SEARCH

SEARCH BY CITATION

Keywords:

  • SLE;
  • Lupus;
  • Children;
  • ECLAM;
  • SLEDAI

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

Objectives

To evaluate the European Consensus Lupus Activity Measurement (ECLAM) for responsiveness to change in disease activity when used in childhood-onset systemic lupus erythematosus (cSLE). To confirm the construct validity and to characterize the measurement properties of the ECLAM by assessing its ability to predict damage and steroid requirements.

Methods

The disease courses of 66 newly diagnosed cSLE patients were reviewed. The ECLAM and Systemic Lupus Erythematosus Disease Activity Index (SLEDAI) were scored for all clinic visits and hospitalizations. Damage was assessed at the end of the followup period using the Systemic Lupus International Collaboration Clinics/American College of Rheumatology Damage Index. Disease activity at the time of diagnosis, 6 months after diagnosis, at the time of first flare, and 6 months after first flare was used to estimate responsiveness of the measures. Responsiveness was measured by the effect size, the effect size index, the standardized response mean, and the relative efficiency index (REI). The measurement properties of the ECLAM and SLEDAI over time were examined by comparing the ability of both measures to predict damage and oral steroid requirement.

Results

The ECLAM and SLEDAI are both responsive to change in disease activity irrespective of the statistic used. The ECLAM is more sensitive than the SLEDAI using the REI (all >1.9). Cumulative disease activity as measured by the SLEDAI or the ECLAM are important predictors of damage. There are no statistically important differences between the 2 measures with regard to their ability to predict steroid requirements.

Conclusions

The ECLAM has construct validity in cSLE and, like the SLEDAI, is highly sensitive to clinically important change in disease activity. The ECLAM may be more responsive. The quantitative properties of the 2 measures are very similar. The SLEDAI likely remains the preferable disease activity measure for cSLE given its overall measurement properties and ease of use.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

Disease activity and damage in systemic lupus erythematosus (SLE) cannot be measured directly by using any individual laboratory value or clinical sign. Thus, disease activity and damage measures have been developed for adult SLE (1–5). Because the clinical and laboratory features associated with childhood-onset SLE (cSLE; defined as SLE with disease onset prior to age 18 years) are different from adult SLE (6–14), all disease indices that have been developed and tested for adult SLE require reevaluation prior to use in children (14). The Systemic Lupus Erythematosus Disease Activity Index (SLEDAI) and the Systemic Lupus International Collaboration Clinics/American College of Rheumatology Damage Index (SDI) were developed to measure, respectively, disease activity and irreversible damage in adult SLE (5, 14–16). Both the SLEDAI and SDI have been validated for use in cSLE, and it has been shown that cumulative disease activity over time as measured by the SLEDAI is one of the single best predictors of disease damage in cSLE (14, 17).

The European Consensus Lupus Activity Measurement (ECLAM) is a disease activity scale developed for adult SLE (17–19) whose usefulness for cSLE has not yet been examined. It has been suggested that the ECLAM may be preferable to the SLEDAI for measuring disease activity because of its superior quantitative properties; the ECLAM has a higher correlation with changes in the physician global assessment of disease activity compared with the SLEDAI (18). If this was truly the case, then changes of ECLAM rather than SLEDAI scores could guide clinical decision making in children with cSLE because ECLAM scores would correlate better with expert assessment of changes in disease activity and, presumably, with decisions regarding patient therapies.

Our study objectives were 1) to assess the responsiveness of the ECLAM in cSLE when compared with the SLEDAI and 2) to test the hypothesis that the ECLAM has better quantitative properties compared with the SLEDAI by comparing the ability of the 2 measures to predict disease damage and patient medication requirements over time.

PATIENTS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

Instruments.

Systemic Lupus Erythematosus Disease Activity Index.

The SLEDAI measures “potentially reversible manifestations of the underlying inflammatory disease process” (5). The scale consists of 24 weighted attributes, grouped into 9 domains, called organ systems (weightings in brackets): central nervous system [8], vascular [8], renal [4], musculoskeletal [4], serosal [2], dermal [2], immunologic [2], constitutional [1], and hematologic [1]. If during a 10-day period prior to the assessment, a patient fulfills an attribute, then the corresponding weighted score will be given. The sum of all weighted attribute scores comprises the final SLEDAI score. Final SLEDAI scores range between 0 and 105, with 0 being no disease activity.

European Consensus Lupus Activity Measurement.

The ECLAM is a measure of disease activity for adult SLE that has been developed in a group process (19, 20). It consists of 34 items grouped into 12 domains. The mucocutaneous, renal, and hypocomplementemia domains have attributes for evolving symptoms, i.e., for items that newly occurred or worsened since the time of the last observation. The time frame during which evolving symptoms have to have (newly) appeared to be included in the ECLAM score is not firmly specified by the developer of the measure. The domains and attributes are as follows (maximum attribute/domain score in brackets): 1) generalized manifestations [0.5]; 2) articular manifestations [1]; 3a) active mucocutaneous manifestations [0.5]; 3b) evolving mucocutaneous manifestations [1]; 4) myositis [2]; 5) pericarditis [1]; 6) intestinal manifestations [2]; 7) pulmonary manifestations [1]; 8) evolving neuropsychiatric manifestations [2]; 9a) renal manifestations [0.5]; 9b) evolving renal manifestations [2]; 10) hematologic manifestations [1]; 11) erythrocyte sedimentation rate [1]; 12a) hypocomplementemia [1]; and 12b) evolving hypocomplementemia [1].

Calculation of the final ECLAM disease activity score.

The final ECLAM disease activity score is always an integer number between 0 and 10 and does not correspond to the sum of the domain scores. Several rules are to be considered when calculating the final ECLAM disease activity score. 1) If disease activity is present in only 1 of the items of the organ systems 1–10 (as numbered above), then 2 scores are added to the sum of the domain score to obtain the final ECLAM disease activity score.2) Sums of noninteger domain scores under 6 are rounded down to the next smaller integer number to obtain the final ECLAM disease activity score. 3) Sums of noninteger domain scores between 6 and 10 are rounded up to the next larger integer number to obtain the final ECLAM disease activity score. 4) All sums of domain scores larger than 10 are rounded down to 10 to obtain the final ECLAM disease activity score.

Special rules for scoring of ECLAM items applied in the current study.

Because not otherwise specified by the developer of the ECLAM, special assumptions were made for the purpose of the current study to allow for standardized scoring of the ECLAM. The item fatigue (ECLAM definition: subjective feeling of extraordinary tiredness) was scored only when fatigue interfered with school activities. Scores for evolving renal manifestations were given when renal manifestations newly emerged or, in the case of worsening, when the following occurred since the last clinic visit: the daily proteinuria increased by at least 500 mg, the creatinine clearance decreased by 50 ml/minute/1.73 m2, the serum creatinine increased by 30%, or the urinalysis showed increases of white blood cells or red blood cells by at least 10 cells/high-power field or any increase in the number of cellular casts. Scores for the item evolving hypocomplementemia were given for complement C3, C4, or CH50 levels that decrease by at least 30% compared with the last clinic visit.

Systemic Lupus International Collaboration Clinics/American College of Rheuma-tology Damage Index.

The SDI measures irreversible damage since the diagnosis of SLE. Irreversibility in the SDI is defined as the presence of any given SDI item for at least 6 months continuously (16). For all 41 items of the SDI, a clear definition is provided (21), a score of 0 is given to patients without irreversible damage, and the maximal possible SDI score is 47. Damage scored by the SDI is grouped into 12 different domains (maximal score per organ system in brackets): ocular [2], neuropsychiatric [6], renal [3], pulmonary [5], cardiovascular [6], peripheral vascular [4], gastrointestinal [7], musculoskeletal [7], and skin [3]. Damage scores are also given for premature gonadal failure [1], diabetes mellitus [1], and malignancy [2].

Patients.

The hospital charts of an inception cohort of all newly diagnosed cSLE patients (from January 1990 to June 1998) followed in the lupus clinic at The Hospital for Sick Children, Toronto (HSC) were reviewed. We have already reported on these patients (17). All patients fulfilled at least 4 of the American College of Rheumatology classification criteria for SLE (22) and were followed for at least 6 months. The disease activity of the patients was measured using the ECLAM and the SLEDAI for all clinic visits and admissions to the HSC. All data necessary to complete the SLEDAI or the ECLAM were extracted from standardized clinic forms used at the HSC and from laboratory databases, which stored the results of standard laboratory testing. Information regarding the use of steroids (oral and intravenous) and of immunosuppressive agents during the study period was obtained. The SDI was scored at the end of the study period.

Measurement of responsiveness to change.

To assess the ability of the ECLAM and the SLEDAI to detect clinically important improvement, disease activity at the time of diagnosis (D) was compared with that 6 months post diagnosis (D6). It was assumed that with therapy the disease activity of the subjects would, on average, significantly improve within 6 months. Similarly, disease activity scores of patients at the time of the first flare (F) should, on average, be higher than 6 months afterwards (F6). For the purpose of the analysis, flare of disease activity was defined as a change in disease activity requiring an increase in steroids (intravenous or oral) or the introduction of new immunosuppressive agents.

Responsiveness of the ECLAM and SLEDAI to clinically important worsening was measured by comparing the disease activity scores at D6 to the disease activity scores at the time of the first flare (D6/F). There is no commonly accepted best statistical approach to determine responsiveness of a disease index. Thus we used 4 different frequently used responsiveness statistics. They are effect size (ES) (23), effect size index (ESI) (24), standardized response mean (SRM) (25), and relative efficiency index (REI) (26). The ES, ESI, and SRM compare the responsiveness of the ECLAM and SLEDAI to an external standard (improvement, flare), whereas the REI assesses the relative responsiveness of the ECLAM and the SLEDAI by comparing the t distributions of the ECLAM scores to those of the SLEDAI. Thus, the F statistic can be used to compare the responsiveness of the 2 indices (as measured by the REI) for significant differences (25). The validity of the results from F statistics is highly influenced by the sample size: for the presented data the power is estimated to be at 80% for the comparison D/D6 (n = 66) and at almost 50% for the comparisons D6/F and F/F6 (n = 30).

Assessment of the quantitative properties of the ECLAM and the SLEDAI.

Two approaches were taken to examine the construct validity and to further assess the measurement properties (quantitative properties) of the ECLAM and the SLEDAI: We examined how well the cumulative disease activity over time can predict the observed damage of a patient; and we tested the ability of disease activity scores to predict the daily oral steroid dose prescribed to a patient at a given clinic visit. The oral steroid dose was considered to be an indicator of the physician's global judgement of patient disease activity.

Cumulative disease activity over time as predictor of disease damage.

The cumulative disease activity over time was determined by calculating the area under the curve of the serial measurements of disease activity using the trapezoidal rule. The algorithm to calculate the area under the curve of the ongoing disease activity is the following (27):

  • equation image

where n = number of assessments 1, 2, …m; DA = ECLAM score or SLEDAI score; AUC[DA(m)] = area under the curve of disease activity between diagnosis and time of assessment (m).

Linear modeling was used to describe the relationship between the cumulative disease activity over time and damage. Linear models were examined for fit and regression diagnostics were done to evaluate the statistical performance (28–30).

Relationship between ongoing disease activity and requirements of oral steroids.

We assumed that the weight-adjusted daily oral steroid dose a physician will prescribe to a patient at the time of a clinic visit will depend on the disease activity at the time of the clinic visit, the change in disease activity compared with the preceding clinic visits, and the time between the two clinic visits, i.e., the time necessary to mount this change in disease activity. Thus we assumed that the weight-adjusted daily steroid dose (outcome variable) can be predicted by the score of the disease activity measure at the time of the clinic visit (predictor 1), the change of the disease activity score since the last visit (predictor 2) when taking into account the time interval between the 2 visits (predictor 3). Possible differences among patients with regard to their response to medications were considered by introducing a patient-specific variable (predictor 4) in the regression models (i.e., a fixed model effect).

Repeated measure analysis considering variable time intervals between visits (unequally spaced data) was performed. Various statistical procedures are available in SAS 8.2 (SAS Institute, Cary, NC) to perform this type of repeated measure analysis. Briefly, these procedures differ in the statistical approach in that model predictions are either derived from least-squares estimations (PROC GLM) or from maximum likelihood estimations (PROC MIXED and PROC GENMOD). The output statistics provided to assess these statistical models for their quality (goodness of fit) and performance (here, how well the prescribed weight-adjusted steroid dose can be predicted) (30, 31) differ depending on whether least-squares or maximum likelihood estimations are used. Models that fit the model assumptions (thus, model results are trustworthy) have lower values of the Bayesian information criterion (BIC) and the Akaike's information criterion (AIC). The minimally statistically or clinically important differences of AIC or BIC values have not yet been established (30, 31). The scaled deviance has also been used to assess model fit (ideal value = 1).

Using least squares estimates, well performing models have high R2 values and statistically significant beta coefficients of the predictor variables. R2 values of models cannot be compared directly for statistical differences at the current time. However, models can be compared for significant differences in their performance by using the model –2 log-likelihood values and chi-square statistics (degrees of freedom [df] = 1).

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

Demographics and clinical features.

Sixty-six patients (56 girls) with various ethnic backgrounds were identified. Their mean age was 12.92 years (SD 2.65 years). All but 1 patient required oral steroids and 46 patients were treated with additional therapies: pulse treatments with methylprednisolone were given to 21 patients, and 43 patients received immunosuppressive agents (azathioprine, cyclophosphamide, methotrexate, and cyclosporine) during the followup period (mean 3.3; SD 2.2; range 0.5–9 years).

Damage.

Twenty-six of the 66 patients had no damage (SDI score = 0) and 40 patients developed damage as scored by the SDI. The mean SDI score was 1.76 (range 0–12, SD 2.64, median 2) (17). A combined total of 116 SDI scores was accumulated by the study group. Damage occurred primarily in the musculoskeletal, ocular, neuropsychiatric, and renal domains of the SDI. Musculoskeletal damage was the most common type of damage (34/116 scores; 29%) and 33 scores were given for ocular damage with small cataracts accounting for 29 of these 33 scores. Seven patients (11%) had neuropsychiatric damage, which accounted for 17% (20/116 scores) of the total damage. Renal damage occurred in 6 patients (11/116 scores; 9.5% of total damage).

Flares.

Only 30 of the 66 patients had a flare of disease activity during the study period. The mean time period between the time of diagnosis and the time of the first flare was 24 months (SD 10.7; median 24; range 6–47 months). At the time of diagnosis, 6 months after diagnosis, at the time of the first flare, and 6 months after flare (D, D6, F, and F6) the mean disease activity scores (SD) were 7.6 (2.4), 1.8 (1.5), 5.2 (1.8), and 2.1 (1.4) as measured by the ECLAM; 17 (10), 5 (4.3), 11 (5.1), and 5.1 (3.8) when using the SLEDAI. At these time points (D, D6, F, and F6) the mean weight-adjusted daily prednisone dosages of the patients were 1.5 mg/kg, 0.7 mg/kg, 1.1 mg/kg, and 0.6 mg/kg, respectively.

Responsiveness statistics.

Effect size.

The ES was calculated according to the formula shown in Table 1. For all 3 comparisons of changes in disease activity (D/D6; F/F6; D6/F), the SLEDAI and ECLAM were both very sensitive to change (an ES of > 0.8 indicates large responsiveness) (23) with the ECLAM being more responsive than the SLEDAI.

Table 1. Responsiveness to change in disease activity of ECLAM and SLEDAI*
Disease activity scaleChange in score mean (SD)/medianESESISRM (95% CI)
  • *

    Maximum total European Consensus Lupus Activity Measurement (ECLAM) score is 10. Maximum total Systemic Lupus Erythematosus Disease Activity Index (SLEDAI) score is 104. ES = effect size, determined as mean change/SD at baseline; ESI = effect size index, determined as mean change/SD at followup; SRM = standardized response mean, determined as mean change/SD of change; 95% CI = 95% confidence interval.

  • Comparison of disease activity at the time of diagnosis (D) with the disease activity 6 months after diagnosis (D6).

  • Comparison of disease activity at the time of the first flare (F) with the disease activity 6 months after first flare (F6).

  • §

    Comparison of disease activity at disease activity 6 months after diagnosis (D6) with the disease activity at the time of the first flare (F).

D/D6 (improvement)    
 ECLAM−5.9 (2.6)/−62.13.62.3 (+1.7, +2.9)
 SLEDAI−12.1 (8.8)/−100.92.91.4 (−0.7, +3.5)
F/F6 (improvement)    
 ECLAM−2.9 (2.0)/−2.81.62.31.4 (+0.9, +1.9)
 SLEDAI−5.7 (5.3)/−4.51.41.51.1 (−0.2, +2.4)
D6/F§ (worsening)    
 ECLAM3.1 (2.8)/3.3−1.9−1.7−1.1 (−1.8, −0.4)
 SLEDAI3.9 (6.9)/2−0.9−0.9−0.6 (−2.3, +1.7)
Effect size index.

Evaluation of the 2 disease activity measures using the ESI supports that both the ECLAM and the SLEDAI are very responsive to change (an ESI > 0.8 indicates large responsiveness) (32). Again, when using this statistic, the ECLAM performs better than the SLEDAI for all 3 comparisons (Table 1).

Standardized response mean.

When assessing the responsiveness of the 2 measures using the SRM, the ECLAM performs better than the SLEDAI and has smaller 95% confidence intervals around the SRM (an SRM > 0.8 indicates large responsiveness) (32). Both indices are moderate to highly responsive to change and the difference in responsiveness between the 2 measures was statistically not significant (as indicated by the overlapping 95% confidence intervals; Table 1).

Relative efficiency index.

The ECLAM appeared to be more sensitive to change than the SLEDAI, as supported by REI > 1 for the 3 comparisons (Table 2). This difference in responsiveness was statistically significant for all 3 comparisons made.

Table 2. Relative efficiency index to compare the responsiveness of the ECLAM with the SLEDAI*
REIREIF statistic
  • *

    Relative efficiency index (REI) determined as (t from paired t-test of ECLAM score)2/(t from paired t-test of SLEDAI score)2; values >1 signify that the ECLAM is more response than the SLEDAI. ECLAM = European Consensus Lupus Activity Measurement; SLEDAI = Systemic Lupus Erythematosus Disease Activity Index; df = degrees of freedom.

  • Comparison of disease activity at the time of diagnosis (D) with the disease activity 6 months after diagnosis (D6).

  • Comparison of disease activity at the time of the first flare (F) with the disease activity 6 months after first flare (F6).

  • §

    Comparison of disease activity at disease activity 6 months after diagnosis with the disease activity at the time of the first flare.

D/D62.7F = 3.36 (64 df); α < 0.0001
F/F61.9F = 2.73 (28 df); α < 0.005
D6/F§1.9F = 2.14 (28 df); α < 0.003

Comparison of other measurement properties of the ECLAM and the SLEDAI.

Cumulative disease activity over time (estimated by the area under the curve of ongoing disease activity) as measured by the ECLAM was a good predictor of damage (R2 = 24%, slope = 47.2, SD of slope 11.0, intercept = 0.221; P of the slope < 0.0001, P of intercept = not significant [NS]). Cumulative disease activity as measured by the SLEDAI was somewhat better in predicting patient damage (R2 = 30%, slope = 22.3, SD = 4.2, intercept = 0.125; P of the slope < 0.0001, P of intercept = NS). When repeating the same analysis using maximum likelihood estimations (instead of least square estimations), the cumulative disease activity as measured by the ECLAM and that measured by the SLEDAI not significantly different from each other in predicting damage (comparison of –2 log likelihood values of the normalized ECLAM and SLEDAI scores: χ2 = 2.88, 1 df, P = NS).

Change of ECLAM and SLEDAI scores relative to oral steroid requirement of the patients.

To further assess the measurement characteristics of the ECLAM and SLEDAI over time repeated measure analysis was done.

Models.

As described above, statistical modeling was performed to predict the weight-adjusted daily oral steroid dose at a given clinic visit by using the 4 predictors: 1) absolute disease activity scores (ECLAM or SLEDAI) at the time of a clinic visit; 2) change of disease activity score (ECLAM or SLEDAI) in response to therapies since the preceding visit; 3) time between the two visits; and 4) patient-specific variable.

Irrespective of the type of repeated measure analysis (PROC GLM, PROC GENMOD, PROC MIXED) and for both the ECLAM and the SLEDAI, the 4 chosen predictors were all highly statistically significant in estimating the weight-adjusted daily steroid dose (P < 0.0001) of a patient at a given clinic visit. This confirmed the initial model assumptions (see above) and also supports that individual patients differ significantly in their response to oral steroids.

Least squares estimations (PROC GLM).

The SLEDAI is similar to the ECLAM in predicting the weight-adjusted daily steroid dose prescribed by the physician at a given clinic visit when least squares estimations were used (ECLAM: R2 = 48.1%, intercept = 0.65, P of the intercept < 0.0001; SLEDAI: R2 = 50.2%; intercept = 0.41, P of the intercept < 0.0001). There were no important differences in model fit. The intercept of both models is statistically significant, suggesting that there are possibly other important factors that influence treatment decisions that are not considered in the model. However, the absolute size of both intercepts is very small and therefore unlikely important in this context.

Maximum likelihood estimation (PROC GENMOD, PROC MIXED).

Compared with the SLEDAI, the ECLAM was significantly better in predicting the weight-adjusted daily steroid dose (χ2 > 6, 1 df, P < 0.014). However, models using the SLEDAI-derived predictors had a better fit (i.e., were in better conformance with the basic model assumptions) than those using ECLAM-derived predictors (AICSLEDAI = 375 < AICECLAM = 467; BICSLEDAI = 380 < BICECLAM = 471; scaled devianceECLAM and scaled devianceSLEDAI both 1.003). Taken together, there appears to be no important difference between the ECLAM and the SLEDAI in predicting the weight-adjusted daily steroid dose of a patient at a given clinic visit.

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

The ECLAM is a useful measure of disease activity for children with lupus. Compared with the SLEDAI, the ECLAM appears marginally more sensitive to changes in disease activity over time. However, the SLEDAI and the ECLAM are very similar in predicting disease damage and the need for steroids of patients with cSLE.

There is no universally accepted best measure of disease activity for adults or children with lupus (14). Such a measure would be desirable to make studies that include measurements of disease activity better comparable to each other. We recently validated the BILAG (British Isles Lupus Assessment Group Index) (2), the SLAM (Systemic Lupus Activity Measure) (3) and the SLEDAI for use in cSLE (14). The present study supports that the ECLAM can also be used for measuring disease activity in children with cSLE. As previously suggested for adults (33), the ECLAM appeared to be slightly more sensitive to changes in disease activity than the SLEDAI when used in cSLE. However, the responsiveness of the ECLAM compared with the SLEDAI was only significantly better when using the REI, but not when using the SRM (overlapping confidence intervals). Thus the superiority of the ECLAM over the SLEDAI for cSLE cannot be concluded on the basis of the results of this study. Another approach to evaluate whether the ECLAM may be preferable to the SLEDAI for use in cSLE is to assess whether there are significant differences between the 2 measures when predicting a clinically relevant surrogate of disease activity, such as the steroid dose chosen by the treating physicians. The ECLAM scores, however, were no different from those of the SLEDAI in estimating the daily weight-corrected dose of oral steroids of a patient. We further examined whether the ECLAM and the SLEDAI differed in the ability to predict damage (17). Again, the ECLAM and the SLEDAI were similarly suitable to predict disease damage in children with cSLE. Thus, the current data do not support that the ECLAM has better quantitative properties compared with the SLEDAI, as was previously suggested (18).

Taken together, the statistical evidence that the ECLAM is superior to the SLEDAI is weak and inconsistent. We previously suggested (14) that the SLEDAI may be the preferable disease activity measure in cSLE. Using statistical criteria, the results of this study do not support that the ECLAM is clearly superior to the SLEDAI for use in cSLE.

Five methodologic criteria are required for a successful disease activity measure: the measure has to have credibility (face validity), accuracy, and sensitivity to change; it should make biological sense; and it must be simple for daily use (34). The results of this and previous studies (14, 17) support that both the ECLAM and the SLEDAI are suitable for measuring disease activity in cSLE. However, the face validity of the ECLAM may be inferior to that of the SLEDAI. This is because, unlike for the SLEDAI, the items of the ECLAM are not exactly defined. For example, there is not an exact definition for the severity of fatigue that should be scored in the ECLAM. Also, the time frame during which symptoms are regarded as “evolving” is unclear and the decrease of complement levels necessary to be scored as “significantly reduced compared to the last observation” (attribute 12b) is not well defined. As opposed to some other lupus disease activity measures (2, 3, 5), the ECLAM lacks a clearly defined time frame during which symptoms have to occur to be included in a certain measurement of disease activity. Although previous studies suggest that both the ECLAM and the SLEDAI are highly reliable when used in prospective and retrospective studies (2, 34–36), the lack in exact definitions may lead to decreased inter- and intrarater reliability of the ECLAM. However, if exact definitions for all items of the ECLAM were generated, then the credibility of the ECLAM could likely be enhanced.

The acceptance of a disease measure in research and clinical practice is also influenced by its ease of use. Because the ECLAM has 34 items but the SLEDAI only has 24 items, completion of the ECLAM requires more time. Opposed to the SLEDAI disease activity score, the final disease activity score of the ECLAM is not equal to the simple sum of the domain scores. Instead, a more complicated scoring scheme is suggested for the calculation of the final ECLAM disease activity score, which is not consistently applied in the literature (37, 38). Thus the scoring scheme for the final ECLAM disease activity score not only increases the time requirements to measure disease activity, but it is also a potential source of error and may not even be necessary. In an exploratory analysis, we evaluated the measurement properties of the ECLAM in our cohort without using the suggested scoring scheme to calculate the final ECLAM disease activity score. There were no important differences in the responsiveness of the ECLAM or its ability to predict the weight-adjusted oral steroid dose or disease damage, irrespective of whether the simple sum of the domain scores or the final ECLAM disease activity score (using the differential scoring scheme) was used (data not shown).

Our study supports the idea that the ECLAM is a useful and responsive measure of disease activity in cSLE. However, the ECLAM does not offer a clear and consistent advantage over the SLEDAI in measuring disease activity in cSLE. We feel that given its ease of use, its wide acceptance in adult SLE, and its excellent measurement properties, the SLEDAI remains the preferred measure of disease activity in cSLE.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES

We would like to thank Bin Huang, PhD and Edward Giannini, MSc, DrPH for reviewing the statistical methods presented in this article.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. Acknowledgements
  8. REFERENCES
  • 1
    Hay E, Gordon C, Emery P. Assessment of lupus: where are we now? Ann Rheum Dis 1993; 52: 16972.
  • 2
    Hay E, Bacon P, Gordon C, Isenberg DA, Maddison P, Snaith ML, et al. The BILAG index: a reliable and valid instrument for measuring clinical disease activity in systemic lupus erythematosus. Q J Med 1993; 86: 44758.
  • 3
    Liang MH, Socker S, Larson MG, Schur PH. Reliability and validity of six systems for the clinical assessment of disease activity in systemic lupus erythematosus. Arthritis Rheum 1989; 32: 110718.
  • 4
    Bencivelli W, Vitali C, Isenberg DA, Smolen JS, Snaith ML, Sciuto M, et al.: Disease activity in systemic lupus erythematosus: report of the consensus study group of the European workshop for rheumatology research. III. Development of a computerized clinical chart and its application to the comparison of different indices of disease activity. Clin Exp Rheumatol 1992; 10: 54954.
  • 5
    Bombardier C, Gladman D, Urowitz M, Committee on Prognosis Studies in Systemic Lupus Erythematosus. The derivation of the SLEDAI: a disease activity index for lupus patients. Arthritis Rheum 1992; 35: 63040.
  • 6
    Press J, Palayew K, Laxer RM, Elkon K, Eddy A, Rakoff D, et al: Antiribosomal P antibodies in pediatric patients with systemic lupus erythematosus and psychosis. Arthritis Rheum 1996; 39: 6716.
  • 7
    Font J, Cervera R, Espinosa G, Parrarés L, Ramos-Casals M, Jiménez S, et al.: Systemic lupus erythematosus (SLE) in childhood: analysis of clinical and immunological findings in 34 patients and comparison with SLE in adults. Ann Rheum Dis 1998; 57: 4569.
  • 8
    Barron KS, Silverman ED, Gonzales J, Reveille JD: Clinical, serologic, and immunogenetic studies in childhood-onset systemic lupus erythematosus. Arthritis Rheum 1993; 36: 34854.
  • 9
    Janiwityanujit S, Totemchokchyakarn K, Versasertniyom O, Vanichapunto M, Vatanasuk M. Age-related differences of clinical and immunological manifestations of SLE. Asian Pac J Allergy Immunol 1995; 13: 1459.
  • 10
    Costallat LTL, Coimbra AMV. Systemic lupus erythematosus: clinical and laboratory aspects related to age at disease onset. Clin Exp Rheumatol 1994; 12: 6037.
  • 11
    Tucker LB, Menon S, Schaller JG, Isenberg DA. Adult- and childhood-onset systemic lupus erythematosus: a comparison of onset, clinical features, serology, and outcome. Br J Rheumatol 1995; 35: 86672.
  • 12
    Hashimoto H, Tsuda H, Hirano T, Takasaki Y, Matsumoto T, Hirose S. Differences in clinical and immunological findings of systemic lupus erythematosus related to age. J Rheumatol 1987; 14: 497501.
  • 13
    Lehman TJ, McCurdy DK, Bernstein BH, King KK, Hanson V. Systemic lupus erythematosus in the first decade of life. Pediatrics 1989; 83: 23541.
  • 14
    Brunner HI, Feldman BM, Bombardier C, Silverman ED. The SLEDAI, SLAM and BILAG are sensitive to clinical change in disease activity in childhood-onset systemic lupus erythematosus. Arthritis Rheum 1999; 42: 135460.
  • 15
    Strand V, Gladman D, Isenberg D, Petri M, Smolen J, Tugwell P. Outcome measures to be used in clinical trials in systemic lupus erythematosus. J Rheumatol 1999; 26: 4907.
  • 16
    Gladman D, Ginzler E, Goldsmith C, Fortin P, Liang M, Urowitz M, et al. The development of the Systemic Lupus International Collaborating Clinics/American College of Rheumatology damage index in patients with systemic lupus erythematosus. Arthritis Rheum 1996; 39: 3639.
  • 17
    Brunner HI, Silverman ED, To T, Bombardier C, Feldman BM. Risk factors for damage in childhood-onset systemic lupus erythematosus (cSLE): cumulative disease activity over time and medication use predict disease damage. Arthritis Rheum 2002; 46: 43644.
  • 18
    Bencivelli W, Vitali C, Isenberg DA, Smolen JS, Snaith ML, Sciuto M, et al. Disease activity in systemic lupus erythematosus: report of the consensus study group of the European workshop for rheumatology research. III. Development of a computerised clinical chart and its application to the comparison clinical chart and its application to the comparison of different indices of disease activity. Clin Exp Rheumatol 1992; 10: 54954.
  • 19
    Bombardieri S, Vitali C, Caponi L, Manca L, Bencivelli W. Activity criteria in systemic lupus erythematosus. Clin Exp Rheumatol 1994; 12: S458.
  • 20
    Vitali C, Bencivelli W, Isenberg DA, Smolen JS, Snaith ML, Sciuto M, et al. Disease activity in systemic lupus erythematosus: report of the consensus study group of the European workshop for rheumatology research. II. Identification of the variables indicative of disease activity and their use in the development of an activity score. Clin Exp Rheumatol 1992; 10: 5417.
  • 21
    Gladman DD, Urowitz MB, Goldsmith CH, Fortin P, Ginzler E, Gordon C, et al. The reliability of the Systemic Lupus Erythematosus International Collaborating Clinics/American College of Rheumatology damage index in patients with systemic lupus erythematosus. Arthritis Rheum 1997; 40: 80913.
  • 22
    Tan EM, Cohen AS, Fries JF, Masi AT, McShane DJ, Rothfield NF, et al. The 1982 revised criteria for the classification of SLE. Arthritis Rheum 1982; 25: 12717.
  • 23
    Kazis L, Anderson J, Meenan R. Effect sizes for interpreting changes in health status. Med Care 1989; 27: S17889.
  • 24
    Liang MH. Evaluating measurement responsiveness. J Rheumatol 1995; 22: 11912.
  • 25
    Cohen J. Statistical power analysis for the behavioral sciences. New York: Academic Press; 1977.
  • 26
    Liang MH, Larson MG, Cullen KE, Schwartz JA. Comparative measurement efficiency and sensitivity of five health status instruments for arthritis research. Arthritis Rheum 1985; 28: 5427.
  • 27
    Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986; 1: 30710.
  • 28
    Hook EB, Regal RR. Validity of methods to model selection weighting for model uncertainty, and small sample size. Am J Epidemiol 1997; 145: 113844.
  • 29
    Fox J. Regression diagnostics. Thousand Oaks (CA), Sage Publications; 1997.
  • 30
    SAS/STAT software: changes and enhancements through release 6.12. Cary (NC): SAS Institute; 1997.
  • 31
    Littell RC, Milliken GA, Stroup WW, Wolfinger RD. SAS systems for mixed models. Cary (NC): SAS Institute; 1996.
  • 32
    Beaton DE, Hogg-Jonson S, Bombardier C. Evaluating changes in health status: reliability and responsiveness of five generic health status measures in workers with musculoskeletal diseases. J Clin Epidemiol 1997; 50: 7993.
  • 33
    Ward MM, Marx AS, Barry NN. Comparison of the validity and sensitivity to change of 5 activity indices in systemic lupus erythematosus. J Rheumatol 2000; 27: 66470.
  • 34
    Hawker G, Gabriel S, Bombardier C. A reliability study of the SLEDAI: a disease activity index for systemic lupus erythematosus. J Rheumatol 1993; 20: 65760.
  • 35
    FitzGerald JD, Grossman JM. Validity and reliability of retrospective assessment of disease activity and flare in observational cohorts of lupus patients. Lupus 1999; 8: 63844.
  • 36
    Mosca M, Bencivelli W, Vitali C, Carrai P, Neri R, Bombardieri S. The validity of the ECLAM index for the retrospective evaluation of disease activity in systemic lupus erythematosus. Lupus 2000; 9: 44550.
  • 37
    Samsonov MY, Tilz GP, Egorova O, Reibnegger G, Balabanova RM, Nassonov EL, et al. Serum soluble markers of immune activation and disease activity in systemic lupus erythematosus. Lupus 1995; 4: 2935.
  • 38
    Caligaris-Cappio F, Bertero MT, Converso M, Stacchini A, Romagnani S, Pizzolo G. Circulating levels of soluble CD30, a marker of cell producing Th2-type cytokines, are increased in patients with systemic lupus erythematosus and correlate with disease activity. Clin Exp Rheumatol 1995; 13: 33943.