Breast cancer is by far the most common cancer in women worldwide, accounting for almost 20% of all malignancies.1 The age-standardized incidence rose by 40% from 1979 to 1992 in England and Wales.2 Use of antidepressant drugs is also quite common and has steadily increased in recent years, especially after the introduction of serotonin reuptake inhibitors (SSRIs), a new class of antidepressants with a perceived better safety profile.3,4 However, several laboratory studies have linked exposure to SSRIs and tricyclic antidepressants (TCA) with an increased risk of breast cancer or promotion of tumor growth.5 The suggested mechanism involves binding of TCAs and SSRIs to growth-regulatory intracellular histamine receptors, associated with antiestrogen-binding sites in microsomes and nuclei.
Several epidemiologic studies have evaluated the association between antidepressant drugs and breast cancer with conflicting results. Although an overall effect seems unlikely, these studies have not been able to rule out a direct association between increased risk of breast cancer and long-term use of TCAs or SSRIs.
Given the broad use of these drugs, the public health implications of this issue could be considerable. Furthermore, antidepressant drugs have been found to be prescribed to as many as 10% to 25% of oncology patients, underscoring the relevance of these findings.6,7
We conducted a cohort study with a secondary nested case-control analysis to assess the association between the risk of breast cancer and use of SSRIs, TCAs, and other antidepressants using the General Practice Research Database.
The General Practice Research Database contains computerized information entered by general practitioners in the United Kingdom.8 Data on over 2 million patients are systematically recorded and sent anonymously to the Medicines and Healthcare Products Regulatory Agency that collects and organizes this information for use in research projects. The computerized information includes demographic characteristics, details from general practitioners’ visits, diagnoses from specialist referrals and hospital admissions, results of laboratory tests, and a free-text section. Prescriptions issued by the general practitioner are directly generated from the computer. Several studies with the General Practice Research Database have documented the validity and completeness of this database,9 and some of them have explored the risk of certain outcomes among users of antidepressant drugs.10–12
The study protocol was approved by the General Practice Research Database Scientific Ethical and Advisory Group (SEAG).
We identified all women in the database who were 30–79 years old between January 1995 and December 2001. Women became members of the study population on the first day of the study period when they met the criteria of at least 1-year enrollment with the general practitioner and 1 year since the first computerized prescription, that date designated as their start date. Study members with a code for cancer before the start date were excluded. We also excluded women who were 70 years and older at the start date if they had a follow up greater than 1 year and no data recorded during their total follow-up time. Our final study cohort comprised 734,899 women.
All study members were followed from start date until the earliest occurrence of one of the following end points: recorded diagnosis of breast cancer, any cancer other than breast, age 80 years, death, or end of study period (December 2001).
Case Ascertainment and Validation
From our study cohort, we identified 4005 patients with a recorded code of breast cancer and manually reviewed their computerized patient profiles. Information included demographic data and all clinical information (with data on personal identifiers removed). Confirmed cases were patients who, after the first diagnosis, met at least one of the following criteria: additional breast cancer diagnosis recorded in subsequent visits, regular attendance at an oncology clinic, confirmatory biopsy, surgical intervention (mastectomy, wide local excision), or breast cancer drug therapy. We excluded 297 women; the main reasons were diagnosis of breast cancer not confirmed (60%) and prevalent cases (31%). We sent questionnaires to the general practitioners for validation of a random sample of 114 cases; of these, 108 (95%) questionnaires were received and all of them confirmed the computer diagnosis. In the end, 3708 patients were considered to be incident cases of breast cancer. The resulting overall incidence of breast cancer was 156 per 100,000 person-years in our population of 30- to 79-year-old women, with annual incidence rates of 80 per 100,00 women below 50 years of age, 247 per 100,000 between 50 and 70 years of age, and 282 per 100,000 above 70 years of age.
Cohort and Nested Case-Control Analyses
All incident cases of breast cancer (n = 3708) identified in the study cohort were used in the nested case–control analysis. We considered their date of initial diagnosis as the index date. For every member of the study cohort, a date during the study period was generated at random. If the random date of a study member was included in her eligible person-time, we used her random date as the index date and marked that woman as an eligible control. This selection mechanism allows the likelihood of being selected as a control to be proportional to the person-time at risk. The same eligibility criteria were applied to controls as to cases. From the list of all eligible controls, we frequency matched 20,000 controls by age (interval of 1 year) and calendar year.
We computed odds ratios (ORs), assumed to be valid estimates of the rate ratio, and 95% confidence intervals (CIs) associated with use of SSRIs, TCAs, and other antidepressants compared with nonuse using unconditional logistic regression. From the database, we collected information on history of depression (depression recorded more than 1 year before the index date) and subjects’ use of health services (visits to the general practitioner, specialist referrals, and hospital admissions) in the 2 years before the index date. We also ascertained patients with previous benign breast disease (breast lump or breast biopsy recorded more than 1 year before the index date). Information on other risk factors such as alcohol intake, body mass index (BMI), and use of nonsteroidal antiinflammatory drugs (NSAIDs) and hormone replacement therapy (HRT) were also ascertained. All odds ratios were adjusted for age, calendar year, depression, BMI, alcohol intake, HRT use, NSAID use, and prior benign breast disease.
We ascertained drug exposure of study subjects at any time before the index date (including medication use recorded before the start date of the study).
We studied 3 groups of antidepressant drugs: SSRIs, TCAs, and other antidepressants. Because women sometimes took more than 1 type of antidepressant, to obtain an unconfounded estimates for 1 group, we always adjusted for the other 2. We defined 3 time windows of exposure for each class of drugs: current use, past use, and no use. “Current use” was defined as use lasting until the index date or ending in the year before the index date based on the supply of drug therapy as prescribed by the general practitioner. “Past use” ended more than 1 year before the index date. Finally, the time window of “no use” was defined as no use of each respective drug group at any time before the index date. Current users were further subdivided into categories by treatment duration. Long-term use was defined as current use that lasted more than 1 year. We evaluated the effect of individual antidepressants as well as the effect of dose among long-term users. A secondary analysis using a 1-year time lag (advancing the index date by 1 year in cases and controls) was also performed.
The associations found between breast cancer and several potential risk factors are shown in Table 1. Higher BMI, alcohol consumption, prior benign breast disease, and HRT therapy were all associated with an increased risk of breast cancer. Smoking was not associated with an increased risk (OR = 0.92; 95% CI = 0.83–1.01).
The association between BMI and breast cancer was modified by age. Among women 55 years and older, those with a BMI of 30 or above had an OR of 1.39 (1.20–1.60) compared with women who had a BMI between 20 and 24. When we compared the same 2 groups among women younger than 55 years, the corresponding OR was 0.99 (0.83–1.18).
Antidepressant drug use was not associated with an increased risk of breast cancer (Table 2). Current users of SSRIs had an OR of 0.98 (0.81–1.19), whereas current users of TCAs had an OR of 0.86 (0.73–1.00).
Longer treatment duration was not associated with an increasing risk of breast cancer (Table 2). When only long-term use (greater than 1 year) was considered, the corresponding estimates for SSRIs and TCAs were 0.76 (0.53–1.09) and 0.87 (0.70–1.09), respectively.
We also evaluated the effect of dose among long-term use. Neither SSRI nor TCA high-dose users had an increased risk of breast cancer (Table 3). Furthermore, we assessed the association for the specific antidepressants most frequently prescribed. None of the individual drugs were associated with a clear increased or reduced risk of breast cancer.
Additionally, we assessed the effect of long-term antidepressant use among women age 55 years and older. Users of SSRIs and TCAs had ORs of 1.01 (0.64–1.61) and 0.85 (0.65–1.11), respectively.
Further adjustment for healthcare utilization (visits to the general practitioner, specialist referrals, and hospital admissions) did not materially modify these results. Also, the analysis using 1-year lag time yielded identical results.
The findings in our study do not support the hypothesis that antidepressants increase the risk of breast cancer. Although most previous studies assessing the association reported overall negative results,13–24 some of these studies found increased risks in subgroup analyses. Kelly et al.20 found a small elevation in risk among users of SSRIs. Cotterchio et al.21 found a modest elevated risk among long-term users of TCAs and a larger but imprecise risk among paroxetine users. Kato et al.,22 using the New York University Women's health study, found an elevated risk (relative risk [RR] = 1.75; 95% CI = 1.06–2.88) among postmenopausal women using any antidepressant, whereas Sharpe et al.,23 using data from Saskatchewan, reported an elevated risk associated with heavy exposure to TCAs. More recently, a study by Moorman et al.24 found that use of SSRI for 36 months or longer was possibly associated with an increased risk of breast cancer (OR = 2.2; 95% CI = 0.8–6.3).
We failed to detect any increased risk associated with either SSRIs or TCAs in our study. The same results were observed regardless of treatment duration or dose category. Data for chronic use were more suggestive of a reduced risk of breast cancer risk than with an increased risk. Unfortunately, we did not have much information on long-term therapy (more than 5 years). Similarly, our findings do not support an increased risk among individual antidepressants. Paroxetine, previously reported as presenting an increased risk, was not associated with breast cancer in our study (0.76, 040–1.45). To evaluate a potential effect in postmenopausal women, we evaluated the effect of antidepressants among women 55 years and older; we did not observe an increased risk associated with antidepressant use in this subgroup of women.
We evaluated documented risk factors for breast cancer as potential confounders. We found modest risk elevations associated with alcohol intake, BMI, and HRT therapy, consistent with existing literature.25 The largest published study looking at the relationship between BMI and breast cancer found that BMI was a risk factor for breast cancer among postmenopausal but not among premenopausal women,26 and our data show the same pattern. This modification of the effect of BMI is most likely the result of the fact that adipose tissue is an important extragonadal source of endogenous estrogens in postmenopausal women.27 Prior benign breast disease has also been found to approximately double the risk of breast cancer in prior studies, in agreement with our results.25
We recorded information regarding the most relevant risk factors for breast cancer, including age, alcohol use, BMI, HRT use, and previous benign breast disease. There are some potential confounders that we were not able to ascertain through our database, including age at menarche, parity, family history, age at first child, and germline mutation. However, it seems highly unlikely that this limitation of our study would have affected the results. Indeed, one would expect the distribution of these potential confounders to be rather evenly distributed among users of various antidepressants as well as among nonusers.
Exposure ascertainment was based on computerized prescription data, and some patients will either not fill the prescription or not take the pills once they are dispensed (although this is less likely with chronic treatment). Patients who did not take their medicines would be incorrectly classified as users, and this would tend to dilute any effect of antidepressant drugs on breast cancer, biasing the results toward the null.
Detection bias could occur if patients with regular health contacts are more likely to be screened and therefore to be subsequently detected with early breast cancer lesions. Because patients with a greater use of health services also share a greater likelihood of receiving antidepressant medication, this could result into a spurious association between breast cancer and antidepressants. To control for this potential bias, we included patients’ health services utilization in the multivariate model; the main results remained unchanged.
This study provides additional evidence to exclude with confidence a clinically relevant increased risk of breast cancer associated with antidepressant use. We did not observe any increased risk among antidepressant users, regardless of daily dose or duration of use, and we did not find any individual antidepressant drug to be clearly associated with an increased risk. Overall, results from most previous observational studies are consistent with this finding. Multiple testing with small numbers could be one of the explanations for the elevated risk observed in a few instances among subgroups of users in some of these studies. Considering that these drugs are among the most widely used by women, the results of this study provide some reassurance from a public health perspective.
We thank the staff at General Practice Research Database and the participating general practitioners for their collaboration. We also thank the Boston Collaborative Drug Surveillance Program for providing access to the database.
1. McPherson K, Steel CM, Dixon JM. ABC of breast diseases. Breast cancer epidemiology, risk factors, and genetics. BMJ
2. Quinn M, Allen E. Changes in incidence of and mortality from breast cancer in England and Wales since introduction of screening. United Kingdom Association of Cancer Registries. BMJ
3. McManus P, Mant A, Mitchell PB, et al. Recent trends in the use of antidepressant drugs in Australia, 1990–1998. Med J Aust
4. Martin RM, Hilton SR, Kerry SM, et al. General practitioners’ perceptions of the tolerability of antidepressant drugs: a comparison of selective serotonin reuptake inhibitors and tricyclic antidepressants. BMJ
5. Brandes LJ, Arron RJ, Bogdanovic RP, et al. Stimulation of malignant growth in rodents by antidepressant drugs at clinically relevant doses. Cancer Res
6. Jaeger H, Morrow GR, Carpenter PJ, et al. A survey of psychotropic drug utilization by patients with advanced neoplastic disease. Gen Hosp Psychiatry
7. Cullivan R, Crown J, Walsh N. The use of psychotropic medication in patients referred to a psycho-oncology service. Psychooncology
8. García Rodríguez LA, Pérez Gutthann S. Use of the UK General Practice Research Database for pharmacoepidemiology. Br J Clin Pharmacol
9. Jick H, Jick SS, Derby LE. Validation of information recorded on general practitioner based computerised data resource in the United Kingdom. BMJ
10. De Abajo FJ, García Rodríguez LA, Montero D. Association between selective serotonin reuptake inhibitors and upper gastrointestinal bleeding: population based case–control study. BMJ
11. Meier CR, Schlienger RG, Jick H. Use of selective serotonin reuptake inhibitors and risk of developing first-time acute myocardial infarction, Br J Clin Pharmacol
12. De Abajo FJ, Jick H, Derby LE, et al. Intracranial hemorrhage and use of selective serotonin reuptake inhibitors, Br J Clin Pharmacol
13. Friedman GD, Habel LA, Boles M, et al. Kaiser Permanent Medical Care Program: Division of Research, Northern California, and Centre for Health Research, Northwest Division. In: Strom BL, ed. Pharmacoepidemiology
. Chichester: John Wiley & Sons, Ltd; 2000:263–283.
14. Van Den Eeden SK, Friedman GD. Prescription drug screening for possible carcinogenicity: eleven to fifteen years of follow-up. Pharmacoepidemiol Drug Safety
15. Selby JV, Friedman GD, Fireman BH. Screening prescription drugs for possible carcinogenicity: eleven to fifteen years of follow-up. Cancer Res
16. Friedman GD, Ury HK. Screening for possible drug carcinogenicity: second report of findings. J Natl Cancer Inst
17. Friedman GD, Ury HK. Initial screening for carcinogenicity of commonly used drugs. J Natl Cancer Inst
18. Wang PS, Walker AM, Tsuang MT, et al. Antidepressant use and the risk of breast cancer: a non-association. J Clin Epidemiol
19. Danielson DA, Jick H, Hunter JR, et al. Nonestrogenic drugs and breast cancer. Am J Epidemiol
20. Kelly JP, Rosenberg L, Palmer JR, et al. Risk of breast cancer according to use of antidepressants, phenothiazines, and antihistamines. Am J Epidemiol
21. Cotterchio M, Kreiger N, Darlington G, et al. Antidepressant medication use and breast cancer risk. Am J Epidemiol
22. Kato I, Zeleniuch-Jacquotte A, Toniolo PG, et al. Psychotropic medication use and risk of hormone-related cancers: the New York University Women's Health Study. J Public Health Med
23. Sharpe CR, Collet J-P, Belzile E, et al. The effects of tricyclic antidepressants on breast cancer risk. Br J Cancer
24. Moorman PG, Grubber JM, Millikan RC, et al. Antidepressant medications and their association with invasive breast cancer and carcinoma in situ of the breast. Epidemiology
25. Singletary SE. Rating the risk factors for breast cancer. Ann Surg
26. Tretli S. Height and weight in relation to breast cancer morbidity and mortality. A prospective study of 570,000 women in Norway. Int J Cancer
27. Clemons M, Goss P. Estrogen and the risk of breast cancer. N Engl J Med