Regular and Low-dose Aspirin, Other Non-steroidal Anti-inflammatory Medications and Prospective Risk of HER2-defined Breast Cancer

The California Teachers Study

Christina A. Clarke; Alison J. Canchola; Lisa M. Moy; Susan L. Neuhausen; Nadia T. Chung; James V. Lacey Jr; Leslie Bernstein


Breast Cancer Res. 2017;19(52) 

In This Article


Among CTS participants who reported their detailed, current use of NSAIDs and then were followed prospectively for a median of 7 years, the association between use of specific NSAIDs and the development of invasive breast cancer or its receptor-defined subtypes differed depending on the NSAID used. In the 23% of women who reported using low-dose aspirin at least three times per week, we observed a modest 20% reduction in risk of developing HR-positive/HER2-negative breast cancer, which is likely responsible for the similar association observed between NSAID use and risk of breast cancer overall. This association persisted after consideration of other breast cancer risk factors including use of HT and prior history of myocardial infarction. This association is intriguing because no such association was observed with the use of regular-dose aspirin (325 mg). We suspect that this could relate to the more regular use of low-dose aspirin for cardioprotection, as opposed to a more sporadic pattern of use of regular-dose aspirin to relieve pain. We did not observe any apparent modification of the low-dose aspirin effect by overweight status, nor did we observe stronger associations in women who were likely to be longer-term users, having reported using aspirin 10 years earlier, at baseline. This association should be re-examined in cohorts with larger numbers of incident breast cancers in which HR and HER2 status are recorded.

The three studies published previously that assessed HR-defined and HER-2-defined subtypes differed in the definitions of medication dose and duration of use, limiting our ability to compare the results of those studies to ours. The Nurses' Health Study[10] conducted the most comprehensive assessment of NSAID use in relation to receptor-defined subtypes, finding no association between breast cancer and use of non-aspirin NSAIDs or acetaminophen. They defined regular users of aspirin as those using two or more reduced tablets per week, but did not distinguish low-dose from regular-dose formulations. Across the four HR/HER2-defined subtypes, the analyses of which were based on fewer cases than in the present study, statistically significant protective effects against HR-positive/HER2-negative breast cancer were detected (n = 341 cases), regardless of the duration of reported use (10+ years of use: relative risk (RR) = 0.66, 95% CI 0.49–0.89, fewer than 10 years of use: RR = 0.75, 95% CI 0.58–0.96). They did not observe any association with the HR-positive/HER2-positive subtype (n = 74 cases, 10+ years of use: RR = 1.47, 95% CI 0.76–2.82; fewer than 10 years of use: RR = 1.40, 95% CI 0.79–2.51) or for the receptor-negative subtypes (n = 174 cases), or for breast cancer overall.[16,17] This pattern of association is similar to what we have shown here for CTS participants.

In the Nashville Breast Health Study,[12] a case-control study, the authors reported statistically significant protective effects of regular use of any NSAID against the risk of all receptor-defined subtypes, with a reduced odds ratio (OR) for HR-positive/HER2-negative cancer (OR = 0.71, 95% CI 0.56–0.88) similar to the HRR reported here; this was limited to overweight women with BMI of at least 25 kg/m2.[12] Our results did not suggest an interaction with overweight status. In the Western New York Exposures and Breast Cancer Study,[11] another case-control study, the authors reported a statistically significant reduction in risk of breast cancer overall associated with aspirin use, which did not persist for any subtype examined, including the four receptor-defined subtypes.

Our key finding is related to low-dose aspirin and not to regular-dose aspirin. Women who reported using low-dose aspirin were more likely to take it more than three times per week or daily, possibly for cardiovascular disease prevention. In our prior analysis of NSAID use reported at baseline by CTS participants and development of subsequent breast cancer,[18] we could not distinguish between low-dose and regular-dose aspirin as we did not inquire about dose in the baseline questionnaire. In that analysis we found no association between use of aspirin or ibuprofen more than once weekly and risk of breast cancer overall, but risk of HR-negative breast cancer was increased with 5 or more years of daily aspirin use (RR = 1.81, 95% CI 1.12–2.92).[18] In the current assessment, we did not see any association between breast cancer and use of regular-dose aspirin, nor did we detect any significant increase in risk of any breast cancer subtype with any aspirin use, regardless of dose. Our assessment of probable long-term users (women who used aspirin at baseline and at the 10-year follow-up) showed that these women had a similar risk of breast cancer overall and of the HR-positive/HER2-negative subtype as those who reported current use at the 10-year follow-up.

Our finding of an inverse association between use of 3+ low-dose aspirin tablets/week and risk of breast cancer overall is consistent with the findings of several other observational studies that did not separately examine HR-defined and HER2-defined subtypes. The largest and most detailed prospective study, the VITAL cohort,[19] found that women taking 81 mg low-dose aspirin for ≥4 days per week had a more pronounced reduction in risk of breast cancer overall (HRR = 0.65, 95% CI 0.43–0.97) after 10 years of follow up[20] than we observed here. The results from a large prospective study using the UK General Practice Database[21] showed a statistically significant decreased risk of breast cancer among women who took low-dose aspirin daily for at least 1 year (OR = 0.67, 95% CI 0.51–0.89), suggesting only a short duration of use was needed for a reduction in risk to become apparent.

In contrast to these studies, in the Women's Health Initiative observational study[22] risk of breast cancer overall was reduced 21% in women who took regular-dose aspirin but not in women who took low-dose aspirin. Our findings are not consistent with the Women's Health Study, a randomized clinical trial of the use of low-dose (100 mg) aspirin every other day, a frequency that would have mapped to the low end of our category defining regular use as at least three tablets weekly. After 10 years, the risk of breast cancer overall was unchanged (RR = 0.98, 95% CI 0.97–1.09, p = 0.68)[3] among the women randomized to low-dose aspirin. In a later sub-analysis the investigators reported no association according to breast tumor characteristics such as size, histology, grade, or differentiation.[17] Altogether, these studies support the notion that use of low-dose aspirin at least three times per week, or perhaps daily, modestly reduces overall breast cancer risk by about 20–25%. Our results add to this evidence base, suggesting that the reduction in risk occurs mainly in the HR-positive/HER2-negative subtype.

The biological mechanism by which low-dose aspirin could function as a chemopreventive agent against HR-positive/HER2-negative breast cancer, but not other breast cancer subtypes, is not yet clear. A consistent lowering of COX-2 and prostaglandin activity could prevent or slow carcinogenesis in a number of ways, at the tumor level, by interfering with DNA adduct formation,[23] inhibiting tumor angiogenesis,[4,24,25] or promoting apoptosis.[26] Recent data examining serum circulating inflammatory markers among healthy subjects ages 55–74 years did not indicate that regular-dose aspirin use is associated with any of 78 circulating markers, calling into question the relevance of circulating levels of immune markers[27] and raising the possibility of a more local immune effect. Alternatively, as prostaglandins may upregulate production of circulating estrogens via aromatase,[28,29] daily use of low-dose aspirin may inhibit aromatase, which could reduce levels of key hormones and thereby impact initiation or promotion of estrogen-sensitive tumors.[18,30,31] This analysis and that from the Nurses' Health Study[10] indicated protective effects of consistent aspirin use against the risk of HR-positive tumors, but only those that are also HER2-negative. On the other hand, elevated COX-2 levels have been detected in triple-negative tumors.[32]

Strengths of this study include the complete and accurate prospective ascertainment of HER2-defined breast cancer development based on routine linkage of the cohort to the statewide cancer registry, linkage to hospital discharge summary data to confirm any previous myocardial infarction, and a median of 7 years of follow up. Our study also had limitations. The total number of breast cancer cases was greater than was available in previously published analyses, but the limited numbers available for subtype-specific analyses meant that these were exploratory. It is possible that women who regularly take low-dose aspirin differed from women who did not, based on important health parameters (other than history of myocardial infarction or comorbid diabetes mellitus), resulting in residual and unmeasured confounding. For instance, women who regularly take low-dose aspirin could engage in more health-conscious behavior than non-users or infrequent-users. Like most observational studies, we were only able to measure NSAID exposure as "snapshots" of exposure at the time women completed their surveys. We were limited in evaluating possible confounding by indication for cardiovascular disease that did not result in hospitalization for myocardial infarction. Finally, this population is not representative of the general population of California women or women across the USA, particularly with respect to educational status; thus, it is uncertain how generalizable our findings are to the broader population, particularly non-white women and women born outside the USA. At cohort inception in 1996, CTS participants had incidence rates of breast cancer that were over 50% higher than those for age-matched and race-matched women in California,[14] which probably reflects higher prevalence of risk factors including hormone therapy use, alcohol consumption and particular reproductive profiles.

In summary, our study strongly supports the need for further, perhaps experimental, study of low-dose aspirin as a widely available, inexpensive chemopreventive option for the most common subtype of breast cancer, the HR-positive/HER2-negative subtype. Our study adds to the existing evidence on this topic, showing that previously reported associations between low-dose aspirin use and risk of breast cancer overall may be driven by a more specific association with this hormone-sensitive and HER2-negative breast cancer subtype. Furthermore, it suggests that previously reported associations using measures that combined low-dose aspirin use (more likely to be daily or more frequent) with regular aspirin use should reassess these associations. Future studies of aspirin and breast cancer must be able to distinguish low-dose from other formulations and to assess risks separately by molecularly defined subtype. Such studies should not only detail this chemopreventive potential but should also quantify any side effects associated with regular low-dose aspirin use.