Objective: Physician-assessed disease activity and damage predict mortality in systemic lupus erythematosus (SLE). Patient-reported outcomes (PROs) are known predictors of mortality in other chronic diseases, but this relationship has not been well examined in SLE. The aim of the present study was to assess whether PROs predict mortality in SLE.
Methods: Data were derived from the University of California at San Francisco Lupus Outcomes Study (n = 728). PROs (Medical Outcomes Study Short Form 36 [SF-36] subscales), self-rated health, and depression (Center for Epidemiologic Studies Depression scale [CES-D]) from 2007 (baseline data [T0]) were used to predict mortality (censored 2015). Univariate Cox regression analyses were completed for each PRO as a predictor of mortality, and multivariate Cox regression with covariates for each PRO separately. Covariates were age, sex, race/ethnicity, poverty, disease duration, disease activity (Systemic Lupus Activity Questionnaire), and damage (Brief Index of Lupus Damage).
Results: The mean ± SD age of patients was 50.6 ± 12.6 years. Ninety-two percent of patients were women and 68.5% were white. There were 71 deaths (9.1%). In univariate analyses, both the SF-36 physical component subscale score and self-rated health were associated with mortality, and the SF-36 mental health subscale and CES-D scores were not associated with mortality. In multivariate analyses, lower scores of SF-36 physical function at T0 independently predicted mortality after controlling for all other covariates (hazard ratio 0.97 [95% confidence interval 0.94-0.99]; P < 0.01).
Conclusion: Patient-reported physical function independently predicted mortality in SLE, even after accounting for demographics (including poverty) and disease (duration, activity, and damage). Because PROs are easy to assess, they may be used to triage, track, and guide early interventions for those at high risk of mortality in SLE.
© 2018, American College of Rheumatology.