Keywords
arthroscopy - articular instability - shoulder - retrospective studies
Introduction
Anterior glenohumeral dislocation is the most common kind of shoulder dislocation.
It results from damage to the capsulolabral complex in the anteroinferior portion
of the glenoid, and is also referred to as Bankart lesion or essential lesion.[1]
[2]
[3]
[4]
Surgical repair, either open or arthroscopic, is indicated for recurrent dislocations.
Among the surgical treatments for shoulder stabilization, arthroscopic repair has
become increasingly popular due to reduced movement loss, milder surgical aggression,
subscapularis preservation and low morbidity compared to open surgery.[5]
Several patient-related factors contributing to the recurrence risk following the
Bankart arthroscopic repair have been reported, including younger age at surgery,
male gender, bilateral instability, ligamentous hyperlaxity, participation in collision
sports, and early return to contact sports.[6] Injury-associated risk factors, such as erosion or glenoid deficiency, Hill-Sachs
lesion size, and anterior glenoid border involvement were also implicated.[7] Accurate identification of the risk factors associated with Bankart arthroscopic
repair failure and capsular dislocation helps the surgeon stratify the risk for an
individual patient, enabling proper counseling.
Therefore, the present study aims to perform a functional assessment and to analyze
the factors influencing the outcome of patients with anterior shoulder instability
submitted to the arthroscopic treatment of anterior shoulder instability.
Materials and Methods
From March to May 2017, 94 patients were retrospectively analyzed through a review
of medical records and subsequent clinical evaluation. All patients underwent arthroscopic
surgical treatment for anterior shoulder instability in two private hospitals and
by four independent surgeons between January 2010 and December 2014. Ten patients
operated on during this period were excluded due to non-attendance at the follow-up.
The medical records had the following information: patient identification, clinical
history characterizing the cause of the instability (traumatic or atraumatic), functional
limitation, preoperative physical examination, preoperative imaging, and surgical
description. The postoperative follow-up time ranged from 24 to 72 months, with a
mean period of 40 months, to assess the functional limitation, decreased lateral rotation,
instability, pain, the and functional scores of Carter-Rowe, University of California
at Los Angeles (UCLA) and Constant-Murley through clinical and physical examinations
performed by an external physician.
The present study included patients with recurrent anterior shoulder instability (dislocation
or subluxation) and Bankart lesion, all operated arthroscopically and with a minimum
postoperative follow-up time of 24 months. The exclusion criteria were traumatic dislocation
associated with neurovascular injury, fracture on other sites of the shoulder girdle,
glenoid bone loss greater than 25%, Hill-Sachs fracture involving more than 1/4 of
the humeral head, previous surgeries at the involved shoulder, and multidirectional
instability.
Two surgeons performed the procedures with the patients in lateral position. The surgical
procedure was performed under general anesthesia and brachial plexus block. The limb
was kept at approximately 70° of abduction and 20° of flexion, and fixed and vertical
longitudinal traction with 4- to 7-kg weights was applied.
The other two surgeons performed the procedures with the patients in the beach chair
position, that is, with a trunk elevation of at least 70°, around 30° of knee flexion,
and gentle lateral inclination opposite to the operated side.
The surgeons used conventional arthroscopic portals (posterior, anterosuperior and
anteroinferior). During the procedure, the joint was inspected to assess the presence
of associated lesions. Subsequently, the anteriorinferior labral lesion and its extension
were observed, followed by capsulolabral release and glenoid border scarification;
next, the labral lesion was fixed with two to four bioabsorbable anchors.
Postoperatively, the patients remained with continuous immobilization on a Velpeau
sling for three weeks. Pendular motion and passive self-motion were then initiated
over the following two weeks; lateral rotation movements beyond 20° were allowed from
the 6th week onwards. Muscle strengthening was dependent on the range of motion of
the operated shoulder, and usually began in the third postoperative month. Return
to contact or collision sports activities was allowed from the sixth month onwards.
In the descriptive analysis, the observed data were presented as tables; the continuous
and quantitative variables were expressed as medians and interquartile ranges (IQRs),
whereas the categorical and qualitative variables were expressed as frequencies (n)
and percentages (%).
The inferential analysis employed the following methods:
-
the association of the Carter-Rowe, UCLA and Constant-Murley scores with the clinical
variables (age, gender, type of injury and surgical position) was assessed by the
Mann-Whitney test for categorical data and the Spearman correlation coefficient for
numerical data; and
-
the association of dislocation recurrence, residual pain, loss of lateral rotation
and apprehension with the clinical variables was analyzed by the Chi-squared (χ2) or Fisher exact test for categorical data and by the Mann-Whitney test for numerical
data.
A nonparametric method was applied because the variables under study did not present
normal (Gaussian) distribution due to the rejection of the normality hypothesis by
the Shapiro-Wilks test. The significance was determined as a 5% level. The statistical
analysis was performed using Statistical Analysis System (SAS, SAS Institute, Inc.,
Cary, North Carolina, US) software, version 6.11.
All patients signed the Informed Consent Form. The present study was approved by the
Committee of Ethics in Research with Human Beings under Opinion no. 2,197,472, CAAE
70807917.6.0000.0023.
Results
In total, 94 patients were evaluated in the present study, with a follow-up period
of 2 to 6 years after surgery; 74 subjects (78,7%) were male, and 20 (21,3%) were
female, with ages at the surgical procedure ranging from 17 to 62 years ([Table 1]). A total of 11 (11.7%) patients presented dislocation recurrence, 35 (37.23%) had
decreased lateral rotation, 48 (51.06%) reported some degree of residual pain, and
23 (24.46%) were positive at the apprehension test ([Table 2]).
Table 1
|
Variable
|
N
|
%
|
|
Gender
|
|
Male
|
74
|
78.7
|
|
Female
|
20
|
21.3
|
|
Age at surgery (years)
|
|
Median value (Q1–Q3)
|
34 (28–43)
|
|
Injury type
|
|
Traumatic injury
|
85
|
90.4
|
|
Atraumatic injury
|
9
|
9.6
|
|
Position
|
|
Beach chair
|
23
|
24.5
|
|
Lateral recumbency
|
71
|
75.5
|
|
Carter-Rowe score (points)
|
|
Median value (Q1–Q3)
|
95 (80–100)
|
|
University of California at Los Angeles (UCLA) score (points)
|
|
Median value (Q1–Q3)
|
33 (31–35)
|
|
Constant-Murley score (points)
|
|
Median value (Q1–Q3)
|
95 (87–100)
|
Table 2
|
Redislocation
|
N
|
%
|
|
Yes
|
11
|
11.7
|
|
No
|
83
|
88.3
|
|
Residual pain
|
|
Yes
|
48
|
51.1
|
|
No
|
46
|
48.9
|
|
External rotation loss
|
|
Yes
|
35
|
37.2
|
|
No
|
59
|
62.8
|
|
Apprehension
|
|
Yes
|
23
|
24.5
|
|
No
|
71
|
75.5
|
Traumatic instability was prevalent, representing 90.4% of cases (85 patients). The
most common surgical positioning was that of lateral decubitus, with 71 of the operated
patients (75.5%) of patients underwent surgery in this position. At the clinical evaluation,
the median Carter-Rowe score was of 95 ([Table 3]), with 67 excellent outcomes (71,3%), 14 good outcomes (14,9%), 2 regular outcomes
(2,1%) and 11 bad outcomes (11,7%) ([Figure 1]). Bad outcomes were associated with dislocation recurrence, and they occurred in
9 male patients (81,8%) and in 2 female patients (18,2%). Although more prevalent
in male patients, our study found no statistically significant difference between
gender and dislocation recurrence. Similarly, there was no significant correlation
between the Carter-Rowe score and age at the time of surgery (rs = 0.162; p = 0.011).
Fig. 1 Carter-Rowe score.
Table 3
|
Variable
|
n
|
Median value
|
Interquartile range
|
Minimum
|
Maximum
|
|
Age at surgery (years)
|
94
|
34
|
27.8
|
–
|
43
|
17
|
62
|
|
Score (points)
|
|
Carter-Rowe
|
94
|
95
|
80
|
–
|
100
|
15
|
100
|
|
University of California at Los Angeles (UCLA)
|
94
|
33
|
31
|
–
|
35
|
23
|
35
|
|
Constant-Murley
|
94
|
95
|
87
|
–
|
100
|
68
|
100
|
In the UCLA score evaluation, there were 83 patients (88,3%) with good/excellent outcomes
and 11 patients (11,7%) with bad/regular outcomes ([Figure 2]). There was a significant direct correlation between the UCLA score and the patient's
age at the time of surgery (rs = 0.250; p = 0.015). Therefore, the older the age, the higher the expected score.
Fig. 2 University of California at Los Angeles (UCLA) score.
In the Constant-Murley score evaluation, there were 65 patients (69,2%) with excellent
outcomes, 18 (19,1%) good outcomes, 10 (10,7%) regular outcomes and 1 poor outcome
(1%) ([Figure 3]). None of the three scores presented a statistically significant correlation with
gender, type of injury or surgical position.
Fig. 3 Constant-Murley score.
In addition, there was no statistically significant correlation between the clinical
variables (gender, age, type of injury and surgical position) and dislocation recurrence,
residual pain, loss of lateral rotation and apprehension ([Tables 4], [5], [6], [7]). The subgroup with residual pain, however, was significantly younger than the subgroup
without residual pain (p = 0.016).
Table 4
|
Variable
|
With redislocation
|
Without redislocation
|
p-value
|
|
Gender
|
|
Male
|
9
|
81.8
|
65
|
78.3
|
0.57
|
|
Female
|
2
|
18.2
|
18
|
21.7
|
|
Age at surgery (years)
|
|
Median value (Q1–Q3)
|
32 (23–38)
|
34 (28–43)
|
0.10
|
|
Injury type
|
|
Traumatic injury
|
11
|
100
|
74
|
89.2
|
0.31
|
|
Atraumatic injury
|
0
|
0
|
9
|
10.8
|
|
Surgical position
|
|
Beach chair
|
4
|
36.4
|
19
|
22.9
|
0.26
|
|
Lateral recumbency
|
7
|
63.6
|
64
|
77.1
|
Table 5
|
Variable
|
With residual pain
|
Without residual pain
|
p-value
|
|
Gender
|
|
Male
|
41
|
85.4
|
33
|
71.7
|
0.085
|
|
Female
|
7
|
14.6
|
13
|
28.3
|
|
Age at surgery (years)
|
|
Median value (Q1–Q3)
|
32 (25–38)
|
37 (30–44)
|
0.016
|
|
Injury type
|
|
Traumatic injury
|
43
|
89.6
|
42
|
91.3
|
0.52
|
|
Atraumatic injury
|
5
|
10.4
|
4
|
8.7
|
|
Surgical position
|
|
Beach chair
|
12
|
25.0
|
11
|
23.9
|
0.9
|
|
Lateral recumbency
|
36
|
75.0
|
35
|
76.1
|
Table 6
|
Variable
|
With external rotation loss
|
Without external rotation loss
|
p-value
|
|
Gender
|
|
Male
|
29
|
82.9
|
45
|
76.3
|
0.45
|
|
Female
|
6
|
17.1
|
14
|
23.7
|
|
Age at surgery (years)
|
|
Median value (Q1–Q3)
|
35 (25–44)
|
34 (28–40)
|
0.53
|
|
Injury type
|
|
Traumatic injury
|
30
|
85.7
|
55
|
93.2
|
0.20
|
|
Atraumatic injury
|
5
|
14.3
|
4
|
6.8
|
|
Surgical position
|
|
Beach chair
|
6
|
17.1
|
17
|
28.8
|
0.20
|
|
Lateral recumbency
|
29
|
82.9
|
42
|
71.2
|
Table 7
|
Variable
|
With apprehension
|
Without apprehension
|
p-value
|
|
Gender
|
|
Male
|
19
|
82.6
|
55
|
77.5
|
0.42
|
|
Female
|
4
|
17.4
|
16
|
22.5
|
|
Age at surgery (years)
|
|
Median value (Q1–Q3)
|
31 (25–40)
|
34 (28–43)
|
0.29
|
|
Injury type
|
|
Traumatic injury
|
23
|
100
|
62
|
87.3
|
0.070
|
|
Atraumatic injury
|
0
|
0
|
9
|
12.7
|
|
Surgical position
|
|
Beach chair
|
5
|
21.7
|
18
|
25.4
|
0.72
|
|
Lateral recumbency
|
18
|
78.3
|
53
|
74.6
|
Discussion
The surgical treatment for anterior shoulder instability must be chosen based on factors
such as age, previous dislocations, occupation, level of physical activity, ligament
laxity, overall health status of the patient, degree of humeral head and glenoid bone
involvement, and the presence of associated conditions, such as upper labral (superior
labral tear from anterior to posterior, SLAP) injuries and rotator cuff tendon rupture.[8] Choosing the most appropriate treatment for each patient can influence the outcomes
and minimize the risk of complications.
The best method for labral lesion repair is still debatable.[9] The anatomical procedures include open and arthroscopic Bankart surgeries, which
aim to restore the original shoulder anatomy and involve labral lesion repair.[10] Bankart open surgery, which is performed with labrum mobilization and fixation,
is still considered the gold standard for anterior instability treatment.[11] However, conditions such as incision size, risk of subscapularis muscle weakness,
loss of shoulder range of motion (especially lateral rotation) and difficulty in accessing
other intra-articular injuries have favored the choice for the arthroscopic procedure.
Other advantages of arthroscopy are reduced blood loss, shorter surgery duration and
greater esthetic satisfaction.[12]
The dislocation recurrence rates after arthroscopic surgery have been decreasing due
to technological advances and a better understanding of previous errors. In 2007,
Balg and Boileau[13] described a 14.5% recurrence rate after Bankart arthroscopic surgery. Castagna et
al.[14] obtained a 23% recurrence rate in a 10-year follow-up of 43 patients submitted to
arthroscopy. In a meta-analysis, Petrera et al[15] compared the open Bankart and arthroscopic surgeries and found a dislocation recurrence
rate 6.3% lower in the arthroscopy group.
Ahmed et al[6] also reported better results with arthroscopic techniques, with a dislocation recurrence
rate of 13.2% in a total of 302 patients. In our study, all recurrence cases were
associated with traumatic instability, with an 11.7% rate.
As for loss of lateral rotation, Bottoni et al[16] demonstrated that the decrease in range of motion was of 2 o to 3o in the arthroscopic group, and of 6.5o in the open surgery group. In our study, loss of lateral rotation was observed mostly
in patients with traumatic instability, which is similar to the results reported by
Ferreira Neto et al,[17] who also found a greater lateral rotation limitation in patients with such a feature.
These results may be justified by the better capsular ligament accommodation in patients
with atraumatic injuries when compared to patients with traumatic instability.[17]
[18]
[19] Almeida Filho et a.,[20] after performing postoperative clinical and radiographic evaluations of 49 patients,
observed that despite the arthrosis-related lateral rotation decrease, it was not
possible to ascertain a cause-effect relationship between them.
Residual pain is a common postoperative complaint in patients surgically treated for
anterior shoulder instability. Several authors[21]
[22]
[23] attribute this complication to injuries associated with instability and that may
not be diagnosed at first, especially partial rotator cuff tears, SLAP injuries, and
joint degenerative changes. In our study, approximately 51% of the patients had some
degree of postoperative residual pain, and there was an inverse association between
age and pain. The younger the age at surgery, the greater the occurrence of postoperative
pain, which may be justified by the higher level of activity of this group of patients.
In addition, as noted by Almeida Filho et al,[20] there is a significant relationship between postoperative glenohumeral arthrosis
development and younger age at the time of surgery and first dislocation, corroborating
our observation.
Some studies[13]
[24]
[25]
[26]
[27] have shown that younger patients are at an increased risk of recurrence after arthroscopic
surgical stabilization, but no age has been defined. According to Ahmed et al,[6] age alone is not sufficient to determine the treatment offered to younger patients,
but it is clearly an important factor to predict the risk of recurrence. In our series,
there was no statistically significant correlation between age and dislocation recurrence;
the median age of patients with and without recurrence was similar (32 versus 34 years
old).
Three functional scores, the UCLA, Carter-Rowe and Constant-Murley scores, were used
for outcome evaluation. The median Carter-Rowe score was 95, with 67 excellent outcomes,
14 good outcomes, 2 regular outcomes and 11 poor outcomes. As for the UCLA score,
there were 83 patients with good/excellent outcomes and 11 patients with bad/regular
outcomes. Using the Constant score, 65 patients presented excellent outcomes, 18 had
good outcomes, 10 had regular outcomes, and 1 had a poor outcome. Regarding the three
scores, all patients with regular/poor outcomes were those with dislocation recurrence.
The results reported by Boileau et al[28] in 91 patients showed a mean Carter-Rowe score of 77.8 points (ranging from 15 to
100 points). Balg and Boileau[13] observed an average Carter-Rowe score of 81.5 points (range: 10-100 points) after
evaluating 131 operated patients. Neri et al,[29] analyzing 11 patients, found a Carter-Rowe score of 74.5 points (range: 35-100 points),
and an average UCLA score of 29.6 points (ranging from 24 to 35 points), with good/excellent
outcomes in 8 patients (72.7%) and regular/poor outcomes in 3 subjects (27.3%). In
a study evaluating 314 shoulders from 302 patients, Godinho et al[30] demonstrated an average Carter-Rowe score of 91.8, ranging from 25 to 100 points,
and an UCLA score of 33.8 points, with 97.6% of good/excellent outcomes. Our series
presented a significant direct correlation between the UCLA score and age (rs = 0.250; p = 0.015), indicating that, the older the age, the higher the expected UCLA score.
Regardless of the method used, the main postoperative complication is dislocation
recurrence, which is associated with the previously mentioned risk factors. Thus,
the use of scores such as the Instability and Severity Index Score (ISIS) may indicate
the technique that provides the greatest benefits for patients. The score includes
age at time of surgery, degree of sports participation, type of sport practiced, ligamentous
laxity, presence of Hill-Sachs lesion, and loss of glenoid contour to generate a total
of 10 points. According to Balg and Boileau,[13] patients with scores of up to six points benefit most from the arthroscopic technique,
while lower scores are indications for open surgeries, especially the Latarjet procedure.
The main limitations of the study include its retrospective nature, which did not
enable a rigorous preoperative functional assessment, the short follow-up period,
and the non-normal data distribution, requiring the use of nonparametric tests.
Conclusions
We conclude that the arthroscopic repair of anterior shoulder instability has satisfactory
outcomes in the functional scores at the short/medium-term, with a low rate of severe,
limiting complications, and it can be a method of choice for most patients. Bad/regular
outcomes were associated with dislocation recurrence. Postoperative pain was associated
with younger age at surgery, whereas lateral rotation limitation was related to the
type of instability (traumatic injury).
