Improving the Assessment of Systemic Lupus Erythematosus Disease Activity

OBJECTIVE I - INITIAL DEVELOPMENT AND VALIDATION OF SSIStudy Population and Sampling Methods: The eligible study population will be compromised of patients seen in the Toronto Lupus Clinic between January 1st, 2010 and December 31, 2012. Patients with active disease, SLEDAI-2K >=6, in whom prednisone was increased to >=15mg/day or initiated at >=15mg/day will be identified in the TLC database. All patients must have follow up visits at 3 and 6 months. N=~185Eligibility CriteriaInclusion Criteria• >=4 American College of Rheumatology (ACR) criteria or 3 ACR criteria plus a typical histological lesion of SLE on renal or skin biopsy• SLEDAI-2K >=6• Clinician's diagnosis based on his/her assessment• Patients from the Toronto Lupus Clinic with regular follow-up, defined as having follow up visits at 3 and 6 months from the baseline visit (1st study visit).Exclusion Criteria• Patients with missing follow up visits at 3 and 6 months from the baseline visit (1st study visit).• Patients with missing data in the charts for all visits.Sampling: In 2012, 65 patients met these criteria and to increase our sample size to 185 we will analyze the data collected since 2010. Based on the COnsensus-based Measurement INstruments (COSMIN) recommendations, a sample size >=100 is recommended in this step(1).Data for patients who meet the inclusion criteria will be extracted from the TLC database. The study start (1st study visit) is defined as the 1st occurrence of SLEDAI-2K>=6 in spite of standard of care where the inclusion and exclusion criteria are met. Follow-up visits at 3 and 6 months will be used in the analysis of the data for this study.Primary Endpoints•   Improved patients (responders) will be identified based on SLEDAI-2K definition of improvement (decrease in the total score by >=4).•   Clinician scoring of disease activity-Improved (external construct) based on predefined definitions.Sample Size/Power Calculations: N =~ 185. No power calculation was performed but the sample size was determined based on the COSMIN recommendations for studied on measurement properties in this step(1). In 2012, 65 patients met the same inclusion criteria used in this study and to increase our sample size to 185 we will analyze the data collected since 2010. Based on the COSMIN recommendations, a sample size >=100 is recommended in this step(1).Hypothesis: Incorporating a new concept, “lupus status overall”, which accounts for disease activity and corticosteroids dose in the assessment of disease activity will enhance the ability to understand the nature of improvement in responders. Data Analysis Considerations - Development of SLEDAI-2K Steroids Index (SSI)Measuring Lupus Status Overall: The corticosteroids dose must be considered when trying to understand the nature of improvement in responders. We hypothesize that incorporating corticosteroid dose in the assessment of disease activity will overcome this issue. We are proposing a new concept “lupus status overall” which accounts for disease activity and corticosteroids dose. A new “organ system” named “corticosteroid” which contains only one descriptor also named “corticosteroid” will be added to SLEDAI's organ systems and descriptors. This new index reflects “lupus status overall” and named SLEDAI-2K Steroid Index (SSI).Scoring of SSI: In the original derivation of SLEDAI, multiple regression models were used to determine the relative importance of the proposed 34 descriptors in the physician global rating of disease activity. As a result, 24 of 34 descriptors were retained and grouped in 9 organ systems with different weighted scores: nervous system (7 descriptors each scored 8), vascular (1 descriptor scored 8), renal (4 descriptors each scored 4), musculoskeletal (2 descriptors each scored 4), serosal (2 descriptors each scored 2). Dermal (3 descriptors each scored 2), immunologic (2 descriptors each scored 2), constitutional (1 descriptor scored 1) and hematologic (2 descriptors each scored 1). The total possible weighted score of SLEDAI ranges between 0-105 with the highest weighted score, 8, being assigned for the nervous system and vasculitis(2). In the new index, the total number of organ systems will be 10 with 25 descriptors. At least 4 models (gross, intermediate, fine and ultra-fine) which assign different scores for the corticosteroid descriptor of SSI based on the dose will be studied to determine the appropriate score for the corticosteroid descriptor. Since the dose of corticosteroids is very important in the assessment of lupus status overall, it was assigned a maximum weighted score of 8 which is equal to the highest weighted organ system in SLEDAI. In the case of absence of use of corticosteroids the weighted score will be 0, and the weighted score for corticosteroids increases with higher doses (table 1).Table 1. SLEDAI-2K Steroid Index (SSI) and the 4 models of scoringGross    Intermediate    Fine    Ultra-fineDose(mg) Score    Dose (mg) Score   Dose (mg) Score   Dose (mg) Score0 0    0 0    0 0    0 0=7.5 2    =5 2    =5 2    =5 28-20 4    5.1-15 4    5.1-10 3    5.1-10 4>21 8    16-30 6    11-15 4    11-15 6    >30 8    16-30 6    16-30 8       >30 8    >30 8Retrospective validation of SSI: Data will be extracted from the TLC database. First, improved patients (responders), defined as a decrease by >=4 in the total score of SLEDAI-2K, will be further studied. In this group of patients, the mean change in SLEDAI-2K will be determined in each patient and at the same time we will be calculating the SSI using the described models in table 1. Descriptive analysis will be used to study the mean change of SLEDAI-2K scores and the mean change of SSI scores in the responders. Second, we will determine if SSI responders are true responders and not false responders. Using the SLEDAI-2K responders as “Gold Standard”, the results will be analyzed by constructing a 2x2 table in which will included: true positives (SLEDAI-2K responders and SSI responders), false positives (SLEDAI-2K non-responders and SSI responders), true negatives (SLEDAI-2K non-responders and SSI non-responders) and false negative (SLEDAI-2K responders and SSI non responders). Third, (concurrent construct validity): Clinician scoring of disease activity: A clinician (external construct) who does not know the patients and who is not aware of the SLEDAI-2K and SSI scores will evaluate each patient's record and assign a clinical activity score for each assessment according to the following scale: improved, same, and worse, using standardized predefined definitions. “Improved,” defined as one of the following: (1) stopping treatment in the presence of improvement of an already active system or in response to complete remission of an active system; (2) a decrease in medication dosage for the above reason; (3) indication of improvement in SLE disease activity in the physician's notes; (4) use of the term improvement in the physician's notes. “Worse,” defined as one of the following: (1) introduction of new treatment in the presence of worsening of an already active system, or in response to activation of a new system; (2) increase in medication dosage for the above reasons; (3) indication of concern regarding SLE disease activity in the physician's notes — arrangement for an earlier appointment/investigation to assess SLE disease activity; (4) the use of the term flare/worsening in the physician's notes; (5) new diagnosis of SLE (new presentation, not just the accumulation of American College of Rheumatology criteria)(3). “Same”, defined as no change in disease activity in patients who did not qualify for the definitions of improved or worse(4,5). The correlation between the external construct and the change in SLEDAI-2K and SSI scores will be studied. Using the clinician scoring of disease activity-Improved as “Gold Standard”, the results of SLEDAI-2K and SSI responders will be analyzed by constructing two 2x2 tables. Forth, using the clinician scoring of disease activity-Improved as “Gold Standard”, the results of SLEDAI-2K and SSI responders will be analyzed by constructing two 2x2 tables as follows:Table 1: True positives (Clinician scoring-Improved and SSI responders), false positives (Clinician scoring-Not Improved and SSI responders), true negatives (Clinician-scoring Not Improved and SSI non-responders) and false negative (Clinician scoring-Improved and SSI non responders). Table 2: True positives (Clinician scoring-Improved and SLEDAI-2K responders), false positives (Clinician scoring-Not Improved and SLEDAI-2K responders), true negatives (Clinician scoring-Not Improved and SLEDAI-2K non-responders) and false negative (Clinician scoring-Improved and SLEDAI-2K non responders). The results of the third and fourth phases will be further studied and compared.Objective I Limitations: This initial validation of SSI lack external validity especially due to the use of a single cohort, TLC.OBJECTIVE II - FURTHER VALIDATION OF SSI USING BLISS TRIAL DATA Study Population and Sampling Methods: Patients in GSK BLISS-52 (N=865) and BLISS-76 (N=819) were assessed at regular intervals and data, including laboratory and clinical parameters important to lupus, medications and in particular corticosteroids was collected and it will be used in this study on validation of SSI. Eligibility CriteriaInclusion/Exclusion Criteria: Not applicable. We will include all patients from BLISS-52 and BLISS-76.Sampling: We will use the same sample size of BLISS-52 and BLISS-76. Patients in GSK BLISS-52 (N=865) and BLISS-76 (N=819). In these trials 1684 adult seropositive patients with SLE who scored >=6 on SLEDAI were enrolled.Primary Endpoints: SRI responders and SRI-modified responders.Sample Size/Power Calculations: Not applicableHypothesis: A more precise outcome measures for SLE, SSI, will enhance the ability to understand the nature of improvement in responders in BLISS studies. Objective II Data Analysis Considerations: For this objective, data from GSK BLISS-52 (N=865; ClinicalTrials.gov NCT00424476) and BLISS-76 (N=819; ClinicalTrials.gov NCT 00410384) will be provided. In these trials 1684 adult seropositive patients with SLE who scored >=6 on SLEDAI were enrolled. Patients were from Europe and Central/North America. Patients were randomized in 1:1:1 ratio to receive 1mg/kg belimumab or 10mg/kg belimumab or placebo intravenously on day 0, 14, 28 and then every 28 days for 72 weeks. The primary efficacy endpoint in both trials was SLE Responder Index (SRI). The SRI incorporates the Safety of Estrogens in Lupus Erythematosus-National Assessment-SLEDAI (SELENA-SLEDAI), British Isles Lupus Assessment Group (BILAG), and Physician Global Assessment (PGA)(9). The SRI utilizes: i) the SELENA-SLEDAI score to determine global improvement defined as decrease in total SELENA-SLEDAI score by >= 4. The BILAG domain scores were used to ensure that no significant worsening in heretofore unaffected organ systems had occurred. The PGA ensured that improvement in disease activity was not achieved at the expense of the patient's overall condition. The SRI was adopted as the primary outcome measure in phase III trials on belimumab (BLISS-52 and BLISS-76)(6,7,8,9). Data analysisOutcomes•   SRI with the 3 components listed above: SLEDAI-2K, BILAG and PGA•   SRI-modified: The SRI-modified will include the same 2nd and 3rd components but with modification of the first component where SELENA-SLEDAI will be replaced by SSI. Endpoints•   First, the percentage of responders in each group, 1mg/kg of belimumab, 10 mg/kg of belimumab and placebo will be analyzed for SRI-modified and compared to the original SRI. The percentage of responders for all 4 SSI models will be compared to the percentage of SRI responders as in table 1. This analysis will allow us to gain insight into which model of SSI is best able to discern differences between treatment groups; placebo and belimumab. •   Second, we will determine if SRI-modified responders (patients with improvement) are true responders and not false responders. Using the SRI as “Gold Standard”, the results will be analyzed by constructing a 2x2 table in which will included: true positives (SRI responders and SRI-modified responders), false positives (SRI non-responders and SRI-modified responders), true negatives (SRI non-responders and SRI-modified non-responders) and false negative (SRI responders and SRI-modified non responders). Further analysis of the data will be conducted to understand the false negatives and false positives responders.Objective II Limitations: The analysis in this study is conducted on an existing prospectively collected data from BLISS. It is very important to evaluate the validity of SSI on a prospectively collected data.OBJECTIVE III ASSESSMENT OF CONCURRENT CONSTRUCT VALIDITY OF SSI PROSPECTIVELY IN THE UNIVERSITY OF TORONTO LUPUS CLINIC:Study Population and Sampling Methods: Patients' selection: All patients visiting the Toronto Lupus Clinic from December 2015 to May 2017. We aim to identify patients with active disease with a flare (increase in SLEDAI-2K by at least 4) that requires an increase in the dose of prednisone to >=15 mg/day or initiation of prednisone at >=15 mg/day. Eligibility CriteriaInclusion Criteria:• =4 American College of Rheumatology (ACR) criteria or 3 ACR criteria plus a typical histological lesion of SLE on renal or skin biopsy• Increase in SLEDAI-2K >=4• Clinician's diagnosis based on his/her assessmentExclusion Criteria:• Patients with missing follow up visits at 3 and 6 months from the baseline visit (1st study visit).• Patients with missing data in the charts for all visits.Sampling: SSI includes the same 24 descriptors of SLEDAI-2K(4,10). At the Toronto Lupus clinic 885 patients will contribute at least to 1770, 2655, or 3540 patients-visit if evaluated 2, 3 or 4 times/year. In our previous validation of SLEDAI-2K Responder Index-50 (S2K RI-50) a total of 62 patients with improvement were analyzed and the correlation with the external construct (7-point Likert scale) was 0.48 (p<0.001)(10). In the validation of SSI, we aim to achieve similar or higher correlation with the external construct. In the lupus clinic, approximately 42 patients are seen weekly and 168 patients monthly. The rate of flares reported in the Toronto Lupus Clinic of the patients in their follow-up clinic visit is 35%(3). Thus, of the 168 patients seen per month it is expected that 58 will have a flare/worsening. Among the patients who flared in 2012, 65 patients required prednisone >=15 mg/day. Our target sample size is 100 patients and to achieve this number at least 18 months will be required and subsequently a 6 months period for follow-up for each patients. Based on the COnsensus-based Measurement INstruments (COSMIN) recommendations, a sample size >=100 is recommended in this step (5).Primary Endpoints• Improved patients (responders) will be identified based on SLEDAI-2K definition of improvement (decrease in the total score by >=4).• SSI improved patients (responders) will be identified (decrease in the total score by >=4).• Physician Global Assessment (PGA).Sample Size/Power Calculations: N=~ 100. No power calculation was performed but the sample size was determined based on the Consensus-based Measurement INstruments (COSMIN) recommendations for studied on measurement properties in this step (5). Our target sample size is 100 patients and to achieve this number at least 18 months will be required and subsequently a 6 months period for follow-up for each patients. Based on the COSMIN recommendations, a sample size =100 is recommended in this step (5).Hypothesis: In the patients with improvement, the mean change SSI will correlate to a greater extent with PGA compared to change SLEDAI-2K. Objective III Data Analysis Considerations:Patient assessment: Patients will be assessed initially (at an anchor visit) and then followed prospectively and assessed every 3 months or sooner if medically required. The analysis will be conducted at 3 and 6 months from the anchor visit. SLEDAI-2K at time 0 (T0: anchor visit) will be determined on the baseline visit. SLEDAI-2K T1 (follow-up visit) and SSI scores will be determined on a follow-up visits at 3 and 6 months. The physician global assessment (PGA) will be completed for all visits on a visual analog scale (VAS) line of 100 mm, with anchors of 0 “no disease activity” and 10 for “very active disease.” Method and analysis: Patients with improvement, defined as a decrease by >= 4 in the total score of SLEDAI-2K at follow-up visits, will be identified. The mean change of SSI scores at follow up visits (change SSI = change SLEDAI-2K + weighted score of corticosteroids dose). The correlation of the mean change SSI will be correlated to the mean change PGA and compared to the correlation of the mean change SLEDAI-2K and mean change PGA. In studies on validation of new tools it is recommended to demonstrate a high correlation with the external construct (in our case it is the PGA), >=0.70, as long as the external construct is the “gold standard” accepted in the field. In the absence of “gold standard” a correlation =<0.7 is justified as long as specific hypotheses are formulated a priori and at least 75% of the results are in agreement with these hypotheses(3). Thus, we hypothesize that in the patients with improvement, the mean change SSI will correlate to a greater extent with PGA compared to change SLEDAI-2K.Objective III Limitations: A single centre study.REFERENCES1. Mokkink LB, Terwee CB, Patrick DL, Alonso J, Stratford PW, Knol DL, et al. The COSMIN checklist for assessing the methodological quality of studies on measurement properties of health status measurement instruments: an international Delphi study. Quality of Life Research. 2010;19(4):539-49.2. Bombardier C, Gladman DD, Urowitz MB, Caron D, Chang CH. Derivation of the SLEDAI. A disease activity index for lupus patients. The Committee on Prognosis Studies in SLE. Arthritis Rheum. 1992;35(6):630-40.3. Tan EM, Cohen AS, Fries JF, Masi AT, McShane DJ, Rothfield NF, et al. The 1982 revised criteria for the classification of systemic lupus erythematosus. Arthritis Rheum. 1982;25(11):1271-7.4. Gladman DD, Urowitz MB, Kagal A, Hallett D. Accurately describing changes in disease activity in Systemic Lupus Erythematosus. J Rheumatol. 2000;27(2):377-9. 5. Touma Z, Gladman DD, Ibanez D, Urowitz MB. Development and initial validation of the systemic lupus erythematosus disease activity index 2000 responder index 50. J Rheumatol. 2011;38(2):275-84.6. Furie R, Petri M, Zamani O, Cervera R, Wallace DJ, Tegzova D, et al. A phase III, randomized, placebo-controlled study of belimumab, a monoclonal antibody that inhibits B lymphocyte stimulator, in patients with systemic lupus erythematosus. Arthritis Rheum. 2011;63(12):3918-30.7. Navarra SV, Guzman RM, Gallacher AE, Hall S, Levy RA, Jimenez RE, et al. Efficacy and safety of belimumab in patients with active systemic lupus erythematosus: a randomised, placebo-controlled, phase 3 trial. Lancet. 2011;377(9767):721-31.8. Smolen JS, Strand V, Cardiel M, Edworthy S, Furst D, Gladman D, et al. Randomized clinical trials and longitudinal observational studies in systemic lupus erythematosus: consensus on a preliminary core set of outcome domains. J Rheumatol. 1999;26(2):504-7.9. Wang J, Hu W, Xie H, Zhang H, Chen H, Zeng C, et al. Induction therapies for class IV lupus nephritis with non-inflammatory necrotizing vasculopathy: mycophenolate mofetil or intravenous cyclophosphamide. Lupus. 2007;16(9):707-12. 10. Petri M, Kim MY, Kalunian KC, Grossman J, Hahn BH, Sammaritano LR, et al. Combined oral contraceptives in women with systemic lupus erythematosus. The New England journal of medicine. 2005;353(24):2550-8.