Abstract
Background: Patellar resurfacing during total knee arthroplasty
remains controversial. We aimed to evaluate the effectiveness of this
technique through an evaluation of the current literature.
Methods: We performed a meta-analysis of randomized controlled
trials comparing total knee arthroplasties performed with and without patellar
resurfacing. Outcomes of interest included the number of reoperations, the
prevalence of postoperative anterior knee pain, and the improvement in various
knee scores.
Results: Ten trials assessing 1223 knees were eligible. The absolute
risk of reoperation was reduced by 4.6% (95% confidence interval, 1.9% to
7.3%) in the patellar resurfacing arm (between-study heterogeneity, p <
0.01; I2 = 60%), implying that one would have to resurface
twenty-two patellae (95% confidence interval, fourteen to fifty-two patellae)
in order to prevent one reoperation. Patellar resurfacing reduced the absolute
risk of postoperative anterior knee pain by 13.8% (95% confidence interval,
6.4% to 21.2%), implying that one would have to resurface seven patellae (95%
confidence interval, five to sixteen patellae) in order to prevent one case of
postoperative anterior knee pain. Only four trials provided adequate data for
a quantitative synthesis of the changes in the various knee scores; on the
basis of those four trials, there was no difference in the mean improvement in
the knee scores (standardized mean difference, 0.03; 95% confidence interval,
-0.50 to 0.56).
Conclusions: The available evidence indicates that patellar
resurfacing reduces the risks of reoperation and anterior knee pain after
total knee arthroplasty. The observed effects are clinically important despite
their modest magnitude. Additional, carefully designed randomized trials are
required to strengthen this claim.
Level of Evidence: Therapeutic Level I. See Instructions
to Authors for a complete description of levels of evidence.
The treatment of the patella at the time of total knee arthroplasty
remains controversial. The early implants for total knee arthroplasty were not
designed for patellar resurfacing. The persistent or recurrent anterior knee
pain that occurred at high rates following treatment with early implants,
along with other complications such as dislocation, maltracking, and
subluxation, typically were attributed to the patellofemoral joint. This
prompted the development and use of tricompartmental replacements that allowed
patellar
resurfacing1,2.
However, as a variety of complications requiring reoperation also were
observed in patients with resurfaced
patellae3-6,
it was proposed that resurfacing should be reserved for patients with
inflammatory arthritis, severe preoperative patellar pain, large or thick
patellae, or multiple knee operations as well as for those who are overweight
or poorly
compliant7-9.
The effectiveness of the surgical options of patellar resurfacing and not
resurfacing has been extensively studied. Randomized controlled trials
constitute the most reliable source of evidence for the evaluation of the
efficacy of a potential
intervention10.
Several randomized controlled trials have compared total knee arthroplasties
performed with and without patellar
resurfacing11-22;
however, there have been conflicting results. Most of these trials had
relatively small sample sizes and probably were underpowered to detect
differences between the two treatment options. Thus, several questions have
been posed that require quantitative synthesis of the accumulated evidence:
(1) Is reoperation less frequently needed when the patella is resurfaced? (2)
How commonly is postoperative anterior knee pain associated with each method
of treatment? (3) What is the mean improvement in the knee scores associated
with each type of treatment? In order to address these questions, we conducted
a meta-analysis of all available randomized trials of which we were aware that
compared total knee arthroplasties performed with and without patellar
resurfacing.
Identification and Eligibility of Relevant Randomized Studies
We searched MEDLINE, EMBASE, and the Cochrane Central Register of
Controlled Trials for randomized trials comparing total knee arthroplasties
performed with and without patellar resurfacing; the last search was performed
in November 2004 (see Appendix). References of retrieved articles were also
screened. We set no language restrictions.
We considered all randomized trials comparing total knee arthroplasties
performed with and without patellar resurfacing, regardless of the indication
for total knee arthroplasty, the surgical technique, or the specific type of
prosthesis used. All nonrandomized trials were excluded. We also excluded
studies of bilateral total knee arthroplasties in which the patella was
resurfaced in one knee and not in the other when the allocation of the two
techniques to the operated knees was not randomized. Whenever reports
pertained to the same patients at different follow-up periods, we retained the
one with the longer duration of follow-up for the main analysis in order to
avoid duplication of information. However, reports with shorter durations of
follow-up were used in subgroup analyses comparing short-term with long-term
follow-up trials. Personal communication with the authors of published
eligible reports was attempted whenever the available published information
was not adequate for the analysis.
Outcomes
The primary outcome of our meta-analysis was the number of reoperations for
any reason, including later additional procedures on the extensor mechanism,
such as lateral retinacular release. Secondary end points were the presence of
postoperative anterior knee pain of any grade as well as symptom and function
scores such as the Knee Society
score23 (including
the combined knee and function components of this score), The Hospital for
Special Surgery
score15, and the
Bristol knee
score24. Each study
comparing total knee arthroplasties performed with and without patellar
resurfacing was included in our meta-analysis, regardless of the type and
number of outcomes used to evaluate the efficacy of either method.
Data Extraction
All three investigators independently extracted the data and then reached
consensus on all extraction items. For each report, we gathered information on
characteristics of the trial as well as of the studied population. We recorded
the number of patients who underwent a secondary operation and the number of
patients with postoperative anterior knee pain of any grade. We also kept
information on the number of patients with different severities of pain (mild
pain or moderate/severe pain), respecting the classification in the individual
trials. Finally, we recorded the average knee symptom and functionality
assessment scores and the corresponding standard deviations preoperatively and
postoperatively (see Appendix).
Meta-Analysis
The knee was the unit of the analyses. For each trial, we calculated the
relative risk and the risk differences for reoperation and the presence of
postoperative anterior knee pain (along with the corresponding 95% confidence
intervals) as well as the standardized mean
difference25 (and
95% confidence interval) in the mean score improvement between the compared
arms. The standardized mean difference expresses the mean score improvement in
standard deviation units and can be used to directly compare different scales
or scores across the individual studies (see Appendix).
The overall summary effect sizes were estimated with
fixed-effects26 and
random-effects27
models. Random effects are more appropriate in the presence of between-study
heterogeneity27,
provided that events are not
rare28. Therefore,
especially for the reoperation outcomes, for which events were relatively
rare, fixed-effects syntheses were employed. We tested for heterogeneity with
the Q statistic (traditionally considered to be significant at p <
0.1025,26)
and quantified its extent with the I2
statistic29
(possible range, 0% to 100%) (see Appendix). We also calculated the number of
knees that needed to be treated in order to avoid a reoperation or
postoperative anterior knee pain. The number needed to treat is defined as the
absolute value of the inverse of the estimated risk difference
(1|RD|).
We performed analyses in a priori defined subgroups of trials in order to
explain the observed between-study heterogeneity and to identify
subgroup-specific differences in the effect of the intervention. As it is
likely that most complications occur after the third or fourth year
postoperatively4, we
conducted separate analyses for trials with less than five years of follow-up
and for trials with at least five years of follow-up. Further subgroup
analyses were based on the year of publication (before 2000 as opposed to 2000
or later), the presence of quality characteristics (adequate descriptions of
blinding and allocation concealment) in the published reports, and the
indication for arthroplasty (osteoarthritis or other). We further tested
whether the individual trial effect sizes for the primary outcome were
dependent on the mean (or median) weight of the enrolled patients with use of
meta-regression30.
Finally, we explored whether larger studies showed systematically different
results from smaller studies; such differences may suggest the presence of
biases (particularly publication bias). For these assessments, we generated
inverted funnel
plots31 and used
the Begg-Mazumdar publication bias
diagnostic32, as
appropriate. We also examined whether there were significant discrepancies
between the largest trial and the overall synthesis of the smaller trials and
calculated the corresponding z-scores.
Unless otherwise specified, random-effects syntheses are presented.
Analyses were conducted with Stata software (version 8.2; Stata, College
Station, Texas) and Meta-Analyst (Joseph Lau, Boston, Massachusetts). All p
values were two-tailed, and the level of significance was set at p <
0.05.
Eligible Studies
Ten independent randomized trials (described in twelve
papers11-22)
were included in the overall meta-analysis, with a cumulative sample size of
1223 involved knees (Table I).
All published reports provided data on reoperations during follow-up. Five
reports provided details on the prevalence of anterior knee pain of any grade
postoperatively16,17,19-21.
Details on the definition of "mild," "moderate," or
"severe" pain were not described. There was not enough information
to perform intention-to-treat analyses. Only four trials provided adequate
data for a quantitative synthesis of the changes in the various knee scores
(see Appendix). Study characteristics and quality items are described in
Table I.
Meta-Analysis
Reoperation—Main Analyses
The rate of reoperations varied across the eligible trials from 0% to 13%
in the resurfacing group and from 0% to 20% in the nonresurfacing group. When
all studies were considered, the frequency of reoperations in the patellar
resurfacing group was 0.48 times lower (95% confidence interval, 0.30 to 0.75
times lower; fixed-effects analysis) than that in the nonresurfacing group,
with little between-study heterogeneity (p = 0.15; I2 = 32%)
(Fig. 1). Similarly, the
absolute risk of reoperation was reduced by 4.6% (95% confidence interval,
1.9% to 7.3%; fixed-effects analysis) in the patellar resurfacing arm, but
with substantial between-study heterogeneity (p < 0.01; I2 =
60%) (Fig. 1), implying that
one would need to resurface twenty-two patellae (95% confidence interval,
fourteen to fifty-two patellae) to prevent one reoperation. Random-effects
syntheses yielded similar results (see Appendix).
Reoperation—Subgroup Analyses
When we examined trials in which patients had been followed for less than
five years, there was no significant difference between the compared arms with
regard to either the relative risks or the risk differences (see Appendix). On
the other hand, among trials with at least five years of follow-up, the odds
of reoperation were significantly smaller (relative risk, 0.19; 95% confidence
interval, 0.08 to 0.46; fixed-effects analysis) in the resurfaced arm
(Fig. 2). The absolute risk was
reduced by 6.7% (95% confidence interval, 3.6% to 9.8%) in the resurfacing
arm, meaning that one would need to resurface fifteen patellae (95% confidence
interval, ten to twenty-eight patellae) to prevent one reoperation. The
summary effects also tended to favor the resurfacing arm across subgroups
defined by the quality characteristics of the published reports (see
Appendix). The effects of patellar resurfacing were very similar to the
overall analyses when only osteoarthritic knees were examined. The number of
patients with rheumatoid arthritis affecting the knee was too small to permit
the estimate of any effects (see Appendix). Finally, the individual trial risk
ratio and risk difference were not dependent on the average weight of the
patients (meta-regression, p = 0.79 and p = 0.34, respectively).
Postoperative Anterior Knee Pain
Among the five reports that provided relevant data, the presence of
postoperative anterior knee pain of any grade was 0.40 times lower (95%
confidence interval, 0.19 to 0.85 times lower) in the resurfacing arm
(Fig. 3). Patellar resurfacing
reduced the absolute risk of knee pain by 13.8% (95% confidence interval, 6.4%
to 21.2%) (Fig. 3), implying
that one would need to resurface seven patellae (95% confidence interval, five
to sixteen patellae) to prevent one case of postoperative anterior knee pain.
Between-study heterogeneity was significant for both of the above comparisons
(see Appendix).
Similar results were drawn from subgroup analyses. Of note, the nonblinded
trials had the most impressive reduction in postoperative anterior knee pain:
the prevalence of pain was approximately five times smaller (relative risk =
0.18; 95% confidence interval, 0.04 to 0.81) in the resurfacing arm, which
means that one would need to resurface seven patellae (95% confidence
interval, four to 125 patellae) to prevent a case of postoperative knee pain
on the basis of the risk differences described in a table in the Appendix.
The relative risk for the emergence of only "mild"
postoperative anterior knee pain was 0.33 (95% confidence interval, 0.22 to
0.50; fixed-effects analysis) when the patella was resurfaced. For patients
with "moderate or severe" pain, the relative risk was 0.19 (95%
confidence interval, 0.09 to 0.38; fixed-effects analysis).
Knee Scores
The Knee Society score (specifically, the combined knee and function
components of this score), the Bristol knee score, and The Hospital for
Special Surgery score were used for the clinical assessment of patients (in
eight trials, one trial, and one trial, respectively). However, summary score
data necessary for quantitative analyses (that is, mean changes and the
standard deviations thereof) were incompletely reported. The corresponding
standardized mean differences in score improvement were calculated for only
four
trials11,13,15,22
(on the basis of the information in the publication for one trial and after
personal communication with the authors for three). Hence, the corresponding
syntheses should be interpreted with caution. The available data were very
heterogeneous and, overall, they showed no significant difference between the
compared arms (standardized mean difference, 0.03; 95% confidence interval,
-0.50 to 0.56; between-study heterogeneity, p < 0.01; I2 = 76%)
(Fig. 4). As reported in the
published papers, the remaining four studies from the group of eight trials
that measured mean improvement in knee scores from which no relevant data
could be extracted tended to favor
resurfacing19,20,
not resurfacing17,
or neither16.
However, the magnitude and the uncertainty of these tendencies could not be
deduced.
Larger Versus Smaller Trials
There was no evidence of publication bias in the overall metaanalyses of
the reoperation rates and the emergence of anterior knee pain (see Appendix).
For the outcome of reoperation, the largest
trial20 (with 474
analyzed knees) yielded estimates that were similar to those of the synthesis
of the remaining studies (p > 0.05 for all pertinent contrasts based on the
z-scores). However, for the outcome of postoperative anterior knee pain, the
largest trial favored patellar resurfacing more strongly than the smaller
trials did (p = 0.05).
This meta-analysis of 1223 randomized knees documents that the rate
of reoperation is lower following total knee arthroplasty with patellar
resurfacing than it is after total knee arthroplasty without patellar
resurfacing. The risk of reoperation was substantially (approximately 50%)
smaller in the resurfacing group, and this finding was more evident in the
subgroup of trials with adequately long follow-up. Despite the large relative
risk estimates, the absolute risk difference was small. Similarly, the
prevalence of anterior knee pain was reduced by >10% in patients with
patellar resurfacing. As calculated, this implies that one would prevent one
reoperation for every twenty-two patellae resurfaced and one case of
postoperative anterior knee pain for every seven patellae resurfaced. Data
were too sparse to allow a solid analysis of the improvement in the various
score scales, and therefore those syntheses should be interpreted with
caution.
Published evidence supports the notion that many complications appear in
the long term, after the third or fourth postoperative
year4. In this
respect, estimates from trials with longer follow-up may be more informative.
The observed difference in the number of reoperations was more evident in
studies with follow-up periods of five years or more. However, the estimated
risk difference was modest; as calculated from this subgroup of studies, one
would prevent a reoperation for every fifteen patellae resurfaced. For the
outcome of postoperative anterior knee pain, similar inferences based on
subgroups of trials with longer or shorter follow-up may be precarious because
of the limited amount of accumulated information.
The findings of our meta-analysis might have been biased by the fact that
the analysis inevitably was "on-treatment" and not by
"intention-to-treat." Intention-to-treat analyses compare patients
according to the groups to which they were initially assigned, irrespective of
the treatment that they actually received. Thus, they reflect a real clinical
situation because they retain the balance in prognostic factors arising from
the original random treatment allocation, give an unbiased estimate of
treatment effect, and admit noncompliance and protocol deviations. On the
other hand, "on-treatment" analyses compare patients according to
the treatment that they actually received, regardless of the initial
allocation. This type of analysis does not take into account dropouts from the
study for systematic reasons and thus allows unknown factors that may be
influencing the treatment to affect the reported outcomes. In the present
study, intention-to-treat analysis was impossible to perform.
Postrandomization exclusions were reported in six trials, but this was not
done systematically, and most studies did not present separate counts across
the compared arms. If anything, formal intention-to-treat analyses would blunt
the calculated effect sizes.
Some limitations of our study should be acknowledged. First, the cumulative
sample size was not very large because the majority of trials included
relatively few knees. This was especially true for most of the subgroup
analyses. With 1200 randomized knees, empirical data from other medical
domains33 suggest
that one would not expect very large fluctuations in the summary risk ratio
with the accumulation of more randomized trials. Second, in some trials, data
on the prevalence of anterior knee pain and the various knee scores were
incompletely reported or were not reported at all, precluding their inclusion
in the calculations. Incomplete reporting or nonreporting of outcomes may be
related to their level of significance in the pertinent
trials34. This
phenomenon has been termed "outcome reporting
bias"34 and
may influence the results of the quantitative synthesis. Moreover, the
classification of anterior knee pain as "mild,"
"moderate," or "severe" was not defined in detail in
the individual trials. Thus, the summary results regarding pain severity
should be interpreted with caution. However, the estimates still favored
patellar resurfacing when a broader definition for the pain outcome (pain of
any grade) was used. Third, some of the included randomized trials did not
adequately describe quality items (e.g., mode of randomization, blinding,
allocation concealment) that are often used for the assessment of overall
trial quality. It has been hypothesized that differences in the individual
trial effect sizes may be due to differences in mean body
weight15,19,
the frequency of preoperative anterior knee
pain35 or patellar
maltracking36, the
type of
arthritis37, the
attention to technical detail during
surgery38, and the
type of prosthesis
used39-41.
Formal testing of these speculations would necessitate detailed data on
individual patients. However, meta-regression and subgroup analyses did not
indicate that the individual trial effect sizes had any significant dependency
on the average patient weight or the specific type of arthritis, factors for
which information was available.
In our search of the literature, we identified a meeting abstract
describing a meta-analytic attempt by Rapuri et
al.42 on the same
subject. The authors reported that the prevalence of anterior knee pain was
lower and the cumulative functional scores were higher for the patellar
resurfacing group, whereas the prevalence of reoperation and the rate of
revision for patellofemoral problems were the same for both groups. However,
the incomplete reporting of eligibility criteria, the absence of a description
of the synthetic methods used to assess the outcomes, the exclusion of studies
published in languages other than English, and the inadequate description of
results precluded any comparison with the outcomes of our meta-analysis. Our
results point to the same direction as a previous decision
analysis43 that
compared the two treatment strategies. That decision analysis identified
patellar resurfacing as a marginally superior strategy.
Details of the research methodology, inverted funnel plots for selected
outcomes, and tables presenting details of the studies analyzed are available
with the electronic versions of this article, on our web site at
(go to
the article citation and click on "Supplementary Material") and on
our quarterly CD-ROM (call our subscription department, at 781-449-9780, to
order the CD-ROM). ?
The authors did not receive grants or outside funding in support of their
research or preparation of this manuscript. They did not receive payments or
other benefits or a commitment or agreement to provide such benefits from a
commercial entity. No commercial entity paid or directed, or agreed to pay or
direct, any benefits to any research fund, foundation, educational
institution, or other charitable or nonprofit organization with which the
authors are affiliated or associated.
Insall JN, Ranawat CS, Aglietti P, Shine
J. A comparison of four models of total knee-replacement prostheses. J
Bone Joint Surg Am.1976;58:
754-65.58754
1976
[PubMed]
Jones EC, Insall JN, Inglis AE, Ranawat
CS. GUEPAR knee arthroplasty results and late complications. Clin
Orthop Relat Res.1979;140:
145-52.140145
1979
Grace JN, Sim FH. Fracture of the
patella after total knee arthroplasty. Clin Orthop Relat Res.1988;230:
168-75.230168
1988
[PubMed]
Boyd AD Jr, Ewald FC, Thomas WH, Poss R,
Sledge CB. Long-term complications after total knee arthroplasty with or
without resurfacing of the patella. J Bone Joint Surg Am.1993;75:
674-81.75674
1993
[PubMed]
Windsor RE, Scuderi GR, Insall JN.
Patellar fractures in total knee arthroplasty. J Arthroplasty.1989;4 Suppl:
S63-7.4S63
1989
[PubMed][CrossRef]
Hernandez-Vaquero D, Alvarez-Gonzalez
UJ, Fernandez-Corona C, Garcia-Sandoval MA, Rubio-Gonzalez A. Patellar
complications after total knee arthroplasty. Int Orthop.1996;20:
103-6.20103
1996
[PubMed][CrossRef]
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:
930-7.76930
1994
Kim BS, Reitman RD, Schai PA, Scott RD.
Selective patellar nonresurfacing in total knee arthroplasty. 10 year results.
Clin Orthop Relat Res.1999;367:
81-8.36781
1999
[PubMed]
Fern ED, Winson IG, Getty CJ. Anterior
knee pain in rheumatoid patients after total knee replacement. Possible
selection criteria for patellar resurfacing. J Bone Joint Surg
Br.1992;74:
745-8.74745
1992
Piantadosi S. Clinical trials: a
methodologic perspective. New York: Wiley;
1997.
1997
Newman JH, Ackroyd CE, Shah NA,
Karachalios T. Should the patella be resurfaced during total knee replacement?
Knee.2000;7:
17-23.717
2000
[CrossRef]
Schroeder-Boersch H, Scheller G, Fischer
J, Jani L. Advantages of patellar resurfacing in total knee arthroplasty.
Two-year results of a prospective randomized study. Arch Orthop Trauma
Surg.1998;117:
73-8.11773
1998
[CrossRef]
Bourne RB, Rorabeck CH, Vaz M, Kramer J,
Hardie R, Robertson D. Resurfacing versus not resurfacing the patella during
total knee replacement. Clin Orthop Relat Res.1995;321:
156-61.321156
1995
[PubMed]
Mayman D, Bourne RB, Rorabeck CH, Vaz M,
Kramer J. Resurfacing versus not resurfacing the patella in total knee
arthroplasty: 8- to 10-year results. J Arthroplasty.2003;18:
541-5.18541
2003
[PubMed][CrossRef]
Feller JA, Bartlett RJ, Lang DM.
Patellar resurfacing versus retention in total knee arthroplasty. J
Bone Joint Surg Br.1996;78:
226-8.78226
1996
Partio E, Wirta J. Comparison of
patellar resurfacing and nonresurfacing in total knee arthroplasty: a
prospective, randomized study. J Orthop Rheumatol.1995;8:
69-74.869
1995
Barrack RL, Bertot AJ, Wolfe MW, Waldman
DA, Milicic M, Myers L. 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:
1376-81.831376
2001
[PubMed]
Barrack RL, Wolfe MW, Waldman DA,
Milicic M, Bertot AJ, Myers L. Resurfacing of the patella in total knee
arthroplasty. A prospective, randomized, double-blind study. J Bone
Joint Surg Am.1997;79:
1121-31.791121
1997
Wood DJ, Smith AJ, Collopy D, White B,
Brankov B, Bulsara MK. Patellar resurfacing in total knee arthroplasty: a
prospective, randomized trial. J Bone Joint Surg Am.2002;84:
187-93.84187
2002
[PubMed]
Waters TS, Bentley G. Patellar
resurfacing in total knee arthroplasty. A prospective, randomized study.
J Bone Joint Surg Am.2003;85:
212-7.85212
2003
[PubMed]
Waikakul S, Vanadurongwan V, Bintachitt
P. The effects of patellar resurfacing in total knee arthroplasty on position
sense: a prospective randomized study. J Med Assoc Thai.2000;83:
975-82.83975
2000
[PubMed]
Kordelle J, Schleicher I, Kaltschmidt I,
Haas H, Gruner MR. [Patella resurfacing in patients without substantial
retropatellar knee pain symptoms?] Z Orthop Ihre Grenzgeb.2003;141: 557-62.
German.141557
2003
[PubMed][CrossRef]
Insall JN, Dorr LD, Scott RD, Scott WN.
Rationale of the Knee Society clinical rating system. Clin Orthop Relat
Res.1989;248:
13-4.24813
1989
Newman JH, Ackroyd CE, Shah NA.
Unicompartmental or total knee replacement? Five-year results of a
prospective, randomised trial of 102 osteoarthritic knees with
unicompartmental arthritis. J Bone Joint Surg Br.1998;80:
862-5.80862
1998
[PubMed][CrossRef]
Cooper H, Hedges L. The handbook
of research synthesis. New York: Russell Sage Foundation;
1994.
1994
Lau J, Ioannidis JP, Schmid CH.
Quantitative synthesis in systematic reviews. Ann Intern Med.1997;127:
820-6.127820
1997
[PubMed]
DerSimonian R, Laird N. Meta-analysis in
clinical trials. Control Clin Trials.1986;7:
177-88.7177
1986
[PubMed][CrossRef]
Sweeting MJ, Sutton AJ, Lambert PC. What
to add to nothing? Use and avoidance of continuity corrections in
meta-analysis of sparse data. Stat Med.2004;23:
1351-75.231351
2004
[PubMed][CrossRef]
Higgins JP, Thompson SG. Quantifying
heterogeneity in a meta-analysis. Stat Med.2002;21:
1539-58.211539
2002
[PubMed][CrossRef]
Higgins JP, Thompson SG. Controlling the
risk of spurious findings from meta-regression. Stat Med.2004;23:
1663-82.231663
2004
[PubMed][CrossRef]
Sterne JA, Egger M. Funnel plots for
detecting bias in meta-analysis: guidelines on choice of axis. J Clin
Epidemiol.2001;54:
1046-55.541046
2001
[CrossRef]
Begg CB, Mazumdar M. Operating
characteristics of a rank correlation test for publication bias.
Biometrics.1994;50:
1088-101.501088
1994
[PubMed][CrossRef]
Ioannidis J, Lau J. Evolution of
treatment effects over time: empirical insight from recursive cumulative
metaanalyses. Proc Natl Acad Sci USA.2001;98:
831-6.98831
2001
[PubMed][CrossRef]
Chan AW, Hrobjartsson A, Haahr MT,
Gotzsche PC, Altman DG. Empirical evidence for selective reporting of outcomes
in randomized trials: comparison of protocols to published articles.
JAMA.2004;291:
2457-65.2912457
2004
[PubMed][CrossRef]
Picetti GD 3rd, McGann WA, Welch RB. The
patellofemoral joint after total knee arthroplasty without patellar
resurfacing. J Bone Joint Surg Am.1990;72:
1379-82.721379
1990
[PubMed]
Shih HN, Shih LY, Wong YC, Hsu RW.
Long-term changes of the nonresurfaced patella after total knee arthroplasty.
J Bone Joint Surg Am.2004;86:
935-9.86935
2004
[PubMed][CrossRef]
Kajino A, Yoshino S, Kameyama S, Kohda
M, Nagashima S. Comparison of the results of bilateral total knee arthroplasty
with and without patellar replacement for rheumatoid arthritis. A follow-up
note. J Bone Joint Surg Am.1997;79:
570-4.79570
1997
[PubMed][CrossRef]
Rand JA. Patellar resurfacing in total
knee arthroplasty. Clin Orthop Relat Res.1990;260:
110-7.260110
1990
[PubMed]
Andriacchi TP, Yoder D, Conley A,
Rosenberg A, Sum J, Galante JO. Patellofemoral design influences function
following total knee arthroplasty. J Arthroplasty.1997;12:
243-9.12243
1997
[PubMed][CrossRef]
Chew JT, Stewart NJ, Hanssen AD, Luo ZP,
Rand JA, An KN. Differences in patellar tracking and knee kinematics among
three different total knee designs. Clin Orthop Relat Res.1997;345:
87-98.34587
1997
[PubMed]
Scuderi GR, Insall JN, Scott NW.
Patellofemoral pain after total knee arthroplasty. J Am Acad Orthop
Surg.1994;2:
239-46.2239
1994
Rapuri V, Parvizi J, Saleh K, Sharkey P,
Hozack W, Mont M. Patellar resurfacing during total knee arthroplasty: a
meta-analysis. J Arthroplasty.2004;19:
262.19262
2004
[CrossRef]
Zangger P, Detsky A. Computer-assisted
decision analysis in orthopedics: resurfacing the patella in total knee
arthroplasty as an example. J Arthroplasty.2000;15:
283-8.15283
2000
[PubMed][CrossRef]