A microtiter plate enzyme immunoassay (EIA) is reported for the measur
ement of levonorgestrel (LNG) in serum and urine samples of human and
non-human primates, and the results are compared to data obtained by r
adioimmunoassay (RIA). Rabbit polyclonal antibodies were raised agains
t the bovine serum albumin conjugate of the 3-O-carboxymethyl oxime (C
MO) derivative of LNG. The enzyme label was produced by the conjugatio
n of horseradish peroxidase to LNG at the 3-position by the same CMO b
ridge used for the immunogen. The assay requires 2.5 hours to perform
using 2,2-azino-di-(3-ethylbenzthiazoline sulfonic acid) diammonium sa
lt as the chromogenic substrate. Serum (100 mu l) is extracted with pe
troleum ether prior to assay, whereas urine samples (25 mu l) are dilu
ted and measured directly. The sensitivity of the assay is 0.25 pg/wel
l with a 50% displacement of label at 7.5-9.5 pg and a linear response
through 250 pg/well. Minimum levels of 8.7 and 10.0 pg/ml can be dete
cted in serum and urine samples, respectively. Changes in serum LNG co
ncentrations were measured in women and non-human primates following L
NG implantation or injection. In the non-human primate study, serum LN
G concentrations began to rise rapidly following i.m. injection of LNG
, with peak levels occurring on days 3 to 5, then decreasing to approx
imately 25-35% of peak levels for the duration of the study. Circulati
ng concentrations of 1.86 +/- 0.18 ng/ml LNG were reached in women the
first week post-insertion of Norplant implants and decreased by 50% a
t 7-10 days, 75% after 14-21 days, followed by a steady decrease durin
g the next 60-70 days to constant low levels that exhibited a high ind
ividual variation. Correlation coefficients of EIA and RIA results wer
e 0.988 for human serum, 0.926 for human urine, and 0.972 for non-huma
n primate serum.