A modern approach to selectivity of proteinuria and tubulointerstitial damage in nephrotic syndrome

The selectivity of proteinuria, introduced in clinical nephrology in 1960 and useful in predicting steroid responsiveness in nephrotic syndrome, found little place in clinical practice in subsequent decades, since its assessment did not appear to help predict histologic diagnosis or determine prognosis. The amount of proteinuria and the degree of tubulointerstitial damage appeared to be better predictors of functional outcome. A correlation between them has been found, referred to some toxicity of proteinuria on tubular cells, but so far no single feature or component of proteinuria has been identified as being responsible for this toxicity.

We evaluated 89 patients with nephrotic syndrome [9 with minimal change disease (MCD), 29 with primary focal segmental glomerulosclerosis (FSGS), and 51 with idiopathic membranous glomerulonephritis (MGN)] to determine if the selectivity of proteinuria was associated with tubulointerstitial damage. A semiquantitative grading of histologic lesions and qualitative evaluation of the "tubular" component of proteinuria expressed as a pattern of sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and as fractional excretion of the low molecular weight (LMW) protein alpha1-microglobulin (FE alpha1m) were used. A second aim of the study was to assess the predictive value on functional outcome [remission or progression to chronic renal failure (CRF)] and response to therapy of the selectivity of proteinuria, considered alone and in combination with FE alpha1m.

Proteinuria was classified as highly selective [selectivity index (SI) < or = 0.10, N = 15], moderately selective (SI > or = 0.11 < or = 0.20, N = 34), or nonselective (SI > or = 0.21, N = 40). A significant relationship was found between the SI and the histologic degree of tubulointerstitial damage (score 0 to 1 vs. score > or =2, P = 0.000), severity of the tubular component of proteinuria (mixed SDS-PAGE pattern with LMW proteins not lower than 23 kD vs. mixed pattern with LMW proteins up to 20 to 10 kD, P = 0.000), and FE alpha1m (values below vs. above a defined cut-off, P = 0.000). The functional outcome was evaluated in 60 patients with baseline normal renal function (serum creatinine 0.97 +/- 0.19 mg/dL). The patients with high, moderate, or nonselective proteinuria had 100, 50, and 29% of complete or partial remission (P = 0.0001) and 0, 25, and 35% of progression to CRF, respectively (P = 0.050). In 45 patients with moderately selective (N = 28) and nonselective (N = 17) proteinuria, according to some arbitrary cutoffs for FE alpha1m (MGN, < or = vs. > 0. 240% of creatinine clearance; FSGS and MCD, < or = vs. > 0.350%), the remission rate was 62 versus 6% in patients with FE alpha1m below or above the cutoffs (P = 0.0001), and progression to CRF was 7 and 69%, respectively (P = 0.0001). The response to therapy (complete or partial remission at the last observation), evaluated retrospectively in 40 patients, was 100, 67, and 33% in high, moderate, and nonselective proteinuria (P = 0.0002); in 30 patients with moderate and nonselective proteinuria, according to an FE alpha1m value that was < or = or > the cutoffs, the response rate was 75 versus 10% (P = 0.001).

There is a significant relationship between selectivity of proteinuria and tubulointerstitial damage. Moreover, the selectivity of proteinuria has a predictive value on functional outcome. When proteinuria is highly selective, the tubulointerstitial damage is rather infrequent, and 100% of patients develop clinical remission. When proteinuria is moderately selective or nonselective, increasing numbers of patients develop tubulointerstitial damage; in these patients, the functional outcome and response to therapy is partly dependent on tubulointerstitial involvement, and the best predictor of functional outcome is the combination of SI and FE alpha1m.