 years.

Results

TKA performed with either PSI and OTS implants demonstrated no difference in obtaining ROM by 3 months (extension 3° short of full extension vs. 0°, p = 0.16) or flexion (114° vs. 115°, p = 0.99) and final ROM was identical [0° extension to 120° flexion (p = 1)]. Although not significant (p = 0.42), 5 (22%) PSI TKA and 2 (10%) OTS implant patients required manipulation under anesthesia. KSS-F and PROMIS T-scores were higher in the PSI versus OTS TKA patients, respectively (90 vs. 73, p = 0.002; 51.6 vs. 44.5, p = 0.01). However, after multivariable analysis, none of these continuous outcome measures were significantly different (p = 0.28 for KSS and p = 0.45 for PROMIS T-score) between the groups.

Conclusion

In a series of TKAs performed with PSI, no difference existed in postoperative ROM, reoperations, or patient-reported outcomes compared to OTS implants at 5 years. Surgeons may utilize the equivocal midterm results during TKA preoperative patient discussion of implant technologies.

1. Introduction

Total knee arthroplasty (TKA) is among the most performed elective surgeries with approximately 600,000 TKAs performed annually in the United States alone.1, 2, 3, 4 In 2015, an estimated 4.7 million Americans were living with at least one TKA.^2,4 Despite success and increasing prevalence, up to 30% of patients have been cited to be unsatisfied with their postoperative outcomes.^5,6 To reduce the rate of dissatisfaction, innovations continue to be created to improve TKA surgical technique including new instrumentation, navigation and robotic technology, and implant designs.^7,8

The ConforMIS implants (ConforMIS Inc, Billerica, MA) represent yet another novel iteration of custom instrumentation whereby a patient-specific implant (PSI) is manufactured in addition to custom cutting guides. Utilizing a computed tomography (CT) scan of the operative extremity, disposable cutting guides and individualized implants (femur, tibia, and polyethylene tibial inserts) are fabricated based upon anatomical and biomechanical axes of the knee.⁹ This allows for detailed preoperative planning and operative efficiency that may decrease instrument trays and optimized technique of bone cuts and soft tissue balancing.

Despite theoretical benefits, the indications for the use of PSI versus conventional or off-the-shelf (OTS) implants for TKA remains unclear. Patient-specific instrument and implant systems have not clearly proven superiority in the acute outcomes of operative time, blood loss, range of motion (ROM), or hospital length of stay.9, 10, 11 Yet, other studies have demonstrated favorable results of reproducible postoperative limb alignment and more natural knee kinematics.^9,10,12, 13, 14 PSI TKA advocates could suggest such outcomes may improve patient satisfaction and implant durability; however, longitudinal data specifically comparing patient reported outcome measures (PROMs) of PSI to OTS is lacking.

We sought to primarily evaluate mid-term PROMs for a consecutive cohort undergoing TKA with either PSI or OTS implants. Secondary aims included reviewing perioperative outcomes and reoperations after TKA of each group. We hypothesized that no significant difference would exist in patient satisfaction or outcomes between the implant systems.

2. Methods

After obtaining Institutional Review Board (IRB) approval, a retrospective review was performed on prospectively collected cohort of TKAs performed with PSI or OTS implants between 2012 and 2015. Patients were included for the indication of osteoarthritis only and after failing to improve with minimum 3-month conservative treatment program. Exclusion criteria included: simultaneous bilateral procedures, ipsilateral lower extremity surgery or contralateral TKA within one-year, severe deformity (>15° fixed varus/valgus deformity or > 10° flexion contracture) body mass index (BMI) > 40 kg/m² or incomplete minimum 2-year follow-up. A total of 43 patients (23 PSI and 20 OTS) were reviewed with mean follow-up of 5.4 (range, 3.3–5.6) years.

All TKA were performed by contemporary technique by a high-volume arthroplasty surgeon (MIO) at the same institution. A standard medial parapatellar approach was utilized and cemented cruciate-retaining prosthesis were implanted—ConforMIS iTotal® knee for PSI group and Stryker Triathlon® knee (Stryker, Kalamazoo, MI) for OTS group (Fig. 1). Both cohorts received a standardized rapid mobilization protocol including multimodal analgesia, physical therapy on postoperative day zero or one, venous thromboembolism prophylaxis, and case management to optimize discharge and outpatient physical therapy. Each patient was routinely followed postoperatively in outpatient clinic at approximately 2 weeks, 2 months, 2– and 5 years whereby clinical and patient-reported outcome measures were collected.

Fig. 1.

Anteroposterior and lateral knee radiographs of ConforMIS iTotal® and Stryker Triathlon® cruciate retaining total knee implants.

Data was collected and reviewed via the institutional electronic medical record. Baseline preoperative patient characteristics (age, sex, race, co-morbidities, prior knee surgical history, and ROM), and operative details (anesthesia type, surgery duration, and surgical techniques) was obtained. Surgical outcomes evaluated included rate of return of ROM during the early (<3 months) and final postoperative follow-up, any significant complication or reoperation, and patient reported outcome measures. Specifically, the Patient-Reported Outcomes Measurement Information System (PROMIS) T-score, Knee Injury and Osteoarthritis outcome (KOOS) score, and Knee Society Score-Function (KSS-F) were selected as the validated PROMs to be consecutively collected throughout the postoperative period.15, 16, 17

A comparison of baseline demographics and operative characteristics between PSI and OTS patients is shown in Table 1. PSI TKA patients were younger (65 vs. 73 years, p = 0.009) and more often male (53% vs. 18%, p = 0.003). There were no other significant differences between the groups including BMI, preoperative pain scores, ROM, or tourniquet time.

Table 1.

Comparison of baseline and operative information according to type of total knee arthroplasty implant used.

	ConforMIS iTotal® (N = 23)a	Stryker Triathlon® (N = 20)a	p-value
Sex			0.004
Female	9 (39%)	16 (85%)
Male	14 (61%)	4 (15%)
Age at Surgery (years)	64.7 (37–86)	72.8 (2046)	<0.001
Body Mass Index (kg/m2)	33.8 (22–43)	29.8 (20–52)	0.19
Preoperative VAS Pain Scoreb	7 (4–10)	8 (4–10)	0.48
Preoperative Range of Motion
Flexion	115° (95°–130°)	113° (90°–130°)	0.47
Extension	5° (−2°–20°)	5° (0°–115°)	0.74
Anesthesia Type			1.00
Spinal/Regional	18 (78%)	15 (75%)
General	5 (22%)	5 (25%)
Tourniquet Time (minutes)	70 (54–104)	68 (41–78)	0.09

^aThe sample median (minimum-maximum) is given for continuous variables.

^bVAS = Visual Analog Scale for pain.

2.1. Statistical analysis

Continuous variables were summarized using the sample median and range. Categorical variables were summarized with number and percentage of patients. Unadjusted comparisons of baseline, operative, and post-operative characteristics between PSI or OTS patients were made using a Wilcoxon rank sum test (continuous variables) or Fisher's exact test (categorical variables). For continuous outcome variables where multivariable analysis was possible, linear regression models were adjusted for any baseline or operative variable that differed between cohorts while accounting for potential confounding variables. A p-value less than 0.05 was considered statistically significant and all statistical tests were two-sided. Statistical analyses were performed using R Statistical Software (version 3.4.2; R Foundation for Statistical Computing, Vienna, Austria).

3. Results

The outcomes of return of ROM, reoperations, and revisions are compared in Table 2. During the acute interval 3 months postoperatively, patients with PSI and OTS implants demonstrated no statistical significance in the ability to obtain extension (3° short of full extension vs. 0°, p = 0.16) or flexion (114° vs. 115°, p = 0.99). Final average range of motion was identical in both groups: 0° extension to 120° flexion (p = 1). Although not significant (p = 0.42), 5 (21.7%) PSI TKA and 2 (10%) OTS implant patients required a reoperation of manipulation under anesthesia (MUA). All patients that underwent MUA achieved acceptable ROM without further sequelae and there were no further reoperation or revision TKAs required for either cohort.

Table 2.

Comparison of range of motion return and reoperations according to type of total knee implant used.

	ConforMIS iTotal® (N = 23)a	Stryker Triathlon® (N = 20)a	p-value
Acute Postoperative Range of Motion
Flexion	114° (85°–130°)	115° (75°–130°)	0.99
Extension	3° (0°–30°)	0° (0°–25°)	0.16
Final Postoperative Range of Motion
Flexion	120° (95°–130°)	120° (90°–130°)	0.65
Extension	0° (0°–5°)	0° (0°–3°)	0.56
Reoperationsb	5 (22%)	2 (10%)	0.42

^aThe sample median (minimum-maximum) is given for continuous variables.

^bManipulation under anesthesia for acute postoperative stiffness was the only type of reoperations required in these patients.

At final follow-up, the KOOS scores were similar between PSI and OTS knees (81.4 vs. 85.2, p = 0.81). Both the KSS-F and PROMIS T-scores, respectively, were significantly higher in the PSI TKA patients (90 vs. 73, p = 0.002; 51.6 vs. 44.5, p = 0.011). However, after multivariable analysis adjusted for patient demographics and length of follow-up, no significant difference (p ≥ 0.16) remained for any PROMs between the groups (Table 3).

Table 3.

Comparison of patient reported clinical outcome measures according to type of total knee implant used.

Final Clinical Outcome Scores	ConforMIS iTotal® (N = 23)a	Stryker Triathlon® (N = 20)a	p-value**	Multivariable Analysis p-value†
PROMIS T-Score	52 (38–69)	45 (37–61)	0.01	0.45
KOOS Score	81 (65–100)	85 (54–100)	0.81	0.16
Knee Society Score	90 (45–100)	73 (45–100)	0.002	0.28
Follow-up (years)	5 (4.3–6.4)	5.2 (4.9–6.6)	0.32

**P-values from Wilcoxon rank sum test for continuous variables.

†Multivariable regression analysis performed adjusting for confounding variables (age, sex, and length of time from surgery); none of these continuous outcome measures were significantly different.

^aThe sample median (minimum-maximum) is given for continuous variables.

4. Discussion

Patient-specific implants for TKA represent an innovative attempt to minimize patient dissatisfaction.^5,6 Longer term data on outcomes and PROMs for PSI is not currently widely reported. According to our results, PSI did not infer significant improvement on clinical or patient reported outcomes compared to OTS implants at 5-year follow-up.

There were no demonstrated significant differences in ROM between implant systems. Indeed, our results indicate the absence of change in ROM between PSI and OTS should not be solely attributed to the implants. Gatha et al. evaluated preoperative range of motion, preoperative KSS, and multiple preoperative and postoperative radiographic findings (patellar height, knee alignment, and patellar thickness) and only preoperative ROM statistically correlated with postoperative ROM.¹⁸ Similarly, Bade et al. also reported that preoperative range of motion is the major predictor of postoperative ROM.¹⁹ These authors noted multiple other variables such as contractions of the extensor mechanism and/or periarticular soft tissues in patients with poor preoperative ROM likely contribute to subsequent reduced postoperative ROM.

Patient-specific instrument and implant systems have not proven superiority in the acute outcomes of operative time, blood loss, or hospital length of stay.9, 10, 11 With the numbers available in our series, PSI similarly did not infer additional benefit on perioperative clinical outcomes compared to OTS implants. Still, novel attempts to improve clinical outcomes and satisfaction through technology continue. Specifically, robotic-assisted TKA is designed to decrease errors in bone cuts, prosthesis positioning, and alignment, thus improving outcomes.²⁰ However, Kim et al. found no differences between robotic-assisted TKA and manual TKA regarding functional outcome scores, rates of aseptic loosening, overall survivorship, and operative complications at minimum 10-year follow-up.²¹ Augmented reality represents the newest technology to improve accuracy and outcomes, but present data remains limited.²² As ongoing technology will invariably continue to develop, we recommend cautious adaptation with realistic expectation setting of patients’ surgical outcomes.

The most widely reported complication with PSI has been demonstrated increased rates of arthrofibrosis requiring MUA.9, 10, 11, 12, 13 Similarly, our study found 22% PSI patients required MUA for arthrofibrosis, though was not statistically significant different compared to a 10% rate in OTS knees. Multiple factors may have existed for the varied rates, but perhaps none more so than the surgical learning curve associated with new implant design and technique. Chinnappa et al. retrospectively reviewed 86 consecutive PSI (guides) TKAs performed by one surgeon, including comparing the first 30 patients to the next 56 patients.²³ The authors were unable to detect a difference in both operative time and postoperative mechanical alignment with the use of PSI in TKA, thus concluding there is a minimal learning curve. In another study, De Gori et al. examined the cumulative summation (CUSUM) method to assess the learning curve for PSI in TKA.²⁴ It was determined that the learning curve for PSI in TKA was significant to achieve accuracy and obtain acceptable mechanical alignment. Although the length of the learning curve for PSI in TKA is controversial, we believe surgeons must individually consider its potential impact when implementing new technique or technologies.

As healthcare system implements value-based care incentives, PROMs have become an increasingly important indicator of patient satisfaction and function after TKA.²⁵ In our study, the KSS-F and PROMIS were improved in the PSI cohort; however, all 5-year PROMs of PSI TKAs were not significantly different than OTS TKA after multivariable analysis. It is important to note that a variety of PROMs and true functional testing (i.e., timed functional activities of daily living) exist and our study did not examine functional testing that might determine finite differences, particularly if larger numbers were available. Nonetheless, recent systematic and meta-analyses have utilized similar PROMs and demonstrated no difference when comparing PSI and OTS KSS function scores at multiple different postoperative time points up to 2 years.^26,27 Other reports have also shown no improved acute PROMs with the use of PSI, though OTS implants had more mechanical alignment outliers with associated inferior outcomes in those outliers.²⁸ Given such findings, PSI for TKA does not appear to significantly impact PROMs at short or midterm follow-up, thus surgeons must determine if other potential advantages (i.e., instrument process cost-savings or technical ease) provide value for PSI utilization.

Several limitations of this study should be noted. The retrospective design introduces inherent potential for selection biases into the data collection. Additionally, the sample size is relatively small; therefore, the possibility of a type II error (i.e., a false-negative finding) must be considered and results interpreted accordingly. This is a limited-surgeon series that utilized a single technique with CR implants, which have been associated with increased rates of MUA. Also, there was a potential learning curve with the utilization of PSI with an impact that is unknown. Despite these limitations, the data adds to a paucity of literature in providing a concise report of equivocal midterm patient-reported outcomes of PSI technology.

In conclusion, TKAs performed with PSI demonstrated no difference in postoperative ROM, reoperations, or PROMs compared to OTS implants at midterm follow-up. Surgeons may utilize the equivocal results during preoperative patient discussion of PSI in TKA. Further well-designed research with longitudinal follow-up will be required for both PSI and future innovative TKA technologies.

Declaration of competing interest

Each author has contributed to this original work by participating in writing.

This work nor a similar manuscript has been previously published.

This work is not currently under consideration by another journal and shall not be submitted for publication elsewhere while under consideration.

This work does not infringe upon any copyright or other proprietary right of any third party.

Authors will not post, archive, or disseminate the article by electronic, or any other means without the explicit permission of the publisher.