Background/Purpose:

Treatment of
systemic lupus erythematosus (SLE) aims at controlling disease activity to
prevent damage.  In a large multi-centre cohort of prevalent SLE patients we previously
described a high proportion of patients with active disease and requiring
steroid treatment. We examined changes in disease activity over time in this
cohort.

Methods:

Predictors of
changes in disease activity over the follow-up period, including symptoms,
sociodemographic characteristics, and treatment were described in our cohort. Four
groups were created based on their SLEDAI score at cohort entry (T0): 1: no or low
(<4); 2: moderate (4 to <6); 3: active (6 to <10) and 4: highly active
(≥10). Linear and mixed regression models) were used to identify
predictors of changes in SLEDAI over time.

Results:

Among 2,048
participants, 1314, 553, 271, 183 and 146 had follow-up data at 1, 2, 3, 4 and
5 years, respectively.  At T0, for all patients, mean (sd) age was 42 years (17), age at
diagnosis was 31 (15), disease duration 11 (10) years. There were 90% females, 84
% had completed high school, and 63% were Caucasians.  At T0, almost half of
the patients were in Group 1 (n=898, 44%; 95%CI: 42-46%), 1/5 were in Group 2 (n=406,
20%; 95%CI: 18-22%) or Group 3 (n=447, 22%; 95%CI: 20-24%), while 287 (14%; 95%CI:
13-16%) were in Group 4. The probability that patients in Group 1 at T0 would
move to a higher disease activity level was 30% at 1 year, 36% at 2 years, 49%
at 3 years, 59% at 4 years and 54% at 5 years; transitions shown in Figure 1.  In
linear regression of changes in the 1^st year, higher SLEDAI at T0
was a very significant predictor of higher SLEDAI at year 1 (0.3 unit increase
for each additional SLEDAI unit at T0; p<0.00001). Higher Income was
associated with lower SLEDAI at Yr 1 (0.5 unit decrease per each higher of 5 income
categories, p=0.002), but there were no demographic or clinical predictors of
disease activity changes. In
mixed model analyses of i. longitudinal data across follow-up, and ii. if/how
variables at previous visits predict changes in SLEDAI scores at the next visit,
higher T0 SLEDAI was the single significant predictor of higher SLEDAI across
the follow-up (1 unit increase in Yr 0 is associated with 0.3 increase in Yr1,
p=0.001; and 0.3 increase per 1 unit increase at previous visit, p<0.0001),
again without significant demographic or clinical predictors of disease
activity changes.

Conclusion:

In our analyses
more than 1/2 the patients had active disease at T0, and 116 (64.1%) of 181
patients with moderate to high disease activity at T0, who had at least 3 years
of follow-up remained in this category for at least 3 years. Higher SLEDAI was
the single major independent predictor of higher SLEDAI over time. Our data
support persistence of active disease over the disease course in a significant
subset of patients, even in those with longstanding disease. This highlights
gaps in the optimal treatment of lupus and the need for additional therapies.