Abstract
Background: In separate Women's Health Initiative randomized trials, combined hormone therapy with estrogen plus progestin reduced colorectal cancer incidence but estrogen alone in women with hysterectomy did not. We now analyze features of the colorectal cancers that developed and examine the survival of women following colorectal cancer diagnosis in the latter trial.
Participants and Methods: 10,739 postmenopausal women who were 50 to 79 years of age and had undergone hysterectomy were randomized to conjugated equine estrogens (0.625 mg/d) or matching placebo. Colorectal cancer incidence was a component of the monitoring global index of the study but was not a primary study endpoint. Colorectal cancers were verified by central medical record and pathology report review. Bowel exam frequency was not protocol defined, but information on their use was collected.
Results: After a median 7.1 years, there were 58 invasive colorectal cancers in the hormone group and 53 in the placebo group [hazard ratio, 1.12; 95% confidence interval (95% CI), 0.77-1.63]. Tumor size, stage, and grade were comparable in the two randomization groups. Bowel exam frequency was also comparable in the two groups. The cumulative mortality following colorectal cancer diagnosis among women in the conjugated equine estrogen group was 34% compared with 30% in the placebo group (hazard ratio, 1.34; 95% CI, 0.58-3.19).
Conclusions: In contrast to the preponderance of observational studies, conjugated equine estrogens in a randomized clinical trial did not reduce colorectal cancer incidence nor improve survival after diagnosis. (Cancer Epidemiol Biomarkers Prev 2008;17(10):2609–18)
Introduction
Colorectal cancer is the second leading cause of cancer death in the United States (1). Interventions including calcium (2), celecoxib (3), aspirin (4), sulindac (5, 6), and the bile acid ursodiol (7) have all reduced colorectal polyp recurrence in various high-risk populations. However, only combined estrogen plus progestin use has been shown to reduce colorectal cancer incidence in a randomized trial setting (8, 9).
In prospective cohort studies, which have generally not distinguished combined estrogen plus progestin from estrogen alone use, postmenopausal hormone therapy has been consistently associated with reduced colorectal cancer risk (10-14). In addition, both significantly improved colorectal cancer–specific survival (15), and overall survival (15, 16) has been seen in women using postmenopausal hormone therapy when their colorectal cancer was diagnosed. These observational study findings received some support from the Women's Health Initiative (WHI) randomized trial evaluating combined hormone therapy with conjugated equine estrogen plus medroxyprogesterone acetate, wherein a statistically significant 44% decrease in colorectal cancer incidence in the combined hormone therapy compared with placebo groups was seen (9).
In the original report of the WHI randomized placebo-controlled trial, conjugated equine estrogen alone did not reduce colorectal cancer incidence compared with placebo (17). The current report provides details of the colorectal cancer screening procedures, characteristics of the diagnosed colorectal cancers, additional subgroup analyses, and additional follow-up information, including survival following colorectal cancer diagnosis in the WHI randomized trial comparing conjugated equine estrogen alone to placebo.
Materials and Methods
Study Design
The WHI trial of conjugated equine estrogen alone enrolled 10,739 women with previous hysterectomy at 40 clinical centers in the United States between 1993 and 1998 under protocols approved by the institutional review boards at each institution (18). Recruitment was predominantly done through mass mailings. Eligibility criteria included ages between 50 to 79 years, postmenopausal status, and providing written informed consent. Women were eligible if they had undergone a previous hysterectomy, with or without oophorectomy. Exclusion criteria included a personal history of breast cancer, medical conditions likely to result in death in the next three years, or colorectal cancer within the past 10 years. Women currently using postmenopausal hormones were eligible after a 3-month washout period.
The trial was a randomized, double-blind, placebo-controlled trial of conjugated equine estrogen (0.625 mg/d) administered in a single tablet (Premarin, Wyeth Ayerst) compared with an identical-seeming placebo. Randomization was determined by the WHI Clinical Coordinating Center and implemented at local centers with a distributed database that maintained blinding of participants and staff to randomization allocation throughout the trial.
Women in the hormone trial also could participate on study entry in the dietary modification trial or could enter a randomized trial of calcium and vitamin D at the first or second annual clinic visit. Approximately 20% and 60% joined the dietary modification and calcium and vitamin D trials, respectively.
Follow-up
Follow-up procedures have been described previously (18, 19). Participants were contacted after six weeks to assess symptoms and promote adherence, at six month intervals for clinical outcomes assessment, and annually for clinic visits. For potential colorectal cancer diagnoses, medical records and pathology reports were reviewed by locally trained physician adjudicators (blind to group assignment) to identify cases; cases underwent a second blind adjudication at the clinical coordinating center and tumor characteristics were coded using the Surveillance, Epidemiology, and End Results program guidelines (20). Proximal colon lesions included those in the cecum, ascending colon (right colon), hepatic flexure, and transverse colon. Distal colon lesions included those in the splenic flexure, descending colon (left colon), and sigmoid and rectosigmoid junction.
Colorectal cancer was not a WHI conjugated equine estrogen primary study outcome and colorectal cancer screening was not protocol defined; participants received bowel exams as determined by their personal health care provider. Self-administered questionnaires or structured telephone interviews were used every six months to monitor the frequencies of rectal exams, fecal occult-blood testings, sigmoidoscopies and colonoscopies (asked as one question), and barium enema exams. Participants were also asked semiannually if they had intestinal or colon polyps diagnosed during that period.
Termination of the Study
After a mean follow-up of 7.1 years, the National Heart, Lung, and Blood Institute stopped the trial early because of increased stroke risk, absence of coronary heart disease benefit, and the low probability that interpretation of the trial would change with further information. In the original report on conjugated equine estrogen trial outcomes, 119 total colorectal cancer cases were identified (17). The current report is based on 120 total colorectal cancers diagnosed before February 29, 2004, the date on which the participants were instructed to discontinue their study medication and follow-up mortality data through April 1, 2005, on those women diagnosed with colorectal cancer.
Statistical Analysis
Primary results were assessed with time-to-event methods based on the intention-to-treat principle. Comparisons of rates of colorectal cancer are presented as hazard ratios with 95% CI and Wald z-statistical P values from Cox regression models, stratified according to age and randomization in the WHI trials of dietary modification, calcium and vitamin D, or both. Because participants in the calcium and vitamin D trial entered that trial one to two years after their entry into the hormone trial, stratification for calcium and vitamin D trial participation was time dependent, based on the calcium and vitamin D randomization date. No adjustments are made for multiple analyses over time because the incidence of colorectal cancer did not influence the decision to stop the trial.
Kaplan-Meier plots were used to examine the rates of colorectal cancer over time. Effects of baseline participant characteristics, including risk factors for colorectal cancer, were assessed in Cox regression models as described above, expanded to include the designated risk factor and randomization assignment (as the main effects) and the interaction between them. P values for interaction were computed with likelihood-ratio tests, and models with and without the interaction term were compared. For variables measured on a continuous scale, P values were computed on the continuous variable. Women with missing risk factor values in a given analysis were excluded from that analysis. Eighteen subgroup comparisons were done; one would be expected to be significant at the 0.05 level by chance alone. The exploratory analyses significance levels were not adjusted for multiple comparisons. Comparisons of survival after colorectal cancer diagnosis between estrogen alone and placebo groups were based on a proportional hazards model that was stratified by age and cancer stage at diagnosis with age as a covariate. Sensitivity analyses were conducted to explore the impact of adherence on outcome, and follow-up was censored six months after participants first became nonadherent (defined as taking <80% of study pills).
Results
There were 5,310 women randomized to conjugated equine estrogen and 5,429 to placebo (total, 10,739) with follow-up during active intervention for an average of 7.1 years (maximum, 10.8 years). Recent outcomes data were available for 10,176 participants (94.8%). Of the women, 54% in the hormone group and 53.5% in the control group stopped study medications at some point; most did not restart. In addition, 5.7% in the hormone group and 9.1% in the control group reported using postmenopausal hormones off protocol. A full consort diagram is published elsewhere (8). The two groups were well balanced for colorectal cancer risk factors, including age, education, anthropometry, personal or family history of colorectal cancer, smoking history, use of anti-inflammatory medications, dietary patterns, use of nutritional supplements, levels of physical activity, previous postmenopausal hormone use, and previous colorectal cancer screening (Table 1).
Characteristics of participants in the WHI conjugated equine estrogen trial at baseline by group assignment
. | CEE . | Placebo . | ||
---|---|---|---|---|
. | n (%) . | n (%) . | ||
Characteristic | ||||
Age at screening | ||||
50-59 | 1,637 (30.8) | 1,673 (30.8) | ||
60-69 | 2,387 (45.0) | 2,465 (45.4) | ||
70-79 | 1,286 (24.2) | 1,291 (23.8) | ||
Ethnicity | ||||
White | 4,007 (75.5) | 4,075 (75.1) | ||
Black | 782 (14.7) | 835 (15.4) | ||
Hispanic | 322 (6.1) | 333 (6.1) | ||
American Indian | 41 (0.8) | 34 (0.6) | ||
Asian or Pacific Islander | 86 (1.6) | 78 (1.4) | ||
Unknown | 72 (1.4) | 74 (1.4) | ||
Education | ||||
0-8 y | 181 (3.4) | 148 (2.7) | ||
Some high school | 354 (6.7) | 370 (6.9) | ||
High school diploma or GED | 1,233 (23.5) | 1,188 (22.1) | ||
School after high school | 2,271 (43.2) | 2,350 (43.7) | ||
College degree or higher | 1,216 (23.1) | 1,327 (24.7) | ||
First-degree relative with colorectal cancer | ||||
None | 3,904 (85.3) | 4,017 (85.2) | ||
1 | 575 (12.6) | 591 (12.5) | ||
≥2 | 98 (2.1) | 109 (2.3) | ||
History of polyps removal | ||||
No | 4,323 (90.4) | 4,336 (90.3) | ||
Yes | 460 (9.6) | 466 (9.7) | ||
History of ulcerative colitis or Crohn disease | ||||
No | 5,145 (98.6) | 5,268 (98.8) | ||
Yes | 74 (1.4) | 64 (1.2) | ||
History of colorectal cancer | ||||
No | 5,233 (99.7) | 5,339 (99.6) | ||
Yes | 16 (0.3) | 24 (0.4) | ||
History of diabetes | ||||
No | 4,806 (90.5) | 4,906 (90.4) | ||
Yes | 502 (9.5) | 519 (9.6) | ||
BMI | ||||
<25 | 1,110 (21.0) | 1,096 (20.3) | ||
25 to <30 | 1,795 (34.0) | 1,912 (35.5) | ||
≥30 | 2,376 (45.0) | 2,383 (44.2) | ||
Waist circumference > 88 cm | ||||
No | 2,304 (43.5) | 2,342 (43.3) | ||
Yes | 2,994 (56.5) | 3,071 (56.7) | ||
Height > 161.3 cm | ||||
No | 2,633 (49.6) | 2,718 (50.1) | ||
Yes | 2,654 (50.0) | 2,688 (49.5) | ||
Physical activity (Metabolic equivalents per week) | ||||
None | 1,081 (22.2) | 1,043 (21.3) | ||
>0-3.75 | 1,089 (22.3) | 1,142 (23.3) | ||
>3.75-8.75 | 914 (18.7) | 895 (18.3) | ||
>8.75-17.5 | 968 (19.8) | 958 (19.5) | ||
>17.5 | 828 (17.0) | 866 (17.7) | ||
NSAID use | ||||
No | 3,601 (67.8) | 3,689 (68.0) | ||
Yes | 1,709 (32.2) | 1,740 (32.1) | ||
Ibuprofen use | ||||
No | 4,702 (88.5) | 4,798 (88.4) | ||
Yes | 608 (11.5) | 631 (11.6) | ||
Aspirin use (≥100 mg) | ||||
No | 4,413 (83.1) | 4,498 (82.9) | ||
Yes | 897 (16.9) | 931 (17.2) | ||
Acetaminophen use | ||||
No | 4,566 (86.0) | 4,654 (85.7) | ||
Yes | 744 (14.0) | 775 (14.3) | ||
Total Energy (kcal) | ||||
≤1,119 | 1,211 (23.9) | 1,197 (23.1) | ||
>1,119-1,414.5 | 980 (19.3) | 1,056 (20.4) | ||
>1,414.5-1,715 | 957 (18.9) | 939 (18.1) | ||
>1,715-2,129.5 | 907 (17.9) | 926 (17.9) | ||
>2,129.5 | 1,021 (20.1) | 1,056 (20.4) | ||
Percent energy from fat | ||||
≤27 | 867 (17.1) | 868 (16.8) | ||
>27-32.5 | 920 (18.1) | 964 (18.6) | ||
>32.5-37 | 1,069 (21.1) | 1,098 (21.2) | ||
>37-41.5 | 1,067 (21.0) | 1,050 (20.3) | ||
>41.5 | 1,153 (22.7) | 1,194 (23.1) | ||
Selenium (μg/d) | ||||
≤74.4 | 1,813 (35.7) | 1,872 (36.2) | ||
74.5-106.3 | 1,596 (31.4) | 1,638 (31.7) | ||
≥106.4 | 1,667 (32.8) | 1,664 (32.2) | ||
Red meat (servings/d) | ||||
0.000 to <0.364 | 1,278 (25.2) | 1,286 (24.9) | ||
0.364 to <0.649 | 1,272 (25.1) | 1,290 (25.0) | ||
0.649 to <1.070 | 1,262 (25.9) | 1,293 (25.0) | ||
≥1.070 | 1,260 (24.8) | 1,301 (25.2) | ||
Fruit and vegetables (servings/d) | ||||
0.0 to ≤3.0 | 2,214 (43.6) | 2,223 (43.0) | ||
3.0 to ≤6.0 | 2,166 (42.7) | 2,195 (42.4) | ||
6.0 to ≤9.0 | 606 (11.9) | 625 (12.1) | ||
>9.0 | 92 (1.8) | 132 (2.6) | ||
Alcohol intake | ||||
None | 1,995 (38.0) | 2,007 (37.2) | ||
<1 drink/wk | 1,768 (33.7) | 1,815 (33.7) | ||
≥1 drink/wk | 1,484 (28.3) | 1,566 (29.1) | ||
Multivitamin use | ||||
No | 3,618 (68.1) | 3,683 (67.8) | ||
Yes | 1,692 (31.9) | 1,746 (32.2) | ||
Calcium supplement use (mg/d) | ||||
None | 3,074 (57.9) | 3,041 (56.0) | ||
≤200 | 887 (16.7) | 918 (16.9) | ||
>200-662 | 745 (14.0) | 812 (15.0) | ||
>662 | 604 (11.4) | 658 (12.1) | ||
Vitamin D supplement use (μg/d) | ||||
None | 3,290 (62.0) | 3,328 (61.3) | ||
≤10 | 1,696 (31.9) | 1,770 (32.6) | ||
>10 | 324 (6.1) | 331 (6.1) | ||
Smoking | ||||
Never smoked | 2,723 (51.9) | 2,705 (50.4) | ||
Past smoker | 1,986 (37.8) | 2,089 (38.9) | ||
Current smoker | 542 (10.3) | 571 (10.6) | ||
Oral contraceptive use, duration | ||||
Non-user | 3,257 (61.4) | 3,377 (62.2) | ||
<5 y | 1,207 (22.8) | 1,198 (22.1) | ||
5 to <10 y | 472 (8.9) | 441 (8.1) | ||
≥10 y | 369 (7.0) | 411 (7.6) | ||
Previous menopausal hormone use | ||||
Non-user | 2,769 (52.1) | 2,770 (51.0) | ||
<5 y | 1,352 (25.5) | 1,412 (26.0) | ||
5 to <10 y | 469 (8.8) | 515 (9.5) | ||
≥10 y | 720 (13.6) | 732 (13.5) | ||
Previous E-alone use | ||||
Non-user | 2,872 (54.1) | 2,891 (53.3) | ||
<5 y | 1,317 (24.8) | 1,368 (25.2) | ||
5 to <10 y | 448 (8.4) | 494 (9.1) | ||
≥10 y | 673 (12.7) | 676 (12.5) | ||
Previous E+progestin use | ||||
Non-user | 5,093 (95.9) | 5,178 (95.4) | ||
<5 y | 144 (2.7) | 158 (2.9) | ||
5 to <10 y | 42 (0.8) | 56 (1.0) | ||
≥10 y | 31 (0.6) | 37 (0.7) | ||
Recency of HRT use | ||||
Non-user | 2,769 (52.2) | 2,770 (51.1) | ||
Past, <5 y | 566 (10.7) | 586 (10.8) | ||
Past, 5-10 y | 257 (4.8) | 271 (5.0) | ||
Past, >10 y | 1,048 (19.7) | 1,091 (20.1) | ||
Current | 669 (12.6) | 708 (13.0) | ||
Bilateral oophorectomy | ||||
No | 2,973 (60.5) | 2,917 (58.0) | ||
Yes | 1,938 (39.5) | 2,111 (42.0) | ||
Previous colonoscopy, sigmoidoscopy, or flex sig | ||||
None | 2,606 (54.0) | 2,609 (53.7) | ||
<5 y ago | 1,216 (25.2) | 1,228 (25.3) | ||
≥5 y ago | 1,008 (20.9) | 1,024 (21.1) |
. | CEE . | Placebo . | ||
---|---|---|---|---|
. | n (%) . | n (%) . | ||
Characteristic | ||||
Age at screening | ||||
50-59 | 1,637 (30.8) | 1,673 (30.8) | ||
60-69 | 2,387 (45.0) | 2,465 (45.4) | ||
70-79 | 1,286 (24.2) | 1,291 (23.8) | ||
Ethnicity | ||||
White | 4,007 (75.5) | 4,075 (75.1) | ||
Black | 782 (14.7) | 835 (15.4) | ||
Hispanic | 322 (6.1) | 333 (6.1) | ||
American Indian | 41 (0.8) | 34 (0.6) | ||
Asian or Pacific Islander | 86 (1.6) | 78 (1.4) | ||
Unknown | 72 (1.4) | 74 (1.4) | ||
Education | ||||
0-8 y | 181 (3.4) | 148 (2.7) | ||
Some high school | 354 (6.7) | 370 (6.9) | ||
High school diploma or GED | 1,233 (23.5) | 1,188 (22.1) | ||
School after high school | 2,271 (43.2) | 2,350 (43.7) | ||
College degree or higher | 1,216 (23.1) | 1,327 (24.7) | ||
First-degree relative with colorectal cancer | ||||
None | 3,904 (85.3) | 4,017 (85.2) | ||
1 | 575 (12.6) | 591 (12.5) | ||
≥2 | 98 (2.1) | 109 (2.3) | ||
History of polyps removal | ||||
No | 4,323 (90.4) | 4,336 (90.3) | ||
Yes | 460 (9.6) | 466 (9.7) | ||
History of ulcerative colitis or Crohn disease | ||||
No | 5,145 (98.6) | 5,268 (98.8) | ||
Yes | 74 (1.4) | 64 (1.2) | ||
History of colorectal cancer | ||||
No | 5,233 (99.7) | 5,339 (99.6) | ||
Yes | 16 (0.3) | 24 (0.4) | ||
History of diabetes | ||||
No | 4,806 (90.5) | 4,906 (90.4) | ||
Yes | 502 (9.5) | 519 (9.6) | ||
BMI | ||||
<25 | 1,110 (21.0) | 1,096 (20.3) | ||
25 to <30 | 1,795 (34.0) | 1,912 (35.5) | ||
≥30 | 2,376 (45.0) | 2,383 (44.2) | ||
Waist circumference > 88 cm | ||||
No | 2,304 (43.5) | 2,342 (43.3) | ||
Yes | 2,994 (56.5) | 3,071 (56.7) | ||
Height > 161.3 cm | ||||
No | 2,633 (49.6) | 2,718 (50.1) | ||
Yes | 2,654 (50.0) | 2,688 (49.5) | ||
Physical activity (Metabolic equivalents per week) | ||||
None | 1,081 (22.2) | 1,043 (21.3) | ||
>0-3.75 | 1,089 (22.3) | 1,142 (23.3) | ||
>3.75-8.75 | 914 (18.7) | 895 (18.3) | ||
>8.75-17.5 | 968 (19.8) | 958 (19.5) | ||
>17.5 | 828 (17.0) | 866 (17.7) | ||
NSAID use | ||||
No | 3,601 (67.8) | 3,689 (68.0) | ||
Yes | 1,709 (32.2) | 1,740 (32.1) | ||
Ibuprofen use | ||||
No | 4,702 (88.5) | 4,798 (88.4) | ||
Yes | 608 (11.5) | 631 (11.6) | ||
Aspirin use (≥100 mg) | ||||
No | 4,413 (83.1) | 4,498 (82.9) | ||
Yes | 897 (16.9) | 931 (17.2) | ||
Acetaminophen use | ||||
No | 4,566 (86.0) | 4,654 (85.7) | ||
Yes | 744 (14.0) | 775 (14.3) | ||
Total Energy (kcal) | ||||
≤1,119 | 1,211 (23.9) | 1,197 (23.1) | ||
>1,119-1,414.5 | 980 (19.3) | 1,056 (20.4) | ||
>1,414.5-1,715 | 957 (18.9) | 939 (18.1) | ||
>1,715-2,129.5 | 907 (17.9) | 926 (17.9) | ||
>2,129.5 | 1,021 (20.1) | 1,056 (20.4) | ||
Percent energy from fat | ||||
≤27 | 867 (17.1) | 868 (16.8) | ||
>27-32.5 | 920 (18.1) | 964 (18.6) | ||
>32.5-37 | 1,069 (21.1) | 1,098 (21.2) | ||
>37-41.5 | 1,067 (21.0) | 1,050 (20.3) | ||
>41.5 | 1,153 (22.7) | 1,194 (23.1) | ||
Selenium (μg/d) | ||||
≤74.4 | 1,813 (35.7) | 1,872 (36.2) | ||
74.5-106.3 | 1,596 (31.4) | 1,638 (31.7) | ||
≥106.4 | 1,667 (32.8) | 1,664 (32.2) | ||
Red meat (servings/d) | ||||
0.000 to <0.364 | 1,278 (25.2) | 1,286 (24.9) | ||
0.364 to <0.649 | 1,272 (25.1) | 1,290 (25.0) | ||
0.649 to <1.070 | 1,262 (25.9) | 1,293 (25.0) | ||
≥1.070 | 1,260 (24.8) | 1,301 (25.2) | ||
Fruit and vegetables (servings/d) | ||||
0.0 to ≤3.0 | 2,214 (43.6) | 2,223 (43.0) | ||
3.0 to ≤6.0 | 2,166 (42.7) | 2,195 (42.4) | ||
6.0 to ≤9.0 | 606 (11.9) | 625 (12.1) | ||
>9.0 | 92 (1.8) | 132 (2.6) | ||
Alcohol intake | ||||
None | 1,995 (38.0) | 2,007 (37.2) | ||
<1 drink/wk | 1,768 (33.7) | 1,815 (33.7) | ||
≥1 drink/wk | 1,484 (28.3) | 1,566 (29.1) | ||
Multivitamin use | ||||
No | 3,618 (68.1) | 3,683 (67.8) | ||
Yes | 1,692 (31.9) | 1,746 (32.2) | ||
Calcium supplement use (mg/d) | ||||
None | 3,074 (57.9) | 3,041 (56.0) | ||
≤200 | 887 (16.7) | 918 (16.9) | ||
>200-662 | 745 (14.0) | 812 (15.0) | ||
>662 | 604 (11.4) | 658 (12.1) | ||
Vitamin D supplement use (μg/d) | ||||
None | 3,290 (62.0) | 3,328 (61.3) | ||
≤10 | 1,696 (31.9) | 1,770 (32.6) | ||
>10 | 324 (6.1) | 331 (6.1) | ||
Smoking | ||||
Never smoked | 2,723 (51.9) | 2,705 (50.4) | ||
Past smoker | 1,986 (37.8) | 2,089 (38.9) | ||
Current smoker | 542 (10.3) | 571 (10.6) | ||
Oral contraceptive use, duration | ||||
Non-user | 3,257 (61.4) | 3,377 (62.2) | ||
<5 y | 1,207 (22.8) | 1,198 (22.1) | ||
5 to <10 y | 472 (8.9) | 441 (8.1) | ||
≥10 y | 369 (7.0) | 411 (7.6) | ||
Previous menopausal hormone use | ||||
Non-user | 2,769 (52.1) | 2,770 (51.0) | ||
<5 y | 1,352 (25.5) | 1,412 (26.0) | ||
5 to <10 y | 469 (8.8) | 515 (9.5) | ||
≥10 y | 720 (13.6) | 732 (13.5) | ||
Previous E-alone use | ||||
Non-user | 2,872 (54.1) | 2,891 (53.3) | ||
<5 y | 1,317 (24.8) | 1,368 (25.2) | ||
5 to <10 y | 448 (8.4) | 494 (9.1) | ||
≥10 y | 673 (12.7) | 676 (12.5) | ||
Previous E+progestin use | ||||
Non-user | 5,093 (95.9) | 5,178 (95.4) | ||
<5 y | 144 (2.7) | 158 (2.9) | ||
5 to <10 y | 42 (0.8) | 56 (1.0) | ||
≥10 y | 31 (0.6) | 37 (0.7) | ||
Recency of HRT use | ||||
Non-user | 2,769 (52.2) | 2,770 (51.1) | ||
Past, <5 y | 566 (10.7) | 586 (10.8) | ||
Past, 5-10 y | 257 (4.8) | 271 (5.0) | ||
Past, >10 y | 1,048 (19.7) | 1,091 (20.1) | ||
Current | 669 (12.6) | 708 (13.0) | ||
Bilateral oophorectomy | ||||
No | 2,973 (60.5) | 2,917 (58.0) | ||
Yes | 1,938 (39.5) | 2,111 (42.0) | ||
Previous colonoscopy, sigmoidoscopy, or flex sig | ||||
None | 2,606 (54.0) | 2,609 (53.7) | ||
<5 y ago | 1,216 (25.2) | 1,228 (25.3) | ||
≥5 y ago | 1,008 (20.9) | 1,024 (21.1) |
NOTE: From χ2 test.
Abbreviations: CEE, conjugated equine estrogen; E, estrogen; NSAID, nonsteroidal anti-inflammatory drug; HRT, hormone replacement therapy.
At the termination of the study intervention, 53.8% of women had already stopped taking study medication and the discontinuation rate was closely comparable in the intervention and placebo groups. Nonprotocol hormone therapy was initiated by 9.1% in the placebo group and 5.7% in the conjugated equine estrogen group through year 6.
Among the 120 women with colorectal cancer, three in the intervention group and two in the placebo group had stage 0 (in situ) disease. There was one squamous cell carcinoma and one neuroendocrine carcinoma in the placebo group, and one malignant melanoma of the colon and one neuroendocrine carcinoma in the hormone group. All presented analyses are limited to the remaining 111 women with invasive colorectal carcinomas of epithelial origin.
By an intention-to-treat analysis, there was no difference between groups in the incidence of invasive colorectal cancer (58 in the conjugated equine estrogen group and 53 in the placebo group; hazard ratio, 1.12; 95% CI, 0.77-1.63; P = 0.55; Table 2). Although fewer rectal cancers were diagnosed in the conjugated equine estrogen group, a total of only 15 rectal cancers were observed, precluding reliable estimation of any conjugated equine estrogen effect. In addition, Kaplan-Meier plots by randomization group revealed no time trends of treatment effects (Fig. 1). In adherence-adjusted sensitivity analysis, the hazard ratio for invasive colorectal cancer changed only slightly to 1.02 (95% CI, 0.64-1.64). Tumor characteristics were also comparable between treatment groups (Table 2).
Incidence (annualized rate) and characteristics of colorectal cancer, according to hormone treatment group
Tumor characteristics . | CEE (n = 5,310) . | . | Placebo (n = 5,429) . | . | Hazard ratio . | 95% CI . | P . | |||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
. | n . | % (rate) . | n . | % (rate) . | . | . | . | |||||||
Invasive colorectal cancer | 58 | 0.15 | 53 | 0.14 | 1.12 | 0.77-1.63 | 0.55 | |||||||
Colon cancer | 53 | 0.14 | 43 | 0.11 | 1.26 | 0.84-1.88 | 0.26 | |||||||
Rectal cancer | 5 | 0.01 | 10 | 0.03 | 0.53 | 0.18-1.56 | 0.25 | |||||||
Histology | ||||||||||||||
Adenocarcinoma | 51 | 0.14% | 50 | 0.13% | 1.04 | 0.71-1.54 | 0.83 | |||||||
Mucinous and serous | 7 | 0.02% | 3 | 0.01% | 2.44 | 0.63-9.43 | 0.20 | |||||||
Tumor grade | ||||||||||||||
Missing | 2 | 0.01% | 2 | 0.01% | Not determined | |||||||||
Well differentiated | 10 | 0.03% | 4 | 0.01% | 2.53 | 0.79-8.07 | 0.12 | |||||||
Moderately differentiated | 36 | 0.10% | 37 | 0.10% | 1.00 | 0.63-1.58 | 0.99 | |||||||
Poorly differentiated or anaplastic | 10 | 0.03% | 10 | 0.03% | 1.03 | 0.43-2.49 | 0.94 | |||||||
Stage of disease | ||||||||||||||
Localized | 19 | 0.05% | 19 | 0.05% | 1.02 | 0.54-1.93 | 0.95 | |||||||
Regional | 34 | 0.09% | 24 | 0.06% | 1.46 | 0.87-2.47 | 0.15 | |||||||
Distant | 5 | 0.01% | 10 | 0.03% | 0.50 | 0.17-1.46 | 0.21 | |||||||
Lymph node involvement | ||||||||||||||
Missing | 7 | 0.02% | 7 | 0.02% | Not determined | |||||||||
No | 29 | 0.08% | 27 | 0.07% | 1.09 | 0.64-1.84 | 0.76 | |||||||
Yes | 22 | 0.06% | 19 | 0.05% | 1.20 | 0.65-2.21 | 0.57 | |||||||
Cancer location | ||||||||||||||
Proximal | 27 | 0.07% | 24 | 0.06% | 1.17 | 0.67-2.03 | 0.58 | |||||||
Distal | 26 | 0.07% | 18 | 0.05% | 1.45 | 0.79-2.64 | 0.23 | |||||||
Unknown | 0 | 0.00% | 1 | <.01% | Not determined | |||||||||
Rectum | 5 | 0.01% | 10 | 0.03% | 0.53 | 0.18-1.56 | 0.25 | |||||||
Tumor size (cm) | 38 | 4.1 ± 2.0 | 37 | 4.0 ± 2.4 | ||||||||||
Missing | 20 | 0.05% | 17 | 0.04% | Not determined | |||||||||
<3.0 | 8 | 0.02% | 11 | 0.03% | 0.78 | 0.31-1.94 | 0.59 | |||||||
3.0-3.9 | 6 | 0.02% | 10 | 0.03% | 0.62 | 0.22-1.70 | 0.35 | |||||||
4.0-5.9 | 15 | 0.04% | 7 | 0.02% | 2.15 | 0.88-5.27 | 0.10 | |||||||
≥6.0 | 9 | 0.02% | 8 | 0.02% | 1.18 | 0.46-3.07 | 0.73 | |||||||
Number + lymph nodes | 49 | 1.6 ± 3.3 | 44 | 0.9 ± 1.8 | ||||||||||
Missing | 10 | 0.03% | 9 | 0.02% | Not determined | |||||||||
None | 26 | 0.07% | 25 | 0.06% | 1.05 | 0.61-1.82 | 0.86 | |||||||
1-2 | 10 | 0.03% | 16 | 0.04% | 0.65 | 0.30-1.44 | 0.29 | |||||||
≥3 | 12 | 0.03% | 3 | 0.01% | 4.06 | 1.15-14.40 | 0.03 |
Tumor characteristics . | CEE (n = 5,310) . | . | Placebo (n = 5,429) . | . | Hazard ratio . | 95% CI . | P . | |||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
. | n . | % (rate) . | n . | % (rate) . | . | . | . | |||||||
Invasive colorectal cancer | 58 | 0.15 | 53 | 0.14 | 1.12 | 0.77-1.63 | 0.55 | |||||||
Colon cancer | 53 | 0.14 | 43 | 0.11 | 1.26 | 0.84-1.88 | 0.26 | |||||||
Rectal cancer | 5 | 0.01 | 10 | 0.03 | 0.53 | 0.18-1.56 | 0.25 | |||||||
Histology | ||||||||||||||
Adenocarcinoma | 51 | 0.14% | 50 | 0.13% | 1.04 | 0.71-1.54 | 0.83 | |||||||
Mucinous and serous | 7 | 0.02% | 3 | 0.01% | 2.44 | 0.63-9.43 | 0.20 | |||||||
Tumor grade | ||||||||||||||
Missing | 2 | 0.01% | 2 | 0.01% | Not determined | |||||||||
Well differentiated | 10 | 0.03% | 4 | 0.01% | 2.53 | 0.79-8.07 | 0.12 | |||||||
Moderately differentiated | 36 | 0.10% | 37 | 0.10% | 1.00 | 0.63-1.58 | 0.99 | |||||||
Poorly differentiated or anaplastic | 10 | 0.03% | 10 | 0.03% | 1.03 | 0.43-2.49 | 0.94 | |||||||
Stage of disease | ||||||||||||||
Localized | 19 | 0.05% | 19 | 0.05% | 1.02 | 0.54-1.93 | 0.95 | |||||||
Regional | 34 | 0.09% | 24 | 0.06% | 1.46 | 0.87-2.47 | 0.15 | |||||||
Distant | 5 | 0.01% | 10 | 0.03% | 0.50 | 0.17-1.46 | 0.21 | |||||||
Lymph node involvement | ||||||||||||||
Missing | 7 | 0.02% | 7 | 0.02% | Not determined | |||||||||
No | 29 | 0.08% | 27 | 0.07% | 1.09 | 0.64-1.84 | 0.76 | |||||||
Yes | 22 | 0.06% | 19 | 0.05% | 1.20 | 0.65-2.21 | 0.57 | |||||||
Cancer location | ||||||||||||||
Proximal | 27 | 0.07% | 24 | 0.06% | 1.17 | 0.67-2.03 | 0.58 | |||||||
Distal | 26 | 0.07% | 18 | 0.05% | 1.45 | 0.79-2.64 | 0.23 | |||||||
Unknown | 0 | 0.00% | 1 | <.01% | Not determined | |||||||||
Rectum | 5 | 0.01% | 10 | 0.03% | 0.53 | 0.18-1.56 | 0.25 | |||||||
Tumor size (cm) | 38 | 4.1 ± 2.0 | 37 | 4.0 ± 2.4 | ||||||||||
Missing | 20 | 0.05% | 17 | 0.04% | Not determined | |||||||||
<3.0 | 8 | 0.02% | 11 | 0.03% | 0.78 | 0.31-1.94 | 0.59 | |||||||
3.0-3.9 | 6 | 0.02% | 10 | 0.03% | 0.62 | 0.22-1.70 | 0.35 | |||||||
4.0-5.9 | 15 | 0.04% | 7 | 0.02% | 2.15 | 0.88-5.27 | 0.10 | |||||||
≥6.0 | 9 | 0.02% | 8 | 0.02% | 1.18 | 0.46-3.07 | 0.73 | |||||||
Number + lymph nodes | 49 | 1.6 ± 3.3 | 44 | 0.9 ± 1.8 | ||||||||||
Missing | 10 | 0.03% | 9 | 0.02% | Not determined | |||||||||
None | 26 | 0.07% | 25 | 0.06% | 1.05 | 0.61-1.82 | 0.86 | |||||||
1-2 | 10 | 0.03% | 16 | 0.04% | 0.65 | 0.30-1.44 | 0.29 | |||||||
≥3 | 12 | 0.03% | 3 | 0.01% | 4.06 | 1.15-14.40 | 0.03 |
NOTE: Cox regression models stratified according to age groups, dietary modification participation, and previous colorectal cancer. Calcium and vitamin D participation is adjusted as a time-dependent covariate.
Among the 18 interactions examined, history of polyp removal, height, and waist circumference differed at the 0.05 level of significance (Table 3) with more colorectal cancer seen in the conjugated equine estrogen group among women with waist circumference of <85 cm or height of <158.7 cm. Although the interaction test was not significant, the hazard ratio of the conjugated equine estrogen to placebo group for invasive colorectal cancer was 0.82 (95% CI, 0.34-1.98) for women 50 to 59 years and 0.83 (95% CI, 0.49-1.43) for women 60 to 69 years old at baseline compared with 2.11 (95% CI, 1.06-4.21) in the 70 to 79 year age group.
Incidence (annualized rate) of invasive colorectal cancer, according to baseline characteristics and treatment group
Variable . | CEE alone . | Placebo . | Hazard ratio* (95% CI) . | P value for interaction† . | ||||
---|---|---|---|---|---|---|---|---|
Number of women (annualized %) | ||||||||
Overall | 58 (0.15%) | 53 (0.14%) | 1.12 (0.77,1.63) | |||||
Age at enrollment (y) | 0.08 | |||||||
50-59 | 9 (0.07%) | 11 (0.09%) | 0.82 (0.34,1.98) | |||||
60-69 | 24 (0.14%) | 30 (0.17%) | 0.83 (0.49,1.43) | |||||
70-79 | 25 (0.29%) | 12 (0.13%) | 2.11 (1.06,4.21) | |||||
Race or ethnic group | 0.40 | |||||||
White | 46 (0.16%) | 39 (0.13%) | 1.20 (0.79,1.85) | |||||
Black | 5 (0.09%) | 6 (0.10%) | 0.92 (0.28,3.02) | |||||
Hispanic | 4 (0.18%) | 5 (0.22%) | 0.66 (0.17,2.56) | |||||
American Indians | 1 (0.36%) | 0 (0.00%) | ND | |||||
Asian or Pacific Islander | 2 (0.36%) | 1 (0.18%) | 1.04 (0.06,16.60) | |||||
Unknown or other | 0 (0.00%) | 2 (0.39%) | ND | |||||
History of polyp removal | 0.00‡ | |||||||
No | 40 (0.13%) | 46 (0.15%) | 0.87 (0.57,1.33) | |||||
Yes | 9 (0.29%) | 0 (0.00%) | ND | |||||
Family history of colorectal cancer | 0.10 | |||||||
No | 41 (0.15%) | 45 (0.16%) | 0.94 (0.62,1.44) | |||||
Yes | 13 (0.23%) | 6 (0.10%) | 2.16 (0.82,5.72) | |||||
BMI | 0.09 | |||||||
<25 | 18 (0.23%) | 9 (0.11%) | 2.02 (0.91,4.50) | |||||
25-29 | 19 (0.15%) | 18 (0.13%) | 1.11 (0.58,2.11) | |||||
≥30 | 21 (0.12%) | 26 (0.15%) | 0.80 (0.45,1.42) | |||||
Height (cm) | 0.03 | |||||||
96.0-158.6 | 26 (0.21%) | 12 (0.10%) | 2.12 (1.07,4.19) | |||||
158.7-163.9 | 18 (0.15%) | 15 (0.12%) | 1.27 (0.64,2.52) | |||||
164.0-188.3 | 14 (0.11%) | 26 (0.20%) | 0.57 (0.29,1.10) | |||||
Waist circumference (cm) | 0.03 | |||||||
37.0-84.9 | 22 (0.17%) | 11 (0.09%) | 2.03 (0.98,4.19) | |||||
85.0-96.9 | 18 (0.15%) | 17 (0.13%) | 1.14 (0.59,2.22) | |||||
97.0-191.6 | 18 (0.14%) | 25 (0.20%) | 0.68 (0.37,1.25) | |||||
Smoking status | 0.92 | |||||||
Never | 33 (0.17%) | 28 (0.14%) | 1.14 (0.69,1.89) | |||||
Past | 20 (0.14%) | 19 (0.12%) | 1.14 (0.61,2.15) | |||||
Current | 5 (0.13%) | 6 (0.15%) | 0.94 (0.28,3.12) | |||||
Alcohol intake | 0.29 | |||||||
None | 20 (0.14%) | 20 (0.14%) | 1.03 (0.55,1.91) | |||||
<1 drink/wk | 22 (0.17%) | 12 (0.09%) | 1.86 (0.92,3.76) | |||||
≥1 drink/wk | 16 (0.15%) | 19 (0.17%) | 0.96 (0.49,1.89) | |||||
Dietary selenium (μg/d) | 0.88 | |||||||
≤74.4 | 21 (0.16%) | 18 (0.14%) | 1.22 (0.65,2.29) | |||||
74.5-106.3 | 18 (0.16%) | 13 (0.11%) | 1.44 (0.71,2.95) | |||||
>106.3 | 19 (0.16%) | 17 (0.14%) | 1.07 (0.55,2.06) | |||||
Treated diabetes | 0.69 | |||||||
Never | 52 (0.15%) | 46 (0.13%) | 1.15 (0.77,1.71) | |||||
Current or past | 6 (0.22%) | 7 (0.25%) | 1.12 (0.36,3.48) | |||||
Use of NSAIDs | 0.96 | |||||||
No | 40 (0.16%) | 37 (0.14%) | 1.11 (0.71,1.74) | |||||
Yes | 18 (0.15%) | 16 (0.13%) | 1.18 (0.60,2.33) | |||||
Previous use of menopausal hormones | 0.78 | |||||||
No | 32 (0.16%) | 30 (0.15%) | 1.07 (0.65,1.76) | |||||
Yes | 26 (0.14%) | 23 (0.11%) | 1.17 (0.67,2.05) | |||||
Previous use of oral contraceptives | 0.90 | |||||||
No | 38 (0.17%) | 34 (0.14%) | 1.15 (0.72,1.82) | |||||
Yes | 20 (0.14%) | 19 (0.13%) | 1.08 (0.57,2.02) | |||||
Bilateral oophorectomy at baseline | 0.35 | |||||||
No | 25 (0.12%) | 28 (0.13%) | 0.89 (0.52,1.52) | |||||
Yes | 28 (0.21%) | 24 (0.16%) | 1.25 (0.72,2.16) |
Variable . | CEE alone . | Placebo . | Hazard ratio* (95% CI) . | P value for interaction† . | ||||
---|---|---|---|---|---|---|---|---|
Number of women (annualized %) | ||||||||
Overall | 58 (0.15%) | 53 (0.14%) | 1.12 (0.77,1.63) | |||||
Age at enrollment (y) | 0.08 | |||||||
50-59 | 9 (0.07%) | 11 (0.09%) | 0.82 (0.34,1.98) | |||||
60-69 | 24 (0.14%) | 30 (0.17%) | 0.83 (0.49,1.43) | |||||
70-79 | 25 (0.29%) | 12 (0.13%) | 2.11 (1.06,4.21) | |||||
Race or ethnic group | 0.40 | |||||||
White | 46 (0.16%) | 39 (0.13%) | 1.20 (0.79,1.85) | |||||
Black | 5 (0.09%) | 6 (0.10%) | 0.92 (0.28,3.02) | |||||
Hispanic | 4 (0.18%) | 5 (0.22%) | 0.66 (0.17,2.56) | |||||
American Indians | 1 (0.36%) | 0 (0.00%) | ND | |||||
Asian or Pacific Islander | 2 (0.36%) | 1 (0.18%) | 1.04 (0.06,16.60) | |||||
Unknown or other | 0 (0.00%) | 2 (0.39%) | ND | |||||
History of polyp removal | 0.00‡ | |||||||
No | 40 (0.13%) | 46 (0.15%) | 0.87 (0.57,1.33) | |||||
Yes | 9 (0.29%) | 0 (0.00%) | ND | |||||
Family history of colorectal cancer | 0.10 | |||||||
No | 41 (0.15%) | 45 (0.16%) | 0.94 (0.62,1.44) | |||||
Yes | 13 (0.23%) | 6 (0.10%) | 2.16 (0.82,5.72) | |||||
BMI | 0.09 | |||||||
<25 | 18 (0.23%) | 9 (0.11%) | 2.02 (0.91,4.50) | |||||
25-29 | 19 (0.15%) | 18 (0.13%) | 1.11 (0.58,2.11) | |||||
≥30 | 21 (0.12%) | 26 (0.15%) | 0.80 (0.45,1.42) | |||||
Height (cm) | 0.03 | |||||||
96.0-158.6 | 26 (0.21%) | 12 (0.10%) | 2.12 (1.07,4.19) | |||||
158.7-163.9 | 18 (0.15%) | 15 (0.12%) | 1.27 (0.64,2.52) | |||||
164.0-188.3 | 14 (0.11%) | 26 (0.20%) | 0.57 (0.29,1.10) | |||||
Waist circumference (cm) | 0.03 | |||||||
37.0-84.9 | 22 (0.17%) | 11 (0.09%) | 2.03 (0.98,4.19) | |||||
85.0-96.9 | 18 (0.15%) | 17 (0.13%) | 1.14 (0.59,2.22) | |||||
97.0-191.6 | 18 (0.14%) | 25 (0.20%) | 0.68 (0.37,1.25) | |||||
Smoking status | 0.92 | |||||||
Never | 33 (0.17%) | 28 (0.14%) | 1.14 (0.69,1.89) | |||||
Past | 20 (0.14%) | 19 (0.12%) | 1.14 (0.61,2.15) | |||||
Current | 5 (0.13%) | 6 (0.15%) | 0.94 (0.28,3.12) | |||||
Alcohol intake | 0.29 | |||||||
None | 20 (0.14%) | 20 (0.14%) | 1.03 (0.55,1.91) | |||||
<1 drink/wk | 22 (0.17%) | 12 (0.09%) | 1.86 (0.92,3.76) | |||||
≥1 drink/wk | 16 (0.15%) | 19 (0.17%) | 0.96 (0.49,1.89) | |||||
Dietary selenium (μg/d) | 0.88 | |||||||
≤74.4 | 21 (0.16%) | 18 (0.14%) | 1.22 (0.65,2.29) | |||||
74.5-106.3 | 18 (0.16%) | 13 (0.11%) | 1.44 (0.71,2.95) | |||||
>106.3 | 19 (0.16%) | 17 (0.14%) | 1.07 (0.55,2.06) | |||||
Treated diabetes | 0.69 | |||||||
Never | 52 (0.15%) | 46 (0.13%) | 1.15 (0.77,1.71) | |||||
Current or past | 6 (0.22%) | 7 (0.25%) | 1.12 (0.36,3.48) | |||||
Use of NSAIDs | 0.96 | |||||||
No | 40 (0.16%) | 37 (0.14%) | 1.11 (0.71,1.74) | |||||
Yes | 18 (0.15%) | 16 (0.13%) | 1.18 (0.60,2.33) | |||||
Previous use of menopausal hormones | 0.78 | |||||||
No | 32 (0.16%) | 30 (0.15%) | 1.07 (0.65,1.76) | |||||
Yes | 26 (0.14%) | 23 (0.11%) | 1.17 (0.67,2.05) | |||||
Previous use of oral contraceptives | 0.90 | |||||||
No | 38 (0.17%) | 34 (0.14%) | 1.15 (0.72,1.82) | |||||
Yes | 20 (0.14%) | 19 (0.13%) | 1.08 (0.57,2.02) | |||||
Bilateral oophorectomy at baseline | 0.35 | |||||||
No | 25 (0.12%) | 28 (0.13%) | 0.89 (0.52,1.52) | |||||
Yes | 28 (0.21%) | 24 (0.16%) | 1.25 (0.72,2.16) |
Cox regression models are stratified according to age groups, dietary modification participation, and previous colorectal cancer. Calcium and vitamin D participation is adjusted as a time-dependent covariate.
Test of linear trend is used for age, BMI, height, and waist circumference.
P value for interaction derived from exact test under conditional logistic regression model.
Based on self-reports, the proportion of women who developed polyps during the trial was comparable between the two randomization groups (16.9% of hormone and 17.1% of placebo users) as was the time to first report of an identified polyp (hazard ratio, 0.97; 95% CI, 0.87-1.07). A higher risk for invasive colorectal cancer was associated with previous polyp removal among those in the hormone compared with the control group; however, the limited number of cases (n = 9) precludes reliable estimation of effect (Table 3).
The frequency of bowel exams, including rates of hemoccult testing, rectal exams, and sigmoidoscopy or colonoscopy, was similar in the conjugated equine estrogen and placebo groups during the study. Screening rates increased during the study, with rates of sigmoidoscopy or colonoscopy ranging from ∼9% to 15.5% per year and other exams (rectal or hemoccult) ranging from ∼16% to 23%. Over the study period, 24.0% of women in the hormone group and 22.9% of the placebo group participants reported no bowel screening. Surveillance sigmoidoscopy or colonoscopy was more frequent in women with previous polyp removal than was screening use of these exams in women without such history (69% versus 44%). Screening with these tests was also more frequent in those with a colorectal cancer family history (57% versus 45%) and somewhat more frequent in women 50 to 59 years old but did not differ by randomization group.
Mortality outcomes for participants who were diagnosed with colorectal cancer were compared between randomization groups. After 3.6 ± 2.7 years (mean ± SD) following a colorectal cancer diagnosis, the cumulative mortality rate for women in the conjugated equine estrogen alone group was 34% (20 of 58 dead) compared with 30% (16 of 53 dead) in the placebo group (hazard ratio, 1.34; 95% CI, 0.58-3.19). Kaplan-Meier plots by randomization group are provided in Fig. 2.
Survival following diagnosis of invasive colorectal cancer by randomization.
In a cross-study comparison, the colorectal cancer annual incidence rates were found to be similar in the placebo groups from the trial of conjugated equine estrogen alone compared with the trial of conjugated equine estrogen plus medroxyprogesterone acetate (0.14% for women in the conjugated equine estrogen alone trial versus 0.16% for women in the conjugated equine estrogen plus medroxyprogesterone acetate trials; P = 0.35 for comparison of incidence between the two placebo groups; ref. 1). Difference between the conjugated equine estrogen plus medroxyprogesterone acetate (9) and conjugated equine estrogen effect on colorectal cancer were explored using a Cox regression model stratified according to cohorts, age, dietary modification participation, history of colorectal cancer, and adjusting for covariates of ethnicity, body mass index (BMI), history of polyp removal, previous colonoscopy or sigmoidoscopy, and family history of colorectal cancer. The ratio of the hazard ratios for the conjugated equine estrogen alone compared with the conjugated equine estrogen plus medroxyprogesterone acetate trial for colorectal cancer incidence was 2.21 (95% CI, 1.29-3.81; P = 0.004), indicating a significant difference in hormone therapy effects in the two trials.
Discussion
In a randomized, double-blind, placebo-controlled trial, conjugated equine estrogen alone for an average of 7.1 years did not influence invasive colorectal cancer incidence or survival following a colorectal cancer diagnosis in postmenopausal women. No differences were observed in the clinical characteristics of the tumors diagnosed in the conjugated equine estrogen and placebo groups. Colorectal cancer risk factors were balanced between treatment groups, as were examinations for colorectal cancer detection.
The lack of an effect of conjugated equine estrogen alone on colorectal cancer incidence contrasts with the results of the WHI trial of conjugated equine estrogen plus medroxyprogesterone acetate, wherein colorectal cancer incidence was significantly lower in the combined hormone group (9). In the randomized Heart and Estrogen/Progestin Replacement Study fewer colorectal cancers were also seen in women randomized to the conjugated equine estrogen and medroxyprogesterone acetate compared with the placebo groups (hazard ratio, 0.69; 95% CI, 0.32-1.49; not significant; ref. 21).
Although participants in the two WHI trials, one evaluating conjugated equine estrogen plus medroxyprogesterone acetate for women with no previous hysterectomy and one evaluating conjugated equine estrogen alone for women with previous hysterectomy, differed in several baseline characteristics (22), the colorectal cancer annual incidence rates were similar in the placebo groups for the two trials. Thus, differences in participant characteristics do not explain the discrepant findings and suggest progestin may modify the influence of estrogen on colorectal cancer. Because colorectal cancers in the WHI combined hormone therapy trial were diagnosed at a higher stage compared with those diagnosed on placebo, a differential influence on cancer detection in the two studies also cannot be excluded as an explanation.
In the predominance of observational studies, postmenopausal hormone therapy use has been consistently associated with reduced colorectal cancer risk (23, 24) and improved cancer-specific and overall survival in some reports (15, 16, 25). However, the observational study literature about menopausal hormone therapy and colorectal cancer have not commonly considered use of estrogen alone and combined hormone therapy separately. Several case-control studies reported that both estrogen plus progestin, as well as estrogen alone, were associated with lower colorectal cancer risk (12, 26). A recent case-control in the United States found an association with lower colorectal cancer risk only in women using estrogen plus progestin (odds ratio, 0.6; 95% CI, 0.5-0.9) but not those using estrogen alone (odds ratio, 0.9; 95% CI, 0.7-1.1; ref. 13). Similarly, in a Swedish cohort (27) and in the large United Kingdom General Practice Research Database, colorectal cancer risk was lower only in women using estrogen plus progestin but not those using estrogen alone (28). Future observational studies of colorectal cancer should clearly differentiate combined hormone use from estrogen alone.
In the current randomized trial, there was no improvement in the post diagnosis survival for women with colorectal cancers in the conjugated equine estrogen group (n = 58 cases) compared with those diagnosed in the placebo group (n = 53 cases; hazard ratio, 1.34; 95% CI, 0.58-3.19). This result stands in contrast to the recent findings of Chan and colleagues (15), who reported that, for women in the Nurse's Health Study cohort, current users of estrogen at the time of diagnosis of colorectal cancer (n = 235 cases) experienced significantly less colorectal cancer–specific mortality and less overall mortality compared with women who never used hormone therapy before a colorectal cancer diagnosis (multivariate hazard ratio for overall mortality, 0.74; 95% CI, 0.56-0.97). In that observational study experience, 75% of the person-time represented estrogen alone use. Other reports of longer survival for women diagnosed with colorectal cancer when they were receiving menopausal hormone therapy do not provide details about type of hormone therapy received (19, 29). Given the findings from the WHI randomized trial, current evidence does not support clinical use of estrogen to improve outcome of postmenopausal women with diagnosed colorectal cancer.
The biological effects of combined estrogen plus progestin and estrogen alone on malignant processes are complex. Use of estrogen alone increases endometrial proliferation and endometrial cancer risk, whereas the addition of progestin to estrogen largely abrogates the estrogen effects (8, 30). Combined estrogen plus progestin use significantly increased breast cancer risk (31), whereas the use of estrogen alone was associated with a nonsignificant reduction in breast cancer incidence (32). Combined conjugated equine estrogen plus medroxyprogesterone acetate use reduced colorectal cancer incidence in the WHI trial for women with no previous hysterectomy (9), but the use of conjugated equine estrogen alone in the currently reported trial did not influence either colorectal cancer incidence or mortality after diagnosis. Taken together, these results suggest an hypothesis that progestins modulate estrogen effects on malignancy and that their influence is organ specific. Alternatively, progestins and/or estrogens alone could simply have different effects on these three cancer types, but clinical evidence for effects related to progestin alone use are sparse.
Study subgroup analyses suggested that conjugated equine estrogen influence on colorectal cancer may vary according to age. Alternatively, given the limited numbers of women in these subgroup analyses and that the test for interaction was not significant, chance alone may explain the age effects observed. One case-control study (33) reporting on mostly unopposed estrogen use found no association with colorectal cancer risk in younger women, but hormone users with ages 60 to 69 years had a lower colorectal cancer risk relative to non-users. A nominally significant interaction among conjugated equine estrogen alone use and colorectal cancer with both waist circumference and height was seen, along with a suggestion of an interaction with BMI. However, observational studies examining BMI modification of hormone therapy association with colorectal cancer have provided conflicting results (16, 34-38), and relationships among BMI, menopausal hormone therapy, and colorectal cancer risk remain unclear.
Although no effect of conjugated equine estrogen alone on colorectal cancer incidence was seen in the current trial, several mechanisms have been proposed as potential mediators of the favorable effect of combined conjugated equine estrogen plus medroxyprogesterone acetate use on colorectal cancer previously reported (9). These include hormone influences on bile acids, (6) insulin (39), and direct effects on the intestinal epithelium (40, 41).
Strengths of the current study include the randomized double-blind design and the long-term follow-up of a large number of ethnically diverse participants with standardized outcome ascertainment and adjudication. The limited number of colorectal cancers is a constraint, as is the lower-than-anticipated adherence to study medications, which likely dilutes any true effect. Although bowel screening was not protocol defined, the frequency of exams was comparable in the two groups.
In summary, in a randomized placebo-controlled clinical trial in 10,749 postmenopausal women, use of conjugated equine estrogen alone did not reduce colorectal cancer incidence or improve the post diagnosis survival of women with this disease.
Women's Health Initiative Investigators
Program Office: (National Heart, Lung, and Blood Institute, Bethesda, MD) Barbara Alving, Jacques Rossouw, Linda Pottern, Shari Ludlam, Joan McGowan, Nancy Geller, and Leslie Ford.
Clinical Coordinating Center: (Fred Hutchinson Cancer Research Center, Seattle, WA) Ross Prentice, Garnet Anderson, Andrea LaCroix, Ruth Patterson, Anne McTiernan, Barbara Cochrane, Julie Hunt, Lesley Tinker, Charles Kooperberg, Martin McIntosh, C.Y. Wang, Chu Chen, Deborah Bowen, Alan Kristal, Janet Stanford, Nicole Urban, Noel Weiss, and Emily White; (Wake Forest University School of Medicine, Winston-Salem, NC) Sally Shumaker, Ronald Prineas, and Michelle Naughton; (Medical Research Laboratories, Highland Heights, KY) Evan Stein and Peter Laskarzewski; (San Francisco Coordinating Center, San Francisco, CA) Steven R. Cummings, Michael Nevitt, and Lisa Palermo; (University of Minnesota, Minneapolis, MN) Lisa Harnack; (Fisher BioServices, Rockville, MD) Frank Cammarata and Steve Lindenfelser; (University of Washington, Seattle, WA) Bruce Psaty and Susan Heckbert.
Clinical Centers: (Albert Einstein College of Medicine, Bronx, NY) Sylvia Wassertheil-Smoller, William Frishman, Judith Wylie-Rosett, David Barad, and Ruth Freeman; (Baylor College of Medicine, Houston, TX) Jennifer Hays, Ronald Young, Jill Anderson, Sandy Lithgow, and Paul Bray; (Brigham and Women's Hospital, Harvard Medical School, Boston, MA) JoAnn Manson, J. Michael Gaziano, Claudia Chae, Kathryn Rexrode, and Caren Solomon; (Brown University, Providence, RI) Annlouise R. Assaf, Carol Wheeler, Charles Eaton, and Michelle Cyr; (Emory University, Atlanta, GA) Lawrence Phillips, Margaret Pedersen, Ora Strickland, Margaret Huber, and Vivian Porter; (Fred Hutchinson Cancer Research Center, Seattle, WA) Shirley A.A. Beresford, Vicky M. Taylor, Nancy F. Woods, Maureen Henderson, and Robyn Andersen; (George Washington University, Washington, DC) Judith Hsia, Nancy Gaba, and Joao Ascensao; (Harbor-UCLA Research and Education Institute, Torrance, CA) Rowan Chlebowski, Robert Detrano, Anita Nelson, and Michele Geller; (Kaiser Permanente Center for Health Research, Portland, OR) Evelyn Whitlock, Victor Stevens, and Njeri Karanja; (Kaiser Permanente Division of Research, Oakland, CA) Bette Caan, Stephen Sidney, Geri Bailey, and Jane Hirata; (Medical College of Wisconsin, Milwaukee, WI) Jane Morley Kotchen, Vanessa Barnabei, Theodore A. Kotchen, Mary Ann C. Gilligan, and Joan Neuner; (MedStar Research Institute, Howard University, Washington, DC) Barbara V. Howard, Lucile Adams-Campbell, Lawrence Lessin, Monique Rainford, and Gabriel Uwaifo; (Northwestern University, Chicago/Evanston, IL) Linda Van Horn, Philip Greenland, Janardan Khandekar, Kiang Liu, Carol Rosenberg; (Rush University Medical Center, Chicago, IL) Henry Black, Lynda Powell, and Ellen Mason; Martha Gulati; (Stanford Prevention Research Center, Stanford, CA) Marcia L. Stefanick, Mark A. Hlatky, Bertha Chen, Randall S. Stafford, and Sally Mackey; (State University of New York at Stony Brook, Stony Brook, NY) Dorothy Lane, Iris Granek, William Lawson, Gabriel San Roman, and Catherine Messina; (The Ohio State University, Columbus, OH) Rebecca Jackson, Randall Harris, Electra Paskett, W. Jerry Mysiw, and Michael Blumenfeld; (University of Alabama at Birmingham, Birmingham, AL) Cora E. Lewis, Albert Oberman, James M. Shikany, Monika Safford, and Mona Fouad; (University of Arizona, Tucson/Phoenix, AZ) Cyndi Thomson, Tamsen Bassford, Marcia Ko, Ana Maria Lopez, and Cheryl Ritenbaugh; (University at Buffalo, Buffalo, NY) Jean Wactawski-Wende, Maurizio Trevisan, Ellen Smit, Susan Graham, and June Chang; (University of California at Davis, Sacramento, CA) John Robbins and S. Yasmeen; (University of California at Irvine, Irvine, CA) F. Allan Hubbell, Gail Frank, Nathan Wong, Nancy Greep, and Bradley Monk; (University of California at Los Angeles, Los Angeles, CA) Howard Judd, David Heber, and Robert Elashoff; (University of California at San Diego, LaJolla/Chula Vista, CA) Robert D. Langer, Michael H. Criqui, Gregory T. Talavera, Cedric F. Garland, and Matthew A. Allison; (University of Cincinnati, Cincinnati, OH) Margery Gass and Suzanne Wernke; (University of Florida, Gainesville/Jacksonville, FL) Marian Limacher, Michael Perri, Andrew Kaunitz, R. Stan Williams, and Yvonne Brinson; (University of Hawaii, Honolulu, HI) J. David Curb, Helen Petrovitch, Beatriz Rodriguez, Kamal Masaki, and Santosh Sharma; (University of Iowa, Iowa City/Davenport, IA) Robert Wallace, James Torner, Susan Johnson, Linda Snetselaar, and Jennifer Robinson; (University of Massachusetts/Fallon Clinic, Worcester, MA) Judith Ockene, Milagros Rosal, Ira Ockene, Robert Yood, and Patricia Aronson; (University of Medicine and Dentistry of New Jersey, Newark, NJ) Norman Lasser, Baljinder Singh, Vera Lasser, John Kostis, and Peter McGovern; (University of Miami, Miami, FL) Mary Jo O'Sullivan, Linda Parker, Timothy DeSantis, Diann Fernandez, and Pat Caralis; (University of Minnesota, Minneapolis, MN) Karen L. Margolis, Richard H. Grimm, Mary F. Perron, Cynthia Bjerk, and Sarah Kempainen; (University of Nevada, Reno, NV) Robert Brunner, William Graettinger, Vicki Oujevolk, and Michael Bloch; (University of North Carolina, Chapel Hill, NC) Gerardo Heiss, Pamela Haines, David Ontjes, Carla Sueta, and Ellen Wells; (University of Pittsburgh, Pittsburgh, PA) Lewis Kuller, Jane Cauley, and N. Carole Milas; (University of Tennessee Health Science Center, Memphis, TN) Karen C. Johnson, Suzanne Satterfield, Raymond W. Ke, Stephanie Connelly, and Fran Tylavsky; (University of Texas Health Science Center, San Antonio, TX) Robert Brzyski, Robert Schenken, Jose Trabal, Mercedes Rodriguez-Sifuentes, and Charles Mouton; (University of Wisconsin, Madison, WI) Gloria E. Sarto, Douglas Laube, Patrick McBride, Julie Mares-Perlman, and Barbara Loevinger; (Wake Forest University School of Medicine, Winston-Salem, NC) Denise Bonds, Greg Burke, Robin Crouse, Mara Vitolins, and Scott Washburn; (Wayne State University School of Medicine/Hutzel Hospital, Detroit, MI) Susan Hendrix, Michael Simon, and Gene McNeeley.
Former Principal Investigators and Project Officers: (Baylor College of Medicine) John Foreyt, Ph.D.; (Emory University) Dallas Hall, M.D.; (George Washington University) Valery Miller, M.D.; (Kaiser, Oakland, CA) Robert Hiatt, M.D.; (Kaiser, Portland, OR), Barbara Valanis, Ph.D.; (National Cancer Institute, Bethesda, MD), Carolyn Clifford13
Deceased.
Disclosure of Potential Conflicts of Interest
Dr. Chlebowski is a consultant for Astra-Zeneca Pharmaceuticals LP (Wilmington, DE), Novartis (Basel, Switzerland), Pfizer, Inc. (New York, NY), Eli Lilly and Co. (Indianapolis, IN), and Organon International (Kenilworth, NJ) and has received research support from Eli Lilly and Co. and Organon International. There are no other relationships to disclose.
Acknowledgments
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.