Validation cohort of a tool to predict proliferative histological class in lupus nephritis based on clinical and...

Primary: To validate a previously developed lupus nephritis (LN) histologic class predictor tool (III or IV±(V) vs V). In this instrument, urinary sediment, anti-dsDNA antibody and serum creatinine were predictor variables. Secondary: To evaluate its accuracy in different cut levels of urinary red cell count and LN patient's profile that may be useful to guide diagnosis and treatment, especially when kidney biopsy is not available.

A retrospective, cross-sectional study of 196 patients with SLE who underwent kidney biopsy, analysing sensitivity, specificity, positive and negative predictive values, accuracy, and positive and negative likelihood ratios.

81.6% of the patients were female, 60.2% were non-Caucasian and the mean age at the time of the biopsy was 31.2±10.4 years. 30 patients presented class III, 104 class IV, 36 class V and 26 mixed classes (7 class III+V and 19 class IV+V). In the validation cohort, sensitivity was 90.6%, specifically was 66.7%, positive predictive value was 92.4% and accuracy was 86.2%, in predicting proliferative classes (class III or IV(±V) vs class V). There was no difference when analysing urinary red cell counts >5 x 10³/mL, >10 x 10³/mL or >20 x 10³/mL, as well as when the tool was applied in the first or recurrent LN and in proliferative LN without class V.

The validation of a tool to predict proliferative histologic class showed good performance, like that found in the development cohort, without difference in accuracy with different cut-off points for urinary red cell count or number of flares of LN. It is easily applied and can be accessed via the internet (https://nefritelupica.medicalcore.com.br).