Biopsy Morphometrics as Predictors of Treatment Response in Primary Nephrotic Syndrome

Clinical outcome of primary nephrotic syndrome (PNS) is highly variable, and predicting an individual patient's treatment response remains difficult. PNS is characterized by means of podocyte injury and loss. We hypothesized that histologic parameters related to podocyte depletion predict treatment response.

Retrospective cohort study.

We analyzed biopsy tissue of 106 patients with PNS (minimal change disease, N = 26; focal segmental glomerulosclerosis, N = 21; and membranous nephropathy [MN], N = 59) and 9 controls. Minimal change disease and focal segmental glomerulosclerosis were considered manifestations of the same entity, defined as idiopathic nephrotic syndrome (iNS), and analyzed as one group. Patients' baseline clinical and follow-up data were recorded. Kidney biopsies, stained for podocyte-specific and fibrosis markers, were quantitatively analyzed.

Glomerular density, glomerulosclerosis, podocyte number, podocyte density, and cortical fibrosis.

Complete remission (CR) and delayed treatment response.

Odds ratios and receiver operating characteristic-the area under the curve (ROC-AUC) values identified predictors.

In patients with iNS, the respective partial remission and CR rates were 29% and 60% during a median follow-up of 40 months. The majority of patients received high-dose corticosteroid treatment. Quantitation of cortical fibrosis had the highest discriminative power (ROC-AUC value, 0.79; 95% CI, 0.655-0.923) to predict CR. Other significant predictors included podocyte density, nonsclerotic glomerular density, and percentage of nonsclerotic glomeruli.In patients with MN, respective partial remission and CR rates were 41% and 54% during a median follow-up of 50 months. The percentage of nonsclerotic glomeruli and nonsclerotic glomerular density were predictors for CR (patients receiving immunosuppressive treatment [ROC-AUC value, 0.71; 95% CI, 0.535-0.893]; patients receiving nonimmunosuppressive treatment alone [ROC-AUC value, 0.80; 95% CI, 0.584-1.000]).

Relatively small cohorts prevented the use of covariates.

In patients with iNS, higher podocyte density and nonsclerotic glomerular density, and lower glomerulosclerosis and cortical fibrosis predicted CR. In patients with MN, lower glomerulosclerosis and higher nonsclerotic glomerular density predicted CR. Biopsy parameters may thus be useful for estimating proteinuria outcome.