Background Damage in SLE is associated with mortality. Not every damage manifestation is associated in the same way. Some studies were made assessing this relation but in small and heterogeneous samples making difficult to obtain meaningful conclusions.
Objectives To evaluate patterns of damage accrual and mortality in a large sample of SLE patients.
Methods SLE patients from RELESSER were studied. After K-means cluster analysis, different clusters of patients with similar characteristics in terms of damage accrual were identified. Kaplan-Meier log-rank test and Cox regression were used to analyse mortality in each group.
Results 3,656 SLE patients from 45 Rheumatology Units across Spain were studied. 90.3% were women. 93.1% Caucasian, 5.2% Latinamerican and 1.6% other races. Mean age (±SD) at SLE diagnosis was 35.1±14.6 years. Mean follow-up time was 120.1±87.6 months. Mean SLICC/ACR damage index (SDI) score was 1.1±1.6. Average number of organ systems affected in terms of damage was 0.6±1.0. 207 (5.6%) patients died.
The SDI organ systems most frequently damaged were: musculoskeletal (MS) (13.7%), ocular (8.5%), cardiovascular (CV) (7.9%) and renal (6.1%). The least frequently present were:gastrointestinal (1.9%), diabetes mellitus (2.4%) and premature gonadal failure (2.5%).
Three clusters (C) were formed. C1 had mild or no damage. Patients in C2 had MS damage but no CV. In C3 all had CV damage.
Among the 3 clusters, there were significant differences in prevalence of damage in each organ system assessed by the SDI, in average SDI score, in number of SDI systems damaged and mortality rate (p<0.001, for all comparisons). Detailed data are shown in Table 1.
C3 patients were older at SLE diagnosis and had higher %of males (p<0.001, for both comparisons).
Comparing survival curves of the 3 clusters, the log-rank test showed significant differences (p<0.001 for the triple and double comparisons). Analysing the survival rate at 10, 20 and 30 years from diagnosis of SLE, we found lower survival in patients of C2 and C3 compared to C1 (p=0.068 when C2 is compared to C1 at 10 years, p<0.01 for all the other cases). Between C2 and C3, there were no significant differences in survival at 10 years and itwas significantly lower in C3 at 20 and 30 years (p=0.025 for both).
Cox regression analysis showed that, compared with C1, the mortality hazard ratio of C2 and C3 was 1.9 and 3.5 higher respectively (p<0.001, for both comparisons).
Conclusions SLE patients can be divided into different homogeneous groups (clusters) based on damage accrual. These clusters have different mortality rates.
Disclosure of Interest J. Pego Grant/research support from: Spanish Society of Rheumatology, FIS/ISCIII (PI11/02857), BIOCAPS from the EU 7th Framework Programme/REGPOT-2012-2013.1 (316265), GSK, Roche, Novartis,UCB., A. Lois: None declared, F. Lόpez: None declared, M. Galindo: None declared, J. Calvo: None declared, J. Uña: None declared, V. Balboa: None declared, A. Olivé: None declared, C. Mouriño: None declared, T. Otόn: None declared, J. Ibañez: None declared, L. Horcada: None declared, A. Sánchez: None declared, R. Blanco: None declared, C. Montilla: None declared, R. Melero: None declared, E. Diez: None declared, M. Fernández: None declared, E. Ruiz: None declared, J. Hernández: None declared, M. Gantes: None declared, B. Hernández: None declared, A. Pecondόn: None declared, N. Lozano: None declared, G. Bonilla: None declared, V. Torrente: None declared, I. Rúa: None declared