- Email: [email protected]
Effect of Implant Design on Knee Flexion
The Journal of Arthroplasty 28 (2013) 429–438
Contents lists available at SciVerse ScienceDirect
The Journal of Arthroplasty journal homepage: www.arthroplastyjournal.org
Effect of Implant Design on Knee Flexion Douglas A. Dennis MD a, R. David Heekin MD b, Charles R. Clark MD c, Jeffrey A. Murphy MS d, Tammy L. O'Dell CCRC, CCRA d, Kimberly A. Dwyer PhD, CCRA d a
Colorado Joint Replacement, Denver, Colorado Heekin Orthopaedics, Jacksonville, Florida University of Iowa Hospitals & Clinics, Orthopaedic Surgery 01075-JPP, Iowa City, Iowa d DePuy Orthopaedics, Warsaw, Indiana b c
a r t i c l e
i n f o
Article history: Received 3 June 2011 Accepted 2 July 2012 Keywords: Total Knee Arthroplasty (TKA) simultaneous bilateral flexion high flexion rotating platform mobile bearing knee knee implant design
a b s t r a c t From March 2006 to August 2008, 93 subjects (186 knees) underwent simultaneous bilateral total knee arthroplasty performed by eight surgeons at North American centers. This randomized study was conducted to determine whether non-weight-bearing passive flexion was superior for knees receiving a posterior stabilized high flexion device compared to a posterior stabilized standard device in the contra-lateral knee. Weight-bearing single leg active flexion was one secondary endpoint. Follow-up compliance was 92.5%. Results show small, but significant superiority in the motion metrics for the high flexion device compared to the standard device 12 months after surgery, especially for a subgroup of patients with pre-operative flexion less than 120° in both knees. Thus, the ideal candidate for the high flexion device may be one with lesser preoperative flexion. © 2013 Elsevier Inc. All rights reserved.
Orthopaedic surgeons are replacing knees in younger, more active subjects who have elevated expectations of implant performance. Expectations include restoration of joint function, including greater knee flexion [1–3]. To address these heightened expectations, many orthopaedic companies have enhanced their total knee arthroplasty (TKA) portfolios by including devices specifically designed to provide high flexion (HF) [4]. Early published results comparing flexion of HF to standard (STD) devices are mixed. Two randomized, controlled trials [5,6] comparing STD to HF knees of fixed tibial bearing, posterior stabilized designs found no significant differences in knee flexion between the groups, although Nutton et al. [6] admitted their unilateral study was underpowered. A third study [4], which was a matched-pair study comparing rotating platform, posterior stabilized STD to HF devices (of the same design as those used in this study) demonstrated enhanced flexion in those subjects implanted with the HF design, particularly in those subjects whose pre-operative knee flexion was less than 120°. Knee flexion after TKA is affected by multiple variables including implant design, surgical technique, physical therapy regimen, postoperative pain management protocol, and subject factors such as preoperative flexion, body habitus, and subject motivation [1,2,7,8]. Therefore, we designed this simultaneous bilateral study in which The Conflict of Interest statement associated with this article can be found at http:// dx.doi.org/10.1016/j.arth.2012.07.019. Reprint requests: Douglas A. Dennis, MD, Colorado Joint Replacement, 2535 S. Downing Street, #100, Denver, CO 80210. 0883-5403/2803-0009$36.00/0 – see front matter © 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.arth.2012.07.019
each subject would receive both devices and therefore, eliminate most of these variables, that is surgical technique, physical therapy regimen, post-operative pain management protocol, body habitus, and subject motivation, from interfering with any flexion differences that may exist between the HF and STD knee designs. Both the HF and STD designs were from the same knee implant family (P.F.C. Sigma Rotating Platform and P.F.C. Sigma Rotating Platform High Flexion, DePuy Orthopaedics, Warsaw, Indiana). We defined ‘simultaneous’ bilateral as implanted during the same bilateral TKA procedure. Consequently, the HF feature of the implant could be evaluated without confounding variables present in non-bilateral studies [9,10]. The primary objective of this study was to determine whether the HF design provided superior non-weight-bearing passive flexion (PF) compared to the STD design 6 and 12 months after surgery. Twelve months was chosen as the maximum time for endpoint comparison because most studies show no change in motion beyond the first year [5,7,11–14]. Secondary study objectives included identification of preoperative factors, such as body mass index (BMI), gender, age, thigh girth, and skin-fold thickness, which may be associated with improved flexion with this HF design. Materials and Methods Each subject received a standard (STD) device in one knee and a high flexion (HF) device in the other, eliminating most confounding variables present in non-bilateral studies [9,10]. From March 2006 to August 2008, data were collected prospectively for 93 subjects (186 knees) who underwent simultaneous bilateral total knee
430
D.A. Dennis et al. / The Journal of Arthroplasty 28 (2013) 429–438
arthroplasty (TKA) performed during the same operative procedure from eight experienced joint replacement specialists across eight North American centers. Allocation of subjects for enrollment was evenly distributed amongst the eight centers; however, there were varying degrees of enrollment attained at each center (see Fig. 1). Criteria for inclusion included males and females age 40–70, inclusive, subjects willing to return for all post-operative follow-ups, suitable bilateral candidates for the devices specified in the protocol, and voluntary, written informed consent. Subjects were excluded if they had a diagnosis of inflammatory arthritis, flexion contracture greater than 20°, involved in litigation or worker's compensation claims, known drug, alcohol or psychological disorders that could affect follow-up care, participation in a clinical investigation with an investigational product in the last three months, previous revision, and pregnant or lactating females. The 93 subjects enrolled in the study were considered the Safety Dataset and underwent adverse event analysis. Twelve of these 93 subjects did not complete the primary endpoint, non-weightbearing passive flexion (PF), 12 months after surgery (see Fig. 1).
One subject withdrew consent 10 months after the bilateral surgeries; the withdrawal was unrelated to either implant. There were two protocol violations in which the STD components (femurs with lugs) were not listed in the study protocol, but the appropriate HF devices were implanted on the contra-lateral side. One of the two subjects with a protocol violation had their eligible HF knee complete the primary endpoint; therefore, that knee was used in unpaired analyses. There were two revision TKA procedures, one HF device and one STD device. The HF device was revised 6 months after the index surgery and the STD at 7 months; both revisions were secondary to deep infection, and were performed at different centers. The remaining seven subjects were lost to follow-up, leaving 81 bilateral subjects available for the Primary Efficacy Dataset and respective analyses (Fig. 1), and provides a 92.5% (86 of 93) subject follow-up compliance rate. There were two protocol deviations in which 71- and 73-year old subjects were inadvertently enrolled when the study maximum specified age was 70. This age criterion was set to help ensure strong subjects were enrolled because of the invasive nature of this bilateral
Fig. 1. Subject accountability flowchart demonstrating two main datasets for analysis: Safety Dataset and Primary Efficacy Dataset.
D.A. Dennis et al. / The Journal of Arthroplasty 28 (2013) 429–438
431
Fig. 2. Secondary endpoint Single Leg Active Flexion (SLAF).
study. The researchers felt these small age deviations were of no real consequence to subject recovery or return to function, as the subjects were otherwise healthy, and do not believe they compromised the scientific validity of the study. Therefore, the two subjects were included in the Primary Efficacy Dataset. Endpoints were measured pre-operatively, for baseline assessment, and post-operatively at 2-weeks, 6-weeks, 6-months, and 12months. Flexion measurements were obtained by a qualified individual who was blinded to which knee received each device and appropriately trained to minimize measurement error. Because early post-operative outcomes are not the focus of this paper (those results are found elsewhere [15]) only results of the latter 2 evaluation periods are provided, along with metrics that assess their change from pre-operative measurements. Paired differences and 95% confidence intervals are included to provide effect size. Secondary endpoints included weight-bearing single leg active flexion (SLAF) (Fig. 2), American Knee Society (AKS) Score [16], Knee injury and Osteoarthritis Outcome Score (KOOS) [17,18], the anterior knee pain question from the patella score [19], and degree of patellofemoral crepitus with passive motion [20]. Because it was anticipated that subjects would be unable or unwilling to participate in SLAF early post-operatively, it was not collected at 2 and 6-weeks. Anterior knee pain [19] was rated using a scale of ‘none’, ‘mild’, ‘moderate’ or ‘severe’. The degree of patellofemoral crepitus was assessed using ‘none’, ‘fine’ or ‘coarse’ categories [20]. Some subjects had missing information for some of the secondary outcomes, especially pre-operative SLAF, as noted by the ‘n’ or denominator statistics reported throughout this paper. The missing information was not considered a limitation of the study because, in most cases, the missing information was shared on both knees within a subject, and the study was very amply powered and could withstand a substantially reduced sample size. The study protocol was approved by each site's Institutional Review Board (IRB), and written informed subject consent was obtained from all subjects prior to enrollment. The study was registered at www.clintrials.gov as study number NCT00380861. Treatment assignment for each knee was based on pre-operative PF measured laterally with subject supine on an examination table, using a full-length 12-inch (30.5 cm) goniometer (provided by the study sponsor for consistency across sites). This measurement was performed by a single qualified individual, at each site, blinded to the device received on each knee. Goniometric measurements of the knee have been shown to be reliable and valid compared to radiographs [20,21]. Pre-operatively, subjects identified which knee they considered their ‘worse’ knee, and the surgeon began the bilateral procedure on that knee. Both the HF and STD used the same patellar components and tibial tray design. The posterior stabilized femoral components and polyethylene inserts differed in their posterior condylar geometry, to accommodate different magnitudes of flexion (Fig. 3). Specifically, the HF tibial
insert post was placed 2 mm posterior compared to the STD insert with the goal of increasing posterior femoral translation, which has been associated with enhanced flexion. All components were cemented. Although not specified in the study protocol, a brief radiographic review was conducted by a blinded, independent radiographic reviewer, to ascertain the bone–cement–implant interface using both anterior posterior (AP) and lateral radiographic films for subjects at their 12-month follow-up interval. Previous intervals (e.g. 6months, 6-weeks) films were assessed, if necessary for confirmation of 12-month findings. Any signs of osteolysis, or radiolucencies and progression of osteolysis or radiolucencies were noted. The average subject age was 61.4 years (range, 44–73 years, standard deviation 5.8). Ninety-nine (92 of 93) percent of subjects had a diagnosis of osteoarthritis, 66% (61 of 93) were female, and the average Body Mass Index (BMI) was 31.6 (range 19–47, standard deviation 5.3). Setting a standard for future flexion studies, race and ethnicity were collected per a US Food and Drug Administration guidance document [22]. Ninety eight percent (91 of 93) were of the Caucasian race, and 99% (92 of 93) were of Non-Hispanic ethnicity. Because body mass index (BMI) [8] does not indicate the location of mass and body habitus is known to have an influence on knee flexion, several knee specific anthropometrics, such as thigh circumference and thigh skin-fold thickness, were collected to assess their
Polyethylene insert
Femoral Component
‘J’ curve for HF ‘J’ curve for STD Tibial Component
Fig. 3. The high flexion (HF — up to 155o of flexion) and standard (STD — up to 120o of flexion) implants came from the same product family. The J-Curve of the HF device has a tighter posterior radius for the HF design compared to the STD design.
432
D.A. Dennis et al. / The Journal of Arthroplasty 28 (2013) 429–438
Table 1 Pre-operative Knee and Subject Specific Anthropometrics and Other Covariates Examined to Assess their Potential Influence and Interaction on Post-Operative Outcomes. Knee Specific Thigh Circumference* (cm) Ankle Circumference (cm) Skin-fold† (mm) Quadriceps Strength‡ Hamstring Strength§ Pre-Op Flexion (PF & SLAF) Pre-Op AKS See Tables 3 & 4 Pre-Op KOOS sub-scores See Tables 3 & 4
Mean
SD (Range)
Subject Specific
Mean
SD (Range)
50.2 24.8 32.3 27.2% full strength 14.7% full strength See Tables 3 & 4
6.3 (37 to 71) 2.7 (16 to 33) 14.6 (4 to 60)
Age (years) Height (inches) Weight (lb) Body Mass Index (BMI) Pre-Op Motivation∥ Raceb Ethnicityb Gender
61.4 67 201.5 31.6 11.4 98% Caucasian 99% Non-Hispanic 66% Female
5.8 (44 to 73) 4.2 (60 to 76) 39 (115 to 305) 5.3 (19 to 47) 3.8 (0 to 17)
SD = Standard Deviation; PF = Passive Flexion; SLAF = Single Leg Active Flexion; AKS = American Knee Society; KOOS = Knee Injury and Osteoarthritis Outcomes Score. * Thigh circumference was measured 10 cm proximal to the superior aspect of the patella bone. † Skin-fold was measured in the middle of the thigh, 10 cm proximal to the superior patellar pole, using calipers. ‡ From a sitting position with knee bent 90o and a resistance band attached to their ankle, subjects attempted to achieve full extension, one knee at a time; 27.2% of the subjects achieved full extension. § From a sitting position with knee fully extended and a resistance band attached to their ankle, subjects attempted to achieve flexion of 90o, one knee at a time; 14.7% of the subjects achieved 90o flexion. ∥ Subjects were asked whether the following 17 flexion dependent activities were important to them pre-operatively: descending stairs, ascending stairs, rising from sitting, bending to floor/pick up an object, getting in/out of car, putting on socks/stockings, rising from bed, taking off socks/stockings, getting in/out of bath, sitting, getting on/off of toilet, heavy domestic duties, squatting, running, jumping, twisting/pivoting, and kneeling. The activities are a subset of the KOOS ADL and Sport and Recreation subscales. If the activity was 'Very' important, then the question was counted. On average, subjects marked 11.4 of these 17 activities as very important to them pre-operatively. b Race and ethnicity were planned to be included as covariates in the statistical modeling, however, because 98% of the subjects enrolled in this study were Caucasian and 99% were Non-Hispanic, there was not enough disparity to model properly.
impact on flexion. These anthropometrics along with a number of other covariates (Table 1) thought to affect flexion were collected preoperatively and examined to assess their potential influence on postoperative outcomes. All operative procedures were performed under regional or general anesthesia, using a standard anterior midline incision and a medial parapatellar arthrotomy. Patellar resection was consistent bilaterally. The technique of ligament balancing was individualized based on surgeon preference but was consistent for each site and identically applied to both knees for each subject. The surgical technique specified in the protocol ensured that close attention was paid to factors affecting the positioning of the implants during surgery. The protocol required that the difference in the thickness of the patellar reconstruction between the bilateral knees was no greater than 4 mm; and the implanted femoral components between the bilateral knees did not differ by more than one implant size. Intra-operatively, the surgeon deliberately resected the tibial posterior slope such that the planned difference between the right and left knees was not greater than 2°. Post-operative care was prescribed as per the surgeon's usual routine; however, the physical therapy and pain management were standardized across all sites. Even if differences existed across sites, the bilateral study design eliminated their impact in comparing the HF and STD knees within each subject. The type and frequency of adverse events were compared for all subjects (i.e., the Safety Dataset), and categorized as intra-operative, post-operative operative site and post-operative systemic. Based on a power analysis, it was determined that 128 subjects were required to significantly differentiate a PF difference of 5°. However, this power estimate was based on a 2-sided 2-sample t-test and was not based on the more appropriate paired (bilateral) data because none was available for rotating platform knees when this study was initially designed. After eighty-four (84) subjects had been accrued, the power analysis was re-calculated as a paired study endpoint using data from the study itself. The study data indicated the power to be greater than 99% to detect a 5° paired PF difference between the HF and STD groups (only 13 subjects were required for 80% power). Thus, it was decided to terminate enrollment at that time. By the time this information had been disseminated to the study sites and been approved by the IRB's, ninety-three (93) subjects had been accrued for the study. A subgroup of subjects with low pre-operative non-weightbearing PF in both knees was identified by placing subjects into
decreasingly lower pre-operative groupings, one degree at a time, until a threshold was found where post-operative PF differences were large between HF and STD, and enough subjects remained for statistical analysis. This algorithm yielded the threshold of 120° PF in both knees, and was supported by the same threshold used by Gupta et al. [4] in his unilateral study of the same devices. Twenty subjects met these criteria, 18 of whom completed the primary endpoint, yielding the Primary Efficacy Subgroup (see Fig. 1). Because of the increased power associated with pairing the data, this subgroup of 18 subjects had more than 85% power to detect a 5° difference in PF at the 12 month follow-up evaluation.
Statistical Methods Paired t-test analyses were conducted on continuous variables, whereas Fisher's Exact tests were used for categorical variables. When the comparison of groups or categories yielded p-values below 5%, it was considered statistically significant, whereas p-values from 5% and 10% were considered ‘borderline’ and identified as such. Regression and longitudinal analyses were carried out on postoperative data using PF and SLAF as dependent variables, with the devices and covariates (Table 1) as the independent variables. Longitudinal analyses utilized all post-operative evaluations. Plot smoothing was used for longitudinal graphs to provide continuity between follow-up intervals. The covariates were included to identify any pre-operative factors that would identify subjects who could benefit most from the HF devices. The MIXED procedure of SAS software (Cary, North Carolina, USA) was utilized for this modeling because it can fit covariate adjusted models with paired and unpaired data and can accommodate missing information. All potentially significant covariates were combined into a full covariate model, where non-parallel covariate effects were nested to accommodate different covariate regression slopes for the two devices. Covariate effects were reduced from this full model stepwise using an individual alpha above 0.05; non-parallel covariate effects were reduced from the model first when possible, followed by main effect covariates. Restricted maximum likelihood estimation was used in the final reduced model. SAS convergence issues were sometimes encountered in model reduction, and in these instances, minimum variance quadratic unbiased estimation was used. An unstructured covariance structure was used to model both the pairing of device measurements by subject and the repeated measurements over time. Either quadratic
D.A. Dennis et al. / The Journal of Arthroplasty 28 (2013) 429–438
433
Table 2 Primary Endpoint non-weight-bearing Passive Flexion for the Standard (STD) and High Flexion (HF) Knee Devices. Standard Metric Passive Flexion (°)
High Flexion
Interval
n
Mean
SD (Range)
n
Mean
SD (Range)
Pre-op 6 Months† 12 Months Δ Pre-op to 6 Months Δ Pre-op to 12 Months
91* 74 81 74 81
122.8 124.0 126.8 1.01 3.83
10.2 (90 to 146) 10.6 (90 to 147) 9.8 (89 to 142) 11.5 (−33 to 25) 11.6 (−38 to 42)
93 75 82‡ 75 82‡
122.5 124.2 127.8 2.16 5.37
11.0 (85 to 144) 11.2 (70 to 142) 9.2 (90 to 145) 14.1 (−58 to 35) 11.7 (−38 to 38)
Paired Difference (HF-STD) −0.20 0.23 1.11 1.11 1.61
95% CI of Difference −2.03 −1.18 −0.15 −1.39 −0.37
to to to to to
1.64 1.64 2.38 3.60 3.60
Significant Difference? No No Borderline No No
Paired t-test 2-sided P-value 0.831 0.747 0.084 0.759 0.109
SD = standard deviation; CI = confidence interval; Δ = change from. * Two STD devices were not among the approved components listed within the study protocol. † The 6-month follow-up interval had a smaller interval window compared to the 12 month interval and this proved challenging for sites to get all of their subjects back for evaluation; hence, the smaller number of subjects with follow-up in this interval. ‡ One subject had 12 Month follow-up for his HF knee, but his STD knee was ineligible. The HF knee is included in unpaired analyses.
or linear longitudinal models were utilized based on the evident shape of responses over time. Source of Funding Funding for this study was provided by DePuy Orthopaedics (Warsaw, Indiana, USA).
represented the predictor variable. Sites with less than 10 subjects were pooled together. Model p-values were 0.103 and 0.544 for the 6and 12-month follow-up intervals, respectively, validating no differences between any of the sites. Balanced results across each side of the patient were investigated using a t-test. Resulting p-values for both 6- and 12-month follow-up intervals were 0.941 and 0.371, respectively, indicating side had no impact on the observed results.
Results Primary Endpoint: Non-weight-bearing Passive Flexion A total of 81 subjects, deemed the Primary Efficacy Dataset (Fig. 1), were evaluated for the primary endpoint of non-weight-bearing passive flexion (PF) 12 months after surgery. An additional five subjects met protocol defined study withdrawal criteria before reaching their 12 month follow-up evaluation, providing a subject follow-up compliance rate of 92.5% (86 of 93). Confirmation of no site to site differences was verified using a oneway analysis of variance (ANOVA) test where paired-differences of the primary endpoint PF was used as the response variable and sites
There were no statistically significant differences between the high flexion (HF) and standard (STD) devices for PF at any of the follow-up intervals, including pre-op, or change metrics according to paired t-test analyses (see Table 2). However, at the 12 month post-operative evaluation interval, there was a paired 1.11° borderline (P = .084) advantage for the knees receiving the HF device. Regression analysis identified pre-operative PF as a significant covariate (see Table 1) at both post-operative evaluations, while pre-op American Knee Society
Table 3 Secondary continuous Endpoints for the Standard (STD) and High Flexion (HF) Knee Devices. SLAF, KOOS, and AKS Outcome Metrics for STD and HF Knee Devices Standard Metric Single Leg Active Flexion (°) KOOS Pain (points) KOOS Symptoms (points) KOOS ADL (points)
KOOS Sport & Rec (points) KOOS Quality of Life (Points) AKS Knee Score (Points)
Interval
N
Mean
Pre-op 6 Months† 12 Months Pre-op 6 Months† 12 Months Pre-op 6 Months† 12 Months Pre-op 6 Months† 12 Months Pre-op 6 Months† 12 Months Pre-op 6 Months† 12 Months Pre-op 6 Months† 12 Months
27* 70 73 86 67 79 86 70 79 79 69 75 82 48 54 90 72 78 91 74 80
124.6 126.0 126.6 45.7 84.0 89.5 50.7 76.8 82.5 49.0 85.4 90.4 19.9 56.5 68.1 19.2 65.0 74.9 48.1 86.1 85.6
SD (Range) 11.1 11.0 10.6 19.2 14.8 14.0 20.2 13.7 11.9 20.2 14.0 11.8 21.8 28.9 24.3 17.3 21.4 22.1 19.0 11.5 11.5
(108 to 148) (90 to 147) (88 to 143) (0 to 94) (42 to 100) (28 to 100) (4 to 93) (46 to 100) (46 to 100) (1 to 100) (38 to 100) (31 to 100) (0 to 100) (0 to 100) (5 to 100) (0 to 94) (13 to 100) (6 to 100) (2 to 100) (48 to 100) (54 to 100)
High Flexion n
Mean
31* 72 74 88 68 80 89 68 82 83 69 75 82 48 55 92 72 78 93 74 81
120.9 125.2 127.5 45.3 83.1 90.0 49.9 75.5 83.1 47.9 85.6 90.1 19.0 55.4 69.3 19.4 64.4 75.6 47.4 84.5 87.0
SD (Range) 13.3 10.8 11.1 18.2 14.9 11.8 18.9 14.5 13.1 19.2 12.7 11.3 19.2 26.7 24.4 15.9 21.2 22.2 17.4 12.5 10.7
(82 to 150) (90 to 142) (82 to 145) (6 to 89) (39 to 100) (53 to 100) (4 to 89) (29 to 93) (43 to 100) (4 to 90) (43 to 100) (43 to 100) (0 to 100) (0 to 100) (5 to 100) (0 to 69) (13 to 100) (19 to 100) (13 to 100) (40 to 100) (56 to 100)
Paired Difference (HF-STD) −2.26 −0.54 1.23 −0.48 −0.95 0.71 −1.08 −0.89 0.54 −1.43 0.15 −0.29 −1.42 −0.53 0.93 0.42 −0.61 0.65 −1.41 −1.55 1.21
95% CI of Difference
Significant Difference?
Paired 2-sided t-test P-value
−6.24 to 1.72 −1.77 to 0.69 0.14 to 2.32 −3.51 to 2.54 −3.09 to 1.19 −1.41 to 2.84 −4.09 to 1.93 −3.47 to 1.68 −0.86 to 1.95 −4.28 to 1.41 −1.59 to 1.89 −1.60 to 1.01 −3.93 to 1.09 −2.45 to 1.39 −1.06 to 2.91 −1.82 to 2.65 −2.63 to 1.41 −1.24 to 2.54 −4.19 to 1.38 −4.46 to 1.35 −1.05 to 3.47
No No Yes No No No No No No No No No No No No No No No No No No
0.254 0.382 0.027 0.751 0.377 0.506 0.478 0.491 0.444 0.319 0.865 0.655 0.264 0.579 0.354 0.712 0.551 0.496 0.318 0.290 0.288
SD = standard deviation; CI = confidence interval; ADL = Activities of daily living; AKS = American Knee Society; SLAF = Single Leg Active Flexion; KOOS = Knee Injury and Osteoarthritis Outcome Score. * Most subjects were unable or unwilling to participate in SLAF pre-operatively because of pain. Because of the study's high power (N99%), the comparison with the reduced number still had adequate power (N80%). † The 6-month follow-up interval had a smaller interval window compared to the 12-month interval and this proved challenging for sites to get all of their subjects back for evaluation; hence, the smaller number of subjects with follow-up in this interval.
434
D.A. Dennis et al. / The Journal of Arthroplasty 28 (2013) 429–438
Table 4 Primary Endpoint non-weight-bearing Passive Flexion for the Standard (STD) and High Flexion (HF) Knee Devices for Subjects with Pre-op Passive Flexion b 120° in Both Knees. Standard (STD)
Passive Flexion (°)
High Flexion (HF)
Interval
n
Mean
SD (range)
n
Mean
SD (range)
Paired Difference (HF-STD)
Pre-op 6 Months* 12 Months Δ Pre-op to 6 Months Δ Pre-op to 12 Months
19 16 18 16 18
111.6 118.9 120.2 7.06 8.22
6.0 (95 to 118) 13.9 (90 to 140) 11.3 (89 to 140) 13.1 (−18 to 25) 11.3 (17 to 31)
20 17 18 17 18
111.4 120.9 124.7 9.18 13.06
5.8 (102 to 119) 12.7 (92 to 142) 9.8 (108 to 145) 12.1 (−16 to 26) 9.6 (−6 to 28)
−0.42 2.06 4.56 2.38 4.83
Metric
95% CI of Difference −2.82 −0.30 1.29 −1.46 1.64
to to to to to
1.97 4.43 7.82 6.21 8.03
Statistically Significant?
Paired 2-sided t-test P-value
No Borderline Yes No Yes
0.716 0.083 0.009 0.207 0.005
SD = standard deviation, CI = confidence interval, Δ = change from. * The 6-month follow-up interval had a smaller interval window compared to the 12 month interval and this proved challenging for sites to get all of their subjects back for evaluation; hence, the smaller number of subjects with follow-up in this interval.
(AKS) Score was a significant covariate at 6 months and skin-fold thickness was a significant covariate at 12 months. The impact of the significant covariates was minimal with the covariate adjusted pvalue for PF at 6 months changing from 0.747 (see Table 2) to 0.724, and the covariate adjusted p-value for PF at 12 months changing from 0.084 to 0.079. Secondary Endpoints Results of continuous secondary endpoints are provided in Table 3. There were no significant differences between the two knees preoperatively for any of the endpoints. Weight-Bearing Single Leg Active Flexion (SLAF) There was no significant difference between the HF and STD devices 6 months post-operatively, however, SLAF was significantly greater (P = .027) in the HF knees at 12 months with a paired difference of 1.23°. Knee Injury and Osteoarthritis Outcome Score (KOOS) and American Knee Society (AKS) There were no significant differences between knees for the AKS knee score or any of the five KOOS subscales pre-operatively, or at 6 or 12 months post-operatively. The AKS functional scores are not included in Table 3 because that outcome involves both knees. Preoperatively, average AKS function scores were 54.9 points (SD = 15.9, range 5 to 100). Post-operatively average AKS function scores were 83.9 points (SD = 18.4, range 15 to 100) at 6 months and 91.6 points (SD = 11.2, range 45 to 100) at 12 months. Anterior Knee Pain There were no statistical differences pre-operatively, or at 6 or 12 months post-operatively for anterior knee pain when comparing the proportions between the two knee devices. As such, pooled statistics between HF and STD are provided here: Pre-operatively 16% (30 of 184) of subjects had severe anterior knee pain, 41% (76 of 184) had moderate, 15% (27 of 184) had mild, and 28% (51 of 184) had no anterior knee pain. Six months after the surgical procedures, 79% (119 of 150) of subjects had no anterior knee pain, 18% (27 of 150) had mild, 2% (3 of 150) had moderate, and b1% (1 of 150) had severe anterior knee pain. At the 12 month post-operative interval, 79% (113 of 143) of subjects had no anterior knee pain, 17% (24 of 143) had mild, 4% (6 of 143) had moderate, and no subject had severe anterior knee pain. Patellar Crepitus There were no statistical differences pre-operatively, or at 6 or 12 months post-operatively in the degree of incidence for crepitus when comparing the two knee devices. Thus, combined HF and STD statistics are as follows: Pre-operatively 48% (87 of 182) of subjects had coarse crepitus, 37% (68 of 182) had fine crepitus, and 15% (27 of 182) had no crepitus. Six months post-operatively, 56% (84 of 150) of subjects had no crepitus, 34% (52 of 150) had fine crepitus, and 9% (14
of 150) had coarse crepitus. At the 12 month interval, 70% (97 of 139) of subjects had no crepitus, 25% (35 of 139) had fine crepitus, and 5% (7 of 139) had coarse crepitus. Longitudinal Analysis Longitudinally (across post-operative time) there was no significant difference between knees for both flexion outcomes: PF and SLAF. Subgroup Analyses Table 4 provides the non-weight-bearing PF results for a subgroup of 18 subjects whose pre-operative PF was less than 120° in both knees (see Fig. 1). The knees with the HF device (124.7°) were statistically superior (P = .009) compared to the STD knee devices (120.2°) 12 months after surgery yielding a paired difference of 4.56° for the Primary Efficacy subgroup. Similar results were observed for the change from pre-operative to 12 months post-op: HF was superior to STD by 4.83° with P = .005. There were no significant covariates. Significant advantages for the HF device were observed for weight-bearing SLAF at both 6 months (means: HF = 121.1°, STD = 118.3°; P = .038) and 12 months (means: HF = 121.3°, STD = 118.9°; P = .040) after surgery. There were no significant differences observed between the STD and HF devices for KOOS and AKS scores. Longitudinally (Fig. 4), the HF knees had borderline superior PF (P = .067) compared to the STD knees. The HF knees were significantly superior compared to the STD knees for SLAF (P = .029). Significant covariates from the longitudinal models included pre-operative PF for the post-operative PF response, and thigh circumference for the post-operative SLAF response. The impact of the latter significant covariate was minimal with the longitudinal covariate adjusted p-value for SLAF changing from 0.029 to 0.069. The impact of significant covariate pre-operative PF on the longitudinal postoperative PF was more dramatic with the p-value changing from 0.067 to 0.477. On average, each 1° pre-operative PF increase yielded a 0.5° increase in post-op PF. Radiographic Review No osteolysis was observed for any of the HF or STD knees. No radiolucencies ≥ 2 mm were observed on the AP view. Two subjects, one involving their HF device and the other involving their STD device, had radiolucencies ≥ 2 mm on their lateral film. These were interpreted as interface gaps that did not progress from immediate to 12-month post-op films; the cause was attributed to sub-optimal cement technique. Adverse Events There was 1 intra-operative adverse event: 1 STD knee required a medial collateral ligament reconstruction with a soft tissue washer and screw, which was resolved at the end of surgery. There were 52 systemic complications (Table 5a). There were 36 post-operative operative site adverse events (Table 5b). The type, frequency,
D.A. Dennis et al. / The Journal of Arthroplasty 28 (2013) 429–438
435
Passive Flexion (PF)
P-value = .067 Significant Covariate: Pre-operative PF
Smoothed line
The impact of significant covariate pre-operative PF on longitudinal post-operative PF was somewhat dramatic with the covariate adjusted P-value changing from .067 to .477. The regression model tells us for each 1 degree pre-operative PF increase an increase of 0.5 degrees is expected in post-operative PF.
Single Leg Active Flexion (SLAF)
P-value = .029 Significant Covariate: Thigh circumference
Smoothed line
SLAF was only measured 6 and 12 months post-operatively because earlier time intervals interfered with subject’s rehabilitation. The impact of significant covariate thigh circumference was minimal with the longitudinal covariate adjusted P-value changing from .029 to .069.
Fig. 4. Longitudinal plots of post-operative flexion metrics in subjects with b 120° pre-operative passive flexion in both knees.
proportion, and 95% exact confidence interval of post-operative systemic and operative site adverse events are noted in Table 5a and 5b respectively. There were no significant differences in operative site adverse events between the two implants as noted by the Fishers Exact Test p-values greater than 0.05 (Table 5b). Revisions and Reoperations Four knees (3 patients) were revised, 2 that received the HF device and 2 that received the STD device. Two of the revisions, 1 HF and 1 STD, were secondary to deep infections that occurred in different subjects 6 months (HF) and 7 months (STD) after the index surgery,
respectively. These two revisions prevented the subjects from reaching the primary endpoint of 12 month post-operative follow-up and are, therefore, accounted for in Fig. 1. The other 2 revisions were secondary to tibial tray loosening and occurred in the same subject in both knees, after the 12 month primary endpoint had been collected. The HF knee was revised 19 months after the index surgery whereas the STD knee was revised 21.5 months after the index surgery. One additional subject had a reoperation where the tibial polyethylene insert was exchanged 7 weeks after the index surgery due to polyethylene ‘spin out’. The ‘spin out’ was discovered radiographically in the HF knee at a routine follow-up visit, as the
436
D.A. Dennis et al. / The Journal of Arthroplasty 28 (2013) 429–438
Table 5a Post-Operative Systemic Adverse Events (includes both knees). Post-operative Systemic Adverse Event
Frequency
Other* Musculoskeletal Respiratory Gastrointestinal Genitourinary Vascular — Pulmonary Embolism Vascular — Deep Vein Thrombosis Cardiac Endocrine/Metabolic Central Nervous System Dermatological Head, Ears, Eyes, Nose and Throat Hematological Peripheral Nervous System Total
Percent of Total Frequency (N = 93)
11 9 8 6 4 3 2 2 2 1 1 1 1 1 52
95% Exact Confidence Interval
11.8% 9.7% 8.6% 6.5% 4.3% 3.2% 2.1% 2.1% 2.1% 1.1% 1.1% 1.1% 1.1% 1.1% 55.9%
6.1–20.2 4.5–17.6 3.8–16.3 2.4–13.5 1.2–10.7 0.7–9.1 0.3–7.6 0.3–7.6 0.3–7.6 0.0–5.9 0.0–5.9 0.0–5.9 0.0–5.9 0.0–5.9 45.2–66.2
* ‘Other’ systemic adverse events included: back spasms, right ankle pain, right wrist carpal tunnel, 2 electrolyte imbalances, sciatica, chest pain, infected PICC line, cellulitis right arm, left hand/arm weakness, right arm sensation at IV site after blood transfusion. Adverse Events are counted based on a per subject basis; that is, duplicate reports for resolving issues within a single subject are not counted twice.
patient was asymptomatic. When questioned, the patient recalled a challenging time in physical therapy several weeks beforehand. The patient received a thicker tibial polyethylene insert during the reoperation and both knees completed the 12-month followup interval. The adverse events associated with these revisions and reoperations are included in Table 5b. Discussion The primary objective of this study was to determine whether the high flexion (HF) design provided superior non-weight-bearing passive flexion (PF) compared to the standard (STD) design 12 months after surgery. This study showed small, but borderline statistically significant improvements in non-weight-bearing PF and weight-bearing single leg active flexion (SLAF) for the knees receiving the HF device compared to the STD device 12 months postoperatively. The small magnitude of improvement in our study may not be clinically significant. The bilateral study design allowed subtle motion improvements in the HF compared to STD to become evident that typical unilateral study designs are unable to detect. Superiority in flexion was more pronounced in a subgroup of subjects with less than 120 o of pre-operative flexion in both knees, suggesting the ideal candidate for an HF TKA may be one with lesser pre-
operative flexion. Biomechanical studies do show that if high flexion is obtained by a subject implanted with an HF design, reduced polyethylene stresses [23,24] and lower patellar ligament strains [25] are incurred when compared to STD designs, which may enhance long term performance. The recently published paper by Choi, Lee et al., which compared the same two knee devices as our study, in an Asian population of unilateral subjects, reported no differences between the devices. The authors noted their high pre-operative motion (126° of active maximal flexion) may have obscured the motion advantage of the high-flexion design [26]. This agrees with our study results where the advantage of the HF device over the STD device was more pronounced in subjects with less than 120° of pre-operative flexion in both knees. Pre-operative PF was a significant covariate for several outcomes (Tables 2 and 4) indicating its large influence on the post-operative measurements, which is consistent with many other reports [6,9,25]. Pre-operative American Knee Society (AKS) scores, skin-fold thickness, and thigh circumference were other covariates with some influence on post-operative outcomes, but their impact was minimal, which may be a result of our homogeneous study group (that is, 98% Caucasian, 99% Non-Hispanic, average body mass index of 31.6). The rest of the covariates tested (Table 1) were not significantly influencing post-operative outcomes.
Table 5b Post-Operative Operative Site Adverse Events for the Standard (STD) and High Flexion (HF) Knee Devices. Standard (STD) Post-operative Operative Site Adverse Event Other Pain Infection Effusion Patellar Crepitus/Clunk Bone Fracture Loosening Polyethylene ‘spin out’ Arthrofibrosis Total
High Flexion (HF)
STD Frequency
STD Percent of Total Frequency (N = 93)
STD 95% Exact Confidence Interval
6* 3 1 0 2 1 1 0 0 14
6.6% 3.2% 1.1% 0.0% 2.2% 1.1% 1.1% 0.0% 0.0% 15.1%
2.4–13.5 0.7–9.1 0.0–5.8 0.0–0.0 0.3–7.6 0.0–5.8 0.0–5.8 0.0–0.0 0.0–0.0 8.5–24.0
HF Frequency 10† 3 3 2 0 1 1 1 1 22
HF Percent of Total Frequency (N = 93) 10.8% 3.2% 3.2% 2.2% 0.0% 1.1% 1.1% 1.1% 1.1% 23.7%
HF 95% Exact Confidence Interval
Fishers Exact Test P-value
5.3–18.9 0.7–9.1 0.7–9.1 0.3–7.6 0.0–0.0 0.0–5.8 0.0–5.8 0.0–5.8 0.0–5.8 15.5–33.6
0.437 1.000 0.621 0.497 0.497 1.000 1.000 1.000 1.000 0.275
Every unique adverse event was reported once, regardless of whether a single knee reported more than one instance of a particular adverse event. For example, if a knee reported ‘infection’ twice, then infection was counted once for that knee. However, if that same knee had ‘effusion’, then that adverse event was counted in addition to the infection adverse event. * STD ‘Other’ operative site adverse events included: stiffness during physical therapy, swelling, contusion, twisted knee w/pain, retained drain fragment, pulled hamstring. † HF ‘Other’ operative site adverse events included: sciatica (right knee), tingling/burning incision area, quadriceps strain in PT, deep orthopaedic drain site bleed, stiffness, leg swelling, abrasion, saphenous neuritis, right MCL instability, fall.
D.A. Dennis et al. / The Journal of Arthroplasty 28 (2013) 429–438
Differences in the AKS score and Knee Injury and Osteoarthritis Outcomes Score (KOOS) scores, as a function of prosthesis design, were not observed at 6 or 12 month post-operative follow-up intervals (Table 3). The AKS Score has been criticized for having a pronounced ceiling effect and inadequate sensitivity, whereas Knee Injury and Osteoarthritis Outcome Score (KOOS) has been touted as being more sensitive [27] in detecting differences in knee devices. Nevertheless, no differences were observed in the current study, perhaps because of the bilateral design where subjects may have trouble articulating differences between their right and left legs. The success of high-flexion knee implant systems has been inconsistent in the literature. In a mixture of retrospective case– control and prospectively randomized control studies, some authors have demonstrated a beneficial effect from 6 o to 13 o [4,9,28] while other authors have demonstrated no difference between HF and STD knee systems [5,6,11,14,26,26,29–31]. With the exception of two studies [5,30], who also studied bilateral subjects, these series have more confounding variables inherent with unilateral studies. Gupta et al.[4] in a matched pair study of subjects implanted with the same devices as the current study, similarly observed the greatest knee flexion gains occurred in HF subjects with pre-operative flexion less than 120°. There are concerns that the increased flexion that may result from HF TKA design modifications may be detrimental to the long term survival of these implant designs[14]. Increased flexion has been shown to increase peak polyethylene stresses in traditional TKA designs, theoretically increasing the risk of premature polyethylene wear. There were no instances of polyethylene wear in the current study, although follow-up duration is short, being 12 months. High flexion magnitudes are also associated with increased tension in the extensor mechanism which could result in increased anterior knee pain or patellar component fixation failure. We saw no differences in anterior knee pain in the current study. Reduced conformity of the posterior aspect of the tibial insert might risk an increased rate of flexion instability. Many of the current HF TKA designs, including the one in this study, require increased posterior femoral condylar bone resection which is a disadvantage should revision arthroplasty be necessary. Lastly, the increased loads of high knee flexion place increased stresses on the fixation interface, theoretically increasing the risk of component loosening. Han et al. [3] reviewed a cohort of 47 subjects implanted with 72 posterior stabilized-flex, fixed-bearing (HF) implants and observed an alarmingly high incidence of aseptic femoral component loosening of 38% (27 of 72) at a follow-up duration of 32-months. The postoperative mean maximum flexion was 136° in the group with femoral component loosening versus 125° in the well-fixed group (P = .022). Cho et al. [32] reviewed a group of 166 patients (218 TKA) implanted with the same HF TKA implants as the Han et al. [3] study with a mean follow-up duration of 51 months. Cho et al. [32] reported 7 of 218 cases (3.2%) required revision for femoral component loosening at a mean follow-up duration of 49months. In the current study, the incidence of loosening (Table 5b) was low, and equal between the HF and STD devices. Because of these concerns, continued long-term follow-up evaluation of these devices is necessary. Strengths of this study are the enrollment of prospective bilateral subjects in whom treatment and control are contained within a single subject, the standardized surgical technique, physical therapy and pain management protocols, and the detailed statistical analysis methods. Limitations include the short-term follow-up duration, the lack of diversity in ethnicity and race, and the limited radiographic review. The study was designed as a 12 month follow-up study because the maximum achievable non-weight-bearing passive flexion, the primary endpoint, was expected to be achieved by almost all subjects before the 12 month follow-up interval, which is supported as a reasonable time endpoint in the published literature [5,7,11–14]. Indeed, one report [7] stated ‘Most studies examining ROM have
437
shown that no significant improvement occurs in ROM beyond the first year.’ The lack of ethnic and racial diversity was not unexpected given the study took place at 8 U.S. centers and was not stratified for race or ethnicity. If this study had included subjects with high flexion activities of daily living, such as Asians, greater differences between the HF and STD may have been realized. Goniometer measurements have been shown to be accurate and reproducible and the strict criteria imposed on the surgical technique would likely not allow component mal-position to be a radiographic concern[21]. A brief radiographic review was conducted where no osteolysis was observed and only two cases of radiolucencies ≥ 2 mm were observed, one of each knee device, which were attributed to suboptimal cement technique. Future studies could consider radiographic evaluation of alignment, specifically femoral flexion and posterior tibial slope (which could affect knee flexion), to further validate knee flexion studies. The findings of this study are limited to the implant devices used and the bilateral subjects utilized, and may not extrapolate to other HF TKA designs, or the more commonly studied unilateral subject population. Lastly, the findings are also limited to the specific, Western, predominantly Caucasian, subject population. Acknowledgments The authors would like to thank the following for their integral contributions: Richard Driessnack, MD (Vanderbilt University School of Medicine, Nashville, TN), Dean Sukin, MD (Montana Orthopaedics and Sports Medicine, Billings, MT), Steven Teeny, MD (Puget Sound Orthopaedics, Lakewood, WA), William Bugbee, MD (Scripps Clinic, LaJolla, CA), Lawrence Housman, MD (Tucson Orthopaedic Institute , Tucson, AZ), James Lesko, PhD (DePuy Orthopaedics, Warsaw, IN) for statistical analysis assistance, Steven Tippett, PhD (Bradley University, Peoria, IL) for the physical therapy protocol, input on data collection and training of the study coordinators on collection of physical measurements, Thomas Gruen, MS (Zonal Concepts, Wesley Chapel, FL). And, the study coordinators for their dedication to the implementation of the protocol: Jane Mang, RN, BSN, MAT (Colorado Joint Replacement, Denver, CO), Heather Whetsell, RN, (Great Plains Sports Medicine & Rehabilitation Center Peoria, IL), Terry Anderson, (Great Plains Sports Medicine & Rehabilitation Center, Peoria, IL), Jennifer Crews, RN, BSN (Heekin Orthopaedics, Jacksonville, FL), Linda Fink, RN, MSN (University of Iowa Hospitals & Clinics, Iowa City, IA), Judy Miller, RN, CCRP (Montana Health Research Institute, Billings, MT), Sally York (Lakewood Orthopaedic Surgeons, Lakewood, WA). References 1. Ranawat CS. Design may be counterproductive for optimizing flexion after TKR. Clin Orthop Relat Res 2003(416):174. 2. Sultan PG, Most E, Schule S, et al. Optimizing flexion after total knee arthroplasty: advances in prosthetic design. Clin Orthop Relat Res 2003(416):167. 3. Han HS, Kang SB, Yoon KS. High incidence of loosening of the femoral component in legacy posterior stabilised-flex total knee replacement. J Bone Joint Surg Br 2007; 89(11):1457. 4. Gupta SK, Ranawat AS, Shah V, et al. The P.F.C. sigma RP-F TKA designed for improved performance: a matched-pair study. Orthopedics 2006;29(9 Suppl):S49. 5. Kim YH, Sohn KS, Kim JS, et al. Range of motion of standard and high-flexion posterior stabilized total knee prostheses. A prospective, randomized study. J Bone Joint Surg Am 2005;87(7):1470. 6. Nutton RW, van der Linden ML, Rowe PJ, et al. A prospective randomised doubleblind study of functional outcome and range of flexion following total knee replacement with the NexGen standard and high flexion components. J Bone Joint Surg Br 2008;90(1):37. 7. Anouchi YS, McShane M, Kelly Jr F, et al. Range of motion in total knee replacement. Clin Orthop Relat Res 1996(331):87. 8. Jette M, Sidney K, Lewis W. Fitness, performance and anthropometric characteristics of 19,185 Canadian Forces personnel classified according to body mass index. Mil Med 1990;155(3):120. 9. Weeden SH, Schmidt R. A randomized, prospective study of primary total knee components designed for increased flexion. J Arthroplasty 2007;22(3):349. 10. Wohlrab D, Ditl J, Herrschelmann R, et al. Does the NexGen LPS flex mobile knee prosthesis offer advantages compared to the NexGen LPS? A comparison of clinical and radiological results. Z Orthop Ihre Grenzgeb 2005;143(5):567.
438
D.A. Dennis et al. / The Journal of Arthroplasty 28 (2013) 429–438
11. Minoda Y, Aihara M, Sakawa A, et al. Range of motion of standard and high-flexion cruciate retaining total knee prostheses. J Arthroplasty 2009;24(5):674. 12. Insall JN, Hood RW, Flawn LB, et al. The total condylar knee prosthesis in gonarthrosis. A five to nine-year follow-up of the first one hundred consecutive replacements. J Bone Joint Surg Am 1983;65(5):619. 13. Schurman DJ, Parker JN, Ornstein D. Total condylar knee replacement. A study of factors influencing range of motion as late as two years after arthroplasty. J Bone Joint Surg Am 1985;67(7):1006. 14. McCalden RW, Macdonald SJ, Bourne RB, et al. A randomized controlled trial comparing "high-flex" vs "standard" posterior cruciate substituting polyethylene tibial inserts in total knee arthroplasty. J Arthroplasty 2009;24(6 Suppl):33. 15. Tippett SR, Mang J, Dwyer KA, et al. Collecting data with Palm technology: comparing preoperative expectations and postoperative satisfaction in patients undergoing total knee arthroplasty. J Bone Joint Surg Am 2010;92(Suppl 2):59. 16. Insall JN, Dorr LD, Scott RD, et al. Rationale of the Knee Society clinical rating system. Clin Orthop Relat Res 1989(248):13. 17. Roos EM, Roos HP, Lohmander LS, et al. Knee Injury and Osteoarthritis Outcome Score (KOOS) — development of a self-administered outcome measure. J Orthop Sports Phys Ther 1998;28(2):88. 18. Roos EM, Lohmander LS. The Knee injury and Osteoarthritis Outcome Score (KOOS). from joint injury to osteoarthritis. Health Qual Life Outcomes 2003;1:64. 19. Feller JA, Bartlett RJ, Lang DM. Patellar resurfacing versus retention in total knee arthroplasty. J Bone Joint Surg Br 1996;78(2):226. 20. Cibere J, Bellamy N, Thorne A, et al. Reliability of the knee examination in osteoarthritis: effect of standardization. Arthritis Rheum 2004;50(2):458. 21. Gogia PP, Braatz JH, Rose SJ, et al. Reliability and validity of goniometric measurements at the knee. Phys Ther 1987;67(2):192. 22. FDA. Guidance for industry. Collection of race and ethnicity used in clinical trials; 2005.
23. Zelle J, der Zanden AC Van, De Waal MM, et al. Biomechanical analysis of posterior cruciate ligament retaining high-flexion total knee arthroplasty. Clin Biomech (Bristol, Avon) 2009;24(10):842. 24. Barink M, De Waal MM, Celada P, et al. A mechanical comparison of high-flexion and conventional total knee arthroplasty. Proc Inst Mech Eng [H ] Part H - J Eng Med 2008;222(3):297. 25. Gejo R, Morita Y, Matsushita I, et al. Intraoperative patellar tendon strain: predicting the range of knee flexion after total knee arthroplasty. J Orthop Sci 2009; 14(1):51. 26. Choi WC, Lee S, Seong SC, et al. Comparison between standard and high-flexion posterior-stabilized rotating-platform mobile-bearing total knee arthroplasties: a randomized controlled study. J Bone Joint Surg Am 2010;92(16):2634. 27. Lingard EA, Katz JN, Wright RJ, et al. Validity and responsiveness of the Knee Society Clinical Rating System in comparison with the SF-36 and WOMAC. J Bone Joint Surg Am 2001;83-A(12):1856. 28. Huang HT, Su JY, Wang GJ. The early results of high-flex total knee arthroplasty: a minimum of 2 years of follow-up. J Arthroplasty 2005;20(5):674. 29. Seon JK, Song EK, Lee JY, et al. Comparison of range of motion of high-flexion prosthesis and mobile-bearing prosthesis in total knee arthroplasty. Orthopedics 2005;28(10 Suppl):s1247. 30. Ng FY, Wong HL, Yau WP, et al. Comparison of range of motion after standard and high-flexion posterior stabilised total knee replacement. Int Orthop 2008;32(6): 795. 31. Malik A, Salas A, Ben AJ, et al. Range of motion and function are similar in patients undergoing TKA with posterior stabilised and high-flexion inserts. Int Orthop 2009. 32. Cho SD, Youm YS, Park KB. Three- to six-year follow-up results after high-flexion total knee arthroplasty: can we allow passive deep knee bending? Knee Surg Sports Traumatol Arthrosc 2010.
Contents lists available at SciVerse ScienceDirect
The Journal of Arthroplasty journal homepage: www.arthroplastyjournal.org
Effect of Implant Design on Knee Flexion Douglas A. Dennis MD a, R. David Heekin MD b, Charles R. Clark MD c, Jeffrey A. Murphy MS d, Tammy L. O'Dell CCRC, CCRA d, Kimberly A. Dwyer PhD, CCRA d a
Colorado Joint Replacement, Denver, Colorado Heekin Orthopaedics, Jacksonville, Florida University of Iowa Hospitals & Clinics, Orthopaedic Surgery 01075-JPP, Iowa City, Iowa d DePuy Orthopaedics, Warsaw, Indiana b c
a r t i c l e
i n f o
Article history: Received 3 June 2011 Accepted 2 July 2012 Keywords: Total Knee Arthroplasty (TKA) simultaneous bilateral flexion high flexion rotating platform mobile bearing knee knee implant design
a b s t r a c t From March 2006 to August 2008, 93 subjects (186 knees) underwent simultaneous bilateral total knee arthroplasty performed by eight surgeons at North American centers. This randomized study was conducted to determine whether non-weight-bearing passive flexion was superior for knees receiving a posterior stabilized high flexion device compared to a posterior stabilized standard device in the contra-lateral knee. Weight-bearing single leg active flexion was one secondary endpoint. Follow-up compliance was 92.5%. Results show small, but significant superiority in the motion metrics for the high flexion device compared to the standard device 12 months after surgery, especially for a subgroup of patients with pre-operative flexion less than 120° in both knees. Thus, the ideal candidate for the high flexion device may be one with lesser preoperative flexion. © 2013 Elsevier Inc. All rights reserved.
Orthopaedic surgeons are replacing knees in younger, more active subjects who have elevated expectations of implant performance. Expectations include restoration of joint function, including greater knee flexion [1–3]. To address these heightened expectations, many orthopaedic companies have enhanced their total knee arthroplasty (TKA) portfolios by including devices specifically designed to provide high flexion (HF) [4]. Early published results comparing flexion of HF to standard (STD) devices are mixed. Two randomized, controlled trials [5,6] comparing STD to HF knees of fixed tibial bearing, posterior stabilized designs found no significant differences in knee flexion between the groups, although Nutton et al. [6] admitted their unilateral study was underpowered. A third study [4], which was a matched-pair study comparing rotating platform, posterior stabilized STD to HF devices (of the same design as those used in this study) demonstrated enhanced flexion in those subjects implanted with the HF design, particularly in those subjects whose pre-operative knee flexion was less than 120°. Knee flexion after TKA is affected by multiple variables including implant design, surgical technique, physical therapy regimen, postoperative pain management protocol, and subject factors such as preoperative flexion, body habitus, and subject motivation [1,2,7,8]. Therefore, we designed this simultaneous bilateral study in which The Conflict of Interest statement associated with this article can be found at http:// dx.doi.org/10.1016/j.arth.2012.07.019. Reprint requests: Douglas A. Dennis, MD, Colorado Joint Replacement, 2535 S. Downing Street, #100, Denver, CO 80210. 0883-5403/2803-0009$36.00/0 – see front matter © 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.arth.2012.07.019
each subject would receive both devices and therefore, eliminate most of these variables, that is surgical technique, physical therapy regimen, post-operative pain management protocol, body habitus, and subject motivation, from interfering with any flexion differences that may exist between the HF and STD knee designs. Both the HF and STD designs were from the same knee implant family (P.F.C. Sigma Rotating Platform and P.F.C. Sigma Rotating Platform High Flexion, DePuy Orthopaedics, Warsaw, Indiana). We defined ‘simultaneous’ bilateral as implanted during the same bilateral TKA procedure. Consequently, the HF feature of the implant could be evaluated without confounding variables present in non-bilateral studies [9,10]. The primary objective of this study was to determine whether the HF design provided superior non-weight-bearing passive flexion (PF) compared to the STD design 6 and 12 months after surgery. Twelve months was chosen as the maximum time for endpoint comparison because most studies show no change in motion beyond the first year [5,7,11–14]. Secondary study objectives included identification of preoperative factors, such as body mass index (BMI), gender, age, thigh girth, and skin-fold thickness, which may be associated with improved flexion with this HF design. Materials and Methods Each subject received a standard (STD) device in one knee and a high flexion (HF) device in the other, eliminating most confounding variables present in non-bilateral studies [9,10]. From March 2006 to August 2008, data were collected prospectively for 93 subjects (186 knees) who underwent simultaneous bilateral total knee
430
D.A. Dennis et al. / The Journal of Arthroplasty 28 (2013) 429–438
arthroplasty (TKA) performed during the same operative procedure from eight experienced joint replacement specialists across eight North American centers. Allocation of subjects for enrollment was evenly distributed amongst the eight centers; however, there were varying degrees of enrollment attained at each center (see Fig. 1). Criteria for inclusion included males and females age 40–70, inclusive, subjects willing to return for all post-operative follow-ups, suitable bilateral candidates for the devices specified in the protocol, and voluntary, written informed consent. Subjects were excluded if they had a diagnosis of inflammatory arthritis, flexion contracture greater than 20°, involved in litigation or worker's compensation claims, known drug, alcohol or psychological disorders that could affect follow-up care, participation in a clinical investigation with an investigational product in the last three months, previous revision, and pregnant or lactating females. The 93 subjects enrolled in the study were considered the Safety Dataset and underwent adverse event analysis. Twelve of these 93 subjects did not complete the primary endpoint, non-weightbearing passive flexion (PF), 12 months after surgery (see Fig. 1).
One subject withdrew consent 10 months after the bilateral surgeries; the withdrawal was unrelated to either implant. There were two protocol violations in which the STD components (femurs with lugs) were not listed in the study protocol, but the appropriate HF devices were implanted on the contra-lateral side. One of the two subjects with a protocol violation had their eligible HF knee complete the primary endpoint; therefore, that knee was used in unpaired analyses. There were two revision TKA procedures, one HF device and one STD device. The HF device was revised 6 months after the index surgery and the STD at 7 months; both revisions were secondary to deep infection, and were performed at different centers. The remaining seven subjects were lost to follow-up, leaving 81 bilateral subjects available for the Primary Efficacy Dataset and respective analyses (Fig. 1), and provides a 92.5% (86 of 93) subject follow-up compliance rate. There were two protocol deviations in which 71- and 73-year old subjects were inadvertently enrolled when the study maximum specified age was 70. This age criterion was set to help ensure strong subjects were enrolled because of the invasive nature of this bilateral
Fig. 1. Subject accountability flowchart demonstrating two main datasets for analysis: Safety Dataset and Primary Efficacy Dataset.
D.A. Dennis et al. / The Journal of Arthroplasty 28 (2013) 429–438
431
Fig. 2. Secondary endpoint Single Leg Active Flexion (SLAF).
study. The researchers felt these small age deviations were of no real consequence to subject recovery or return to function, as the subjects were otherwise healthy, and do not believe they compromised the scientific validity of the study. Therefore, the two subjects were included in the Primary Efficacy Dataset. Endpoints were measured pre-operatively, for baseline assessment, and post-operatively at 2-weeks, 6-weeks, 6-months, and 12months. Flexion measurements were obtained by a qualified individual who was blinded to which knee received each device and appropriately trained to minimize measurement error. Because early post-operative outcomes are not the focus of this paper (those results are found elsewhere [15]) only results of the latter 2 evaluation periods are provided, along with metrics that assess their change from pre-operative measurements. Paired differences and 95% confidence intervals are included to provide effect size. Secondary endpoints included weight-bearing single leg active flexion (SLAF) (Fig. 2), American Knee Society (AKS) Score [16], Knee injury and Osteoarthritis Outcome Score (KOOS) [17,18], the anterior knee pain question from the patella score [19], and degree of patellofemoral crepitus with passive motion [20]. Because it was anticipated that subjects would be unable or unwilling to participate in SLAF early post-operatively, it was not collected at 2 and 6-weeks. Anterior knee pain [19] was rated using a scale of ‘none’, ‘mild’, ‘moderate’ or ‘severe’. The degree of patellofemoral crepitus was assessed using ‘none’, ‘fine’ or ‘coarse’ categories [20]. Some subjects had missing information for some of the secondary outcomes, especially pre-operative SLAF, as noted by the ‘n’ or denominator statistics reported throughout this paper. The missing information was not considered a limitation of the study because, in most cases, the missing information was shared on both knees within a subject, and the study was very amply powered and could withstand a substantially reduced sample size. The study protocol was approved by each site's Institutional Review Board (IRB), and written informed subject consent was obtained from all subjects prior to enrollment. The study was registered at www.clintrials.gov as study number NCT00380861. Treatment assignment for each knee was based on pre-operative PF measured laterally with subject supine on an examination table, using a full-length 12-inch (30.5 cm) goniometer (provided by the study sponsor for consistency across sites). This measurement was performed by a single qualified individual, at each site, blinded to the device received on each knee. Goniometric measurements of the knee have been shown to be reliable and valid compared to radiographs [20,21]. Pre-operatively, subjects identified which knee they considered their ‘worse’ knee, and the surgeon began the bilateral procedure on that knee. Both the HF and STD used the same patellar components and tibial tray design. The posterior stabilized femoral components and polyethylene inserts differed in their posterior condylar geometry, to accommodate different magnitudes of flexion (Fig. 3). Specifically, the HF tibial
insert post was placed 2 mm posterior compared to the STD insert with the goal of increasing posterior femoral translation, which has been associated with enhanced flexion. All components were cemented. Although not specified in the study protocol, a brief radiographic review was conducted by a blinded, independent radiographic reviewer, to ascertain the bone–cement–implant interface using both anterior posterior (AP) and lateral radiographic films for subjects at their 12-month follow-up interval. Previous intervals (e.g. 6months, 6-weeks) films were assessed, if necessary for confirmation of 12-month findings. Any signs of osteolysis, or radiolucencies and progression of osteolysis or radiolucencies were noted. The average subject age was 61.4 years (range, 44–73 years, standard deviation 5.8). Ninety-nine (92 of 93) percent of subjects had a diagnosis of osteoarthritis, 66% (61 of 93) were female, and the average Body Mass Index (BMI) was 31.6 (range 19–47, standard deviation 5.3). Setting a standard for future flexion studies, race and ethnicity were collected per a US Food and Drug Administration guidance document [22]. Ninety eight percent (91 of 93) were of the Caucasian race, and 99% (92 of 93) were of Non-Hispanic ethnicity. Because body mass index (BMI) [8] does not indicate the location of mass and body habitus is known to have an influence on knee flexion, several knee specific anthropometrics, such as thigh circumference and thigh skin-fold thickness, were collected to assess their
Polyethylene insert
Femoral Component
‘J’ curve for HF ‘J’ curve for STD Tibial Component
Fig. 3. The high flexion (HF — up to 155o of flexion) and standard (STD — up to 120o of flexion) implants came from the same product family. The J-Curve of the HF device has a tighter posterior radius for the HF design compared to the STD design.
432
D.A. Dennis et al. / The Journal of Arthroplasty 28 (2013) 429–438
Table 1 Pre-operative Knee and Subject Specific Anthropometrics and Other Covariates Examined to Assess their Potential Influence and Interaction on Post-Operative Outcomes. Knee Specific Thigh Circumference* (cm) Ankle Circumference (cm) Skin-fold† (mm) Quadriceps Strength‡ Hamstring Strength§ Pre-Op Flexion (PF & SLAF) Pre-Op AKS See Tables 3 & 4 Pre-Op KOOS sub-scores See Tables 3 & 4
Mean
SD (Range)
Subject Specific
Mean
SD (Range)
50.2 24.8 32.3 27.2% full strength 14.7% full strength See Tables 3 & 4
6.3 (37 to 71) 2.7 (16 to 33) 14.6 (4 to 60)
Age (years) Height (inches) Weight (lb) Body Mass Index (BMI) Pre-Op Motivation∥ Raceb Ethnicityb Gender
61.4 67 201.5 31.6 11.4 98% Caucasian 99% Non-Hispanic 66% Female
5.8 (44 to 73) 4.2 (60 to 76) 39 (115 to 305) 5.3 (19 to 47) 3.8 (0 to 17)
SD = Standard Deviation; PF = Passive Flexion; SLAF = Single Leg Active Flexion; AKS = American Knee Society; KOOS = Knee Injury and Osteoarthritis Outcomes Score. * Thigh circumference was measured 10 cm proximal to the superior aspect of the patella bone. † Skin-fold was measured in the middle of the thigh, 10 cm proximal to the superior patellar pole, using calipers. ‡ From a sitting position with knee bent 90o and a resistance band attached to their ankle, subjects attempted to achieve full extension, one knee at a time; 27.2% of the subjects achieved full extension. § From a sitting position with knee fully extended and a resistance band attached to their ankle, subjects attempted to achieve flexion of 90o, one knee at a time; 14.7% of the subjects achieved 90o flexion. ∥ Subjects were asked whether the following 17 flexion dependent activities were important to them pre-operatively: descending stairs, ascending stairs, rising from sitting, bending to floor/pick up an object, getting in/out of car, putting on socks/stockings, rising from bed, taking off socks/stockings, getting in/out of bath, sitting, getting on/off of toilet, heavy domestic duties, squatting, running, jumping, twisting/pivoting, and kneeling. The activities are a subset of the KOOS ADL and Sport and Recreation subscales. If the activity was 'Very' important, then the question was counted. On average, subjects marked 11.4 of these 17 activities as very important to them pre-operatively. b Race and ethnicity were planned to be included as covariates in the statistical modeling, however, because 98% of the subjects enrolled in this study were Caucasian and 99% were Non-Hispanic, there was not enough disparity to model properly.
impact on flexion. These anthropometrics along with a number of other covariates (Table 1) thought to affect flexion were collected preoperatively and examined to assess their potential influence on postoperative outcomes. All operative procedures were performed under regional or general anesthesia, using a standard anterior midline incision and a medial parapatellar arthrotomy. Patellar resection was consistent bilaterally. The technique of ligament balancing was individualized based on surgeon preference but was consistent for each site and identically applied to both knees for each subject. The surgical technique specified in the protocol ensured that close attention was paid to factors affecting the positioning of the implants during surgery. The protocol required that the difference in the thickness of the patellar reconstruction between the bilateral knees was no greater than 4 mm; and the implanted femoral components between the bilateral knees did not differ by more than one implant size. Intra-operatively, the surgeon deliberately resected the tibial posterior slope such that the planned difference between the right and left knees was not greater than 2°. Post-operative care was prescribed as per the surgeon's usual routine; however, the physical therapy and pain management were standardized across all sites. Even if differences existed across sites, the bilateral study design eliminated their impact in comparing the HF and STD knees within each subject. The type and frequency of adverse events were compared for all subjects (i.e., the Safety Dataset), and categorized as intra-operative, post-operative operative site and post-operative systemic. Based on a power analysis, it was determined that 128 subjects were required to significantly differentiate a PF difference of 5°. However, this power estimate was based on a 2-sided 2-sample t-test and was not based on the more appropriate paired (bilateral) data because none was available for rotating platform knees when this study was initially designed. After eighty-four (84) subjects had been accrued, the power analysis was re-calculated as a paired study endpoint using data from the study itself. The study data indicated the power to be greater than 99% to detect a 5° paired PF difference between the HF and STD groups (only 13 subjects were required for 80% power). Thus, it was decided to terminate enrollment at that time. By the time this information had been disseminated to the study sites and been approved by the IRB's, ninety-three (93) subjects had been accrued for the study. A subgroup of subjects with low pre-operative non-weightbearing PF in both knees was identified by placing subjects into
decreasingly lower pre-operative groupings, one degree at a time, until a threshold was found where post-operative PF differences were large between HF and STD, and enough subjects remained for statistical analysis. This algorithm yielded the threshold of 120° PF in both knees, and was supported by the same threshold used by Gupta et al. [4] in his unilateral study of the same devices. Twenty subjects met these criteria, 18 of whom completed the primary endpoint, yielding the Primary Efficacy Subgroup (see Fig. 1). Because of the increased power associated with pairing the data, this subgroup of 18 subjects had more than 85% power to detect a 5° difference in PF at the 12 month follow-up evaluation.
Statistical Methods Paired t-test analyses were conducted on continuous variables, whereas Fisher's Exact tests were used for categorical variables. When the comparison of groups or categories yielded p-values below 5%, it was considered statistically significant, whereas p-values from 5% and 10% were considered ‘borderline’ and identified as such. Regression and longitudinal analyses were carried out on postoperative data using PF and SLAF as dependent variables, with the devices and covariates (Table 1) as the independent variables. Longitudinal analyses utilized all post-operative evaluations. Plot smoothing was used for longitudinal graphs to provide continuity between follow-up intervals. The covariates were included to identify any pre-operative factors that would identify subjects who could benefit most from the HF devices. The MIXED procedure of SAS software (Cary, North Carolina, USA) was utilized for this modeling because it can fit covariate adjusted models with paired and unpaired data and can accommodate missing information. All potentially significant covariates were combined into a full covariate model, where non-parallel covariate effects were nested to accommodate different covariate regression slopes for the two devices. Covariate effects were reduced from this full model stepwise using an individual alpha above 0.05; non-parallel covariate effects were reduced from the model first when possible, followed by main effect covariates. Restricted maximum likelihood estimation was used in the final reduced model. SAS convergence issues were sometimes encountered in model reduction, and in these instances, minimum variance quadratic unbiased estimation was used. An unstructured covariance structure was used to model both the pairing of device measurements by subject and the repeated measurements over time. Either quadratic
D.A. Dennis et al. / The Journal of Arthroplasty 28 (2013) 429–438
433
Table 2 Primary Endpoint non-weight-bearing Passive Flexion for the Standard (STD) and High Flexion (HF) Knee Devices. Standard Metric Passive Flexion (°)
High Flexion
Interval
n
Mean
SD (Range)
n
Mean
SD (Range)
Pre-op 6 Months† 12 Months Δ Pre-op to 6 Months Δ Pre-op to 12 Months
91* 74 81 74 81
122.8 124.0 126.8 1.01 3.83
10.2 (90 to 146) 10.6 (90 to 147) 9.8 (89 to 142) 11.5 (−33 to 25) 11.6 (−38 to 42)
93 75 82‡ 75 82‡
122.5 124.2 127.8 2.16 5.37
11.0 (85 to 144) 11.2 (70 to 142) 9.2 (90 to 145) 14.1 (−58 to 35) 11.7 (−38 to 38)
Paired Difference (HF-STD) −0.20 0.23 1.11 1.11 1.61
95% CI of Difference −2.03 −1.18 −0.15 −1.39 −0.37
to to to to to
1.64 1.64 2.38 3.60 3.60
Significant Difference? No No Borderline No No
Paired t-test 2-sided P-value 0.831 0.747 0.084 0.759 0.109
SD = standard deviation; CI = confidence interval; Δ = change from. * Two STD devices were not among the approved components listed within the study protocol. † The 6-month follow-up interval had a smaller interval window compared to the 12 month interval and this proved challenging for sites to get all of their subjects back for evaluation; hence, the smaller number of subjects with follow-up in this interval. ‡ One subject had 12 Month follow-up for his HF knee, but his STD knee was ineligible. The HF knee is included in unpaired analyses.
or linear longitudinal models were utilized based on the evident shape of responses over time. Source of Funding Funding for this study was provided by DePuy Orthopaedics (Warsaw, Indiana, USA).
represented the predictor variable. Sites with less than 10 subjects were pooled together. Model p-values were 0.103 and 0.544 for the 6and 12-month follow-up intervals, respectively, validating no differences between any of the sites. Balanced results across each side of the patient were investigated using a t-test. Resulting p-values for both 6- and 12-month follow-up intervals were 0.941 and 0.371, respectively, indicating side had no impact on the observed results.
Results Primary Endpoint: Non-weight-bearing Passive Flexion A total of 81 subjects, deemed the Primary Efficacy Dataset (Fig. 1), were evaluated for the primary endpoint of non-weight-bearing passive flexion (PF) 12 months after surgery. An additional five subjects met protocol defined study withdrawal criteria before reaching their 12 month follow-up evaluation, providing a subject follow-up compliance rate of 92.5% (86 of 93). Confirmation of no site to site differences was verified using a oneway analysis of variance (ANOVA) test where paired-differences of the primary endpoint PF was used as the response variable and sites
There were no statistically significant differences between the high flexion (HF) and standard (STD) devices for PF at any of the follow-up intervals, including pre-op, or change metrics according to paired t-test analyses (see Table 2). However, at the 12 month post-operative evaluation interval, there was a paired 1.11° borderline (P = .084) advantage for the knees receiving the HF device. Regression analysis identified pre-operative PF as a significant covariate (see Table 1) at both post-operative evaluations, while pre-op American Knee Society
Table 3 Secondary continuous Endpoints for the Standard (STD) and High Flexion (HF) Knee Devices. SLAF, KOOS, and AKS Outcome Metrics for STD and HF Knee Devices Standard Metric Single Leg Active Flexion (°) KOOS Pain (points) KOOS Symptoms (points) KOOS ADL (points)
KOOS Sport & Rec (points) KOOS Quality of Life (Points) AKS Knee Score (Points)
Interval
N
Mean
Pre-op 6 Months† 12 Months Pre-op 6 Months† 12 Months Pre-op 6 Months† 12 Months Pre-op 6 Months† 12 Months Pre-op 6 Months† 12 Months Pre-op 6 Months† 12 Months Pre-op 6 Months† 12 Months
27* 70 73 86 67 79 86 70 79 79 69 75 82 48 54 90 72 78 91 74 80
124.6 126.0 126.6 45.7 84.0 89.5 50.7 76.8 82.5 49.0 85.4 90.4 19.9 56.5 68.1 19.2 65.0 74.9 48.1 86.1 85.6
SD (Range) 11.1 11.0 10.6 19.2 14.8 14.0 20.2 13.7 11.9 20.2 14.0 11.8 21.8 28.9 24.3 17.3 21.4 22.1 19.0 11.5 11.5
(108 to 148) (90 to 147) (88 to 143) (0 to 94) (42 to 100) (28 to 100) (4 to 93) (46 to 100) (46 to 100) (1 to 100) (38 to 100) (31 to 100) (0 to 100) (0 to 100) (5 to 100) (0 to 94) (13 to 100) (6 to 100) (2 to 100) (48 to 100) (54 to 100)
High Flexion n
Mean
31* 72 74 88 68 80 89 68 82 83 69 75 82 48 55 92 72 78 93 74 81
120.9 125.2 127.5 45.3 83.1 90.0 49.9 75.5 83.1 47.9 85.6 90.1 19.0 55.4 69.3 19.4 64.4 75.6 47.4 84.5 87.0
SD (Range) 13.3 10.8 11.1 18.2 14.9 11.8 18.9 14.5 13.1 19.2 12.7 11.3 19.2 26.7 24.4 15.9 21.2 22.2 17.4 12.5 10.7
(82 to 150) (90 to 142) (82 to 145) (6 to 89) (39 to 100) (53 to 100) (4 to 89) (29 to 93) (43 to 100) (4 to 90) (43 to 100) (43 to 100) (0 to 100) (0 to 100) (5 to 100) (0 to 69) (13 to 100) (19 to 100) (13 to 100) (40 to 100) (56 to 100)
Paired Difference (HF-STD) −2.26 −0.54 1.23 −0.48 −0.95 0.71 −1.08 −0.89 0.54 −1.43 0.15 −0.29 −1.42 −0.53 0.93 0.42 −0.61 0.65 −1.41 −1.55 1.21
95% CI of Difference
Significant Difference?
Paired 2-sided t-test P-value
−6.24 to 1.72 −1.77 to 0.69 0.14 to 2.32 −3.51 to 2.54 −3.09 to 1.19 −1.41 to 2.84 −4.09 to 1.93 −3.47 to 1.68 −0.86 to 1.95 −4.28 to 1.41 −1.59 to 1.89 −1.60 to 1.01 −3.93 to 1.09 −2.45 to 1.39 −1.06 to 2.91 −1.82 to 2.65 −2.63 to 1.41 −1.24 to 2.54 −4.19 to 1.38 −4.46 to 1.35 −1.05 to 3.47
No No Yes No No No No No No No No No No No No No No No No No No
0.254 0.382 0.027 0.751 0.377 0.506 0.478 0.491 0.444 0.319 0.865 0.655 0.264 0.579 0.354 0.712 0.551 0.496 0.318 0.290 0.288
SD = standard deviation; CI = confidence interval; ADL = Activities of daily living; AKS = American Knee Society; SLAF = Single Leg Active Flexion; KOOS = Knee Injury and Osteoarthritis Outcome Score. * Most subjects were unable or unwilling to participate in SLAF pre-operatively because of pain. Because of the study's high power (N99%), the comparison with the reduced number still had adequate power (N80%). † The 6-month follow-up interval had a smaller interval window compared to the 12-month interval and this proved challenging for sites to get all of their subjects back for evaluation; hence, the smaller number of subjects with follow-up in this interval.
434
D.A. Dennis et al. / The Journal of Arthroplasty 28 (2013) 429–438
Table 4 Primary Endpoint non-weight-bearing Passive Flexion for the Standard (STD) and High Flexion (HF) Knee Devices for Subjects with Pre-op Passive Flexion b 120° in Both Knees. Standard (STD)
Passive Flexion (°)
High Flexion (HF)
Interval
n
Mean
SD (range)
n
Mean
SD (range)
Paired Difference (HF-STD)
Pre-op 6 Months* 12 Months Δ Pre-op to 6 Months Δ Pre-op to 12 Months
19 16 18 16 18
111.6 118.9 120.2 7.06 8.22
6.0 (95 to 118) 13.9 (90 to 140) 11.3 (89 to 140) 13.1 (−18 to 25) 11.3 (17 to 31)
20 17 18 17 18
111.4 120.9 124.7 9.18 13.06
5.8 (102 to 119) 12.7 (92 to 142) 9.8 (108 to 145) 12.1 (−16 to 26) 9.6 (−6 to 28)
−0.42 2.06 4.56 2.38 4.83
Metric
95% CI of Difference −2.82 −0.30 1.29 −1.46 1.64
to to to to to
1.97 4.43 7.82 6.21 8.03
Statistically Significant?
Paired 2-sided t-test P-value
No Borderline Yes No Yes
0.716 0.083 0.009 0.207 0.005
SD = standard deviation, CI = confidence interval, Δ = change from. * The 6-month follow-up interval had a smaller interval window compared to the 12 month interval and this proved challenging for sites to get all of their subjects back for evaluation; hence, the smaller number of subjects with follow-up in this interval.
(AKS) Score was a significant covariate at 6 months and skin-fold thickness was a significant covariate at 12 months. The impact of the significant covariates was minimal with the covariate adjusted pvalue for PF at 6 months changing from 0.747 (see Table 2) to 0.724, and the covariate adjusted p-value for PF at 12 months changing from 0.084 to 0.079. Secondary Endpoints Results of continuous secondary endpoints are provided in Table 3. There were no significant differences between the two knees preoperatively for any of the endpoints. Weight-Bearing Single Leg Active Flexion (SLAF) There was no significant difference between the HF and STD devices 6 months post-operatively, however, SLAF was significantly greater (P = .027) in the HF knees at 12 months with a paired difference of 1.23°. Knee Injury and Osteoarthritis Outcome Score (KOOS) and American Knee Society (AKS) There were no significant differences between knees for the AKS knee score or any of the five KOOS subscales pre-operatively, or at 6 or 12 months post-operatively. The AKS functional scores are not included in Table 3 because that outcome involves both knees. Preoperatively, average AKS function scores were 54.9 points (SD = 15.9, range 5 to 100). Post-operatively average AKS function scores were 83.9 points (SD = 18.4, range 15 to 100) at 6 months and 91.6 points (SD = 11.2, range 45 to 100) at 12 months. Anterior Knee Pain There were no statistical differences pre-operatively, or at 6 or 12 months post-operatively for anterior knee pain when comparing the proportions between the two knee devices. As such, pooled statistics between HF and STD are provided here: Pre-operatively 16% (30 of 184) of subjects had severe anterior knee pain, 41% (76 of 184) had moderate, 15% (27 of 184) had mild, and 28% (51 of 184) had no anterior knee pain. Six months after the surgical procedures, 79% (119 of 150) of subjects had no anterior knee pain, 18% (27 of 150) had mild, 2% (3 of 150) had moderate, and b1% (1 of 150) had severe anterior knee pain. At the 12 month post-operative interval, 79% (113 of 143) of subjects had no anterior knee pain, 17% (24 of 143) had mild, 4% (6 of 143) had moderate, and no subject had severe anterior knee pain. Patellar Crepitus There were no statistical differences pre-operatively, or at 6 or 12 months post-operatively in the degree of incidence for crepitus when comparing the two knee devices. Thus, combined HF and STD statistics are as follows: Pre-operatively 48% (87 of 182) of subjects had coarse crepitus, 37% (68 of 182) had fine crepitus, and 15% (27 of 182) had no crepitus. Six months post-operatively, 56% (84 of 150) of subjects had no crepitus, 34% (52 of 150) had fine crepitus, and 9% (14
of 150) had coarse crepitus. At the 12 month interval, 70% (97 of 139) of subjects had no crepitus, 25% (35 of 139) had fine crepitus, and 5% (7 of 139) had coarse crepitus. Longitudinal Analysis Longitudinally (across post-operative time) there was no significant difference between knees for both flexion outcomes: PF and SLAF. Subgroup Analyses Table 4 provides the non-weight-bearing PF results for a subgroup of 18 subjects whose pre-operative PF was less than 120° in both knees (see Fig. 1). The knees with the HF device (124.7°) were statistically superior (P = .009) compared to the STD knee devices (120.2°) 12 months after surgery yielding a paired difference of 4.56° for the Primary Efficacy subgroup. Similar results were observed for the change from pre-operative to 12 months post-op: HF was superior to STD by 4.83° with P = .005. There were no significant covariates. Significant advantages for the HF device were observed for weight-bearing SLAF at both 6 months (means: HF = 121.1°, STD = 118.3°; P = .038) and 12 months (means: HF = 121.3°, STD = 118.9°; P = .040) after surgery. There were no significant differences observed between the STD and HF devices for KOOS and AKS scores. Longitudinally (Fig. 4), the HF knees had borderline superior PF (P = .067) compared to the STD knees. The HF knees were significantly superior compared to the STD knees for SLAF (P = .029). Significant covariates from the longitudinal models included pre-operative PF for the post-operative PF response, and thigh circumference for the post-operative SLAF response. The impact of the latter significant covariate was minimal with the longitudinal covariate adjusted p-value for SLAF changing from 0.029 to 0.069. The impact of significant covariate pre-operative PF on the longitudinal postoperative PF was more dramatic with the p-value changing from 0.067 to 0.477. On average, each 1° pre-operative PF increase yielded a 0.5° increase in post-op PF. Radiographic Review No osteolysis was observed for any of the HF or STD knees. No radiolucencies ≥ 2 mm were observed on the AP view. Two subjects, one involving their HF device and the other involving their STD device, had radiolucencies ≥ 2 mm on their lateral film. These were interpreted as interface gaps that did not progress from immediate to 12-month post-op films; the cause was attributed to sub-optimal cement technique. Adverse Events There was 1 intra-operative adverse event: 1 STD knee required a medial collateral ligament reconstruction with a soft tissue washer and screw, which was resolved at the end of surgery. There were 52 systemic complications (Table 5a). There were 36 post-operative operative site adverse events (Table 5b). The type, frequency,
D.A. Dennis et al. / The Journal of Arthroplasty 28 (2013) 429–438
435
Passive Flexion (PF)
P-value = .067 Significant Covariate: Pre-operative PF
Smoothed line
The impact of significant covariate pre-operative PF on longitudinal post-operative PF was somewhat dramatic with the covariate adjusted P-value changing from .067 to .477. The regression model tells us for each 1 degree pre-operative PF increase an increase of 0.5 degrees is expected in post-operative PF.
Single Leg Active Flexion (SLAF)
P-value = .029 Significant Covariate: Thigh circumference
Smoothed line
SLAF was only measured 6 and 12 months post-operatively because earlier time intervals interfered with subject’s rehabilitation. The impact of significant covariate thigh circumference was minimal with the longitudinal covariate adjusted P-value changing from .029 to .069.
Fig. 4. Longitudinal plots of post-operative flexion metrics in subjects with b 120° pre-operative passive flexion in both knees.
proportion, and 95% exact confidence interval of post-operative systemic and operative site adverse events are noted in Table 5a and 5b respectively. There were no significant differences in operative site adverse events between the two implants as noted by the Fishers Exact Test p-values greater than 0.05 (Table 5b). Revisions and Reoperations Four knees (3 patients) were revised, 2 that received the HF device and 2 that received the STD device. Two of the revisions, 1 HF and 1 STD, were secondary to deep infections that occurred in different subjects 6 months (HF) and 7 months (STD) after the index surgery,
respectively. These two revisions prevented the subjects from reaching the primary endpoint of 12 month post-operative follow-up and are, therefore, accounted for in Fig. 1. The other 2 revisions were secondary to tibial tray loosening and occurred in the same subject in both knees, after the 12 month primary endpoint had been collected. The HF knee was revised 19 months after the index surgery whereas the STD knee was revised 21.5 months after the index surgery. One additional subject had a reoperation where the tibial polyethylene insert was exchanged 7 weeks after the index surgery due to polyethylene ‘spin out’. The ‘spin out’ was discovered radiographically in the HF knee at a routine follow-up visit, as the
436
D.A. Dennis et al. / The Journal of Arthroplasty 28 (2013) 429–438
Table 5a Post-Operative Systemic Adverse Events (includes both knees). Post-operative Systemic Adverse Event
Frequency
Other* Musculoskeletal Respiratory Gastrointestinal Genitourinary Vascular — Pulmonary Embolism Vascular — Deep Vein Thrombosis Cardiac Endocrine/Metabolic Central Nervous System Dermatological Head, Ears, Eyes, Nose and Throat Hematological Peripheral Nervous System Total
Percent of Total Frequency (N = 93)
11 9 8 6 4 3 2 2 2 1 1 1 1 1 52
95% Exact Confidence Interval
11.8% 9.7% 8.6% 6.5% 4.3% 3.2% 2.1% 2.1% 2.1% 1.1% 1.1% 1.1% 1.1% 1.1% 55.9%
6.1–20.2 4.5–17.6 3.8–16.3 2.4–13.5 1.2–10.7 0.7–9.1 0.3–7.6 0.3–7.6 0.3–7.6 0.0–5.9 0.0–5.9 0.0–5.9 0.0–5.9 0.0–5.9 45.2–66.2
* ‘Other’ systemic adverse events included: back spasms, right ankle pain, right wrist carpal tunnel, 2 electrolyte imbalances, sciatica, chest pain, infected PICC line, cellulitis right arm, left hand/arm weakness, right arm sensation at IV site after blood transfusion. Adverse Events are counted based on a per subject basis; that is, duplicate reports for resolving issues within a single subject are not counted twice.
patient was asymptomatic. When questioned, the patient recalled a challenging time in physical therapy several weeks beforehand. The patient received a thicker tibial polyethylene insert during the reoperation and both knees completed the 12-month followup interval. The adverse events associated with these revisions and reoperations are included in Table 5b. Discussion The primary objective of this study was to determine whether the high flexion (HF) design provided superior non-weight-bearing passive flexion (PF) compared to the standard (STD) design 12 months after surgery. This study showed small, but borderline statistically significant improvements in non-weight-bearing PF and weight-bearing single leg active flexion (SLAF) for the knees receiving the HF device compared to the STD device 12 months postoperatively. The small magnitude of improvement in our study may not be clinically significant. The bilateral study design allowed subtle motion improvements in the HF compared to STD to become evident that typical unilateral study designs are unable to detect. Superiority in flexion was more pronounced in a subgroup of subjects with less than 120 o of pre-operative flexion in both knees, suggesting the ideal candidate for an HF TKA may be one with lesser pre-
operative flexion. Biomechanical studies do show that if high flexion is obtained by a subject implanted with an HF design, reduced polyethylene stresses [23,24] and lower patellar ligament strains [25] are incurred when compared to STD designs, which may enhance long term performance. The recently published paper by Choi, Lee et al., which compared the same two knee devices as our study, in an Asian population of unilateral subjects, reported no differences between the devices. The authors noted their high pre-operative motion (126° of active maximal flexion) may have obscured the motion advantage of the high-flexion design [26]. This agrees with our study results where the advantage of the HF device over the STD device was more pronounced in subjects with less than 120° of pre-operative flexion in both knees. Pre-operative PF was a significant covariate for several outcomes (Tables 2 and 4) indicating its large influence on the post-operative measurements, which is consistent with many other reports [6,9,25]. Pre-operative American Knee Society (AKS) scores, skin-fold thickness, and thigh circumference were other covariates with some influence on post-operative outcomes, but their impact was minimal, which may be a result of our homogeneous study group (that is, 98% Caucasian, 99% Non-Hispanic, average body mass index of 31.6). The rest of the covariates tested (Table 1) were not significantly influencing post-operative outcomes.
Table 5b Post-Operative Operative Site Adverse Events for the Standard (STD) and High Flexion (HF) Knee Devices. Standard (STD) Post-operative Operative Site Adverse Event Other Pain Infection Effusion Patellar Crepitus/Clunk Bone Fracture Loosening Polyethylene ‘spin out’ Arthrofibrosis Total
High Flexion (HF)
STD Frequency
STD Percent of Total Frequency (N = 93)
STD 95% Exact Confidence Interval
6* 3 1 0 2 1 1 0 0 14
6.6% 3.2% 1.1% 0.0% 2.2% 1.1% 1.1% 0.0% 0.0% 15.1%
2.4–13.5 0.7–9.1 0.0–5.8 0.0–0.0 0.3–7.6 0.0–5.8 0.0–5.8 0.0–0.0 0.0–0.0 8.5–24.0
HF Frequency 10† 3 3 2 0 1 1 1 1 22
HF Percent of Total Frequency (N = 93) 10.8% 3.2% 3.2% 2.2% 0.0% 1.1% 1.1% 1.1% 1.1% 23.7%
HF 95% Exact Confidence Interval
Fishers Exact Test P-value
5.3–18.9 0.7–9.1 0.7–9.1 0.3–7.6 0.0–0.0 0.0–5.8 0.0–5.8 0.0–5.8 0.0–5.8 15.5–33.6
0.437 1.000 0.621 0.497 0.497 1.000 1.000 1.000 1.000 0.275
Every unique adverse event was reported once, regardless of whether a single knee reported more than one instance of a particular adverse event. For example, if a knee reported ‘infection’ twice, then infection was counted once for that knee. However, if that same knee had ‘effusion’, then that adverse event was counted in addition to the infection adverse event. * STD ‘Other’ operative site adverse events included: stiffness during physical therapy, swelling, contusion, twisted knee w/pain, retained drain fragment, pulled hamstring. † HF ‘Other’ operative site adverse events included: sciatica (right knee), tingling/burning incision area, quadriceps strain in PT, deep orthopaedic drain site bleed, stiffness, leg swelling, abrasion, saphenous neuritis, right MCL instability, fall.
D.A. Dennis et al. / The Journal of Arthroplasty 28 (2013) 429–438
Differences in the AKS score and Knee Injury and Osteoarthritis Outcomes Score (KOOS) scores, as a function of prosthesis design, were not observed at 6 or 12 month post-operative follow-up intervals (Table 3). The AKS Score has been criticized for having a pronounced ceiling effect and inadequate sensitivity, whereas Knee Injury and Osteoarthritis Outcome Score (KOOS) has been touted as being more sensitive [27] in detecting differences in knee devices. Nevertheless, no differences were observed in the current study, perhaps because of the bilateral design where subjects may have trouble articulating differences between their right and left legs. The success of high-flexion knee implant systems has been inconsistent in the literature. In a mixture of retrospective case– control and prospectively randomized control studies, some authors have demonstrated a beneficial effect from 6 o to 13 o [4,9,28] while other authors have demonstrated no difference between HF and STD knee systems [5,6,11,14,26,26,29–31]. With the exception of two studies [5,30], who also studied bilateral subjects, these series have more confounding variables inherent with unilateral studies. Gupta et al.[4] in a matched pair study of subjects implanted with the same devices as the current study, similarly observed the greatest knee flexion gains occurred in HF subjects with pre-operative flexion less than 120°. There are concerns that the increased flexion that may result from HF TKA design modifications may be detrimental to the long term survival of these implant designs[14]. Increased flexion has been shown to increase peak polyethylene stresses in traditional TKA designs, theoretically increasing the risk of premature polyethylene wear. There were no instances of polyethylene wear in the current study, although follow-up duration is short, being 12 months. High flexion magnitudes are also associated with increased tension in the extensor mechanism which could result in increased anterior knee pain or patellar component fixation failure. We saw no differences in anterior knee pain in the current study. Reduced conformity of the posterior aspect of the tibial insert might risk an increased rate of flexion instability. Many of the current HF TKA designs, including the one in this study, require increased posterior femoral condylar bone resection which is a disadvantage should revision arthroplasty be necessary. Lastly, the increased loads of high knee flexion place increased stresses on the fixation interface, theoretically increasing the risk of component loosening. Han et al. [3] reviewed a cohort of 47 subjects implanted with 72 posterior stabilized-flex, fixed-bearing (HF) implants and observed an alarmingly high incidence of aseptic femoral component loosening of 38% (27 of 72) at a follow-up duration of 32-months. The postoperative mean maximum flexion was 136° in the group with femoral component loosening versus 125° in the well-fixed group (P = .022). Cho et al. [32] reviewed a group of 166 patients (218 TKA) implanted with the same HF TKA implants as the Han et al. [3] study with a mean follow-up duration of 51 months. Cho et al. [32] reported 7 of 218 cases (3.2%) required revision for femoral component loosening at a mean follow-up duration of 49months. In the current study, the incidence of loosening (Table 5b) was low, and equal between the HF and STD devices. Because of these concerns, continued long-term follow-up evaluation of these devices is necessary. Strengths of this study are the enrollment of prospective bilateral subjects in whom treatment and control are contained within a single subject, the standardized surgical technique, physical therapy and pain management protocols, and the detailed statistical analysis methods. Limitations include the short-term follow-up duration, the lack of diversity in ethnicity and race, and the limited radiographic review. The study was designed as a 12 month follow-up study because the maximum achievable non-weight-bearing passive flexion, the primary endpoint, was expected to be achieved by almost all subjects before the 12 month follow-up interval, which is supported as a reasonable time endpoint in the published literature [5,7,11–14]. Indeed, one report [7] stated ‘Most studies examining ROM have
437
shown that no significant improvement occurs in ROM beyond the first year.’ The lack of ethnic and racial diversity was not unexpected given the study took place at 8 U.S. centers and was not stratified for race or ethnicity. If this study had included subjects with high flexion activities of daily living, such as Asians, greater differences between the HF and STD may have been realized. Goniometer measurements have been shown to be accurate and reproducible and the strict criteria imposed on the surgical technique would likely not allow component mal-position to be a radiographic concern[21]. A brief radiographic review was conducted where no osteolysis was observed and only two cases of radiolucencies ≥ 2 mm were observed, one of each knee device, which were attributed to suboptimal cement technique. Future studies could consider radiographic evaluation of alignment, specifically femoral flexion and posterior tibial slope (which could affect knee flexion), to further validate knee flexion studies. The findings of this study are limited to the implant devices used and the bilateral subjects utilized, and may not extrapolate to other HF TKA designs, or the more commonly studied unilateral subject population. Lastly, the findings are also limited to the specific, Western, predominantly Caucasian, subject population. Acknowledgments The authors would like to thank the following for their integral contributions: Richard Driessnack, MD (Vanderbilt University School of Medicine, Nashville, TN), Dean Sukin, MD (Montana Orthopaedics and Sports Medicine, Billings, MT), Steven Teeny, MD (Puget Sound Orthopaedics, Lakewood, WA), William Bugbee, MD (Scripps Clinic, LaJolla, CA), Lawrence Housman, MD (Tucson Orthopaedic Institute , Tucson, AZ), James Lesko, PhD (DePuy Orthopaedics, Warsaw, IN) for statistical analysis assistance, Steven Tippett, PhD (Bradley University, Peoria, IL) for the physical therapy protocol, input on data collection and training of the study coordinators on collection of physical measurements, Thomas Gruen, MS (Zonal Concepts, Wesley Chapel, FL). And, the study coordinators for their dedication to the implementation of the protocol: Jane Mang, RN, BSN, MAT (Colorado Joint Replacement, Denver, CO), Heather Whetsell, RN, (Great Plains Sports Medicine & Rehabilitation Center Peoria, IL), Terry Anderson, (Great Plains Sports Medicine & Rehabilitation Center, Peoria, IL), Jennifer Crews, RN, BSN (Heekin Orthopaedics, Jacksonville, FL), Linda Fink, RN, MSN (University of Iowa Hospitals & Clinics, Iowa City, IA), Judy Miller, RN, CCRP (Montana Health Research Institute, Billings, MT), Sally York (Lakewood Orthopaedic Surgeons, Lakewood, WA). References 1. Ranawat CS. Design may be counterproductive for optimizing flexion after TKR. Clin Orthop Relat Res 2003(416):174. 2. Sultan PG, Most E, Schule S, et al. Optimizing flexion after total knee arthroplasty: advances in prosthetic design. Clin Orthop Relat Res 2003(416):167. 3. Han HS, Kang SB, Yoon KS. High incidence of loosening of the femoral component in legacy posterior stabilised-flex total knee replacement. J Bone Joint Surg Br 2007; 89(11):1457. 4. Gupta SK, Ranawat AS, Shah V, et al. The P.F.C. sigma RP-F TKA designed for improved performance: a matched-pair study. Orthopedics 2006;29(9 Suppl):S49. 5. Kim YH, Sohn KS, Kim JS, et al. Range of motion of standard and high-flexion posterior stabilized total knee prostheses. A prospective, randomized study. J Bone Joint Surg Am 2005;87(7):1470. 6. Nutton RW, van der Linden ML, Rowe PJ, et al. A prospective randomised doubleblind study of functional outcome and range of flexion following total knee replacement with the NexGen standard and high flexion components. J Bone Joint Surg Br 2008;90(1):37. 7. Anouchi YS, McShane M, Kelly Jr F, et al. Range of motion in total knee replacement. Clin Orthop Relat Res 1996(331):87. 8. Jette M, Sidney K, Lewis W. Fitness, performance and anthropometric characteristics of 19,185 Canadian Forces personnel classified according to body mass index. Mil Med 1990;155(3):120. 9. Weeden SH, Schmidt R. A randomized, prospective study of primary total knee components designed for increased flexion. J Arthroplasty 2007;22(3):349. 10. Wohlrab D, Ditl J, Herrschelmann R, et al. Does the NexGen LPS flex mobile knee prosthesis offer advantages compared to the NexGen LPS? A comparison of clinical and radiological results. Z Orthop Ihre Grenzgeb 2005;143(5):567.
438
D.A. Dennis et al. / The Journal of Arthroplasty 28 (2013) 429–438
11. Minoda Y, Aihara M, Sakawa A, et al. Range of motion of standard and high-flexion cruciate retaining total knee prostheses. J Arthroplasty 2009;24(5):674. 12. Insall JN, Hood RW, Flawn LB, et al. The total condylar knee prosthesis in gonarthrosis. A five to nine-year follow-up of the first one hundred consecutive replacements. J Bone Joint Surg Am 1983;65(5):619. 13. Schurman DJ, Parker JN, Ornstein D. Total condylar knee replacement. A study of factors influencing range of motion as late as two years after arthroplasty. J Bone Joint Surg Am 1985;67(7):1006. 14. McCalden RW, Macdonald SJ, Bourne RB, et al. A randomized controlled trial comparing "high-flex" vs "standard" posterior cruciate substituting polyethylene tibial inserts in total knee arthroplasty. J Arthroplasty 2009;24(6 Suppl):33. 15. Tippett SR, Mang J, Dwyer KA, et al. Collecting data with Palm technology: comparing preoperative expectations and postoperative satisfaction in patients undergoing total knee arthroplasty. J Bone Joint Surg Am 2010;92(Suppl 2):59. 16. Insall JN, Dorr LD, Scott RD, et al. Rationale of the Knee Society clinical rating system. Clin Orthop Relat Res 1989(248):13. 17. Roos EM, Roos HP, Lohmander LS, et al. Knee Injury and Osteoarthritis Outcome Score (KOOS) — development of a self-administered outcome measure. J Orthop Sports Phys Ther 1998;28(2):88. 18. Roos EM, Lohmander LS. The Knee injury and Osteoarthritis Outcome Score (KOOS). from joint injury to osteoarthritis. Health Qual Life Outcomes 2003;1:64. 19. Feller JA, Bartlett RJ, Lang DM. Patellar resurfacing versus retention in total knee arthroplasty. J Bone Joint Surg Br 1996;78(2):226. 20. Cibere J, Bellamy N, Thorne A, et al. Reliability of the knee examination in osteoarthritis: effect of standardization. Arthritis Rheum 2004;50(2):458. 21. Gogia PP, Braatz JH, Rose SJ, et al. Reliability and validity of goniometric measurements at the knee. Phys Ther 1987;67(2):192. 22. FDA. Guidance for industry. Collection of race and ethnicity used in clinical trials; 2005.
23. Zelle J, der Zanden AC Van, De Waal MM, et al. Biomechanical analysis of posterior cruciate ligament retaining high-flexion total knee arthroplasty. Clin Biomech (Bristol, Avon) 2009;24(10):842. 24. Barink M, De Waal MM, Celada P, et al. A mechanical comparison of high-flexion and conventional total knee arthroplasty. Proc Inst Mech Eng [H ] Part H - J Eng Med 2008;222(3):297. 25. Gejo R, Morita Y, Matsushita I, et al. Intraoperative patellar tendon strain: predicting the range of knee flexion after total knee arthroplasty. J Orthop Sci 2009; 14(1):51. 26. Choi WC, Lee S, Seong SC, et al. Comparison between standard and high-flexion posterior-stabilized rotating-platform mobile-bearing total knee arthroplasties: a randomized controlled study. J Bone Joint Surg Am 2010;92(16):2634. 27. Lingard EA, Katz JN, Wright RJ, et al. Validity and responsiveness of the Knee Society Clinical Rating System in comparison with the SF-36 and WOMAC. J Bone Joint Surg Am 2001;83-A(12):1856. 28. Huang HT, Su JY, Wang GJ. The early results of high-flex total knee arthroplasty: a minimum of 2 years of follow-up. J Arthroplasty 2005;20(5):674. 29. Seon JK, Song EK, Lee JY, et al. Comparison of range of motion of high-flexion prosthesis and mobile-bearing prosthesis in total knee arthroplasty. Orthopedics 2005;28(10 Suppl):s1247. 30. Ng FY, Wong HL, Yau WP, et al. Comparison of range of motion after standard and high-flexion posterior stabilised total knee replacement. Int Orthop 2008;32(6): 795. 31. Malik A, Salas A, Ben AJ, et al. Range of motion and function are similar in patients undergoing TKA with posterior stabilised and high-flexion inserts. Int Orthop 2009. 32. Cho SD, Youm YS, Park KB. Three- to six-year follow-up results after high-flexion total knee arthroplasty: can we allow passive deep knee bending? Knee Surg Sports Traumatol Arthrosc 2010.