objectives: to identify initial predictive factors of renal involvement in children and adolescents with henoch-sch？nlein purpura. methods: we reviewed the medical records of 142 patients admitted to our university hospital over a 21-year period with a diagnosis of henoch-sch？nlein purpura. the initial predictive factors assessed, observed during the first 3 months, included: demographic data, clinical manifestations (persistent palpable purpura, arthritis, abdominal pain, severe abdominal pain, gastrointestinal bleeding, orchitis, central nervous system involvement and pulmonary hemorrhage), laboratory tests (serum iga levels) and treatment given (corticosteroids, intravenous immunoglobulin and immunosuppressive drugs). patients were divided into two groups (presence or absence of nephritis) and assessed by univariate and multivariate analysis. results: evidence of nephritis was detected in 70 patients (49.3%). the univariate analysis revealed that severe abdominal pain (p = 0.0049; or = 1.6; 95%ci 1.18-2.21), gastrointestinal bleeding (p = 0.004; or = 1.6; 95%ci 1.10-2.26) and corticosteroid use (p = 0.0012; or = 1.7; 95%ci 1.28-2.40) were all associated with increased incidence of renal involvement. in the multivariate analysis, logistic regression demonstrated that the only independent variable that predicted nephritis was intense abdominal pain (p < 0.012; or = 2.593; 95%ci 1.234-5.452). conclusions: severe abdominal pain was a significant predictor of nephritis in henoch-sch？nlein purpura. consequently, pediatric patients exhibiting this clinical manifestation should be rigorously monitored, due to the increased risk of renal involvement.