Predictive value of residual active histologic lesions on renal flare in lupus nephritis patients with clinical remission

ABSTRACT Background Renal flare in lupus nephritis (LN) is a crucial contributing factor to poor kidney outcomes. This study aimed at evaluating the predictive value of residual active histologic lesions on renal flare in proliferative LN patients with clinical remission. Methods We retrospectively enrolled LN patients with class III/IV ± V (biopsy 1) who had undergone a protocol repeat biopsy (biopsy 2) at 7.3 (IQR: 6.5, 8.4) months after induction therapy with clinical remission and experienced renal flare within 3 years or had been followed up for at least 3 years without renal flare after biopsy 2 with maintenance therapy from two kidney units in China. Results A total of 114 eligible patients were included, 28 (24.6%) of whom developed a renal flare. Activity index (AI) at biopsy 2 was significantly associated with LN flare (P < .0001). If AI > 1, the OR for LN flare was 23.1 (95%CI, 5.1–103.8, P < .001). For patients with partial clinical remission compared with those with complete clinical remission, the OR for LN flare was 3.0 (95%CI: 1.1–8.3, P = .029). Multivariate analysis showed that anti-dsDNA positivity, presence of cellular/fibrocellular crescent, and endocapillary hypercellularity at biopsy 2 were independent risk factors for LN flare. When residual active histologic lesions were added to clinical variables, the area under the curve of the prediction model for LN flare significantly increased and the misclassification rate significantly decreased. Conclusions Renal flare in LN patients with clinical remission is strongly associated with the residual active histologic lesions.