Clinical predictors of active LN development in children – evidence from the UK JSLE Cohort Study

Abstract

Background Juvenile-onset systemic lupus erythematosus (JSLE) patients may develop lupus nephritis (LN) during their initial presentation, or later in their disease. This study aimed to assess whether clinical/demographic factors characterize patients with LN within the United Kingdom JSLE Cohort Study, and whether such factors predict subsequent LN development. Methods Univariate logistic regression modelling compared clinical/demographic factors in patients with and without LN at baseline. For those who subsequently developed LN, Cox proportional-hazard modelling was used to test the association between such factors and time to LN development. Covariates with p < 0.2 univariately were included within a multiple-regression model. Results A total of 121/331 (37%) patients presented with active LN at baseline, with first American College of Rheumatology (ACR) score ( p < 2.0 × 10–16), severe hypertension ( p = 0.0006), proteinuria ( p < 2.0 × 10–16), creatinine ( p = 1.0 × 10–16), erythrocyte sedimentation rate ( p = 1.0 × 10–16), neutrophils ( p < 2.0 × 10–16), complement 3 (C3) ( p = 4.0 × 10–16) and ethnicity ( p = 3.0 × 10–13) differing between those with and without LN. Of the 210 individuals without active LN at baseline, 13 patients had a single visit and were excluded from further analysis. Thirty-four of 197 (17%) developed LN after a median of 2.04 years (interquartile range, 0.8–3.7), with higher ACR scores ( p = 0.014 , hazard ratio (HR) = 1.45, 95% confidence interval (CI) = 1.08–1.95) and lower C3 levels ( p = 0.0082 , HR = 0.27, 95% CI = 0.10–0.68) demonstrated as predictors of subsequent LN. Conclusions Clinical and demographic factors can help to characterize patients at increased risk of LN.