CT-Based Evaluation of Diffuse Idiopathic Skeletal Hyperostosis in Adult Population; Prevalence, Associations and Interobserver Agreement

Abstract

Diffuse idiopathic skeletal hyperostosis (DISH), being an asymptomatic condition, is generally discovered incidentally on imaging and it has not received much attention for research on clinical grounds. We assessed the prevalence of DISH, its associated factors, and interobserver agreement for computed tomography (CT)-based diagnosis of DISH. CT scans of chest, abdomen, and pelvis performed for various clinical indications were retrospectively reviewed. Resnick criteria were used for the diagnosis of DISH. Moreover, enthesopathy along with comorbidities was assessed. CT scans were observed by 3 observers having different experience levels. Out of total 416 patients, the prevalence of DISH was 30.8%. Strong positive agreement was observed between observer 1 and 2 (k = 0.89), observer 1 and 3 (k = 0.91), and observer 2 and 3 (k = 0.94). Reporting rate of DISH was 59.3%. Regression analyses showed that enthesopathy was 2.45 times (adjusted odds ratio [AOR]: 2.45, 95% confidence intervals [CI]: 1.48–4.05), diabetic patients were 4.74 times (AOR: 4.74, 95% CI: 2.89–7.78) while hypertensive patients were 2.17 times (AOR: 2.17, 95% CI: 1.30–3.62) more likely to have DISH in comparison to those who do not have DISH. A high prevalence of DISH was observed in our cohort. Enthesopathy and comorbidities like diabetes and hypertension were significant factors associated with DISH. Moreover, excellent agreement was observed in defining DISH on CT according to Resnick criteria.