THU0284 WHAT DO HEMATOLOGICAL ABNORMALITIES TELL US IN SLE? RESULTS FROM TWO INDEPENDENT MULTICENTER EUROPEAN SLE COHORTS

Background Detailed analysis of hematological manifestations are limited and their clinical impact on disease remain obscure. Objectives To scrutinize factors associated with different hematological abnormalities (HA) in SLE patients and their impact on infections, bleeding and damage accrual. Methods A dataset (GIPTSLE) originated from SLE patients in Europe was studied. The dataset consisted of six monthly visits of each patients for at least 2 years. Results were compared with another well-established SLE cohort from Sweden (Swedish SLE network). Patients in both cohorts fulfilled the ACR1982 SLE criteria. Variables collected at each visit included CBC, presence or absence of haemolytic anemia, all domains of SLEDAI-2K, medications, infection or bleeding episodes, Systemic Lupus International Collaborating Clinics/ACR damage index (SDI). HA occurred at the beginning or during follow-up were defined according to the ACR criteria. HA was only considered in patients with no prior use of immunsuppressives or biologic agents at least 3 months before the first HA events. The Swedish confirmation cohort included all the variables of interest. We excluded all haematological domains from SLEDAI-2K (aSLEDAI) for the analysis. Based on six monthly change in SDI, each visit was labelled either “damage transition”or “non-damage transition”. Generalized estimating equations (GEE), multiple logistic regression, Chi-square test and independent samples t test were used where it was appropriate. Results Of the 1430 visits of 286 patients analysed from the first cohort (89.5% female, 95.8% Caucasian, 64.7% dsDNA positive) 60% had at least one HA. At the enrolment, the median (range) disease duration was 8 yrs (0-38), aSLEDAI-2K was 6 (0-25) and SDI was 0(0-6). Upon GEE analysis, lymphopenia was significant for SDI (OR=2.5, 1.0-5.9).Upon GEE analysis after further adjusting prednisolone exposure, HA found be significant for SDI was thrombocytopenia (OR=3.6, 1.04-12.1) but no significant evidence for the dependence of this relationship on time. The effect of lymphopenia was attenuated. HR for bleeding was significant when platelets were below 50K (HR=2.6, 1.15-6.1) Of the 1395 patients analysed from the second set (86.8%female, 93.0% Caucasian, 61.6 dsDNA positive), 63.4% had HA. At the last visit, the median (range) disease duration was 14yrs (1-65) and SDI was 1(0-14). Upon multiple regression analysis after adjusting age, sex and prednisolone exposure both thrombocytopenia (OR=2.2, 1.3-3.6) and lymphopenia (OR=1.8, 1.2-2.6) had significant effect on SDI. HA at the time of diagnosis or occurring at any time during disease course did not contribute to mortality. Conclusion Lymhopenia contributes independently damage accrual in European background. Lymphopenia and thrombocytopenia at early disease appear to be independent risk factors for subsequent damage. References [1] Seminar in Arthritis and Rheumatism 45(2016):675-83 Disclosure of Interests Sule Yavuz: None declared, Dondu Cansu: None declared, Dionisis Nikolopoulos: None declared, Francesca Crisafulli: None declared, Ana M Arantunes: None declared, Christina Adamichou: None declared, Sara Reid: None declared, Chiara Satgnaro: None declared, Laura Andreoli: None declared, Angela Tincani Consultant for: UCB, Pfizer, Abbvie, BMS, Sanofi, Roche, GSK, AlphaSigma, Lillly, Jannsen, Cellgene, Novartis, Maria Moraes-Fontes: None declared, George Bertsias: None declared, Marta Mosca Paid instructor for: GlaxoSmithKline, Lilly, UCB, Dag Leonard: None declared, Antonis Fanouriakis Paid instructor for: Amgen, GSK, Speakers bureau: Abbvie, Enorasis, Genesis Pharma, Lars Ronnblom: None declared, SWEDISHSLE NETWORK: None declared