Annals of Surgical Oncology

, Volume 15, Issue 9, pp 2526–2532

Menopausal Hormone Therapy and Breast Cancer Phenotype: Does Dose Matter?

Authors

    • Department of SurgeryUniversity of California San Francisco
  • Anjali S. Kumar
    • Department of SurgeryKaiser Permanente Oakland Medical Center
  • Veronica Shim
    • Department of SurgeryKaiser Permanente Oakland Medical Center
Breast Oncology

DOI: 10.1245/s10434-008-0019-7

Cite this article as:
Garwood, E.R., Kumar, A.S. & Shim, V. Ann Surg Oncol (2008) 15: 2526. doi:10.1245/s10434-008-0019-7

Abstract

Background

Duration and type of menopausal hormone therapy (HT) has been associated with increased breast cancer risk and the development of estrogen receptor (ER)-positive tumors. The effect of HT dose on breast cancer tumor characteristics remains undefined. We sought to determine if HT dosing regimens influence breast cancer phenotype.

Methods

We conducted a retrospective review of incident female breast cancers occurring in the year 2003 listed in the Kaiser Permanente Northern California Cancer Registry. Type of HT, dose, number of tablets dispensed, tumor phenotype, stage, grade, and histology were obtained from electronic records for women aged ≥50 years who had more than 1 year of uninterrupted pharmacy data (= 1701). A dose index of HT exposure was created and odds ratios were used to determine if tumor phenotype varied between exposure groups. These results were compared with a previously published analysis of HT duration on tumor phenotype conducted with the same dataset.

Results

The cumulative effect of estrogen and progesterone hormone therapy as calculated by factoring both dose and duration of HT use prior to breast cancer diagnosis did not reveal any new associations that were not previously identified by analysis of HT duration of exposure alone. Low-dose-index combination-HT users were less likely to have tumors with an ER-positive phenotype. An overall trend developed in which low- and high-dose-index exposed women had the lowest rates of ER- and progesterone receptor (PR) -positive tumors.

Conclusion

Duration of use is an adequate surrogate for determining overall exposure to HT when considering the effect of HT on breast cancer phenotype.

Keywords

Breast CancerHormone replacement therapyHormone therapyEstrogen receptorHistologyMenopause

Menopausal hormone therapy (HT) has been implicated both in the development of breast cancer and, more controversially, in determining hormone receptor status and other characteristics of tumor aggression. The year 2003 saw a 6.7% decline of incident breast cancer cases in women over the age of 50 years in the USA.1,2 This decline in breast cancers paralleled the timing of a nationwide reduction in menopausal HT use, catalyzed by the Women’s Health Initiative reports of 2002. These reports indicated that combination HT confers an increased risk of cardiovascular events and breast cancers. An immediate 38% reduction in HT use was observed within 1 year (mostly combination HT and estrogen-only HT).3 More controversial is the hypothesis that breast cancers developing while exposed to hormone administration may be more likely to elaborate hormone receptor positive tumors and develop less aggressive tumor characteristics.4 Interestingly, estrogen receptor (ER)-positive cancers experienced the sharpest decline in 2003: 14.7%, versus a 1.7% decrease in ER-negative tumors among women aged 50–69 years.2 This selective decline in ER-positive tumors proportional to the decrease in HT use rekindles interest in unraveling the interactions between hormone therapy exposure and host tumor characteristics.

Greater risk of developing breast cancer has been observed in women exposed to HT for long durations.5 However, these women were less likely to die from their disease.6 This observation has given rise to a number of studies examining the relationship between HT exposure and tumor characteristics to determine what differences exist in this group of tumors arising under the influence of exogenous HT.

Previous studies of the relationship between HT exposure and tumor characteristics including grade, stage, nodal status, and hormone receptor phenotype have yielded conflicting results, likely because hormonal influences on breast cancer seem to be modulated by multiple host characteristics, most notably determinants of endogenous hormonal levels (age and menopausal status), and multiple exposure characteristics (HT type, duration, and dosing). Although four published studies have observed a significant relationship between HT or combination-HT use and a predilection towards receptor-positive tumor development, the majority of published case–control and cohort studies do not.710 Notably, the Women’s Health Initiative (WHI) randomized trial found no difference between tumor grade, histologic type (ductal versus lobular) or receptor phenotype (ER and PR) between HT and placebo groups. Invasive breast cancers associated with estrogen plus progestin use were larger, more likely to be node positive, and were diagnosed at a significantly more advanced stage.3 One study found that women currently exposed to HT (estrogen alone or combination) were more likely to have high S-phase cancers and PR-positive disease.6 Endogenous sex hormone levels measured in the plasma of post-menopausal women are associated with increasing breast cancer risk in a dose–response relationship but it is unknown how hormone therapy dose may affect breast cancer receptor status or histology.11 The Million Women study found higher rates of lobular and tubular breast carcinoma histology occurring among current hormone therapy users as compared with those who had never used the therapy, indicating that hormone therapy may be more closely associated with the development of lobular and tubular cancer histologies than with ductal.12

To date, the relationship between dose of exogenous hormone replacement therapy and breast cancer characteristics has not been examined, although duration of HT use has been studied extensively. In a retrospective cohort analysis of incident cases of breast cancer during 2003, our group found that prolonged (>6 months) use of combination HT versus estrogen or progesterone-only HT, increased the odds of presenting with ER-positive tumors (odds ratio [OR] 1.65; 95% confidence interval [CI] 1.07–2.5) and that users of estrogen only HT were statistically more likely to present with low-grade and early-stage tumors.13

Moving forward in defining these relationships, it is important to determine if our current surrogate measure of HT exposure, duration of use, is an adequate marker of cumulative exogenous exposure as estrogen and progesterone HT dosing regimens vary widely. In our dataset, prescribed courses ranged from 0.02 to 10 mg per day for estrogen and from 0.3 to 200 mg per day for progesterone, such that high-dose HT may provide more than 500–650 times greater drug exposure than low-dose HT. As HT dosing is not uniform, we hypothesize that a dose index (summation of tablet dose multiplied by the number of tablets dispensed) may be a more useful marker than duration of use when examining the influence of HT use on breast cancer. Standard-dose PremarinTM sales (0.625 mg/day) have been decreasing since the WHI reports, however, the sales of low-dose Premarin preparations (0.45 mg/day) continue to rise.14 It is important to understand how exogenous hormone dose (HT) may interact with breast cancer phenotype and histology when a post-menopausal woman develops breast cancer while exposed to HT.

Methods

We conducted a retrospective review on incident cases of female breast malignancy occurring in the year 2003 listed in the Kaiser Permanente Northern California Cancer Registry (n = 2,830) as previously described. This study was approved by the Kaiser Permanente, Northern California, Institutional Review Board. Type of HT, dose, number of tablets dispensed, tumor receptor phenotype, stage, grade, and histology were obtained from electronic records for women over 50 years old who had continuous pharmacy data for more than one year (n = 1,701).13

HT Use

The type of HT prescribed was recorded from the electronic pharmacy logs. Patients classified as taking combination HT were simultaneously prescribed separate preparations of estrogen and progesterone, as the pharmacy does not carry single formulations of combination estrogen and progesterone. Additionally, to be classified as “combination HT” the total estrogen and progesterone tablet count could not differ by >100 tablets. All tablets were prescribed to be taken once per day, which we used as a measure of duration and compliance. From these data, a dose index was generated by multiplying the number of tablets prescribed by the dose of each tablet. The index values for estrogen and progesterone tablets were divided into four categories: none (0), low (1–199 for estrogen, 1–999 for progesterone), moderate (200–999 for estrogen, 1,000–2,850 for progesterone), and high (>1,000 for estrogen, >2,900 for progesterone). The cutoffs for these categories were assigned to divide women with a wide range of dose index exposure values roughly into thirds. This assignment strategy, as opposed to using quartiles, was employed to maximize the power to detect a difference in tumor phenotype between exposure categories by allowing a greater number of patients per exposure category.

For combination-HT use, patients must have been prescribed both estrogen and progesterone with <100 tablet difference between the dispensed medication. Figure 1 displays the allocation of combination-HT dose-index categories and the distribution of patients within this framework. Combination HT (C-HT) use was divided into three dose-index categories: the low C-HT dose index category, which included low progesterone/low estrogen or low progesterone/moderate estrogen users and moderate progesterone/low estrogen users (n = 153); moderate C-HT, comprised entirely of patients taking moderate estrogen/moderate progesterone (n = 89); and high C-HT, including high progesterone/moderate estrogen or high progesterone/high estrogen (n = 61) users.
https://static-content.springer.com/image/art%3A10.1245%2Fs10434-008-0019-7/MediaObjects/10434_2008_19_Fig1_HTML.gif
Fig. 1

Scatterplot of estrogen HT dose index versus progesterone HT dose index in women prescribed comparable amounts of estrogen and progesterone, demonstrating the allocation of combination hormone therapy dose-index categories.

Receptor Phenotype and Histology

Estrogen and progesterone receptor phenotype was documented in the data registry as the ratio of cells staining positive for estrogen and progesterone receptors versus those unstained. Tumors with less than 5% nuclear staining were considered negative for the respective antibody. All tumor specimens from this registry are read by a team of five pathologists at a single pathology laboratory.

Pathology determined tumor grade based on histologic features and was coded in the registry as the highest grade present: I (low grade), II (moderate grade), and III (high grade). Stage was recorded according to the Surveillance, Epidemiology, and End Results (SEER) guidelines: in situ carcinoma (IS), localized malignancy (LOC), regional malignancy–lymph node metastasis (REG), distant metastasis (DIS).

Statistical Analyses

Summary statistics for continuous variables include mean, standard deviation (SD), median, minimum and maximum value; categorical variables are presented as counts and percentages. Standard baseline characteristics are summarized for each group. Logistic regression was used to estimate the effects of predictor variables (age and HT use) on dichotomous outcome variables (estrogen and progesterone receptor status). Stata version 9 (College Station,TX) was used to perform all analyses and prepare all figures.

Results

The analysis included 1,701 women over 50 years of age. Patient demographic information and tumor characteristics are shown in Table 1A, B. Patient age at the time of primary breast tumor diagnosis ranged from 50 to 95.7 years (mean ± standard deviation 66 ± 10.1 years). Fifty-five percentage of women were taking any type of HT, whereas 45% were not. Among the 951 women on any HT, 30% were prescribed estrogen only, 4% progesterone only, and 66% were prescribed any combination of estrogen and progesterone. Of the HT population, 70% used any progesterone with a mean dose index of 1,875 and a wide range of 12.5–195,300. Of HT users, 97% were prescribed any estrogen. A wide range of dose indices was also seen amongst estrogen users (mean 599, range 3.75–3,962). Although 66% of women over 50 years old who were prescribed HT received some combination of estrogen and progesterone, only 303 (32%) received combination HT, defined as comparable use of both estrogen and progesterone (within 100 tablets).
Table 1a

Patient demographics

 

n

Age, mean (range), years

66.2 (50–96)

 

n (%)

Hormone therapy (HT) use

    None

759 (45)

    Any HT

942 (55)

    Any estrogen HT

911 (97)

    Any progesterone HT

664 (70)

    Combination HT

303 (32)

Race/ethnicity

    Caucasian

1338 (79)

    African American

111 (7)

    Hispanic

98 (6)

    Asian

88 (5)

    Filipino

47 (3)

Table 1b

Disease characteristics

 

n (%)

Hormone receptor status

    ER positive

1406 (83)

    PR positive

1106 (65)

Tumor stage

    In situ

181 (11)

    Localized

1019 (60)

    Regional

437 (26)

    Distant

48 (3)

    Unknown

16 (1)

Age-adjusted odds ratios for ER- and PR-positive tumors in women with low, medium, and high dose indices for estrogen exposure, progesterone exposure, and combination HT were compared with those who were not exposed to HT. These results were compared with our previous analysis performed on the same dataset that allocated HT exposure categories based on duration alone. The odds of developing an ER-positive cancer while taking various types of hormone therapy are outlined in Fig. 2A–C and Tables 2A–C. The odds of developing a PR-positive cancer are outlined in Fig. 3A–C and Tables 3A–C. Tumors with an ER-positive phenotype were distributed evenly among women at any dose index level of estrogen or progesterone menopausal hormone therapy as compared with women not exposed to hormone therapy. Women with a moderate-dose-index exposure to combination hormone therapy had significantly increased odds of having ER-positive tumor.
https://static-content.springer.com/image/art%3A10.1245%2Fs10434-008-0019-7/MediaObjects/10434_2008_19_Fig2_HTML.gif
Fig. 2

Age-adjusted odds ratios of having an ER-positive tumor among women currently taking HT when HT is categorized by dose index (black circles) versus duration only (gray triangles): (A) any estrogen, (B) any progesterone, or (C) combination HT. The reference population is women who were not taking any HT.

Table 2a

Age-adjusted odds ratios of having an ER-positive tumor among women currently taking any estrogen HT when HT is categorized by dose index versus duration only. The reference population is women who were not taking any HT

HT type

 

n

% ER+

OR of ER+ tumor (95% CI)

P value

None (ref.)

 

759

84%

1.0

 

Any estrogen HT

Hormone therapy dose index

 

  Low

234

80%

0.85 (0.58, 1.23)

0.40

  Moderate

512

84%

1.06 (0.78, 1.44)

0.69

  High

165

79%

0.81 (0.53, 1.24)

0.34

Hormone therapy duration

 

  Low

106

73%

0.56 (0.35, 0.89)

0.02

  Moderate

722

84%

1.09 (0.86, 1.45)

0.52

  High

76

75%

0.61 (0.35, 1.06)

0.80

Bold indicates statiscally significant P value.

Table 2b

Age-adjusted odds ratios of having an ER-positive tumor among women currently taking any progesterone HT when HT is categorized by dose index versus duration only. The reference population is women who were not taking any HT

HT type

 

n

% ER+

OR of ER+ tumor (95% CI)

P value

None (Ref)

 

759

84%

1.0

 

Any progesterone HT

Hormone therapy dose index

 

  Low

180

80%

0.86 (0.57, 1.31)

0.48

  Moderate

253

85%

1.24 (0.84, 1.57)

0.28

  High

231

83%

1.05 (0.71, 1.57)

0.79

Hormone therapy duration

 

  Low

150

77%

0.76 (0.49, 1.17)

0.21

  Moderate

490

84%

1.15 (0.84, 1.57)

0.39

  High

20

90%

1.89 (0.43, 8.25)

0.40

Table 2c

Age-adjusted odds ratios of having an ER-positive tumor among women currently taking combination HT when HT is categorized by dose index versus duration only. The reference population is women who were not taking any HT

HT type

 

n

% ER+

OR of ER+ tumor (95% CI)

P value

None (ref.)

 

759

84%

1.0

 

Combination HT

Hormone therapy dose index

 

  Low

153

80%

0.87 (0.56, 1.36)

0.55

  Moderate

89

93%

3.05 (1.30, 7.17)

0.01

  High

61

80%

0.87 (0.45, 1.70)

0.69

Hormone therapy duration

 

  Low

72

71%

0.52 (0.30, 0.89)

0.02

  Moderate

223

88%

1.57 (1.00, 1.05)

0.05

  High

7

100%

N/A

N/A

Bold indicates statiscally significant P value.

https://static-content.springer.com/image/art%3A10.1245%2Fs10434-008-0019-7/MediaObjects/10434_2008_19_Fig3_HTML.gif
Fig. 3

Age-adjusted odds ratios of having a PR-positive tumor among women currently taking HT when HT is categorized by dose index (black circles) versus duration only (gray triangles): (A) any estrogen, (B) any progesterone, or (C) combination HT. The reference population is women who were not taking any HT.

Table 3a

Age-adjusted odds ratios of having a PR-positive tumor among women currently taking any estrogen HT when HT is categorized by dose index versus duration only. The reference population is women who were not taking any HT

HT type

 

n

% PR+

OR of PR+ tumor (95% CI)

P value

None (ref.)

 

759

68%

1.0

 

Any estrogen HT

Hormone therapy dose index

 

  Low

234

58%

0.67 (0.50, 0.91)

0.01

  Moderate

512

63%

0.86 (0.68, 1.09)

0.20

  High

165

62%

0.85 (0.59, 1.21)

0.36

Hormone therapy duration

 

  Low

106

51%

0.51 (0.34, 0.77)

<0.01

  Moderate

722

64%

0.88 (0.71, 1.09)

0.25

  High

76

55%

0.60 (0.37, 0.97)

0.04

Bold indicates statiscally significant P value.

Table 3b

Age-adjusted odds ratios of having a PR-positive tumor among women currently taking any progesterone HT when HT is categorized by dose index versus duration only. The reference population is women who were not taking any HT

HT type

 

n

% PR+

OR of PR+ tumor (95% CI)

P value

None (ref.)

 

759

68%

1.0

 

Any progesterone HT

Hormone therapy dose index

 

  Low

180

53%

0.59 (0.42, 0.82)

<0.01

  Moderate

253

65%

0.97 (0.71, 1.32)

0.84

  High

231

62%

0.85 (0.62, 1.15)

0.29

Hormone therapy duration

 

  Low

150

52%

0.55 (0.38, 0.79)

<0.01

  Moderate

490

64%

0.87 (0.69, 1.11)

0.28

  High

20

75%

1.49 (0.53, 4.14)

0.45

Bold indicates statiscally significant P value.

Table 3c

Age-adjusted odds ratios of having a PR-positive tumor among women currently taking combination HT when HT is categorized by dose index versus duration only. The reference population is women who were not taking any HT

HT Type

 

n

% PR+

OR of PR+ tumor (95% CI)

P value

None (ref.)

 

759

68%

1.0

 

Combination HT

Hormone therapy dose index

 

  Low

153

56%

0.65 (0.46, 0.93)

0.02

  Moderate

89

75%

1.56 (0.94, 2.59)

0.09

  High

61

62%

0.89 (0.52, 1.55)

0.69

Hormone therapy duration

 

  Low

72

47%

0.44 (0.27, 0.73)

<0.01

  Moderate

223

69%

1.10 (0.79, 1.52)

0.58

  High

7

71%

1.20 (0.23, 6.24)

0.83

Bold indicates statiscally significant P value.

Investigation of PR receptor phenotype showed that women with a low-dose-index exposure to either estrogen HT, progesterone HT, or combination HT were less likely to have PR-positive tumors (P = 0.01 for estrogen, P < 0.01 for progesterone, and P = 0.02 for C-HT). There was an overall trend toward a U-shaped distribution, with women exposed to low-dose-index estrogen or progesterone and high-dose-index estrogen or progesterone having the lowest rates of ER- and PR-positive tumors. These findings echoed the HT duration analysis that we previously reported in which women exposed to estrogen or progesterone for less than 6 months experienced a significant reduction in PR-positive tumor occurrence [P < 0.01 (OR 0.51, 95% CI 0.34–0.77) for estrogen and P < 0.01 (OR 0.55, 95% CI 0.38–0.79) for progesterone]. As in our previous duration-only analysis, we did not find any significant differences in ductal versus lobular tumor type among low-, moderate-, and high-dose-index estrogen or progesterone exposed women. Ductal cancers predominated, comprising 68% of tumors in women taking any estrogen, 70% in those taking any progesterone, and 69% in those taking combination HT, with no statistically significant differences in tumor type occurring between dose-index categories when compared with those not exposed to hormone therapy.

Discussion

In light of the recent studies regarding HT and breast cancer risk, particularly ER-positive breast cancer, the specific effects of HT on breast cancer histology and estrogen receptor status have been scrutinized. It is important to know whether the cumulative effect of high exposure to estrogen or progesterone HT affects breast cancer type. When examining the association between hormone therapy, breast cancer development, and characteristics of these cancers that arise in women taking hormone therapy, our analyses indicate that factoring in dose does not provide any further insight than consideration of duration of exposure alone. This is an important finding given the wide variations in hormone therapy doses prescribed.

In vitro, breast cancer cells respond to hormone exposure. Estrogen receptor-α is a transcription factor that regulates genes responsible for the maintenance of estrogen-responsive tissues. Gene expression of estrogen receptor-α is modifiable by a number of compounds, including hormones and growth factors. Experiments in nonreproductive estrogen-responsive tissues demonstrate upregulation of ER expression in response to estradiol exposure.15 MCF-7 human breast adenocarcinoma cell lines exposed to estradiol increase their mRNA expression of ER but decrease ER protein.15 The relationship is likely complex and exactly how HT interacts with breast cancer ER expression remains unknown. Breast cancers that elaborate estrogen receptors may originate from the small population of “normally” ER-positive human breast ductal epithelial cells at higher risk for malignancy due to their proliferative response to estrogen,7 or various agents may induce estrogen receptor expression in breast tissue, expanding the population of cells at risk.7 It is also possible that HT exposure in itself does not modulate breast cancer receptor status, but selectively promotes the growth of pre-existing ER-positive occult cancers.6

One strength of this study lies in the electronic pharmacy database, which allowed for precise calculation of total exposure to the duration and dose of HT use, which was accurately reflected in the number of prescriptions dispensed, as few Kaiser patients fill prescriptions outside the Kaiser formulary system. This produces a more accurate estimate of total HT exposure than self-reporting surveys and provides an advantage over randomized trials, since total time on HT is known rather than estimated, which allows us the opportunity to explore the association between HT use and its impact on breast cancer hormonal phenotypes.

Currently, many routes of hormone therapy administration exist including oral, parenteral, and transdermal, all with different potencies and pharmacokinetic profiles. Route of HT administration is a variable that remains to be thoroughly investigated in the context of breast cancer. As nearly all of the women in our dataset were prescribed oral hormone therapy, route was not addressed in this analysis.

Conclusion

We analyzed the cumulative effect of exogenous estrogen and progesterone as calculated by factoring both dose and duration of HT use prior to breast cancer diagnosis and compared these results with our previous analysis of HT duration alone. The addition of dose information to this calculation did not reveal any new associations. We conclude that duration of use is an adequate surrogate for determining overall exposure to HT in the context of breast cancer development.

Copyright information

© Society of Surgical Oncology 2008