Secondary Logo

Journal Logo

Advanced Search
Pharmacoepidemiology

Low-dose Aspirin, Nonsteroidal Anti-inflammatory Drugs, Selective COX-2 Inhibitors and Breast Cancer Recurrence

Cronin-Fenton, Deirdre P.; Heide-Jørgensen, Uffe; Ahern, Thomas P.; Lash, Timothy L.; Christiansen, Peer; Ejlertsen, Bent; Sørensen, Henrik T.

Author Information

From the aDepartment of Clinical Epidemiology, Aarhus University, Aarhus, Denmark; bDepartments of Surgery and Biochemistry, University of Vermont, Burlington, VT; cDepartment of Epidemiology, Rollins School of Public Health, Emory University, Atlanta, GA; dBreast and Endocrine Section, Department of Surgery, Aarhus University Hospital, Aarhus, Denmark; eDanish Breast Cancer Cooperative Group, Copenhagen, Denmark; and fRigshospitalet, Copenhagen, Denmark.

Submitted 22 June 2015; accepted 21 March 2016.

The study was supported by grants from the Danish Cancer Society (R73-A4284-13-S17) (H.T.S.); the Aarhus University Research Foundation (H.T.S.); Program for Clinical Research Infrastructure (PROCRIN) established by the Lundbeck Foundation and the Novo Nordisk Foundation (H.T.S.); the Elvira & Rasmus Riisforts Fonden (D.C.F.); the Lundbeck Foundation (R167-2013–15861) (D.C.F.); Susan G. Komen for the Cure (CCR 13264024) (T.P.A.); the Mary Kay Foundation (003-14) (T.P.A.); and the US National Cancer Institute at the National Institutes of Health (R01CA166825) (T.L.L., D.C.F.). The funding agencies had no role in the design of the study; the collection, analysis, and interpretation of the data; the writing of the article; or the decision to submit the article for publication.

D.C.F., H.T.S. conceived the study idea, and along with all authors, designed the study. B.E. and P.C. collected the data. D.C.F. reviewed the literature. D.C.F., U.H.J. directed the analyses, which were carried out by U.H.J. All authors participated in the discussion and interpretation of the results. D.C.F. organized the writing and wrote the initial drafts. All authors critically revised the manuscript for intellectual content and approved the final version. H.T.S. is the guarantor.

The authors report no conflicts of interest. However, the Department of Clinical Epidemiology is involved in studies with funding from various companies as research grants to (and administered by) Aarhus University. None of these studies has any relation to the present study.

Supplemental digital content is available through direct URL citations in the HTML and PDF versions of this article (www.epidem.com).

Correspondence: Deirdre P. Cronin-Fenton, Department of Clinical Epidemiology, Aarhus University, Olof Palmes Alle 43–45, 8200 Aarhus N, Denmark. E-mail: [email protected].

doi: 10.1097/EDE.0000000000000480

Abstract

Breast cancer accounts for 23% of female cancers and is the most frequent cause of cancer-related death among women worldwide.1 Aspirin, nonsteroidal anti-inflammatory drugs (NSAIDs), and selective COX-2 inhibitors, hereafter referred to as COX-2 inhibitors, are effective analgesic, antipyretic, and anti-inflammatory drugs. They exert pleiotropic effects, including the prevention of cardiovascular disease2,3 and cancer,4 although the latter may be confined to subpopulations.5,6 Aspirin, NSAIDs, and COX-2 inhibitors target the cyclooxygenase enzymes, COX-1 and COX-2.7,8 These enzymes promote angiogenesis and prevent apoptosis.9 COX-1 is ubiquitously expressed, but COX-2 is expressed only during inflammation and in cancer.10 Elevated COX-2 and its products, prostaglandins, correlate with shorter breast cancer survival.11–15 Cell line and animal model research suggests that NSAIDs, aspirin, and COX-2 inhibitors impede breast cancer cell growth.9,16,17

Findings from epidemiologic studies of the association of these drugs with breast cancer prognosis are inconsistent.18–27 At least 3 years of prediagnostic aspirin use has been associated with a 20% reduction in breast cancer mortality.23 Postdiagnostic aspirin use (both low- and high-dose) has been associated with a 50% reduction in breast cancer recurrence/mortality in some,18,28 but not all studies.19,21–25 Research on the effect of NSAIDs on breast cancer recurrence or mortality also report conflicting findings.20–22,29,30 Decreased recurrence has been correlated with perioperative use of ketorolac29,30 and use of any NSAID.20,21 However, protective associations have not been found in all studies,22 and there is no evidence of a dose–response effect.20

Ongoing clinical trials are investigating potential survival benefits associated with use of pre- and postoperative COX-2 inhibitors in cancer patients (www.clinicaltrials.gov). A trial of 37 breast cancer patients suggested that preoperative use of the COX-2 inhibitor celecoxib decreased Ki67 expression, a marker of tumor proliferation rate, but did not evaluate patient survival.31 No observational study has investigated the association between COX-2 inhibitors and survival outcomes in breast cancer patients.21

The discordant findings regarding the association of aspirin or NSAIDs with outcomes in breast cancer patients may stem from differences in the measured confounders, or variation in the methods used to assess drug exposure, i.e., self-reported versus prescription based. Many studies investigated mortality18–22,25 rather than disease recurrence.18,21 Given the established protective effect of low-dose aspirin against cardiovascular disease,32 assessing mortality rather than recurrence mixes the potential effect of the drug on cancer recurrence—which predisposes patients to mortality—with its direct effect on mortality. Furthermore, several published aspirin studies did not adjust for statins,18,20–22,24,28,33,34 which are frequently prescribed with aspirin to prevent cardiovascular disease, and have been consistently associated with better survival in breast cancer patients.35,36

To address the limitations of previous research, we conducted a large population-based study with prospectively collected data to investigate the association of aspirin (primarily low-dose aspirin, as high-dose aspirin is rarely used in Denmark), NSAIDs, and COX-2 inhibitors with the rate of breast cancer recurrence among breast cancer survivors in Denmark.

METHODS

This study was approved by the Danish Data Protection Agency (Record 2012-41-0793), the Danish Medicines Agency, and the Danish Breast Cancer Cooperative Group (DBCG). This study is based on routinely collected registry data, therefore separate ethical approval was not necessary.

Source Population and Data Collection

We included all women in Denmark diagnosed with incident invasive nonmetastatic breast cancer between 1996 and 2008 registered in the DBCG (n = 34,188). Since its establishment in 1976, the DBCG has registered most cases of invasive breast cancer in Denmark.37 Registration completeness has increased over time from 87% in 1986 to 96% in 1997.38,39 Prespecified data on tumor, treatment, and patient characteristics are prospectively obtained from treating physicians. DBCG patients with operable disease undergo follow-up exams to detect recurrences twice-yearly for the first 5 years and annually up to 10 years after diagnosis. Follow-up exams include a clinical evaluation and, if indicated, a chest x-ray, computed tomographic scan, bone scan, or other investigation to detect disease recurrence.40 Patients who develop recurrent disease between follow-up exams are also reported to the DBCG. We used the DBCG registry to ascertain data on age, menopausal status at diagnosis, world health organization histological tumor type and grade, lymph node status, tumor estrogen receptor (ER) status, type of primary surgery (mastectomy or breast-conserving surgery), chemotherapy, radiation therapy, endocrine therapy (ET), and date and anatomical site of recurrence. We used the Danish Civil Registration System to retrieve information on mortality and emigration.41

The Danish National Prescription Registry (NPR), maintained by Statistics Denmark, has recorded all prescriptions dispensed at Danish pharmacies since 1995. Recorded data include the redemption date, prescribed drug (classified by Anatomical Therapeutic Chemical [ATC] codes), and fill quantity.42 Via Statistics Denmark, we linked prescription data from the NPR to the clinical cohort using the civil personal registration number, a unique personal identification number assigned to each Danish citizen at birth or emigration.43

We used the NPR to ascertain information on prescriptions for aspirin (low dose as B01AC06 and N02BA01 in tablet sizes of 75, 100, or 150 mg, and high dose as N02BA51 and N02BA01 in tablet sizes of 500 mg), NSAIDs, and COX-2 inhibitors, as well as potentially confounding comedications including one or more prescriptions for simvastatin, angiotensin-converting enzyme inhibitors (ACE inhibitors), beta-adrenergic blocking drugs (beta-blockers), and postmenopausal hormone replacement therapy (HRT; eAppendix 1; https://links.lww.com/EDE/B38).

We obtained data on comorbid diseases from the Danish National Patient Registry, which contains data on all nonpsychiatric hospital admissions since 1977 and on outpatient hospital contacts since 1995.44 At discharge, the Danish National Patient Registry records the civil personal registration number, admission, and discharge dates, and up to 20 discharge diagnoses. We ascertained prevalent comorbid diseases at index date, including rheumatoid arthritis, osteoarthritis, diabetes, cancer, liver disease, arrhythmia, angina, peripheral and cerebral vascular disease, myocardial infarction, and congestive heart failure (eAppendix 2; https://links.lww.com/EDE/B38).

Analytic Variables

Age at diagnosis was treated as a continuous variable in the multivariable Cox regression models. Person-time at risk for recurrence was defined as the number of days between the date of primary surgery and the date of breast cancer recurrence, death, emigration, or 1 January 2013, whichever occurred first. We used the DBCG definition of breast cancer recurrence as any local, regional, or distant recurrence, or contralateral breast cancer.40 Histological grade was classified as low, moderate, or high. We defined surgery as either mastectomy or breast-conserving surgery with radiation therapy. Adjuvant chemotherapy was treated as a dichotomous variable. We summarized ER and ET as a joint variable (ER+/ET+, ER+/ET−, ER−/ET+, ER−/ET−).

The exposure drugs—aspirin, NSAIDs, and COX-2 inhibitors—were modeled as time-dependent post-diagnosis exposures to at least one prescription, updated daily during follow-up and lagged by 1 year. We also modeled exclusive use of each of the exposure drugs (i.e., aspirin but not NSAIDs or COX-2 inhibitors; NSAIDs, but not aspirin or COX-2 inhibitors; and COX-2 inhibitors, but not aspirin or NSAIDs), and use of any of the three drug types as a combined category. In the latter analyses, all patients were included in these models and the reference category represented women who had not used any of the drug types. In sensitivity analyses, we changed the definition of exposure from filling ≥1 to filling ≥2 prescriptions during follow-up, and changed the lag time from one to 2 years.45

We assessed a potential dose–response effect of the exposure drugs on recurrence according to the number of prescriptions for each drug over the entire follow-up period. We assessed the mean number of days per prescription and then estimated 1 year supply for each drug. A 1 year supply equated to three prescriptions for aspirin, five prescriptions for NSAIDs, and six prescriptions for COX-2 inhibitors.

We modeled the association of prediagnostic aspirin use, and, in post-hoc analyses, modeled the association of prediagnostic use of NSAIDs and COX-2 inhibitors with the rate of breast cancer recurrence. For these analyses, we restricted the study cohort to women with at least 2 years of prescription history at the time of diagnosis (n = 32,024). Prediagnostic use of each drug was defined as redemption of one or more prescriptions before breast cancer diagnosis.

Prescriptions for potentially confounding drugs—simvastatin, ACE inhibitors, and beta-blockers—were modeled as time-varying covariates lagged by 1 year. HRT was modeled as a baseline covariate.18,36

Statistical Analyses

We examined the frequency and proportion of subjects ever- versus never-prescribed aspirin, NSAIDs, and COX-2 inhibitors within categories of covariates (Table 1 and eTables 1a–1c; https://links.lww.com/EDE/B38). Cox regression models were used to compute the rate of breast cancer recurrence and site of recurrent disease up to 10 years after diagnosis, expressed in terms of hazard ratios (HRs) and associated Wald 95% confidence intervals (95% CIs), with and without adjustment for age, menopausal status, histological grade, ER/ET status, stage, primary surgery type, chemotherapy, prevalent comorbidities, baseline HRT use, and postdiagnosis use of simvastatin, ACE inhibitors, and beta-blockers. The 1-year lag in the Cox models allowed for a sufficient induction period to examine drug effects on recurrence, and guarded against the possibility that subclinical recurrences affected prescribing patterns. We used Cox regression models to evaluate a dose–response association between the exposure drugs and breast cancer recurrence and mortality. We investigated effect measure modification stratifying analyses by stage and ER status. Cox models were used to evaluate the association between prediagnostic drug use and breast cancer recurrence, and to evaluate the association of the exposure drugs with the rate of all-cause mortality.

T1-20
TABLE 1:
Baseline Characteristics of Patients Diagnosed with Breast Cancer in Denmark, 1996 to 2008 (n = 34,188), According to Postdiagnosis Aspirin Use, NSAIDs, and COX-2 Inhibitor Prescription Use

We also conducted analyses in a cohort restricted to patients with at least 5 years prescription history, without any prediagnostic prescriptions for the exposure drugs, and used Cox models to analyze “new use” of the exposure drugs. In these analyses, the exposure drugs were treated as time-varying exposures lagged by 1 year. All analyses were performed using SAS v.9.3 (SAS Institute, Cary, NC).

RESULTS

The study included 34,188 breast cancer patients diagnosed between 1996 and 2008. Overall, 17%, 42%, and 17% were ever users of aspirin, NSAIDs, or COX-2 inhibitors, respectively (Table 1). Aspirin users were older than nonusers, and accordingly more frequently postmenopausal at diagnosis. More users of aspirin or COX-2 inhibitors (62% vs. 56%, and 59% vs. 56%, respectively), but not NSAIDs (55% vs. 57%), underwent mastectomy compared with nonusers of each drug type. Aspirin, NSAIDs, and COX-2 inhibitor users were also more likely to have used simvastatin than nonusers of each drug type. Users of the exposure drugs were less likely to receive chemotherapy. Compared with nonusers, aspirin users had a higher frequency of heart, vascular, or liver diseases; users of NSAIDs and COX-2 inhibitors had a higher frequency of rheumatoid arthritis (eTables 1a–1c; https://links.lww.com/EDE/B38). Over 99% of aspirin prescriptions were for low-dose formulations. Ibuprofen (46%) and diclofenac (69%) were the most frequently prescribed NSAID and COX-2 inhibitors, respectively (eAppendix 3; https://links.lww.com/EDE/B38).

Overall, 5,325 (16%) patients developed recurrent disease. Compared with nonuse, use of aspirin, NSAIDs, and COX-2 inhibitors had almost no effect on recurrence in models where drug exposure was lagged by 1 year: adjusted HRaspirin = 1.0, 95% CI = 0.90, 1.1; HRNSAIDs = 0.99, 95% CI = 0.92, 1.1; HRsCOX-2 inhibitors = 1.1, 95% CI = 0.98, 1.2 (Table 2). Similar near-null findings were evident in preplanned sensitivity analyses with drug exposures lagged by 2 years and with the exposure threshold set to ≥2 prescriptions annually (eAppendix 4; https://links.lww.com/EDE/B38). Results also suggested no evidence of an association between use of any of the three drug types with breast cancer recurrence (HR = 1.0, 95% CI = 0.94, 1.1). We noted similar findings for exclusive use of aspirin, or NSAIDs, or COX-2 inhibitors, or several exposure drugs (HRaspirin = 1.0, 95% CI = 0.90, 1.2; HRNSAIDs = 0.99, 95% CI = 0.91, 1.1; HRsCOX-2 inhibitors = 1.0, 95% CI = 0.90, 1.2; HRseveral = 1.0, 95% CI = 0.90, 1.2).

T2-20
TABLE 2:
Breast Cancer Recurrence Up to 10 Years After Diagnosis, Hazard Ratios, and Associated 95% CIs for Stages I, II, or III Breast Cancer Patients in Denmark from 1996 to 2008, by Aspirin, NSAID, or COX-2 Inhibitor Use

Analyses stratified by ER/ET status and by stage at diagnosis yielded little change in the effect estimates for the exposure drugs (Table 3). Findings for aspirin and COX-2 inhibitors were also near null in analyses accounting for cumulative number of prescriptions (Table 2). We note a reduced rate of recurrence with increasing number of NSAIDs prescriptions (HR1–5 prescriptions = 1.0, 95% CI = 0.95, 1.1; HR6–10 prescriptions = 0.88, 95% CI = 0.77, 1.0; HR>10 prescriptions = 0.92, 95% CI = 0.81, 1.0), but this was accompanied by increased all-cause mortality (HR1–5 prescriptions = 1.0, 95% CI = 1.0, 1.2; HR6–10 prescriptions = 1.3, 95% CI = 1.1, 1.4; HR>10 prescriptions = 1.7, 95% CI = 1.6, 1.9).

T3-20
TABLE 3:
HRs and 95% CIs Associating Postdiagnosis Prescriptions for Aspirin, NSAIDs, or COX2 Inhibitors with Breast Cancer Recurrence Up to 10 Years After Diagnosis, Stratified by ER/ET Status and Stage Among Women with Stage I, II, or III Breast Cancer in Denmark (1996–2008)

Most recurrent events occurred in the bone, followed by the ipsilateral breast, contralateral breast, lymph nodes, and lungs. Associations for aspirin remained near null by site of recurrence (data not presented). Use of prediagnostic aspirin, NSAIDs, and COX-2 inhibitors correlated with a decreased rate of breast cancer recurrence (adjusted HRaspirin = 0.92, 95% CI = 0.82, 1.0; adjusted HRNSAIDs = 0.86, 95% CI = 0.81, 0.91; adjusted HRsCOX-2 inhibitors = 0.89, 95% CI = 0.83, 0.95).

DISCUSSION

Evidence from our large clinical cohort study does not support a protective effect of postdiagnostic use of aspirin, NSAIDs, or COX-2 inhibitors on breast cancer recurrence. These near-null findings were not modified by stratification on ER status/receipt of ET or stage at diagnosis, and also remained null in analyses examining drug exposure and site-specific cancer recurrence. We observed a slightly protective association between NSAIDs and breast cancer recurrence, but higher mortality among patients with increasing numbers of NSAIDs prescriptions, suggesting confounding by indication. Our findings from preplanned analyses of prediagnostic aspirin use, and post-hoc analyses of prediagnostic NSAIDs and COX-2 inhibitor use, suggest a protective association with breast cancer recurrence.

The near-null associations we observed for postdiagnostic aspirin exposure agree with several published studies on the effect of postdiagnostic aspirin prescriptions and breast cancer-specific mortality.19,21,23–25 However, our findings contrast with the strong protective associations for aspirin observed in two US-based studies (the Nurses’ Health Study and the Iowa Women’s Health Study) and a Scottish cohort study.18,20,28 The mortality reductions observed in the US-based studies18,20 could be attributed to a higher prevalence of high-dose aspirin use compared with the European studies. Over one-third of aspirin users in the Nurses’ Health Study cohort took high-dose aspirin.18 High-dose aspirin inhibits prostaglandin production to exert an analgesic effect, whereas low-dose aspirin exerts an antiplatelet effect.19,46 Still, if the reduced mortality rates in the US-based studies stem from high-dose aspirin inhibiting prostaglandins, we would expect to observe a similar decreased rate of recurrence associated with postdiagnostic NSAID and COX-2 inhibitor use in our study. The near-null associations that we and others22 observed for postdiagnostic NSAIDs and COX-2 inhibitor use do not support such a protective effect of prostaglandin inhibition on breast cancer recurrence. Our findings for postdiagnostic aspirin use also seem at odds with results from studies nested within randomized trials of aspirin and vascular events26 and a meta-analysis of trials and observational studies.26,27 Both suggest that cancers that developed in aspirin users were less likely to have distant metastases compared with cancers that developed in nonusers. However, these studies do not isolate the specific effect of aspirin and NSAIDs on the risk of distant metastases in breast cancer patients.

The Scottish study suggested substantially reduced breast cancer mortality associated with aspirin exposure. However, the study had large proportions of missing data (40% for some variables), seemingly contradictory effects of pre- and postdiagnostic aspirin use on breast cancer mortality (over two-fold risk vs. half the risk, respectively), and substantial differences between crude and adjusted estimates. The last of these suggests implausibly strong confounding. The Scottish study also may be prone to confounding by indication due to the inclusion of patients with stage IV disease.

Our results for NSAIDs contrast with two other US-based cohort studies, which observed almost 50% reduction in breast cancer-specific or all-cause mortality associated with NSAID use.20,21 However, drug exposure in both studies was based on questionnaire data ascertained at only one time point, and in one study was not modeled as a time-varying factor.21

Contrary to most published studies, our study investigated breast cancer recurrence rather than cancer-specific or all-cause mortality. Due to the definition of recurrent disease in the DBCG, a proportion (16%) of the DBCG-designated “recurrent disease” is actually contralateral breast cancer, and as such, may represent de novo disease. However, our near-null effect estimates were consistent regardless of site of disease recurrence.

The slight reduction in risk of recurrence associated with prediagnostic aspirin use is in keeping with a more pronounced (20%) reduction in breast cancer mortality at 5 years observed by Barron et al.23 among Irish breast cancer patients. The reduced rates of recurrence we observed for prediagnostic use of NSAIDs and COX-2 inhibitors are intriguing. They may indicate different tumor phenotypes in women routinely prescribed these drugs,47 but may also be attributable to detection bias, via more frequent healthcare access, and accordingly earlier stage at cancer diagnosis with a resulting lower likelihood for recurrence. The biologic mechanism(s) underlying any protective effects of prediagnostic drug use are difficult to deduce. Our findings suggest that both a platelet inhibitory effect (mediated by low-dose aspirin), and a prostaglandin inhibitory effect (mediated by NSAIDs and sCOX-2 inhibitors) may have a role in preventing recurrence. Further exploration of these findings is necessary to determine any potential causal associations.

Several issues should be considered when interpreting our findings. The large size of our study and prospectively collected data ensured that exposure assessment occurred before outcome, eliminating recall bias. Use of the DBCG and Danish registries ensured high quality data on breast cancer diagnosis, treatment, and recurrence, and virtually complete follow-up.37 Due to the comprehensive registry network, a strength of our study was the possibility of adjusting for several potential confounders. However, as our crude estimates were similar to the adjusted estimates, there was little evidence of confounding. The prospective prescription data enabled us to assess any evidence of a dose–response effect of the exposure drugs on recurrence. Our study substantially extends existing knowledge as we assessed the outcome of breast cancer recurrence rather than mortality. This advantage allowed us to investigate the specific effect of aspirin, NSAIDs, and COX-2 inhibitors on breast cancer, rather than on breast cancer mortality—which may be misclassified—or on all-cause mortality. We linked our cohort to other population-based and medical registries to retrieve comprehensive data on comorbid diseases and coprescriptions for each patient. In sensitivity analyses, we assessed the possibility of a latent effect of drug exposure on our outcome data, minimizing the potential for reverse causation. Some research suggests that the anticancer effect of aspirin and NSAIDs only manifests in patients with ER−/HER2− tumors.48 However, we saw little difference in results when stratified by ER/ET status. As HER2 testing was only implemented in Denmark in 2006, we had insufficient data to also stratify by HER2.

Our study has some limitations. We lacked information on prescription compliance, which may lead to exposure misclassification. However, in Denmark, patients pay for part of their redeemed prescriptions, so our estimates are likely to reflect actual use. Another concern is that all aspirin formulations are available over the counter in Denmark.49 However, low-dose aspirin, which targets platelets and has anticancer effects,50 is almost exclusively indicated for cardiovascular disease prevention. Over the counter analgesics are only available in pill packs, not in large quantity bottles. Regularly used low-dose aspirin is therefore usually prescribed by physicians and reimbursable via the Danish national health insurance system.49 Research suggests that use of nonprescription low-dose aspirin is about 8% in Denmark.51 As 17% of women in our cohort were prescribed postdiagnostic aspirin, such nonprescription low-dose aspirin use may have contributed to our null findings, but only if used regularly, which is unlikely as described above. We had no information on use of nonprescription low-dose ibuprofen (200 mg tablets), which accounts for about one-third of ibuprofen use in Denmark.51 However, nonprescription use of ibuprofen is also likely to be sporadic and short term, for treatment of transient pain like headaches or acute injuries. The Danish prescription registry data, therefore, are a valid source of regular exposure to aspirin and nonaspirin NSAIDs, particularly for chronic indications. We adjusted our analyses for rheumatoid arthritis and migraine as both may correlate with over-the-counter use of analgesics. We had no information on the indication for the exposure drug prescriptions, but adjustment for several potentially confounding comorbid conditions did not alter the effect estimates.

We note that research on colorectal cancer suggests that anticancer effects of aspirin depend on genetic factors (for example BRAF mutations),5,6 and some individuals may be intrinsically resistant to the anticancer effects of aspirin. Accordingly, aspirin users who develop cancer may be a selected population with intrinsic resistance. Unfortunately, we had no genetic data on patients included in our study. However, we did explore the possibility of acquired resistance to the exposure drugs by examining recurrence rates among new users of the drugs, but our findings remained near null.

The potential for generic drugs such as aspirin, NSAIDs, and COX-2 inhibitors to delay, mitigate, or prevent breast cancer mortality is compelling, with global implications for increasing access to affordable medicine to treat cancer.8,52 Several researchers have called for clinical trials to investigate the association between aspirin and breast cancer survival.18,24,52 Our study of over 34,000 breast cancer survivors provides little evidence to support such a protective effect of postdiagnostic use of aspirin, NSAIDs, or COX-2 inhibitors on breast cancer recurrence. However, our observed reduction in recurrence rates associated with prediagnostic use of these drugs necessitates further exploration.

ACKNOWLEDGMENTS

The authors thank the Danish Breast Cancer Cooperative Group for preparation of the initial dataset.

REFERENCES

1. Jemal A, Bray F, Center MM, Ferlay J, Ward E, Forman D. Global cancer statistics. CA Cancer J Clin. 2011;61:69–90.
2. Antithrombotic Trialists’ Collaboration. Collaborative meta-analysis of randomised trials of antiplatelet therapy for prevention of death, myocardial infarction, and stroke in high risk patients. BMJ. 2002;324:71–86.
3. Collins R, Baigent C, Sandercock P, Peto R. Antiplatelet therapy for thromboprophylaxis: the need for careful consideration of the evidence from randomised trials. Antiplatelet Trialists’ Collaboration. BMJ. 1994;309:1215–1217.
4. Cuzick J, Otto F, Baron JA, et al. Aspirin and non-steroidal anti-inflammatory drugs for cancer prevention: an international consensus statement. Lancet Oncol. 2009;10:501–507.
5. Nishihara R, Lochhead P, Kuchiba A, et al. Aspirin use and risk of colorectal cancer according to BRAF mutation status. JAMA. 2013;309:2563–2571.
6. Pasche B. Differential effects of aspirin before and after diagnosis of colorectal cancer. JAMA. 2013;309:2598–2599.
7. Smith WL, DeWitt DL, Garavito RM. Cyclooxygenases: structural, cellular, and molecular biology. Annu Rev Biochem. 2000;69:145–182.
8. Ulrich CM, Bigler J, Potter JD. Non-steroidal anti-inflammatory drugs for cancer prevention: promise, perils and pharmacogenetics. Nat Rev Cancer. 2006;6:130–140.
9. Tsujii M, DuBois RN. Alterations in cellular adhesion and apoptosis in epithelial cells overexpressing prostaglandin endoperoxide synthase 2. Cell. 1995;83:493–501.
10. Elwood PC, Gallagher AM, Duthie GG, Mur LA, Morgan G. Aspirin, salicylates, and cancer. Lancet. 2009;373:1301–1309.
11. Wang D, Dubois RN. Cyclooxygenase-2: a potential target in breast cancer. Semin Oncol. 2004;31(1 suppl 3):64–73.
12. Howe LR. Inflammation and breast cancer. Cyclooxygenase/prostaglandin signaling and breast cancer. Breast Cancer Res. 2007;9:210.
13. Kim HS, Moon HG, Han W, et al. COX2 overexpression is a prognostic marker for Stage III breast cancer. Breast Cancer Res Treat. 2012;132:51–59.
14. van Nes JG, de Kruijf EM, Faratian D, et al. COX2 expression in prognosis and in prediction to endocrine therapy in early breast cancer patients. Breast Cancer Res Treat. 2011;125:671–685.
15. Lucci A, Krishnamurthy S, Singh B, et al. Cyclooxygenase-2 expression in primary breast cancers predicts dissemination of cancer cells to the bone marrow. Breast Cancer Res Treat. 2009;117:61–68.
16. Allen JE, Patel AS, Prabhu VV, et al. COX-2 drives metastatic breast cells from brain lesions into the cerebrospinal fluid and systemic circulation. Cancer Res. 2014;74:2385–2390.
17. Choi BH, Chakraborty G, Baek K, Yoon HS. Aspirin-induced Bcl-2 translocation and its phosphorylation in the nucleus trigger apoptosis in breast cancer cells. Exp Mol Med. 2013;45:e47.
18. Holmes MD, Chen WY, Li L, Hertzmark E, Spiegelman D, Hankinson SE. Aspirin intake and survival after breast cancer. J Clin Oncol. 2010;28:1467–1472.
19. Murray LJ, Cooper JA, Hughes CM, Powe DG, Cardwell CR. Post-diagnostic prescriptions for low-dose aspirin and breast cancer-specific survival: a nested case-control study in a breast cancer cohort from the UK Clinical Practice Research Datalink. Breast Cancer Res. 2014;16:R34.
20. Blair CK, Sweeney C, Anderson KE, Folsom AR. NSAID use and survival after breast cancer diagnosis in post-menopausal women. Breast Cancer Res Treat. 2007;101:191–197.
21. Kwan ML, Habel LA, Slattery ML, Caan B. NSAIDs and breast cancer recurrence in a prospective cohort study. Cancer Causes Control. 2007;18:613–620.
22. Li Y, Brasky TM, Nie J, et al. Use of nonsteroidal anti-inflammatory drugs and survival following breast cancer diagnosis. Cancer Epidemiol Biomarkers Prev. 2012;21:239–242.
23. Barron TI, Flahavan EM, Sharp L, Bennett K, Visvanathan K. Recent prediagnostic aspirin use, lymph node involvement, and 5-year mortality in women with stage I-III breast cancer: a nationwide population-based cohort study. Cancer Res. 2014;74:4065–4077.
24. Holmes MD, Olsson H, Pawitan Y, et al. Aspirin intake and breast cancer survival - a nation-wide study using prospectively recorded data in Sweden. BMC Cancer. 2014;14:391.
25. Barron TI, Murphy LM, Brown C, Bennett K, Visvanathan K, Sharp L. De novo post-diagnosis aspirin use and mortality in women with stage I-III breast cancer. Cancer Epidemiol Biomarkers Prev. 2015;24:898–904.
26. Rothwell PM, Wilson M, Price JF, Belch JF, Meade TW, Mehta Z. Effect of daily aspirin on risk of cancer metastasis: a study of incident cancers during randomised controlled trials. Lancet. 2012;379:1591–1601.
27. Algra AM, Rothwell PM. Effects of regular aspirin on long-term cancer incidence and metastasis: a systematic comparison of evidence from observational studies versus randomised trials. Lancet Oncol. 2012;13:518–527.
28. Fraser DM, Sullivan FM, Thompson AM, McCowan C. Aspirin use and survival after the diagnosis of breast cancer: a population-based cohort study. Br J Cancer. 2014;111:623–627.
29. Forget P, Vandenhende J, Berliere M, et al. Do intraoperative analgesics influence breast cancer recurrence after mastectomy? A retrospective analysis. Anesth Analg. 2010;110:1630–1635.
30. Forget P, Bentin C, Machiels JP, Berliere M, Coulie PG, De Kock M. Intraoperative use of ketorolac or diclofenac is associated with improved disease-free survival and overall survival in conservative breast cancer surgery. Br J Anaesth. 2014;113(Suppl 1):i82–i87.
31. Brandão RD, Veeck J, Van de Vijver KK, et al. A randomised controlled phase II trial of pre-operative celecoxib treatment reveals anti-tumour transcriptional response in primary breast cancer. Breast Cancer Res. 2013;15:R29.
32. Sørensen HT, Mellemkjaer L, Blot WJ, et al. Risk of upper gastrointestinal bleeding associated with use of low-dose aspirin. Am J Gastroenterol. 2000;95:2218–2224.
33. Kwan ML, Habel LA, Flick ED, Quesenberry CP, Caan B. Post-diagnosis statin use and breast cancer recurrence in a prospective cohort study of early stage breast cancer survivors. Breast Cancer Res Treat. 2008;109:573–579.
34. Chae YK, Valsecchi ME, Kim J, et al. Reduced risk of breast cancer recurrence in patients using ACE inhibitors, ARBs, and/or statins. Cancer Invest. 2011;29:585–593.
35. Ahern TP, Lash TL, Damkier P, Christiansen PM, Cronin-Fenton DP. Statins and breast cancer prognosis: evidence and opportunities. Lancet Oncol. 2014;15:e461–e468.
36. Ahern TP, Pedersen L, Tarp M, et al. Statin prescriptions and breast cancer recurrence risk: a Danish nationwide prospective cohort study. J Natl Cancer Inst. 2011;103:1461–1468.
37. Jensen AR, Storm HH, Møller S, Overgaard J. Validity and representativity in the Danish Breast Cancer Cooperative Group–a study on protocol allocation and data validity from one county to a multi-centre database. Acta Oncol. 2003;42:179–185.
38. Møller S, Jensen MB, Ejlertsen B, et al.; Danish Breast Cancer Cooperative Group. The clinical database and the treatment guidelines of the Danish Breast Cancer Cooperative Group (DBCG); its 30-years experience and future promise. Acta Oncol. 2008;47:506–524.
39. Danish Breast Cancer Cooperative Group. Danish Breast Cancer Cooperative Group Kvalitetsindikatorrapport for Brystkræft 2012. Available at: www.dbcg.dk. Updated 20122014. Accessed 28th February 2015.
40. Christiansen P, Al-Suliman N, Bjerre K, Møller S; Danish Breast Cancer Cooperative Group. Recurrence pattern and prognosis in low-risk breast cancer patients–data from the DBCG 89-A programme. Acta Oncol. 2008;47:691–703.
41. Schmidt M, Pedersen L, Sørensen HT. The Danish civil registration system as a tool in epidemiology. Eur J Epidemiol. 2014;29:541–549.
42. Gaist D, Sørensen HT, Hallas J. The Danish prescription registries. Dan Med Bull. 1997;44:445–448.
43. Pedersen CB, Gøtzsche H, Møller JO, Mortensen PB. The Danish civil registration system. A cohort of eight million persons. Dan Med Bull. 2006;53:441–449.
44. Andersen TF, Madsen M, Jørgensen J, Mellemkjoer L, Olsen JH. The Danish national hospital register. A valuable source of data for modern health sciences. Dan Med Bull. 1999;46:263–268.
45. Salvan A, Stayner L, Steenland K, Smith R. Selecting an exposure lag period. Epidemiology. 1995;6:387–390.
46. Weksler BB, Pett SB, Alonso D, et al. Differential inhibition by aspirin of vascular and platelet prostaglandin synthesis in atherosclerotic patients. N Engl J Med. 1983;308:800–805.
47. Chan AT, Cook NR. Are we ready to recommend aspirin for cancer prevention? Lancet. 2012;379:1569–1571.
48. Brasky TM, Bonner MR, Moysich KB, et al. Non-steroidal anti-inflammatory drugs (NSAIDs) and breast cancer risk: differences by molecular subtype. Cancer Causes Control. 2011;22:965–975.
49. Schmidt M, Johansen MB, Lash TL, Christiansen CF, Christensen S, Sørensen HT. Antiplatelet drugs and risk of subarachnoid hemorrhage: a population-based case-control study. J Thromb Haemost. 2010;8:1468–1474.
50. Bambace NM, Holmes CE. The platelet contribution to cancer progression. J Thromb Haemost. 2011;9:237–249.
51. Schmidt M, Hallas J, Friis S. Potential of prescription registries to capture individual-level use of aspirin and other nonsteroidal anti-inflammatory drugs in Denmark: trends in utilization 1999-2012. Clin Epidemiol. 2014;6:155–168.
52. Holmes MD, Chen WY. Hiding in plain view: the potential for commonly used drugs to reduce breast cancer mortality. Breast Cancer Res. 2012;14:216.

Supplemental Digital Content

Copyright © 2016 Wolters Kluwer Health, Inc. All rights reserved.