Abstract
OBJECTIVE: To examine the conventional wisdom that declines in antral follicle count accelerate after approximately 37 years of age, which has influenced clinical decision making. We critically examine evidence for this sudden acceleration. DESIGN: Cross-sectional study. SETTING: Academic setting. PATIENT(S): Two hundred fifty-two white women aged 25 to 45 years with regular cycles, a community-based study. INTERVENTION(S): Antral follicle counts were measured by transvaginal ultrasound in early follicular phase. MAIN OUTCOME MEASURE(S): Determination of the rate of antral follicle count decline with age. RESULT(S): The relationship of antral follicle count with age was determined by comparing a linear model with four nonlinear models (biphasic, quadratic, spline, and power). The linear model estimates the follicle decline as 0.97 follicles per year. The biphasic model had estimates of 0.76 follicles per year before 39.6 years of age and 1.92 follicles per year thereafter. At age 35 years the linear, quadratic, broken line, spline, and power models gave estimated declines (respectively) of 0.97, 0.94, 0.76, 0.17, and 0.90 follicles per year. At age 43 years the values were 0.97, 1.51, 1.92, 1.78, and 1.64 follicles per year. The quality of model fit was comparable for all models. CONCLUSION(S): Antral follicle count decline with age in a white population is best described as a gradual acceleration in decline with age. Therefore antral follicle count alone should not be used to determine aggressive treatment because of fear of rapid loss of follicles.