[The neonatal form of Bartter's syndrome: current findings in etiology and physiopathology].

I. Time has come to distinguish "Bartter syndrome" from "Bartter disease". The latter is an autosomal recessive renal tubulopathy which manifests itself mostly during infancy and childhood. II. Bartter disease is caused neither by a primary renal potassium loss nor by a primary renal hyperprostaglandinism. All evidence is in favor of a defect in the chloride pump located at the thick ascending limb of Henle's loop. III. The most severe expression of Bartter disease is its neonatal form which is characterized by polyhydramnios, premature delivery and a life threatening sodium chloride loss during the early weeks of life. It takes several weeks before sodium wasting turns into renal potassium wasting. IV. Polyhydramnios not associated with echographically detectable fetal malformation is highly suggestive of Bartter disease. Prenatal diagnosis is based on the combination of fetal polyuria and elevated chloride in the amniotic fluid. V. In this setting the administration of indomethacin is useless and even dangerous from the 32nd week of gestation on. Similarly, indomethacin should not be given to the newborn Bartter patient for the first weeks and months of life. Treatment at that stage consist mainly of the administration of large amounts of fluid and sodium chloride. VI. Indomethacin can be used as soon as children with Bartter disease stop growing normally and preferably after the age of 18 months when kidney maturation is established. The daily dose should not exceed 2.5 mg/kg body weight. VII. Hypercalciuria is part of (the neonatal form of) Bartter disease and it is so severe that nephrocalcinosis seems to be the rule. This hypercalciuria is the direct consequence of the chloride reabsorption defect in Henle's loop. Research is needed to find an adequate solution to this problem.