Background Previous studies have suggested an association between idiopathic acute anterior uveitis (AAU) and spondyloarthritis (SpA), including showing that at least 40% of AAU patients have undiagnosed SpA. However, the clinical factors to predict the incident diagnosis or future development of SpA remain poorly recognized.

Objectives The objectives of this study were: 1) to describe the patient characteristics of AAU patients with and without SpA diagnosis, and 2) to identify the predictive factors of SpA diagnosis among AAU patients using real-world data.

Methods Adult patients with at least one diagnosis of AAU (ICD-9-CM: 364.01 or 364.02) were selected from a large US insurance claims database (01/01/2008–06/30/2015). The first AAU diagnosis was defined as the index date. Patients were required to have at least 6 months of continuous data availability before the index date. Patients with intraocular surgery, penetrating or blunt eye trauma, or a diagnosis of rheumatoid arthritis or SpA on or prior to the index date were excluded. Potential predictive factors for subsequesnt diagnosis of SpA included demographic characteristics, type of AAU (primary [first diagnosis] vs. recurrent), SpA-related comorbidities, and healthcare resource utilization prior to the index date. Factors predictive of SpA diagnosis were selected into a multivariable Cox proportional hazards model based on statistical significance and clinical relevance. Hazards ratios (HR) and p-values were estimated for each factor.

Results A total of 48,822 patients with AAU were included, and among them, 1,032 patients were newly diagnosed with SpA during the follow-up period which was 24 months on average. Patients with SpA were younger (45.7 vs. 50.4 years), more likely to be male (52% vs. 42%), more likely to have recurrent AAU (44% vs. 29%), back pain (21% vs. 13%), SpA-related comorbidities including inflammatory bowel disease (IBD) (4% vs. 1%) and psoriasis (3% vs. 1%), and more likely to use corticosteroids (45% vs 40%). Predictive factors in the final Cox model were: male vs. female (HR=1.55; p-value<0.01), age<45 vs. ≥45 years old (1.65; <0.01), recurrent vs. primary AAU (1.94; <0.01), back pain under age 45 vs. no back pain (1.90; <0.01), back pain above age 45 vs. no back pain (1.46; <0.01), psoriasis (5.16; <0.01), IBD (2.50; <0.01), chiropractor/physical therapist visits (1.29; 0.01), conjunctivitis (1.23; 0.02), nonsteroidal anti-inflammatory drugs use (1.44; <0.01), use of conventional synthetic disease-modifying anti-rheumatic drugs (1.32; 0.05), joint pain (1.14; 0.19), imaging test use (1.03; 0.74), corticosteroids use (1.07; 0.33), and enthesitis/heel pain (1.13; 0.36). When sacroiliitis was included in the model, results remained largely similar.

Conclusions There are significant differences among isolated AAU patients and AAU patients that developed SpA later. The most predictive factors of SpA diagnosis were male, age<45 years, recurrent AAU, back pain, and other extra-articular manifestations of SpA such as IBD and psoriasis. Since delayed diagnosis is common among SpA patients, identifying such predictive factors can help inform risk stratification.

Disclosure of Interest M. Haroon Grant/research support from: AbbVie and Pfizer, K. Betts Consultant for: AbbVie, F. Mu Consultant for: AbbVie, M. Skup Employee of: AbbVie, J. Anderson Employee of: AbbVie, A. Joshi Employee of: AbbVie