The Journal of Arthroplasty 30 (2015) 216–222

Contents lists available at ScienceDirect

The Journal of Arthroplasty

journal homepage: www.arthroplastyjournal.org

Selective Patellar Resurfacing in Total Knee Arthroplasty: A Prospective,

Randomized, Double-Blind Study
Donald W. Roberts, MD a, T. David Hayes, MD a, Christine T. Tate, PT b, James P. Lesko, PhD c
a
Northwest Surgical Research Foundation, Vancouver, Washington
b
Department of Physical Therapy, Southwest Washington Medical Center, Vancouver, Washington
c
Biostatistics and Outcomes Research, DePuy Orthopaedics, Warsaw, Indiana

a r t i c l e i n f o a b s t r a c t

Article history: 350 knees were evaluated in a prospective, randomized, double-blinded study of selective patellar resurfacing in
Received 2 June 2014 primary total knee arthroplasty. Knees with exposed bone on the patellar articular surface were excluded. 327
Accepted 16 September 2014 knees were evaluated at a mean follow-up of 7.8 years. 114 knees followed for greater than 10 years were ana-
lyzed separately. Satisfaction was higher in patients with a resurfaced patella. In patients followed for at least 10
Keywords:
years, no significant difference was found. No difference was found in KSS scores or survivorship. No complica-
knee arthroplasty/replacement
patella
tions of patellar resurfacing were identified. The vast majority of patients with remaining patellar articular carti-
resurfacing lage do very well with total knee arthroplasty regardless of patellar resurfacing. Patient satisfaction may be
treatment outcome slightly higher with patellar resurfacing.
prospective study © 2014 Elsevier Inc. All rights reserved.
patient satisfaction

Whether or not to resurface the patella at the time of total knee
arthroplasty has been controversial since the development of the first
patellar prosthesis. The earliest knee arthroplasties did not include pa-
tellar resurfacing. The frequent complaint of anterior knee pain stimu-
lated the development of patellar resurfacing components [1]. A trend
toward routine resurfacing was accompanied by the rapid recognition
of patellofemoral complications related to resurfacing [2,3]. Patellar
malalignment, dislocation, avascular necrosis, extensor mechanism fail-
ure, loosening, anterior knee pain, and difficulties with fracture man-
agement of the resurfaced patella were reported as the most common
complications of total knee arthroplasty [4,5]. Given the prevalence of
these complications, the benefits of patellar resurfacing are debatable.
Even as recently as 2002 a randomized trial of 220 knees found that
10% of the resurfaced patellae required revision [6]. Prior studies were
either inconclusive or failed to agree in the quest for resolution of the
resurfacing dilemma [6–21]. Design features of early generations of
arthroplasty components appear to have compromised the function
and durability of the patellofemoral joint [22,23].
We questioned whether a modern implant with a more accommo-
dating trochlear design could produce a predictably well functioning
arthroplasty without the complications previously associated with pa-
tellar resurfacing. Additionally, previous randomized studies of patellar
resurfacing had included knees with exposed bone in the patellofemoral
The Conflict of Interest statement associated with this article can be found at http://dx.
doi.org/10.1016/j.arth.2014.09.012.
Reprint requests: Donald W. Roberts, M.D., Northwest Surgical Research Foundation,
200 Mother Joseph Place, Suite 210, Vancouver, WA 98664.
joint [13,16,24–26]. This subset of knees with severe patellar damage
may have less optimal outcomes if the patella is not resurfaced. We
chose to study whether or not the patella should be resurfaced in the
population of knees that had remaining articular cartilage on the patel-
lar articular surface. Considering the reported risk of resurfaced patellar
complications, we hypothesized that the not resurfaced patella should
have a better outcome compared to the resurfaced patella when knees
with exposed bone on the patellar articular surface are excluded.
Methods
Enrollment
From July 1996 to April 2001, the patients of two surgeons at one
center were enrolled in a prospective, blinded, randomized study of se-
lective patellar resurfacing. All patients undergoing primary total knee
arthroplasty for a primary diagnosis of osteoarthritis were recruited to
participate in this study. Patients with inflammatory arthritis, avascular
necrosis, previous patellar fracture or osteotomy, or who were undergo-
ing revision knee arthroplasty were excluded from participation. Pa-
tients who were found at the time of surgery to have any exposed
bone on the patellar articular surface were excluded. Institutional Re-
view Board approval was obtained prior to the initiation of the study
and informed consent was obtained from each patient.
An independent physical therapist examined the lower extremity
and obtained a medical history and Knee Society Scores preoperatively.
The same researcher, blinded to the treatment allocation, performed

http://dx.doi.org/10.1016/j.arth.2014.09.012
0883-5403/© 2014 Elsevier Inc. All rights reserved.
 D.W. Roberts et al. / The Journal of Arthroplasty 30 (2015) 216–222 217

Total Knee Patients Screened

496 Patients / 616 Knees
Excluded
226 Patients / 266 Knees
79 Patients: Inflam. Arth, AVN, PTO, Revision, Fx.
70 Patients: Chose not to participate
77 Patients: Excluded in surgery - exposed bone

Enrollment
270 Patients / 350 Knees
172 NR 178 RS

Withdrew
5 Patients / 7 Knees
Lost Contact: Traveler 1 NR 6 RS
1 Patient / 2 Knees
2 NR 0 RS
Incapacity to Return: Limited Health/Cognition
3 Patients / 6 Knees
Deaths 2 NR 4 RS
6 Patients / 8 Knees
4 NR 4 RS

Final Follow-up: Patients with Minimum 2 year Final

Follow-up (Average 7.8 Years)
255 Patients / 327 Knees

Subset of Patients with Minimum 10 Year Final

Follow-up (Average 10.4 Years)
88 Patients / 114 Knees

Revisions Beyond 2 Year Follow-up Deaths Beyond 2 Year Follow-up

14 Patients / 14 Knees 56 Patients / 73 Knees
9 NR 5 RS 36 NR 37 RS

Fig. 1. Study enrollment and knee accountability flowchart.

postoperative evaluations at 6 weeks, 3 months, 1 year, 2 years and
every subsequent 2-year interval.
Knee evaluations used for this study were a pre-operative evaluation
and a final follow-up (minimum 2-year) evaluation. All knees with a
minimum 10-year follow-up were also analyzed separately to provide
a comparison to other reports with 10-year follow-up.
The primary outcomes in this study were patient satisfaction, revi-
sion, Knee Society score and Knee Society function score. Secondary out-
comes included active and passive range of motion, presence of anterior
knee pain and stair climbing ability. Satisfaction was documented at
each follow-up visit on an ordinal scale [27].
Radiographs were evaluated preoperatively, one year postoperatively
and at the time of final evaluation. The radiographs were evaluated for
coronal alignment (whether the patella was located centrally, medially
or laterally in the trochlea on the Merchant radiograph) Insall ratio, the
presence of articular cartilage space on the un-resurfaced patella, whether
the patella was congruent to the trochlea and the angle of patellar tilt.
Surgical Procedure
The implant was the DePuy Sigma fixed bearing cruciate-retaining
knee system. The femoral component was externally rotated three de-
grees from the posterior condylar axis and was placed toward the lateral
side of the resected femoral surface.
Patellar osteophytes were excised. When the patella was resurfaced
the composite patellar thickness was restored to within 2 mm of the
pre-resection thickness. The patellar component was an all-
polyethylene dome-shaped implant with three fixation pegs. The patel-
lar surface was prepared with standard cementing technique. A lateral
retinacular release was performed when the patella was not centered
in the trochlea with the knee flexed 45° and the medial capsular retinac-
ulum unapproximated. The superior lateral geniculate artery was iden-
tified and preserved when possible.
Postoperatively, a continuous passive motion machine was used
for the duration of the hospitalization. Weight bearing as tolerated
was allowed immediately; no immobilization devices were used.
Physical therapy was prescribed three times a week for four to
six weeks.

The surgical procedure was performed under spinal anesthesia. No Randomization

peripheral nerve blocks were used. We employed a midline incision
and a medial parapatellar arthrotomy. The patella was everted and the
patellofemoral joint was inspected. If exposed bone was found on the
patellar articular surface or grossly evident chondrocalcinosis, the patel-
la was resurfaced and the patient was not included in the study. If no ex-
posed bone was found on the patellar articular surface, an envelope was
opened instructing the surgeon whether or not to resurface the patella.
If the patient was undergoing a simultaneous bilateral total knee
arthroplasty, only a single envelope was opened and both patellae
were treated the same.
Prior to initiating the study, the assignment of patellar resurfacing
had been made by a random number generator. The assignments
were placed in opaque envelopes and the envelopes were taken in con-
secutive order. The envelopes were opened in the operating room after
evaluation of the patellar surface. At that time the patient was assigned
to the patellar resurfacing or non-resurfacing group. No discrepancies of
randomization occurred. The treatment allocation was concealed from
the patient. An independent observer who had no knowledge of the
treatment allocation performed clinical evaluations.
218 D.W. Roberts et al. / The Journal of Arthroplasty 30 (2015) 216–222

Table 1
Baseline Demographics and Treatment Group Characteristics (Pre-Operative, Intra-Operative, and Follow-Up).

Not Resurfaced Resurfaced

N Mean S.D. Range N Mean S.D. Range P Value

Age (yrs, at time of surgery) 172 71.3 7.4 48–90 178 70.2 8.7 48–89 0.199
BMI (at time of surgery) 172 29.2 5.0 19.5–44.5 178 29.5 5.4 20.1–43.0 0.584
Pre-op Passive ROM 170 119.3 15.4 74–152 176 116.8 16.6 75–150 0.153
Pre-op Active ROM 171 109.1 15.6 59–141 175 107.8 17.6 47–145 0.494
Pre-op KS Score 164 32.8 18.3 0–87 170 33.5 18.3 0–88 0.740
Pre-op KS Function Score 172 50.2 18.2 0–100 178 51.2 17.4 0–100 0.597
Pre-op KS Stair Sub-score 172 30.7 7.62 0–50 178 31.38 7.32 0–50 0.396
Pre-op Insall Ratio 147 1.07 0.19 0.70–1.80 162 1.05 0.16 0.72–1.61 0.290
Final Follow-up (yrs): Knees with Min 2 yr Follow-up 163 7.4 2.8 1.9–12.2 164 8.2 2.4 1.9–12.0 0.006
Final Follow-up (yrs), Subset with Min 10 yr Follow-up 47 10.5 0.7 9.5–12.2 67 10.3 0.7 9.6–12.0 0.163

Count (%) Count (%)

Gender (Males:Females) 172 89 Males (51.7%) 178 81 Males (45.5%) 0.285

Simultaneous Bilateral Patients 134 34 (25.4%) 138 39 (28.3%) 0.682
Lateral Retinacular Release 172 41 (23.8%) 178 45 (25.3%) 0.804
Pre-op Effusion 170 47 (27.7%) 174 47 (27.0%) 0.904
Pre-op Pain, Walking 172 171 (99.4%) 178 174 (97.8%) 0.372
Pre-op Anterior Pain, Walking 172 30 (17.4%) 178 31 (17.4%) 1.00
Pre-op Pain, at Rest 172 131 (76.2%) 178 138 (77.5%) 0.801
Pre-op Anterior Pain, at Rest 172 18 (10.5%) 178 17 (9.6%) 0.859
Deaths (None related to Study) 172 39 (22.7%) 178 41 (23.0%) 1.00

Statistical Methods
For the purpose of statistical analysis, knees were assumed to be in-
dependent and all analyses of knees were carried out on a per knee basis
rather than a per patient basis. Means for interval variables were com-
pared with a two-sample t-test. Dichotomous variables were compared
with Fisher’s exact test. A P value less than 0.05 was deemed significant.
Baseline demographic variables, preoperative and intra-operative
knee characteristics, average follow-up times and the proportion of
deaths during the 12.5-year study period were compared to confirm
treatment group independence.
Postoperatively, study endpoint measurements were compared
across treatment groups for knees with minimum 2-year final follow-
up, and on the subset of knees with minimum 10-year final follow-up.
Patient satisfaction was converted to an interval scale, 1 = ‘Definitely
Not’ through 5 = ‘Definitely Yes’.
Anterior knee pain while walking was considered as a response var-
iable in three separate logistic regression models evaluating treatment
group with age, gender and BMI as predictor variables. Interaction
terms were confirmed to be not significant. Each of the Knee Society
and range of motion measurements was considered as the response
variable in an analysis of covariance (ANCOVA) model with treatment
group as a predictor variable and age, gender and BMI as covariates.
The purpose of these ANCOVA models was to investigate the impact of
age, gender and BMI on each of the Knee Society and ROM measure-
ments, particularly to see if there was a difference across treatment
groups after adjusting for these covariate effects. Covariates that were
not significant were reduced from respective models stepwise.
A Kaplan Meier survivorship analysis was carried out to obtain sur-
vivorship estimates and to compare survivorship curves across treat-
ment groups.
Results
During the enrollment period, total knee arthroplasties were per-
formed on 616 knees in 496 patients. Two hundred sixty-six knees in

Table 2
Final Follow-Up Measurements on Knees With Minimum 2-Year Follow-Up.

Not Resurfaced Resurfaced

N Mean S.D. Range N Mean S.D. Range P Value

Minimum 2-year Final Follow-up

Passive ROM 140 126.2 10.1 94–150 139 125.1 12.7 78–152 0.438
Active ROM 139 116.6 10.5 86–142 138 116.2 12.6 72–145 0.774
KS Score 138 84.0 13.2 25–100 135 83.7 12.3 12–100 0.867
KS Function Score 162 60.0 28.8 0–100 164 63.0 27.4 0–100 0.318
KS Stair Sub-score 162 36.0 13.1 0–50 158 36.6 11.2 0–50 0.631
Increase in Passive ROM 138 8.1 12.9 −16 to 49 139 7.7 13.6 −27 to 43 0.832
Increase in Active ROM 138 8.9 13.1 −27 to 47 137 7.3 14.9 −30 to 48 0.371
Increase in KS Score 130 50.1 22.4 −7 to 88 130 51.5 21.4 −6 to 93 0.616
Increase in KS Function Score 161 9.3 26.6 −65 to 70 164 10.9 24.3 −50 to 60 0.558
Satisfaction 162 4.73 0.57 2–5 153 4.85 0.47 2–5 0.039
Insall Ratio 144 1.10 0.18 0.70–1.60 150 1.08 0.17 0.68–1.59 0.366

Count (%) Count (%)

Effusion 145 3 (2.1%) 131 10 (7.6%) 0.044

Pain, Walking 162 20 (12.4%) 158 18 (11.4%) 0.864
Anterior Pain, Walking 162 5 (3.1%) 158 5 (3.2%) 1.00
Pain, at Rest 162 21 (13.0%) 158 21 (13.3%) 1.00
Anterior Pain, at Rest 162 4 (2.5%) 158 5 (3.2%) 0.748
 D.W. Roberts et al. / The Journal of Arthroplasty 30 (2015) 216–222 219

Table 3
Final Follow-Up Measurements on Subset With Minimum 10-Year Follow-Up.

Not Resurfaced Resurfaced

N Mean S.D. Range N Mean S.D. Range P Value

Minimum 10-year Follow-up

Passive ROM 43 126.2 10.2 105–142 58 125.5 11.5 83–142 0.748
Active ROM 43 115.4 10.1 97–134 58 116.7 11.4 80–136 0.572
KS Score 42 86.6 11.9 60–100 54 88.0 9.0 62–100 0.506
KS Function Score 47 59.8 26.3 0–100 67 65.6 28.0 0–100 0.266
KS Stair Sub-score 47 36.8 11.1 0–50 66 38.1 10.2 0–50 0.522
Increase in Passive ROM 42 11.1 14.5 −16 to 49 58 10.4 12.9 −12 to 42 0.807
Increase in Active ROM 42 9.6 14.4 −27 to 42 58 10.3 14.3 −25 to 48 0.804
Increase in KS Score 38 53.7 21.2 5 to 88 51 54.6 17.5 10 to 90 0.825
Increase in KS Function Score 47 6.6 25.7 −50 to 60 67 12.9 25.5 −50 to 60 0.198
Satisfaction 47 4.79 0.41 4–5 64 4.91 0.29 4–5 0.096
Insall Ratio 46 1.09 0.21 0.70–1.60 65 1.07 0.18 0.68–1.59 0.679

Count (%) Count (%)

Effusion 44 1 (2.3%) 55 4 (7.3%) 0.378

Pain, Walking 47 6 (12.8%) 66 5 (7.6%) 0.521
Anterior Pain, Walking 47 1 (2.1%) 66 2 (3.0%) 1.00
Pain, at Rest 47 5 (10.6%) 66 5 (7.6%) 0.739
Anterior Pain, at Rest 47 0 (0.0%) 66 2 (3.0%) 0.510

226 patients were excluded. Of these seventy-nine patients had inflam-
matory arthritis, avascular necrosis, previous patellar fracture or
osteotomy, or were undergoing revision knee arthroplasty. Seventy pa-
tients chose not to participate. Seventy-three patients who consented to
participate in the study were excluded at the time of the surgery due to
finding exposed bone on the articular surface of the patella. Four pa-
tients were excluded based on finding chondrocalcinosis at the time of
the surgery. Of the excluded patients, 40 had simultaneous bilateral
total knee arthroplasties.
Three hundred fifty knees in two hundred seventy patients were in-
cluded in the study: 178 resurfaced knees (denoted RS) in 138 patients,
and 172 not resurfaced knees (denoted NR) in 134 patients. There were
80 patients who had both knees enrolled into this study.
The flowchart in Fig. 1 shows study enrollment and treatment allo-
cation, indicating knees lost to follow-up throughout the course of the
study. In some instances, the patient could not return for a follow-up
visit. These patients were interviewed by telephone. Of the primary out-
comes of the study, the occurrence of reoperation, Knee Society function
score and patient satisfaction scores could be obtained, but range of mo-
tion and complete Knee Society scores were not available for these pa-
tients. Sample sizes for Knee Society and ROM measurements are
indicated in the flowchart.
Of 327 knees with greater than 2-year final follow-up, 273 had an
evaluable Knee Society score (138 not resurfaced and 135 resurfaced).
A post-hoc power analysis with Knee Society standard deviation of 13
showed that this sample size of 273 knees was large enough to have de-
tected an average Knee Society score difference of 4.5 or larger with 80%
power and 95% confidence.
Table 1 displays means and proportions across trefsatment groups
for baseline demographic variables, preoperative and intra-operative
knee characteristics, average follow-up times, and the proportion of
deaths during the 12.5-year study duration. Other than average final
follow-up time for knees with minimum 2-year final follow-up, none
of the means or proportions displayed in Table 1 was significantly differ-
ent across treatment groups. The average BMI for both groups was 29.2
The average final follow-up time for knees with minimum 2-year
follow-up was slightly greater (8.2 years) for resurfaced knees com-
pared to not resurfaced knees (7.4 years) (P value = 0.006); the aver-
age follow-up time for knees with minimum 10-year follow-up was
not significantly different across treatment groups.
Final follow-up means and proportions are presented in Tables 2 and
3 for knees with minimum 2-year follow-up and the subset of knees
with minimum 10-year follow-up, respectively. These results include
means for active ROM, passive ROM, Knee Society score, Knee society
function score, and the mean increases in these measurements from
preoperative measurements. In all results displayed in these tables,
the only significant differences across treatment groups were for aver-
age satisfaction and the proportion of knees exhibiting effusion in
knees with minimum 2-year follow-up. Average satisfaction was
lower for knees with not resurfaced patella in knees with minimum 2-
year follow-up (P = 0.039); the difference was not statistically signifi-
cant for the subset of knees with minimum 10-year follow-up. Neither
the means of Knee Society and ROM measurements nor the mean in-
creases from preoperative measurements were significantly different
across treatment groups.
Three logistic regression models were carried out with anterior knee
pain while walking as the response variable (knees with 2-year mini-
mum follow-up); predictor variables were treatment group and age,
treatment group and gender, and treatment group and BMI in the re-
spective models. Treatment group was not a significant predictor in
any of these models.
ANCOVA models of Knee Society Function Score, Knee Society Score,
active ROM and passive ROM as response variables (knees with 2-year
minimum follow-up) were carried out with treatment group as a pre-
dictor variable and age, gender and BMI as regression covariates. The co-
variate adjusted treatment group means were not significantly different
in any of these models.
A Kaplan–Meier survivorship analysis was carried out to analyze revi-
sions for any reason. Knees were followed from the time of surgery until
the time of revision or the time of censorship, which was the last known
follow-up time or the time of death. Nine of 172 (5.8%) not resurfaced
knees were revised during the study duration, and five of 178 (2.8%)
resurfaced knees. The results demonstrated a 90.4% survival for not
resurfaced knees at 10.5 years and a 91.4% survival for resurfaced knees
at 10.4 years (the times when 20 subjects in each respective treatment
group remained uncensored; all revisions happened prior to these
times). The difference in survivorship between treatment groups was
not statistically significant (log-rank P value = 0.207) (Fig. 2).
Complications
Manipulation
Six manipulations under anesthesia were performed for limitation of
range of motion in the early postoperative period. Three of these had pa-
tella resurfacing and three had not resurfaced patellae.
220 D.W. Roberts et al. / The Journal of Arthroplasty 30 (2015) 216–222
Fig. 2. Survivorship curves for not resurfaced and resurfaced patellae.
Loosening
No patellar component was found to loosen clinically or radiograph-
ically. One knee had asymptomatic loosening of the tibial component
detected radiographically. One patient with a resurfaced patella
underwent an arthroscopy for removal of a fragment of bone cement
unrelated to the patellar resurfacing.
Fracture
One patient with a nonresurfaced patella sustained an undisplaced
patellar fracture two years after his knee arthroplasty. This healed un-
eventfully with immobilization. No patient with a resurfaced patella
sustained a fracture.
Infection
There was one superficial infection in the immediate postoperative
period treated with incision and drainage and antibiotics. There was
one deep infection that occurred 2.5 years postoperatively and was
treated successfully with debridement, insert exchange and a six-
week course of antibiotics. Both infections were in patients with
resurfaced patellae.
Revisions
There were 14 revisions (9 not resurfaced knees and 5 resurfaced
knees), all of which occurred more than 2 years post-operatively. Four
knees were revised for anterior knee pain. One patient with a not
resurfaced patella had anterior knee pain with stairs that started shortly
postoperatively. Her patella was resurfaced 3.0 years after her index
arthroplasty. Her pain did not improve after secondary patellar
resurfacing. A second patient with a not resurfaced patella underwent
patellar resurfacing for anterior knee pain 2.9 years after her index
arthroplasty. She was found to have exposed bone on the undersurface
of the patella. Her anterior knee pain was not improved after
resurfacing. A third patient with a not resurfaced patella had anterior
knee pain with stairs and arising from chairs. At the time of secondary
patella resurfacing, 8.5 years after her index arthroplasty she was
found to have exposed bone on the patellar articular surface. Her knee
pain resolved after patellar resurfacing. A fourth patient with a non-
resurfaced patella had patella baja and impingement of the inferior
pole of the patella on the tibial polyethylene. The inferior pole of the pa-
tella was debrided and the polyethylene insert was trimmed with reso-
lution of his symptoms. The patella was not resurfaced at the time of the
second surgery.
The remaining ten revisions were for chronic effusions and synovitis
secondary to polyethylene wear at an average of 7.4 years post-
operatively (range 3.6–10.4 years). These patients did not have anterior
knee pain. In five knees the patella had been resurfaced and in five
knees the patella had not been resurfaced at the index operation. All
of the resurfaced patellae had stable, intact implants present at the
time of revision. Considering our absence of complications with patellar
resurfacing, we elected to resurface the previously not resurfaced patel-
lae at the time of revision for polyethylene wear. All ten patients had re-
lief of their chronic effusions and have not required further surgery.
Although we found that there were more effusions in the resurfaced
knees than the non-resurfaced knees, this finding is of uncertain clinical
significance. These effusions may be due to increased wear debris from
the cumulative effect of an additional polyethylene articulating surface
in resurfaced knees.
Discussion
When our randomized study was initiated in 1996, patellar compli-
cations were considered to be the most common cause of failure in total
knee arthroplasty [16,22]. There was controversy whether or not to re-
surface the patella during total knee arthroplasty. Since that time, sever-
al well-designed, long-term, randomized studies of patellar resurfacing
with mean follow-up of greater than 5 years have been published
[13,16,24–26,28]. These authors reached differing conclusions regard-
ing the advisability of routine patellar resurfacing. One feature of these
reports is that all knees were enrolled in the randomization regardless
of the severity of the articular cartilage loss in the patellofemoral articu-
lation. Many surgeons who selectively resurface the patella do so based
on the appearance of patellar articular cartilage at the time of surgery.
Our study investigates the outcome of this approach to patellar
resurfacing. We randomized only those knees with no exposed bone
on the undersurface of the patella at the time of knee arthroplasty to
determine whether the outcome of patellar resurfacing outweighed
the potential advantages of leaving the patella not resurfaced in those
knees with patellar articular cartilage remaining. We chose this criterion
for inclusion as it was a simple, reproducible distinction that can
be made at the time of surgery for stratifying patellar articular
cartilage damage.
Our results show that the vast majority of patients with remaining
patellar articular cartilage do very well with total knee arthroplasty at
an average follow-up of 7.8 years whether or not the patella was
resurfaced. No statistically significant difference was found whether or
not the patella was resurfaced with regard to anterior knee pain, stair
climbing ability, active or passive range of motion, Knee Society scores
or survivorship. However, we did find a statistically significant
 D.W. Roberts et al. / The Journal of Arthroplasty 30 (2015) 216–222
improvement in patient satisfaction at final follow-up (average
7.8 years) in patients with resurfaced patellae. Although this difference
was statistically significant, the clinical significance may be minimal.
We separately examined all 114 knees that could be followed for
greater than ten years. This is the largest group of patients reported
with 10-year follow-up in a randomized study of patellar resurfacing.
These knees were not different than the larger group of patients with
shorter follow-up except that the improved patient satisfaction in
resurfaced patients no longer achieved statistical significance.
As with other reports [13,24,29], our results demonstrated that pa-
tients whose patella was not resurfaced at the index total knee
arthroplasty tended to have a higher revision rate although this differ-
ence did not achieve statistical significance. Additionally, our results
confirm that secondary patellar resurfacing for anterior knee pain in
the setting of a previously not resurfaced patella does not necessarily re-
sult in the resolution of the pain [30–38]. As noted by Burnett [39] there
are many causes of anterior knee pain other than the patellofemoral
joint. Therefore, in the setting of a not resurfaced patella, the surgeon
may recommend patellar resurfacing for persistent anterior knee pain
assuming that the pain is coming from the not resurfaced patella, but
if the patella has been previously resurfaced, no operation may be sug-
gested. Managing a patient with a not resurfaced patella and persistent
anterior knee pain may result in a secondary resurfacing because it is an
available option. There is, however, little to offer a patient if the patella is
already resurfaced and other clinical and imaging findings are unre-
markable [24]. This may account for the greater prevalence of reopera-
tion in patients with not resurfaced patellae and also for the finding that
in many patients secondary patellar resurfacing does not lead to resolu-
tion of their anterior knee pain.
We found no complications of patellar resurfacing at follow-up as
long as twelve years. Other reports have described patellar fracture, sy-
novial impingement, osteonecrosis, instability, loosening and wear as a
result of patellar resurfacing [40–43]. However, many of these results
were reported prior to understanding the drawbacks of metal-backed
patellar implants and the importance of restoring patellar thickness
and appropriate patellar tracking.
Several authors have suggested that a randomized study of selective
patellar resurfacing may help resolve the controversy of the indications
for patellar resurfacing in total knee arthroplasty [44,45]. We know of
only one other prospective randomized study that addressed selective
patellar resurfacing [46]. These authors stratified patients by
Outerbridge classification. Patients were separately randomized to
resurfaced or not resurfaced patella. The only outcome measure exam-
ined was reoperation for anterior knee pain. They reported that the
risk of patellar resurfacing was 21 times greater when the patella was
not resurfaced in the Outerbridge 4 knees compared to the Outerbridge
1–3 knees. These authors concluded that the patella need not be
resurfaced in Outerbridge 1–3 knees, but should be resurfaced in
Outerbridge 4 knees. Since these authors only evaluated the risk of reop-
eration and did not examine knee pain or function measures, these results
have limited utility for the surgeon deciding whether or not to resurface
the patella in patients with remaining patellar articular cartilage.
A limitation of any study of patellar resurfacing is that the findings
may be specific to the implant design that was studied. The shape and
depth of the trochlear groove of the femoral component are important
factors in the contact forces of the articular cartilage of the not resurfaced
patella [47–50]. This may explain why Waters et al [13] found a much
higher prevalence of anterior knee pain in patients randomized to not
resurfaced patellae. These authors employed an implant with a shallow
trochlear groove that was less accommodating to the native patellar car-
tilage. The design used in our study has a deeper, more capacious trochlea
that may be more accommodating to the native patella.
In conclusion, we aimed to answer the question of whether or not to
resurface the patella at the time of total knee arthroplasty in those knees
with remaining patellar articular cartilage. Unlike previous studies, we
specifically excluded knees with loss of patellar cartilage down to
221
exposed bone. We hypothesized that leaving the native patellar articu-
lar cartilage in this selected population would result in a better outcome
than previous reports of randomized patellar resurfacing that were not
selective as to the condition of the patellar articular cartilage. However,
despite employing an implant with a trochlear groove more accommo-
dating to the native patella and excluding knees with exposed bone in
the patellar articular surface, resurfaced knees had slightly better pa-
tient satisfaction than not resurfaced knees. While we demonstrated
that excellent results can be achieved with not resurfacing the patella,
the absence of complications we found with resurfacing suggests that
the surgeon can confidently resurface the patella without the risks pre-
viously associated with patellar resurfacing using older techniques or
earlier component design.
Acknowledgments
The authors would like to recognize the contribution and support of
DePuy Orthopedics, PeaceHealth Southwest Washington Medical Cen-
ter and our study coordinators: Lynette Alber R.N., Sherri Tzvetcoff
and Charlanne Sappington R.N.
References
1. Insall JN, Ranawat CS, Aglietti P, et al. A comparison of four models of total knee-
replacement prostheses. J Bone Joint Surg Am 1976;58(6):754.
2. Picetti III GD, McGann WA, Welch RB. The patellofemoral joint after total knee
arthroplasty without patellar resurfacing. J Bone Joint Surg Am 1990;72(9):1379.
3. Healy WL, Wasilewski SA, Takei R, et al. Patellofemoral complications following total
knee arthroplasty. Correlation with implant design and patient risk factors. J
Arthroplasty 1995;10(2):197.
4. Dennis DA. Extensor mechanism problems in total knee arthroplasty. Instr Course
Lect 1997;46:171.
5. Rand JA. The patellofemoral joint in total knee arthroplasty. J Bone Joint Surg Am
1994;76(4):612.
6. Wood DJ, Smith AJ, Collopy D, et al. Patellar resurfacing in total knee arthroplasty: a
prospective, randomized trial. J Bone Joint Surg Am 2002;84-A(2):187.
7. Boyd Jr AD, Ewald FC, Thomas WH, et al. Long-term complications after total knee
arthroplasty with or without resurfacing of the patella. J Bone Joint Surg Am 1993;
75(5):674.
8. Feller JA, Bartlett RJ, Lang DM. Patellar resurfacing versus retention in total knee
arthroplasty. J Bone Joint Surg (Br) 1996;78(2):226.
9. Keblish PA, Varma AK, Greenwald AS. Patellar resurfacing or retention in total knee
arthroplasty. A prospective study of patients with bilateral replacements. J Bone
Joint Surg (Br) 1994;76(6):930.
10. Barrack RL, Bertot AJ, Wolfe MW, et al. Patellar resurfacing in total knee arthroplasty.
A prospective, randomized, double-blind study with five to seven years of follow-up. J
Bone Joint Surg Am 2001;83-A(9):1376.
11. Burnett RSJ, Boone JL, McCarthy KP, et al. A prospective randomized clinical trial of
patellar resurfacing and nonresurfacing in bilateral TKA. Clin Orthop 2007;464:65.
http://dx.doi.org/10.1097/BLO.0b013e31812f783b.
12. Ogon M, Hartig F, Bach C, et al. Patella resurfacing: no benefit for the long-term out-
come of total knee arthroplasty. A 10- to 16.3-year follow-up. Arch Orthop Trauma
Surg 2002;122(4):229.
13. Waters TS, Bentley G. Patellar resurfacing in total knee arthroplasty. A prospective,
randomized study. J Bone Joint Surg Am 2003;85-A(2):212.
14. Peng CW, Tay BK, Lee BPH. Prospective trial of resurfaced patellaversus non-
resurfaced patella in simultaneous bilateral total knee replacement. Singapore Med
J 2003;44(7):347.
15. Pakos EE, Ntzani EE, Trikalinos TA. Patellar resurfacing in total knee arthroplasty. A
meta-analysis. J Bone Joint Surg Am 2005;87(7):1438.
16. Campbell DG, Duncan WW, Ashworth M, et al. Patellar resurfacing in total knee re-
placement: a ten-year randomised prospective trial. J Bone Joint Surg (Br) 2006;88
(6):734.
17. Seo SS, Kim CW, Moon SW. A comparison of patella retention versus resurfacing for
moderate or severe patellar articular defects in total knee arthroplasty: minimum
5-year follow-up results. Knee Surg Relat Res 2011;23(3):142.
18. Park S-J, Jung Y-B, Jeong HJ, et al. Long-term results of primary total knee arthroplasty
with and without patellar resurfacing. Acta Med Okayama 2010;64(5):331.
19. Chen K, Li G, Fu D, et al. Patellar resurfacing versus nonresurfacing in total knee
arthroplasty: a meta-analysis of randomised controlled trials. Int Orthop 2013;37
(6):1075.
20. Pilling RWD, Moulder E, Allgar V, et al. Patellar resurfacing in primary total knee re-
placement: a meta-analysis. J Bone Joint Surg Am 2012;94(24):2270.
21. Lygre SHL, Espehaug B, Havelin LI, et al. Failure of total knee arthroplasty with or
without patella resurfacing. Acta Orthop 2011;82(3):282.
22. Brick GW, Scott RD. The patellofemoral component of total knee arthroplasty. Clin
Orthop 1988;231:163.
23. Freeman MA, Samuelson KM, Elias SG, et al. The patellofemoral joint in total knee
prostheses. Design considerations. J Arthroplasty 1989;4:S69 [Suppl].
222 D.W. Roberts et al. / The Journal of Arthroplasty 30 (2015) 216–222
24. Burnett RSJ, Boone JL, Rosenzweig SD, et al. Patellar resurfacing compared with
nonresurfacing in total knee arthroplasty. A concise follow-up of a randomized
trial. J Bone Joint Surg Am 2009;91(11):2562.
25. Mayman D, Bourne RB, Rorabeck CH, et al. Resurfacing versus not resurfacing the pa-
tella in total knee arthroplasty: 8- to 10-year results. J Arthroplasty 2003;18(5):541.
26. Breeman S, Campbell M, Dakin H, et al. Patellar resurfacing in total knee replacement:
five-year clinical and economic results of a large randomized controlled trial. J Bone
Joint Surg Am 2011;93(16):1473.
27. American Academy of Orthopedic Surgeons Council of Musculoskeletal Specialty So-
cieties. Outcomes data collection package; 1997.
28. Beaupre L, Secretan C, Johnston DWC, et al. A randomized controlled trial comparing
patellar retention versus patellar resurfacing in primary total knee arthroplasty: 5–10
year follow-up. BMC Res Notes 2012;5:273. http://dx.doi.org/10.1186/1756-0500-5-
273.
29. Burnett RS, Haydon CM, Rorabeck CH, et al. Patella resurfacing versus nonresurfacing
in total knee arthroplasty: results of a randomized controlled clinical trial at a mini-
mum of 10 years’ follow-up. Clin Orthop 2004;428:12.
30. Campbell DG, Mintz AD, Stevenson TM. Early patellofemoral revision following total
knee arthroplasty. J Arthroplasty 1995;10(3):287.
31. Garcia RM, Kraay MJ, Goldberg VM. Isolated resurfacing of the previously
unresurfaced patella total knee arthroplasty. J Arthroplasty 2010;25(5):754.
32. Mockford BJ, Beverland DE. Secondary resurfacing of the patella in mobile-bearing
total knee arthroplasty. J Arthroplasty 2005;20(7):898.
33. Khatod M, Codsi M, Bierbaum B. Results of resurfacing a native patella in patients
with a painful total knee arthroplasty. Am J Knee Surg 2004;17(3):151.
34. Karnezis IA, Vossinakis IC, Rex C, et al. Secondary patellar resurfacing in total knee
arthroplasty: results of multivariate analysis in two case-matched groups. J
Arthroplasty 2003;18(8):993.
35. Muoneke HE, Khan AM, Giannikas KA, et al. Secondary resurfacing of the patella for
persistent anterior knee pain after primary knee arthroplasty. J Bone Joint Surg (Br)
2003;85(5):675.
36. Clements WJ, Miller L, Whitehouse SL, et al. Early outcomes of patella resurfacing in
total knee arthroplasty. Acta Orthop 2010;81(1):108.
37. Parvizi J, Mortazavi SMJ, Devulapalli C, et al. Secondary resurfacing of the patella after
primary total knee arthroplasty does the anterior knee pain resolve? J Arthroplasty
2012;27(1):21.
38. Daniilidis K, Vogt B, Gosheger G, et al. Patellar resurfacing as a second stage procedure
for persistent anterior knee pain after primary total knee arthroplasty. Int Orthop
2012;36(6):1181.
39. Burnett RS, Bourne RB. Indications for patellar resurfacing in total knee arthroplasty.
Instr Course Lect 2004;53:167.
40. Chalidis BE, Tsiridis E, Tragas AA, et al. Management of periprosthetic patellar frac-
tures. A systematic review of literature. Injury 2007;38(6):714.
41. Pagnano MW, Trousdale RT. Asymmetric patella resurfacing in total knee
arthroplasty. Am J Knee Surg 2000;13(4):228.
42. Lindstrand A, Robertsson O, Lewold S, et al. The patella in total knee arthroplasty:
resurfacing or nonresurfacing of patella. Knee Surg Sports Traumatol Arthrosc
2001;9(Suppl. 1):S21.
43. Rader CP, Löhr J, Wittmann R, et al. Results of total knee arthroplasty with a metal-
backed patellar component. A 6-year follow-up study. J Arthroplasty 1996;11(8):923.
44. Gross AE. Patellar resurfacing for total knee replacement: ultimate answer may lie in
subgroups: commentary on an article. In: Breeman Suzanne, et al, editors. Patellar
resurfacing in total knee replacement: five-year clinical and economic results of a
large randomized controlled trial, 93(16). , J Bone Joint Surg Am; 2011. p. e94 1.
45. Bourne RB. To resurface the patella or not? Better assessments needed to address the
benefits for total knee replacement: commentary on an article. In: Pavlou George,
et al, editors. Patellar resurfacing in total knee arthroplasty: does design matter? A
meta-analysis of 7075 cases, 93(14). , J Bone Joint Surg Am; 2011. p. e82.
46. Rodríguez-Merchán EC, Gómez-Cardero P. The outerbridge classification predicts the
need for patellar resurfacing in TKA. Clin Orthop 2010;468(5):1254.
47. Whiteside LA, Nakamura T. Effect of femoral component design on unresurfaced pa-
tellas in knee arthroplasty. Clin Orthop 2003;410:189. http://dx.doi.org/10.1097/01.
blo.0000063605.67412.4d.
48. Matsuda S, Ishinishi T, Whiteside LA. Contact stresses with an unresurfaced patella in
total knee arthroplasty: the effect of femoral component design. Orthopedics 2000;23
(3):213.
49. Tanzer M, McLean CA, Laxer E, et al. Effect of femoral component designs on the con-
tact and tracking characteristics of the unresurfaced patella in total knee arthroplasty.
Can J Surg 2001;44(2):127.
50. Petersilge WJ, Oishi CS, Kaufman KR, et al. The effect of trochlear design on
patellofemoral shear and compressive forces in total knee arthroplasty. Clin Orthop
1994;309:124.