| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Department of Diagnostic Radiology and Organ Imaging, The Chinese University
of Hong Kong, Prince of Wales Hospital, 30-32 Ngan Shing St., Shatin, Hong
Kong, SAR, China.
2 Department of Orthopaedics and Traumatology, The Chinese University of Hong
Kong, Prince of Wales Hospital, Hong Kong, SAR, China.
3 Centre for Epidemiology and Biostatistics, Postgraduate Education Centre,
Faculty of Medicine, Prince of Wales Hospital, Hong Kong, SAR, China.
Received August 9, 2007;
accepted after revision November 6, 2007.
Address correspondence to J. F. Griffith
(griffith{at}cuhk.edu.hk).
Abstract
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
SUBJECTS AND METHODS. Two hundred eighteen patients with single or recurrent anterior shoulder dislocation underwent shoulder CT examination. Fifteen patients had bilateral dislocation. Prevalence and severity of glenoid bone loss and glenoid fracture were assessed. CT examinations of 56 control subjects without shoulder dislocation were evaluated for glenoid contour and side-to-side variation in glenoid width.
RESULTS. Glenoid bone loss was present in 27 (41%) of 66 patients
with first-time unilateral dislocation and 118 (86%) of 137 patients with
recurrent unilateral dislocation. Glenoid bone loss ranged from –0.3% to
–33% (mean, –10.8% ± 7.9%). Seventy-four (51%) of 145
patients had 
10% glenoid bone loss, 54 (37%) had between 10% and 20%,
eight (6%) had between 20% and 25% glenoid bone loss, and nine (6%) had 
25% glenoid bone loss. Glenoid rim fractures were present in 49 (21%) of 233
dislocated shoulders. The number of dislocations correlated moderately with
the severity of glenoid bone loss (r = 0.56). The normal side-to-side
glenoid width variation was small (0.46 ± 0.81 mm).
CONCLUSION. Glenoid bone loss is common in anterior shoulder dislocation. It is probably multifactorial in origin, is usually mild in degree, and has a maximum observed severity of –33%. Dislocation frequency cannot accurately predict the degree of bone loss.
Keywords: CT • dislocation • glenoid • shoulder
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Glenoid and humeral bone loss can be evaluated by radiography and CT [2, 5, 6, 11, 12]. CT examination shows that glenoid bone loss leads to straightening of the normally curved anterior glenoid rim (i.e., an anterior straight line) and reduction in glenoid width of the dislocated side compared with the contralateral, nondislocated side [5, 12]. These features may be evaluated on a CT image obtained en face to the glenoid articular surface [5, 12]. This CT quantification method rests on the observation that normal side-to-side variation in glenoid width is small [5]. CT has high sensitivity (93%) and specificity (78%) for detecting glenoid bone loss and has good agreement (r = 0.79) with shoulder arthroscopy regarding the severity of glenoid bone loss [12].
Because methods of quantifying glenoid bone loss have only recently been described, the prevalence, spectrum, and pattern of glenoid bone loss in a large number of patients with anterior shoulder dislocation are not known. This information is central to the detection and interpretation of bone defects on cross-sectional imaging and to understanding the pathophysiology and consequences of recurrent dislocation.
The primary aim of this study was to prospectively study a large cohort of patients with anterior shoulder dislocation to establish the prevalence, spectrum, and severity of glenoid bone loss. Secondary aims were to investigate the etiology of glenoid bone loss and to test the reliability of the CT technique used.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
|
|
Subjects Without Shoulder Dislocation
Fifty-six control subjects (24 males, 32 females; mean age, 45 years;
range, 14–76 years) with no history of shoulder dislocation, who had
undergone high-resolution CT of the upper thorax and both shoulder girdles in
an identical manner to the patient group (i.e., with the arms positioned by
the chest wall), were evaluated as a reference for normal glenohumeral
morphology and side-to-side variation in glenoid width and length. Only
patients with disease not likely to affect the size or shape of the glenoid
fossa or proximal humerus were defined as healthy subjects. The indication for
CT examination in these 56 subjects was sternomanubrial disease in 38 (68%)
subjects, clavicle or proximal humeral fractures in 11 (20%), and upper
thoracic or proximal humeral tumors in seven (12%). Informed consent was not
required by the ethics committee for review of these control CT
examinations.
CT Examination Technique and Analysis
Simultaneous CT examination of both shoulders was undertaken in both
patients and healthy subjects with either a single-detector helical CT scanner
(HiSpeed Advantage, GE Healthcare) using 1-mm acquisitions with 200 mA, 120
kV, and a pitch of 1:1) (n = 118 examinations) or an MDCT scanner
(LightSpeed 16 Plus, GE Healthcare) using 16 x 0.625 mm acquisitions
with 400 mA, 120 kV, and a pitch of 1:1.75 (n = 100 examinations).
The scanning plane extended from the acromion to just below the glenoid, with
the patient's arms positioned by the chest wall and the palms pronated. Double
oblique reconstruction of each glenoid was used to obtain oblique sagittal
images en face to the glenoid articular surface (Advantage Windows,
version 4.2, GE Healthcare) (Figs.
1A,
1B,
2A,
2B,
3A,
3B,
4).
|
|
|
|
|
Presence and Quantification of Glenoid Bone Loss
Glenoid bone loss in patients with unilateral dis location was diagnosed
only after two criteria were met. First, an anterior straight line to the glen
oid rim had to be present; and second, there had to be a reduced glenoid width
in the dislocated shoulder compared with the normal side. In other words, an
anterior straight line or reduction in glenoid width in isolation was not
sufficient to constitute glenoid bone loss. Both measures were combined
because an anterior straight line adds specificity to an isolated measurement
of reduced glenoid width [12].
Percentage of glenoid bone loss was measured by dividing the difference in
glenoid width between the affected and unaffected sides by the glenoid width
of the unaffected side (Figs.
1A,
1B,
2A,
2B,
3A,
3B). No quan titative measure
of glenoid bone loss was possible in patients with bilateral dislocation
because there was no normal glenoid for comparison.
Presence and Type of Glenoid Rim Fracture
The presence and pattern of glenoid rim fracture were determined on axial
and reformatted images through the glenoid (Figs.
4 and
5A,
5B,
5C,
5D,
5E). Glenoid rim fractures
were diagnosed by observing either a free bone fragment adjacent to the
glenoid rim (a detached fracture) or a focal deformity of the glenoid contour
consistent with either a partially detached fracture or a reattached fracture
(attached fracture) (Figs. 3A,
3B and
4).
|
|
|
|
|
Reliability of CT Measurements
Inter- and intraobserver agreement was established on 40 CT examinations
(80 shoulders) comprising both healthy individuals and patients with
dislocation. For interobserver agreement, two investigators, one with 19 and
one with 2 years of working CT experience, independently recon structed and
measured the glenoid length and width on 40 CT data sets. For intraobserver
agreement, the same 40 CT data sets were recon structed and remeasured, and
these second blinded readings were compared with the initial readings. The
interval between initial and second CT measurements for interreader
reliability was 3–4 weeks.
Data Analysis
SPSS for Windows, version 11.5 (SPSS), was used for statistical analyses.
Variables were expressed as mean ± 2 SDs unless otherwise stated.
Pearson's chi-square test was used to test for side-to-side or sex differences
in normal glenoid width and for length as well as prevalence of glenoid bone
loss, Hill-Sachs deformity, and fracture in patients with single or recurrent
dislocation. Spearman's cor relation was used to examine the relationship
between glenoid bone loss, Hill-Sachs deformity, and number of dis locations.
CART (classification and regression tree), a binary tree-structured stati
stical method that can help search for hidden structure in data, was applied
to help determine a critical level of glenoid bone loss—that is, the
level beyond which the patient was more likely to develop increased frequency
of dislocation. Intraclass cor relation coefficient was used to assess
interreader and intrareader reliability. Intraclass correlation coef ficients
were expressed as means (± 95% CIs). A 5% significance level was
applied for all tests (p < 0.05).
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
|
|
|
|
Seventy-four (51%) of the 145 patients with glenoid bone loss incurred
10% glenoid bone loss, 54 (37%) incurred > 10% but < 20%, and eight (6%)
incurred > 20% but < 25% glenoid bone loss
(Fig. 7). Nine (6%) patients
incurred > 25% glenoid bone loss (Figs.
3A,
3B and
6). CART analysis indicated a
critical level of glenoid bone loss to be 13.4%. Below this level, the average
number of dislocations was 6.3, whereas above this level the average number of
dislocations rose to 10.1. The greatest bone loss seen in this cohort was
–33% (Figs. 6 and
8).
Increased length of the anterior straight line correlated positively with a reduction in glenoid width (r = 0.72), indicating that as glenoid width diminished, the normally curved anterior edge of the glenoid became progressively straighter (Figs. 1A, 1B, 2A, 2B, 3A, 3B). With severe bone loss, an anterior concavity to the anterior glenoid was observed (Fig. 3A, 3B). Such anterior concavity was present in 23 (10%) of 233 dislocated shoulders. If an anterior concavity is present, glenoid bone loss is generally severe (22.8 ± 7.4%).
Glenoid rim fracture was present in 49 (21%) of 233 dislocated shoulders (Table 1). The relative prevalence of detached or attached glenoid rim fracture is shown in Table 1. Glenoid rim fracture was more common in recurrent than single dislocation, although this difference was not significant (p = 0.39) (Table 1).
Prevalence of Hill-Sachs deformity—Hill-Sachs deformity was present in 200 (86%) of 233 dislocated shoulders (Table 1). Hill-Sachs deformity was not more common (p = 0.21) (Table 1) but was more severe (p = 0.008) in recurrent than in single dislocations. A weak positive correlation was seen between glenoid bone loss and increasing severity of Hill-Sachs deformity (r = 0.24, p = 0.03).
Anterior straight line and Hill-Sachs deformity—An anterior straight line was observed less frequently than Hill-Sachs deformity in single dislocation (p < 0.001) and more frequently than Hill-Sachs deformity in recurrent dislocation (Table 1), although this difference was not significant (p = 0.12). Either a humeral (Hill-Sachs deformity) or glenoid (anterior straight line) bone defect was evident in all shoulders with recurrent dislocation.
Subjects Without Shoulder Dislocation
Glenoid width, length, and side-to-side variation in normal shoulders
without dislocation—In the 56 subjects without dislocation (i.e.,
112 normal shoulders), the normal glenoid fossa width was found to be 26.5
± 4.6 mm and the glenoid fossa length, 40.1 ± 7.6 mm.
Side-to-side variation in glenoid width was 0.46 ± 0.81 mm.
Side-to-side variation in glenoid length was 0.78 ± 0.68 mm. Both
glenoid width and length were significantly larger in males than females
(p < 0.0001). For each sex, no significant side-to-side difference
in either glenoid width (males, p = 0.25; females, p = 0.28)
or length (males, p = 0.53; females, p = 0.78) was
found.
Prevalence and pattern of anterior straight line and Hill-Sachs-type deformity in normal shoulders without dislocation—An anterior straight line to the glenoid rim was present in seven (2%) of 282 nondislocated shoulders (comprising 112 shoulders in healthy subjects without dislocation and 170 nonaffected shoulders in patients with unilateral dislocation). This anterior straight line measured 5.6–11.9 mm (mean, 7.9 mm). A Hill-Sachs-type deformity was present in two (0.7%) of 282 nondislocated shoulders. This Hill-Sachs-type deformity was classified as mild in both cases.
Inter- and Intraobserver Reliability of CT Measurements
Interobserver agreement, as determined by the intraclass correlation
coefficient, was 0.914 for glenoid width and 0.858 for glen oid length
measurements. The intraobserver measurement difference was –0.24
± 2.27 mm (mean ± 2 SDs) for glenoid width and 0.09 ±
3.27 mm for glenoid length. Intraobserver agreement was 0.958 for glenoid
width and 0.790 for glenoid length measurements. The interobserver measurement
difference was –0.14 ± 1.40 mm for glenoid width and 0.22
± 3.94 mm for glenoid length.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
This study reveals that glenoid bone loss is common, being evident in 41% of single and 86% of recurrent dislocated shoulders. This is in agreement with the radiographic study of Edwards et al. [6], who undertook a fluoroscopically guided radiographic analysis ("Bernageau profile view") of 107 recurrent dislocated shoulders. Although no quantitative analysis was undertaken, a glenoid bone defect was present in 75% of cases, including a fracture in 51%; loss of the anterior angle ("cliff sign") in 16%; and blunting of the anterior angle ("blunted angle sign") in 33% [6]. Rowe et al. [10], in a surgical study of 162 recurrent dislocations, showed glenoid rim damage in 73% comprising eburnation, erosion, or fracture. Three previous CT-based studies have addressed glenoid bone loss [3–5]. Griffith et al. [5], in a study of 50 patients, analyzed which methods optimally quantify glenoid bone loss. Sugaya et al. [3] applied a best-fit circle to the inferior glenoid on 3D reconstructed CT images. Glenoid contour deformity was present in 90 (90%) of recurrent dislocated shoulders, comprising a fracture in 50% and a contour deformity in 40%. The size of the bone fragment was used as a quantitative measure of bone loss in patients with fracture, although no quantitative analysis was undertaken in patients without fracture [3]. Saito et al. [4] studied the location of the glenoid defect on CT but did not undertake a quantitative assessment of the defect.
Our study, therefore, is the first large-scale quantitative study of the spectrum of glenoid bone loss occurring in anterior shoulder dislocations. It shows that glenoid bone loss is very common, is usually mild (< 10%) in degree, and has a maximum severity of approximately 33%. Bone loss greater than one third of the normal glenoid width was not observed in any patient. This finding is to be expected because more severe glenoid bone loss would involve erosion of the thicker middle third of the glenoid. More severe degrees of bone loss result in a concavity to the anterior glenoid rim rather than progressive flattening of the anterior glenoid. Only modest correlation was present between frequency of dislocation and severity of glenoid bone loss, indicating that one cannot accurately predict severity of glenoid bone loss on the basis of the number of dislocations alone. Indeed, several of the most severe cases of glenoid bone loss occurred in patients with only one or a few dislocations. The inverse exponential relationship observed between the number of dislocations and the severity of glenoid bone loss is as expected because bone loss will proceed more quickly initially through the thinner, peripheral parts of the glenoid and then more slowly through the thicker, central parts of the glenoid.
This study defines the spectrum of glenoid bone loss, which is the first step to knowing the optimal level of bone loss at which bone augmentation procedures should be considered [8, 9]. In this cross-sectional study using CART regression analysis, we attempted to define a critical level of glenoid bone loss by determining that level above which the patient is more likely to have experienced more frequent dislocations. This critical level was found to be 13.4%. Beyond this level, the num ber of dislocations experienced by the patient rose steeply from six to 10 dislocations. This does not necessarily imply that this is the level at which bone augmentation surgery should be undertaken. That particular level has not yet been determined because it is only recently that means of quantifying glenoid bone loss have been available, and very little data exist on the prevalence and severity of glenoid bone loss. Some orthopedic surgeons have recommended 25%, others 33%, whereas others recommend bone grafting be undertaken if there is a large glenoid defect [8, 11, 13–15]. Our study, through defining the spectrum of glenoid bone loss, provides baseline data for defining the optimal level of surgical intervention in the future.
The frequency of 86% for Hill-Sachs deformity in this CT-based study is comparable to the 76–90% reported from radiographic studies [6, 10, 16]. Hill-Sachs deformity results from impaction of the posterosuperior aspect of the humerus against the glenoid [2, 6]. The origin of glenoid bone loss is less clear. It is not likely to be solely due to fracture because in our study, the prevalence of glenoid rim fracture was only 21%. Similarly, it is unlikely to be the glenoid equivalent of a Hill-Sachs deformity because only a weak association between the severity of Hill-Sachs deformity and glenoid bone loss was found. Nevertheless, it was not rare to find glenoid bone loss of 5–10% in patients with only a single dislocation in the absence of a glenoid rim fracture. Overall, it seems most likely that glenoid bone loss is multifactorial in origin, resulting from glenoid rim erosion or fracture during dislocation and humeral head impaction after dislocation. Some form of bone defect—either Hill-Sachs deformity or an anterior straight line—was present in all patients with recurrent dislocation.
Our study confirms the usefulness of CT in quantifying glenoid bone loss. CT allows both shoulders to be simultaneously examined. As shown, little side-to-side difference in normal glenoid width exists so both glenoids can be readily compared. The examination and analysis are easy to perform and the repeatability of the test is high. The main limitations of CT are that it cannot be used to accurately quantify glenoid bone loss in bilateral dislocations (because there is no normal shoulder for reference), and it involves irradiation of an area close to the thyroid and breast, although this can be minimized by ensuring a scanning plane limited to the glenoid region only.
MRI is the usual method of investigating anterior shoulder dislocation. Although MR images obtained directly en face to the glenoid fossa should be comparable to reconstructed CT images, MRI may not allow quantification of glenoid bone loss because the contralateral glenoid is not included for comparison. However, MRI should allow one to select those patients who would benefit from CT quantification of glenoid bone loss.
Our study has some limitations. First, as in all published studies of this nature, the frequency of dislocation was established retrospectively. As a result, the number of dislocations cited by patients with frequent dislocations may not be entirely accurate. Second, when analyzing CT examinations, observers were not blinded to whether the shoulder had previously dislocated. It is difficult to remove this bias because CT shows contour deformities indicative of dislocation. Third, no arthroscopic quantification of glenoid bone loss was available, although a previous study has shown good correlation between CT and arthroscopy in this respect [12]. Fourth, an ordinal measure of Hill-Sachs deformity was used only after several quantitative methods (e.g., humeral cross-sectional area, maximum dimensions of defect, best circular fit) proved unreliable.
In conclusion, glenoid bone loss is almost as common as Hill-Sachs deformity in anterior shoulder dislocation. Glenoid bone loss is probably multifactorial in origin and is usually mild in degree, with a maximum observed severity of –33%. Occasionally, only a single or a few dislocations may result in a large bone defect. Dislocation frequency cannot accurately predict the degree of bone loss, which helps to justify the use of CT.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
