Content uploaded by Daniel Schwartz
Author content
All content in this area was uploaded by Daniel Schwartz on Feb 25, 2020
Content may be subject to copyright.
The arthroscopic Latarjet: a multisurgeon
learning curve analysis
Epaminondas M. Valsamis, MB BChir, MA(Cantab), MRCS, PGCert Med Ed
a,
*,
Jean Kany, MD
b
, Nicolas Bonnevialle, MD, PhD
c
, Roberto Castricini, MD
d
,
Alexandre L€
adermann, MD
e,f
, Gregory Cunningham, MD
e,g
, Daniel G. Schwartz, MD
h
,
George S. Athwal, MD, FRCSC
i
,
Joideep Phadnis, FRCS(Tr&Orth), Dip Sports Med, MBChB
j,k
a
Nuffield Orthopaedic Centre, Oxford University Hospitals NHS Foundation Trust, Oxford, UK
b
Shoulder Department, Clinique de l’Union, Saint Jean, France
c
Trauma and Orthopaedics Department, Centre Hospitalier Universitaire de Toulouse, Toulouse, France
d
Department of Orthopaedic and Trauma Surgery, Maria Cecilia Hospital, GVM Care and Research, Ravenna, Italy
e
Division of Orthopaedics and Trauma Surgery, Geneva University Hospitals, Geneva, Switzerland
f
Division of Orthopaedics and Trauma Surgery, La Tour Hospital, Meyrin, Switzerland
g
Shoulder Center, Hirslanden Clinique la Colline, Geneva, Switzerland
h
The Polyclinic, Seattle, WA, USA
i
St. Joseph’s Health Care, Western University, Roth McFarlane Hand and Upper Limb Center, London, ON, Canada
j
Brighton and Sussex Medical School, Brighton, UK
k
Trauma and Orthopaedics Department, Brighton and Sussex University Hospitals NHS Trust, Brighton, UK
Background: The open Latarjet procedure is a standard surgical treatment option for anterior shoulder
instability in patients with a high risk of failure following soft tissue stabilization. The arthroscopic tech-
nique has potential advantages of minimal invasiveness, reduced postoperative stiffness, and faster reha-
bilitation but is regarded as technically challenging with concern over surgical risk during the learning
curve. The aim of this study was to undertake a multisurgeon, large-volume learning curve analysis of
the arthroscopic Latarjet procedure using continuous learning curve analysis.
Methods: Individual patient data from 12 surgeons across 5 countries were retrospectively reviewed. A
total of 573 patients undergoing the arthroscopic Latarjet procedure were included. Outcome measures of
learning were collected, including operative time, computed tomography (CT) bone-block positioning,
complications, and patient-reported outcome measures (PROMs). A segmented linear regression
modeling technique was used for learning curve analysis.
Results: High-volume surgeons converged to an operative time steady state after 30-50 cases. Surgeons
completing fewer procedures demonstrated a constant reduction in operative time without reaching a
plateau. Low-volume surgeons completing fewer than 14 operations did not demonstrate a reduction
in operative time. Accuracy of bone-block positioning on postoperative CT demonstrated constant
Institutional Review Board approval was not required for this methodology
study.
*Reprint requests: Epaminondas M. Valsamis, MB BChir, MA(C-
antab), MRCS, PGCert Med Ed, Nuffield Orthopaedic Centre, Oxford
University Hospitals NHS Foundation Trust, Oxford OX3 7LD, UK.
E-mail address: markosvalsamis@gmail.com (E.M. Valsamis).
J Shoulder Elbow Surg (2019) -,1–8
www.elsevier.com/locate/ymse
1058-2746/$ - see front matter Ó2019 Journal of Shoulder and Elbow Surgery Board of Trustees. All rights reserved.
https://doi.org/10.1016/j.jse.2019.10.022
improvement without reaching a plateau after 53 cases. There was no change in PROMs or complications
with increased operative volume.
Conclusion: Specialist shoulder surgeons require 30-50 arthroscopic Latarjet procedures to attain
steady-state operative efficiency, during which there is improvement in bone-block positioning. Only sur-
geons expecting to undertake the arthroscopic Latarjet in high volume should consider adopting this
procedure.
Level of evidence: Educational Methodology Study; Learning Curve
Ó2019 Journal of Shoulder and Elbow Surgery Board of Trustees. All rights reserved.
Keywords: Shoulder surgery; arthroscopic Latarjet; learning curve; anterior instability; coracoid process
transfer; surgeon experience
The Latarjet procedure is an established surgical treat-
ment option for anterior shoulder instability with bone loss
or other risk factors for failure following soft tissue
stabilization.
15,20
The procedure has classically been an
open technique with excellent long-term functional results,
low rates of recurrent instability, and high rates of return to
sport.
23
A drawback to the Latarjet technique is its more
serious complication profile including nerve injury,
nonunion, implant failure, and glenohumeral arthritis to
name a few.
11,17,22
The expansion of arthroscopic shoulder
surgery has led to some authors developing and switching
to an all-arthroscopic technique. The proponents of an
arthroscopic technique cite minimal invasiveness, ability to
assess and treat concurrent lesions, reduced postoperative
stiffness, and faster rehabilitation and return to sport as
potential advantages.
18
However, the arthroscopic tech-
nique is regarded as technically challenging, with concerns
raised regarding the potential for increased surgical risk and
less accurate graft positioning, particularly during the
learning curve.
7,10
A learning curve is a graphical representation of the
relationship between learning effort and learning
outcome.
28
Several statistical methods have been used to
analyze learning curves in surgical practice, but these have
been mainly descriptive and have lacked mathematical
rigor.
25
The most common technique is the ‘‘split-group’’
method: data are chronologically split into 2 or more
consecutive groups, and the means of the groups are
compared using ttests (or equivalent tests). This technique
is prone to bias, including but not limited to the arbitrary
selection of group size that may obscure change.
13
Ekhtiari
et al
9
found that all studies that have investigated the
learning curve of the arthroscopic Latarjet procedure to
date have used rudimentary statistical techniques (the
group-splitting method in particular) and called for future
studies using continuous case data to ‘‘characterize a true
learning curve and identify the number of cases that
represent the inflection point,’’ that is, the point at which a
surgeon reaches a steady-state level of performance.
2,4,7,8,12
The primary aim of the current study was to undertake a
multisurgeon, large-volume learning curve analysis of the
arthroscopic Latarjet procedure, employing a segmented
linear regression modeling technique to allow continuous
case data analysis. By comparing the fit of multiple
learning models, this method can reliably provide a simple,
quantitative description of the learning curve.
30
The num-
ber of cases required to attain proficiency in the arthro-
scopic Latarjet may thus be estimated. The secondary aim
was to compare the learning curve between multiple
surgeons and to identify trends that are applicable to those
considering adoption of the arthroscopic Latarjet.
Materials and methods
Surgeon and patient data
Individual patient data from 12 centers across 5 countries (Can-
ada, USA, Italy, France, and Switzerland) were retrospectively
reviewed. The arthroscopic Latarjet procedure was performed in
573 patients by 12 shoulder surgeons between 2008 and 2019. The
first and all subsequent consecutive procedures for all surgeons
were included. None of the surgeons had prior real patient clinical
experience using the arthroscopic Latarjet. All surgeons were
fellowship-trained shoulder or sports surgeons performing the
open technique before their transition to the arthroscopic tech-
nique. The median number of previous open Latarjet procedures
per surgeon was 100 (range, 10-500). Each surgeon is identified
here on with a letter from Ato L.
Patients were selected for the Latarjet procedure based on in-
dividual factors, including age, type of sport, level of sport, bone
loss, and previous instability episodes.
5,19
Institutional research board and ethical committee approval
was undertaken for all data collection in the individual centers.
Surgical technique
There was some variation in the exact surgical technique between
surgeons but not within the same series of patients collected by
each surgeon. Two surgeons (surgeons Aand B) used the original
arthroscopic Latarjet technique as described by Lafosse
et al.
19
Five surgeons (surgeons C,D,E,F, and G) used a modified
Lafosse technique.
2
Surgeon H further modified the Lafosse
technique using 5 portals instead of 7 (posterior, anterolateral,
anterior, suicide and superior to access the superior
coracoid).
8
Four surgeons (surgeons I,J,K, and L) used a double-
button fixation method and a guided approach to transfer the
coracoid through the subscapularis muscle as described by
2 E.M. Valsamis et al.
Boileau.
3,4
Dedicated instruments (Latarjet Guiding System;
Smith & Nephew Inc, Andover, MA, USA; or Latarjet Disposable
Kit, DePuy Mitek, Raynham, MA, USA) were used by all
surgeons.
Outcome measures
The primary outcome measure was operative time, defined as the
time from incision to skin closure, and was recorded by all 12
surgeons. The operative time for each of the 5 successive steps
(joint evaluation and exposure; subscapularis split; coracoid graft
harvesting; graft transfer; graft fixation) of the procedure was
recorded by surgeon B.
Secondary outcome measures included bone-block positioning,
complications, and patient-reported outcome measures (PROMs).
Postoperative computed tomography (CT) scanning to assess graft
positioning was undertaken by 5 surgeons for a total of 201 cases
(surgeons A,I,J,K, and L).
4,12
CT scan analysis was performed
with OsiriX imaging software (Pixmeo, Geneva, Switzerland),
allowing for multiplanar reconstruction from the original
data.
26
Surgeon Aevaluated bone-block positioning in accordance
with Burkhart et al.
6
Four surgeons (surgeons I, J, K, and
L) evaluated the position of the bone-block in accordance with
Kraus et al.
14
The same technique for CT analysis was used for all
cases within a surgeon’s data set.
Six surgeons (surgeons C,D,E,F,G, and H) collected
complication data, including but not limited to graft fracture,
nonunion, nerve and vessel injury, subscapularis rupture, early
recurrent instability, hematoma, infection, and metalwork-related
complications.
Surgeons B(n ¼30) and H(n ¼28) collected PROMs data
preoperatively and postoperatively using either the Rowe score
27
or Walch-Duplay scores.
33
Surgeon Bevaluated postoperative
patient satisfaction using a Likert-type scale.
7
Segmented linear regression technique
A validated segmented linear regression modeling method was
employed as previously described by Valsamis et al.
30,31
This
technique uses progressively more complex models consisting of
combinations of linear segments with variable adjoining points.
The models consist of the following:
1. Plateau: This is a single average value of the data set. A
‘‘plateau’’ indicates no learning has taken place.
2. Line: This is a single line of nonzero gradient. A ‘‘line’’ in-
dicates that learning is taking place at a constant rate.
3. Line-plateau: This is a line that is followed by an adjoining
plateau. A ‘‘line-plateau’’ indicates that learning is taking place
at a constant rate, but that a steady state is subsequently
achieved.
4. Line-line (a line followed by an adjoining line): This is a line
that is followed by an adjoining line of different slope. A ‘‘line-
line’’ indicates that learning is initially taking place at a con-
stant rate up until a point where the rate changes. This accounts
for the possibility that the rate of learning changes but does not
reach a steady state as in the ‘‘line-plateau’’ model.
5. Line-line-plateau: This is a line-line that is subsequently fol-
lowed by a plateau. It may identify more complex trends that
demonstrate more than 2 phases of learning.
6. Line-line-line: This is a line-line that is subsequently followed
by a line. It may identify more complex trends that demonstrate
more than 2 phases of learning.
To select the best model, all models are fitted to the data and F
tests are mutually conducted between models on the basis they are
nested models (Fig. 1).
1
The model that prevails is the simplest
one, unless a more complex model offers a significantly better fit,
as tested and confirmed by Pvalues through a tabular method
(Table I). We assume a significant Pvalue when P<.05. A
dedicated MatLab program (Mathworks Inc, Natick, MA) was
used for this analysis.
Results
Operative time
Surgeon Awho completed 288 cases demonstrated a line-
line-plateau in operative time with experience. There was
a decrease in operative time of 1.2 minutes per case for the
first 43 cases, followed by a decrease of 0.17 minutes per
case for the next 91 cases, after which a plateau of 62.5
minutes was attained (Fig. 2).
The line-plateau model fit best for 2 surgeons
completing 30 (surgeon B) and 50 (surgeon I) cases. The
operative time for surgeon Bdecreased by 4.2 minutes per
case for 16 cases, reaching a plateau at 99 minutes (Fig. 3).
The operative time for surgeon Idecreased by 1.8 minutes
per operation for 28 cases, reaching a plateau at 78 minutes
(Fig. 4).
The line-line model fit best for 2 surgeons completing 50
(surgeon C) and 16 (surgeon F) cases. For surgeon C, the
operative time decreased by 3.2 minutes per case for the
first 17 cases, followed by a decrease of 0.7 minutes per
case for the next 33 cases (Fig. 5). For surgeon F, the
operative time decreased by 9.3 minutes per case for the
first 8 cases, followed by a decrease of 1.7 minutes per case
for the next 8 cases.
The line model fit best for 5 surgeons completing 28
(surgeon H), 26 (surgeon E), 24 (surgeon J), 24 (surgeon
D), and 14 (surgeon G) cases. The rate of decrease of
operative time was 4.7, 1.4, 2.5, 1.8, and 5.6 minutes per
case, respectively, without reaching a plateau.
Two surgeons who undertook 10 (surgeon L) and 12
(surgeon K) operations each showed no evidence of
learning. The plateau model fit best at 97.5 and 150 mi-
nutes, respectively.
Analysis of the operative time for the 5 individual steps
of the arthroscopic Latarjet procedure for surgeon B(n ¼
30) demonstrated the following results:
Step 1 (joint evaluation and exposure): The line-line
model fit best with a reduction in operative time of 7.4
minutes per operation for the first 4 cases and 0.3 mi-
nutes per operation for the next 26.
Step 2 (subscapularis split): The line model fit best with
a constant reduction in operative time of 0.7 minutes per
operation.
Arthroscopic Latarjet learning curve 3
Step 3 (coracoid graft harvesting): The line model fit
best with a constant reduction in operative time of 0.6
minutes per operation.
Step 4 (graft transfer): The plateau model fit best with no
change demonstrated.
Step 5 (graft fixation): The line model fit best with a
constant reduction in operative time of 0.3 minutes per
operation.
CT-guided bone-block positioning
From surgeon A’s data, the accuracy of bone block posi-
tioning in the sagittal/oblique and coronal/oblique planes
and joint violation by the screw showed no change with
experience over 89 cases, the plateau model fitting best for
all measures.
The rate of accurate subequatorial coronal posi-
tioning of the bone-block improved for 4 surgeons at
0.004 per case for 53 cases (0.4% improvement per
case).
Complications
The rate of complications was unchanged with
experience for 6 surgeons, the plateau model
fitting best for all surgeons. The overall
Figure 1 All models fit to the same data (surgeon A), after which the best model is chosen using the tabular method (see Table I). Plateau:
y¼71.06. Line: y¼0.1373xþ90.90. Line-plateau: line of y¼–1.165xþ129.1, plateau of y¼65.05 joining at k¼55. Line-line: 1st
line of y¼1.286xþ131.2, 2nd line of y¼0.0548(x55) þ72.04 joining at k¼55. Line-line-plateau: 1st line of y¼1.210xþ
130.1, 2nd line of y¼0.1702(x43) þ78.02 joining at k¼43, plateau of y¼62.53 joining at k¼134. Line-line-line: 1st line of y¼
1.211xþ130.1, 2nd line of y¼0.1691(x43) þ77.99 joining at k¼43, 3rd line of y¼0.0011(x134) þ62.6 joining at k¼134
(virtually indistinguishable from line-line-plateau).
Table I Tabular method using Pvalues of mutually conducted Ftests between models: Example using operative time data from
surgeon A
Plateau Line Line-plateau Line-line Line-line-plateau
Line <.0001
Line-plateau <.0001 <.0001
Line-line <.0001 <.0001 .0003
Line-line-plateau <.0001 <.0001 <.0001 .0035
Line-line-line <.0001 <.0001 <.0001 .0140 .961
The best model is found as that model whose row featuring a contiguous set of significant Pvalues extends furthest to the right. Significant Pvalues are
shown in bold. When 2 such rows tie (as in this example), select the simplest model. In this example, the line-line-plateau is the selected model (shown
in italic).
4 E.M. Valsamis et al.
complication rate for the 242 cases included was
12.4% (Tab l e I I).
PROMs
There was no change in the preoperative to postoperative
improvement in Walch-Duplay scores or Rowe scores, or
patient satisfaction with experience. The plateau model fit
best for all surgeons.
Discussion
This multicenter, large-volume study investigating the
learning curve of the arthroscopic Latarjet using segmented
linear regression identified that between 30 and 50 cases are
required to reach a steady state in operative time for
experienced shoulder surgeons. There appeared to be an
improvement in the accuracy of bone-block positioning
with experience, although no plateau was reached after 53
cases. During this period, there was no significant evidence
Figure 2 Learning curve of operative time in minutes for sur-
geon Acompleting 288 consecutive arthroscopic Latarjet pro-
cedures. The line-line-plateau model fit best. y¼1.210xþ
130.1 for the first 43 cases, y¼–0.1702(x43) þ78.02 for the
next 91 cases, and y¼62.53 for the plateau after 134 cases.
Figure 3 Learning curve of operative time in minutes for sur-
geon Bcompleting 30 consecutive arthroscopic Latarjet proced-
ures. The line-plateau model fit best. y¼4.179xþ165.9 for the
first 16 cases and y¼99.05 for the plateau.
Figure 4 Learning curve of operative time in minutes for a
surgeon Icompleting 50 consecutive arthroscopic Latarjet pro-
cedures. The line-plateau model fit best. y¼1.758xþ127.1 for
the first 28 cases and y¼77.83 for the plateau.
Figure 5 Learning curve of operative time in minutes for sur-
geon Ccompleting 50 consecutive arthroscopic Latarjet proced-
ures. The line-line model fit best. y¼3.236xþ142.6 for the
first 17 cases and y¼0.725xþ97.32 for the next 33 cases.
Arthroscopic Latarjet learning curve 5
of learning based on PROMs and complications, indicating
that surgeons gained operative efficiency in completing safe
procedures.
When analyzing surgical learning curves, operative time
is the most commonly investigated learning outcome
measure and provides some information about surgical
performance.
24
In isolation, however, it is a weak proxy of
learning, and therefore we analyzed several other secondary
learning outcomes to investigate the relationships between
learning outcomes and to draw conclusions that are as
robust as possible.
Operative time
Previous studies suggested that 15,
7
20,
8,21
or 30
4
cases are
required to attain an operative time plateau, whereas our
study found that this number may be substantially greater.
The number of consecutive cases for the highest-volume
surgeon in this study (n ¼288) is considerably greater than
those included in other studies (n ¼104). Indeed, surgeon
A, who undertook the greatest number of procedures (n ¼
288), completed a rapid learning phase after 43 cases but
continued to demonstrate improvement at a much slower
rate to attain a plateau after 134 cases. Surgeon B, who
reached a plateau of 99 minutes after 16 cases, reports that
after completing more than 100 cases (a datum that was not
available for this study) his operative time varied between
70 and 80 minutes. This suggests that if data from all his
cases were available, he may have reached a lower plateau
much later. Surgeon Cmay not have attained a plateau after
50 cases but demonstrated a decline in the rate of learning
after 17 cases, suggesting an impending plateau.
Analysis of the operative time for the 5 individual steps
of the procedure could be undertaken for only 1 surgeon
(surgeon B), and suggested that apart from graft transfer,
the operative time for all steps shortens with surgeon
experience. Joint evaluation and exposure demonstrated the
greatest improvement (7.4 minutes per operation) over the
first few cases and appears to be the rate-limiting step for
the procedure early in the learning curve. The subscapularis
split had the second greatest rate of reduction in operative
time of 0.7 minutes per operation. An adequate surgical
approach is indeed both demanding and important to allow
subsequent accurate bone-block positioning.
16
Our finding that surgeons who completed fewer than 14
cases showed no learning may have been due to the large
variability in operative time between cases, meaning that
more cases are required to confidently conclude a
decreasing trend. The other explanation is that a low
operative frequency does not allow for learning over a
small number of operative cases.
Secondary outcome measures
Bonnevialle et al
4
found that the accuracy of subequatorial
bone-block positioning improved after 30 cases, and the
current analysis confirmed a steady improvement of 0.4%
per case for all 53 cases. This finding suggests that accu-
racy of bone-block positioning may continue to improve
after operative time reaches a plateau, and a greater number
of consecutive data are required to determine the level of
experience where positioning plateaus. However, this
improvement in bone-block positioning is small and,
although statistically significant, its effect on clinical
outcome is not known.
We did not find a significant effect of surgeon experi-
ence on complication rate. However, the overall rate of
complications was 12.4%, meaning a change in this
outcome measure may have been undetectable with our
sample size. Reassuringly however, the rates are low
enough to not be a major concern for surgeons adopting this
procedure. In agreement with Castricini et al,
7
we found no
change in PROMs with surgical experience.
Learning curve methodology
All previous studies
2,4,7,8,12
used the group-splitting method
to analyze the learning curve of the arthroscopic Latarjet.
This method splits the data chronologically into arbitrarily
sized groups that are then compared using ttests. Although
popular, the technique does not provide any information
about the shape of the learning curve or the precise number
of procedures required to attain a steady-state
plateau.
28
Bonnevialle et al
4
also used a Spearman corre-
lation to analyze the learning curve. Although this does add
an element of continuity to the analysis, it does not test the
significance of the correlation against other learning models
or allow for a change in the rate of improvement to be
detected.
Our method for learning curve analysis uses segmented
linear regression techniques testing multiple different
Table II Complications
Complication Number (%) of patients
Graft fracture 10 (4.1)
Early recurrent instability 4 (1.7)
Hardware removal surgery 4 (1.7)
Screw backout/bending/failure 3 (1.2)
Delayed failure 2 (0.8)
Nerve injury (transient) 2 (0.8)
Excessive fluid extravasation 1 (0.4)
Infection 1 (0.4)
Mal-positioned screw 1 (0.4)
Nonunion 1 (0.4)
Vascular injury 1 (0.4)
Nerve injury (permanent) 0 (0.0)
Complications for the 242 cases where complications were recorded
and subsequently analyzed. The overall rate of complications was
12.4%.
6 E.M. Valsamis et al.
learning models. This allows the best model to be chosen,
an inflection point to be identified if present, and an ac-
curate, robust conclusion to be drawn about the learning
curve of the procedure. Segmented linear regression has
been previously used to evaluate the learning curves of
imageless navigation total hip and total knee re-
placements
30,32
and to evaluate change in retrospective
studies.
31,29
Furthermore, although other studies have grouped sur-
geon data in their analysis,
2,4
we analyzed each surgeon’s
data individually, allowing trends for each surgeon to be
revealed. Analyzing data from several surgeons allows the
identification and comparison of individual trends between
surgeons working at different centers, allowing a pragmatic
description of learning.
Limitations
Although the multicenter design of this study allows a
pragmatic conclusion to be drawn about the learning curve
of the procedure, it does present a limitation. The different
clinical setting in each center may mean that the absolute
operative times are less comparable from surgeon to sur-
geon, and the learning curve may have been influenced by
the operating room team’s expertise and other external
factors. Therefore, when comparing surgeons, we placed
greater emphasis on the model choice and timing of a
plateau as opposed to absolute operative time values. Data
collection techniques including the choice of
PROM instruments varied between centers, although these
were consistent throughout the study period within each
center. Not all surgeons undertook the same number of
procedures, and the frequency of procedures also varied.
This may have contributed to the variable rate of decrease
in operative time observed between surgeons. Finally, there
was some variation in the exact surgical technique used by
each surgeon, and although the main steps of the procedure
remained identical, these differences may have influenced
the data.
Recommendations
Surgeons must be aware that for experienced, specialist
shoulder surgeons with prior experience of the open
Latarjet technique, 30-50 cases are required to achieve a
steady state in operative efficiency. This has implications
for patients in the early phase of a surgeon’s learning curve
who are exposed to longer operative times. Therefore, we
feel that only surgeons expecting to undertake the arthro-
scopic Latarjet in high volume should consider learning this
procedure, and that measures such as cadaveric training,
joint-consultant operating, and surgical visitations should
be mandatory prior to adoption. During training, there
should be an emphasis on mastering joint exposure and
graft placement, which were found to be the rate-limiting
steps and main technical factors for improvement.
Conclusions
Specialist shoulder surgeons should expect to complete
30-50 arthroscopic Latarjet procedures to attain steady-
state operative efficiency. Shoulder surgeons must be
aware of this considerable learning curve before
deciding to introduce this demanding arthroscopic
technique into their practice.
Acknowledgments
The authors would like to acknowledge Dr. Charlie
Getz, Dr. Robert Meislin, Dr. Paul Favorito, Dr. David
Weinstein and Dr. Charles Edouard Thelu for providing
operative data.
Disclaimer
Nicolas Bonnevialle reports personal fees from Smith &
Nephew, during the conduct of the study.
Alexandre L€
adermann reports royalties from Wright
and consultancy fees from Arthrex, Wright, and
Medacta.
George S. Athwal reports other from DePuy Mitek,
other from ConMed Linvatec, outside the submitted
work; in addition, Dr. Athwal has a patent Coracoid drill
guide and method of use pending.
The other authors, their immediate families, and any
research foundations with which they are affiliated have
not received any financial payments or other benefits
from any commercial entity related to the subject of this
article.
References
1. Alexopoulos EC. Introduction to multivariate regression analysis.
Hippokratia 2010;14:23-8.
2. Athwal GS, Meislin R, Getz C, Weinstein D, Favorito P. Short-term
complications of the arthroscopic Latarjet procedure: a North Amer-
ican experience. Arthroscopy 2016;32:1965-70. https://doi.org/10.
1016/j.arthro.2016.02.022
3. Boileau P, Gendre P, Baba M, Th
elu C, Baring T, Gonzalez J, et al. A
guided surgical approach and novel fixation method for arthroscopic
Latarjet. J Shoulder Elbow Surg 2016;25:78-89. https://doi.org/10.
1016/j.jse.2015.06.001
4. Bonnevialle N, Th
elu CE, Bouju Y, Vogels J, Agout C, Duriez P, et al.
Arthroscopic Latarjet procedure with double-button fixation: short-
term complications and learning curve analysis. J Shoulder Elbow
Surg 2018;27:e189-95. https://doi.org/10.1016/j.jse.2017.12.007
Arthroscopic Latarjet learning curve 7
5. Burkhart SS, De Beer JF. Traumatic glenohumeral bone defects and
their relationship to failure of arthroscopic Bankart repairs: signifi-
cance of the inverted-pear glenoid and the humeral engaging Hill-
Sachs lesion. Arthroscopy 2000;16:677-94.
6. Burkhart SS, De Beer JF, Barth JRH, Criswell T, Roberts C,
Richards DP. Results of modified Latarjet reconstruction in patients
with anteroinferior instability and significant bone loss. Arthroscopy
2007;23:1033-41. https://doi.org/10.1016/j.arthro.2007.08.009
7. Castricini R, De Benedetto M, Orlando N, Rocchi M, Zini R, Pirani P.
Arthroscopic Latarjet procedure: analysis of the learning curve.
Musculoskelet Surg 2013;97(Suppl 1):93-8. https://doi.org/10.1007/
s12306-013-0262-3
8. Cunningham G, Benchouk S, Kherad O, L€
adermann A. Comparison of
arthroscopic and open Latarjet with a learning curve analysis. Knee
Surg Sport Traumatol Arthrosc 2016;24:540-5. https://doi.org/10.
1007/s00167-015-3910-3
9. Ekhtiari S, Horner NS, Bedi A, Ayeni OR, Khan M. The learning
curve for the Latarjet procedure: a systematic review. Orthop J Sport
Med 2018;6. 2325967118786930. https://doi.org/10.1177/2325967118
786930.
10. Gracitelli MEC, Ferreira AA, Benegas E, Malavolta EA, Sunada EE,
Assunc¸~
ao JH. Arthroscopic Latarjet procedure: safety evaluation in
cadavers. Acta Ortop Bras 2013;21:139-43. https://doi.org/10.1590/
S1413-78522013000300002
11. Griesser MJ, Harris JD, McCoy BW, Hussain WM, Jones MH,
Bishop JY, et al. Complications and re-operations after Bristow-
Latarjet shoulder stabilization: A systematic review. J Shoulder
Elbow Surg 2013;22:286-92. https://doi.org/10.1016/j.jse.2012.09.009
12. Kany J, Flamand O, Grimberg J, Guinand R, Croutzet P,
Amaravathi R, et al. Arthroscopic Latarjet procedure: is optimal
positioning of the bone block and screws possible? A prospective
computed tomography scan analysis. J Shoulder Elbow Surg 2016;25:
69-77. https://doi.org/10.1016/j.jse.2015.06.010
13. Khan N, Abboudi H, Khan MS, Dasgupta P, Ahmed K. Measuring the
surgical ‘‘learning curve’’: methods, variables and competency. BJU
Int 2014;113:504-8. https://doi.org/10.1111/bju.12197
14. Kraus TM, Graveleau N, Bohu Y, Pansard E, Klouche S, Hardy P.
Coracoid graft positioning in the Latarjet procedure. Knee Surg Sport
Traumatol Arthrosc 2016;24:496-501. https://doi.org/10.1007/s00167-
013-2651-4
15. L€
adermann A. Editorial commentary: Arthroscopic Latarjet shoulder
stabilization: Where are we? Where are we going? Arthroscopy 2017;
33:2136-8. https://doi.org/10.1016/j.arthro.2017.08.277
16. L€
adermann A, Denard PJ, Arrigoni P, Narbona P, Burkhart SS, Barth J.
Level of the subscapularis split during arthroscopic Latarjet.
Arthroscopy 2017;33:2120-4. https://doi.org/10.1016/j.arthro.2017.06.
013
17. L€
adermann A, Lubbeke A, Stern R, Cunningham G, Bellotti V,
Gazielly DF. Risk factors for dislocation arthropathy after Latarjet
procedure: a long-term study. Int Orthop 2013;37:1093-8. https://doi.
org/10.1007/s00264-013-1848-y
18. Lafosse L, Boyle S. Arthroscopic Latarjet procedure. J Shoulder
Elbow Surg 2010;19(Suppl 2):2-12. https://doi.org/10.1016/j.jse.2009.
12.010
19. Lafosse L, Lejeune E, Bouchard A, Kakuda C, Gobezie R, Kochhar T.
The arthroscopic Latarjet procedure for the treatment of anterior
shoulder instability. Arthroscopy 2007;23:1242.e1-e5. https://doi.org/
10.1016/j.arthro.2007.06.008
20. Latarjet M. Treatment of recurrent dislocation of the shoulder. Lyon
Chir 1954;49:994-7.
21. Leuzinger J, Brzoska R, M
etais P, Clavert P, Nourissat G, Walch G,
et al. Learning Curves in the Arthroscopic Latarjet Procedure: A
Multicenter Analysis of the First 25 Cases of 5 International Surgeons.
Arthroscopy 2019;35:2304-11. https://doi.org/10.1016/j.arthro.2019.
03.035
22. Longo UG, Loppini M, Rizzello G, Ciuffreda M, Berton A,
Maffulli N, et al. Remplissage, humeral osteochondral grafts, Weber
osteotomy, and shoulder arthroplasty for the management of humeral
bone defects in shoulder instability: systematic review and quantitative
synthesis of the literature. Arthroscopy 2014;30:1650-66. https://doi.
org/10.1016/j.arthro.2014.06.010
23. Mizuno N, Denard PJ, Raiss P, Melis B, Walch G. Long-term results of
the Latarjet procedure for anterior instability of the shoulder. J
Shoulder Elbow Surg 2014;23:1691-9. https://doi.org/10.1016/j.jse.
2014.02.015
24. Ramsay CR, Grant AM, Wallace SA, Garthwaite PH, Monk AF,
Russell IT. Assessment of the learning curve in health technologies. A
systematic review. Int J Technol Assess Health Care 2000;16:1095-8.
https://doi.org/10.1017/s0266462300103149
25. Ramsay CR, Grant AM, Wallace SA, Garthwaite PH, Monk AF,
Russell IT. Statistical assessment of the learning curves of health
technologies. Health Technol Assess 2001;5:1-79.
26. Rosset A, Spadola L, Ratib O. OsiriX: an open-source software for
navigating in multidimensional DICOM images. J Digit Imaging
2004;17:205-16. https://doi.org/10.1007/s10278-004-1014-6
27. Rowe CR, Patel D, Southmayd WW. The Bankart procedure. A long-
term end-result study. J Bone Joint Surg Am 1978;60:1-16.
28. Valsamis EM, Chouari T, O’Dowd-Booth C, Rogers B, Ricketts D.
Learning curves in surgery: Variables, analysis and applications.
Postgrad Med J 2018;94:525-30. https://doi.org/10.1136/post-
gradmedj-2018-135880
29. Valsamis EM, Husband H, Chan GK. Segmented Linear Regression
Modelling of Time-Series of Binary Variables in Healthcare. Comput
Math Methods Med 2019:1-7. https://doi.org/10.1155/2019/3478598
30. Valsamis EM, Golubic R, Glover TE, Husband H, Hussain A,
Jenabzadeh AR. Modeling learning in surgical practice. J Surg Educ
2018;75:78-87. https://doi.org/10.1016/j.jsurg.2017.06.015
31. Valsamis E, Ricketts D, Husband H, Rogers B. Segmented linear
regression models for assessing change in retrospective studies in
healthcare. Comput Math Methods Med [Internet] 2019;2019:
9810675. https://doi.org/10.1155/2019/9810675
32. Valsamis EM, Ricketts D, Hussain A, Jenabzadeh AR. Imageless
navigation total hip arthroplasty – an evaluation of operative time.
SICOT J 2018;4:18. https://doi.org/10.1051/sicotj/2018016
33. Walch G. The Walch-Duplay score for instability of the shoulder. Di-
rections for the use of the quotation of anterior instabilities of the
shoulder. In: Abstract presented at: First Open Congress of the European
Society of Surgery of the Shoulder and Elbow; 1987. p. 51-5. Paris.
8 E.M. Valsamis et al.
