A novel statistical approach to analysis of bleeding patterns during continuous hormone replacement therapy.

Previous studies have shown that continuous oral oestrogen-progestogen therapy, which is a relatively new treatment regimen, does not induce endometrial hyperstimulation. However, bleeding disturbances are common during the early months of therapy and in the present study we used both conventional methods of statistical analysis and a stochastic model to describe the bleeding patterns. Four groups of 15 post-menopausal women were given different oral formulations continuously for 1 yr. The oestrogen component in all cases was 2 mg 17 beta-estradiol. The progestogen used was norethisterone acetate at a dose of either 1 mg or 0.5 mg (Groups A and B) or megestrol acetate at a dose of either 5 mg and 2.5 mg (Groups C and D). Each woman kept a daily record of all bleeding episodes, 80% of which occurred during the first 4 mth of therapy. Analysis of variance showed that the high doses of the two progestogens were associated with less spotting and menstrual-like bleeding than the low doses. Stochastic analysis of the bleeding data confirmed that women on the high progestogen doses experienced fewer bleeding episodes than those on the low doses. It also showed that women receiving the high progestogen doses who weighed under 67 kg or had had their last menstruation over 5 yr previously bled less than the other women in the high-dose groups.