Information gathered prospectively by Treloar et al from white, college-educated American women purportedly showed that the length of menstrual cycles dropped slightly between age 20 and early perimenopause and, in the last 4 years before menopause, rose to 57 days. It is possible that these findings were biased by incorrectly identified dates of menopause; classifying mean values by calendar year; and/or excluding cycles that straddled two calendar years.
The present investigators conducted a secondary analysis of prospectively collected data in 120 women from the original Treloar study who were taking part in the Tremin Research Program on Women’s Health, an ongoing study of health and menstrual cycles covering the reproductive lifespan. Menopause was defined as the first day of the final menstrual period (FMP), calculating full 1-year intervals working back from the FMP. Straddling and nonstraddling cycles were assigned to new 12-month intervals based on the date when each cycle ended. Mean cycle lengths then were recalculated for each of the 10 years preceding menopause.
Mean estimates of cycle length for each of the 4 years preceding menopause were 30.5, 35.0, 45.15, and 80.2 days, respectively. The corresponding figures from the original analysis were 33.6, 43.9, 55.9, and 54.6 days. The original and revised data differ significantly only for the year just before menopause. In this year, a majority of women are spending 75% or more of their time in cycles lasting longer than 40 days. The corrected version of the data demonstrates a greater difference between the second-to-last year and the last year before menopause.
The proportion of a year spent in menstrual cycles exceeding 40 days may help to determine the time until menopause begins. This often is a critical factor when choosing between hormone therapy and hysterectomy for women with perimenopausal menorrhagia. This figure may also help to predict such postmenopausal states as an increased risk of osteoporosis.