Association of mortality with disease severity in rheumatoid arthritis, independent of comorbidity

Abstract

Objective: To measure the extent to which mortality in rheumatoid arthritis (RA) is associated with disease severity, independent of the presence of coexistent diseases (comorbidity).

Methods: We measured disease severity and comorbidity among RA patients attending scheduled appointments at a rheumatology clinic. We used the Duke Severity of Illness Checklist (DUSOI), a clinical judgment-based measure of the severity of disease. We disaggregated the DUSOI into an RA component (RADUSOI), which we used to measure RA disease severity, together with a physician-rated 10-point global RA severity assessment. We measured comorbidity using the non-RA component of the DUSOI (COMDUSOI) and using the Charlson Comorbidity Index. Patients were contacted periodically for up to 6 years, during which we recorded deaths. We estimated the effect of disease severity and comorbidity on mortality using Kaplan-Meier survival curves, Cox proportional hazards models, and logistic regression with receiver operating characteristics (ROC) curve analysis.

Results: The sample comprised 779 patients. Followup ranged from 0.1 year to 6.3 years (mean 2.52 years), for an observation period of 2,315 patient-years. Seventy-five patients died (9.6%), for a total mortality of 3.2 per 100 patient-years (95% confidence interval 2.6-4.1). Both disease severity and comorbidity displayed significant bivariate associations with mortality. In multivariate Cox models adjusted for age, sex, and disease duration, the RADUSOI and global RA severity scores were associated with mortality independent of either the COMDUSOI or Charlson Comorbidity scores. The area under the ROC curve for a logistic model on mortality increased from 0.79 with age, sex, and disease duration included, to 0.84 after adding RA severity and comorbidity (P < or = 0.005 for the increase in ROC area).

Conclusion: RA disease severity is significantly associated with mortality regardless of the presence of comorbid disease. Combined with each patient's age, sex, and disease duration, information on RA severity and comorbidity allows an accurate prediction of mortality among patients with RA.