Medial Patellofemoral Ligament
Reconstruction Combined With Distal
Realignment for Recurrent Dislocations
of the Patella
5-Year Results of a Randomized Controlled Trial
MBBS, Murray Blythe,
FRACS(Orth), David Wysocki,
MBBS, and Peter Annear,
Investigation performed at the Perth Orthopaedic & Sports Medicine Centre, Perth, Australia
Background: Tibial tubercle transfer (TTT) and medial patellofemoral ligament (MPFL) reconstruction have both shown, either in
isolation or in combination, to provide improved patellofemoral joint (PFJ) stability. There are few studies that provide evidence
that this remains true in the long term.
Purpose: To compare the long-term results of patellar instability after TTT with and without MPFL reconstruction in 2 randomized
Study Design: Randomized controlled trial; Level of evidence, 1.
Methods: A total of 34 patients (36 knees) were randomized to 2 groups. The first group underwent lateral release (LR) and TTT
for confirmed maltracking of the patella (control group). The second group underwent MPFL reconstruction in addition to TTT and
LR (reconstruction group). Patients were followed up with validated questionnaires (Kujala score, Tegner activity score), a visual
analog scale (VAS) assessing their insecurity, and a clinical assessment at a minimum of 5 years postoperatively. Participants also
underwent quantitative computed tomography (CT) at 1 year for comparison. Two patients in the control group and 1 patient in
the reconstruction group were lost to follow-up at 5 years.
Results: There were no significant differences in the Kujala (P= .75), Tegner (P= .36), or VAS (P= .75) scores at any time period.
One patient in the control group sustained a patellar redislocation at 3 years. Five patients in the control group and 2 in the recon-
struction group had functional failures and required reoperations; however, this was not statistically significant (P= .30). There
were no significant differences between groups in the time to return to school or work (P= .65) or sports (P= .38) after surgery.
Overall patient satisfaction was higher in the reconstruction group compared with the control group (P= .04), and quantitative CT
scans showed that the reconstruction group had a statistically significant improvement in the mean patellar tilt (6°vs 28°, respec-
tively; P= .03) and mean congruence angle (13°vs 211°, respectively; P= .03) in the quadriceps-contracted state compared with
the control group.
Conclusion: Reconstruction of the MPFL in addition to TTT and LR resulted in improved alignment parameters (congruence
angle, patellar tilt angle) as well as patient satisfaction. The Kujala and Tegner scores were no different between the 2 groups
at any time period. There was insufficient evidence to conclude that the addition of MPFL reconstruction to TTT results in fewer
redislocations or reoperations. This study concludes that MPFL reconstruction improves PFJ alignment and patient satisfaction;
however, further studies with larger patient numbers are required to satisfy its significance with respect to redislocation rates and
functional scores in the long term.
Keywords: patellofemoral instability; recurrent patellar dislocation; MPFL reconstruction; tibial tubercle transfer
The nonoperative management of patellofemoral instabil-
ity has been reported to have long-term recurrence rates
of up to 49%.
Half of those who do not suffer a further
dislocation fail to return to their chosen sport by 6
The condition is known to be multifactorial,
with malalignment, patellofemoral dysplasia, patella alta,
soft tissue imbalance, and ligamentous laxity contributing
variably in individual cases. Operations to address these
factors can be grouped into proximal soft tissue balancing,
The American Journal of Sports Medicine, Vol. XX, No. X
Ó2016 The Author(s)
distal bony procedures, and trochleoplasty. The modified
Elmslie-Trillat procedure consisting of medial rotation tib-
ial tubercle transfer (TTT) combined with lateral release
(LR) addresses both soft tissue and bony abnormalities.
Case series have generally reported satisfactory outcomes,
although long-term recurrence rates of up to 13% have been
Autograft reconstruction of the medial patellofemoral
ligament (MPFL) has become an increasingly common pro-
cedure in the last decade. Three systematic reviews
reported satisfactory functional outcomes and low redislo-
cation rates but noted that available evidence consists
mainly of case series of varying techniques and participants
of differing age ranges.
Two randomized controlled tri-
als (RCTs) have compared isolated MPFL repair to nonoper-
ative management in first-time dislocators and reported
conflicting outcomes for redislocation rates and functional
There has been no RCT of MPFL reconstruc-
tion compared with other surgical treatments.
This study investigated the additional benefit of MPFL
reconstruction for patients undergoing TTT and LR for
recurrent patellar instability. Our primary hypothesis
was that MPFL reconstruction would improve subjective
outcome measures. Additionally, we aimed to examine
the effect of MPFL reconstruction on postoperative redislo-
cations, functional failures, patient satisfaction, and post-
operative patellar kinematics based on dynamic
computed tomography (CT).
We designed a prospective RCT. Eighty-seven knees in 84
consecutively referred patients with recurrent lateral
patellar dislocations were assessed for enrollment in the
study between December 2007 and November 2010.
Approval was obtained from a hospital ethics committee,
and all participants gave informed consent. Patients were
not informed of their allocation until the conclusion of
The inclusion criteria were (1) 3 lateral patellar dislo-
cations, (2) no congenital or habitual dislocations, (3)
abnormal patellar tracking as determined by the presence
of ‘‘J’’ tracking and lateral subluxation of the patella
through a qualitative assessment of CT scans, (4) skeletal
maturity, (5) no previous patellofemoral realignment pro-
cedure (bony or soft tissue), (6) no significant ligamentous
knee injury, (7) absent or minor patellofemoral joint (PFJ)
degenerative arthropathy, and (8) competence to consent
to the trial and follow-up period.
A total of 39 patients (39 knees) were excluded (Figure
1): 15 had undergone previous realignment surgery, 8
had fewer than 3 dislocations, 6 were skeletally immature,
4 had other significant ligamentous knee injuries, 4 had
moderate to severe degenerative PFJ arthropathy, 1 did
not have subluxation on CT, and 1 was a habitual disloca-
tor. Eleven patients (12 knees) declined involvement in the
study. Thirty-six knees in 34 consecutive patients with
recurrent lateral patellar dislocations were randomized to
undergo TTT and LR (control group) or TTT, LR, and
MPFL reconstruction (reconstruction group). Patients
were randomized to each group by computer-generated
instructions placed into sealed, opaque envelopes. The
envelopes were then opened in the operating theater once
general anesthesia had been administered.
Seventeen patients (18 knees) were randomized to the
control group, and all underwent TTT and LR as allo-
cated. Two patients (2 knees) were lost to follow-up. One
patient was not able to be contacted by any means at all
follow-up intervals. One patient declined to attend fol-
low-up visits after the 3-month follow-up and before post-
operative CT because of an interstate move. The
remaining 16 knees were observed at a minimum of 5
Address correspondence to Iswadi Damasena, MBBS, Orthopaedic Surgery Department, Sir Charles Gairdner Hospital, Hospital Avenue, Nedlands,
Perth 6008151, Western Australia (email: firstname.lastname@example.org).
Orthopaedic Surgery Department, Sir Charles Gairdner Hospital, Perth, Australia.
Southern Cross Orthopaedic Group, Perth, Australia.
Perth Orthopaedic & Sports Medicine Centre, Perth, Australia.
Orthopaedic Surgery Department, Royal Perth Hospital, Perth, Australia.
One or more of the authors has declared the following potential conflict of interest or source of funding: P.A. has received institutional support from
Corin Australia Pty Ltd and Smith & Nephew Australia.
Assessed for eligibility
(84 patients/87 knees)
(34 patients/36 knees)
Excluded (50 patients/
• 39 did not meet inclusion
• 11 declined participation
TTT + LR
(17 patients/18 knees)
TTT + LR + MPFL
(17 patients/18 knees)
15 patients (16 knees);
2 lost to follow-up
• 1 unable to be reached
• 1 declined
16 patients (17 knees);
1 lost to follow-up
• 1 unable to be reached
15 patients (16 knees) 16 patients (17 knees)
Figure 1. Patient flowchart. LR, lateral release; MPFL,
medial patellofemoral ligament; TTT, tibial tubercle transfer.
2Damasena et al The American Journal of Sports Medicine
Seventeen patients (18 knees) were randomized to the
reconstruction group, and all underwent TTT, LR, and
MPFL reconstruction as allocated. One patient could not be
contacted at any follow-up interval. The remaining 17 knees
were observed at a minimum of 5 years postoperatively.
Based on a comparison of 2 independent groups, a sample
size of 16 participants per group was estimated to have suf-
ficient power (b= 0.2) to detect a difference of 7 points in
the Kujala score. This difference was based on the Kujala
scores in a case series of modified Elmslie-Trillat proce-
dures published in the literature
compared with an
unpublished case series from this center of a modified Elm-
slie-Trillat procedure combined with MPFL reconstruction.
Assuming a 10% rate of a loss to follow-up, a sample size of
18 in each group was recruited.
All procedures were performed by the senior author (P.A.).
An arthroscopic assessment of the knee was performed and
additional pathological conditions addressed as required. If
present, patellofemoral chondropathy was assessed for
severity by the Outerbridge classification and the pattern
according to Pidoriano et al.
LR was performed arthroscopically using radiofre-
quency ablation (VAPR; DePuy Mitek). This was per-
formed in the inferolateral retinaculum in all patients to
prevent tethering of the extensor mechanism with TTT.
When there was less than 1 quadrant of medial patellar
glide, the release was extended to the level of the proximal
pole of the patella. TTT was then performed through
a 5-cm transverse incision centered over the middle third
of the tibial tuberosity extending medially to the pes anser-
inus using longitudinal osteotomy oriented for anteromedi-
alization. The osteotomy site was translated approximately
8 to 10 mm and temporarily held with a drill bit. Tracking
through a range of motion was reassessed arthroscopically.
Normal tracking was defined as patellar central engage-
ment on the trochlea 40°of knee flexion viewed through
an anterolateral portal. The osteotomy procedure was
adjusted if required. When confirmed appropriate, the
tubercle was fixed with 2 fully threaded cancellous AO
4.0-mm screws (DePuy Synthes).
In the reconstruction group, the graft was harvested
before TTT. The semitendinosus tendon was harvested
using a custom tendon stripper, leaving the pes anserinus
insertion preserved. The free end was whipstitched by
passage and tensioning. Through a 3-cm transverse medial
patellar incision, a medial-to-anterolateral bone tunnel was
created at the midpoint of the patella using a 4.5-mm drill
bit. The graft was passed subcutaneously from the first inci-
sion to the second, through the bone tunnel from anterolat-
eral to medial, and then deep to the medial retinaculum to
a tunnel based 3 mm proximal to the medial epicondyle.
The graft was not fixed within the patellar tunnel. The
femoral tunnel was reamed to 7 mm over a guide wire
directed 30°proximally and anteriorly exiting the lateral
femoral cortex to allow free tensioning of the graft. The graft
was initially tensioned in extension with the patella manu-
ally reduced to the center of the trochlea and then adjusted
to allow 1 patellar quadrant of lateral glide. Tracking was
arthroscopically reassessed through a range of motion,
with the patella confirmed to be engaging the trochlea at
\40°of knee flexion. If there was medial tilt or translation
in terminal extension or excessive tension in deep flexion,
then the graft tension was adjusted. When confirmed appro-
priate, the knee was placed in full extension and the graft
secured with a 7-mm interference screw (Guardsman;
Conmed Linvatec) in the tunnel at the medial epicondyle.
Rehabilitation was identical for both groups. Active flex-
ion exercises began on the first day postoperatively.
Patients were fully weightbearing immediately but wore
an extension splint when ambulating for the first 3 weeks.
Return to sport was permitted when the rehabilitation
goals had been achieved.
Patients were assessed preoperatively; at 6 weeks, 3
months, and 12 months; and at a minimum of 5 years at
final follow-up. The clinical assessment at the 6-week
follow-up was performed by the senior author (P.A.). The
assessments at all other time periods were performed by 3
authors (I.D., M.B., D.W.) who were independent of the sur-
gical procedures and patient care. The assessors could not
be blinded, as the pattern of incisions indicates whether
MPFL reconstruction had been performed. At each time
period, patients were assessed for passive knee extension,
active knee flexion, apprehension, tenderness over the
osteotomy and hamstring donor sites, dislocations, return
to work or school, and return to sports in addition to self-
administered scores consisting of the Kujala patellofemoral
the Tegner activity level score,
satisfaction score, and an ‘‘insecurity’’ visual analog scale
(VAS) score. For the VAS, the patients made a mark on
a 100-mm line indicating how insecure they perceived their
patella to be, as described by Watanabe et al
pletely secure’’ at 0 mm to ‘‘dislocating’’ at 100 mm). Patient
satisfaction was assessed on a 5-point Likert-type scale,
with participants asked to rate the outcome of their surgery
as excellent (1), good (2), fair (3), poor (4), and worse (5). Pri-
mary outcomes were measured as the redislocation rate and
functional failures requiring reoperations. Secondary out-
comes were measured as the Kujala score, Tegner score,
and VAS score.
Patients underwent quantitative patellofemoral CT preop-
eratively and postoperatively between the 3- and 12-month
follow-ups. CT was performed in a standardized manner.
The patellar height measurement was adapted from a tech-
nique described for magnetic resonance imaging (MRI)
AJSM Vol. XX, No. X, XXXX MPFL Reconstruction and Distal Realignment for Patellar Dislocations 3
allow all measurements to be performed from the same
The Kujala scores for both groups were compared at all
postoperative time periods after adjusting for baseline var-
iables using analysis of covariance (repeated-measures
analysis of covariance [ANCOVA]). These comparisons
were repeated for the Tegner activity score; VAS score;
range of motion data; and change in patellar tilt, patellar
height, congruency angle, and trochlear dysplasia based
on CT findings. The necessary assumptions for the analy-
sis were assessed and deemed appropriate for the data.
The association between group and categorical data was
evaluated by means of a Pearson chi-square test or Fisher
exact test depending on which was most suitable. Subjec-
tive patient satisfaction (Likert-type scales with 2 groups)
was analyzed using a Wilcoxon rank-sum test. A 2-sided P
value of \.05 indicated statistical significance.
The groups were comparable for sex, age, side, age of first
dislocation, other knee injuries, and previous operations
(Table 1). One patient in the control group had a nonopera-
tively managed posterior cruciate ligament injury 3 years
after her initial patellar dislocation. She had a stable pos-
terior drawer test result before her procedure. All previous
operations on the ipsilateral knee consisted of arthroscopic
removal of loose bodies or osteochondral fragments and
debridement. All patients had a positive preoperative
patellar apprehension test result. The groups were compa-
rable for patellofemoral and other intra-articular patholog-
ical conditions (Table 2).
The Tegner, Kujala, and VAS scores for each time period
are presented in Figures 2, 3, and 4, respectively. The
ANCOVA indicated no significant differences between
the groups for any of the scores. At the 6-week follow-up,
there was a trend for the reconstruction group to have
a lower insecurity VAS score (mean 6SD, 13 68.5 mm)
compared with the control group (27 619.8 mm), although
this was not statistically significant (P= .07). At 5 years,
functional scores were maintained or improved in compar-
ison to those at the 12-month follow-up, although there
was no statistical significance between the 2 groups (P=
.75 [Kujala] and .36 [Tegner]).
Demographic and Knee Characteristics
(n = 16)
(n = 17)
Sex, male/female, n 5/11 3/14
Age, y, mean (range) 16 (14-29) 21 (12-47)
Age of first dislocation, y,
16 (12-19) 17 (7-29)
Other ipsilateral knee injuries, n 1 0
Other ipsilateral knee operations, n 4 4
Arthroscopic Intraoperative Findings
at the Time of Reconstruction
(n = 16)
(n = 17)
Patellar chondropathy, n 11 9
Outerbridge grade, mean (range) 2.1 (1-3) 2.3 (1-3)
Chondropathy pattern, mean (range)
2.6 (1-4) 2.8 (1-4)
Other intra-articular injuries, n 3 2
According to Kujala et al.
6 weeks 3 months 12 months 5+ years
Figure 2. Tegner activity score. MPFL, medial patellofemoral
ligament; TTT, tibial tubercle transfer.
6 weeks 3 months 12 months 5+ years
Figure 3. Kujala score. MPFL, medial patellofemoral liga-
ment; TTT, tibial tubercle transfer.
4Damasena et al The American Journal of Sports Medicine
At final follow-up, 5 of 16 patients in the control group and
2 of 17 in the reconstruction group were found to have
a ‘‘functional failure’’ (ie, a positive apprehension test
result, a history of subluxation episodes after surgery, or
a PFJ dislocation). There was no statistical difference
between the 2 groups (P= .30). Patient satisfaction scores
(Figure 5) were higher in the reconstruction group, 88%
(15/17) excellent compared with 56% (9/16) in the control
group, and this was statistically significant (P= .04).
Patients returned to work or school at a mean of 1.5 6
2.6 months in the control group and 1.2 60.7 months in
the reconstruction group (P= .65). Three (19%) in the con-
trol group and 6 (35%) in the reconstruction group did not
return to sport by the 1-year follow-up, but this was not
statistically significant (P= .43). At 5 years, all patients
had returned to a sporting activity; however, the majority
of them (90%) had not returned to their preinjury level of
sport. Of the patients who did return to sport, this occurred
at a mean of 3.8 63.2 months in the control group and 3.0
61.8 months in the reconstruction group (P= .38).
There were no significant differences in range of motion
preoperatively or postoperatively at any follow-up period
(Table 3). At the 6-week follow-up, there was a trend
toward greater flexion in the control group than the recon-
struction group, with a mean of 136.5°613.4°compared
with 120.4°620.5°, respectively (P= .07).
The quantitative CT findings are shown in Table 4. The
preoperative CT scans showed that the 2 groups had a sim-
ilar but wide variation of tibial tubercle (TT) lateralization,
the congruence angle, and the patellar tilt. There was no
difference in the patellar height (tendon length/patellar
length) between the 2 groups preoperatively (P= .26) or
postoperatively (P= .13). Trochlear dysplasia was
assessed on preoperative CT, and patients were deemed
to have a shallow dysplastic trochlea if they had a troch-
lear angle (TA) of .145°or a trochlear sulcus depth
(TD) of \3 mm. Both groups were comparable (Table 5),
and no statistically significant difference was found
between the 2 groups for the TA (P= .90) or TD (P=
.06). Postoperative assessments showed a similar magni-
tude of TT medialization between the groups. For the
patellar tilt, there was more (worsened) tilt postopera-
tively in the control group by a mean of 1°in the quadri-
ceps-relaxed state and 8°in the contracted state. The
reconstruction group showed a mean 5°and 6°less tilt
in the relaxed and contracted states, respectively. The
change in patellar tilt between the groups was not signif-
icant in the relaxed state (P= .16) but was significant in
the contracted state (P= .03).
For the congruence angle, there was more subluxation
postoperatively in the control group, with a mean 1°
increase in the quadriceps-relaxed state and 11°in the con-
tracted state. The reconstruction group showed a mean 11°
and 13°less subluxation in the relaxed and contracted
states, respectively. The change in congruence angle
between the groups was not significant in the relaxed state
(P= .24) but was significant in the contracted state (P=
.03). Typical postoperative lateral radiographs are dis-
played in Figure 6.
One patient in the control group sustained a patellar disloca-
tion 3 years after surgery while playing sports; the reconstruc-
tion group had no dislocations. There were 7 other
complications in the control group: 2 superficial wound infec-
tions, 1 patient requiring removal of TT screws, 2 patients
with TT screw irritation but not requiring removal, 1 pares-
thesia of the infrapatellar branch of the saphenous nerve,
6 weeks 3 months 12 months 5+ years
Figure 4. Visual analog scale (VAS) score for insecurity.
MPFL, medial patellofemoral ligament; TTT, tibial tubercle
Fair Poor Worse
(No. of Patients)
Figure 5. Subjective patient satisfaction at 5 years. MPFL,
medial patellofemoral ligament; TTT, tibial tubercle transfer.
Range of Motion for Both Groups at Each Time Period
Control Group Reconstruction Group
Preoperatively 2.8 65.9 142.8 67.3 4.1 64.6 145.1 68.5
6 weeks –0.4 61.3 136.5 613.4 –0.8 62.7 120.4 620.5
3 months 1.6 64.6 142.8 69.3 3.0 64.0 136.3 69.7
1 year 1.7 64.3 141.7 66.2 3.1 64.6 141.4 611.5
5 years 1.5 64.2 141.9 66.3 2.3 63.6 142.1 67.6
Data are reported as mean 6SD.
AJSM Vol. XX, No. X, XXXX MPFL Reconstruction and Distal Realignment for Patellar Dislocations 5
and 1 postoperative vasospasm that resolved spontaneously.
There were 2 complications in the reconstruction group: 1
deep infection requiring arthroscopic lavage and 1 patient
requiring removal of TT screws. The patient with a deep infec-
tion was included in the analyses at all follow-ups.
The challenge facing surgeons in managing patients with
recurrent patellar dislocations has been well documented.
Decision making is multifactorial, and clear guidelines
for optimal treatment are yet to be determined. MPFL
reconstruction has proved popular in the past decade,
with patients undergoing the procedure having improved
functional results and low redislocation rates.
Most long-term studies however are case series, use varying
methods for reconstructing the MPFL, and often combine
them with other procedures, both bony and soft tissue.
Reported results often have no comparison group, leaving
surgeons with the difficult choice between several treatment
options. To our knowledge, this is the first RCT to add
MPFL reconstruction to another surgical procedure.
Both the control and reconstruction groups showed
improved functional scores at a minimum 5-year follow-
up. Tegner activity, Kujala, VAS, and patient satisfaction
scores improved for both groups. The clinically significant
change in the VAS score for patellar insecurity is unknown,
but Crossley et al
reported that the minimum clinically
important change in the VAS score for patellofemoral pain
for both patients and clinicians is 20 of 100 mm. Other
measures of rehabilitation progress such as time to return
to work or school and return to sports were not different.
CT demonstrated an improvement in both the congruence
angle and patellar tilt in the reconstruction group but not in
the control group. This was statistically significant in the
quadriceps-contracted state but not in the relaxed state.
The MPFL graft was tensioned intraoperatively, such that
there was at least 1 quadrant of lateral glide possible pas-
sively. Therefore, the effect of MPFL reconstruction would
only be expected in the contracted state, as was found.
Although the difference in the functional scores was not sta-
tistically significant between the 2 groups, the reconstruction
group had better patient satisfaction, fewer episodes of insta-
bility or ‘‘functional failures,’’ and fewer reoperations. It
remains to be seen if these findings reflect improved patello-
femoral maltracking and hence contact pressures within the
PFJ, thereby reducing long-term chondral wear and eventual
osteoarthritis. Further long-term results will be required to
better answer this question.
The reconstruction group showed a trend toward less
flexion and lower functional scores at 6 weeks, which was
Preoperative Computed Tomography Results
of Trochlear Dysplasia
Trochlear angle, deg 143.6 68.4 143.8 67.7 .90
Trochlear sulcus depth, mm 5.1 61.6 6.2 61.6 .06
Data are reported as mean 6SD.
Figure 6. Postoperative lateral radiographs for the (A) con-
trol and (B) reconstruction groups.
Preoperative and Postoperative Quantitative Computed Tomography Results
(n = 16)
(n = 17)
(n = 15)
(n = 16)
Trochlear angle, deg 143 68.6 144 67.9
TT-TG distance, mm 16 63.2 15 63.5 10 65.0 12 64.3
Congruence angle, deg
Relaxed 24 617.0 17 618.3 25 618.0 6 619.5
Contracted 38 623.1 41 622.6 49 615.3 29 622.1
Patellar tilt angle, deg
Relaxed 0 611.7 6 67.9 –1 69.9 10 68.8
Contracted –5 614.8 –2 612.8 –12 610.2 3 614.1
TL/PL (patella alta), mm 1.40 60.15 1.34 60.15 1.44 60.22 1.33 60.18
Data are reported as mean 6SD. PL, patellar length; TL, tendon length; TT-TG, tibial tubercle–trochlear groove.
6Damasena et al The American Journal of Sports Medicine
not found at further follow-up. This may reflect greater pain
due to the additional procedure, which was not specifically
quantified. Another explanation could be excessive con-
straint in deep flexion caused by a nonanatomic MPFL posi-
tion. In this study, the femoral tunnel was positioned at the
medial epicondyle. We have since modified our technique by
using image intensification to more accurately replicate the
anatomic origin proximal and posterior to the medial
That there was little change in the control group’s radio-
graphic alignment is surprising. The measured TT medial-
ization, which does not measure anteriorization, was small
but similar between the groups and so unlikely to explain
the difference in postoperative alignment. All postopera-
tive and most preoperative CT scans were assessed by
the same radiologist. When the patient presented to the
clinic with CT already performed at another facility, it
was thought impractical to repeat the scans; as such, there
may have been an interobserver error, although this would
not be expected to be selective.
The redislocation rates and functional scores reported
for isolated MPFL reconstruction have generally been
excellent. The mean postoperative Kujala scores in 2
recent systematic reviews ranged from 83 to 96.
ever, it was noted that there was significant heterogeneity
in the inclusion criteria for the studies and the surgical
technique. One of the reviews included first-time patellar
dislocators together with redislocators.
did not exclude patients if they had undergone previous
knee surgery or if they were habitual dislocators, for exam-
Furthermore, most of the reported studies were only
of a short-term follow-up, making a direct comparison
between our results and these difficult.
There are a number of weaknesses in our study. First,
the assessors were not blinded to the treatment. The addi-
tional incisions required for MPFL reconstruction made it
impossible to blind the assessors, leading to a risk of mea-
surement bias. Patients themselves were not informed of
which arm of the study they were in; however, it must
again be deduced that this may be assumed from the sur-
gical scars. Similarly, although the radiologist assessing
the CT scans was not informed of patient allocation, this
can be determined from the images.
The study did not have a group of patients that underwent
MPFL reconstruction alone. Unfortunately, at the time of
recruitment, this was not the standard practice of the leading
surgeon, the decision to perform TTT being based on an
assessment of maltracking, not the TT–trochlear groove
(TT-TG) distance. Isolated MPFL reconstruction is generally
accepted for TT-TG distances \20 mm, above which TTT is
The mean preoperative TT-TG distance for
the control and reconstruction groups was 16 mm and
15 mm, respectively. It could be argued that these patients
only required MPFL reconstruction. A recent study by
Stephen et al
questioned this distance and found that
patellar tracking and contact pressures can be restored to
normal by isolated MPFL reconstruction up to a TT-TG dis-
tance of 15 mm, whereas those patients with TT-TG distan-
ces in excess of this may benefit from an additional TT
medialization procedure. Camp et al
also noted that the
TT-TG distance was not an accurate predictor of patellar
instability and devised a more patient-specific method that
takes patient size and individualized bony anatomy into
account. Thaunat and Erasmus
went a step further and sug-
gested that their failed MPFL reconstructions were caused by
unaddressed bony pathological abnormalities, including
abnormal TT-TG distances.
Although both groups were comparable for CT meas-
ures of trochlear dysplasia, the accuracy of this may be
questioned. Some authors have previously reported diffi-
culty in measuring the trochlear sulcus angle at 0°of flex-
ion, as was done in this study.
The interobserver and
intraobserver reliability of this measure was much better
at 20°of flexion. Ideally, trochlear dysplasia would have
been measured at 20°of flexion on preoperative radio-
graphs (classified by Dejour) or MRI; however, this was
not possible in our study.
There has been a trend toward quantitative radiology-
directed realignment protocols to tailor operations to indi-
This is in contrast with the standardized
surgical technique used in this study after selection by
clinical and qualitative CT assessments. The former
approach may help identify subcategories of patellofemoral
instability, particularly the severity of dysplasia, in which
the addition of MPFL reconstruction is beneficial or is
appropriate alone. Achieving sufficiently powered studies
of such subgroups will be difficult.
Sillanpaa et al
were the first to report on the long-
term results of MPFL reconstruction compared with distal
realignment for recurrent patellar dislocations. In their
retrospective case series, 47 Finnish military servicemen
were reviewed at a mean follow-up time of 10.2 years.
Eighteen underwent adductor magnus tenodesis and 29
a Roux-Goldthwait procedure for distal realignment. These
authors noted that 5 patients demonstrated patellofemoral
osteoarthritis in the Roux-Goldthwait group and none in
the adductor magnus tenodesis group. They concluded
that MPFL reconstruction reduces the risk of osteoarthri-
tis compared with distal realignment surgery. Although
this study was not randomized, used only male partici-
pants, and did not include preoperative results, their find-
ings suggest that MPFL reconstruction may improve
patellofemoral kinematics and hence reduce chondral
wear. There is not enough evidence in the literature how-
ever to comment on isolated MPFL reconstruction in
patients with documented TT-TG distances .15 mm. In
these patients, larger contact pressures in the PFJ result
from greater tension in the MPFL graft.
As a result,
the increased joint reaction forces and elevated articular
contact pressures may predispose to early chondral wear
in the PFJ. We therefore recommend realignment proce-
dures be considered in these patients in addition to
MPFL reconstruction to reduce wear patterns and poten-
tially patellofemoral osteoarthritis.
Reconstruction of the MPFL in addition to TTT and LR
resulted in improved alignment parameters (congruence
AJSM Vol. XX, No. X, XXXX MPFL Reconstruction and Distal Realignment for Patellar Dislocations 7
angle, patellar tilt angle) as well as patient satisfaction.
The Kujala and Tegner activity scores were no different
between the 2 groups at any time period. There was insuf-
ficient evidence to conclude that the addition of MPFL
reconstruction to TTT results in fewer redislocations or
reoperations. This study concludes that MPFL reconstruc-
tion improves PFJ alignment and patient satisfaction;
however, further studies with larger patient numbers are
required to satisfy its significance with respect to redisloca-
tion rates and functional scores in the long term.
The authors thank David Hille for providing support with
1. Atkin DM, Fithian DC, Marangi KS, Stone ML, Dobson BE, Mendel-
sohn C. Characteristics of patients with primary acute lateral patellar
dislocation and their recovery within the first 6 months of injury. Am J
Sports Med. 2000;28(4):472-479.
2. Bitar A, Demange M, D’Elia C, Camanho G. Traumatic patellar dislo-
cation: nonoperative treatment compared with MPFL reconstruction
using patellar tendon. Am J Sports Med. 2011;40(1):114-122.
3. Buckens CFM, Saris DBF. Reconstruction of the medial patellofe-
moral ligament for treatment of patellofemoral instability: a systematic
review. Am J Sports Med. 2010;38(1):181-188.
4. Camp C, Heidenreich M, Dahm D, Stuart M, Levy B, Krych A. Individ-
ualizing the tibial tubercle-trochlear groove distance patellar instabil-
ity ratios that predict recurrent instability. Am J Sports Med. 2016;
5. Carney JR, Mologne TS, Muldoon M, Cox JS. Long-term evaluation
of the Roux-Elmslie-Trillat procedure for patellar instability: a 26-
year follow-up. Am J Sports Med. 2005;33(8):1220-1223.
6. Christiansen SE, Jakobsen BW, Lund B, Lind M. Isolated repair of the
medial patellofemoral ligament in primary dislocation of the patella:
a prospective randomized study. Arthroscopy. 2008;24(8):881-887.
7. Crossley KM, Bennell KL, Cowan SM, Green S. Analysis of outcome
measures for persons with patellofemoral pain: which are reliable and
valid? Arch Phys Med Rehabil. 2004;85(5):815-822.
8. Dejour H, Walch G, Nove-Josserand L, Guier C. Factors of patella
instability: an anatomic radiographic study. Knee Surg Sports Trau-
matol Arthrosc. 1994;2(1):19-26.
9. Delgado-Martinez A, Rodriguez-Merchan E, Bellesteros R, Luna J.
Reproducibility of patellofemoral CT scan measurements. Int Orthop.
10. Fithian D, Neyret P, Servien E. Patellar instability: the Lyon experi-
ence. Tech Knee Surg. 2007;6(2):112-123.
11. Inoue M, Shino K, Hirose H, Horibe S, Ono K. Subluxation of the
patella: computed tomography analysis of patella-femoral congru-
ence. J Bone Joint Surg Am. 1988;70:1331-1337.
12. Kujala UM, Jaakkola LH, Koskinen SK, Taimela S, Hurme M,
Nelimarkka O. Scoring of patellofemoral disorders. Arthroscopy.
13. Kumar A, Jones S, Bickerstaff DR, Smith TW. Functional evaluation
of the modified Elmslie-Trillat procedure for patello-femoral dysfunc-
tion. Knee. 2001;8(4):287-292.
14. Lind M, Jakobsen BW, Lund B, Christiansen SE. Reconstruction of
the medial patellofemoral ligament for treatment of patellar instability.
Acta Orthop. 2008;79(3):354-360.
15. Mackay N, Smith N, Parsons N, Spalding T, Thompson P, Sprowson
A. Medial patellofemoral ligament reconstruction for patellar disloca-
tion: a systematic review. Orthop J Sports Med. 2014;2(8):232596711
16. Maenpaa H, Huhtala H. Recurrence after patellar dislocation. Acta
Orthop Scand. 1997;68(5):424-426.
¨H, Lehto MU. Patellar dislocation: the long-term results of
nonoperative management in 100 patients. Am J Sports Med.
18. Nakagawa K, Wada Y, Minamide M, Tsuchiya A, Moriya H. Deterio-
ration of long-term clinical results after the Elmslie-Trillat procedure for
dislocation of the patella. J Bone Joint Surg Br. 2002;84(6):861-864.
19. Pidoriano AJ, Weinstein RN, Buuck DA, Fulkerson JP. Correlation of
patellar articular lesions with results from anteromedial tibial tubercle
transfer. Am J Sports Med. 1997;25(4):533-537.
¨ttle PB, Schmeling A, Rosenstiel N, Weiler A. Radiographic
landmarks for femoral tunnel placement in medial patellofemoral lig-
ament reconstruction. Am J Sports Med. 2007;35(5):801-804.
21. Shabshin N, Schweitzer ME, Morrison WB, Parker L. MRI criteria for
patella alta and baja. Skeletal Radiol. 2004;33(8):445-450.
22. Sillanpaa P, Matilla V, Visuri T, Maenpaa H, Pihlajamaki H. Ligament
reconstruction versus distal realignemnt for patella dislocation. Clin
Orthop Relat Res. 2008;466(6):1475-1484.
23. Smith TO, Walker J, Russell N. Outcomes of medial patellofemoral
ligament reconstruction for patellar instability: a systematic review.
Knee Surg Sports Traumatol Arthrosc. 2007;15(11):1301-1314.
24. Spindler KP. Reconstruction of the medial patellofemoral ligament
was effective for traumatic patellar dislocation. J Bone Joint Surg
25. Stephen J, Dodds A, Lumpaopong P, Kader D, Williams A, Amis A.
The ability of medial patellofemoral ligament reconstruction to cor-
rect patellar kinematics and contact mechanics in the presence of
a lateralised tubercle. Am J Sports Med. 2015;43(9):2198-2207.
26. Tegner Y, Lysholm J. Rating systems in the evaluation of knee liga-
ment injuries. Clin Orthop Relat Res. 1985;198:43-49.
27. Thaunat M, Erasmus PJ. Recurrent patellar dislocation after medial
patellofemoral ligament reconstruction. Knee Surg Sports Traumatol
28. Trillat A, Dejour H, Couette A. [Diagnosis and treatment of recurrent
dislocations of the patella]. Rev Chir Orthop Reparatrice Appar Mot.
29. Watanabe T, Muneta T, Ikeda H, Tateishi T, Sekiya I. Visual analog
scale assessment after medial patellofemoral ligament reconstruction:
with or without tibial tubercle transfer. J Orthop Sci. 2008; 13(1):32-38.
For reprints and permission queries, please visit SAGE’s Web site at http://www.sagepub.com/journalsPermissions.nav.
8Damasena et al The American Journal of Sports Medicine