Since 1973, 178 children with portal hypertension (PH) have been seen at Instituto da Criança of the University of São Paulo Medical School. Fifty-five of these children were excluded from this analysis for various reasons, including no treatment required, death before treatment, or incomplete data. From the remaining 123 children with esophageal varices, only 96 (76.1%) of them had at least one episode of upper gastrointestinal hemorrhage. Eighty-eight children were submitted to injection sclerotherapy; 26 treated prophylactically, and 62 for treatment of previous bleeding. Eleven (42.3%) children from the prophylactic group bled from esophageal varices during the treatment. They were all successfully managed thereafter. Satisfactory results were achieved in 53 (85.4%) children in the therapeutic group. Twenty-eight (45.1%) children had at least one episode of bleeding after beginning of sclerotherapy, 19 of whom eventually had successful control of the variceal bleeding. From 1973 to 1984, distal splenorenal shunt (DSS) was the procedure of choice for the treatment of bleeding esophageal varices. Forty-two children have undergone DSS during this period. Only one child was shunted prophylactically. Since 1985, injection sclerotherapy has been the first choice for the treatment and only seven children with sclerotherapy failure have since been treated by DSS. Characteristically these children had very similar splenoportographic pattern with huge esophageal and gastric varices and deviation of portal vein blood flow toward the left gastric vein.(ABSTRACT TRUNCATED AT 250 WORDS)