OXFORD, England, April 19 -- Use of hormone replacement therapy for at least five years added up to 1,300 additional ovarian cancers over 14 years in Britain and 1,000 extra deaths, researchers here reported.
Current HRT users, treated for at least five years, were on average 23% more likely to die of ovarian cancer than those who had never used HRT, according to results of the United Kingdom Million Women Study reported online in the The Lancet.
For every 1,000 HRT users, 1.6 women died compared with 1.3 non-users. This amounted to one extra ovarian cancer death in roughly 3,300 users, said Valerie Beral, MBBS, of the Cancer Research United Kingdom Epidemiology Unit here, and colleagues.
The researchers also found that after women stopped taking HRT, their risk returned to that of never-users. In addition, no residual risk was seen in women who had taken HRT at some time in the past or for fewer than five years.
The effect of HRT on ovarian cancer, Dr. Beral said, should not be viewed in isolation especially because HRT also affects the risk of breast and endometrial cancer. In total, these three cancers account for 39% of all cancers in women in Britain.
"The total incidence of these three cancers in the [Million Women] study population is 63% higher in current users of HRT than in never-users (31 versus 19 per 1,000 over five years)," the researchers wrote.
Thus, they said, when ovarian, endometrial, and breast cancers are taken together, use of HRT results in a material increase in these common cancers.
Ovarian cancer, Dr. Beral said, is the fourth most common cancer in the United Kingdom with some 6,700 new cases every year. However, published results of the relationship between HRT and the risk of ovarian cancer have been inconclusive, she said.
To investigate the HRT risk in a large cohort study, the researchers studied 948,576 postmenopausal women in the Million Women Study, half of whom took HRT.
The women did not have previous cancer or bilateral oophorectomy and were followed for an average of 5.3 years for incident ovarian cancer and 6.9 years for death.
Relative risks for ovarian cancer were calculated, stratified by age and hysterectomy status, and adjusted for a variety of confounders, including, age, socioeconomic status, BMI, time since menopause, alcohol consumption, and use of oral contraceptives.
When they last reported HRT use, 287,143 women (30%) were current users and 186,751 (20%) were past users. During follow-up, 2,273 incident ovarian cancers and 1,591 deaths were recorded.
The findings were:
- Current users were significantly more likely to develop ovarian cancer (relative risk 1.20 [95% CI 1.09-1.32, P=0.0002]) and to die (relative risk 1.23 [1.09-1.38, P=0.0006]).
- For current HRT, the incidence of ovarian cancer increased with increasing duration of use, but did not differ significantly by type of preparation used, its components, or mode of administration.
- Risks associated with HRT varied significantly according to tumor histology (P<0.0001). In women with epithelial tumors, the relative risk for current versus never-users was greater for serous than for mucinous, endometroid, or clear cell tumors.
- Past users of HRT were not at an increased risk (0.98 [0.88-1.11]) for incident disease and (0.97 [0.84-1.11]) for death.
- Over five years, the standardized incidence rates for ovarian cancer in current and never-users was 2.6 (2.4-2.9) and 2.2 (2.1-2.3) per 1,000, respectively. These numbers amounted to one extra ovarian cancer in about 2,500 users.
- Over five years, death rates in current and never-users were 1.6 (1.4-1.8) and 1.3 (1.2-1.4) per 1,000 respectively. These numbers amounted to one extra ovarian cancer death in roughly 3,300 users.
These results, Dr. Beral wrote, are broadly consistent with previous reports. However, a finding that the excess risk varied significantly according to tumor histology is difficult to interpret at present, she said. The international pooling of data now planned should help clarify the role of HRT in tumors of different histology.
Stratification for hysterectomy as well as by age in these analyses ensured that no direct comparisons were made between women who had had a hysterectomy and those who did not, the investigators said
Mentioning a study limitation, the researchers said that because the ovarian cancers in these analyses were diagnosed on average 2.4 years after use of HRT was last reported, it is possible that some women's use at the time of diagnosis could have been misclassified. However, they added, it is unlikely that over 2.4 years, users and never-users would have changed categories.
The method of ovarian carcinogenesis is not well understood, the investigators wrote, suggesting that further study of the variation in hormone-associated risks across histological types of ovarian cancer might help clarify the mechanisms.
In an accompanying comment, Steven A. Narod, M.D., of the University of Toronto, said a relative risk of 1.2 might be thought of as small, but enormous numbers of women have been exposed. In the Million Women Study alone, half a million women had taken HRT.
However, he said, because current use was found to be the main risk factor, the number of new cases attributable to HRT should be a function of the number of women taking the drug at any given time.
Use of hormone replacement, Dr. Narod said, has declined greatly in the UK as well as elsewhere since the 2003 report of the Women's Health Initiative, and is thought to be responsible for a recent reduction in breast cancer rates in the U.S. (The ovarian cancer incidence has decreased by about 0.7% per year since 1985, according to The American Cancer Society.)
"With these new data on ovarian cancer, we expect the use of HRT to fall further. We hope that the number of women dying of ovarian cancer will decline as well," Dr. Narod concluded.