Delayed lupus nephritis in the course of systemic lupus erythematosus is associated with a poorer treatment response: a multicentre, retrospective cohort study in Japan

Abstract

Objective: The objective of this study was to investigate possible differences in treatment responses between two categories for the onset of lupus nephritis.

Methods: We performed a multicentre, retrospective cohort study of class III-V lupus nephritis patients diagnosed between 1997 and 2014. The renal responses to initial induction therapy were compared between patients who developed lupus nephritis within one year from diagnosis of systemic lupus erythematosus (early (E-) LN) and the remainder (delayed (D-) LN) using the Kaplan-Meier method. We determined the predictors of renal response as well as renal flares and long-term renal outcomes using multivariate Cox regression analyses.

Results: A total of 107 E-LN and 70 D-LN patients were followed up for a median of 10.2 years. Log-rank tests showed a lower cumulative incidence of complete response in D-LN compared with E-LN patients. Multivariate analysis identified D-LN (hazard ratio (HR) 0.48, 95% confidence interval (CI) 0.33-0.70), nephrotic syndrome at baseline, and a chronicity index greater than 2 as negative predictors of complete response. D-LN patients were more likely to experience renal flares. D-LN (HR 2.54, 95% CI 1.10-5.83) and decreased renal function were significant predictors of chronic kidney disease at baseline.

Conclusion: D-LN was a predictor of poorer treatment outcomes, in addition to renal histology and severity of nephritis at lupus nephritis onset.

Keywords: Lupus nephritis; delayed lupus nephritis; systemic lupus erythematosus; treatment response.

Figure 1

Cumulative incidence of CR over 24 months from initial induction therapy between E-LN and D-LN groups in all patients (a), and those who received cyclophosphamide or mycophenolate mofetil as induction therapy (b). Also shown is the cumulative incidence of CR between the E-LN and D-LN groups with class III/IV or class III+V/IV+V (c). Unadjusted HR (D-LN to E-LN) (95% CI): (a) 0.60 (0.43–0.84), P = 0.003; (b) 0.58 (0.33–1.02), P = 0.057; (c) 0.74 (0.44–1.23), P = 0.250 in class III/IV and 0.77 (0.42–1.43), P = 0.413 in class III+V/IV+V. Multivariate adjusted HR (D-LN to E-LN) (95% CI): (a) 0.48 (0.33–0.70), P < 0.001; (b) 0.54 (0.30–0.99), P = 0.047; (c) 0.68 (0.39–1.18), P = 0.175 in class III/IV and 0.44 (0.20–1.01), P = 0.053 in class III+V/IV+V. CR: complete response; E-LN: early lupus nephritis; D-LN: delayed lupus nephritis; HR: hazard ratio; CI: confidence interval.

Figure 2

Renal relapse-free survival over 20 years from initial induction therapy between the E-LN and D-LN groups in all patients (a), and those who achieved CR within 6 months (b). Also shown is survival without CKD over 20 years from initial induction therapy between the E-LN and D-LN groups in all patients (c) and the associations among D-LN, renal response status at 6 months, renal flares and long-term renal outcomes analyzed using mediation analysis with a generalized structural equation model (d) (see Supplementary Table 10 and Supplementary Figure 1, for more details). Unadjusted HR (D-LN to E-LN) (95% CI) (a) 2.10 (1.23–3.59), P = 0.007; (b) 1.73 (0.78–3.81), P = 0.176; (c) 1.70 (0.81–3.57), P = 0.160. Multivariate adjusted HR (D-LN to E-LN) (95% CI) (a) 1.71 (0.95–3.08), P = 0.072; (b) 1.68 (0.70–4.05), P = 0.247; (c) 2.54 (1.10–5.83), P = 0.028. E-LN: early lupus nephritis; D-LN: delayed lupus nephritis; CR: complete response; CKD: chronic kidney disease; HR: hazard ratio; CI: confidence interval; Coef: coefficient value.