Medical therapy for women in the perimenopausal period is controversial, in part due to varying degrees of ovarian hormone secretion characteristic of this time of life. To extend our understanding of the reproductive endocrine milieu of perimenopausal women, we studied 6 cycling women, aged 47 yr and older, for 6 months with daily collections of first morning voided urine. Five additional older reproductive aged (43-47 yr old) women were studied with daily urine and serum sampling for a single menstrual cycle; their urinary hormone data were combined with the former group for menstrual cycle comparisons. Urine was assayed for LH, FSH, estrone conjugates, and pregnanediol glucuronide and normalized for creatinine (Cr). Eleven midreproductive aged (19-38 yr old) normally cycling women, 5 women with well defined premature ovarian failure, and 5 women aged 54 yr and older who were at least 1 yr postmenopausal were used for comparison. Perimenopausal women had shorter follicular phases (11 +/- 2 days vs. 14 +/- 1 days; P = 0.031) and, hence, shorter menstrual cycles than midreproductive aged controls. FSH excretion in perimenopausal women was greater than that in younger women (range of means, 4-32 vs 3-7 IU/g Cr; P = 0.0005). LH secretion was overall greater than that in younger normal subjects (range of means, 1.4-6.8 vs. 1.1-4.2 IU/g Cr; P < 0.026). Overall mean estrone conjugate excretion was greater in the perimenopausal women compared to that in the younger women [76.9 ng/mg Cr (range, 13.1-135) vs. 40.7 ng/mg Cr (range, 22.8-60.3); P = 0.023] and was similarly elevated in both follicular and luteal phases. Luteal phase pregnanediol excretion was diminished in the perimenopausal women compared to that in younger normal subjects (range for integrated pregnanediol, 1.0-8.4 vs. 1.6-12.7 microg/mg Cr/luteal phase; P = 0.015). Compared to postmenopausal women, perimenopausal women had more overall estrone excretion (2.5-6.2 ng/mg Cr in postmenopausal women; P = 0.02) and lower mean FSH (range of means for postmenopause, 24-85 IU/g Cr; P = 0.017) and LH (range for postmenopause, 4.3-14.8 IU/g Cr; P = 0.041). Compared to women with premature menopause, perimenopausal women again had lower FSH (range of means for premature menopause, 36-82 IU/g Cr; P = 0.0022), lower LH (range of means for premature menopause, 5.5-23.8 IU/g Cr; P = 0.0092), borderline higher mean estrone conjugates (range of means for premature menopause, 4-44 ng/mg Cr; P = 0.064), and far longer periods of ovarian activity (one to two cycles in prematurely menopausal women vs. three to six cycles in perimenopausal women). We conclude that altered ovarian function in the perimenopause can be observed as early as age 43 yr and include hyperestrogenism, hypergonadotropism, and decreased luteal phase progesterone excretion. These hormonal alterations may well be responsible for the increased gynecological morbidity that characterizes this period of life.