Background/Purpose: Poor sleep quality is commonly observed in patients with SLE. We hypothesize that poor sleep contributes to worsening SLE. The aims of this study are to evaluate the relationship between subjective sleep measures and SLE activity over time.

Methods: A prospective, longitudinal, observational study evaluated the relationship between sleep and SLE disease activity. 151 patients were enrolled. Pittsburgh Sleep Quality Index (PSQI), Epworth Sleepiness Scale (ESS), Patient Reported Outcomes Measurement Instrument System (PROMIS)-Sleep Related Impairment (SRI), and PROMIS-Sleep Disturbance (SD) survey instruments measured patient reported sleep quality. The population mean for the PROMIS instruments is 50. The SLEDAI-2000 Responder Index-50 (S2K RI-50) was used to define active SLE as S2K RI-50 >4 and worsening SLE at subsequent visits as an increase in S2K RI-50 >=4. Baseline comparisons were calculated using non-parametric tests. Kaplan-Meier and Cox proportional hazards methods examined the relationship between poor sleep and worsening SLE activity over time.

Results: At baseline, the median age was 42, 90.7% were female, 54.3% were African American, 24.5% were on prednisone doses >7.5 mg/day, and 36.4% had active SLE. Patients with active SLE had significantly higher SRI scores (median 64.3) vs inactive SLE (median 56.6) as well as significantly higher SD scores (median 58.3 vs 52.2), whereas PSQI and ESS were not significantly different.

Data from 109 patients with >= 2 visits were used for longitudinal studies. Kaplan-Meier analysis, stratified by SRI T-score of >60 vs <=60 demonstrated that worse sleep (SRI >60) at the previous visit predicted worsening SLE activity at the next visit (Figure 1). Over a 12 month period, the probability of SLE activity worsening was 21.4% overall, 34.2% for SRI >60, and 15.0% for SRI <=60 (p=0.024). Cox proportional hazards regression analysis showed that SRI >60 (hazard ratio (HR) 3.06), male sex (HR 4.66), and prednisone use (>7.5 mg/day) (HR 3.59), but not age (HR 0.97), were significantly associated with worsening SLE.

Conclusion: Our study reinforces that patients with SLE report worse subjective sleep compared to the general population, and patients with active SLE have worse sleep than patients with inactive SLE. Our longitudinal data demonstrate that poor sleep predicts worsening SLE disease activity. Thus, variation in subjective sleep may have an important role in SLE flares.