Original Article

Outcomes of Posterior-Stabilized Compared

with Cruciate-Retaining Total Knee Arthroplasty
Umile Giuseppe Longo, MD, MSc, PhD1 Mauro Ciuffreda, MD1 Nicholas Mannering, MD2
Valerio D’Andrea, MD1 Joel Locher, MD1 Giuseppe Salvatore, MD1 Vincenzo Denaro, MD1

1 Department of Orthopaedic and Traumatology, Campus Bio Medico Address for correspondence Mauro Ciuffreda, MD, Department of
University, Rome, Italy Orthopaedic and Traumatology, Campus Bio Medico University,
2 Department of Orthopaedic, Melbourne Institute, University of Via Alvaro del Portillo 21, Rome 00155, Italy
Melbourne, Victoria, Australia (e-mail: maurociuffreda@gmail.com).

J Knee Surg

Abstract The aim of this systematic review is to compare clinical outcome scores, rate of
complications, and range of motion (ROM) of posterior-stabilized (PS) and cruciate-
retaining (CR) total knee arthroplasties (TKAs) both pre- and postoperatively to
establish which of the two kinds of implants have the best efficiency. A comprehensive

Downloaded by: University of Michigan. Copyrighted material.

search was performed of studies comparing CR and PS TKAs on PubMed, OVID/
Medline, Cochrane, CINAHL, Google scholar, and Embase databases. Finally, 37 studies
were selected with a total of 5,407 knees in 4,445 patients. For the PS knees, the Knee
Society functional score (KSFS) improved from 44.6 to 77.6 (p ¼ 0.04), extension
decreased from 6.6 to 1.8 degrees (p-value), and flexion increased from 115.2 to 119.4
degrees (p < 0.00001), compared with the CR knees. No significant difference in the
Knee Society objective score (KSOS) (p ¼ 0.82) or complication rates (p ¼ 0.29) was
found. The overall complication rate was 3.9%, 213 in 5,407 knees. Surgeons must be
careful in interpreting these results, as an improved ROM for the PS group may not
correlate to better patient outcomes. This meta-analysis has demonstrated that PS TKA
has a statistically significant greater postoperative improvement of KSFS (p ¼ 0.04),
flexion (p < 0.00001), and extension (p ¼ 0.02), compared with the CR group. These
findings seem to lead the surgeons to prefer the PS design for TKAs especially to
achieve a higher postoperative ROM in patients with high functional demands. On the
contrary, the CR and PS TKAs have similar results in terms of complications and most
Keywords of clinical outcomes analyzed in the included studies. Therefore, the long-term follow-
► knee up of high-quality randomized controlled trials is needed to clarify which of the two
► arthroplasty types of prosthesis provide the better clinical outcome and the lower rate of
► cruciate-retaining complications for osteoarthritis patients in particular cohorts. This is a systematic
► posterior-stabilized review (level II).

Total joint replacement surgery is an important option in
patients with severe symptomatic osteoarthritis (OA) who
have significant impairment in their quality of life and who
have failed to respond to nonpharmacological and pharma-
cological management.1 Total knee arthroplasty (TKA) is the
most common definitive surgical procedure for end-stage
OA. It provides marked pain relief and functional improve-
ment in patients with severe knee OA.2–4
TKA was born in 1974, and this field has been in contin-
uous evolution yet. However, one of the most important

received Copyright © by Thieme Medical DOI https://doi.org/

September 10, 2016 Publishers, Inc., 333 Seventh Avenue, 10.1055/s-0037-1603902.
accepted after revision New York, NY 10001, USA. ISSN 1538-8506.
May 2, 2017 Tel: +1(212) 584-4662.
Outcomes of PS Compared with CR TKA Longo et al.
challenges of TKA is the maintenance or the removal of the
posterior cruciate ligament (PCL). Both posterior cruciate-
retaining (CR) and posterior-stabilized (PS) TKAs have shown
to provide valid pain relief and similar outcomes evaluated
with the main clinical scores (Knee Society score, Western
Ontario and McMaster Universities Arthritis Index [WOMAC]
score, Knee Society functional score [KSFS], Hospital for
Special Surgery [HSS], and Oxford Knee Score [OKS]).1 The
main difference between the two surgical procedures was
previously shown to be postoperative knee range of motion
(ROM), although the results varied depending on the
study.5–11
In the physiological knee, the PCL has different kinematic
functions. During flexion, it guides rollback of the femoral
condyles on the tibial plateau, prevents posterior subluxa-
tion of the tibia on the femur in flexion, and plays a decisive
secondary role in varus/valgus stability. The retention of the
PCL provides possible kinematic and proprioceptive benefits
such as the maintenance of a central stabilizer and joint line
position, and the avoidance of the patellar clunk syndrome
and of the stress inherent in posterior cruciate substituting
knees.12 Moreover, the PCL can absorb shear forces that
otherwise would be transmitted to the bone–implant inter-
face which could lead to premature loosening.1
Advocates of PS TKA instead support that substituting the
PCL with a post and cam improves ROM secondary to
mechanical enforcement of femoral rollback. The mechan-
ism behind this is that posterior translation of the femur
creates more clearance over the tibia, and, theoretically,
more flexion and stair-climbing ability.13 Moreover, the
sacrifice of PCL could reduce patellofemoral peak and
mean pressure and produce lower range of axial rotation
and condylar translation. Potential advantages of these de-
signs incorporate a less technically demanding procedure, a
more stable component interface (with no anterior transla-
tion under weight-bearing conditions and femoral rollback
with passive flexion) and increased ROM.14
Finally, the choice of either removal or retention of the
implant also depends on the degenerative status of the PCL,
the type of prosthesis, and the surgeon’s experience. Due to
these inconsistencies, there is no clear evidence on the
efficiency of PS or CR implants.
Therefore, the aim of this systematic review is to compare
clinical outcome scores, rate of complications, and ROM of PS
and CR TKAs both pre- and postoperatively to establish
which of the two kinds of implants have the best efficiency.
Materials and Methods
A systematic review of the literature was performed accord-
ing to the PRISMA (Preferred Reporting Items for Systematic
Reviews and Meta-analyses) guidelines with a PRISMA
checklist and algorithm.15 The search algorithm according
to the PRISMA guidelines is shown in ►Fig. 1. A comprehen-
sive search of PubMed, Medline, CINAHL, Cochrane, Embase,
and Google Scholar databases was performed using combi-
nations of the keywords “Knee,” “Arthroplasty,” “Cruciate,”
“Retaining,” “Posterior,” “Stabilized,” “Treatment,” “Clinical,”
The Journal of Knee Surgery
Idenficaon
Records identified through
database searching
(n = 142)
Records after duplicates removed
(n = 119)
Screening
Records screened Records excluded
(n = 111) (n = 64)
Eligibility
Full-text articles Full-text articles
excluded, with
assessed for
reasons (n = 10)
eligibility (n = 47)
Included
Studies included in
Downloaded by: University of Michigan. Copyrighted material.
quantitative synthesis
(n = 37)
Fig. 1 PRISMA 2009 flow diagram.
and “Outcome” since inception of databases to February 25,
2016. Two independent investigators (U.G.L. and M.C.)
separately conducted the search. For inclusion in the
study, articles had to be published in a peer-reviewed
journal. All articles were manually screened for relevance.
The two investigators (U.G.L. and M.C.) separately reviewed
the abstract of each publication and then performed close
reading of all articles and extracted data, to minimize selec-
tion bias. The last search was performed on February 25,
2016.
According to the Oxford Centre for Evidence-Based Med-
icine, levels I to IV articles were found in the literature and
included in our study. Given the linguistic capabilities of the
authors, articles in English, French, Spanish, German, or
Italian were included. Articles that reported clinical out-
come, including ROM, after CR or PS TKA were included.
Articles were excluded if there was missing data pertinent to
these parameters. Literature reviews, animal studies cada-
veric or in vitro studies, biomechanical reports, technical
notes, letters to editors, and instructional courses were
excluded.
Finally, to avoid bias, both the selected articles and
excluded articles were reviewed, assessed, and discussed
by all the authors. All the investigators independently ex-
tracted the following data: demographics, bone lesions be-
fore surgery, gender, body mass index (BMI), time of follow-
up, type of prosthesis, type of surgery, outcome measure-
ment, ROM, and complications.
Statistical Analysis
Descriptive statistics were used to summarize the clinical
outcomes, ROM, and complications of the included studies.

Categorical variable data were reported as frequency with
percentage. Continuous variable data were reported as mean
value, with the range between minimum and maximum
values. In all studies, p < 0.05 was considered statistically
significant.
Moreover, a meta-analysis was performed of all rando-
mized level I comparative studies that investigated the two
types of prosthesis: CR and PS TKAs in terms of Knee Society
objective score (KSOS),16 KSFS,16 complications, extension,
and flexion after surgery. Review Manager (RevMan, version
5 for Windows; Cochrane Informatics and Knowledge Man-
agement Department) was used to calculate the magnitude
of treatment effect. The test used for meta-analysis to stratify
or match the categorical data is the Mantel–Haenszel’s test. It
allows an investigator to test the association between a
binary predictor or treatment and a binary outcome such
as case or control status while taking into account the
stratification. I2 was calculated as a measure of heterogene-
ity for the main analysis. An I2 value represents the percen-
tage of total variation across studies caused by heterogeneity
rather than by chance. It was considered a low I2 value to be
25% or lower and a high I2 value to be 75% or higher.
Quality Assessment
To assess the quality of the evidence in the included studies
and to evaluate the strength of recommendation of the
intervention that is proposed in the published article, the
Grading of Recommendations Assessment, Development and
Evaluation (GRADE) was used. The GRADE is used to establish
the quality of the evidence through four factors: study de-
sign, study quality, consistency, and directness. The combi-
nation of these four factors determines the quality of
strength of recommendation that is given through a quali-
tative assessment of the evidence: high quality, moderate
quality, low quality, and very low quality.
Results
The literature search and cross-referencing resulted in a
total of 142 articles (►Fig. 1); 23 duplicates had been
identified and removed, leaving 119 articles. Total 111
articles were screened, of which 64 were rejected because
of off-topic abstracts (22 articles), failure to fulfill the
inclusion criteria (14 articles), or both (28 articles). After
reading the remaining 47 full-text articles, another 10
articles were excluded because of insufficient details and
uncertain diagnosis and outcome measures. Finally, 37
articles on the treatment of patients with primary OA
were included.2,3,5–9,14,17–46
Demographics
A total of 5,407 knees in 4,445 patients were
included,2,3,5–9,14,17–46 with a median age at surgery of
69.1 years for the CR group and 69.2 years for the PS group,
ranging from 54.544 to 778 years (►Table 1). The average BMI
was 29.2 and 29.0 kg/m2 for CR and PS, respectively. Patients
were assessed at follow-up for an average of 41.0 months,
ranging from 66 to 147 months.37 There were 2,591 (47.9%)
Outcomes of PS Compared with CR TKA Longo et al.
TKAs completed using the CR method and 2,816 (52.1%)
completed using PS method.
Outcome Measurements
Several outcome measures were reported in the included
studies (►Tables 2 and 3). The most frequently reported
outcomes were the KSOS and the ROM, both used in 32
(86.5%) of 37 studies. Other less consistently reported scor-
ing systems were the KSFS, used in 21 (56.8%) of 37studies;
the WOMAC used in 12 (32.4%) of 37 studies; the HSS score
used in 7 (18.9%) of 37 studies; and the OKS used in 4 (10.8%)
of 37 studies.
Outcome Scores
The mean KSOS improved from 45.2 (preoperatively) to 90.3
(postoperatively) for the CR group. Similar results were
obtained in the PS group, showing a mean KSOS increase
from 43.7 (preoperatively) to 90.8 (postoperatively). The
preoperative and postoperative mean values of KSFS were
45.6 and 76.6, respectively, for the CR group, and were 44.6
Downloaded by: University of Michigan. Copyrighted material.
and 77.6, respectively, for the PS group. A meta-analysis
was performed including the randomized level I com-
parative studies that investigated the postoperative
KSOS3,6,9,20,26,29,34,40,45 (►Fig. 2) and KSFS3,6,9,20,22,29,40
(►Fig. 3) after CR and PS procedures. No statistical differ-
ence was found between the two techniques in case of KSOS
(95% confidence interval [CI], 1.24 to 1.57; p ¼ 0.82).
Instead, in the case of KSFS, PS highlighted a higher im-
provement of this clinical score than CR (95% CI, 5.36
to 0.18; p ¼ 0.04). A low heterogeneity (I2 ¼ 24% for
KSOS and I2 ¼ 15% for KSFS) across the study results was
found in both cases (►Figs. 2 and 3).
Range of Motion and Complications
ROM was evaluated in 32
studies.3,7,9,19–22,24,25,27,28,30,33,43,46 The mean preoperative
and postoperative flexions in the CR group had a mean 0.6
degrees decrease (115–114.4 degrees), and the PS group had
a mean 4.3 degrees increase in flexion (115.2 and 119.4
degrees). The meta-analysis performed on the postoperative
ROM highlighted a higher improvement of both flexion (95%
CI, 5.81 to 3.60; p < 0.00001)6,9,20,22,29,38,40,45 and ex-
tension (95% CI, 1.05 to 0.09; p ¼ 0.02)3,22 for PS proce-
dures compared with CR procedures. A high heterogeneity
(I2 ¼ 52% for flexion and I2 ¼ 83% for extension) across the
study results was found in both cases (►Figs. 4 and 5).
Complications were reported in only 17 of the 37 studies
(45.9%).3,5–7,14,19,20,22,28,30,32,33,35–37,44,45 There was an
overall complication rate of 3.9% (213 complications in
5,407 knees). The meta-analysis3,6,20,22,29,35,45 showed
that no statistical difference was found between the two
procedures (95% CI, 0.78–2.31; p ¼ 0.29). A low heterogene-
ity (I2 ¼ 0%) across the study results was found (►Fig. 6).
Quality Assessment
The quality of the evidence of managing arthrosis of knee
with the two types of prosthesis, CR and PS, was found to be
high (►Fig. 7).
The Journal of Knee Surgery
 Table 1 Demographics

Authors Study design (level) Number of Number Number Number Mean age, y (range) BMI (kg/m2) Sex M:F Bone defects Mean Prosthesis type
patients of knees of CR of PS follow-up
CR PS
(y)

Akasaki et al Level III (retrospec- 15 20 8 12 72 (56–84) – CR(3:5); OA 5.1 6 Kirschner Medi-

The Journal of Knee Surgery

(2009)17 tive cohort study) PS (2:10) (3.0–7.9) cal Co., 2 Encore
Foundation Sys-
tem, 10 Nexgen
LPS Zimmer, 2
Scorpio Superflex
Striker

Ang et al (2014) 18 Level III 95 100 37 63 CR: 67.7 PS: 67.6 CR: 26.2 CR (7:30); OA, Type II 2 –
Outcomes of PS Compared with CR TKA

(retrospective) (49.9–84.2) (45.4–82.5) (18.8–32.7) PS (7:56) valgus knees

PS: 27.3
(19.1–43.8)

Bin Abd Razak et al Level III 195 195 112 83 CR 66 PS 67 CR: 28.4 (3.6) CR OA 2 PFC Depuy
(2013)19 (50–84) (54–40) PS: 26.9 (3.6) (22:90); PS Orthopedic
Longo et al.

(18:65) International

Carvalho et al Level II (prospective 31 38 14 24 – – 5:26 OA 2.55 NexGen (Zimmer)

(2014)25 study)

Catani et al (2004)20 Level I (RCT) 40 40 20 20 CR: 70  6 PS: 71  7 – CR (7:13); OA 2 OPTETRAK

(60–82) (48–80) PS (5:15) Exactech

Cates et al (2008)21 Level III (retrospec- 30 30 15 15 CR: PS: CR: 27.4  3.6 CR (5:10); OA, RA, AVN 0.5 NexGen CR-Flex
tive study) 67.8  7.1 66.9  10.5 PS: 28.6  4.1 PS (5:10) high-flexion (Zim-
mer) or Legacy
LPS-Flex high-flex-
ion (Zimmer) by a
single surgeon
(Cates)

Chaudhary et al Level I (RCT, double 78 78 40 38 CR: 69.2 PS: 70.2 CR: 32.4 (5.7) CR OA 2 SCORPIO Stryker
(2008)22 blind) (9.1) (8.4) PS: 30.9 (4.3) (27
females);
PS (22
females)
5
Chen et al (2015) Level III 133 133 33 100 CR: 65  8 PS: 64  7 CR: 28.2  5 CR (5:28); OA 2 67 NexGen, 48
(51–80) (51–81) (19.6–38.8) PS (27:73) PFC, 4 Scorpio, 6

Downloaded by: University of Michigan. Copyrighted material.

 Table 1 (Continued)

Authors Study design (level) Number of Number Number Number Mean age, y (range) BMI (kg/m2) Sex M:F Bone defects Mean Prosthesis type
patients of knees of CR of PS follow-up
CR PS
(y)

PS: 29  5 Genesis II, 8

(19.4–44.4) Columbus

Clark et al (2001)23 Level I (RCT) 128 128 69 59 CR: 57–8,9 PS: 71.2 CR: 83.3 (31.6) – OA, PTA, CR: 2; PS: 3 AMK DePuy/John-
71.8 (12.2) (13.6) PS: 82.1 (38.4) psoriasis son & Johnson
(resurfacing of
the patella was
routine)

Conditt et al Level III (retrospec- 49 49 28 21 CR: 72  3 PS: 69  3 – CR OA 1 AMK DePuy

(2004)24 tive study) (45%:55%);
PS
(40%:60%)
2
Delport (2013) Prospective study 943 1152 561 591 Women, 71 (range: 35–92); 27 242:701 OA 15 –
Level II men, 69.7 (range: 31–96)

Fantozzi et al Level I (randomized 23 23 10 13 CR: 73.9 PS: 67.7 CR: 29.3 CR (2:8); OA GRADE: CR CR 3.9 Optetrak
(2006)26 prospective) (65–82) (50–81) PS: 28.1 PS (2:11) 1.8 (1–3)/PS (1.5–5.75), Exactech
2.5 (1–3) PS 1.85
(0.6–4.3)
27
Hamai et al (2015) Level III 19 24 12 12 CR: 70  9 PS: 75  9 CR: 26.5  3.6 – OA, RA CR: 2, PS: CR TKA (Founda-
PS: 24.9  3.0 2.75 tion knee, Encore
Medical Co.), PS
TKA (NexGen
Complete Knee
Solution Legacy
PS, Knee, Zimmer
Inc.)
28
Han et al (2012) Retrospective cohort 127 186 92 94 CR: PS: CR: 27.5  3.4 – OA 2 NexGen Zimmer
study Level III 67.5  5.7 68.0  5.8 PS: 27.4  4.0
Outcomes of PS Compared with CR TKA

Harato et al (2008)3 Level I (RCT) 189 192 99 93 CR: 68.3 PS: 66.0 CR: 29.8 CR OA 5.5 Genesis II Smith &
(49–89) (44–83) (19.7–43.6); PS: (34:65); PS (5.0–7.3) Nephew
31.4 (21.7–48.5) (32:61)

Kim et al (2009) 29 Level I (randomized 250 500 250 250 CR: PS: 71.6  6 CR: 26.8  3.2; 10:240 OA 2.3 Zimmer
prospective) 71.6  6 PS: 26.8  3.2
Longo et al.

The Journal of Knee Surgery

(Continued)

Downloaded by: University of Michigan. Copyrighted material.

 Table 1 (Continued)

Authors Study design (level) Number of Number Number Number Mean age, y (range) BMI (kg/m2) Sex M:F Bone defects Mean Prosthesis type
patients of knees of CR of PS follow-up
CR PS
(y)
14
Kolisek et al (2009) Level II (prospective 91 91 46 45 CR: 64 PS: 66 (45– CR: 32 (26–43); CR OA, ON, RA 5 (4.5–5.75) SCORPIO Stryker

The Journal of Knee Surgery

study) (40–77) 81) PS: 32 (23–47) (20:26); PS
(17:28)

Lee et al (2012) 30 Retrospective study 131 179 45 134 CR: 65.7 PS: 67.2 CR: 27  2.5; PS: CR (2:43); OA CR: 8.1 57 knees with the
Level III 27  3.3; F-PS: PS (3:37); (5–9.7), PS: Scorpio CR im-
28  4.2 F-PS (7:87) 8.0 (5–12) plants, 50 knees
with the Scorpio
PS implants, and
118 knees with
Outcomes of PS Compared with CR TKA

the Scorpio Flex

PS (F-PS) implants

Liu et al (2015)32 Level II 32 64 32 32 65.6 (59–78) – 12:20 OA 2 Genesis II Smith &

Nephew
Longo et al.

Lützner et al Level II 39 78 39 39 68.9 (10.2) 32.8 (5.9) 31 Females OA 1 –

(2015)33 (79.5%)

Maruyama et al Level III (prospective 20 40 20 20 74.3 (65–84) – 8:12 OA 2.6 (2–4.4) PFC DePuy John-
(2004)7 study) son & Johnson

Matsumoto et al Level I (randomized 41 41 19 22 CR: PS: – – Varus-type OA 5 NexGen CR Flex

(2012)34 prospective) 73.5  1.3 74.4  0.9 patients (Zimmer, Inc.);
NexGen LPS Flex
(Zimmer)
35
Misra et al (2003) Level I (RCT) 103 105 51 54 CR: 66.8 PS: 67.2 – CR OA (46 CR and 4.75 (4.6–5) PFC DePuy/John-
(55–83) (59–82) (17:34); PS 51 PS), RA (5 son & Johnson
(22:32) CR and 3 PS)

Mouttet and Sourdet Level IV (compara- 104 114 71 43 CR: 76  5 PS: 77  7 CR: 28  4 CR OA: (63CR/ 4.5 EUROP Euros SAS
(2014)8 tive study) (57–87) (57–91) (18–40); PS: (19:47); PS 35PS), PTA:
28  3 (21–36) (11:29) (3CR/4PS),
ON (5 CR/
3PS), RA:
(1PS)

Peters et al (2014) 36 Level III 382 468 240 228 CR: 62  10 PS: 62  10 CR: 33.3  8.8 163:305 CR: OA 215 2 –
(23–86) (23–82) (17.7–79.5); PS: (90%);AS: 201
31.7  7.9 OA (88%)
(17.7–87.9)

Downloaded by: University of Michigan. Copyrighted material.

 Table 1 (Continued)

Authors Study design (level) Number of Number Number Number Mean age, y (range) BMI (kg/m2) Sex M:F Bone defects Mean Prosthesis type
patients of knees of CR of PS follow-up
CR PS
(y)
37
Sando et al (2015) Level III 360 414 143 271 CR: PS: CR: 31.8  4.8; CR OA 12.3 Genesis Smith &
(prospective) 68.4  8.0 69.3  9.08 PS: 33.0  6.8 (48:95); PS (10.2–14.4) Nephew
(121:150)

Seon et al (2011) 38 Level I (Prospective) 95 95 48 47 CR: PS: CR: 25.8  3.4; CR (4:44); OA 2.25 Zimmer
68.2  7 69.2  6.7 PS: 23.7  2.8 PS (5:42)

Snider and Level I (randomized 200 200 100 100 – – – – OA 2 AMK CR and PS
Macdonald (2009)39 prospective) DePuy, Genesis II
CR and PS Smith &
Nephew

Tanzer et al (2002) 40 Level I (RCT, double 37 40 20 20 CR: 68 PS: 66 CR: 30.7; PS: 30 CR (5:15); OA (36), RA 2 NexGen Zimmer
blind) (range, 51– (range, 52 PS (4:16) (1CR, 2PS), AN (CR) and Legacy
86) -57) (1 PS) Zimmer (PS)

Thomsen et al Level I (RCT) 33 66 33 33 67.2 29.4 (19–41) 14:19 OA 1 Cemented CR

(2013)41 AGCTKA (Biomet-
Merck) in one
knee and cemen-
ted LPS-Flex TKA
(Zimmer) in other
knee

Tsuneizumi et al Level II 40 44 22 22 CR: PS: CR: 26.9  2.4 Female:40 OA 2 –

(2008)42 72.5  7.3 72.5  3.9 PS: 27.3  4.

van den Boom et al Level II (prospective) 21 21 9 12 CR: 72 (SD 8) PS: 75 (SD 6) < 35 CR (7:2); OA 0.75 –
(2014)43 PS (5:7)

Vermesan et al Level I (Randomized) 50 50 25 25 CR: PS: CR: 32.6  7.1 CR OA, RA 0.5 Biomet, Zimmer
(2015)6 68.8  6.9 68.4  6.3 PS: 33.4  7.5 (15:10); PS
(22:3)

Victor et al (2005)9 Level I (RCT) 44 44 22 22 CR: 70  7 PS: 70  3 CR: 34.4 CR: (5:17); OA 5 Genesis II Smith &
Outcomes of PS Compared with CR TKA

PS: 32.8 PS: (4:18) Nephew

Wang et al (2004) 44 Level II (prospective 228 267 157 110 CR: 54.5 PS: 55 (20– CR: 28.1 CR (CR) OA 152, 3.5 (2–5.5) PFC Johnson &
clinical study) (31–69) 83) PS: 27.2 (27:110); RA 3, ON 2, Johnson
PS (18:73) OA 91 (PS)
Longo et al.

The Journal of Knee Surgery

RA17, ON 2

(Continued)

Downloaded by: University of Michigan. Copyrighted material.

Outcomes of PS Compared with CR TKA Longo et al.

Discussion

stabilized; GRADE, Grading of Recommendations Assessment, Development and Evaluation; LPS, legacy knee posterior stabilized; OA, osteoarthritis; ON, osteonecrosis; PFC, press-fit condylar; PS, posterior
Abbreviations: AGC, anatomical graded component; AMK, anatomic modular knee; AS, anterior stabilized; AVN, avascular necrosis; BMI, body mass index; CR, cruciate retaining; F-PS, high flexion posterior
PCF DePuy John-
Prosthesis type

son & Johnson

The most important results of this systematic review is that a
statistical difference was found in terms of postoperative

Zimmer
improvement in KSFS (p ¼ 0.04), flexion (p < 0.00001), and

–
extension (p ¼ 0.02), which show greater improvements for
the PS group compared with the CR group. This finding is
follow-up

3.4 (41.0
consistent with the literature, which unanimously reports
1.5–4.4 better ROM after PS TKAs.5–7,9,46
Mean

mo)
(y)

On the contrary, there was no significant difference

between the CR and PS methods in terms of postoperative
Bone defects

complications (p ¼ 0.29) or the KSOS (p ¼ 0.82). This raises

the possibility that a statistical difference in one outcome
measure may not translate to a clinical difference overall.
OA

OA

Since the KSOS and the KSFS scores are subjective measures,

stabilized; PTA, posttraumatic arthritis; RA, rheumatoid arthritis; RCT, randomized controlled trial; SD, standard deviation; TKA, total knee arthroplasty.
there is need for more objective tools that assess patient
Sex M:F

function which can potentially delineate a reproducible

4:25

6:12

difference.
–

The patient demographics of this systematic review for

CR: 26.3  3; PS:

CR: 29.2 (kg/m )

PS: 29.0 (kg/m2)
2

the CR and PS groups are very similar. The similar usage of

Downloaded by: University of Michigan. Copyrighted material.

BMI (kg/m2)

CR method (47.9%) compared with the PS method (52.1%),

26.3  3

in conjunction with the statistically similar KSOS and

complication rates, further gives weight to the supposition
–

that the choice of constraint depends mainly on individual

patient factors in addition to surgeon preference. A study on
74.3  7.2

PS: 69.2 y

the trends of TKAs for treatment of OA reported that in

Mean age, y (range)

2012, 38% of TKA procedures were CR, having lowered from

PS:
PS

50% in 2003.47 This was compared with 53% using a PS

method, up from 31% in 2001. Despite multiple studies
73.8 (62–84)
74.3  7.2

CR: 69.1 y

demonstrate that neither method has superiority, this in-

crease in the use of the PS method is postulated to have
CR:
CR

been multifactorial.
There may be, however, particular primary TKA patient
Number

2,816

cohorts to which either the CR or PS method would be more

of PS

76.1
29

20

amenable. As CR TKAs have been found to maintain greater

stability due to retention of the cruciate ligament, this
Number

technique would theoretically be most suited to patients

2,591
of CR

70.0
29

20

who are at greater risk of falls. Although OA and osteoporosis

are seen in different patient cohorts, the unlucky patient
of knees
Number

with both will have a higher risk of fracture if the technique

146.1

5,407

proven to be less stable is employed.

58

40

Various studies have demonstrated increased femoral

Number of

rollback associated with the PS method.25,35 Although this

patients

is thought to lead directly to greater instability, this causative

120.1

4,445
29

20

relationship has not been objectively demonstrated in vivo.

Furthermore, Lützner et al33 found similar intraoperative
Study design (level)

Level II (prospective
Level I (Prospective

stability between the two constructs, with no statistical

difference. Despite the difference in femoral rollback, both
randomized)

PS and CR TKAs have been proven to provide a similar

improvement in stability both intraoperatively and for out-
study)

Mean

Total

come scores postoperatively.

In the search for better perioperative surgical manage-
Table 1 (Continued)

ment of TKAs, studies have compared the length of surgical

time for PS compared with CR TKAs. It was found that the CR
Yagishita et al

Yoshiya et al

implant required significantly less time (p ¼ 0.0037).6 It was

(2012)45

(2005)46
Authors

postulated that this would consequently lead to reduced

perioperative blood loss with the CR implant; however, this
was not found to be the case.

The Journal of Knee Surgery

 Table 2 Outcomes and complications

Authors Scoring systems Outcomes score Complications

WOMAC  SD OKS Others  SD CR PS

CR PS CR PS CR PS

Akasaki et al KSS preop and postop, – – – – – – – –

(2009) 17 pain, ROM, stability

Ang et al (2014)18 SF-36, KSS, OKS – – Preop 41 (26–49); Preop 38 (20–58); – – – –

postop 19 (12–38) postop 19 (12–38)

Bin Abd Razak et al KSS (knee and function – – Preop 34.4  8.0; Preop 34.4  9.0; SF-36 preop SF-36 preop – –
(2013) 19 score), OKS, SF-36, ROM 2 y postop 2 y postop 33.7  10.9; 2 y 33.4  10.9; 2 y
18.7  4.7 19.3  5.3 postop 47.3  9.5 postop
46.9  10.2

Carvalho et al ROM, femoral rollback – – Femoral rollback Femoral rollback – –

(2014) 25 (5.8  6.5 mm (13.2  10.5 mm
[ 5.6 to 18.2 [ 8.3 to 27.2
mm]) mm])

Catani et al RSA, KSS, HSS, ROM – – – – Postop HSS: Postop HSS: 1 lateral release 2 lateral release
(2004) 20 86  8 89  7 and patella resur- and patella resur-
facing for anterior facing for anterior
knee pain knee pain

Cates et al HSS, KSS, KSFS, ROM – – – – Preop HSS: Preop HSS: – –

(2008) 21 59.6  6.4; 55.4  7.0;
postop HSS: postop HSS:
93.3  3.3 92.8  3.6

Chaudhary et al ROM, KSFS, WOMAC, pain Pain score (preop: Pain score (preop: – – – – 1 deep infection 1 stiff knee requir-
(2008) 22 score, complication 51.0  17.9, 47.4  16.7, requiring removal ing manipulation
postop: postop: of hardware
85.1  17.6); 83.3  16.7);
functional score functional score
(preop: (preop:
51.6  5.6, 55.8  16.3,
Outcomes of PS Compared with CR TKA

postop: postop:
76.5  0.9) 77.8  17.4)

5
Chen et al (2015) KSFS, KKS, OKS, ROM – – Preop 38 (28.43); Preop 36 (30.43); – – – –
2 y postop 17 2 y 18 (16.23)
Longo et al.

The Journal of Knee Surgery

(16.23)

(Continued)

Downloaded by: University of Michigan. Copyrighted material.

 Table 2 (Continued)

Authors Scoring systems Outcomes score Complications

WOMAC  SD OKS Others  SD CR PS

CR PS CR PS CR PS

Clark et al (2001) 23 KSS, ROM, SF-12, WOMAC Preop Preop – – – – – –

The Journal of Knee Surgery

(47.2  31.0), (50.4  33.8),
postop postop
(18.5  32.9); (22.8  35.4)

Conditt et al SF-36,KSS, ROM – – – – – – – –

(2004) 24

Delport (2013) 2 Clinical KSS, KSFS – – – – – – – –

Fantozzi et al KSS, KSFS, ROM – – – – – – – –

Outcomes of PS Compared with CR TKA

(2006) 26

Hamai et al KSS,KSFS, ROM, tibial – – – – – – – –

(2015) 27 posterior slope (deg)

Han et al (2012) 28 KSS, HSS, WOMAC, ROM, Preop 51.5  6.2; Preop 52.3  7.7; – – Flexion contrac- Flexion contrac- 2 anteroposterior 1 mediolateral
Longo et al.

complication 2 y postop 2 y postop ture preop 3.2 ture preop 3.6 instability (less laxity (less than 10
9.2  91 11.9  9.6 deg  5.4; 2 y deg  4.9; 2 y than 10 mm) deg)
postop 0.2 postop 0.2
deg  1.5 deg  1.1

Harato et al ROM, KSS, KSFS, WOMAC, Preop Preop – – SF-12 (m) SF-12 (m) 1 infection, 7 stiff 3 infection, 1 stiff
(2008) 3 pain score, complication 26.4  16.3, 23.2  11.9, 54.0  10.9; SF-12 54.7  12.1; SF-12 knee, 2 hemar- knee, 1 hemar-
postop postop 8.5  12.3 (f) 30.3  7.1 (f) 30.4  7.8 throsis, 5 knee throsis, 2 knee
10.4  13.4 pain, 1 lucent line, pain, 1 lucent line,
0 DVT 1 DVT

29
Kim et al (2009) KSS, HSS, WOMAC, flex- 5.5  3.77 4.9  2.98 – – Postop HSS: Postop HSS: 2 anterior femoral 3 anterior femoral
ion, radiographs, 90  19.7 91  16.5 notching, 1 super- notching, 1 super-
complications ficial wound ficial wound
infection infection

Kolisek et al KSS, KSFS, ROM – – – – – – – –

(2009) 14

Lee et al (2012)30 KSS, HSS, WOMAC Preop: 59  15.4; Preop: 53  16.1; – – – – 1 aseptic loosening 2 aseptic loosening
postop: 18  12.3 postop: 15  9.1

Liu et al (2015) 32 – –

Downloaded by: University of Michigan. Copyrighted material.

 Table 2 (Continued)

Authors Scoring systems Outcomes score Complications

WOMAC  SD OKS Others  SD CR PS

CR PS CR PS CR PS

WOMAC, KSS, ROM, Only graph Only graph Mediolateral laxity Mediolateral laxity 1 necrosis, 0 infec- 2 necrosis, 0 infec-
Kellgren–Lawrence displayed displayed preop 13.2, postop preop 13.4, postop tion, 1 hemarthro- tion, 1 hemarthro-
grading 13.3; anteropos- 13.8; anteropos- sis, 1 DVT, 1 lucent sis, 0 DVT, 1 lucent
terior laxity preop terior laxity preop line, 2 anterior line, 3 anterior
9.1, postop 8.9; 8.9, postop 7.7; knee pain, 2 stiff knee pain, 0 stiff
valgus preop 3.7, valgus preop 3.5 knee knee
postop 6.5 postop 6.3

Lützner et al KSS, OKS, UCLA activity – – 19.0–30.0 at 3 mo/ – UCLA (maximum level 10) demonstrated 2 delayed wound healing. 1 mobilization
(2015) 33 score, ROM 19.0–34.4 at 1 y minor improvement in self-assessed activ- under anesthesia was necessary due to
ity from median level 3 preoperatively to restricted ROM at the 3-mo follow-up. DVT
level 4 postop (1 y) (n ¼ 4), myocardial infarction (n ¼ 1),
implantation of a cardiac pacemaker
(n ¼ 1), postoperative delirium (n ¼ 1),
and gastrointestinal problems (n ¼ 6)

Maruyama et al KSS, ROM, joint line – – – – – – None 1 superficial

(2004) 7 wound infection

Matsumoto et al KSS, KSFS, ROM – – – – – – – –

(2012) 34

Misra et al (2003) 35 HSS, ROM, satisfaction – – – – HSS: preop (36.2), HSS: preop (38.6), 3 instability, 1 in- 3 instability, 3
rollback loosening postop (81.4) postop (83.6) fection, 2 aseptic aseptic loosening,
loosening, 2 2 stiffness
stiffness

Mouttet and KSS (functional, walking, – – – – – – – –

Sourdet (2014)8 stair, knee score, pain,
flexion)

Peters et al KSS (total, clinical and – – – – – – 1 death, 4 manip- 2 deaths, 3 manip-

(2014) 36 functional), complication ulations, 1 reo- ulations, 1 reo-
Outcomes of PS Compared with CR TKA

peration, 21 peration, 7
revision, 7 septic, revision, 5 septic, 2
14 aseptic, 6 in- aseptic, 0 instabil-
stabilities, 3 ities, 1 loosening
loosening
Longo et al.

The Journal of Knee Surgery

(Continued)

Downloaded by: University of Michigan. Copyrighted material.

 Table 2 (Continued)

Authors Scoring systems Outcomes score Complications

WOMAC  SD OKS Others  SD CR PS

CR PS CR PS CR PS

Sando et al KSS (function and knee Function preop Function preop – – – – 2 patella resurfa- 3 patella resurfa-

The Journal of Knee Surgery

(2015) 37 score), WOMAC (pain, 42.2  16.1; 10 y 43.0  14.2, 10 y cing, 1 patellar cing, 1 joint stiff-
stiffness, and function postop postop crepitus,1 joint ness, 1 dislocation,
score), ROM 62.6  26.1 72.5  21.7 stiffness 1 osteolysis

38
Seon et al (2011) HSS, WOMAC, ROM, Postop: Postop: – – HSS: 94.7  4.3 HSS: 93.9  4.7 None detected None detected
Radiographs 28.4  13.8 27.9  12.2

Snider and Macdo- KSS, lateral method, a-p – – – – – – – –

nald (2009) 39 method
Outcomes of PS Compared with CR TKA

Tanzer et al KSS, KSFS, ROM – – – – – – None detected None detected

(2002) 40

Thomsen et al SF-36, ROM, knee pain – – – – – – – –

(2013) 41
Longo et al.

Tsuneizumi et al KSS, KSFS – – – – – – – –

(2008) 42

van den Boom et al KSS, ROM, WOMAC, knee Preop: 53  20; Preop: 46  18; – – – – – –
(2014) 43 moment (nm/kg) postop 15  10 postop: 15  12

Vermesan et al KSS, ROM, complications – – – – – – 3 stiff knees 1 stiff knee, 1

(2015) 6 infection (drained
an antibiotics)

9
Victor et al (2005) KSS, KSFS, ROM, pain Pain (preop Pain (preop – – – – – –
score, WOMAC 8.8  3.2, postop 10.8  3.3,
1.8  3.1); func- postop 4.1  6.2),
tion (preop function (preop
34.3  10.2, 38.5  11.0,
postop 8.7  6.9) postop
2.1  14.9)

Wang et al ROM, KSS, KSFS, pain – – – – – – – –

(2004) 44 score, complications

Yagishita et al KSS, KSFS, ROM, radio- – – – – – – 1 DVT

(2012) 45 graphs, complications

Downloaded by: University of Michigan. Copyrighted material.

 Outcomes of PS Compared with CR TKA Longo et al.

Other studies have also focused on bleeding as a measurable

Abbreviations: CR, cruciate retaining; DVT, deep vein thrombosis; HSS, Hospital for Special Surgery knee score; KSFS, Knee Society functional score; KSS, Knee Society score; OKS, Oxford Knee Score; Postop,
postoperative; Preop, preoperative; PS, posterior stabilized; ROM, range of motion; RSA, Roentgen stereophotogrammetric analysis; SD, standard deviation; SF-36, Short Form health survey; UCLA, University of
Overall complications: 213 in 5,407 knees
perioperative outcome. While Vermesan et al (2015)6 and

Number of studies reporting complica-

Cankaya et al48 were not able to find a significant difference
in blood loss, Mähringer-Kunz et al49 reported a statistically
significant difference (p ¼ 0.032). Greater blood loss was
PS

–
demonstrated in the PS group (548  216 mL in the PS group

tions: 17 of 37 (45.9%)
compared with 502  186 mL in the CR group),49 based on
the pre- and postoperative hematocrit levels at discharge in
Complications

240 CR patients and 233 PS patients. The authors proposed

that this could be explained by the additional preparation of

(3.9%)
CR

the femoral box for the cam–post mechanism used during

–

the implant of PS prosthesis.

These studies are, therefore, in favor of the CR implant,
which takes less operative time and causes less blood loss,
compared with the PS prosthesis. However, these studies
demonstrate there are varying results when comparing
PS

blood loss. Furthermore, despite the significant difference

in bleeding shown by Mähringer-Kunz et al,49 this did not
translate into any difference in transfusion requirements
Others  SD

between the two groups (PS 0.41 vs. CR 0.37). Once again,

Downloaded by: University of Michigan. Copyrighted material.

this potentially demonstrates that a statistical difference
may not imply any clinical relevance.
CR

This systematic review has several limitations. Some

studies did not report adequately on complication or revision
rates. Furthermore, many studies reported only the number
of complications in all TKAs and did not specify how many
occurred in either the CR or PS group. Unfortunately, the
PS

literature reporting of complications postoperatively was

–

OKS: 4 of 37 (10.8%)

found to be inadequate for further statistical analysis. There

were also different types and severities of complications,
California at Los Angeles Activity Score; WOMAC, Western Ontario McMasters Osteoarthritis Index.

making the impact on the patient a nonhomogenous para-

meter. However, the inclusion of this parameter was deemed
OKS

important to compare the two methods of PCL management

CR

in TKAs more holistically.

Even though a statistically significant difference in
ROM between PS and CR groups was detected, both the
underreporting of results and the difference in follow-up
times may influence our results. For example, there were
WOMAC: 11 of 37 (32.4%)
PS

differences in the definition of postoperative, most reporting

–

at 2 years5,8,19,23,28,30,32,33,36,38,40,42,43,45 and some at

5 years.2,3,9,14,17 Also, it is difficult to know at what time
Outcomes score

WOMAC  SD

assessment was undertaken postoperatively for those who

did not clarify their methods of follow-up.
The variability of available TKA designs, as well as differ-
CR

ences in surgical technique, such as metal backed or all

polyethylene, is a potential source for confounding. Ideally,
KSS (femorotibial contact

the only variable in our study would be the use of CR or PS,

allowing for simultaneous control of all other variables,
Studies reporting
Scoring systems

position), ROM

including patient anthropometrics and surgical design. How-

ever, this level of control is impossible if large sample sizes
are to be analyzed. Finally, the use of many level III evidence
studies naturally constrains our systematic review to the
Table 2 (Continued)

limitations found within this level of evidence.

It will be important for further research to focus on which
particular populations, and not simply the overall TKA
Yoshiya et al
(2005) 46

treated population, perform better with either the CR or

Authors

PS method. Despite there being no statistically significant

difference in KSOS or complications postoperatively for the

The Journal of Knee Surgery

 Table 3 Mean outcome scores

Authors KSS clinical (SD) KSFS (functional) (SD) ROM (deg) (SD)

CR PS CR PS CR PS

Preop Postop Preop Postop Preop Postop Preop Postop Preop Postop Preop Postop

Akasaki et al 68 85 53 94 – – – – 115 105 132 120

The Journal of Knee Surgery

(2009)17

Ang et al 23 88 (< 0.001)a 30 87 (< 0.001)a 45 60 (< 0.001) a 45 65 (< 0.001)a 132 114 (< 0.001) a 132 126 (0.007)a
18
(2014)

Bin Abd Razak 36.8 (19.8) 86.2 (13.7) 42.2 (19.5) 86.4 (10.3) 55 (18.1) 72.7 (21.4) 56 (19.3) 71.8 (15.5) 118 (19) 114 (15) 121 (18) 122 (10)
et al (2013)19

Carvalho et al – – – – – – – – – 106.43 (9) – 105.4 (11.7)

(2014)25
Outcomes of PS Compared with CR TKA

Catani et al 52 (17) 89 (10) 47 (12) 90 (9) 55 (25) 81 (17) 51 (17) 76 (19) 106 (12) 97 (15) 106 (21) 114 (21)
(2004)20

Cates et al 48.7 (12.3) 97.3 (3.0) 44.1 (12.9) 96.2 (3.0) 31.0 (14.3) 81.0 (24.0) 25.0 (12.2) 92.0 (10.5) 117 (15) 126 (9) 110 (17) 125 (10)
(2008)21
Longo et al.

Chaudhary et – – – – 51.6 (15.6) 76.5 (20.9) 55.8 (16.3) 77.8 (17.4) Flex: 116.4 Flex: 105.9 (13.0), Flex: 111.5 Flex: 105.8
al (2008)22 (13.8), Ext: 6.6 Ext: 1.2 (2.5) (15.3), Ext: 7.2 (13.5), Ext: 2.2
(6.2) (8.4) (3.5)

b b b b b b b b b b b
Chen et al 43 (18, 57) 84 (72, 92) 40 (26, 55) 86 (81, 90) 55 (45, 65) 75 (58, 80) 50 (45, 60) 70 (55, 80) 101 (90, 116) 110 (90, 118) 115 (100, 121) 118 (108, 125)b
5
(2015)

Clark et al – – – – – – – – 110.6 108.5 111.6 113.6

(2001)23

Conditt et al – 94.6 (2.4) – 89.4 (3.7) – – – – 109 (8) 121.4 (1.9) 100 (5) 121.9 (1.8)
(2004)24

Delport 33 96 38 99 42 94 42 97 – – – –
(2013)2

Fantozzi et al – 91.5 – 94.7 – 84 – 84.6 – 103 – 124.1

(2006)26

Hamai et al – 93 (6) – 93 (6) – 78 (15) – 75 (14) – Flex: 119 (11), Ext: – Flex: 125 (7),
(2015)27 1 (2) Ext: 3 (3)

Han et al 51.7 (8.5) 97.9 (3.1) 53.7 (10.0) 97.6 (3.2) – – – – 124.5 (15.3) 131 (10.5) 128.4 (13.4) 132.7 (7.0)
(2012)28

46.7 (16.9) 90.8 (13) 44.3 (17.6) 90.4 (15.7) 47.2 (17.2) 69.6 (19.7) 50.8 (13.7) 74.9 (18.7)

Downloaded by: University of Michigan. Copyrighted material.

 Table 3 (Continued)

Harato et al Flex: 113.3 Flex: 113.7 (12.8), Flex: 110.1 Flex: 117 (13.5),
(2008)3 (11.6), Ext: 5.6 Ext: 0.8 (2.1) (13.6), Ext: 5.3 Ext: 1.6 (1.5)
(5.7) (6.4)

Kim et al – 94 (14.5) – 95 (15.3) – 80.2 (19.7) – 83.7 (19.6) – 124.3 (9.2) – 129 (5.2)
(2009)29

Kolisek et al 42 (20, 73)c 93 (55, 100)c 38 (20, 70)c 94 (60, 100) c 36 (10, 60) c 71 (15, 100)c 32 (10, 70)c 73 (32, 100)c – 125 (100, 140)c – 118 (87, 135) c
(2009)14

Lee et al 40 (17.8) 93 (9.7) 38 (17.3) 94 (5.6) 42 (15.4) 73 (25.0) 42 (14.3) 80 (19.5) 112 (14.6) 115 (15.8) 109 (20.7) 121 (11.9)
(2012)30

Liu et al 46 89.5 46 90.1 45.2 65.6 43.7 67.0 Flex: 110.6, Ext: Flex: 110, Ext: 0.7 Flex: 114.4, Ext: Flex: 116.6, Ext:
(2015)32 5.1 4.8 0.7

Lützner et al OTP: preop 43.5 (18.3), postop 86.0 (14.6) OTP: preop 41.5 (18.0), postop 62.9 (22.5) OTP Flex: 105.0 (14.1) OTP Flex: 113.5 (14.0)
(2015)33

Maruyama 42.8 (9.9) 89.8 (7.2) 43.6 (10.1) 89.5 (8.9) OTP: preop 55.3 (11.8), postop 83.3 (11.3) 112.5 (20.9) 122.2 (14.8) 112.2 (22.5) 129.6 (13.9)
et al (2004)7

Matsumoto et 55.4 (28, 74)c 95.7 (84, 100) c 58.0 (11, 76)c 92.9 (81, 100)c 54.7 (35, 82) c 88.6 (72, 100)c 55.9 (30, 82)c 84.8 (72, 100)c 114.2 (75, 140)c 125.3 (95, 140)c 114.2 (75, 140)c 121.6 (100,
al (2012)34 135)c

Misra et al – – – – – – – – 82 (30, 112) c 107.5 (60, 125)c 84.5 (30, 110)c 105.2 (60, 115) c
35
(2003)

Mouttet and 31 (14) 85 (12) 26 (15) 87 (8) 35 (16) 70 (27) 32 (13) 67 (28) 115 (11) 112 (11) 113 (16) 115 (10)
Sourdet
(2014)8

Peters et al 53 (18) 92 (10) 52 (12) 88 (19) 56 (22) 85 (23) 58 (23) 81 (27) – – – –

(2014)36

Sando et al 41.9 (14.1) 92.4 (12.2) 40.2 (13.6) 93.0 (10.4) 44.7 (15.2) 56.5 (32.5) 47.7 (13.7) 69.5 (25.0) Flex: 112.5 Flex: 113.5 (13.4), Flex: 109.1 Flex: 116.7
(2015)37 (13.0), Ext: 3.9 Ext: 0.5 (2.0) (17.4), Ext: 6.4 (12.8), Ext: 0.5
(4.7) (8.3) (2.3)

Seon et al – – – – – – – – 133.8 (9.4) 115 (15.1) 129.2 (12.3) 126.3 (14.1)

(2011)38
Outcomes of PS Compared with CR TKA

Snider and – – – – – – – – – – – –
Macdonald
(2009)39

Tanzer et al 47 (17) 90 (12) 44 (11) 93 (11) 34 (13) 73 (24) 38 (13) 76 (28) 110 (12) 112 (13) 101 (23) 111 (17)
Longo et al.

The Journal of Knee Surgery

(2002)40

(Continued)

Downloaded by: University of Michigan. Copyrighted material.

 Table 3 (Continued)

Thomsen et al – – – – – – – – 116 (70, 140)c 120 (104, 146)c 118 (80, 140)c 127 (107, 146)c
41
(2013)

Tsuneizumi – 91.7 (3.4) – 91.6 (7.1) – 81.8 (12.8) – 80.0 (8.1) – 116.3 (11.0) – 118.9 (11.2)
et al (2008)42

van den Boom 55 (13) 77 (10) 52 (8) 83 (7) – – – – 119 (16) 113 (11) 119 (13) 120 (7)

The Journal of Knee Surgery

et al (2014)43

Vermesan et al – 83.4 (8.5) – 86.1 (5.7) – – – – – 100 (10) – 110 (15)

(2015)6

Victor et al 38.0 (15.1) 82.2 (2.2) 37.2 (21.7) 77.9 (13.8) 50.5 (13.4) 83.6 (15.3) 39.1 (18.5) 74.8 (27.9) 109 (11) 114 (10) 103 (9) 117 (7)
(2005)9

Wang et al 51.7 (20.3) 90.7 (5.3) 45.9 (20.7) 91.0 (4.8) 41.0 (19.9) 84.2 (20.8) 39.2 (21.9) 87 (19.6) Flex: 114 (21.1), Flex: 110, Ext: 1.0 Flex: 110 (20), Flex: 112,
Outcomes of PS Compared with CR TKA

(2004)44 Ext: 6.8 (7.0) Ext: 9.4 (9.6) Ext: 2.8

Yagishita et al 46.6 (9.7) 95.4 (4.1) 48.6 (11.3) 93.5 (5.5) – – – – 132.9 (16.5) 125.4 (10.9) 133.1 (18.5) 129.3 (11.7)
(2012)45

Yoshiya et al – – – – – – – – 113 (19) 121 (16) 113 (21) 131 (12)

Longo et al.

(2005)46

Mean 45.2 90.3 43.7 90.8 45.6 76.6 44.6 77.6 115.0 114.4 115.2 119.4

Mean 45.1 47.1 31.0 33.0  0.6 4.3

improvement

Statistical ana- PS vs. CR KSS improvement (OR, 0.17; 95% CI, 1.24 to 1.57; p ¼ 0.82) PS vs. CR KSFS improvement (OR, 2.77; 95% CI, 5.36 to 0.18; PS vs. CR flexion improvement (OR, 4.70; 95% CI, 5.81 to 3.60; p < 0.00001a);
a
lysis: PS vs. CR p ¼ 0.04 ) PS vs. CR extension improvement (OR, 0.57; 95% CI, 1.05 to 0.09; p ¼ 0.02a)

Number of KSS: 32 of 37 studies (86.5%) KSFS: 21 of 37 studies (56.8%) ROM: 32 of 37 studies (86.5%)
studies

Abbreviations: CI, confidence interval; CR, cruciate retaining; Ext, extension; Flex, flexion; KSFS, Knee Society functional score; KSS, Knee Society Score; OR, odds ratio; OTP, only total provided; Postop,
postoperative; Preop, preoperative; PS, posterior stabilizing; ROM, range of motion; SD, standard deviation.
a
p-Value (statistically significant) of each outcome measure compared with preoperative.
b
Interquartile range.
c
Range.

Downloaded by: University of Michigan. Copyrighted material.

 Outcomes of PS Compared with CR TKA Longo et al.

Fig. 2 Forest plot of postoperative KSOS after CR and PS procedures. CI, confidence interval; CR, cruciate-retaining; KSOS, Knee Society
objective score; M-H, Mantel–Haenszel; PS, posterior-stabilized; SD, standard deviation.

Downloaded by: University of Michigan. Copyrighted material.

Fig. 3 Forest plot of postoperative KSFS after CR and PS procedures. CI, confidence interval; CR, cruciate-retaining; KSFS, Knee Society
functional score; M-H, Mantel–Haenszel; PS, posterior-stabilized; SD, standard deviation.

Fig. 4 Forest plot of postoperative flexion after CR and PS procedures. CI, confidence interval; CR, cruciate-retaining; M-H, Mantel–Haenszel; PS,
posterior-stabilized; SD, standard deviation.

Fig. 5 Forest plot of postoperative extension after CR and PS procedures. CI, confidence interval; CR, cruciate-retaining; M-H, Mantel–Haenszel;
PS, posterior-stabilized; SD, standard deviation.

Fig. 6 Forest plot of postoperative complications after CR and PS procedures. CI, confidence interval; CR, cruciate-retaining; M-H, Mantel–
Haenszel; PS, posterior-stabilized.

The Journal of Knee Surgery

Outcomes of PS Compared with CR TKA Longo et al.
Fig. 7 GRADE. Summary of findings. CI, confidence interval; CR, cruciate-retaining; GRADE, Grading of Recommendations Assessment,
Development and Evaluation; KSFS, Knee Society functional score; KSS, Knee Society score; PS, posterior-stabilized; ROM, range of motion.
two methods, certain cohorts may be suited to one method
over the other. This is clinically important information for
surgeons to offer evidence-based appropriate health care for
their patients. Randomized controlled trials observing more
objective measures such as biomechanical, kinetic perfor-
mance of CR TKAs compared with PS TKAs, as well as out-
comes in select cohorts, such as patients with altered
proprioception, imbalance issues, or additional orthopathol-
ogy, will yield very interesting results.
Conclusion
This meta-analysis has demonstrated that PS TKA has a
statistically significant greater postoperative improvement
of KSFS (p ¼ 0.04), flexion (p < 0.00001), and extension
(p ¼ 0.02), compared with the CR group. These findings
seem to lead the surgeons to prefer the PS design for TKAs
especially to achieve a higher postoperative ROM in patients
with high functional demands. On the contrary, the CR and PS
TKAs have similar results in terms of complications and most
of clinical outcomes analyzed in the included studies. There-
fore, the long-term follow-up of high-quality randomized
controlled trials is needed to clarify which of the two types
of prosthesis provide the better clinical outcome and the lower
rate of complications for OA patients in particular cohorts.
10
References
1 Kawakami Y, Matsumoto T, Takayama K, et al. Intermediate-term
11
comparison of posterior cruciate-retaining versus posterior-sta-
bilized total knee arthroplasty using the new knee scoring
system. Orthopedics 2015;38(12):e1127–e1132
The Journal of Knee Surgery
Downloaded by: University of Michigan. Copyrighted material.
2 Delport HP. The advantage of a total knee arthroplasty with
rotating platform is only theoretical: prospective analysis of
1,152 arthroplasties. Open Orthop J 2013;7:635–640
3 Harato K, Bourne RB, Victor J, Snyder M, Hart J, Ries MD. Midterm
comparison of posterior cruciate-retaining versus -substituting total
knee arthroplasty using the Genesis II prosthesis. A multicenter
prospective randomized clinical trial. Knee 2008;15(03):217–221
4 Waddell DD, Sedacki K, Yang Y, Fitch DA. Early radiographic and
functional outcomes of a cancellous titanium-coated tibial com-
ponent for total knee arthroplasty. Musculoskelet Surg 2016;100
(01):71–74
5 Chen JY, Lo NN, Chong HC, et al. Cruciate retaining versus poster-
ior stabilized total knee arthroplasty after previous high tibial
osteotomy. Knee Surg Sports Traumatol Arthrosc 2015;23(12):
3607–3613
6 Vermesan D, Trocan I, Prejbeanu R, et al. Reduced operating time
but not blood loss with cruciate retaining total knee arthroplasty.
J Clin Med Res 2015;7(03):171–175
7 Maruyama S, Yoshiya S, Matsui N, Kuroda R, Kurosaka M. Func-
tional comparison of posterior cruciate-retaining versus posterior
stabilized total knee arthroplasty. J Arthroplasty 2004;19(03):
349–353
8 Mouttet A, Sourdet V. EUROP total knee prosthesis with or with-
out posterior cruciate ligament retention? Comparative study at
mid-term follow-up. Orthop Traumatol Surg Res 2014;100(08):
895–900
9 Victor J, Banks S, Bellemans J. Kinematics of posterior cruciate
ligament-retaining and -substituting total knee arthroplasty: a
prospective randomised outcome study. J Bone Joint Surg Br
2005;87(05):646–655
Atzori F, Sabatini L, Deledda D, Schirò M, Lo Baido R, Massè A.
Evaluation of anterior knee pain in a PS total knee arthroplasty:
the role of patella-friendly femoral component and patellar size.
Musculoskelet Surg 2015;99(01):75–83
Notarnicola A, Moretti L, Tafuri S, Vacca A, Marella G, Moretti B.
Postoperative pain monitor after total knee replacement. Muscu-
loskelet Surg 2011;95(01):19–24

12 Rosenberg AG. Posterior cruciate ligament retention in total knee
arthroplasty. In: Scott WN, ed. Insall & Scott Surgery of the Knee,
5 ed. London, England: Churchill Livingstone; 2012:1125–1131
13 Bercik MJ, Joshi A, Parvizi J. Posterior cruciate-retaining versus
posterior-stabilized total knee arthroplasty: a meta-analysis.
J Arthroplasty 2013;28(03):439–444
14 Kolisek FR, McGrath MS, Marker DR, et al. Posterior-stabilized
versus posterior cruciate ligament-retaining total knee arthro-
plasty. Iowa Orthop J 2009;29:23–27
15 Liberati A, Altman DG, Tetzlaff J, et al. The PRISMA statement for
reporting systematic reviews and meta-analyses of studies that
evaluate health care interventions: explanation and elaboration.
J Clin Epidemiol 2009;62(10):e1–e34
16 Insall JN, Dorr LD, Scott RD, Scott WN. Rationale of the Knee Society
clinical rating system. Clin Orthop Relat Res 1989;(248):13–14
17 Akasaki Y, Matsuda S, Miura H, et al. Total knee arthroplasty
following failed high tibial osteotomy: mid-term comparison of
posterior cruciate-retaining versus posterior stabilized prosthe-
sis. Knee Surg Sports Traumatol Arthrosc 2009;17(07):795–799
18 Ang CL, Fook S, Chia SL, Chin PL, Lo NN, Yeo SJ. Unconstrained
arthroplasty in type II valgus knees: posterior stabilized or
cruciate retaining? Knee Surg Sports Traumatol Arthrosc 2014;
22(03):666–673
19 Bin Abd Razak HR, Pang HN, Yeo SJ, Tan MH, Lo NN, Chong HC.
Joint line changes in cruciate-retaining versus posterior-stabi-
lized computer-navigated total knee arthroplasty. Arch Orthop
Trauma Surg 2013;133(06):853–859
20 Catani F, Leardini A, Ensini A, et al. The stability of the cemented
tibial component of total knee arthroplasty: posterior cruciate-
retaining versus posterior-stabilized design. J Arthroplasty 2004;
19(06):775–782
21 Cates HE, Komistek RD, Mahfouz MR, Schmidt MA, Anderle M. In
vivo comparison of knee kinematics for subjects having either a
posterior stabilized or cruciate retaining high-flexion total knee
arthroplasty. J Arthroplasty 2008;23(07):1057–1067
22 Chaudhary R, Beaupré LA, Johnston DW. Knee range of motion
during the first two years after use of posterior cruciate-stabilizing
or posterior cruciate-retaining total knee prostheses. A randomized
clinical trial. J Bone Joint Surg Am 2008;90(12):2579–2586
23 Clark CR, Rorabeck CH, MacDonald S, MacDonald D, Swafford J,
Cleland D. Posterior-stabilized and cruciate-retaining total knee
replacement: a randomized study. Clin Orthop Relat Res 2001;
(392):208–212
24 Conditt MA, Noble PC, Bertolusso R, Woody J, Parsley BS. The PCL
significantly affects the functional outcome of total knee arthro-
plasty. J Arthroplasty 2004;19(07, Suppl 2):107–112
25 Carvalho LH Jr, Temponi EF, Soares LFM, Gonçalves MJ. Relation-
ship between range of motion and femoral rollback in total knee
arthroplasty. Acta Orthop Traumatol Turc 2014;48(01):1–5
26 Fantozzi S, Catani F, Ensini A, Leardini A, Giannini S. Femoral
rollback of cruciate-retaining and posterior-stabilized total knee
replacements: in vivo fluoroscopic analysis during activities of
daily living. J Orthop Res 2006;24(12):2222–2229
27 Hamai S, Okazaki K, Shimoto T, Nakahara H, Higaki H, Iwamoto Y.
Continuous sagittal radiological evaluation of stair-climbing in
cruciate-retaining and posterior-stabilized total knee arthroplas-
ties using image-matching techniques. J Arthroplasty 2015;
30(05):864–869
28 Han CW, Yang IH, Lee WS, Park KK, Han CD. Evaluation of post-
operative range of motion and functional outcomes after cruci-
ate-retaining and posterior-stabilized high-flexion total knee
arthroplasty. Yonsei Med J 2012;53(04):794–800
29 Kim YH, Choi Y, Kwon OR, Kim JS. Functional outcome and range
of motion of high-flexion posterior cruciate-retaining and high-
flexion posterior cruciate-substituting total knee prostheses. A
prospective, randomized study. J Bone Joint Surg Am 2009;91
(04):753–760
Outcomes of PS Compared with CR TKA Longo et al.
30 Lee SM, Seong SC, Lee S, Choi WC, Lee MC. Outcomes of the
different types of total knee arthroplasty with the identical
femoral geometry. Knee Surg Relat Res 2012;24(04):214–220
31 Li N, Tan Y, Deng Y, Chen L. Posterior cruciate-retaining versus
posterior stabilized total knee arthroplasty: a meta-analysis of
randomized controlled trials. Knee Surg Sports Traumatol Ar-
throsc 2014;22(03):556–564
32 Liu HG, Zhu W, Zhang ZX. Comparison of outcomes after bilateral
simultaneous total knee arthroplasty using posterior-substitut-
ing versus cruciate-retaining prostheses. Saudi Med J 2015;
36(02):190–195
33 Lützner J, Firmbach FP, Lützner C, Dexel J, Kirschner S. Similar
stability and range of motion between cruciate-retaining and
cruciate-substituting ultracongruent insert total knee arthro-
plasty. Knee Surg Sports Traumatol Arthrosc 2015;23(06):
1638–1643
34 MatsumotoT, Muratsu H, Kubo S, Matsushita T, Kurosaka M, Kuroda
R. Intraoperative soft tissue balance reflects minimum 5-year
midterm outcomes in cruciate-retaining and posterior-stabilized
total knee arthroplasty. J Arthroplasty 2012;27(09):1723–1730
35 Misra AN, Hussain MR, Fiddian NJ, Newton G. The role of the
posterior cruciate ligament in total knee replacement. J Bone Joint
Surg Br 2003;85(03):389–392
Downloaded by: University of Michigan. Copyrighted material.
36 Peters CL, Mulkey P, Erickson J, Anderson MB, Pelt CE. Comparison
of total knee arthroplasty with highly congruent anterior-stabi-
lized bearings versus a cruciate-retaining design. Clin Orthop
Relat Res 2014;472(01):175–180
37 Sando T, McCalden RW, Bourne RB, MacDonald SJ, Somerville LE.
Ten-year results comparing posterior cruciate-retaining versus
posterior cruciate-substituting total knee arthroplasty. J Arthro-
plasty 2015;30(02):210–215
38 Seon J-K, Park J-K, Shin Y-J, Seo H-Y, Lee K-B, Song E-K.
Comparisons of kinematics and range of motion in high-flexion
total knee arthroplasty: cruciate retaining vs. substituting de-
signs. Knee Surg Sports Traumatol Arthrosc 2011;19(12):
2016–2022
39 Snider MG, Macdonald SJ. The influence of the posterior cruciate
ligament and component design on joint line position after
primary total knee arthroplasty. J Arthroplasty 2009;24(07):
1093–1098
40 Tanzer M, Smith K, Burnett S. Posterior-stabilized versus cruciate-
retaining total knee arthroplasty: balancing the gap. J Arthro-
plasty 2002;17(07):813–819
41 Thomsen MG, Husted H, Otte KS, Holm G, Troelsen A. Do patients
care about higher flexion in total knee arthroplasty? A rando-
mized, controlled, double-blinded trial. BMC Musculoskelet
Disord 2013;14(01):127
42 Tsuneizumi Y, Suzuki M, Miyagi J, et al. Evaluation of joint laxity
against distal traction force upon flexion in cruciate-retaining and
posterior-stabilized total knee arthroplasty. J Orthop Sci 2008;
13(06):504–509
43 van den Boom LG, Halbertsma JP, van Raaij JJ, Brouwer RW, Bulstra
SK, van den Akker-Scheek I. No difference in gait between poster-
ior cruciate retention and the posterior stabilized design after
total knee arthroplasty. Knee Surg Sports Traumatol Arthrosc
2014;22(12):3135–3141
44 Wang CJ, Wang JW, Chen HS. Comparing cruciate-retaining total
knee arthroplasty and cruciate-substituting total knee arthro-
plasty: a prospective clinical study. Chang Gung Med J 2004;
27(08):578–585
45 Yagishita K, Muneta T, Ju Y-J, MoritoT, Yamazaki J, Sekiya I. High-flex
posterior cruciate-retaining vs posterior cruciate-substituting de-
signs in simultaneous bilateral total knee arthroplasty: a prospec-
tive, randomized study. J Arthroplasty 2012;27(03):368–374
46 Yoshiya S, Matsui N, Komistek RD, Dennis DA, Mahfouz M,
Kurosaka M. In vivo kinematic comparison of posterior cruci-
ate-retaining and posterior stabilized total knee arthroplasties
The Journal of Knee Surgery
Outcomes of PS Compared with CR TKA Longo et al.
under passive and weight-bearing conditions. J Arthroplasty
2005;20(06):777–783
47 Nguyen LC, Lehil MS, Bozic KJ. Trends in total knee arthroplasty
implant utilization. J Arthroplasty 2015;30(05):739–742
48 Cankaya D, Ozkurt B, Aydin C, Tabak AY. No difference in blood loss
between posterior-cruciate-ligament-retaining and posterior-
The Journal of Knee Surgery
cruciate-ligament-stabilized total knee arthroplasties. Knee
Surg Sports Traumatol Arthrosc 2014;22(08):1865–1869
49 Mähringer-Kunz A, Efe T, Fuchs-Winkelmann S, Schüttler KF,
Paletta JR, Heyse TJ. Bleeding in TKA: posterior stabilized vs.
cruciate retaining. Arch Orthop Trauma Surg 2015;135(06):
867–870
Downloaded by: University of Michigan. Copyrighted material.