Absence of an association of rheumatoid arthritis and diffuse idiopathic skeletal hyperostosis: a case-control study

To compare the frequency of diffuse idiopathic skeletal hyperostosis (DISH), based on an evaluation of chest radiographs, in 2 populations of patients: those with and without rheumatoid arthritis (RA).

Cases of RA were obtained from the Wichita Arthritis Center databank and controls from a private cardiology practice. Patients with RA and controls had to be older than 50 years and have a chest radiograph available. All radiographs were read, in haphazard order, by a single reader who classified them as definitely DISH, probably DISH (these 2 categories were considered as having DISH), uncertain, probably not DISH, and definitely not DISH (these 3 groups defined the absence of DISH). Logistic regression was used to investigate the probability of DISH while controlling for age and sex. Analyses were conducted to take into account the sensitivity and specificity of chest radiographs for the diagnosis of DISH.

Among 748 radiographs (357 RA and 391 controls), 29 were diagnosed as DISH (21 controls and 8 RA). The mean age for the RA group was 66.2 yrs (SD = 11.9 and 108 (30.2%) were men. Control subjects had a mean age of 69.6 yrs (SD = 11.4 and 139 (35.4%) were men. The DISH patients, compared with those without DISH, were more likely to be male (p = 0.004) and older (p < 0.001). After accounting for age and sex, the relative risk of DISH in RA versus the controls was 0.40 (95% confidence interval: 0.20 to 1.01). After taking into account the sensitivity and specificity in this population, the adjusted prevalence of DISH for the RA group was 1% (95% confidence interval: 0.1 to 3.7) and 3.1% (95% confidence interval: 0.2 to 7.9) for the control group.

We confirmed the increase in frequency of DISH with increasing age and male sex, but failed to support an increased association of DISH with RA.