Factors Associated with Radiological Hip Joint Involvement in Patients with Ankylosing Spondylitis

Background: The associated factors for hip involvement in patients with ankylosing spondylitis (AS) are poorly known. This study was to analyze the clinical data of patients with AS and to explore the potential associated factors of AS-related radiological hip joint. Methods: This was a cross-sectional study of patients diagnosed with AS and treated at the Beijing Jishuitan Hospital between 01/2013 and 12/2019. A BASRI-hip score ≥ 2 was dened as radiological hip joint involvement. Univariable and multivariable logistic regression analyses were performed to analyze the factors associated with radiological hip joint involvement. Results: A total of 350 AS patients were included. Patients with radiological hip joint involvement (BASRI-hip ≥ 2) accounted for 50.6% (177/350). The proportion of men was 83.7% (293/350). The mean age was 35.0±12.7 years old. The mean duration of the disease was 10.8±8.6 years. The HLA-B27 positive rate was 90.9% (318/350). The multivariable analysis showed that the juvenile onset (OR=4.955, 95%CI: 2.464-9.961, P<0.001), bone mass lower than peers (OR=2.862, 95%CI: 1.593-5.142, P<0.001), BMI <18.5 kg/m2 (OR=2.832, 95%CI: 1.321-6.069, P=0.007), BASFI (OR=1.278, 95%CI: 1.069-1.527, P=0.007), and continuous NSAIDs treatment (OR=0.400, 95%CI: 0.200-0.799, P=0.009) were independently associated with radiological hip joint involvement in patients with AS. Conclusion: worse and lower bone density. BMI: body mass index; CRP: C-reactive protein; ESR: erythrocyte sedimentation rate; BASDAI: Bath Ankylosing Spondylitis Disease Activity Index; BASFI: Bath Ankylosing Spondylitis Functional Index; TNFi: tumor necrosis factor inhibitor; NSAIDs: non-steroidal anti-inammatory drugs; BASRI: Bath Ankylosing Spondylitis radiological index.


Background
Ankylosing spondylitis (AS) is a chronic in ammatory disease that mainly damages the central axis joints [1]. It is more common in young men and has an incidence of 0.1%-0.5% [1]. The main features are in ammatory low back pain, attachment point in ammation, sacroiliitis, new bone formation in the spine, and high correlation with the HLA-B27 genetic marker [1]. The complications of AS are low bone density, osteoporosis, and fracture [1]. The majority of patients AS report the chronic use of various types of pain killers [1].
Previous studies reported that the proportion of AS-related hip involvement is high and that about 25%-33% of patients with AS also have hip joints damage [2]. About 5% of patients with AS will need hip arthroplasty [2]. In humans, the hip joints are the full-motion joints with the greatest weight-bearing demand. It is the most important joint for maintaining body balance and limb movements. Hip joint damage is also one of the main causes of disability in AS [3]. The methods for evaluating AS-related hip joint involvement include clinical symptoms, joint examination, and imaging results (X-ray radiography and magnetic resonance imaging). Among them, the Bath Ankylosing Spondylitis Radiology Index (BASRI)-hip score based on X-rays is commonly used in the studies of AS-related hip joint damage, and the score is relatively objective [4]. While new bone formation is mainly involved in AS-related damage in the spine, AS-related hip damage mainly involves synovial in ammation, bone erosion, and joint space narrowing [5].
A previous study showed that a high body mass index (BMI) and advanced hip arthritis at baseline were associated with hip arthroplasty in patients with AS [5]. Another study showed that the use of anti-tumor necrosis factor (anti-TNF) decreased the rate of hip arthroplasty by 40% [6]. Notwithstanding, the associated factors for hip involvement in patients with AS are not well known, and exploring the associated factors of AS-related hip joint involvement could guide not only clinical work but also help basic research.
Therefore, the aim of this retrospective study was to analyze the clinical data, including the BASRI-hip score, of patients with AS and to explore the potential risk factors of AS-related radiological hip joint. The results could help identify the patients who might require a closer follow-up because they might require hip arthroplasty.

Patients
This was a cross-sectional study of patients diagnosed with AS and treated at the rheumatology department of Beijing Jishuitan Hospital between January 2013 and December 2019. The inclusion criteria were: 1) >18 years of age; 2) met the modi ed New York criteria (1984) for AS classi cation; and 3) duration of disease ≥1 year. The exclusion criteria were: 1) incomplete data; 2) bone tumor, bone metastasis, or hematological cancer; or 3) other rheumatic diseases, such as rheumatoid arthritis, gouty arthritis, or infectious arthritis.
The study was approved by the Ethics Committee of Beijing Jishuitan Hospital (No: 202003-13). The need for individual consent was waived by the committee because of the retrospective nature of the study.

Data collection
The variables included sex, age, age at onset, duration of disease, smoking history, family history, BMI (classi ed into <18.5 kg/m 2 and ≥18.5 kg/m 2 ) [7], Schober's test, peripheral arthritis (physician found swelling or tenderness in peripheral joints or joint effusion and synovitis on imaging examination), and iritis (diagnosed by an ophthalmologist). The BASDAI and BASFI scores of all patients were recorded [8,9]. The erythrocyte sedimentation rate (ESR) was recorded (Italian ALIFAX Test-1 automatic rapid erythrocyte sedimentation rate analyzer; normal reference range, 0-20 mm/h), as well as C-reactive protein (CRP) (Beckman IMAGE800 analyzer and matching kit, immunoturbidimetry; normal reference range, 0-8 mg/L), HLA-B27 status, and bone mass. Bone mass was determined at the calcaneus using an Ultrasonic bone intensity meter (GE Healthcare, Waukesha, WI, USA). A Z-value ≤-2 was de ned as bone mass lower than that of peers. The drugs (anti-TNF, non-steroid anti-in ammatory drugs (NSAIDs), sulfasalazine, methotrexate, thalidomide, and glucocorticoids) were recorded. All indicators were collected by the specialists at the department of rheumatology and immunology.

Outcomes
The outcome index was the radiographic hip joint involvement assessed by the BASRI-hip score [10]. The BASRI-hip score ranges from 0 to 4 points: 0) normal, i.e., no radiological hip joint damage; 1) suspected hip joint damage, i.e., limited joint space stenosis; 2) mild hip joint damage, i.e., with an obvious hip joint lesion, but the hip joint space >2 mm; 3) moderate hip joint damage, with a de nite hip joint lesion, hip joint space ≤2 mm and articular bony interface ≤2 cm; and 4) severe hip joint damage, i.e., with hip joint fusion or articular bony interface ≥2 cm, or indication for total hip replacement. A BASRI-hip score ≥2 was de ned as radiological hip joint involvement.
Statistical analysis SPSS 22.0 (IBM, Armonk, NY, USA) was used for all analyses. Continuous data were tested with the Kolmogorov-Smirnov test for normal distribution. Normally distributed continuous data are expressed as means ± SD, and non-normally distributed continuous data are expressed as medians (Q1, Q3). Categorical variables were presented as frequencies.
Normally distributed continuous data were tested using the Student t-test, while non-normally distributed continuous data were analyzed by the Mann-Whitney U test. Univariable and multivariable (forward (LR) method) logistic regression analyses were performed to analyze the factors associated with radiological hip joint involvement. The variables with P<0.1 in the univariable analyses were included in the multivariable analysis. Two-sided P-values <0.05 were considered statistically signi cant.

Results
Characteristics of the patients A total of 350 AS patients were included. The basic information, medical history information, and objective examination results are shown in Table 1

Univariable analyses
The results of the univariable analyses are shown in

Multivariable analysis
The multivariable analysis of the factors associated with radiological hip joint involvement is shown in Table 3. The variables with P < 0.1 in the univariable analyses were included in the multivariable analysis.
After adjusting for sex, iritis, smoking history, family history, and TNFi > 3 months, the results showed that the juvenile onset (age

Discussion
The results strongly suggest that AS with radiological hip joint involvement had worse body function and lower bone density. The independently associated factors with radiological hip joint involvement in patients with AS included juvenile-onset, bone mass lower than peers, thin body size, BASFI, and continuous NSAIDs drug treatment.
The results of the multivariable analysis showed that a juvenile-onset, bone mass lower than peers, and thin body size were associated with radiological hip joint involvement. The occurrence rate of radiological hip involvement in AS with juvenile-onset was 5.0 times that of non-juvenile onset. A study pointed out that a juvenile-onset of AS might be more serious, the occurrence rate of radiological hip joint damage might be higher, and the requirements for total hip replacement could be increased [11]. Another previous study indicated that spinal arthritis with onset during childhood was less likely to affect the axial bones, but it was more likely to involve the hip joints [12]. Another study divided AS patients into three groups according to the age of onset and found that the degree of radiological hip joint damage in patients with AS and juvenile onset was signi cantly more severe than that in AS patients with adult-onset [13]. Therefore, the patients could be required to be screened for hip joint damage, and the physicians should pay attention to the progression of hip joint damage during follow-up.
Osteoporosis is a common complication of AS, and the occurrence of both low bone mass and osteoporosis is high in AS patients [14,15]. Of all the 350 AS patients in this study, patients with low bone mass accounted for 25.7% (90/350). Exercise could induce osteoclast differentiation to initiate bone reconstruction, which might increase bone mass [16].
The BASFI of patients with AS and radiological hip joint involvement in this study was signi cantly higher than in those without radiographic hip joint involvement. Compared with healthy people, AS patients have a fat-free mass (FFM) of 3 kg lower than the mean value, and appendicular lean mass (ALM) of 1 kg/m less than the mean value [17]. The data in this study showed that the mean BMI of all 350 AS patients was 23.0 ± 4.4 kg/m 2 , which was lower than the mean level of Chinese adults (24.7 ± 3.5 kg/m 2 ) [18]. The multivariate analysis also showed that lean body shape (BMI < 18.5 kg/m 2 ) was an independent risk factor for radiological hip joint involvement in patients with AS.
The results of this study showed that the cumulative use of slow-acting drugs such as sulfasalazine, methotrexate, and thalidomide for more than 6 months was not protective for radiological hip joint damage in patients with AS. Some studies also reported that anti-TNF could slow the progression of hip joint damage in patients with AS [19,20]. One study even reported that six AS patients with radiological hip joint involvement had increased hip joint space after anti-TNF treatment, and the BASRI score of these six patients decreased from 3 to 2 points [21], but the sample size was small. The present study showed that the usage rate of anti-TNF in AS patients with radiological hip joint damage was 19.7%, which was higher than that of 9.5% in AS patients without radiological hip joint damage, but the multivariable analysis showed that anti-TNF was not a protective factor for AS radiological hip joint damage. The therapeutic effect of NSAIDs on AS had been con rmed by many studies, and the 2019 American College of Rheumatology (ACR) update on AS and non-radiographic axial spondyloarthritis recommended continuous usage of NSAIDs in patients with AS [1]. A 2-year follow-up randomized controlled study showed that continuous non-steroidal anti-in ammatory drug treatment could reduce the progression of spine imaging in patients with AS [22]. Previous studies also reported that continuous NSAID treatment could reduce the risk of fracture in patients with AS [23].
In this study, the BASRI-hip score was used to evaluate AS-related hip joint damage. BASRI-hip ≥ 2 was de ned as radiological hip joint involvement. This method was relatively objective based on the X-ray examination, but the radiological hip joint damage assessed by this method is already in a more advanced stage [24], which could not be considered because of the retrospective nature of the study. In addition, only the variables that were routinely collected in the clinical setting could be analyzed. In addition, the sample size was relatively small and limited to a single center.

Conclusions
In conclusion, patients with AS and radiological hip joint involvement had worse body function and lower bone density. The risk of radiological hip joint involvement in patients with AS and juvenile-onset and thin body size could be signi cantly increased. Screening and monitoring of hip joint damage should be conducted in patients with AS and these characteristics. Continuous NSAIDs drug treatment was a protective factor for radiological hip joint involvement in patients with AS. It might be recommended that NSAID treatment should be continued if the patients are without contraindications. Abbreviations