The predictive value of urinary erythrocyte morphology for proliferative glomerular kidney disease

Background and aim Glomerular erythrocyturia (GlomEry) is usually associated with proliferative kidney diseases. In our retrospective cohort, we aimed to validate the predictive value of GlomEry criteria ≥ 40% dysmorphic erythrocytes (DysEry) or ≥ 5% acanthocytes (AcaEry) or at least 1 erythrocytic cast (CastEry) and of two new indices - the count of DysEry per high power field (HPF) and per microliter of urine (Stansfeld-Webb (SW)) method, for proliferative disease. Materials and methods We included patients with erythrocyturia from 2015 to 2016. Based on renal histology, we divided them into a proliferative and a non-proliferative disease group. Urine erythrocyte count was done using SW and urinary sediment examination was carried out by skilled nephrologists. Sensitivity, specificity, and cutoff values were determined using ROC curves. Results We included 90 patients (33% women), median age of 63 (IQR 51, 71) years. In the proliferative group, proteinuria was lower (2.4 vs. 6.6 g/day), and SW erythrocyturia was higher (174 (IQR 60, 353) vs. 44 (IQR 20, 67) × 106/L) than in the non-proliferative group. The threshold to differentiate between the proliferative and non-proliferative group was determined at ˃ 43% of DysEry (sensitivity 73%, specificity 79%, AUC 0.808) and at ˃ 2% AcaEry (sensitivity 71%, specificity 56%, AUC 0.647). No significant difference in CastEry was found between groups. Among tested parameters, the calculated number of DysEry/HPF > 6.7 (sensitivity 77%, specificity 92%, AUC 0.878), followed by DysEry/SW > 28 × 106/L (sensitivity 76%, specificity 86%, AUC 0.879), discriminated those two groups best. Conclusion In concordance with known GlomEry criteria, > 43% of DysEry predicted proliferative kidney disease, whereas CastEry did not, and AcaEry predicted poorly. The best predictor of proliferative glomerular disease was DysEry/HPF, closely followed by DysEry/SW.