Systemic Lupus Erythematosus Disease

Activity Index 2000

DAFNA D. GLADMAN, DOMINIQUE IBAÑEZ, and MURRAY B. UROWITZ

ABSTRACT. Objective. To describe the Systemic Lupus Erythematosus Disease Activity Index 2000 (SLEDAI-2K),
a modification of SLEDAI to reflect persistent, active disease in those descriptors that had previously
only considered new or recurrent occurrences, and to validate SLEDAI-2K against the original SLEDAI
as a predictor for mortality and as a measure of global disease activity in the clinic.
Methods. All visits in our cohort of 960 patients were used to correlate SLEDAI-2K against the origi-
nal SLEDAI, and the whole cohort was used to validate SLEDAI-2K as a predictor of mortality. A sub-
group of 212 patients with SLE followed at the Lupus Clinic who had 5 regular visits, 3–6 months apart,
in 1991-93 was also included. An uninvolved clinician evaluated each patient record and assigned a
clinical activity level. The SLEDAI score was calculated from the database according to both the orig-
inal and modified definitions.
Results. SLEDAI-2K correlated highly (r = 0.97) with SLEDAI. Both methods for SLEDAI scoring
predicted mortality equally (p = 0.0001), and described similarly the range of disease activity as recog-
nized by the clinician.
Conclusion. SLEDAI-2K, which allows for persistent activity in rash, mucous membranes, alopecia,
and proteinuria, is suitable for use in clinical trials and studies of prognosis in SLE. (J Rheumatol
2002;29:288–91)

Key Indexing Terms:

SYSTEMIC LUPUS ERYTHEMATOSUS DISEASE ACTIVITY SLEDAI OUTCOME

The Systemic Lupus Erythematosus Disease Activity Index However, this group made other modifications to the SLEDAI
(SLEDAI) was developed and validated as a clinical index for such as adding scleritis and episcleritis as descriptors of active
the measurement of disease activity in SLE and has been used disease with high activity weighting of 8. These manipula-
as a global measure of disease activity in SLE since its intro- tions were made without validation of the new index with the
duction in 19851,2. This index was modeled on clinicians’ same rigorous methodological approach as the derivation of
global judgment. It was developed with a panel of experi- the original index, nor did they validate this index against the
enced rheumatologists with expertise in SLE, using well original SLEDAI.
established group techniques and index development method- Our aims were: (1) to describe SLEDAI 2000 (SLEDAI-
ology. The index has been used successfully by both expert 2K), a modification of SLEDAI to reflect persistent, active dis-
clinicians3 and trainees4, and has been shown to be valuable in ease in those descriptors that had previously only considered
both research and clinical settings5,6. It has also been shown to new or recurrent occurrences; and (2) to validate SLEDAI-2K
be time sensitive7. against the original SLEDAI, (a) as a predictor for mortality
The variables proteinuria, rash, alopecia, and mucous and (b) as a measure of global disease activity in the clinic.
membrane lesions are counted as active in the SLEDAI only
at their first occurrence, or upon recurrence. This was done to MATERIALS AND METHODS
distinguish active from chronic lesions, the latter more likely University of Toronto Lupus Clinic. All visits for patients registered in our
to represent damage. cohort were included to compare the scores calculated for the original
Intuitively, physicians would prefer to use these descriptors SLEDAI and SLEDAI-2K. The whole cohort was also used to validate the
SLEDAI-2K against the original SLEDAI at presentation as a predictor of
as active any time they are present. Indeed the SELENA trial
mortality (Table 1).
group modified the SLEDAI to insure that the descriptors of
Table 1. Characteristics of total cohort.
organ system involvement reflected ongoing disease activity8.
Patients (n) 960
From The University of Toronto Lupus Clinic, Centre for Prognosis Female/male (%) 842 (87.7)/118 (12.3)
Studies in the Rheumatic Diseases, Toronto Western Hospital, Toronto, Alive/dead (%) 792 (82.5)/168 (17.5)
Ontario, Canada. Race: Caucasian/Black/other (%) 778 (81.5)/71 (7.48)/106 (11.0)
D.D. Gladman, MD, FRCPC, Professor of Medicine, University of Age at diagnosis, yrs 32.9 (13.68) [8–83]*
Toronto, Deputy Director; D. Ibañez, MSc, Biostatistician; M.B. Urowitz, Age at first visit, yrs 36.3 (13.60) [10–84]*
MD, FRCPC, Professor of Medicine, University of Toronto, Director, Assessments (n) 18,636
Centre for Prognosis Studies in the Rheumatic Diseases. Mean assessments per patient 19.2 (19.8) [1–106]*
Address reprint requests to Dr. D.D. Gladman, Toronto Western Hospital, Duration of followup, yrs 8.2 (7.4) [0-30.5]*
399 Bathurst Street, MP 1-318, Toronto, Ontario M5T 2S8, Canada.
Submitted May 28, 2001; revision accepted August 30, 2001. *Mean (SD) [range].

Personal non-commercial use only. The Journal of Rheumatology Copyright © 2002. All rights reserved.
288 The Journal of Rheumatology 2002; 29:2

Downloaded on February 11, 2022 from www.jrheum.org

Table 2. SLEDAI– 2K data collection form.

As of December 2000 there were 960 patients registered in the cohort, of and who had 5 consecutive regular visits between 1991 and 1993. A clinician
whom 168 have died. All deaths are recorded according to a standardized who did not know the patients and was blind to their SLEDAI score evaluat-
form in the database, and we have only a 10% lost to followup rate and can ed each patient record and assigned a clinical activity score for each assess-
account for the vast majority of deaths9. We have previously shown that ment according to the following scale: 0 = no activity; 1 = mild activity with
SLEDAI at presentation predicts mortality. The survival analysis was repeat- no therapeutic intervention; 2 = activity, but improvement from previous visit;
ed using the original SLEDAI and SLEDAI-2K in the entire cohort. 3 = persistent activity/refractory to treatment; 4 = flare, defined as one of the
To test the validity of the modified index in describing changes in global following: the introduction of new treatment in the presence of worsening of
disease activity in our clinic, we used patients who attended the clinic in 1992 an already active system or in response to the activation of a new system; an

Personal non-commercial use only. The Journal of Rheumatology Copyright © 2002. All rights reserved.
Gladman, et al: SLEDAI 2000 289

Downloaded on February 11, 2022 from www.jrheum.org

increase in medication dosage for the above reasons; indication of concern in Table 4. Characteristics of SLE patients used in the evaluation of disease
the physician’s notes (arrangement for an earlier appointment for the assess- activity.
ment of SLE disease activity or investigation related to SLE including mea-
surement of anti-DNA antibodies and serum complement levels, and 24 hour Patients (n) 212
urine protein determination); the use of the term flare in the physician’s notes Female/Male (%) 191 (90.1)/21 (9.9)
(i.e., treating physician’s impression of increased disease activity); new diag- Race: Caucasian/Black/other (%) 163 (76.9)/23 (10.8)/26 (12.3)
nosis of SLE. Age at diagnosis, yrs 31.0 (13.5) [8–75]*
Modification of SLEDAI (SLEDAI-2K). The original SLEDAI was used as Age at first visit, yrs 33.9 (12.7) [16–75]*
described1,2. Among the SLEDAI descriptors, alopecia, mucous membrane Age at study, yrs 41.2 (12.7) [19–75]*
lesions, and rash had been scored only if they were new or recurrent, and in Disease duration at first visit, yrs 2.9 (4.6) [0–25]*
the case of proteinuria if there was new onset or a recent increase of more than Disease duration at study, yrs 10.2 (7.6) [0–41]*
0.5 g/24 h. We modified the definition of these descriptors to include the pres- Assessments during study (n) 1060
ence of any rash, alopecia, or mucosal ulcers and new, recurrent, or persistent Length of followup during study, mo 15.5 (5.0) [4–27]*
proteinuria > 0.5 g/24 h (Table 2). SLEDAI at first visit to clinic 10.2 (8.3) [0–55]*
Statistical analysis. Descriptive statistics were calculated for the characteris- SLEDAI-2K at first visit to clinic 10.3 (8.3) [0–55]*
tics of both cohorts. T tests were used to compare mean SLEDAI and
SLEDAI-2K across mortality. They were also used to compare the mean * Mean (SD) [range].
change in SLEDAI to the mean change in SLEDAI-2K for each level of sub-
jective categories. Linear regression was used to test for trend in change in (Table 5). As expected, SLEDAI-2K scores were minimally
SLEDAI and in SLEDAI-2K across the subjective categories. higher at a number of levels, reflecting the 4 modified vari-
ables (rash, alopecia, mucous membrane ulcers, and protein-
RESULTS uria). A regression model for SLEDAI and SLEDAI-2K
SLEDAI-2K compared to SLEDAI in the entire cohort. between the activity categories indicates that the scores of the
SLEDAI calculations in the entire cohort were available for individual categories are significantly different (p = 0.0001).
18,636 visits. SLEDAI and SLEDAI-2K were the same in Median SLEDAI and SLEDAI-2K values for the activity lev-
14,532 visits (78%). In the remaining 4104 visits the differ- els were similar (Table 5). The change from the previous visit
ences were due to proteinuria in 1355 (7.3% of all visits), in SLEDAI and SLEDAI-2K over the range of disease activi-
rash: 1378 (7.4%), alopecia: 1116 (6.0%), and mucous mem- ty level was also very similar (Table 6). Linear regressions
brane ulcers: 752 (4.0%) (more than one item was different in revealed that all levels of disease activity were different for
some visits). The average magnitude of change when present both SLEDAI and SLEDAI-2K. SLEDAI-2K thus reflects the
was 2.91, with SLEDAI-2K being greater than SLEDAI. The change in disease activity as reported for SLEDAI10.
correlation between SLEDAI-2K and SLEDAI across all vis-
its was r = 0.97 (p = 0.0001). DISCUSSION
SLEDAI-2K compared to SLEDAI at presentation as a pre- In the derivation of the SLEDAI, investigators were focused
dictor for mortality. At presentation the mean (SD) SLEDAI on describing pure disease activity as opposed to damage from
for the whole group was 10.2 (8.6), ranging from 0 to 56. The disease or therapy. Therefore, in the glossary of the original
mean (SD) SLEDAI-2K was 10.4 (8.7) with a similar range. SLEDAI, items were not scored as active if they could repre-
Both the original SLEDAI and SLEDAI-2K at presentation sent chronicity or damage. With wider use clinicians have
were equally significant predictors for all-cause mortality in interpreted chronic active lesions in the skin, hair, and mucous
our entire cohort (Table 3). membranes or persistent proteinuria as equivalent to active
SLEDAI-2K compared to SLEDAI as a measure of global dis- lesions. Furthermore, in clinical trials if these lesions persist-
ease activity in the clinic. Two hundred twelve patients seen in ed and were not scored as active, this would cause an erro-
1992 had 5 regular visits, 3–6 months apart, between 1991 and neous interpretation of improvement related to the therapy
1993. There were 191 women and 21 men, with a mean age of under investigation. This led the investigators in the SELENA
41 years at the time of the study, and a mean disease duration trial to modify some aspects of the SLEDAI, but without any
of 10 years. The majority of the patients were Caucasian validation of their modification. For that reason we set out to
(76.9%). Table 4 lists the characteristics of this patient cohort. modify the SLEDAI to include persistent disease activity in
SLEDAI-2K describes a similar progression over a range the features noted above, and to validate the modification.
of disease activity and of the same magnitude as SLEDAI We propose modifications to the SLEDAI such that persis-
tent active disease in the items alopecia, mucous membrane
ulcers, rash, and proteinuria would be scored. The revised
Table 3. Correlation of SLEDAI and SLEDAI-2K with mortality in 960
SLEDAI will be identified as SLEDAI-2K to differentiate it
patients with SLE at presentation. Data are mean (SD).
from the original index. We validated SLEDAI-2K against the
Instrument Alive, n = 792 Dead, n = 168 p original SLEDAI in 18,636 patient visits in our Lupus Clinic
and found that the score was different in only 4104 instances
SLEDAI 9.44 (7.88) 13.95 (10.76) 0.0001 (22%). SLEDAI-2K is comparable to the original SLEDAI as
SLEDAI-2K 9.69 (7.93) 14.01 (10.80) 0.0001
a predictor of mortality. SLEDAI-2K describes disease activ-

Personal non-commercial use only. The Journal of Rheumatology Copyright © 2002. All rights reserved.
290 The Journal of Rheumatology 2002; 29:2

Downloaded on February 11, 2022 from www.jrheum.org

 Table 5. SLEDAI, SLEDAI-2K, and clinical activity levels in 1060 assessments.

Activity Level No. (%) SLEDAI* SLEDAI-2K* p

0–No activity 465 (43.9) 1.95 (2.29) [2] 2.23 (2.34) [2] 0.066
1–Mild activity 231 (21.8) 3.80 (2.85) [4] 4.42 (2.93) [4] 0.023
2–Activity, but
improvement 87 (8.2) 6.74 (3.93) [6] 7.45 (3.93) [8] 0.233
3–Persistent activity 165 (15.6) 7.78 (4.05) [8] 8.80 (4.05) [8] 0.023
4–Flare 112 (10.6) 9.37 (6.13) [9] 9.76 (6.15) [9] 0.632

*Mean score (SD) [median].

Table 6. Change (∆) in SLEDAI and SLEDAI-2K by activity levels. REFERENCES

1. Committee on Prognosis Studies. Prognosis studies in SLE: an activity
Activity Level N ∆ SLEDAI* ∆ SLEDAI-2K p index [abstract]. Arthritis Rheum 1986;29 Suppl 4:S93.
2. Bombardier C, Gladman DD, Urowitz MB, Caron D, Chang CH, and
0–No activity 380 –0.71 (3.30) –0.75 (3.27) 0.860 the Committee on Prognosis Studies in SLE. The development and
1–Mild activity 178 –0.87 (4.02) –0.78 (4.10) 0.835 validation of the SLE Disease Activity Index (SLEDAI). Arthritis
2–Activity, but Rheum 1992;35:630-40.
improvement 76 –2.59 (5.03) –2.57 (4.96) 0.974 3. Gladman DD, Goldsmith CH, Urowitz MB, et al. Cross-cultural
3–Persistent activity 125 –0.52 (3.96) –0.23 (3.77) 0.556 validation of three disease activity indices in systemic lupus
4–Flare 89 4.48 (5.21) 4.57 (5.11) 0.908 erythematosus. J Rheumatol 1992;19:608-11.
4. Hawker G, Gabriel S, Bombardier C, Goldsmith C, Caron D, Gladman
* Mean score (SD). DD. A reliability study of SLEDAI: A disease activity index for
systemic lupus erythematosus. J Rheumatol 1993;20:657-60.
5. McLaughlin JR, Bombardier CB, Farewell VT, Gladman DD, Urowitz
MB. Kidney biopsy in systemic lupus erythematosus. III. Survival
analysis controlling for clinical and laboratory variables. Arthritis
ity at various activity levels in a comparable manner to the Rheum 1994;37:559-67.
6. Petri M, Hellmann D, Hochberg M. Validity and reliability of lupus
original SLEDAI. SLEDAI-2K is equivalent to SLEDAI in activity measures in the routine clinic setting. J Rheumatol
describing changes in disease activity from one visit to the 1992;19:53-9.
next. 7. Gladman DD, Goldsmith CH, Urowitz MB, et al. Sensitivity to
Therefore, SLEDAI-2K, which allows for persistent activ- change of 3 systemic lupus erythematosus disease activity indices:
ity in rash, mucous membrane ulcers, alopecia, and protein- international validation. J Rheumatol 1994;21:1468-71.
8. Petri M, Buyon J, Kim M. Classification and definition of major flares
uria, is suitable for use in clinical trials and studies of progno- in SLE clinical trials. Lupus 1999;8:685-91.
sis in SLE. We recommend the use of the modified SLEDAI, 9. Gladman DD, Koh DR, Urowitz MB, Farewell VT. Lost-to-follow-up
SLEDAI-2K, for the assessment of global disease activity in study in SLE. Lupus 2000;9:363-7.
SLE. 10. Gladman DD, Urowitz MB, Kagal A, Hallett D. Accurately describing
disease activity in SLE. J Rheumatol 2000;27:377-9.

Personal non-commercial use only. The Journal of Rheumatology Copyright © 2002. All rights reserved.
Gladman, et al: SLEDAI 2000 291

Downloaded on February 11, 2022 from www.jrheum.org