From 1980 to 1985, we performed biopsies on 87 adults with nephrotic syndrome (NS). The patients were tested for whether serologic studies obtained routinely at biopsy added to clinical diagnostic accuracy. Using history, physical examination, complete blood cell count (CBC), chemistry panel, urinalysis, and urine creatinine and protein, four nephrologists each predicted whether the patient had primary NS (PNS) or secondary NS (SNS), and the most likely histopathologic entity. Six months later, each nephrologist used this information, with results of tests of sera for fluorescent antinuclear antibody (FANA), rheumatoid factor (RF), complement components, hepatitis B surface antigen (HBsAg), venereal disease research laboratory serology (VDRI), cryoglobulins and protein electrophoresis (SPEP), with an erythrocyte sedimentation rate (ESR) and protein electrophoresis of the urine (UPEP), to make identical predictions. Histopathology was established by renal biopsy. We analyzed the concordance between nephrologists' choices and biopsy results both before and after serologic tests were available with a kappa statistic. Preserology concordance was moderate (kappa = 0.52), and identical to postserology concordance (kappa = 0.51) for both PNS versus SNS and actual histopathology. Serologies were rarely abnormal without clinical suspicion. These results suggest routine serologic testing does not improve diagnostic accuracy in adult NS.