1) To evaluate the relationships between urinary sodium (UNa), potassium (UK), and calcium (UCa) excretion in the pediatric population; and 2) to determine the effect of increasing potassium intake in patients with idiopathic hypercalciuria and investigate whether this intervention can be offered as another mode of therapy in this patient population. Prospectively, we determined UNa, UK, UCa, and creatinine (Cr) concentrations in randomly collected urine samples from children on initial evaluation for urinary frequency, dysuria, hematuria, enuresis, or kidney stones to identify children with hypercalciuria. The outpatient renal clinic of an academic hospital. Twenty-three black children (13 girls and 10 boys) and 77 white children (44 girls and 33 boys) 3.92 to 16.67 years of age. Eleven children with hypercalciuria were given potassium supplementation or placed on a high-potassium diet for at least 2 weeks. UNa to UK, UNa to Cr, UK to Cr, and UCa to Cr ratios were calculated from measured levels of urinary minerals. These were repeated in 11 hypercalciuric patients after 2 weeks of increased potassium intake. A total of 100 urine samples were analyzed. The UCa/Cr ratio in blacks 0.04 +/- 0.06 (mean +/- standard deviation) was significantly lower than in whites 0.16 +/- 0.12. There were 21 hypercalciuric white children versus only 1 black child. Linear regression analysis revealed a positive direct correlation between UNa/Cr and UCa/Cr in all 100 subjects and in whites alone but not in blacks. An inverse relationship existed between UK/Cr and UCa/Cr in all subjects and in whites and showed a strong trend in blacks. A marked direct relationship was found between UNa/K and UCa/Cr in all subjects (r = .43) as well as in whites (r = .59) and blacks (r = .49). One black child and 10 white hypercalciuric children were treated with "extra" K for at least 2 weeks. The UNa/K decreased from 4.73 +/- 2.28 to 1.98 +/- 1.09, and the UCa/Cr decreased from 0. 31 +/- 0.10 to 0.14 +/- 0.07, with resolution or improvement of the patients' symptoms. In our patient population with urinary symptoms, the UCa/Cr ratio in black children is lower and hypercalciuria less common than in white children. In both white and black populations, the UNa/K ratio had the strongest association with the UCa/Cr ratio, indicating an opposing role of UNa and UK on the UCa/Cr ratio. Increased potassium intake was found to be beneficial for hypercalciuric children by decreasing the UNa/K ratio and, consequently, the UCa/Cr ratio.