Results
Characteristics of All Newly Postmenopausal Women Before and After the WHI
Overall, 1,229 cohort members reached menopause in the pre-WHI period, and 562 cohort members reached menopause in the post-WHI period. The latter had higher educational levels, reported lower household incomes, and were more likely to be smokers, to do physical exercise, to have two or more children, and to have used oral contraception than those who reached menopause in the pre-WHI period ( Table 1 ). They were also more likely to be retired and were significantly younger at retirement than those who reached menopause in the pre-WHI period (52.6 [5.2] vs 55.1 [3.3] y, P < 0.001). A comparison of the health characteristics of all postmenopausal women showed that those who reached menopause in the post-WHI period reported better self-rated health, were more likely to report simple hot flushes, and were more likely to experience more than two psychological symptoms than those who reached menopause in the pre-WHI period. No significant differences were observed for troublesome hot flushes, vaginal dryness, number of general symptoms, or number of osteoarticular symptoms between the two periods ( Table 1 ).
Characteristics of New HT Users Before and After the WHI
In the pre-WHI period, 652 newly postmenopausal women (53.1%) began HT less than 1 year after reaching menopause, and 577 did not. Among the 562 women who reached menopause in the post-WHI period, only 72 were new HT users (12.8%). A comparison of health characteristics assessed before menopause for women starting HT during the two periods ( Table 2 ) showed essentially the same differences as those previously observed for all newly postmenopausal women between the two periods, except for troublesome hot flushes and psychological symptoms. Specifically, whereas the frequency of troublesome hot flushes did not differ significantly between the pre-WHI and post-WHI periods among all postmenopausal women, troublesome hot flushes were significantly more frequent in women who started HT in the post-WHI period than in women who started HT in the pre-WHI period (12.9% vs 5.8%, P = 0.02). Psychological symptoms were more frequent in the post-WHI period than in the pre-WHI period, and that difference was more pronounced among new HT users, of whom 41.7% experienced more than two psychological symptoms before menopause in the post-WHI period versus 27.7% in the pre-WHI period (P = 0.01).
Multivariable Comparison of Trends in the Characteristics of Women Who Did and Did Not Start HT
We conducted multivariable analyses of health characteristics with significantly different distributions for pre-WHI and post-WHI new HT users (ie, simple and troublesome hot flushes, self-rated health, and psychological symptoms).
After we took into account the characteristics associated with simple hot flushes before menopause, we found these symptoms to be more frequent in the post-WHI period than in the pre-WHI period, both among new HT users (odds ratio [OR], 1.9; 95% CI, 1.2-3.3) and the other women (OR, 1.5; 95% CI, 1.2-2.0; interaction test between HT initiation and period, P = 0.51; Table 3 ). Results of the multivariable analysis of troublesome hot flushes confirmed the differences previously seen among new HT users, who were more likely to report these symptoms before menopause in the post-WHI period than in the pre-WHI period (OR, 2.3; 95% CI, 1.0-5.4), whereas no difference was observed among HT nonusers (OR, 0.9; 95% CI, 0.5-1.6; interaction test between HT initiation and period, P = 0.04; Table 3 ).
Multivariable analysis confirmed that new HT users were less likely to report poor self-rated health before menopause in the post-WHI period than in the pre-WHI period (OR, 0.4; 95% CI, 0.2-1.0). The same result was true for the other women (OR, 0.6; 95% CI, 0.4-0.8; Table 4 ), and the ORs did not differ significantly (interaction test, P = 0.45). Multivariable analysis of psychological symptoms showed that new HT users were more likely to report more than two symptoms before menopause in the post-WHI period than in the pre-WHI period (OR, 1.8; 95% CI, 1.0-3.0), whereas the difference was less marked among HT nonusers (OR, 1.2; 95% CI, 0.9-1.7). However, the interaction test between HT initiation and period did not reach statistical significance (P=0.27; Table 4 ).
Reasons for HT Use
Reasons for HT use differed in the two periods: in the post-WHI period, a higher proportion of women reported using HT because of hot flushes (68.1% vs 52.9%, P = 0.02) and sleep disorders (11.1% vs 2.8%, P = 0.001), and a clearly lower proportion of women reported using HT for general prevention (5.6% vs 24.1%, P = 0.001) and osteoporosis prevention (12.5% vs 43.9%, P < 0.001; Table 5 ). Before the WHI publication, nearly half of women starting HT (47%) did not report hot flushes as a reason for using HTVa proportion that dropped to less than one third (32%) afterward ( Table 5 ). HT use was a personal choice for 28% of new HT users before 2002 and for 30% of new HT users after 2002, and was the doctor's choice for 57% of new HT users before 2002 versus 53% of new HT users after 2002 (proportions do not differ significantly between the two periods; data not shown).