Background/Purpose: Patients with systemic lupus erythematosus (SLE) have poorer body image (BI) than age matched controls. Few studies have been done looking at the effect of disease activity, damage, sleep, stress, pain, fatigue, function, medications, depression and fibromyalgia (FM) in SLE patients. These 10 variables are frequently correlated, and their individual effect on BI is difficult to tease out. To improve BI in SLE, we need to better understand where we need to focus our attention among these 10 variables to guide development of specific interventions. We aimed to evaluate the relative role of disease activity, damage, sleep, stress, pain, fatigue, function, medications, depression and FM on BI in SLE patients.

Methods: 115 SLE patients receiving rheumatology care at two academic medical centers were recruited. Each patient completed the following: Functional Assessment of Chronic Illness Therapy-Fatigue (FACIT), Perceived Stress Scale (PSS), Patient Health Questionnaire (PHQ-9 for depression), Insomnia Severity Index (ISI), pain inventory and LupusPro.  Body image and physical function was measured using the BI (BILS) and Physical health (PH) domains of LupusPRO.  Disease activity and damage were assessed using SELENA-SLEDAI (SS) and SLICC/ACR Damage Index (SDI). Charts were reviewed to evaluate if the patient had been diagnosed with FM by a rheumatologist.  Multivariate regression analyses (including stepwise modelling) were conducted with BI as the dependent variable for (a) all patients, and for patients (b) without and (c) with FM.

Results: 115 SLE patients with mean (SD) age of 40 (14) yrs.  Ninety percent were women, and 15% had concomitant FM. Over 60% were currently on prednisone. Mean (SD) SS and BILS score were 4.8 (4.1) and 74.5 (27.7) respectively.  For all patients, with all 10 variables in the regression analysis, PSS remained the single independent inverse predictor of BI after adjusting for other 9 variables. On stepwise regression, PSS and Pain were predictors of BI (Table 1). Among SLE patients without FM, the SS, PSS, PHQ were independent predictors of poor BI, after adjusting for the other 6 variables.  On stepwise modeling, only PSS, PHQ, and SS were predictors of BI among non-FM SLE patients. When SS was replaced by SS arthritis, rash, alopecia items in this model, PSS, PHQ and active rash were found to be independent predictors. Among SLE patients with FM, none of the 9 variables independently predicted poor BI. On stepwise modelling PH was a predictor of BI in SLE patients with FM.

Conclusion: PSS (and not FM or steroid use) is an important and independent predictor of BI for all SLE patients. Among SLE patients without FM, besides PSS, active disease (rash specially) and depression are also important drivers of poor BI.