Thyroid hormone and breast carcinoma
Primary hypothyroidism is associated with a reduced incidence of primary breast carcinoma
To investigate the role of primary hypothyroidism (HYPT) on breast carcinogenesis, the authors evaluated 1) the association between HYPT and a diagnosis of invasive breast carcinoma and 2) the clinicopathologic characteristics of breast carcinoma in patients with HYPT.
For this retrospective chart review study, 1136 women with primary breast carcinoma (PBC) were identified from the authors' departmental data base. These women (cases) were frequency-matched for age (± 5 years) and ethnicity with 1088 healthy participants (controls) who attended a breast carcinona screening clinic. Women with HYPT who were receiving thyroid-replacement therapy before they were diagnosed with breast carcinoma or before the screening visit were identified.
The mean ages of cases and controls (51.6 years vs. 51.0 years, respectively; P = 0.30) and their menopausal status (65.4% premenopausal vs. 62% postmenopausal; P = 0.10) were comparable. Two hundred forty-two women in the case group (10.9%) with HYPT were identified. The prevalence of this condition was significantly greater the control group compared with the case group (14.9% vs. 7.0%, respectively; P < 0.001). PBC patients were 57% less likely to have HYPT compared with their healthy counterparts (odds ratio, 0.43l 95% confidence interval, 0.33–0.57). Seventy-eight white patients with PBC had HYPT and, compared with women who were euthyroid, they were older at the time of diagnosis (58.8 years vs. 51.1 years; P < 0.001), were more likely to have localized disease (95.0% vs. 85.9% clinical T1 or T2 disease, respectively; P = 0.025), and were more likely to have no pathologic lymph node involvement (62.8% vs. 54.4%; P = 0.15).
Primary HYPT was associated with a reduced risk for PBC and a more indolent invasive disease. These data suggest a possible biologic role for thyroid hormone in the etiology of breast carcinoma and indicate areas of research for the prevention and treatment of breast carcinoma. Cancer 2005. © 2005 American Cancer Society.
Hypothyroidism (HYPT) is a clinical entity that results from a deficiency of thyroid hormones or, more rarely, from impaired biologic activity at the tissue level.1 It is a common condition, with a prevalence of 1.9% in women, and the prevalence increases with age. HYPT can be congenital or acquired, primary or secondary, chronic or transient.1 Primary HYPT is caused by disease or treatment that destroys the thyroid gland or interferes with thyroid hormone biosynthesis in the thyroid gland. Autoimmune thyroiditis is the predominant cause of primary HYPT in Western countries, in which iodine intake generally is adequate or even high, whereas severe iodine deficiency is the major cause in iodine-deplete areas. Another cause of primary HYPT, chronic or transient, is previous radioactive iodine or surgical treatment of HYPT or thyroid tumors, including carcinoma.
Since Beatson2 described the use of thyroid extracts to treat patients with metastatic breast carcinoma more than a century ago, many studies have investigated the influence of thyroid hormones (triiodothyronine [T3] and prohormone thyroxin [T4]) on this malignancy. In preclinical models, it has been found that T3 is able to sustain serum-free proliferation of several cell lines, including breast carcinoma cells.3, 4 In rodents, mammary gland development and physiology are sensitive to T3.5
The issue relative to possible associations between disease of the thyroid gland and breast carcinoma has been debated for decades and remains controversial. Epidemiologic studies have provided conflicting results regarding the association between either thyroid disorders or treatment for thyroid disorders and breast carcinoma risk.6–9 In the 1950s, clinical investigators reported that, although breast carcinoma seldom occurred in hyperthyroid women, it seemed to occur more frequently than expected in hypothyroid women.10 The cause of this relation was not known. Itoh and Maruchi suggested in 1975 that Japanese women with Hashimoto thyroiditis had a higher incidence of breast carcinoma compared with women who did not have thyroid disorders.11 Those data were confirmed a few years later by European investigators.12 Conversely, Smith et al.13 conducted a series of studies investigating the relation between thyroid volume, the presence of thyroid-peroxidase autoantibodies (TPO.Ab), and the possible prognostic implications in patients with invasive breast carcinoma and breast disorders. Survival analysis in a group of 142 women with breast carcinoma demonstrated that high levels of TPO.Ab were associated with a significantly longer disease-free survival and overall survival compared with women who were TPO.Ab negative, suggesting a protective role.13
Overall, these studies have suggested a clinical but yet unrecognized biologic correlation between thyroid disease (mainly HYPT) and invasive breast carcinoma risk. We performed a retrospective, hospital-based, case–control study with the objective of defining the association between primary HYPT and breast carcinoma risk. We also sought to describe the patient characteristics and pathologic features of invasive breast carcinoma in women with primary HYPT compared with women who were not affected by this condition.
MATERIALS AND METHODS
This was a hospital-based, case–control study. The study population was comprised of 2226 women who were evaluated for breast carcinoma or who underwent their screening at the University of Texas M. D. Anderson Cancer Center between January 1997 and December 2000. The 1136 cases were comprised of a consecutive series of women with newly diagnosed, primary invasive breast carcinoma (PBC) who had their initial assessment at the Nellie B. Connally Breast Center of M. D. Anderson Cancer Center between July 1997 and April 1998. These women were identified from the breast cancer clinical database. The 1088 controls were identified among women without a history of invasive carcinoma or preinvasive breast disease (e.g., ductal carcinoma in situ) who attended the breast cancer prevention and screening clinic between January 1997 and December 2000. The controls were frequency-matched with cases based on age (± 5 years) and ethnicity. The study was approved by the Institutional Review Board.
Using a standardized data abstraction form, trained study personnel extracted from the medical records detailed medical history information for all participants. In addition, the clinical and pathologic characteristics of all cases were abstracted. Two sources of information were available in the medical records. The first was an information sheet that was completed by the patient at the time of first clinical encounter associated with the most recent screening or first diagnostic visit. This form included information regarding the patient's medical history, comorbid conditions, and active medications. The second information source was a detailed clinical evaluation of women that was completed by medical personnel (e.g., physician, physician assistant). Data collection particularly was detailed concerning recognized breast carcinoma risk factors in the study population, including 1) first-degree relatives diagnosed with invasive breast carcinoma (genetic factors), 2) the use of hormone-replacement therapy (HRT) at the time of the interview (environmental factors), and 3) menopausal status.
Patients with clinically relevant thyroid disorders were classified as hyperactive or hypoactive (hypothyroidism) hyperactive or hypoactive hypothyroidism (HYPT) by the medical personnel based on information from the medical chart. HYPT was classified as either primary or secondary according to a description of contributing factors (e.g., previous thyroid ablative surgery or radioactive treatment or a spontaneous [possibly autoimmune] condition). Only women with thyroid disorders who were diagnosed before a breast carcinoma diagnosis (cases) or prior to attending the screening clinic (controls) and who used thyroid supplements (e.g., levothyroxine or equivalents) were recorded as “positive.” In addition, all cases had self-reported symptoms of HYPT, and laboratory confirmation by determination of serum thyroid-stimulating hormone (TSH) and free-T4 was not available from the medical records. This retrospective analysis did not allow for determination of the incidence of subclinical HYPT in the study population. Women with hyperthyroid conditions and subsequent hypothyroid disease were excluded from the study analysis
Age and tumor size were analyzed as continuous variables. The Student t test was used to evaluate differences between groups. Chi-square tests were used to compare distributions of categorical variables (e.g., family history, ethnicity) among cases and controls. Odds ratios (ORs) and corresponding 95% confidence intervals (95% CIs) were computed using SPSS (version 9.0; SPSS, Inc., Chicago, IL) unconditional logistic regression models. Multivariate logistic regression models were developed to allow the evaluation of the effects of HYPT with simultaneous adjustment for know breast carcinoma risk factors (e.g., family history of breast carcinoma, history of pregnancy, HRT use, and menopausal status).
Patient characteristics for the 2226 participants are provided in Table 1. The majority of women were white (77.8% of cases vs. 76.7% of controls); the remainder of the population consisted of Hispanics (11.2% of cases vs. 10.8% of controls) and African Americans (11.0% of cases vs. 12.9% of controls). The mean ages of cases and controls, as expected, did not differ significantly (51.6 years vs. 51.0 years, respectively; P = 0.3), and menopausal status also was similar (65.4% vs. 62% postmenopausal, respectively; P = 0.11). The comparison between the 2 groups revealed that women who had PBC were less likely to report HRT use compared with women who did not have a diagnosis of invasive disease (34.2% vs. 44%, respectively; P < 0.001) or a positive family history of breast carcinoma among first-degree relatives (17.2% vs. 24.8%, respectively; P < 0.001). Instead, they were more likely to have a history of pregnancy (89.9% vs. 83.4%, respectively; P < 0.001), but there were no significant differences between cases and controls with regard to age of menarche (12.7 ± 1.5 years vs. 12.7 ± 1.6 years, respectively; P = 0.6) and age of first pregnancy (data not shown). Furthermore, there were no differences noted with regard to body mass index (BMI) between the two groups (Table 1).
Table 1. Clinical Characteristics of the Study Population
|Ethnicity|| || || |
| White||884 (77.8)||830 (76.3)|| |
| African American||127 (11.2)||118 (10.8)|| |
| Hispanic||125 (11.0)||140 (12.9)||0.4a|
|Mean age ± SD (yrs)||51.6 ± 12.6||51.0 ± 11.6||0.3b|
| White||52.4 ± 12.7||51.5 ± 11.7||0.2b|
| African American||49.2 ± 11.6||50.6 ± 10.4||0.6b|
| Hispanic||48.7 ± 12.1||47.9 ± 11.5||0.3b|
|Menopausal status|| || || |
| Premenopausal||392 (34.6)||411 (37.8)|| |
| Postmenopausal||741 (65.4)||675 (62.2)||0.1a|
|First-degree relative with breast carcinoma|| || || |
| Yes||195 (17.2)||252 (24.8)|| |
| No||941 (82.8)||766 (75.2)||< 0.001a|
|History of HRT|| || || |
| Yes||388 (34.2)||465 (44.0)|| |
| No||748 (65.8)||593 (56.0)||< 0.001a|
|History of pregnancy|| || || |
| Yes||954 (89.9)||851 (83.4)|| |
| No||107 (10.1)||169 (16.6)||< 0.001a|
|BMI (kg/m2)|| || || |
| Underweight (< 18)|| 10 (1.1)|| 8 (0.9)|| |
| Normal (18–24.9)||356 (40.9)||362 (41.4)|| |
| Overweight (25–29.9)||290 (33.3)||263 (30.1)|| |
| Obese (≥ 30)||214 (24.6)||242 (27.7)||0.36a|
In total, 272 women with history of thyroid disorder were identified among the 2226 participants (Table 2). The vast majority of women with thyroid disorders (242 women; 10.9%) had been diagnosed previously with symptomatic, primary HYPT. The remaining 30 women who were excluded from the analysis included women with Hashimoto thyroiditis (n = 10 women), benign adenoma (n = 9 women), asymptomatic multinodular goiter (n = 5 women), and HYPT (n = 6 women). All patients with primary HYPT were receiving thyroid replacement therapy. The prevalence of HYPT was significantly higher in the control group compared with the case group (14.9% vs. 7.0%, respectively; P < 0.001) (Table 2). The current results indicate that patients with PBC are 57% less likely to report a history of primary HYPT compared with healthy women (crude OR, 0.43; 95% CI, 0.33–0.57]. Using multivariate logistic regression analysis to adjust for known confounders of breast carcinoma risk (i.e., family history of breast carcinoma, history of pregnancy, menopausal status, age, and HRT use), the association between PBC and HYPT essentially was unchanged and remained a strong protective factor against a diagnosis of invasive breast carcinoma (adjusted OR, 0.44; 95% CI, 0.32–0.60). These data strongly suggest that primary HYPT may have an independent association with breast carcinoma risk. This association was similar in all ethnic groups investigated, although the number of cases with HYPT was very small among African-American and Hispanic women (Table 2). White women had a higher prevalence of HYPT compared with African-American and Hispanic women combined both in the control group (16.4% vs. 10.1%, respectively; P = 0.001) and in the case group (8.4% vs. 2.4%, respectively; P = 0.013).
Table 2. Prevalence of Primary Hypothyroidism in the Study Population
|Overall|| || || || |
| Hypothyroidism||80 (7.0)||162 (14.9)|| || |
| Normal thyroid||1056 (93.0)||926 (85.1)||0.44 (0.32–0.60)||< 0.001|
|White|| || || || |
| Hypothyroidism||74 (8.4)||136 (16.4)|| || |
| Normal thyroid||810 (91.6)||694 (83.6)||0.47 (0.34–0.65)||< 0.001|
|African American|| || || || |
| Hypothyroidism||1 (0.8)||16 (11.4)|| || |
| Normal thyroid||124 (99.2)||124 (88.6)||0.08 (0.01–0.68)||0.02|
|Hispanic|| || || || |
| Hypothyroidism||5 (3.9)||10 (8.5)|| || |
| Normal thyroid||122 (96.1)||108 (91.5)||0.30 (0.08–1.07)||0.06|
We restricted our analysis to the 1136 women with PBC (cases) to explore further the relation between HYPT and the clinicopathologic characteristics of invasive breast carcinoma (Tables 3, 4). These patients were stratified based on the diagnosis of self-reported primary hypothyroidism (HYPT positive vs. HYPT negative). Eighty patients had a history of HYPT, and clinicopathologic data were available for 78 of those patients (74 patients were white). Compared with patients who had no diagnosed thyroid disease, patients with HYPT were older at diagnosis (58.8 years vs. 51.1 years, respectively; P < 0.001) and were more likely to be postmenopausal (82% vs. 64.1%, respectively; P = 0.001). We then restricted the comparison to white women (92.5% of the HYPT-positive cases) to have a more homogeneous study population and to eliminate the influence of ethnicity (Table 4). The current analysis demonstrated that women with HYPT were more likely to be diagnosed with early-stage disease (AJCC, pathologic Stage I and II: 95.0% vs. 85.9%, respectively; P = 0.025) and without pathologic lymph node involvement (63.9% vs. 55.9%, respectively; P = 0.15), although this distinction was not significant. Moreover, having a history of HYPT also was associated with smaller pathologic tumor size, because most patients with HYPT (72.5%) had tumors that measured ≤ 2.0 cm (T1) compared with 55% of patients who were not treated for HYPT (P = 0.002).
Table 3. Clinical Comparison of Women with Primary Breast Carcinoma with and without Hypothyroidism
|Mean age ± SD (yrs)||58.8 ± 12.5||51.1 ± 12.4||< 0.001b|
|Ethnicity|| || || |
| White||74 (92.5)||810 (76.7)|| |
| Hispanic||5 (6.3)||122 (11.6)|| |
| African American||1 (1.3)||124 (11.7)||0.003|
|First-degree relative with breast carcinoma|| || || |
| Yes||66 (82.5)||875 (82.9)|| |
| No||14 (17.5)||181 (17.1)||0.9|
|History of HRT|| || || |
| Yes||37 (46.3)||711 (67.3)|| |
| No||43 (53.8)||345 (32.7)||< 0.001|
|Menopausal status|| || || |
| Premenopausal||14 (17.7)||378 (35.9)|| |
| Postmenopausal||65 (82.3)||676 (64.1)||0.001|
Table 4. Clinical Comparison of White Women with Breast Carcinoma with and without Hypothyroidism
|Mean age ± SD (yrs)||51.6 ± 12.6||58.8 ± 12.5||51.1 ± 12.4|| ||< 0.001a|
|Pathologic stage|| || || || || |
| 0/I||374 (42.8)||39 (53.4)||335 (41.8)||1.0|| |
| II||394 (45.1)||31 (42.5)||363 (45.3)||0.7 (0.4–1.1)||0.16|
| III||106 (12.1)||3 (4.1)||103 (12.9)||0.3 (0.1–0.8)||0.01b|
|Pathologic lymph node involvement|| || || || || |
| Yes||376 (43.4)||26 (36.1)||350 (44.1)||1.0|| |
| No||490 (56.6)||46 (63.9)||444 (55.9)||0.7 (0.4–1.1)||0.15|
|Mean tumor size ± SD||2.44 ± 1.9||2.09 ± 1.74||2.47 ± 1.91|| ||0.1a|
| ≤ 2.0 cm||466 (58.3)||50 (72.5)||416 (57.0)||1.0|| |
| > 2.0 cm||333 (41.7)||19 (27.5)||314 (43.0)||0.4 (0.2–0.7)||0.002|
|Estrogen receptor status|| || || || || |
| Positive||522 (69.2)||49 (79.0)||473 (68.4)||1.0|| |
| Negative||232 (30.8)||13 (21.0)||219 (31.6)||0.6 (0.3–1.1)||0.08|
|Progesterone receptor status|| || || || || |
| Positive||609 (64.2)||43 (69.4)||441 (63.7)||1.0|| |
| Negative||270 (35.8)||19 (30.6)||251 (36.3)||0.7 (0.4–1.2)||0.2|
HYPT is a common disorder with an incidence of approximately 3–4% of symptomatic disease in the general population that increases to up to 13–14% among individuals age ≥ 65 years.14 This retrospective study found a prevalence of HYPT in women with breast carcinoma comparable to the values reported in the general population. The results of this study demonstrated that women with a diagnosis of primary, symptomatic HYPT who are on thyroid supplementation are less likely to be diagnosed with invasive breast carcinoma. These results were demonstrated when the women were compared with a control group of women who had a clearly higher incidence of recognized genetic and environmental risk factors for breast carcinoma, as represented by a positive family history and use of HRT.15, 16 It is interesting to note that, when the confounding effects of these factors were adjusted for by multivariate logistic regression analysis, the protective effect remained significant, suggesting that HYPT may be an independent risk factor. Patients with primary HYPT had a 61% reduction in the risk of developing invasive disease. The subsequent analysis of clinical and pathologic characteristics of women with invasive disease further strengthens the possibility of a significant role of thyroid hormones on breast carcinoma biology. In fact, patients with a concomitant diagnosis of invasive disease and hypothyroid condition may have a more indolent disease, because they were significantly more likely to have pathologically smaller tumors. The data suggest that lymph node involvement and hormone receptor-negative tumors (e.g., estrogen receptor negative) were also less frequent among HYPT-positive patients, although these differences were not statistically significant, probably because of the small number of patients.
There are several limitations but also some strengths of this study that need to be mentioned. The retrospective nature of the investigation and the definition of cases with HYPT based on reported information on medical records, theoretically, may introduce misclassification bias that may affect the results. However, because data for both cases and controls were collected similarly by the same study personnel from medical charts at the same institution, we expect that the bias introduced would be nondifferential and, thus, would bias the observed odds ratio toward 1.0 (the null hypothesis).
Previous, large, prospective studies have failed to demonstrate clearly any correlation between any benign thyroid condition and the risk of breast carcinoma.7–9 To our knowledge, the largest and most complete study published to date that addressed the association between thyroid disorder and breast carcinoma was reported by Simon et al.7 In that population-based, case–control study, 4575 women (cases) with breast carcinoma (2953 whites and 1622 African Americans) and 4682 controls (3021 whites and 1661 African American) between the ages of 35–64 years were interviewed. The conclusions from Simon et al. were in line with previous retrospective reports but did not clarify the interaction between the occurrence of clinical and/or subclinical hypothyroid disease and the risk of breast carcinoma.14 In fact, they defined thyroid disease to include several different clinical conditions (e.g., Grave disease, goiter, nodules, hypoactive thyroid, hyperactive thyroid), although the hormonal milieu associated with each of those conditions is not equivalent. In fact, approximately 11% of women with PBC reported a history of HYPT, and 13% were receiving thyroid replacement therapy. It is interesting to note that, with regard to the breast carcinoma risk factors, the two study groups were quite similar. They differed significantly only in the occurrence of invasive disease in first-degree relatives (favoring the cases). In another study by Altieri et al.,9 the control population was represented by age-matched women who attended the general medicine clinics for acute and nonneoplastic conditions, making their conclusions more difficult to compare with our study. That study did not demonstrate a statistically significant association between any benign thyroid condition and the risk of breast carcinoma.
The observed association between hypothyroidism and breast carcinoma may be due to the biologic effect of T3 at the cellular level through either a direct interaction with the thyroid receptor (TR), a member of the nuclear receptor family, or a modulation of the thyroid-stimulating hormone receptor (TSH-R). Another intriguing possible protective role may be ascribed to iodide based on its antioxidant mechanism.17, 18 This theory has been postulated based on the capacity of breast tissue to transport and concentrate iodide.18 This latter observation requires more detailed confirmatory studies.
A pathophysiologic role of HYPT is supported by in vivo data indicating a role of thyroid hormones in mammary gland development by stimulating ductal branching and alveolar budding.5 Moreover, it has been found that hypothyroid mice have reduced levels of epidermal growth factors receptors compared with euthyroid mice.19 The role of TR and TSH-R on these regulatory mechanisms has not been clarified to date. Normal mammary epithelial cells express significant amounts of TR,20, 21 and breast carcinoma cells contain similar levels of T3-binding activity.22–25 To our knowledge, until recently, little was known regarding how T3 and TSH act on mammary epithelial cells, but several investigations suggested that the 2 hormones, particularly T3, may stimulate proliferation in some experimental models and may inhibit cell growth in others.26–29 Moreover, TR and the nuclear hormone receptors RXR and estrogen receptor can modulate each other's transcriptional activities. In addition, the expression of TSH-R has been described in breast tissue of rodents.32 In animal models, TSH-R appears to be down-regulated during pregnancy, indirectly suggesting a possible protective biologic role in the mammary gland. Taken together, these data indicate that T3 and TSH may play a significant role, which has not been recognized completely, in mammary gland development, and these actions occur mainly through action on TR and possibly indirectly through the modulation of TSH-R.
The results of this retrospective study should prompt prospective investigations directed at demonstrating the differential levels of expression of TR and TSH-R in primary breast carcinoma in euthyroid and hypothyroid conditions and correlations with the expression and function of the estrogen receptor. In addition, the role of intracellular iodide concentration in the various clinical conditions should be planned. These studies potentially may indicate areas of intervention for targeted preventive and therapeutic purposes.