OBJECTIVE The transport of immunoglobulin G and its subclasses 1 to 4 was investigated in the in vitro-perfused isolated cotyledon of the human placenta. METHODS An in vitro system with separate perfusion of the villous capillary system (fetal compartment) and the corresponding intervillous space (maternal compartment) was set up in an isolated cotyledon of human term placenta. After a 2-hour control phase with both compartments perfused in a closed circuit with NCTC-135 tissue culture medium together with Earl's balanced salt solution (2:1), media were exchanged in both circuits and for the experimental phase immunoglobulin G (Sandoglobulin) together with carbon 14-labeled bovine serum albumin (5-10 microCi) was added to the maternal compartment at a concentration of 6 gm/L. During the experimental phase, lasting between 2 and 5 hours, samples were taken from the maternal and fetal compartments every 30 minutes up to 2 hours and every 60 minutes thereafter. RESULTS During the control phase immunoglobulin G appeared in the maternal perfusate and reached a plateau at 60 to 80 mg/L, whereas the concentration in the fetal perfusate did not exceed 20 mg/L. A similar pattern of release was observed for hemoglobin, suggesting a washout of remains of blood from the intervillous space and the villous vascular compartment. After addition of immunoglobulin G to the maternal circuit during the first 2 hours in three of four experiments, no change in immunoglobulin G concentration was seen in the fetal circuit, and only in the fourth and fifth hours did the fetal concentration increase to 0.6% of the maternal concentration. In contrast, carbon 14-labeled bovine serum albumin was already detectable in the fetal circuit after 1 hour, but the level remained constant at 0.1% of the maternal concentration. Total immunoglobulin G transfer was estimated at 0.5% of the amount added to the maternal circulation, which was five times higher than total transfer of bovine serum albumin. Transfer was shown for all four subclasses. At the end of the experiment the ratio of immunoglobulin G1 to immunoglobulin G2 in the fetal perfusate was significantly higher than in the maternal perfusate (3.8 vs 1.8), suggesting preferential transfer of immunoglobulin G1. CONCLUSIONS Transfer of all four immunoglobulin G subclasses of a commercially available immunoglobulin G preparation across the human placenta from the maternal to the fetal side was demonstrated by the dual in vitro perfusion system. There is a preferential transfer for immunoglobulin G1.