Background/Purpose: Although SLE patients are thought to follow different patterns of disease courses, no information is available about the longitudinal disease activity or the number of possible different disease courses. This study sought to: 1) determine the longitudinal disease activity trajectory in childhood-onset SLE (cSLE) patients; 2) determine the number of latent classes of disease activity trajectories and 3) identify factors predictive of membership in different disease trajectories.   

Methods: Single centre longitudinal inception cohort of cSLE patients (onset <18 years) diagnosed and followed from Jan 1985 to Sep 2011. Clinical data from childhood to adulthood was obtained: pediatric data from our institutional cSLE database, adult data from the Toronto Lupus database or extracted from clinical charts from rheumatologists’ offices. Longitudinal disease trajectory was constructed using data from every clinic visit in the 1^st 10 years after diagnosis. Longitudinal SLE activity is a latent construct that is imperfectly measured with SLE disease activity index 2000 (SLEDAI2K) and prednisone exposure. Longitudinal mean population trajectories of SLEDAI2K and prednisone use were first fitted with the mixed effects model. SLEDAI2K and prednisone use were then jointly modeled in a Bayesian growth mixture model (GMM). 

Results: 473 patients were included. 82% were females, median age of diagnosis was 14.1 years. There were 11992 visits, 2666 patient years. By the end of 10 years, 67% of the population had transferred to adult care. Mean population SLEDAI2K and prednisone trajectories of cSLE patients showed rapid decline to low activity levels within 2 years after diagnosis. However, joint GMM showed 5 latent classes were present in this cohort of cSLE patients in the following distributions: 30(6%), 57 (12%), 79 (17%), 92(19%), 215 (45%). Class1 patients have chronic moderate-high disease activity, class 2 had moderate initial disease activity and continued moderate long-term prednisone use, class 3 had initial high disease activity but achieved long-term remission, class 4 had high initial disease activity but relapsed later, class 5 had chronic low-grade disease activity. Across all classes, there was chronic use of prednisone (at least 5-10 mg/day) among cSLE patients in the first 10 years after diagnosis. Baseline major organ involvement, ethnicity, age at diagnosis and the number of baseline ACR criteria predicted membership in different classes.             

Conclusion: cSLE patients could be subclassified into 5 distinct classes of disease activity trajectories. Baseline organ involvement and personal demographic factors could predict membership in the distinct disease activity trajectory classes.

« Back to 2015 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/from-childhood-to-adulthood-identifying-latent-classes-of-disease-activity-trajectories-in-childhood-onset-systemic-lupus-erythematosus-patients/