Document resource
Objectives To identify clinical and serological predictors present at the diagnosis of systemic lupus erythematosus (SLE) associated with mortality, and to describe the distribution of causes of death, assessing potential temporal changes over the past two decades (pre-2014 vs post-2014) overall and according to lupus nephritis (LN) and ethnicity in a retrospective cohort of 860 patients.Methods Contingency tables were constructed for qualitative variables, and their distributions were compared using chi-square or likelihood-ratio tests, as appropriate. Associations between each baseline variable and mortality were evaluated using logistic regression models, reporting odds ratios and corresponding 95% confidence intervals (OR, 95% CI). A multivariable logistic regression model was also fitted, including variables with p less than 0.10 in bivariate analyses plus age at diagnosis.Results In bivariate analyses, higher mortality was associated with leukopenia (OR 2.72; 95% CI 1.91–3.88), thrombocytopenia (OR 2.25; 95% CI 1.48–3.43), serositis (OR 2.23; 95% CI 1.59–3.13), central nervous system involvement (OR 2.13; 95% CI 1.44–3.13), and proteinuria (OR 1.69; 95% CI 1.19–2.39). Anti-Ro negativity was also associated with higher mortality (OR 2.03; 95% CI 1.41–2.93). In the multivariable model, independent associations were observed for proteinuria (p=0.0023), leukopenia (p=0.0019), thrombocytopenia (p=0.0038), rash (p=0.0083), serositis (p=0.0017), anti-Ro (p=0.0195), and age at diagnosis (p<0.0001). Lymphopenia (p=0.0528) and lupus anticoagulant (p=0.0604) were significant at the 10% level. A total of 172 deaths were recorded; the leading causes were infection (25.6%), malignancy (21.5%), and cardiovascular events (18.6%). No significant differences in the distribution of causes of death were observed by LN status (p=0.2999) or ethnicity (p=0.8580). In the subgroup with available information on year of last visit (n=41; 13 pre-2014 vs 28 post-2014), no significant temporal changes in the distribution of causes were detected (p=0.1072), including after stratification by LN or ethnicity.Conclusions In this SLE cohort, mortality was independently associated with baseline manifestations suggesting greater systemic/renal burden (proteinuria), cytopenias (leukopenia, thrombocytopenia), serositis, rash, anti-Ro profile, and older age at diagnosis. Deaths were mainly due to infection, malignancy, and cardiovascular events, with no statistically significant differences according to LN status, ethnicity, or pre-/post-2014 period in the available data.