Background Hip involvement is the most frequent extraspinal arthritic manifestation of Ankylosing Spondylitis (AS) and a common cause of disability. It is present in 24% to 36% of AS patients, leading to total hip replacement (THR) in 5%.

Objectives To examine the prevalence of hip arthritis in patients with AS under TNF-inhibitor (TNFi) treatment, to identify predictors of its development and possible gender differences.

Methods Consecutive TNFi-naïve AS patients (fulfilling the modified New York criteria) who were eligible for TNFi treatment were included. At baseline (before the start of TNFi), disease activity and function were recorded by BASDAI, ASDAS-CRP/ESR, BASFI and BASMI. Anteroposterior x-rays of the pelvis and lateral x-rays of cervical and lumbar spine were obtained at baseline and scored according to the BASRI-hip scoring system and mSASSS. Hip involvement was assessed both clinically (as pain, reduced range of motion and intermalleolar distance) and radiographically (BASRI-h score ≥2 was defined as definitive hip involvement). Mann-Whitney, two-sample t-test and logistic regression analysis were applied. The groups' averages were expressed as mean±SD, or median (IQR) according to the normality of data.

Results 298 consecutive AS patients (214 men, age: 49±12years (mean±SD), disease duration: 23.8±11.8 years) were included. Definite hip involvement was detected in 113/298 (38%) patients. Bilateral THR and unilateral THR underwent 10/298 (3.4%) and 9/298 (3%) patients respectively. No gender difference in the prevalence of hip arthritis was observed (females: 26/84 (31%) vs. males 87/214 (40.8%). The patients with hip arthritis compared to those without had significantly higher disease activity scores (BASDAI (6.1±1.7 vs. 5.5±1.9, p=0.008), ASDAS-CRP (3.9±0.8 vs. 3.4±0.9, p=0.0001), CRP [16 (7.7–32) median (IQR) vs. 7 (2.9–22), p=0.0001], ESR [26 (10–42) vs. 15.5 (7–33), p=0.004], higher BASFI-scores (6.3±2.1vs. 4.8±2.3, p<0.0001), BASMI-scores (5.1±2.3vs. 3.4±2, p<0.0001) and reduced intermalleolar distance (89±23 vs. 104.8±19 cm, p<0.0001). AS patients with hip arthritis also had significantly higher mSASSS-scores [15 (3- 39) vs. 4 (0–16), p<0.0001] more often syndesmophytes [75/107 (70.1%) vs.75/181 (41.4%), p<0.0001] and peripheral arthritis [54/111 (48.7%) vs. 68/184 (37%), p=0.048). According to multivariate logistic regression analysis, independent risk factors for hip arthritis were: raised CRP (OR: 1.01 CI: 1.001–1.02), presence of syndesmophytes (OR: 2.6, CI: 1.45–4.8) and a high BASFI (OR: 1.33, CI: 1.1–1.5).

Conclusions The prevalence of hip arthritis in AS is very high (1/3) and significantly related to high disease activity and a high mSASSscore. No gender difference in prevalence of hip arthritis was found. Considering the large impact on function, this manifestation might need more attention.

Disclosure of Interest None declared