Background Rheumatoid arthritis (RA) is a chronic inflammatory disease associated with increased mortality and reduced life expectancy compared with the general population. This mortality gap has increased in the last years since mortality rates for RA have remained constant throughout time while mortality rates for the general population have decline. Excess mortality has been associated with disease activity. Radiographic joint destruction reflects the cumulative burden of inflammation and it is conceived as an objective measure of RA severity.
Objectives The objective of our study is to analyze the influence of radiological joint damage in the mortality rate in a cohort of RA patients.
Methods We included 783 RA patients in a retrospective longitudinal study, from May 1993 to November 2013, attending the rheumatology outpatient clinic of the Hospital Clínico San Carlos (Madrid, Spain). Subjects were included at the moment of their first X-ray, until October 2012, and followed until patients' death, loss of follow up or November 2013. Clinical records were examined and demographic and clinical data was collected. Radiographic joint damage of hands and wrists was assessed with the Sharp vand-der-Heijde score [total (SHS), erosion (ES) and narrowing/(sub)luxation (NSLS) components]. Survival techniques were applied to estimate the mortality rate (MR; expressed per 1000 patients-years with a 95% of Confidence Interval [95% CI]). Cox bivariate and multivariate regression models were conducted to examine risk factors for death. Interaction terms between radiological damage and rheumatoid factor (RF) positivity, and the elapsed time from RA onset to X-ray, were introduce in the models. Proportional hazard assumption was tested using Schoenfeld residuals and the scaled Schoenfeld residuals. Results were expressed as hazard ratio (HR) and 95% CI.
Results Most of the patients included were women (74%), with a median age of 61 years old (interquartile range [IQR]: 47-71), 67% were RF positive, and the median (IQR) elapsed time between RA symptoms onset and the X-ray was 2 (0-7) years. The median (range) followed up time per patient was 5 [0.4-20] years. 92 patients died during a follow up time of 4758 person-years. Mortality rate was 19 per 1000 patient-year [95% CI 16-24]. We observed in the bivariate analysis that older age, male sex, higher elapsed time from RA onset to X-ray, SHS, ES, NSLS, number of hospital admissions (used as a surrogate measure of comorbidity), basal Health Assessment Questionnaire, RF positivity, earlier RA onset (in calendar time), and no treatment with biological therapy, were associated with a higher MR. 3 multivariate models were constructed, using SHS, ES or NSLS as measures for joint destruction, and adjusted by the previous variables. In none of the models radiographic damage was associated with MR. However, we observed that the interaction between ES and RF positivity and with elapsed time from RA onset to X-ray were both significant (p=0.001, p=0.03, respectively): the lower elapsed time, the greater the effect of ES in the MR. On the other hand, ES was associated to MR only in RF negative patients.
Conclusions Erosive joint damage seems to be a risk factor for all cause mortality among RF negative RA patients. Early radiographic damage seems to better predict mortality.
Disclosure of Interest None declared