The main side effect associated with the use of Norplant contraceptive implants is a disruption of the menstrual bleeding pattern. To explore the relationship between bleeding and hormonal changes, we analyzed the estradiol (E2) and progesterone (P) patterns that preceded bleeding episodes or that corresponded to periods of amenorrhea in 103 cycles observed among 82 women using Norplant subdermal implants. Five different bleeding patterns were defined: 'normal' (24-45 day cycles), oligomenorrhea (46-90 day cycles), amenorrhea (over 90 day cycles), irregular/frequent bleeding (less than 25 day cycles), and prolonged bleeding (continuous bleeding/spotting for more than 10 days). All 'normal' cycles were associated with a rise followed by a fall in E2 levels preceding bleeding. In half of the 'normal' cycles (28/54), a rise and fall of P was also observed. The same pattern was found in oligomenorrheic cycles, but only two of 12 cycles had a rise and fall of both E2 and P. None of the subjects with amenorrhea had luteal activity. Six of the nine amenorrheic cycles displayed persistently low E2 levels (below 75 pg/ml). The remaining three had a moderate elevation in E2 levels during the sampling period. Sixty percent of the subjects who showed irregular/frequent bleeding (15/25) had low E2 levels (less than 75 pg/ml), without luteal activity, and bleeding occurred without clear evidence of a further drop in E2 levels. In the remaining 40%, bleeding was preceded by a rise and drop of E2 without luteal activity, with the exception of one women, who exhibited a rise and fall of both E2 and P. Samples were obtained in only three subjects during continuous bleeding. One had low E2 levels, and the remaining two bled continuously, in spite of having E2 levels in the normal range. We conclude that ovarian hormones continue to influence endometrial shedding during the use of Norplant contraceptive implants.785-94