To evaluate the clinical and laboratory findings in pediatric patients with urolithiasis in a case control study and to obtain variables predicting a low risk of having urolithiasis. Retrospective analysis of 24 cases of children with urolithiasis (age range 3 to 17 years old) admitted to the pediatric emergency room or the hospital ward. Clinical and laboratory information was obtained from the clinical charts and compared with a control group of 29 patients with history of abdominal pain admitted to the hospital. Binary logistic regression and recursive partitioning analysis were used to obtain variables predicting a low risk of having urolithiasis. There was no difference in age, gender, dysuria or nausea (p > 0. 2) between groups. A difference regarding the presence of fever (p = 0.007), haematuria (p = 0.001), costovertebral angle tenderness (p = 0.004), family history (p = 0.007) and abdominal pain lasting more than 48 hs (p = 0.04) was detected. After logistic regression and using recursive partitioning, the presence or history of fever, absence of gross haematuria and no family history of urolithiasis showed a 100% negative predictive value and sensitivity (95% CI 84.7 to 100 and 89.2 to 100 respectively). Children with abdominal pain present a low risk of having urolithiasis if there is a history or presence of fever, absence of gross haematuria and a negative family history of urolithiasis in a first degree relative.