Giant omphalocele is a major clinical challenge for pediatric surgeons. Whereas small- to medium-sized defects can be repaired primarily, larger omphaloceles cannot be closed at birth because the liver and small bowel have lost the right of domain to the abdomen. Two divergent strategies have evolved for treating these giant defects: (1) use of a silon chimney with gradual reduction of the contents of the sac, and (2) initial nonoperative management (epithelialization) of the omphalocele followed by repair of the residual ventral hernia. In an 18-year retrospective study, we have reviewed our experience with these treatment methods. Ninety-four infants underwent treatment for omphalocele between 1975 and 1993. Primary closure (PC) was possible in 55 patients, silon chimney (SC) was used in 15, and 7 had nonoperative management (NM) with epithelialization. In the remaining 17 infants, surgery was believed to be inappropriate because of the lethality of their associated anomalies. Major (but potentially survivable) anomalies were present in 26% of PC, 13% of SC, and 71% of the NM group patients. The majority of the liver was present in 73% of SC- and 86% of NM-treated omphaloceles. There was a decrease in length of stay, time to enteral feeding, and mortality over the 18-year period. However, those patients whose defects could not be closed primarily had consistently longer hospital stays. This was particularly true for the SC patients. The decreased use of total parenteral nutrition seems to reflect a shift from SC to NM rather than a decrease in the interval to full enteral feeding in any given treatment group over time.(ABSTRACT TRUNCATED AT 250 WORDS)