The H2[18O] tracer dilution method was applied to quantify total body water in 6 normal nonpregnant women, 8 normal pregnant women (gestation 29.3 +/- 6.0 weeks, range 21-37 weeks), and 8 insulin-dependent diabetic pregnant subjects (gestation 29.4 +/- 6.1 weeks, range 20-37 weeks). Plateau or equilibrium enrichment of 18O in the breath CO2 was achieved at approximately 2 h in nonpregnant normal subjects. In contrast, in pregnant subjects equilibrium plateau was reached later, at approximately 3 h. In order to examine whether the amniotic fluid/fetal compartment had also reached equilibrium with the maternal water compartment, amniotic fluid samples were obtained from 5 women undergoing elective cesarean section, approximately 5 h after the administration of tracer H2[18O]. The 18O enrichment of amniotic fluid was similar to that of the mother. Total body water was similar in normal and diabetic pregnant subjects (40.4 +/- 4.4 vs. 40.1 +/- 5.8 kg) and represented 55.8 +/- 5.4 and 58.7 +/- 5.5% of body weight, respectively. These data demonstrate the usefulness of the H2[18O] tracer method to measure total body water in pregnancy and emphasize the importance of an adequate equilibrium period 4-5 h in order to measure the dilution of administered tracer. The measurements by H2[18O] tracer are similar to those previously reported with deuterium oxide. Insofar as total body water reflects lean body mass, the data suggest that body compositional changes in normal and diabetic pregnancy are qualitatively similar.