Comment on: Short- and long-term outcomes of patients with pure membranous lupus nephritis compared with patients with proliferative disease

Dear Editor, We read with great interest the recent study by Kharouf et al. published in Rheumatology (2025) [1], which compared short- and long-term outcomes between patients with pure membranous lupus nephritis (MLN) and proliferative lupus nephritis (PLN) [1]. The authors concluded that MLN is not a benign disease and may be associated with poor long-term outcomes, challenging the conventional paradigm. While this work provides valuable insights into lupus nephritis management, we have identified several methodological concerns that may affect the validity and generalizability of the results. First, the multivariable Cox proportional hazard models used to assess long-term outcomes (e.g. end-stage kidney disease [ESKD] and mortality) appear to be at high risk of overfitting. Established statistical guidelines recommend ∼10 events per variable (EPV) to ensure reliable estimates in a regression model [2–4]. In this study, the number of events for key outcomes is limited: for instance, ESKD occurred in only 28 patients (4 in MLN and 24 in PLN), while the multivariable models in Table 3 included up to 10 covariates per outcome. With an EPV likely below the recommended threshold, the model estimates (e.g. hazard ratios) may be unstable and exaggerated, leading to spurious associations. This issue is particularly critical for outcomes with fewer events, such as cardiovascular events (34 events) and death (27 events), where the models may lack sufficient power.