Long-term use of tamoxifen reduces the risk of dementia: a nationwide population-based cohort study

Long-term use of tamoxifen reduces the risk

of

dementia: a nationwide population-based

cohort study

L.-M. Sun

*, H.-J. Chen

, J.-A. Liang

* and C.-H. Kao

Introduction

Breast cancer is the most frequent cancer among women worldwide.

1 Approximately two-thirds of patients with breast cancer

had a hormone-receptor positive disease, with an even larger proportion among elderly women.2 Tamoxifen, one of the most

widely used selective estrogen-receptor modulators for hormone-receptor-positive breast cancer, has been well documented to contribute to an increased life expectancy in patients with breast cancer.3 Although tamoxifen treatment is generally

well tolerated, with side effects primarily attributable to estrogen deprivation, it has adverse effects that are of clinical

concern.4

Dementia is a decline in mental ability that affects longterm thinking and social abilities severely enough to interfere

with a person’s daily functioning. The age-standardized prevalence of dementia for those aged _60 years varies between 5%

and 7% in most world regions. It has been estimated that 35.6 million people lived with dementia worldwide in 2010, with numbers expected to nearly double every 20 years, to 65.7 million in 2030 and 115.4 million in 2050.5 The increased

incidence of dementia in the aging population makes it a major public health problem. In Taiwan, the percentage of

aged persons (_65 years) in the total population has risen continually over the past three decades from 4.1% in 1980 to

6.1% in 1990, 8.6% in 2000 and 11.5% in 2013.6 The current estimated

prevalence of dementia in Taiwan is 5.7%.7 When very

Among known causes of dementia, drug etiology is underconsidered.9

Previous studies have suggested that estrogen may have an

effect on cognitive function in women,10,11 and studies in experimental

animal models have indicated that estrogens exert

neurotrophic and neuroprotective action in the brain and provide a convincing rationale for the role of estrogen replacement therapy in treating and preventing dementia.12,13 However, clinical

studies have yielded inconsistent results in its role in preventing Alzheimer’s disease (AD)14,15 Tamoxifen binds to the

estrogen receptor as a partial agonist or antagonist, depending on the target tissue.16 The potential for tamoxifen to impair cognitive

function and increase the risk of dementia has raised concern;17 however, studies have yielded inconclusive results. 16,18–20 To clarify this concern, we conducted a nationwide

population-based cohort study to determine the possible relationship between tamoxifen use and risk of dementia among female

patients with breast cancer in Taiwan.

Materials and methods

Data source

Taiwan National Health Insurance (NHI) is a nationwide, singlepayer insurance established in 1995, and has covered nearly

99% of Taiwanese citizens since 1998. The National Health Research Institutes (NHRI) compiles all reimbursement claim data to establish the National Health Insurance Research Database (NHIRD), and maintains the database. Each person in Taiwan has a unique personal identification number (PIN). The NHRI manages the claims data and provides scrambled random identifications of insured persons to ensure patient privacy. All NHI datasets can be interlinked with the PIN of each person. In this study, we used two datasets, namely, the Registry for Longitudinal Health Insurance Database 2000 (LHID2000) and the Registry for Catastrophic Illness Patient Database (RCIPD) of Taiwan. The NHIRD contains patient’s sex, date of birth and all records of clinical visits, including prescription details and diagnosed diseases coded according to the International Classification

of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM). This study was approved by the Institutional Review Board of China

Medical University (CMU-REC-101-012).

Study population

This study is a retrospective population-based cohort study. On the basis of the registry of RCIPD, 24 197 women patients age 20 years and older with newly diagnosed breast cancer (ICD-9-CM code 174), but without a previous diagnosis of dementia or any cancers during the 2000–04 period, were divided into two groups according to their tamoxifen use status. Among them, 16 556 patients used tamoxifen therapy and 7641 did not. The date of first receiving tamoxifen therapy was defined as the index date. For each patient with breast cancer included in the final group, four study women without dementia or any cancers were randomly selected from the registry of LHID2000 and frequency

matched by 5-year age interval and index year. A total of 96 788 subjects formed the cancer-free group.

Age was categorized into two levels: 20–54 and _55 (in years). Comorbidity was considered as a possible confounding factor. The comorbidities included diabetes (ICD-9-CM code 250), hyperlipidemia CM code 272), hypertension (ICD-9-CM code 401-405), coronary artery disease (CAD, ICD-9-(ICD-9-CM code 410-414), heart failure (HF, ICD-9-CM code 428), stroke (ICD-9-CM code 430-438) and head injury (ICD-9-(ICD-9-CM code 850-854 and 959.01) before the index date. Among patients with breast cancer, we also considered other potential confounding factors

including surgery, chemotherapy, radiotherapy, aromatase inhibitor therapy and benzodiazepine.

All patients were followed up until they were diagnosed with dementia (ICD-9-CM code 290, 294.1 and 331.0), or until they were censored because of loss to follow-up, death, withdrawal from the NHI system, or the end of 2011.

Statistics

Summary statistics are expressed as frequencies and percentages for categorical data and medians6(standard error, SE) for

continuous variables, as appropriate. The proportionate distributions of age group and comorbidity between the breast cancer

and cancer-free groups were compared and evaluated using the chi-squared test. The median ages were measured and evaluated using Wilcoxon rank-sum test, because the data were not

normally distributed. The incidence density rates of dementia were calculated for both groups (per 1000 person-years), and the

incidence density rate of dementia was calculated using the number of dementia incidents, divided by person-years at risk in both groups.

Extended Cox proportional hazards model with

Lunn–McNeil approach was performed, which is a modified Cox proportional hazards model that considers competing risks and allows for multivariable adjustment.21 Extended Cox proportional

hazard models with competing risk of death served to

evaluate the association between tamoxifen therapy and dementia incidence rates. We also performed age-stratified analyses to assess the association between tamoxifen therapy and the risk of dementia. We calculated hazard ratios (HRs) and 95% confidence intervals (CIs) after adjusting for age and comorbidity (including diabetes, hyperlipidemia, hypertension, CAD, HF, stroke and head injury). Furthermore, the extended Cox proportional hazards model, adjusting for variables including age,

comorbidity, use of benzodiazepine, operations on the breast, aromatase inhibitor, chemotherapy and radiotherapy, were used to evaluated the risk of dementia associated with tamoxifen therapy cumulative exposed year among patients with

breast cancer. Finally, we examined the combined effects of

tamoxifen therapy and an aromatase inhibitor on the risk of dementia in patients with breast cancer.

All P values originated from the two-sided tests, and the

statistical significance was set at P<0.05. All analyses were conducted using SAS Version 9.3 (SAS Institute Inc., Cary, NC).

Results

A total of 24 197 and 96 788 subjects were included in the breast cancer and cancer-free groups, respectively. Our analysis results of the demographics and comorbidities for the breast cancer and cancer-free groups are shown in Table 1. The median ages of the breast cancer and cancer-free groups were 49.58 years (SE¼0.07 years) and 49.49 years (SE¼0.04 years), respectively. Women with breast cancer had a high prevalence of diabetes (10.48% vs. 9.01%), and a low prevalence of stroke (1.59% vs. 1.84%) and head injury (2.83% vs. 3.69%). The mean follow-up

years were 7.41 years (SD¼3.02 years) in the breast cancer group and 8.42 years (SD¼2.12 years) in the cancer-free group. During an average follow-up of 8.22 years, 527 (2.18%) patients in the breast cancer group and 2549 (2.63%) patients in the cancer-free group developed dementia. Among patients with breast cancer, 160 (2.09%) tamoxifen users and 367 (2.22%) persons not using tamoxifen developed dementia. The incidence density rate of dementia was lower in the breast cancer

group than that in the cancer-free cohort (2.94 vs. 3.13 per 1000 person-years), with a HR of 0.95 (95% CI¼0.86–1.04) when adjusting for age and comorbidities (Table 2). Compared with

cancer-free patients, patients with breast cancer receiving and not receiving tamoxifen were not associated with the risk of dementia (adjusted HR¼0.93, 95% CI¼0.83–1.04 and 0.99,

0.84–1.16, respectively). Among patients with breast cancer, tamoxifen users were associated with the risk of dementia (adjusted HR¼0.83, 95% CI¼0.69–0.98). By stratifying age, we

observed that both younger (20–54 years) and older (_55 years) age groups showed a null association between breast cancer and dementia. Among patients with breast cancer, neither age group of tamoxifen users showed a significant difference in risk of dementia compared with their counterparts without tamoxifen therapy.

Among patients with breast cancer, those with _5 years of

tamoxifen therapy had a significantly lower risk of dementia compared with persons not using tamoxifen (adjusted

HR¼0.47, 95% CI¼0.32–0.69), as shown in Table 3. We observed that patients with both tamoxifen therapy and aromatase inhibitor therapy had a significantly lower risk of dementia than

those without tamoxifen therapy and aromatase inhibitor therapy (adjusted HR¼0.57, 95% CI¼0.44–0.73), as shown in Table 4.

Discussion

In this population-based cohort study, we observed that tamoxifen users had a significant 17% lower risk of dementia compared with persons not using tamoxifen among women

diagnosed with breast cancer. Tamoxifen use of 5 years or more significantly decreased the risk of dementia among patients with breast cancer. Furthermore, joint-effect analysis showed

that patients with breast cancer using both tamoxifen and aromatase inhibitor had a significantly lower risk of dementia. Our

data did not support developing dementia as a possible side effect of tamoxifen use.

Animal studies have established the role of estrogen in

regenerating and preserving neuronal elements within the central nervous system (CNS) that are analogous to those regions of the brain most sensitive to neurodegenerative changes associated with AD,12 the most common cause of dementia.22 Based

on a rat experiment, Wise suggested that estrogen therapy protects against brain-cell death by influencing the expression of

genes that suppress apoptotic cell-death pathways.13 Although

basic science studies have indicated the possible effects of hormone therapy on brain function, the clinical relevance of these

seemingly critical differences must be established for dementia and cognitive aging.23 The Women’s Health Initiative Memory

Study (WHIMS) conducted a double-blind controlled trial among women in the Women’s Health Initiative hormone therapy trials who were at least 65 years of age. They found that neither the estrogen-plus-progestogen trial nor the estrogen-alone trial showed the expected reductions in all-cause dementia. By contrast, the dementia rate increased in women allocated to active

treatment.23–25 However, epidemiological data support the efficacy

of early postmenopausal use of estrogens for delaying or preventing AD.26 Zhao et al.11 suggested that select estrogens

within the complex formulation of conjugated equine estrogens contribute to its neuroprotective efficacy.

Tamoxifen is the most commonly prescribed hormone therapy for patients with breast cancer.27 Based on studies suggesting

the effect of estrogens on brain functioning, and the

recognition that tamoxifen has differential tissue-dependent effects on estrogen-receptor function,19,28 whether it has estrogenic

or antiestrogenic qualities on brain tissue remains

unknown, and a large population-based study might clarify this uncertainty. Therefore, we performed this study to identify the possible relationship between tamoxifen use and the risk of dementia.

adverse and beneficial effects of tamoxifen on cognition and brain function.20,29,30 In postmenopausal women,

tamoxifen likely exhibits the most estrogen agonist CNS effects However, competition with estrogen likely exists in women of

reproductive age, resulting in antagonist effects.29 Our study

showed that a significantly lower risk of dementia for tamoxifen users among breast cancer patients, but the difference became non-significant if we stratified patients by age. This finding is partially consistent with a study from Breuer et al.31 who elucidated

a relationship of tamoxifen with a low prevalence of AD from a cross-sectional study on elderly nursing home residents. Other clinical studies have shown inconsistent results regarding tamoxifen use and dementia.18 A study from Denmark showed

no clinically relevant association between tamoxifen use and risk of dementia.16 A population-based case control study conducted

in Los Angeles County suggested that current use of tamoxifen may adversely affect cognition, but no differences were observed among people who have never used and past users of tamoxifen, suggesting limited long-term sequelae.20 In

addition, a small cross-sectional study suggested that ‘tamoxifen’ use in premenopausal women with breast cancer may exert a widespread negative influence on ‘cognitive’ abilities.32

Many patients with breast cancer might use anti-anxiety drugs, and early researchers suggested that benzodiazepine use may increase dementia risk.33,34 We adjusted benzodiazepine to

eliminate its possible confounding effect between the tamoxifen use and dementia risk. Our analyses indicated that longterm (_5 years) tamoxifen users were at a significantly lower risk of developing dementia compared with patients with breast cancer who were not prescribed tamoxifen. No similar finding has previously been reported that implies the protective role against dementia in long-term tamoxifen use. Patients treated with tamoxifen may have a longer life expectancy than breast cancer patients not treated with tamoxifen, and we took into account death as a competing risk. Besides the possible underlying pharmacologic mechanism of a protective role against dementia, we assume that long-term tamoxifen users may tend to

which could be helpful in maintenance of normal brain function. Although more than a 5-year prescription as an adjuvant therapy is uncommon in current clinical practice, researchers have found that extended adjuvant tamoxifen for breast cancer may produce a further reduction in recurrence and mortality.35

In addition to selective estrogen-receptor modulators, researchers are interested in exploring the possible relationship

between aromatase inhibitors and cognitive function, and the results are inconclusive.18,19 We found that tamoxifen and aromatase

inhibitors had a joint effect, and combined users exhibited a significantly decreased risk of dementia compared

with nonusers in patients with breast cancer. However, whether it is really a joint effect or just a chance finding remains undetermined. Chemotherapy may also affect cognitive function

in patients with breast cancer.36 We adjusted this factor in addition

to other treatments in the analyses to eliminate this possible confounding factor.

The strength of this study is the population-based nationwide database; however, several limitations should be considered in interpreting the results. First, selection bias might be present based on the assumption that patients diagnosed with breast cancer or treated with tamoxifen tend to be monitored closely, with opportunities for detecting dementia. However, our results revealed a null association between breast cancer and dementia, and showed an opposite effect among tamoxifen users. Second, information regarding the lifestyle or behavior of patients is lacking in the NHIRD; thus, adjusting for health behavior-related factors such as smoking and alcohol consumption,

which can increase the risk of breast cancer, was not

possible.37 Conversely, smoking is suggested to be a risk factor

for dementia, and moderate alcohol consumption tends to be protective against cognitive decline and dementia.38 Third, dementia

is more likely to occur in elderly or menopausal women,

and tamoxifen may exert various CNS effects for women of reproductive age and postmenopausal women29; however, menopause

information is not recorded in the NHIRD and we could

not precisely determine any distinct effects of tamoxifen on dementia between premenopausal and postmenopausal women.

We matched our cancer-free group to the breast cancer group according to age. In addition, we stratified by age and used 55 years as the cut-off age. We assumed that most women 55 years or older are menopausal. Despite the limitations of the administrative data, the information regarding the tamoxifen use of patients with breast cancer and breast cancer and dementia

diagnoses were highly reliable.

In summary, women diagnosed with breast cancer did not exhibit an association with developing subsequent dementia. Tamoxifen users exhibited a significantly lower risk of dementia compared with persons not using tamoxifen among women diagnosed with breast cancer, but the significant difference was limited to those with use of 5 years or more. Additional largescale studies are required to improve our understanding.