Antral follicle count (AFC) predicts natural menopause in a population based cohort: the coronary artery risk development in young adults (CARDIA) and cardia women's study (CWS)

Abstract

Tools to predict future fertility are needed for contraceptive decision-making as large numbers of women are approaching perimenopause. We evaluated the ability of AFC and other ovarian reserve markers to predict natural menopause (NatMen) in US women ages 34-49. CARDIA is a multicenter (Chicago, IL; Birmingham, AL; Minneapolis, MN; Oakland, CA) observational cohort study begun in 1985-86. In 2002-04, CWS evaluated 1163 women with ≥1 ovary with transvaginal ultrasound and FSH levels. The exam was targeted to the follicular phase unless irregular cycles or OCP use was present. Subjects were re-examined in 2005-06 and followed-up annually by phone through 2009-10. AFC (total # of 2-10mm follicles on 2 visualized ovaries) was assessed in 498 CWS subjects with follow-up data and no history of ovarian/uterine surgeries or menopause. We used C-statistics to determine optimal age, FSH, and AFC cutpoints for NatMen. Cox regression was used to predict NatMen hazard over 7 years' follow up adjusting for race, smoking, OCP use and menstrual history.Tabled 1Risk for NatMen over 7 Years of Follow-upVariableHazard Ratio95% CIAFC £4 vs. >41.72(1.10-2.68)Age >45 vs. £ 454.91(3.13-7.69)Black vs. White0.63(0.42-0.95)Current Smoker vs. Not1.68(1.09-2.59)FSH >13 vs. £132.39(1.56-3.65)Regular vs. Irregular Menses0.59(0.38-0.85)OCP use P>0.05 Open table in a new tab OCP use P>0.05 Median(Q1-Q3) AFC and FSH were 5(2-9) and 8(6-11) at a mean age of 42 (range 34-49). 107 women reported NatMen by 2009-10. Cutpoints at Age>45, FSH>13, and AFC≤4 yielded the highest c-statistic for NatMen with a PPV = 63% (29/47) and a NPV = 83% (373/451) through 2010. In Cox models, age, race, AFC, and FSH were independent predictors of NatMen. AFC is an independent predictor of natural menopause over 7 years of follow-up after controlling for the more established markers of age, race, and FSH in our community-based sample.