Ultrasound features of shoulder involvement in patients with ankylosing spondylitis: a case–control study

During Ankylosing spondylitis (AS) courses, shoulder involvement is common [1, 2]. On the basis on AS clinical assessment, prevalence of shoulder pain varied from 3.5 to 33% [1‐4]. Despite its common occurrence in patients with AS, shoulder involvement is not frequently disabling [2]. However, etiologies of shoulder pain in patients with AS remain to be defined, physicians presumed that shoulder pain may be due to synovitis, bursitis, or structural joint damage. There have been very few studies which examined the etiology of shoulder pain in patients with AS, and usually not in a controlled format [1, 2, 4, 5]. Some cross-sectional plain radiographic studies have described joint space narrowing, erosion, and bony proliferation in the acromio-clavilcular joint, the gleno-humeral joint, and around the rotator cuff insertion [1, 2, 5].

Nowadays, it is widely known that enthesitis is the primary clinical feature in AS [6, 7]. Enthesitis has been demonstrated in various peripheral locations such as the Achilles tendon, and the knee [8‐10], but it has not yet been formally evaluated in the shoulder. For the clinical evaluation of enthesitis in AS, several index have been developed [11‐16], but the shoulder is not usually included in these index.

Over the last few years, ultrasound has proved to be a reliable method for assessing tendon and joint involvement with a high sensitivity. Several musculoskeletal features of AS such as enthesitis, synovitis, erosion, bursitis, and tenosynovitis could be visualized by ultrasound [17]. Recently, there is an increasing interest for the use of ultrasound for AS evaluation. Many ultrasound index were developed to assess enthesitis in patients with AS: the Glasgow Ultrasound Enthesitis Scoring System (GUESS) [18], Sonographic Entheseal Index (SEI) [19], and Ultrasound Enthesis Score [20]. None of these indices include the scanning of shoulder enthesis, which due probably to the lack of sonographic study of shoulder in patients with AS.

The purpose of this study was therefore to assess ultrasound features of shoulder in AS patients compared with age and sex-matched healthy controls, furthermore, we aim to determine predictive factors of shoulder enthesitis in patients with AS.

Thirty eight patient with AS and 38 age and sex-matched healthy controls were enrolled. All patients with AS fulfilled the modified New York criteria for ankylosing spondylitis [21]. We collected demographic and clinical data including; disease duration, form of AS, presence of coxitis and previous or current shoulder pain.

AS activity was assessed by the bath ankylosing spondylitis disease activity index (BASDAI) [22], and we used the bath ankylosing spondylitis function index (BASFI) [23] to evaluate function. Each patient and control subject provided informed written consent to participate in this study at baseline and the study received approval from the ethic committee of biomedical research of Faculty of Medicine and Pharmacy of Rabat.

Each group of patients with AS and control subjects comprised similarly 22 men (57.9%) and 16 women (42.1%), the mean age was respectively 36.8 ± 10.7 and 36.9 ± 10.6 years (p = 0.95). AS patients had mean disease duration of 9.6 ± 7.2 years. The mean BASDAI was 4.9 ± 1.9. Among 38 patients with AS, 21 had coxitis (55%) and 19 had previous or current shoulder pain (50%) (Table 1).

Ultrasound abnormalities of shoulders were significantly more frequent in patients with AS (28/38 patients (73.7%)) than controls (14/38 controls (36.8%)). The most frequent disorder in patients with AS was enthesitis. Among 76 shoulders examined in each group (Table 2), 43 AS shoulders presented enthesitis (56.6%) versus 8 shoulders in control group (10.5%). In both groups, the most frequent involved entheses in shoulder were supraspinatus e.g. (Figure 1) followed by subscapularis e.g. (Figure 2) and infraspinatus. Table 3 shows aspects of enthesitis in AS shoulders compared with control shoulders.

In addition, entheseal involvement was multiple in 21 of 76 AS shoulders and in 14 patients with AS (36.8%), however, there was not multiple entheseal involvement in control shoulders. Seventeen patients from 38 had bilateral enthesitis in shoulders (44.7%) while no bilateral enthesitis was found in control subjects.

Involvement of rotator cuff tendons was significantly higher in patients with AS compared with control subjects (16/38 (42.1%) versus 6 (15.2%) respectively). This rotator cuff involvement was significantly frequently bilateral in patients with AS in comparison with control subjects (Table 2). Description of rotator cuff tendon abnormalities is presented in Table 3. Other ultrasound abnormalities such as synovitis and joint involvement were similarly less frequents in both groups (Tables 2 and 3).

In the group of patients with AS, 23/38 patients had previously known entheseal involvement. Patients with previously known enthesitic form presented similar frequency of shoulder enthesis than patients without enthesitic form (16/23 (69.5%) and 10/15 (66.6%) respectively, p = 1). Similarly, no difference was found between patients with isolated axial disease and mixed forms (axial disease with peripheral joint involvement and/or enthesitic form).

Among 76 AS shoulders, thirty three shoulders were currently or previously painful. These painful shoulders presented significantly more ultrasound abnormalities (75.8% (25/33)) and more enthesitis (24/33 (72.7%)) compared with no painful AS shoulders (20/43 shoulders (46.5%) presented US abnormalities and 19/43 (44.2%) enthesitis).

At univariate analysis, the presence of shoulder enthesitis was associated to long onset disease, coxitis, high disease activity (BASDAI), and poor function (high BASFI). After multivariate analysis, the only factor which still strongly associated to shoulder enthesitis was coxitis (OR = 9.4 with IC 95% (1.10; 81.9)) (Table 4).

To our knowledge, this work represents the first controlled ultrasound evaluation of shoulders in patient with AS. There were several important findings in this study: Firstly, gleno-humeral synovitis was uncommon in our patients with AS. Secondly, the high frequency of enthesitis in AS shoulders with multiple entheseal involvement in shoulder compared with controls, and the bilateral aspect of enthesitis in AS shoulders. Thirdly, the frequency of enthesitis in shoulders was similar in all AS subtypes. Fourthly, coxitis was the only predictive factor for this enthesitis.

Despite of the commonly involvement of shoulders during AS course [2], very few studies focused to study the etiology of shoulder pain in these patients [1, 2, 4, 5]. Using plain radiography, shoulders of 26 AS patients were examined in a previous not controlled study [2]. 31% of patients presented evident radiologic abnormalities and the most commonly involved site was the acromio-clavicular joint, 9 shoulders presented sever radiologic changes in the gleno-humeral joint. One other study, which enrolled 52 patients with AS, found in 29 shoulders plain radiologic abnormalities including acromio-clavicular joint narrowing, sclerosis of the greater tuberosity and cystic changes in the humeral head [1]. In our study, ultrasound involvement of acromio-clavicular and gleno-humeral joints was rare. Recently, a MRI study included 15 patients with AS with shoulder pain, and a control group of 92 individuals with nonspecific shoulder pain. Bone marrow edema at any entheseal site was noted in significantly more AS shoulders (70.6%) than in control (19.1%) shoulders [4].

Some clinical scores have been developed to assess enthesitis in patients with AS, but they don’t all include the shoulder [11‐16]. The most recent index has been developed by the Spondyloarthritis Research Consortium of Canada (SPARCC), it evaluates only the peripheral entheses and it includes the shoulder [16]. According to this SPARCC index, the most frequently affected sites were the greater trochanter and supraspinatus insertion. Despite the lack of specificity of enthese clinical evaluation, the finding of this index is relatively consistent with our ultrasound results, supraspinatus insertion was also the most involved enthese site in the shoulders of our patients.

For enthese evaluation, the accuracy of clinical examination and X ray is still uncertain. So, new imaging techniques such as ultrasound and magnetic resonance imaging (MRI) were recently used to examine enthese in patients with AS. Over the last few years, several studies have highlighted the value of ultrasound in assessing entheses in AS. Ultrasound B-mode aspects of lower limb enthesitis of AS was firstly described by Lehtinen et al. [10] in 1994, and in 2003 Balint et al. developped a ultrasound scoring system designed to assess five entheseal sites in the lower limb (GUESS) [18]. Thereafter, D’Agostino et al. used B-mode US combined with power Doppler (PDUS) to show a high frequency of abnormal peripheral enthesitis among patients with AS in comparison with controls affected by rheumatoid arthritis or degenerative spinal disease in a cross-sectional study [31]. Recently, other ultrasound scoring system were developed, the Spanish enthesis index [19, 20] which is based on the use of ultrasound B-mode only, and the Madrid sonographic enthesitis index (MASEI) [20], which combines abnormalities in grey scale and Doppler . However, none of these scores included the shoulder, and actually there is no clear agreement on which structures to examine while assessing enthesitis in patients with AS. Also, there is no consensus of the definition of enthesitis in AS, so definition varied between authors [30]. In 2005, an OMERACT-EULAR working group on Ultrasound was constituted to address validity issues, and to establish international consensus and scoring systems in the use of US in musculoskeletal diseases [32].

On the other hand, the landmark of US enthesitis in patients with AS was the presence of abnormal vascularization at entheses [31]. In our work, abnormal vascularization at entheses in shoulder was rarely found, that could be due to the recruitment based on patients with or without current or previous shoulder pain. In addition, the high frequency of asymptomatic enthesitis is in agreement with previous findings in other enthese sites in patients with AS [10, 18, 31]. It is known that peripheral enthesitis may be observed in all forms of AS and all phases of disease evolution [33], also D’agostino et al. found no major differences in the frequency of Ultrasound enthesitis between the different AS subtypes [31], which is consistent with our findings.

In our work, synovitis of gleno-humeral joint was rare. On the other hand, patients with AS presented a high prevalence of coxitis in our study. Previous studies, which aimed to describe clinical aspects of AS in our area, had also found a high prevalence of coxitis [34‐36]. In the present paper, hip involvement was the only significant factor associated with enthesitis in shoulders. These results seem to be near to previous data suggesting that involvement of the hip joint in AS is higher in patients with shoulder involvement. Although, disease duration was not a significant factor of shoulder involvement in our study, which is in contrast with previous finding which suggested that disease duration could predict shoulder involvement [2].

In conclusion, in patients with established AS, ultrasound abnormalities of shoulder are common, particularly enthesitis. The significant association between coxitis and shoulder enthesitis led us to suggest that ultrasound evaluation of shoulder should be performed in patients with AS with hip involvement. Finally, there is a great interest to evaluate the diagnostic value of shoulder enthesitis by performing large studies in early onset AS.