The Breast Cancer Prevention Trial (BCPT, NSABP P-1) was a double-blind, randomized, placebo-controlled trial with a primary objective to determine whether 5 years of tamoxifen therapy (20 mg/day) would reduce the incidence of invasive breast cancer in women at high risk for the disease (see INDICATIONS AND USAGE). Secondary objectives included an evaluation of the incidence of ischemic heart disease; the effects on the incidence of bone fractures; and other events that might be associated with the use of tamoxifen, including: endometrial cancer, pulmonary embolus, deep-vein thrombosis, stroke, and cataract formation and surgery (see WARNINGS).
The Gail Model was used to calculate predicted breast cancer risk for women who were less than 60 years of age and did not have lobular carcinoma in situ (LCIS). The following risk factors were used: age; number of first-degree female relatives with breast cancer; previous breast biopsies; presence or absence of atypical hyperplasia; nulliparity; age at first live birth; and age at menarche. A 5 year predicted risk of breast cancer of ≥ 1.67% was required for entry into the trial.
In this trial, 13,388 women of at least 35 years of age were randomized to receive either tamoxifen or placebo for five years. The median duration of treatment was 3.5 years. As of January 31, 1998, follow-up data is available for 13,114 women. Twenty-seven percent of women randomized to placebo (1,782) and 24% of women randomized to tamoxifen (1,596) completed 5 years of therapy. The demographic characteristics of women on the trial with follow-up data are shown in Table 2.
Table 2: Demographic Characteristics of Women in the NSABP P-1 Trial| Characteristic | Placebo | Tamoxifen |
|---|
| # | % | # | % |
|---|
| Age (yrs.) |
| 35 to 39 | 184 | 3 | 158 | 2 |
| 40 to 49 | 2,394 | 36 | 2,411 | 37 |
| 50 to 59 | 2,011 | 31 | 2,019 | 31 |
| 60 to 69 | 1,588 | 24 | 1,563 | 24 |
| ≥ 70 | 393 | 6 | 393 | 6 |
| Age at first live birth (yrs.) |
| Nulliparous | 1,202 | 18 | 1,205 | 18 |
| 12 to 19 | 915 | 14 | 946 | 15 |
| 20 to 24 | 2,448 | 37 | 2,449 | 37 |
| 25 to 29 | 1,399 | 21 | 1,367 | 21 |
| ≥ 30 | 606 | 9 | 577 | 9 |
| Race |
| White | 6,333 | 96 | 6,323 | 96 |
| Black | 109 | 2 | 103 | 2 |
| Other | 128 | 2 | 118 | 2 |
| Age at menarche |
| ≥ 14 | 1,243 | 19 | 1,170 | 18 |
| 12 to 13 | 3,610 | 55 | 3,610 | 55 |
| ≤ 11 | 1,717 | 26 | 1,764 | 27 |
| # of first degree relatives with breast cancer |
| 0 | 1,584 | 24 | 1,525 | 23 |
| 1 | 3,714 | 57 | 3,744 | 57 |
| 2+ | 1,272 | 19 | 1,275 | 20 |
| Prior hysterectomy |
| No | 4,173 | 63.5 | 4,018 | 62.4 |
| Yes | 2,397 | 36.5 | 2,464 | 37.7 |
| # of previous breast biopsies |
| 0 | 2,935 | 45 | 2,923 | 45 |
| 1 | 1,833 | 28 | 1,850 | 28 |
| ≥ 2 | 1,802 | 27 | 1,771 | 27 |
| History of atypical hyperplasia in the breast |
| No | 5,958 | 91 | 5,969 | 91 |
| Yes | 612 | 9 | 575 | 9 |
| History of LCIS at entry |
| No | 6,165 | 94 | 6,135 | 94 |
| Yes | 405 | 6 | 409 | 6 |
| 5 year predicted breast cancer risk (%) |
| ≤ 2.00 | 1,646 | 25 | 1,626 | 25 |
| 2.01 to 3.00 | 2,028 | 31 | 2,057 | 31 |
| 3.01 to 5.00 | 1,787 | 27 | 1,707 | 26 |
| ≥ 5.01 | 1,109 | 17 | 1,162 | 18 |
| Total | 6,570 | 100.0 | 6,544 | 100.0 |
Results are shown in Table 3. After a median follow-up of 4.2 years, the incidence of invasive breast cancer was reduced by 44% among women assigned to tamoxifen (86 cases-tamoxifen, 156 cases-placebo; p < 0.00001; relative risk (RR) = 0.56, 95% CI: 0.43 to 0.72). A reduction in the incidence of breast cancer was seen in each prospectively specified age group (≤ 49, 50 to 59, ≥ 60), in women with or without LCIS, and in each of the absolute risk levels specified in Table 3. A non-significant decrease in the incidence of ductal carcinoma in situ (DCIS) was seen (23 tamoxifen, 35 placebo; RR = 0.66, 95% CI: 0.39 to 1.11).
There was no statistically significant difference in the number of myocardial infarctions, severe angina, or acute ischemic cardiac events between the two groups (61 tamoxifen, 59 placebo; RR = 1.04, 95% CI: 0.73 to 1.49).
No overall difference in mortality (53 deaths in tamoxifen group vs. 65 deaths in placebo group) was present. No difference in breast cancer-related mortality was observed (4 deaths in tamoxifen group vs. 5 deaths in placebo group).
Although there was a non-significant reduction in the number of hip fractures (9 on tamoxifen, 20 on placebo) in the tamoxifen group, the number of wrist fractures was similar in the two treatment groups (69 on tamoxifen, 74 on placebo). A subgroup analysis of the P-1 trial, suggests a difference in effect in bone mineral density (BMD) related to menopausal status in patients receiving tamoxifen. In postmenopausal women there was no evidence of bone loss of the lumbar spine and hip. Conversely, tamoxifen was associated with significant bone loss of the lumbar spine and hip in premenopausal women.
The risks of tamoxifen therapy include endometrial cancer, DVT, PE, stroke, cataract formation, and cataract surgery (see Table 3). In the NSABP P-1 trial, 33 cases of endometrial cancer were observed in the tamoxifen group vs. 14 in the placebo group (RR = 2.48, 95% CI: 1.27 to 4.92). Deep-vein thrombosis was observed in 30 women receiving tamoxifen vs. 19 in women receiving placebo (RR = 1.59, 95% CI: 0.86 to 2.98). Eighteen cases of pulmonary embolism were observed in the tamoxifen group vs. 6 in the placebo group (RR = 3.01, 95% CI: 1.15 to 9.27). There were 34 strokes on the tamoxifen arm and 24 on the placebo arm (RR =1.42, 95% CI: 0.82 to 2.51). Cataract formation in women without cataracts at baseline was observed in 540 women taking tamoxifen vs. 483 women receiving placebo (RR = 1.13, 95% CI: 1.00 to 1.28). Cataract surgery (with or without cataracts at baseline) was performed in 201 women taking tamoxifen vs. 129 women receiving placebo (RR = 1.51, 95% CI: 1.21 to 1.89) (see WARNINGS).
Table 3 summarizes the major outcomes of the NSABP P-1 trial. For each endpoint, the following results are presented: the number of events and rate per 1,000 women per year for the placebo and tamoxifen groups; and the relative risk (RR) and its associated 95% confidence interval (CI) between tamoxifen and placebo. Relative risks less than 1.0 indicate a benefit of tamoxifen therapy. The limits of the confidence intervals can be used to assess the statistical significance of the benefits or risks of tamoxifen therapy. If the upper limit of the CI is less than 1.0, then a statistically significant benefit exists.
For most participants, multiple risk factors would have been required for eligibility. This table considers risk factors individually, regardless of other co-existing risk factors, for women who developed breast cancer. The 5 year predicted absolute breast cancer risk accounts for multiple risk factors in an individual and should provide the best estimate of individual benefit (see INDICATIONS AND USAGE).
Table 3. Major Outcomes of the NSABP P-1 Trial| TYPE OF EVENT | # OF EVENTS | RATE/1,000 WOMEN/YEAR | | 95% CI |
|---|
| PLACEBO | TAMOXIFEN | PLACEBO | TAMOXIFEN | RR | LIMITS |
|---|
| Invasive Breast Cancer | 156 | 86 | 6.49 | 3.58 | 0.56 | 0.43 to 0.72 |
| Age ≤ 49 | 59 | 38 | 6.34 | 4.11 | 0.65 | 0.43 to 0.98 |
| Age 50 to 59 | 46 | 25 | 6.31 | 3.53 | 0.56 | 0.35 to 0.91 |
| Age ≥ 60 | 51 | 23 | 7.17 | 3.22 | 0.45 | 0.27 to 0.74 |
| Risk Factors for Breast Cancer History, LCIS |
| No | 140 | 78 | 6.23 | 3.51 | 0.56 | 0.43 to 0.74 |
| Yes | 16 | 8 | 12.73 | 6.33 | 0.50 | 0.21 to 1.17 |
| History, Atypical Hyperplasia |
| No | 138 | 84 | 6.37 | 3.89 | 0.61 | 0.47 to 0.80 |
| Yes | 18 | 2 | 8.69 | 1.05 | 0.12 | 0.03 to 0.52 |
| No. First Degree Relatives |
| 0 | 32 | 17 | 5.97 | 3.26 | 0.55 | 0.30 to 0.98 |
| 1 | 80 | 45 | 5.81 | 3.31 | 0.57 | 0.40 to 0.82 |
| 2 | 35 | 18 | 8.92 | 4.67 | 0.52 | 0.30 to 0.92 |
| ≥ 3 | 9 | 6 | 13.33 | 7.58 | 0.57 | 0.20 to 1.59 |
| 5 Year Predicted Breast Cancer Risk (as calculated by the Gail Model) |
| ≤ 2.00% | 31 | 13 | 5.36 | 2.26 | 0.42 | 0.22 to 0.81 |
| 2.01 to 3.00% | 39 | 28 | 5.25 | 3.83 | 0.73 | 0.45 to 1.18 |
| 3.01 to 5.00% | 36 | 26 | 5.37 | 4.06 | 0.76 | 0.46 to 1.26 |
| ≥ 5.00% | 50 | 19 | 13.15 | 4.71 | 0.36 | 0.21 to 0.61 |
| DCIS | 35 | 23 | 1.47 | 0.97 | 0.66 | 0.39 to 1.11 |
| Fractures (protocol-specified sites) | 92 Two women had hip and wrist fractures | 76
| 3.87 | 3.20 | 0.61 | 0.83 to 1.12 |
| Hip | 20 | 9 | 0.84 | 0.38 | 0.45 | 0.18 to 1.04 |
| Wrist Includes Colles' and other lower radius fractures | 74 | 69 | 3.11 | 2.91 | 0.93 | 0.67 to 1.29 |
| Total Ischemic Events | 59 | 61 | 2.47 | 2.57 | 1.04 | 0.71 to 1.51 |
| Myocardial Infarction | 27 | 27 | 1.13 | 1.13 | 1.00 | 0.57 to 1.78 |
| Fatal | 8 | 7 | 0.33 | 0.29 | 0.88 | 0.27 to 2.77 |
| Nonfatal | 19 | 20 | 0.79 | 0.84 | 1.06 | 0.54 to 2.09 |
| Angina Requiring angioplasty or CABG | 12 | 12 | 0.50 | 0.50 | 1.00 | 0.41 to 2.44 |
| Acute Ischemic Syndrome New Q-wave on ECG; no angina or elevation of serum enzymes; or angina requiring hospitalization without surgery | 20 | 22 | 0.84 | 0.92 | 1.11 | 0.58 to 2.13 |
| Uterine Malignancies (among women with an intact uterus) Updated long-term follow-up data (median 6.9 years) from NSABP P-1 study added after cut-off for the other information in this table. | 17 | 57 | | | | |
| Endometrial Adenocarcinoma
| 17 | 53 | 0.71 | 2.20 | | |
| Uterine Sarcoma
| 0 | 4 | 0.0 | 0.17 | | |
| Stroke Seven cases were fatal; three in the placebo group and four in the tamoxifen group | 24 | 34 | 1.00 | 1.43 | 1.42 | 0.82 to 2.51 |
| Transient Ischemic Attack | 21 | 18 | 0.88 | 0.75 | 0.86 | 0.43 to 1.70 |
| Pulmonary Emboli Three cases in the tamoxifen group were fatal | 6 | 18 | 0.25 | 0.75 | 3.01 | 1.15 to 9.27 |
| Deep-Vein Thrombosis All but three cases in each group required hospitalization | 19 | 30 | 0.79 | 1.26 | 1.59 | 0.86 to 2.98 |
| Cataracts Developing on Study Based on women without cataracts at baseline (6,230 Placebo, 6,199 Tamoxifen) | 483 | 540 | 22.51 | 25.41 | 1.13 | 1.00 to 1.28 |
| Underwent Cataract Surgery
| 63 | 101 | 2.83 | 4.57 | 1.62 | 1.18 to 2.22 |
| Underwent Cataract Surgery All women (6,707 Placebo, 6,681 Tamoxifen) | 129 | 201 | 5.44 | 8.56 | 1.58 | 1.26 to 1.97 |
Table 4 describes the characteristics of the breast cancers in the NSABP P-1 trial and includes tumor size, nodal status, ER status. Tamoxifen decreased the incidence of small estrogen receptor positive tumors, but did not alter the incidence of estrogen receptor negative tumors or larger tumors.
Table 4: Characteristics of Breast Cancer in NSABP P-1 Trial| Staging Parameter | Placebo | Tamoxifen | Total |
|---|
| N = 156 | N = 86 | N = 242 |
|---|
| Tumor size: |
| T1 | 117 | 60 | 177 |
| T2 | 28 | 20 | 48 |
| T3 | 7 | 3 | 10 |
| T4 | 1 | 2 | 3 |
| Unknown | 3 | 1 | 4 |
| Nodal status: | | | |
| Negative | 103 | 56 | 159 |
| 1 to 3 positive nodes | 29 | 14 | 43 |
| ≥ 4 positive nodes | 10 | 12 | 22 |
| Unknown | 14 | 4 | 18 |
| Stage: |
| I | 88 | 47 | 135 |
| II: node negative | 15 | 9 | 24 |
| II: node positive | 33 | 22 | 55 |
| III | 6 | 4 | 10 |
| IV | 2 One participant presented with a suspicious bone scan but did not have documented metastases. She subsequently died of metastatic breast cancer. | 1 | 3 |
| Unknown | 12 | 3 | 15 |
| Estrogen receptor: |
| Positive | 115 | 38 | 153 |
| Negative | 27 | 36 | 63 |
| Unknown | 14 | 12 | 26 |
Interim results from 2 trials in addition to the NSABP P-1 trial examining the effects of tamoxifen in reducing breast cancer incidence have been reported.
The first was the Italian Tamoxifen Prevention trial. In this trial women between the ages of 35 and 70, who had had a total hysterectomy, were randomized to receive 20 mg tamoxifen or matching placebo for 5 years. The primary endpoints were occurrence of, and death from, invasive breast cancer. Women without any specific risk factors for breast cancer were to be entered. Between 1992 and 1997, 5,408 women were randomized. Hormone Replacement Therapy (HRT) was used in 14% of participants. The trial closed in 1997 due to the large number of dropouts during the first year of treatment (26%). After 46 months of follow-up there were 22 breast cancers in women on placebo and 19 in women on tamoxifen. Although no decrease in breast cancer incidence was observed, there was a trend for reduction in breast cancer among women receiving protocol therapy for at least 1 year (19 placebo, 11 tamoxifen). The small numbers of participants along with the low level of risk in this otherwise healthy group precluded an adequate assessment of the effect of tamoxifen in reducing the incidence of breast cancer.
The second trial, the Royal Marsden Trial (RMT) was reported as an interim analysis. The RMT was begun in 1986 as a feasibility study of whether larger scale trials could be mounted. The trial was subsequently extended to a pilot trial to accrue additional participants to further assess the safety of tamoxifen. Twenty-four hundred and seventy-one women were entered between 1986 and 1996; they were selected on the basis of a family history of breast cancer. HRT was used in 40% of participants. In this trial, with a 70 month median follow-up, 34 and 36 breast cancers (8 noninvasive, 4 on each arm) were observed among women on tamoxifen and placebo, respectively. Patients in this trial were younger than those in the NSABP P-1 trial and may have been more likely to develop ER (-) tumors, which are unlikely to be reduced in number by tamoxifen therapy. Although women were selected on the basis of family history and were thought to have a high risk of breast cancer, few events occurred, reducing the statistical power of the study. These factors are potential reasons why the RMT may not have provided an adequate assessment of the effectiveness of tamoxifen in reducing the incidence of breast cancer.
In these trials, an increased number of cases of deep-vein thrombosis, pulmonary embolus, stroke, and endometrial cancer were observed on the tamoxifen arm compared to the placebo arm. The frequency of events was consistent with the safety data observed in the NSABP P-1 trial.
Tamoxifen citrate tablets are indicated to reduce the incidence of breast cancer in women at high risk for breast cancer. This effect was shown in a study of 5 years planned duration with a median follow-up of 4.2 years. Twenty-five percent of the participants received drug for 5 years. The longer-term effects are not known. In this study, there was no impact of tamoxifen on overall or breast cancer-related mortality (see BOXED WARNING at the beginning of the label).
Tamoxifen citrate tablets are indicated only for high-risk women. "High risk" is defined as women at least 35 years of age with a 5 year predicted risk of breast cancer ≥ 1.67%, as calculated by the Gail Model.
Examples of combinations of factors predicting a 5 year risk ≥ 1.67% are:
Age 35 or older and any of the following combination of factors:
- One first degree relative with a history of breast cancer, 2 or more benign biopsies, and a history of a breast biopsy showing atypical hyperplasia; or
- At least 2 first degree relatives with a history of breast cancer, and a personal history of at least 1 breast biopsy; or
- LCIS
Age 40 or older and any of the following combination of factors:
- One first degree relative with a history of breast cancer, 2 or more benign biopsies, age at first live birth 25 or older, and age at menarche 11 or younger; or
- At least 2 first degree relatives with a history of breast cancer, and age at first live birth 19 or younger; or
- One first degree relative with a history of breast cancer, and a personal history of a breast biopsy showing atypical hyperplasia.
Age 45 or older and any of the following combination of factors:
- At least 2 first degree relatives with a history of breast cancer and age at first live birth 24 or younger; or
- One first degree relative with a history of breast cancer with a personal history of a benign breast biopsy, age at menarche 11 or less and age at first live birth 20 or more.
Age 50 or older and any of the following combination of factors: