Predicting first-year relapses in children with nephrotic syndrome

Objective. More than half of the children diagnosed with nephrotic syndrome will have relapses. These can be infrequent relapses (IRs: <2 in 6 months or <3 in a year) or frequent relaspses (FRs: >2 in 6 months or >3 in a year ). Patients who relapse while on alternate day steroids or within 1 month o f discontinuation of steroid therapy are considered steroid-dependent (SD; J Pediatr. 1982; 101:514-518). Patients with an IR course have a better lon g-term prognosis, and many of them have minimal-change disease without mesa ngial hypercellularity or sclerosis. The purpose of our study was to identi fy factors at initial presentation that could predict the relapse pattern i n the first year after diagnosis, without taking into consideration the his topathology found on renal biopsy. Design. We analyzed the medical records of children who were seen by us bef ore March 1997 and followed for at least 1 year. Variables selected in the study were age, sex, race, presence or absence of hematuria, and days to re mission (defined as protein-free) at the initial presentation, because they could relate to the pattern of relapses (ie, IR, FR, and SD). Results. Of 70 patients, 14 were excluded because of insufficient data. The re were 38 males (67.9%) and 18 females (32.1%), giving a male: female rati o of 1.8:1. Median age at presentation was 3.25 years (range: 1.5-13), and 76.9% were white, 8.9% black, 7.1% Hispanic, and 7.1% other. Of all the pat ients, 23 were IR (41.1%), 9 were FR (16.1%), and 24 were SD (42.9%). Media n days to remission were 10 (range: 2-60), on Prednisone 60 mg/M-2 daily. H ematuria was present initially in 26 patients (46.4%), and absent in 30 (53 .6%). Age, sex, race, and hematuria, as independent variables, were not pre dictors of relapses in the first year. However, using a stratified analysis based on the presence or absence of hematuria, we found that if the remiss ion occurred within the first week of therapy, the patients without hematur ia were more likely to be IR. The sensitivity and specificity of this findi ng were 67% and 89%, respectively, with a positive predictive value of 94%. Conclusion. We conclude that of all the presenting features, the rapidity o f initial response to steroid therapy combined with the presence of hematur ia, could predict future relapses and should be well documented.