Session Information
Session Type: ACR Concurrent Abstract Session
Session Time: 2:30PM-4:00PM
Background/Purpose: 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 examined in SLE. This analysis assessed whether PROs predict mortality in patients with SLE.
Methods:
Data from 728 participants in the UCSF Lupus Outcomes Study were evaluated to determine if PROs at one point in time (“baseline”, 2007) would predict subsequent mortality. Mortality was determined as of December 2015. PROs included the 8 subscales of the SF-36 (Physical Function, Role Physical, Pain, General Health, Vitality, Social Function, Role Emotional, and Mental Health), and depressive symptoms measured with the Center for Epidemiologic Studies Depression scale (CESD). Covariates were age, gender, race/ethnicity, poverty, disease duration, self-reported disease activity (Systemic Lupus Activity Questionnaire, SLAQ), and self-reported disease damage (Brief Index of Lupus Damage, BILD). BILD has previously shown good correspondence with physician-assessed disease damage. Univariate Cox regression analyses first examined each PRO as a predictor of subsequent mortality. Multivariate Cox regression analyses including covariates were then conducted for each PRO separately.
Results: Mean (SD) age was 50.6 (12.6) years; 671 (92.2%) participants were women. Ethnic composition was 68.5% Caucasian, 9.2% Hispanic, 7.3% African American, and 9.5% Asian. Baseline demographics, disease and PROs are shown in Table 1A. Mean (SD) follow up was 74.6 (23.2) months. There were 71 (9.1%) deaths. In univariate analyses, all PROs except the SF-36 Mental Health subscale and CESD were associated with mortality. In multivariate analysis, patient-reported physical function at initial screening independently predicted mortality after controlling for all other covariates (Table 1B), such that the odds of death were 3.5% lower for every increased point rating in physical health on the SF-36 Physical Function score[Hazard Ration (HR) 0.97, 95%CI (0.94,0.99) p < 0.01].
Conclusion: Self-reported physical function was independently predictive of mortality in SLE, even after adjusting for demographics (including poverty) and disease (duration, activity and damage). Tracking PROs, particularly patient-reported function, in clinical settings may add important information to improve patient long-term outcomes.
|
Table 1. A. Demographics and general characteristics (n=728) B. Multivariate Cox Proportional hazard regression analysis of mortality at 8 years |
||||
|
M |
SD |
N |
% |
|
|
A. |
|
|
|
|
|
Female |
|
|
671 |
92.2 |
|
Age (years) |
50.6 |
12.6 |
|
|
|
Ethnicity |
|
|
|
|
|
Caucasian |
|
|
499 |
68.5 |
|
Hispanic |
|
|
67 |
9.2 |
|
African American |
|
|
53 |
7.3 |
|
Asian |
|
|
69 |
9.5 |
|
Other |
|
|
40 |
5.5 |
|
Below Poverty |
|
|
80 |
11.0 |
|
BMI |
26.8 |
6.8 |
|
|
|
Disease activity |
4.2 |
2.7 |
|
|
|
Disease duration |
16.7 |
8.4 |
|
|
|
Disease damage |
2.3 |
2.2 |
|
|
|
Deceased |
|
|
71 |
9.8 |
|
SF36 Physical Function |
39.4 |
12.7 |
|
|
|
B. |
|
|
|
|
|
Predictor |
Hazard Ratio |
95% CI |
p |
|
|
Female |
2.51 |
1.18, 5.35 |
.02 |
|
|
Age |
1.04 |
1.02, 1.07 |
.00 |
|
|
Below Poverty |
086 |
.40, 1.84 |
.69 |
|
|
Disease Duration |
1.02 |
.99, 1.04 |
.28 |
|
|
Disease Activity |
.10 |
.89, 1.12 |
.98 |
|
|
Disease Damage |
1.23 |
1.12, 1.36 |
.00 |
|
|
SF-36 Physical Function |
.97 |
.94, .99 |
.01 |
|
|
Note: A. Relates to demographics and general characteristics. B. Relates to Multivariate Cox Proportional hazard regression analysis of mortality at 8 years. SLE disease activity is rated as 10-point Likert scale of patient rated disease activity; Disease damage is rated on the Brief Index of Lupus Damage (BILD); Below poverty calculated by US federal government guidelines; 36-Item Short Form Survey (SF-36) including only physical function domain. All scores were rated at Time 1. Deceased status was evaluated throughout 8-year study progression. HR = Hazard Ration. 95% CI = 95% Confidence Interval. SLE disease activity is rated as 10-point Likert scale of patient rated disease activity; Disease damage is rated on the Brief Index of Lupus Damage (BILD); Below poverty calculated by US federal government guidelines; 36-Item Short Form Survey (SF-36) including only physical function domain. All scores were rates at T1. |
||||
To cite this abstract in AMA style:
Azizoddin DR, Jolly M, Katz PP, Yelin EH. Patient-Reported Outcomes and Damage Predict Mortality in Lupus [abstract]. Arthritis Rheumatol. 2017; 69 (suppl 10). https://acrabstracts.org/abstract/patient-reported-outcomes-and-damage-predict-mortality-in-lupus/. Accessed .« Back to 2017 ACR/ARHP Annual Meeting
ACR Meeting Abstracts - https://acrabstracts.org/abstract/patient-reported-outcomes-and-damage-predict-mortality-in-lupus/