The transfer of levonorgestrel (LNG) from the maternal plasma via breast milk to the infant was studied in 38 fully lactating and breast-feeding women at 4-6 weeks postpartum, for a duration of 28 days. These volunteers were provided with LNG contraceptive treatment delivered through three, different routes of drug delivery system: (i) intrauterine devices impregnated with LNG (LNG-IUD); (ii) subdermal implant (Norplant (R)-2); and (iii) minipills (LNG 30 micrograms daily). On the first day after either the LNG-IUD (n = 14 women) or Norplant (R)-2 (n = 14 women) insertion, the maternal blood and breast milk samples were collected at 2, 4 and 8 hourly intervals. This was followed by daily collection of these samples as well as infant's blood from days 2 to 4 and thereafter on days 7, 14 and 28. For infant's blood samples from LNG minipill users (n = 10 women), only a single 4-hour sample was collected on the first day and no samples were collected on days 3 and 4. The rest of the schedule for collection of maternal blood and breast milk as well as infant's blood samples were the same in minipill users as for the other two treatment groups. The study revealed a lower LNG percentage transfer from maternal sera to breast milk--11.8 +/- 2, 7 +/- 2 and 8 +/- 1 and relatively higher percentage LNG transfer from breast milk to infant's sera--75 +/- 17, 68 +/- 20 and 32 +/- 3, in LNG-IUD, Norplant (R)-2 and minipill users, respectively. Therefore, LNG contraceptive steroid is transferred into the infant's circulation, the biological significance of which remains to be established.