Although renal biopsy is the definitive investigation in kidney disorders and is particularly helpful in distinguishing steroid-responsive nephrotic syndrome (SRNS) from focal glomerulosclerosis (FGS), it is attended by a small risk to the patient. Accordingly, noninvasive tests have been used to predict the response to steroids and the underlying renal histologic diagnosis in nephrotic syndrome. The performance of these tests has, however, not been encouraging. We have therefore compared the reliability of the conventional selectivity index (SI) of serum and urinary transferrin and immunoglobulin G (IgG) against other tests of urinary proteins, sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS PAGE), and isoelectric focusing (IEF). SI, SDS PAGE, and IEF were carried out in children with nephrotic syndrome within 2 months of clinical presentation. Thirty-one children who had SRNS were compared with 26 who had biopsy-proved FGS and who were steroid resistant. SDS PAGE and IEF revealed excretion of albumin and transferrin only, with homogeneous anionic charge, respectively, in SRNS but unrestricted excretion of additional proteins IgG, beta2-microglobulin, and lysozyme with heterogeneity of electrical charge in FGS. With SDS PAGE and IEF we were able to predict all children who had SRNS and FGS; the SI test predicted all steroid-resistant patients with FGS but was able to predict only 41.7% of the patients with SRNS. Therefore the negative predictive value for steroid response was 58.8% by SI and 100% by SDS PAGE and IEF; the positive predictive value was 100% by SI, SDS PAGE, and IEF. We illustrate the value of SDS PAGE in guiding management in a further seven children with FGS in whom there was either an initial discordance between renal biopsy results and steroid responsiveness or when biopsy was delayed. Accordingly, SDS PAGE and IEF of urinary proteins appear to be useful tests in the diagnosis and management of SRNS and FGS.