P155 Identifying a population of patients for intensive first-line therapy in SLE: a clinical and biomarker model to predict the need for intensive therapy

• Published in: Lupus science & medicine Vol. 11; no. Suppl 1; pp. A160 - A161
• Main Authors: Intapiboon, Porntip, Hassan, Sabih Ul, Arnold, Jack, Mahmoud, Khaled, Vital, Edward M, Md Yusof, Md Yuzaiful
• Format: Journal Article
• Language: English
• Published: London Lupus Foundation of America 14.03.2024; BMJ Publishing Group LTD; BMJ Publishing Group
• Subjects: Biomarkers; Lupus; Poster Presentations
• ISSN: 2053-8790
• DOI: 10.1136/lupus-2024-el.209

ObjectiveTo investigate the clinical and biomarkers factors predicting a requirement for intensive therapy or time to intensive therapy from diagnosis of Systemic Lupus Erythematosus (SLE).MethodsWe conducted a retrospective longitudinal study of all patients with a diagnosis of SLE from two Leeds Cohort databases (CONVAS and DEFINITION) for over 30 years follow-up. Data collection included demographics, clinical characteristics, the 2019 EULAR/ACR classification criteria score, the SLEDAI-2K score, and routine immunological tests. The primary endpoint was the time from SLE diagnosis to initiation of intensive therapy (cyclophosphamide, rituximab, belimumab, or other biologic agents). Univariable analysis (UVA) and multivariable (MVA) Cox proportional hazards regression models were used to test the potential predictors of the primary endpoint. MVA was done using forward selection and backward elimination with p<0.1 associated with the deviance used for inclusion in and exclusion from the model.Results229 SLE patients were included, and baseline characteristics are summarised in table 1. Total follow-up 3448 patient-years; median (IQR) follow up time was 11.6 (6.7, 21.4) years. Intensive therapy was initiated in 110 (48%) patients. The median (IQR) time to intensive therapy was 2 (0.5,8) years. Rituximab was the most common intensive therapy followed by cyclophosphamide (51.8% and 40.9%, respectively). In UVA, factors associated with increased risk of intensive therapies requirement were antibodies positivity for anti-Ro, anti-Sm and anti-RNP, cumulative number of Ab positivity, low complement levels, 2019 EULAR/ACR criteria score≥20, and higher cSLEDAI-2K score. While in MVA, anti-Ro+, low complement levels, and higher cSLEDAI-2K were associated with increased risk of intensive therapies requirement.ConclusionsNearly half of SLE patients required intensive therapy and this was predicted by anti-Ro+, low complement levels, and high clinical-SLEDAI-2K score at SLE diagnosis. At present, it is unclear whether patients should receive initial antimalarials, then immunosuppressants, then biologic therapies, or whether some patients should receive a first-line biologic therapy. Our data suggest that patients with these predictive factors develop more severe SLE, fail conventional therapies, and therefore are suitable for first-line biologic therapy. If our results can be validated, then such a strategy may prevent severe SLE.Abstract P155 Table 1Baseline characteristics, serology, and biomarkers in patients with Systemic Lupus Erythematosus Variables Total N=229 (%) Intensive therapy-given n=110 (%) Intensive therapy-not given n=119 (%) p-value Female gender 209 (91.3) 100 (90.9) 109 (91.6) 0.854 Mean age at diagnosis (SD) 38.8 (14.7) 35.67 (13.97) 41.70 (14.87) 0.002 Age by category Q1: under 28 63 (27.5) 38 (34.5) 25 (21.0) 0.032 Q2: 29 – 38 56 (24.5) 25 (22.7) 31 (26.1) 0.667 Q3: 39 – 51 55 (24.0) 28 (25.5) 27 (22.7) 0.738 Q4: over 51 55 (24.0) 19 (17.3) 36 (30.0) 0.032 Ancestry European 158 (69) 76 (69.1) 82 (68.9) South-Asian 31 (13.5) 17 (15.5) 14 (11.8) Oriental 5 (2.2) 2 (1.8) 3 (2.5) African/Caribbean 17 (7.4) 7 (6.4) 10 (8.4) Mixed 11 (4.8) 7 (6.4) 4 (3.4) Unknown 7 (3.1) 1 (0.9) 6 (5) Intensive therapy given Cyclophosphamide - 45 (40.9) - Rituximab - 57 (51.8) - Belimumab - 4 (3.6) - Ocrelizumab - 2 (1.8) - Efalizumab - 1 (0.9) - Clinical trial drug - 1 (0.9) - Median time to censored; year (IQR) 5.9 (1.7,11.6) 2 (0.5,8) 8.6 (5.2,15.3) < 0.001 Autoantibodies positivity Anti-dsDNA Ab 124 (54.4) 63 (57.8) 61 (51.3) 0.322 Anti-Ro Ab 108 (47.2) 66 (60) 42 (35.3) < 0.001 Anti-La Ab 34 (14.8) 23 (20.9) 11 (9.2) 0.013 Anti-Sm Ab 38 (16.6) 24 (21.8) 14 (11.8) 0.041 Anti-Sm/RNP Ab 55 (24) 30 (27.3) 25 (21) 0.268 Anti-RNP Ab 36 (15.7) 24 (21.8) 12 (10.1) 0.015 Anti-chromatin Ab 81 (35.4) 44 (40) 37 (31.1) 0.159 Anti-ribosomal P Ab 8 (3.5) 5 (4.5) 3 (2.5) 0.486 Mean cumulative number of Ab positivity (SD) 2.1 (1.6) 2.6 (1.8) 1.7 (1.3) < 0.001 Low complement (C3 and/or C4) 66 (28.8) 43 (39.1) 23 (19.3) < 0.001 aPL positivity 50 (21.8) 29 (26.4) 21 (17.6) 0.111 Median 2019 EULAR/ACR criteria score (IQR) 16 (12,22) 20 (16,25) 14 (11.5,18) < 0.001 2019 EULAR/ACR criteria score≥20 79 (34.5) 57 (51.8) 22 (18.5) < 0.001 Median SLEDAI-2K score (IQR) 10 (6,14) 13 (10,19.8) 7 (6,10) < 0.001 Median cSLEDAI-2K score (IQR) 8 (5,12) 12 (8,17) 6 (4,8) < 0.001 SLEDAI-2K≥10 118 (51.5) 85 (77.3) 33 (27.7) < 0.001 Abstract P155 Table 2Univariable and Multivariable Cox regression analysis of factors predicting the intensive therapies requirement Variables Univariable Hazard ratio (95% CI) Univariable p-value Multivariable Hazard Ratio (95% CI) Multivariable p-value Age under 28 vs >28 years 1.1 (0.73–1.70) 0.66 Not included Not included European ancestry vs non-European 0.85 (0.52–1.4) 0.52 Not included Not included Anti-dsDNA ab positivity 1.2 (0.81–1.7) 0.48 Not included Not included Anti-Ro Ab positivity 1.7 (1.20–2.60) 0.005 1.47 (1.00–2.19) 0.052 Anti-La Ab positivity 1.2 (0.75–1.9) 0.45 Not included Not included Anti-Sm Ab positivity 1.6 (1.00–2.60) 0.036 Included in MVA but removed from final model as p>0.1 Anti-RNP Ab positivity 1.5 (0.94–2.30) 0.093 Included in MVA but removed from final model as p>0.1 Cumulative number of Ab positivity 1.2 (1.1–1.3) 0.003 Excluded due to collinearity Low complement (C3 and/or C4) 2.4 (1.60–3.50) <0.0001 1.99 (1.32–2.97) <0.001 2019 EULAR/ACR criteria score>20 2.1 (1.50–3.10) <0.0001 Included in MVA but removed from final model as p>0.1 cSLEDAI-2K score 1.1 (1.07–1.12) <0.0001 1.10 (1.07–1.12) <0.001