Abstract
Background
We examined the role of post-diagnostic coffee and tea consumption in relation to breast cancer-specific and all-cause mortality among women with breast cancer in prospective cohort studies.
Methods
We identified 8900 women with stage I–III breast cancer from 1980 through 2010 in the Nurses’ Health Study (NHS) and from 1991 through 2011 in the NHSII. Post-diagnostic coffee and tea consumption was assessed by a validated food frequency questionnaire every 4 years after diagnosis.
Results
During up to 30 years of follow-up, we documented 1054 breast cancer-specific deaths and 2501 total deaths. Higher post-diagnostic coffee consumption was associated with a lower breast cancer-specific mortality: compared with non-drinkers, >3 cups/day of coffee was associated with a 25% lower risk (hazard ratio (HR) = 0.75, 95% confidence interval (CI) = 0.59–0.96; Ptrend = 0.002). We also observed a lower all-cause mortality with coffee consumption: compared with non-drinkers, >2 to 3 cups/day was associated with a 24% lower risk (HR = 0.76, 95% CI = 0.66–0.87) and >3 cups/day was associated with a 26% lower risk (HR = 0.74, 95% CI = 0.63–0.87, Ptrend < 0.0001). Post-diagnostic tea consumption was associated with a lower all-cause mortality: compared with non-drinkers, >3 cups/day was associated with a 26% lower risk (HR = 0.74, 95% CI = 0.58–0.95; Ptrend = 0.04).
Conclusions
Among breast cancer survivors, higher post-diagnostic coffee consumption was associated with better breast cancer and overall survival. Higher post-diagnostic tea consumption may be related to better overall survival.
Subject terms: Breast cancer, Outcomes research
Background
Coffee and tea are among the world’s most widely consumed beverages. They consist of many bioactive compounds, including high amounts of caffeine and polyphenols. There are reports of beneficial effects of coffee and caffeine intake on insulin sensitivity1 and inflammation,2–5 in addition to antioxidant activity,6 which may contribute to breast cancer prognosis.7–10 However, epidemiological studies assessing the relation between coffee consumption and breast cancer mortality are limited and inconsistent.11–16 Tea has also been hypothesised to be anti-carcinogenic because of its effects on sex hormone levels,17 antioxidative properties,18–20 and role in improving insulin sensitivity.21 However, pre-diagnostic tea consumption was not associated with breast cancer survival in the Swedish Mammography Cohort.13 Pre-diagnostic caffeine intake, a well-known constituent in both coffee and black tea, was not associated with breast cancer-specific or all-cause mortality among Swedish women with breast cancer.13 Nonetheless, the effects of coffee, tea and caffeine consumption after breast cancer diagnosis on survival remain unclear.11,12
In this study, we investigated the associations between coffee and tea consumption and breast cancer survival using dietary assessments every 4 years after diagnosis in the women diagnosed with breast cancer in the Nurses’ Health Study (NHS) and the Nurses’ Health Study II (NHSII). We were able to account for pre-diagnostic coffee and tea consumption and investigate the associations by oestrogen receptor (ER) status, insulin receptor (IR) status, molecular subtypes and stage of the disease.
Methods
Study population
We used data from the NHS, an ongoing cohort study initiated in 1976 with an enrolment of 121,700 US female registered nurses aged 30–55 years, and the NHSII, an ongoing cohort study initiated in 1989 with an enrolment of 116,429 US female registered nurses aged 25–42 years. From 1980 through 2010 in the NHS and from 1991 through 2011 in the NHSII, we selected women with confirmed invasive breast cancer. For this study, we excluded women with the following: missing diet information at least 12 months after diagnosis; total energy intake <600 or >3500 kcal/day; leaving blank more than 70 food items on the food frequency questionnaire (FFQ), leaving blank all coffee or tea items on the FFQ, any other cancers diagnosis (except non-melanoma skin cancer) before breast cancer; stage IV disease at diagnosis; and missing information on disease stage. Thus, 8900 women with stage I–III were eligible for the analysis (Supplementary Fig. S1).
Assessment of dietary intake
In the NHS, frequency of consumption of food items over the past year was evaluated through a 61-item semiquantitative FFQ administered in 1980 and 116–130-item FFQs administered in 1984, 1986, and every 4 years thereafter until 2010. In the NHSII, dietary intake was assessed through ~130-item FFQs administered in 1991 and every four years thereafter until 2011 (questionnaires available at http://www.nurseshealthstudy.org/participants/questionnaires). For this study, we collected the data on “regular coffee”, “decaffeinated coffee”, “regular tea” and “decaffeinated tea”. Total coffee consumption was calculated as the sum of regular and decaffeinated coffee, and total tea consumption was calculated as the sum of regular and decaffeinated tea. The validity of the FFQ has been documented by comparing the mean daily coffee consumption, using FFQs and 28-day diet records. The mean of corrected correlation coefficients for coffee consumption was 0.77, calculated using the FFQ and average of the 28-day diet records.22–24 Greater coffee consumption in these cohorts has also related to lower risk of type 2 diabetes25 and total mortality.14
The amounts of caffeine and energy in foods were calculated using the Harvard University Food Composition Database, updated every 4 years to account for changes in the food supply. Caffeine intake was energy-adjusted using the residuals from the regression of caffeine intake on total energy intake.26 To reduce random within-person variation and best represent long-term intake, we calculated cumulative averages of coffee, tea, caffeine, total energy and alcohol intake from the FFQs reported at least 12 months after diagnosis and every 4 years thereafter. We additionally calculated the pre-diagnostic coffee and tea consumption from the last FFQ reported before breast cancer diagnosis.
Ascertainment of breast cancer and death
Breast cancers were identified through self-reported diagnosis on the biennial questionnaires. In order to confirm the diagnosis and obtain information on tumour characteristics, stage of the disease and treatment, we asked permission to access medical records and pathology reports. We collected breast tumour tissue for ~70% of women. Tissue microarray (TMA) technique was performed, and ER, progesterone receptor (PR), human epidermal growth factor receptor 2 (HER2), cytokeratin 5/6 (CK5/6), Ki-67 and epidermal growth factor receptor (EGFR) in tumours were measured using immunohistochemistry.27–29 If TMAs were not available, ER, PR and HER2 status were obtained from medical records. Molecular subtypes were determined as follows: Luminal A [ER-positive and/or PR-positive, HER2-negative and Ki-67-negative (or histologic grade 1 or 2)]; luminal B [ER-positive and/or PR-positive, and HER2-positive; or ER-positive and/or PR-positive, HER2-negative and Ki-67-positive (or histologic grade 3)]; HER2-enriched (ER-negative, PR-negative and HER2-positive); basal-like (ER-negative, PR-negative, HER2-negative, and CK5/6-positive and/or EGFR-positive); and unclassified tumours lacked expression for all five markers. IR expression (cytoplasmic and membranous) was quantified using Definiens image analysis software (Tissue Studio, Definiens AG, Munich, Germany) in the NHS.30
If deaths were reported by family members or the postal service or determined through a search in the National Death Index, the death certificate and medical record were reviewed by a study physician to determine the cause of death.
Covariates
From questionnaires women reported every 2 years after diagnosis, we obtained data on post-diagnostic aspirin use, smoking status, physical activity, and body mass index (BMI). To reduce treatment effects on lifestyle factor assessment, only data collected at least 12 months after diagnosis were considered and updated at the beginning of each-2-year follow-up interval, if available. Because of possible reverse causation, we calculated the cumulative averages using 4-year-lagged data for both post-diagnostic BMI and physical activity. The change in BMI from prior to and after cancer diagnosis was calculated by subtracting the last reported BMI before diagnosis from the 4-year-lagged cumulative average of post-diagnostic BMI. Data were also collected on oral contraceptive use, menopausal status, age at menopause and postmenopausal hormone use that women reported before breast cancer diagnosis. In addition, using supplemental questionnaires and medical records, we collected data for breast cancer characteristics, including age at diagnosis, stage of the disease and self-reported treatment, including radiotherapy, chemotherapy and hormonal treatment.
Statistical analysis
We followed up women from the return date of the first FFQ after diagnosis through the end of the study period (June 1, 2014, for the NHS and June 1, 2015, for the NHSII) or death, whichever came first. The outcomes of the study were mortality due to breast cancer and all-cause.
Data from the NHS and NHSII were combined and stratified Cox proportional hazards regression models with strata defined by cohort were performed to estimate hazard ratios (HRs) and 95% confidence intervals (CIs). Patients diagnosed with breast cancer were divided into categories according to the cumulative averages of coffee and tea consumption. Model 1 was stratified by cohort and adjusted for age at diagnosis and calendar year of diagnosis. In Model 2 (multivariable model), we additionally adjusted for the time between diagnosis and first FFQ after diagnosis, calendar year at the start of follow-up of each-2-year questionnaire cycle, potential predictors of breast cancer survival including pre-diagnostic BMI, BMI change after diagnosis, post-diagnostic physical activity, post-diagnostic aspirin use, post-diagnostic alcohol consumption, post-diagnostic total energy intake, pre-diagnostic oral contraceptive use, menopausal status, age at menopause, and postmenopausal hormone use, race, tumour characteristics (stage of the disease, ER/PR status) and treatment (radiotherapy, chemotherapy, hormonal treatment). Women with unknown menopausal status at the time of diagnosis were grouped as premenopausal if they were younger than 46 years and smoking or younger than 48 years and never smoking and grouped as postmenopausal if they were older than 54 years and smoking or older than 56 years and never smoking.31 We examined the dose–response relationship between coffee consumption and breast cancer-specific and all-cause mortality using restricted cubic spline analyses.32 We additionally evaluated the associations of pre-diagnostic coffee and tea consumption. In sensitivity analyses, we adjusted for coffee or tea consumption prior to cancer diagnosis, as well as post-diagnostic modified alternate healthy eating index (AHEI) score (not including alcohol score), dietary glycaemic index, intakes of fruits and vegetables, added sugar, sugar-sweetened beverages (SSBs) and caffeine from sources other than coffee. Considering the role of pre-diagnostic coffee consumption in breast cancer survival, we evaluated the breast cancer-specific and all-cause mortality with cross-classification of pre- and post-diagnostic coffee consumption, with pre- and post-diagnostic non-drinkers as the reference category. We also used Fine-Gray method to conduct competing risk analyses for causes of death: breast cancer-specific mortality versus other causes of death.33,34 In addition, we evaluated associations without considering at least a 12-month gap between the time of diagnosis and first post-diagnostic FFQ, considering left truncation time since diagnosis, and with simple update instead of cumulative average.
To examine potential effect modification of clinical-pathological factors, we evaluated the associations between coffee and tea consumption and mortality among women classified by ER status (ER-positive versus ER-negative), IR status (IR-positive versus IR-negative) and molecular subtypes (luminal A/luminal B/HER2-enriched/basal-like). We also evaluated the associations stratified by age at diagnosis (<60 versus ≥ 60 years), post-diagnostic BMI (<25 versus ≥25 kg/m2), post-diagnostic smoking status (ever versus never), post-diagnostic alcohol consumption (<3.5 versus ≥3.5 g/day), stage of the disease (I/II/III), and AHEI (< versus ≥median score). The HR of mortality was calculated per 1 cup/day increment of coffee or tea consumption. The P value for interaction was calculated using the Wald test. All analyses were conducted using SAS software version 9.4 (SAS Institute, Cary, NC) with a two-sided P value of <0.05.
Results
We documented 1054 deaths due to breast cancer and 2501 deaths from all causes among 8900 women with breast cancer after a median of 11.5 years of follow-up (up to 30 years) from the return of the first post-diagnostic FFQ. On average, women drank 1.7 cups/day of coffee and 0.7 cup/day of tea after diagnosis. Women with higher coffee consumption after diagnosis (measured at the first FFQ after diagnosis) were more likely to drink more alcohol and smoke, consume more animal fat, and use aspirin more frequently after diagnosis as well as report higher coffee consumption prior to diagnosis (Table 1). Furthermore, women with higher coffee consumption were less likely to have radiotherapy or chemotherapy. Participants with higher tea consumption after diagnosis tended to drink less alcohol, consume less animal fat, smoke less, and use aspirin less frequently (Table 1). Women with higher tea consumption were more likely to have chemotherapy or hormonal treatment.
Table 1.
Age-standardized characteristics of women with breast cancer in the combined Nurses’ Health Study and Nurses’ Health Study II, according to total coffee and total tea consumption measured at the first FFQ after diagnosis.
| Total coffee consumption | Total tea consumption | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Nondrinker | >0 to 1 cup/day | >1 to 2 cups/day | >2 to 3 cups/day | >3 cups/day | Nondrinker | >0 to 1 cup/day | >1 to 2 cups/day | >2 to 3 cups/day | >3 cups/day | |
| Number | 1586 | 2875 | 759 | 2625 | 1055 | 2644 | 4567 | 488 | 816 | 385 |
| Mean | ||||||||||
| Intake from the first FFQ after diagnosis | ||||||||||
| Coffee, cups/day | 0 | 0.7 | 1.5 | 2.5 | 4.7 | 1.9 | 1.7 | 1.2 | 1.1 | 1.1 |
| Tea, cups/day | 1.2 | 0.8 | 0.7 | 0.6 | 0.5 | 0 | 0.4 | 1.5 | 2.5 | 4.6 |
| Sugar-sweetened beverages, servings/day | 0.3 | 0.3 | 0.2 | 0.2 | 0.2 | 0.3 | 0.3 | 0.2 | 0.3 | 0.3 |
| Total fruits, servings/day | 1.6 | 1.6 | 1.7 | 1.5 | 1.6 | 1.4 | 1.6 | 1.8 | 1.7 | 1.7 |
| Total vegetables, servings/day | 3.0 | 3.1 | 3.3 | 3.1 | 3.2 | 2.8 | 3.1 | 3.6 | 3.3 | 3.5 |
| Alcohol, g/day | 3.0 | 5.2 | 6.1 | 7.4 | 6.8 | 6.6 | 5.4 | 5.5 | 5.3 | 4.3 |
| Animal fat, % energy/day | 13.7 | 14.1 | 14.0 | 14.8 | 15.5 | 15.0 | 14.3 | 12.6 | 14.1 | 13.8 |
| Total fat, % energy/day | 30.2 | 30.1 | 30.2 | 31.2 | 31.6 | 30.6 | 30.6 | 31.1 | 30.6 | 31.3 |
| Total energy, kcal/day | 1667 | 1692 | 1773 | 1735 | 1802 | 1620 | 1735 | 1866 | 1796 | 1853 |
| Alternate Healthy Eating Index, score | 55.5 | 56.6 | 57.7 | 56.5 | 55.0 | 54.1 | 56.6 | 62.0 | 56.8 | 58.5 |
| Intake from the last FFQ before diagnosis | ||||||||||
| Coffee, cups/day | 0.3 | 1.4 | 2.0 | 2.6 | 3.9 | 2.1 | 1.9 | 1.5 | 1.4 | 1.3 |
| Tea, cups/day | 1.1 | 0.7 | 0.7 | 0.6 | 0.5 | 0.2 | 0.6 | 1.2 | 1.8 | 2.5 |
| Age at diagnosis, years | 56.7 | 60.4 | 59.5 | 58.5 | 56.7 | 59.6 | 58.5 | 58.3 | 57.8 | 56.6 |
| BMI, kg/m2 | 27.1 | 27.0 | 26.5 | 26.1 | 25.8 | 26.5 | 26.6 | 26.7 | 26.2 | 26.7 |
| Physical activity, MET-h/wk | 17.2 | 17.2 | 19.1 | 18.5 | 17.2 | 16.6 | 18.1 | 21.0 | 17.9 | 16.8 |
| % | ||||||||||
| Current smokers | 4 | 6 | 6 | 12 | 21 | 14 | 8 | 4 | 6 | 8 |
| Past smokers | 31 | 44 | 47 | 49 | 46 | 45 | 43 | 42 | 42 | 42 |
| Ever used oral contraceptives | 58 | 59 | 56 | 57 | 55 | 57 | 57 | 66 | 56 | 58 |
| Ever used postmenopausal hormone | 47 | 49 | 49 | 48 | 49 | 48 | 47 | 51 | 48 | 48 |
| Current use of aspirin | 39 | 42 | 48 | 47 | 48 | 44 | 45 | 43 | 42 | 40 |
| Premenopausal at diagnosis | 27 | 26 | 25 | 26 | 25 | 25 | 26 | 28 | 27 | 27 |
| Stage of breast cancer | ||||||||||
| I | 58 | 60 | 61 | 61 | 61 | 61 | 59 | 60 | 61 | 60 |
| II | 32 | 30 | 29 | 29 | 29 | 30 | 30 | 30 | 29 | 29 |
| III | 10 | 10 | 10 | 10 | 10 | 9 | 11 | 10 | 10 | 11 |
| Oestrogen receptor status | ||||||||||
| Positive | 76 | 78 | 79 | 76 | 76 | 76 | 77 | 78 | 77 | 77 |
| Negative | 19 | 16 | 16 | 17 | 16 | 17 | 17 | 19 | 17 | 17 |
| Missing | 5 | 6 | 5 | 7 | 8 | 7 | 6 | 3 | 6 | 6 |
| Treatment | ||||||||||
| Radiotherapy | 58 | 58 | 56 | 58 | 49 | 53 | 58 | 70 | 54 | 53 |
| Chemotherapy | 50 | 47 | 46 | 43 | 43 | 42 | 47 | 55 | 45 | 50 |
| Hormonal treatment | 68 | 72 | 69 | 68 | 66 | 67 | 70 | 78 | 68 | 70 |
In the multivariable model, compared with non-drinkers, breast cancer-specific mortality was 25% lower for women consuming >3 cups/day of coffee (HR = 0.75; 95% CI = 0.59–0.96; Ptrend = 0.002) (Table 2). Lower all-cause mortality was also observed among coffee drinkers: compared with non-drinkers, the risk for >2–3 and >3 cups/day of coffee was significantly lower by 24% (HR = 0.76; 95% CI = 0.66–0.87) and 26% (HR = 0.74; 95% CI = 0.63–0.87) (Ptrend < 0.0001), respectively. These associations were approximately linear (Supplementary Fig. S2a, b). The associations with post-diagnostic coffee intake were notably stronger after additional adjustment for pre-diagnostic coffee consumption (Table 2). The associations with post-diagnostic coffee intake remained significant after additional adjustment for post-diagnostic intakes of fruits and vegetables, dietary glycaemic index, added sugar, caffeine from sources other than coffee, SSBs or AHEI. We performed competing risk analyses for causes of death: breast cancer-specific mortality versus other causes of mortality and did not observe changes in the associations. Starting in 2007, women reported the consumption of dairy coffee drink such as cappuccino. Results did not change when it was added to the sum of coffee consumption.
Table 2.
Post-diagnostic coffee and tea consumption in relation to mortality after breast cancer diagnosis (n = 8900) in the Nurses’ Health Study and Nurses’ Health Study II.
| Consumption levels | Ptrend | Per 1 cup/day increase | |||||
|---|---|---|---|---|---|---|---|
| Nondrinker | >0 to 1 cup/day | >1 to 2 cups/day | >2 to 3 cups/day | >3 cups/day | |||
| Total coffee | |||||||
| Breast cancer-specific mortality | |||||||
| No. of deaths | 166 | 328 | 187 | 250 | 123 | ||
| Person-years | 14,207 | 27,648 | 17,913 | 27,478 | 14,383 | ||
| Model 1 | 1 | 0.95 (0.78–1.14) | 0.76 (0.62–0.94) | 0.69 (0.57–0.84) | 0.60 (0.48–0.76) | <0.0001 | 0.87 (0.82–0.91) |
| Model 2 | 1 | 0.99 (0.82–1.20) | 1.09 (0.88–1.36) | 0.83 (0.67–1.02) | 0.75 (0.59–0.96) | 0.002 | 0.92 (0.87–0.97) |
| Model 2+pre-diagnostic total coffee | 1 | 0.89 (0.71–1.12) | 0.92 (0.71–1.20) | 0.66 (0.51–0.85) | 0.54 (0.40–0.73) | <0.0001 | 0.85 (0.79–0.91) |
| All-cause mortality | |||||||
| No. of deaths | 355 | 786 | 509 | 565 | 286 | ||
| Person-years | 14,207 | 27,648 | 17,913 | 27,478 | 14,383 | ||
| Model 1 | 1 | 0.94 (0.83–1.07) | 0.91 (0.79–1.04) | 0.70 (0.61–0.80) | 0.68 (0.58–0.80) | <0.0001 | 0.89 (0.86–0.92) |
| Model 2 | 1 | 0.96 (0.84–1.09) | 1.03 (0.89–1.19) | 0.76 (0.66–0.87) | 0.74 (0.63–0.87) | <0.0001 | 0.91 (0.88–0.95) |
| Model 2+pre-diagnostic total coffee | 1 | 0.88 (0.75–1.03) | 0.91 (0.76–1.08) | 0.64 (0.54–0.77) | 0.58 (0.47–0.70) | <0.0001 | 0.86 (0.83–0.90) |
| Regular coffee | |||||||
| Breast cancer-specific mortality | |||||||
| No. of deaths | 339 | 356 | 99 | 199 | 61 | ||
| Person-years | 28,860 | 31,999 | 12,209 | 21,455 | 7107 | ||
| Model 1 | 1 | 0.91 (0.78–1.05) | 0.62 (0.49–0.77) | 0.77 (0.65–0.92) | 0.67 (0.51–0.88) | <0.0001 | 0.89 (0.83–0.94) |
| Model 2 | 1 | 1.02 (0.88–1.19) | 1.07 (0.84–1.36) | 0.87 (0.73–1.05) | 0.71 (0.53–0.94) | 0.008 | 0.93 (0.87–0.99) |
| Model 2 + pre-diagnostic regular coffee | 1 | 1.01 (0.85–1.20) | 0.97 (0.75–1.26) | 0.77 (0.62–0.95) | 0.57 (0.41–0.78) | <0.0001 | 0.87 (0.81–0.94) |
| All-cause mortality | |||||||
| No. of deaths | 769 | 856 | 316 | 419 | 141 | ||
| Person-years | 28,860 | 31,999 | 12,209 | 21,455 | 7107 | ||
| Model 1 | 1 | 0.98 (0.88–1.08) | 0.95 (0.84–1.09) | 0.77 (0.68–0.87) | 0.80 (0.67–0.96) | <0.0001 | 0.91 (0.88–0.95) |
| Model 2 | 1 | 1.03 (0.93–1.14) | 1.09 (0.95–1.25) | 0.81 (0.72–0.92) | 0.76 (0.63–0.91) | <0.0001 | 0.93 (0.89–0.97) |
| Model 2 + pre-diagnostic regular coffee | 1 | 1.02 (0.91–1.15) | 1.04 (0.89–1.22) | 0.77 (0.66–0.89) | 0.68 (0.55–0.84) | <0.0001 | 0.89 (0.85–0.94) |
| Decaffeinated coffee | |||||||
| Breast cancer-specific mortality | |||||||
| No. of deaths | 509 | 385 | 160 | ||||
| Person-years | 44,453 | 40,368 | 16,808 | ||||
| Model 1 | 1 | 0.75 (0.66–0.86) | 0.69 (0.58–0.83) | 0.0006 | 0.90 (0.84–0.97) | ||
| Model 2 | 1 | 0.97 (0.85–1.12) | 0.94 (0.78–1.12) | 0.49 | 0.98 (0.91–1.06) | ||
| Model 2 + pre-diagnostic decaffeinated coffee | 1 | 0.93 (0.80–1.08) | 0.93 (0.75–1.16) | 0.65 | 0.99 (0.91–1.08) | ||
| All-cause mortality | |||||||
| No. of deaths | 1160 | 938 | 403 | ||||
| Person-years | 44,453 | 40,368 | 16,808 | ||||
| Model 1 | 1 | 0.80 (0.73–0.87) | 0.77 (0.68–0.86) | 0.0001 | 0.90 (0.86–0.95) | ||
| Model 2 | 1 | 0.89 (0.82–0.98) | 0.88 (0.78–0.99) | 0.07 | 0.95 (0.91–1.00) | ||
| Model 2+pre-diagnostic decaffeinated coffee | 1 | 0.87 (0.79–0.96) | 0.84 (0.74–0.97) | 0.06 | 0.94 (0.89–0.99) | ||
| Total tea | |||||||
| Breast cancer-specific mortality | |||||||
| No. of deaths | 260 | 583 | 101 | 75 | 35 | ||
| Person-years | 20,758 | 58,059 | 11,055 | 7883 | 3875 | ||
| Model 1 | 1 | 0.79 (0.68–0.92) | 0.73 (0.58–0.92) | 0.77 (0.60–1.00) | 0.78 (0.55–1.11) | 0.13 | 0.94 (0.87–1.01) |
| Model 2 | 1 | 0.94 (0.81–1.10) | 1.04 (0.82–1.31) | 0.83 (0.64–1.08) | 0.80 (0.56–1.14) | 0.17 | 0.97 (0.91–1.04) |
| Model 2+pre-diagnostic total tea | 1 | 0.96 (0.81–1.13) | 1.05 (0.81–1.35) | 0.83 (0.62–1.10) | 0.78 (0.53–1.14) | 0.18 | 0.98 (0.90–1.06) |
| All-cause mortality | |||||||
| No. of deaths | 594 | 1400 | 264 | 171 | 72 | ||
| Person-years | 20,758 | 58,059 | 11,055 | 7883 | 3875 | ||
| Model 1 | 1 | 0.90 (0.82–0.99) | 0.95 (0.82–1.10) | 0.84 (0.71–0.99) | 0.81 (0.63–1.03) | 0.09 | 0.97 (0.92–1.01) |
| Model 2 | 1 | 0.92 (0.83–1.02) | 1.00 (0.86–1.17) | 0.85 (0.71–1.01) | 0.74 (0.58–0.95) | 0.04 | 0.97 (0.93–1.02) |
| Model 2+pre-diagnostic total tea | 1 | 0.92 (0.83–1.03) | 1.00 (0.85–1.17) | 0.82 (0.68–0.99) | 0.71 (0.55–0.93) | 0.02 | 0.97 (0.92–1.02) |
Note: Model 1 was stratified by cohort and adjusted for age at diagnosis (year) and calendar year of diagnosis.
Model 2 was stratified by cohort and adjusted for age at diagnosis (year), calendar year of diagnosis, time between diagnosis and first FFQ (year), calendar year at start of follow-up of each-2-year questionnaire cycle, pre-diagnostic BMI (<20, 20 to <22.5, 22.5 to <25, 25.0 to <30, 30 to <35, ≥35 kg/m2, missing), BMI change after diagnosis (no change (≥−0.5 to ≤0.5 kg/m2), decrease (<−0.5 kg/m2), increase (>0.5–2 kg/m2), increase (>2 kg/m2), missing), post-diagnostic smoking (never, past, current 1–14 cigarettes/day, current 15–24 cigarettes/day, current ≥25 cigarettes/day, missing), post-diagnostic physical activity (<5, 5 to <11.5, 11.5 to <22, ≥22 MET-h/week, missing), oral contraceptive use (ever, never), post-diagnostic alcohol consumption (<0.15, 0.15 to <2.0, 2.0 to 7.5, ≥7.5 g/day), post-diagnostic total energy intake (quintiles, kcal/day), pre-diagnostic menopausal status, age at menopause and postmenopausal hormone use status (premenopausal; postmenopausal, age at menopause <50 year and never postmenopausal hormone use; postmenopausal, age at menopause <50 year and past postmenopausal hormone use; postmenopausal, age at menopause <50 year and current postmenopausal hormone use; postmenopausal, age at menopause ≥50 year and never postmenopausal hormone use; postmenopausal, age at menopause ≥50 year and past postmenopausal hormone use; postmenopausal, age at menopause ≥50 year and current postmenopausal hormone use; missing), post-diagnostic aspirin use (never, past, current, missing), race (non-Hispanic white, other), stage of disease (I, II, III), ER/PR status (ER/PR-positive, ER-positive and PR-negative, ER/PR-negative, missing), radiotherapy (yes, no, missing), chemotherapy (yes, no, missing) and hormonal treatment (yes, no, missing).
For each one cup/day greater intake, the HR was 7% lower for regular coffee (95% CI = 0.87–0.99) and 2% lower for decaffeinated coffee (95% CI = 0.91–1.06) (Table 2). We observed lower all-cause mortality with greater intake of both regular and decaffeinated coffee: for each one cup/day greater intake, the HR was 7% lower for regular coffee (95% CI = 0.89–0.97) and 5% lower for decaffeinated coffee (95% CI = 0.91–1.00), though there were fewer consumers of decaffeinated coffee (Table 2). When regular and decaffeinated coffee intakes were included together in the model, we observed inverse associations with breast cancer-specific mortality for both regular and decaffeinated coffee, although the association for decaffeinated coffee was not significant (per one cup/day greater intake, HR = 0.92, 95% CI = 0.87–0.99 for regular coffee; HR = 0.96; 95% CI = 0.89–1.04, for decaffeinated coffee). For all-cause mortality, significant inverse associations were observed for both forms of coffee, mutually adjusted (per one cup/day greater intake, HR = 0.91; 95% CI = 0.87–0.95 for regular coffee; HR = 0.93; 95% CI = 0.88–0.97, for decaffeinated coffee).
High consumption of tea after diagnosis was associated with a lower all-cause mortality: compared with non-drinkers, consuming >3 cups/day was associated with a 26% (HR = 0.74; 95% CI = 0.58–0.95) lower risk (Ptrend = 0.04). However, high consumption of tea after diagnosis was not associated with breast cancer-specific mortality (Table 2). When coffee and tea intakes were included together in the model, we observed an inverse association with breast cancer-specific mortality for the consumption of coffee (>3 cups/day versus nondrinker, HR = 0.73; 95% CI = 0.57–0.94; Ptrend = 0.001) but not for tea (>3 cups/day versus nondrinker, HR = 0.76; 95% CI = 0.53–1.09; Ptrend = 0.08). For all-cause mortality, significant inverse associations were observed for both coffee and tea, mutually adjusted (>3 cups/day versus nondrinker, HR = 0.71; 95% CI = 0.60–0.84; Ptrend < 0.0001 for coffee; HR = 0.70; 95% CI = 0.55–0.90; Ptrend = 0.005 for tea).
We also examined the associations of coffee and tea consumption using only the first FFQ after diagnosis. The results were weaker than we observed using the cumulative average of post-diagnostic coffee consumption for breast cancer-specific and all-cause mortality (Supplementary Table S1).
When we considered caffeine intake from all sources, higher post-diagnostic caffeine intake was associated with lower breast cancer-specific and all-cause mortality (Supplementary Table S2).
High coffee and tea consumption before breast cancer diagnosis was not associated with breast cancer-specific or all-cause mortality (Supplementary Table S3). When we evaluated the risk of breast cancer-specific and all-cause mortality with cross-classification of pre- and post-diagnostic intake (compared with pre- and post-diagnostic non-drinkers), breast cancer-specific mortality was higher among women with high pre-diagnositic (>2 cups/day) and no post-diagnositic coffee consumption (HR = 1.91; 95% CI = 1.16–3.14) and also among women who consumed >2 cups/day of coffee before diagnosis and >0 to 2 cups/day of coffee after diagnosis (HR = 1.34; 95% CI = 1.05–1.71) (Table 3). We did not observe significant associations among women with moderate or high pre- and high post-diagnostic coffee consumption and breast cancer-specific mortality. Lower all-cause mortality was observed among women who consumed coffee before diagnosis (>0 to 2 and >2 cups/day) and consumed more than two cups/day after diagnosis (HR = 0.72; 95% CI = 0.57–0.91 and HR = 0.82; 95% CI = 0.71–0.96, respectively) (Table 3).
Table 3.
Changes in total coffee consumption from pre- to post-diagnosis in relation to mortality after breast cancer diagnosis in the Nurses’ Health Study and Nurses’ Health Study II (n = 8513).
| Post-diagnostic total coffee consumption | ||||||
|---|---|---|---|---|---|---|
| Nondrinker | >0 to 2 cups/day | >2 cups/day | ||||
| No. of deaths/person-year | HR (95% CI) | No. of deaths/person-year | HR (95% CI) | No. of deaths/person-year | HR (95% CI) | |
| Breast cancer-specific mortality | ||||||
| Pre-diagnostic total coffee consumption | ||||||
| Nondrinker | 110/11,157 | 1 | 43/4178 | 0.94 (0.65–1.34) | 12/1084 | 0.84 (0.46–1.55) |
| >0 to 2 cups/day | 27/1681 | 1.31 (0.85–2.02) | 243/23,413 | 1.04 (0.83–1.31) | 45/5908 | 0.72 (0.51–1.03) |
| >2 cups/day | 19/689 | 1.91 (1.16–3.14) | 192/15,555 | 1.34 (1.05–1.71) | 294/32,530 | 0.94 (0.75–1.19) |
| All-cause mortality | ||||||
| Pre-diagnostic total coffee consumption | ||||||
| Nondrinker | 252/11,157 | 1 | 101/4178 | 0.90 (0.71–1.14) | 20/1084 | 0.68 (0.42–1.08) |
| >0 to 2 cups/day | 57/1681 | 1.28 (0.95–1.71) | 637/23,413 | 1.00 (0.86–1.17) | 106/5908 | 0.72 (0.57–0.91) |
| >2 cups/day | 27/689 | 1.43 (0.95–2.13) | 476/15,555 | 1.15 (0.98–1.35) | 670/32,530 | 0.82 (0.71–0.96) |
Note. Models were stratified by cohort and adjusted for age at diagnosis (year), calendar year of diagnosis, time between diagnosis and first FFQ (year), calendar year at start of follow-up of each-2-year questionnaire cycle, pre-diagnostic BMI (<20, 20 to <22.5, 22.5 to <25, 25.0 to <30, 30 to <35, ≥35 kg/m2, missing), BMI change after diagnosis (no change (≥−0.5 to ≤0.5 kg/m2), decrease (<−0.5 kg/m2), increase (>0.5–2 kg/m2), increase (>2 kg/m2), missing), post-diagnostic smoking (never, past, current 1–14 cigarettes/day, current 15–24 cigarettes/day, current ≥25 cigarettes/day, missing), post-diagnostic physical activity (<5, 5 to <11.5, 11.5 to <22, ≥22 MET-h/week, missing), oral contraceptive use (ever, never), post-diagnostic alcohol consumption (<0.15, 0.15 to <2.0, 2.0 to 7.5, ≥7.5 g/day), post-diagnostic total energy intake (quintiles, kcal/day), pre-diagnostic menopausal status, age at menopause and postmenopausal hormone use status (premenopausal; postmenopausal, age at menopause <50 year and never postmenopausal hormone use; postmenopausal, age at menopause <50 year and past postmenopausal hormone use; postmenopausal, age at menopause <50 year and current postmenopausal hormone use; postmenopausal, age at menopause ≥50 year and never postmenopausal hormone use; postmenopausal, age at menopause ≥50 year and past postmenopausal hormone use; postmenopausal, age at menopause ≥50 year and current postmenopausal hormone use; missing), post-diagnostic aspirin use (never, past, current, missing), race (non-Hispanic white, other), stage of disease (I, II, III), ER/PR status (ER/PR-positive, ER-positive and PR-negative, ER/PR-negative, missing), radiotherapy (yes, no, missing), chemotherapy (yes, no, missing) and hormonal treatment (yes, no, missing).
We also examined whether the association between total coffee and total tea consumption and mortality differed by ER status, IR status, and molecular subtypes (Table 4). Coffee consumption was associated with similar lower breast cancer-specific mortality and all-cause mortality among women with both ER-positive and ER-negative tumours although the association for ER-negative breast cancer was not quite significant due to lower numbers of ER-negative tumours (n = 1519). Coffee consumption was associated with a lower breast cancer-specific mortality among women with IR-negative tumours (per one cup/day increase, HR = 0.88, 95% CI = 0.78–0.99), but not with IR-positive tumours (per one cup/day increase, HR = 1.06, 95% CI = 0.93–1.22; P-heterogeneity = 0.008). Furthermore, coffee consumption was associated with lower mortality among women with luminal A and B tumours but was not clearly associated among women with HER2-enriched or basal-like tumours, although there was no significant heterogeneity (P-heterogeneity = 0.86 for breast cancer-specific mortality and P-heterogeneity = 0.77 for all-cause mortality) (Table 4).
Table 4.
Post-diagnostic total coffee and total tea consumption in relation to mortality after breast cancer diagnosis in the Nurses’ Health Study and Nurses’ Health Study II, stratified by oestrogen receptor status (n = 8353), insulin receptor status (n = 2488) and molecular subtypes (n = 5294).
| Breast cancer subtype | No. of deaths | Breast cancer-specific mortality HR (95% CI) per 1 cup/day | No. of deaths | All-cause mortality HR (95% CI) per 1 cup/day |
|---|---|---|---|---|
| Total coffee | ||||
| Oestrogen receptor status | ||||
| Oestrogen receptor-positive | 756 | 0.91 (0.86–0.97) | 1831 | 0.91 (0.88–0.95) |
| Oestrogen receptor-negative | 209 | 0.97 (0.85–1.09) | 441 | 0.95 (0.87–1.04) |
| P for heterogeneity | 0.71 | 0.99 | ||
| Insulin receptor status | ||||
| Insulin receptor-positive | 174 | 1.06 (0.93–1.22) | 466 | 0.99 (0.91–1.08) |
| Insulin receptor-negative | 211 | 0.88 (0.78–0.99) | 544 | 0.92 (0.85–0.99) |
| P for heterogeneity | 0.008 | 0.12 | ||
| Molecular subtype | ||||
| Luminal A | 307 | 0.88 (0.80–0.98) | 804 | 0.90 (0.85–0.96) |
| Luminal B | 154 | 0.83 (0.72–0.97) | 345 | 0.88 (0.80–0.97) |
| HER2-enriched | 43 | 1.30 (0.92–1.84) | 80 | 1.05 (0.84–1.32) |
| Basal-like | 45 | 0.93 (0.67–1.28) | 99 | 0.91 (0.74–1.13) |
| P for heterogeneity | 0.86 | 0.77 | ||
| Total tea | ||||
| Oestrogen receptor status | ||||
| Oestrogen receptor-positive | 756 | 0.99 (0.92–1.08) | 1831 | 0.97 (0.92–1.02) |
| Oestrogen receptor-negative | 209 | 0.89 (0.75–1.05) | 441 | 0.94 (0.84–1.05) |
| P for heterogeneity | 0.28 | 0.69 | ||
| Insulin receptor status | ||||
| Insulin receptor-positive | 174 | 0.85 (0.69–1.04) | 466 | 0.91 (0.81–1.03) |
| Insulin receptor-negative | 211 | 0.99 (0.82–1.18) | 544 | 0.96 (0.85–1.07) |
| P for heterogeneity | 0.41 | 0.50 | ||
| Molecular subtype | ||||
| Luminal A | 307 | 0.96 (0.84–1.09) | 804 | 0.90 (0.83–0.98) |
| Luminal B | 154 | 1.02 (0.85–1.24) | 345 | 0.96 (0.84–1.09) |
| HER2-enriched | 43 | 0.77 (0.46–1.30) | 80 | 0.91 (0.66–1.27) |
| Basal-like | 45 | 1.18 (0.82–1.70) | 99 | 1.13 (0.89–1.44) |
| P for heterogeneity | 0.42 | 0.06 | ||
Note: Models were stratified by cohort and adjusted for age at diagnosis (year), calendar year of diagnosis, time between diagnosis and first FFQ (year), calendar year at start of follow-up of each-2-year questionnaire cycle, pre-diagnostic BMI (<20, 20 to <22.5, 22.5 to <25, 25.0 to <30, 30 to <35, ≥35 kg/m2, missing), BMI change after diagnosis (no change (≥−0.5 to ≤0.5 kg/m2), decrease (<−0.5 kg/m2), increase (>0.5–2 kg/m2), increase (>2 kg/m2), missing), post-diagnostic smoking (never, past, current 1–14 cigarettes/day, current 15–24 cigarettes/day, current ≥25 cigarettes/day, missing), post-diagnostic physical activity (<5, 5 to <11.5, 11.5 to <22, ≥22 MET-h/week, missing), oral contraceptive use (ever, never), post-diagnostic alcohol consumption (<0.15, 0.15 to <2.0, 2.0 to 7.5, ≥7.5 g/day), post-diagnostic total energy intake (quintiles, kcal/day), pre-diagnostic menopausal status, age at menopause, and postmenopausal hormone use status (premenopausal; postmenopausal, age at menopause<50 year and never postmenopausal hormone use; postmenopausal, age at menopause <50 year and past postmenopausal hormone use; postmenopausal, age at menopause <50 year and current postmenopausal hormone use; postmenopausal, age at menopause ≥50 year and never postmenopausal hormone use; postmenopausal, age at menopause ≥50 year and past postmenopausal hormone use; postmenopausal, age at menopause ≥50 year and current postmenopausal hormone use; missing), post-diagnostic aspirin use (never, past, current, missing), race (non-Hispanic white, other), stage of disease (I, II, III), ER/PR status (ER/PR-positive, ER-positive and PR-negative, ER/PR-negative, missing), radiotherapy (yes, no, missing), chemotherapy (yes, no, missing), and hormonal treatment (yes, no, missing). For oestrogen receptor status and molecular subtype analyses, we did not adjust for ER/PR status.
Stronger inverse associations for coffee consumption and breast cancer-specific and all-cause mortality were observed among women who never smoked (P-heterogeneity = 0.01 for breast cancer-specific mortality and P-heterogeneity = 0.006 for all-cause mortality). Furthermore, stronger inverse associations for coffee consumption and breast cancer-specific mortality were observed among women who drank alcohol less than 3.5 g/day after diagnosis. However, we did not observe significant interaction with age at adiagnosis, post-diagnostic BMI, stage of the disease (Supplementary Table S4) or modified AHEI (P-heterogeneity >0.05) (data not shown).
When we evaluated associations without considering at least a 12-month gap between time of diagnosis and first post-diagnostic FFQ (Supplementary Table S5), accounting for left truncation time since diagnosis (Supplementary Table S6), and with simple update instead of cumulative average (Supplementary Table S7), similar associations were observed.
Discussion
In two large prospective cohorts, we observed significant inverse dose–response associations between repeated measures of coffee consumption after breast cancer diagnosis and both breast cancer-specific and all-cause mortality. These associations were independent of other dietary factors, as well as pre-diagnostic coffee intake. Consumption of regular coffee was associated with a lower breast cancer-specific mortality, for decaffeinated coffee, power was limited given lower intake. Lower all-cause mortality was observed for both regular and decaffeinated coffee consumption. Post-diagnostic tea consumption was associated with lower all-cause mortality, but not breast cancer-specific mortality.
Associations between coffee consumption and breast cancer mortality were evaluated in several other studies.11–16 In two of them, post-diagnostic coffee consumption was evaluated.11,12 In a cohort of 576 women, ≥2 cups/day of coffee consumption after breast cancer diagnosis was only associated with a significantly decreased risk for early breast cancer events in tamoxifen-treated patients with ER-positive tumours.12 In contrast, Lehrer et al. reported that coffee consumption after breast cancer diagnosis was associated with increased mortality; however, this was based on only 96 women with breast cancer.11 Associations of pre-diagnostic coffee consumption and breast cancer mortality were evaluated among women with breast cancer13 as well as healthy populations.14–16 Although pre-diagnostic coffee consumption was not associated with breast cancer-specific or all-cause mortality among Swedish women with breast cancer,13 among healthy populations in the Cancer Prevention Study II, pre-diagnostic coffee consumption was associated with a lower breast cancer mortality in non-smokers.15 In contrast, in other prospective studies among healthy populations, including separate analyses in the NHS and NHSII, pre-diagnostic coffee consumption was not associated with breast cancer mortality.14,16 This null association may be due to the assessment of coffee consumption only before cancer diagnosis. Consistent with those studies,13,14,16 we found that pre-diagnostic coffee intake was not associated with breast cancer survival, and the inverse association between post-diagnostic intake and survival was independent of coffee consumption before breast cancer diagnosis.
Coffee is rich in biologically active compounds, including caffeine and polyphenolic antioxidants, which may improve survival through several plausible mechanisms. Hyperinsulinemia and inflammation may play important roles in reducing survival in patients with breast cancer.8–10 Both regular and decaffeinated coffee was associated with lower serum levels of C-peptide, a marker of insulin resistance,1 and inflammation;2–5 thus, the insulin-sensitising and anti-inflammatory effects of coffee may benefit patients with breast cancer. Given that added sugar is associated with hyperglycaemia and hyperinsulinemia,35 not accounting for this may underestimate the benefit of coffee. However, in our study, the association with post-diagnostic coffee intake remained significant after additional adjustment for post-diagnostic added sugar intake.
Although we found a lower risk of mortality among ER-positive breast cancer as well as luminal A and B subtypes of breast cancer, there were no significant interactions. Non-significant associations among ER-negative tumours might be due to a small number of women with ER-negative tumour and limited power of the study. Given the insulin-sensitising effect of coffee, we hypothesised that high intake of coffee might decrease the risk of mortality among IR-positive to a greater extent than IR-negative cancer. However, our findings support a greater benefit of coffee consumption for reduction of mortality among women with IR-negative tumour. These findings warrant further investigation.
Coffee consumption was associated with several potentially confounding variables. In this study, higher coffee consumption was significantly associated with higher alcohol consumption and a higher prevalence of smoking. Low-alcohol drinkers appeared to benefit from coffee consumption, as well as lifetime non-smokers. In addition, high coffee consumption was associated with a lower risk of all-cause mortality among women with different levels of coffee consumption before diagnosis, from non-drinking through more than 2 cups/day. However, a higher risk of breast cancer mortality was observed among women who decreased coffee consumption after diagnosis. This increased risk might be due to the elimination of coffee or unmeasured confounding and measurement errors despite controlling for several breast cancer mortality risk factors. A clinical trial study is recommended to confirm our findings.
Our study has several strengths. In this large prospective study with long duration of follow-up, we were able to comprehensively assess coffee and tea consumption and lifestyle factors before and after a breast cancer diagnosis. We also used a standardised method to review medical records of reported breast cancer.
Our study has also some limitations. Because this study is observational, residual confounding is possible, although our sensitivity analyses showed evidence supporting the robustness of the associations. Moreover, we adjusted for several potential predictors of breast cancer survival such as breast cancer characteristics and treatment, and other lifestyle factors associated with survival, including BMI, physical activity and aspirin intake. We also observed stronger associations after controlling for pre-diagnostic coffee consumption. The consumption of decaffeinated coffee in this population was substantially lower than the consumption of regular coffee, limiting our ability to examine the former with the same precision. Also, the majority of our cohort participants were non-Hispanic white, educated females, and therefore our findings may not be generalisable to other populations.
In conclusion, our findings suggest that coffee consumption after a breast cancer diagnosis may improve both breast cancer-specific and overall survival in a dose-dependent manner. Both regular and decaffeinated coffee appear to contribute to these lower risks. High tea consumption may improve overall survival. Confirmation of these results and investigations into potential mechanisms would be useful. These results add to a body of literature showing general benefits of coffee drinking14,25 and can provide reassurance to breast cancer survivors who are habitual coffee drinkers.
Supplementary information
Acknowledgements
We would like to thank the participants and staff of the NHS and NHSII for their valuable contributions as well as the following state cancer registries for their help: AL, AZ, AR, CA, CO, CT, DE, FL, GA, ID, IL, IN, IA, KY, LA, ME, MD, MA, MI, NE, NH, NJ, NY, NC, ND, OH, OK, OR, PA, RI, SC, TN, TX, VA, WA and WY.
Author contributions
M.S.F.: study concept and design, statistical analysis, funding acquisition, interpretation of data, drafting of the paper, critical revision of the paper for important intellectual content, and approval of final paper for submission. N.D.S.: interpretation of the data, critical revision of the paper for important intellectual content and approval of final paper for submission. B.A.R.: statistical analysis, interpretation of the data, critical revision of the paper for important intellectual content and approval of final paper for submission. W.C.W.: interpretation of the data, critical revision of the paper for important intellectual content and approval of final paper for submission. A.H.E.: interpretation of the data, critical revision of the paper for important intellectual content and approval of the final manuscript for submission. M.D.H.: interpretation of the data, critical revision of the paper for important intellectual content and approval of final paper for submission.
Ethics approval and consent to participate
The study protocol was approved by the institutional review boards of Brigham and Women’s Hospital and Harvard T.H. Chan School of Public Health (Boston, MA, USA).
Consent to publish
Completion of the questionnaire was considered to imply informed consent when the study protocol was approved in 1976 (NHS) and 1989 (NHSII) by the institutional review boards of the Brigham and Women’s Hospital (Boston, MA, USA) and Harvard T.H. Chan School of Public Health (Boston, MA, USA), and those of participating registries as required. The studies were conducted in accordance with recognised ethical guidelines (Declaration of Helsinki).
Data availability
The datasets used and analysed during this study are available from the corresponding author on reasonable request.
Competing interests
M.D.H. reported grants from FHI Solutions, nonfinancial support from Bayer AG (Bayer supplies aspirin and placebo for the Aspirin after Breast Cancer trial) and personal fees from Arla Foods (participated in a systematic review of dietary intake in Nigerian children for this company) outside the submitted work. The remaining authors declare no competing interests.
Funding information
The study was supported by the National Institutes of Health Grants (U01 CA176726, UM1 CA186107), American Institute for Cancer Research (AICR) to M.S.F., and the Breast Cancer Research Foundation (BCRF) to W.C.W. The study sponsors were not involved in the study design and collection, analysis and interpretation of the data, or the writing of the article or the decision to submit it for publication. The authors were independent of study sponsors.
Footnotes
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
These authors contributed equally: A. Heather Eliassen, Michelle D. Holmes
Supplementary information
The online version contains supplementary material available at 10.1038/s41416-021-01277-1.
References
- 1.Wu T, Willett WC, Hankinson SE, Giovannucci E. Caffeinated coffee, decaffeinated coffee, and caffeine in relation to plasma C-peptide levels, a marker of insulin secretion, in U.S. women. Diabetes Care. 2005;28:1390–1396. doi: 10.2337/diacare.28.6.1390. [DOI] [PubMed] [Google Scholar]
- 2.Lopez-Garcia E, van Dam RM, Qi L, Hu FB. Coffee consumption and markers of inflammation and endothelial dysfunction in healthy and diabetic women. Am. J. Clin. Nutr. 2006;84:888–893. doi: 10.1093/ajcn/84.4.888. [DOI] [PubMed] [Google Scholar]
- 3.Kempf K, Herder C, Erlund I, Kolb H, Martin S, Carstensen M, et al. Effects of coffee consumption on subclinical inflammation and other risk factors for type 2 diabetes: a clinical trial. Am. J. Clin. Nutr. 2010;91:950–957. doi: 10.3945/ajcn.2009.28548. [DOI] [PubMed] [Google Scholar]
- 4.Jacobs S, Kroger J, Floegel A, Boeing H, Drogan D, Pischon T, et al. Evaluation of various biomarkers as potential mediators of the association between coffee consumption and incident type 2 diabetes in the EPIC-Potsdam Study. Am. J. Clin. Nutr. 2014;100:891–900. doi: 10.3945/ajcn.113.080317. [DOI] [PubMed] [Google Scholar]
- 5.Yamashita K, Yatsuya H, Muramatsu T, Toyoshima H, Murohara T, Tamakoshi K. Association of coffee consumption with serum adiponectin, leptin, inflammation and metabolic markers in Japanese workers: a cross-sectional study. Nutr. Diabetes. 2012;2:e33. doi: 10.1038/nutd.2012.6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Martini D, Del BoC, Tassotti M, Riso P, Del Rio D, Brighenti F, et al. Coffee consumption and oxidative stress: a review of human intervention studies. Molecules. 2016;21:979. doi: 10.3390/molecules21080979. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Saintot M, Mathieu-Daude H, Astre C, Grenier J, Simony-Lafontaine J, Gerber M. Oxidant-antioxidant status in relation to survival among breast cancer patients. Int. J. Cancer. 2002;97:574–579. doi: 10.1002/ijc.10099. [DOI] [PubMed] [Google Scholar]
- 8.Duggan C, Irwin ML, Xiao L, Henderson KD, Smith AW, Baumgartner RN, et al. Associations of insulin resistance and adiponectin with mortality in women with breast cancer. J. Clin. Oncol. 2011;29:32–39. doi: 10.1200/JCO.2009.26.4473. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Goodwin PJ, Ennis M, Pritchard KI, Trudeau ME, Koo J, Madarnas Y. Fasting insulin and outcome in early-stage breast cancer: results of a prospective cohort study. J. Clin. Oncol. 2002;20:42–51. doi: 10.1200/JCO.2002.20.1.42. [DOI] [PubMed] [Google Scholar]
- 10.Pierce BL, Ballard-Barbash R, Bernstein L, Baumgartner RN, Neuhouser ML, Wener MH, et al. Elevated biomarkers of inflammation are associated with reduced survival among breast cancer patients. J. Clin. Oncol. 2009;27:3437–3444. doi: 10.1200/JCO.2008.18.9068. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Lehrer S, Green S, Rosenzweig KE. Coffee consumption associated with increased mortality of women with breast cancer. J. Caffeine Res. 2013;3:38–40. doi: 10.1089/jcr.2013.0001. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Simonsson M, Soderlind V, Henningson M, Hjertberg M, Rose C, Ingvar C, et al. Coffee prevents early events in tamoxifen-treated breast cancer patients and modulates hormone receptor status. Cancer Causes Control. 2013;24:929–940. doi: 10.1007/s10552-013-0169-1. [DOI] [PubMed] [Google Scholar]
- 13.Harris HR, Bergkvist L, Wolk A. Coffee and black tea consumption and breast cancer mortality in a cohort of Swedish women. Br. J. Cancer. 2012;107:874–878. doi: 10.1038/bjc.2012.337. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Ding M, Satija A, Bhupathiraju SN, Hu Y, Sun Q, Han J, et al. Association of coffee consumption with total and cause-specific mortality in 3 large prospective cohorts. Circulation. 2015;132:2305–2315. doi: 10.1161/CIRCULATIONAHA.115.017341. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Gapstur SM, Anderson RL, Campbell PT, Jacobs EJ, Hartman TJ, Hildebrand JS, et al. Associations of coffee drinking and cancer mortality in the cancer prevention study-II. Cancer Epidemiol. Biomark. Prev. 2017;26:1477–1486. doi: 10.1158/1055-9965.EPI-17-0353. [DOI] [PubMed] [Google Scholar]
- 16.Gunter MJ, Murphy N, Cross AJ, Dossus L, Dartois L, Fagherazzi G, et al. Coffee drinking and mortality in 10 European countries: a multinational cohort study. Ann. Intern Med. 2017;167:236–247. doi: 10.7326/M16-2945. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Kuruto-Niwa R, Inoue S, Ogawa S, Muramatsu M, Nozawa R. Effects of tea catechins on the ERE-regulated estrogenic activity. J. Agric Food Chem. 2000;48:6355–6361. doi: 10.1021/jf0008487. [DOI] [PubMed] [Google Scholar]
- 18.Trevisanato SI, Kim YI. Tea and health. Nutr. Rev. 2000;58:1–10. doi: 10.1111/j.1753-4887.2000.tb01818.x. [DOI] [PubMed] [Google Scholar]
- 19.Mujtaba T, Dou QP. Black tea polyphenols inhibit tumor proteasome activity. Vivo. 2012;26:197–202. [PMC free article] [PubMed] [Google Scholar]
- 20.Hudlikar RR, Venkadakrishnan VB, Kumar R, Thorat RA, Kannan S, Ingle AD, et al. Polymeric black tea polyphenols (PBPs) inhibit benzo(a)pyrene and 4-(methylnitrosamino)-1-(3-pyridyl)-1- butanone-induced lung carcinogenesis potentially through down-regulation of p38 and Akt phosphorylation in A/J mice. Mol. Carcinog. 2017;56:625–640. doi: 10.1002/mc.22521. [DOI] [PubMed] [Google Scholar]
- 21.Otton R, Bolin AP, Ferreira LT, Marinovic MP, Rocha ALS, Mori M. Polyphenol-rich green tea extract improves adipose tissue metabolism by down-regulating miR-335 expression and mitigating insulin resistance and inflammation. J. Nutr. Biochem. 2018;57:170–179. doi: 10.1016/j.jnutbio.2018.03.024. [DOI] [PubMed] [Google Scholar]
- 22.Salvini S, Hunter DJ, Sampson L, Stampfer MJ, Colditz GA, Rosner B, et al. Food-based validation of a dietary questionnaire: the effects of week-to-week variation in food consumption. Int. J. Epidemiol. 1989;18:858–867. doi: 10.1093/ije/18.4.858. [DOI] [PubMed] [Google Scholar]
- 23.Willett, W. & Lenart, E. Reproducibility and validity of food frequency questionnaires. In Nutritional Epidemiology (ed. Willett, W.) 96–141 (Oxford University Press, 2013).
- 24.Yuan C, Spiegelman D, Rimm EB, Rosner BA, Stampfer MJ, Barnett JB, et al. Relative Validity of nutrient intakes assessed by questionnaire, 24-hour recalls, and diet records as compared with urinary recovery and plasma concentration biomarkers: findings for women. Am. J. Epidemiol. 2018;187:1051–1063. doi: 10.1093/aje/kwx328. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Bhupathiraju SN, Pan A, Malik VS, Manson JE, Willett WC, van Dam RM, et al. Caffeinated and caffeine-free beverages and risk of type 2 diabetes. Am. J. Clin. Nutr. 2013;97:155–166. doi: 10.3945/ajcn.112.048603. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Willett, W. Implications of total energy intake for epidemiologic analyses. In Nutritional Epidemiology (ed. Willett, W.) 261–283 (Oxford University Press, 2013).
- 27.Hirko KA, Willett WC, Hankinson SE, Rosner BA, Beck AH, Tamimi RM, et al. Expression of estrogen receptor, progesterone receptor, and Ki67 in normal breast tissue in relation to subsequent risk of breast cancer. NPJ Breast Cancer. 2016;2:16032. doi: 10.1038/npjbcancer.2016.32. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Tamimi RM, Baer HJ, Marotti J, Galan M, Galaburda L, Fu Y, et al. Comparison of molecular phenotypes of ductal carcinoma in situ and invasive breast cancer. Breast Cancer Res. 2008;10:R67. doi: 10.1186/bcr2128. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Collins LC, Marotti JD, Baer HJ, Tamimi RM. Comparison of estrogen receptor results from pathology reports with results from central laboratory testing. J. Natl Cancer Inst. 2008;100:218–221. doi: 10.1093/jnci/djm270. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Wang J, Zhang X, Beck AH, Collins LC, Chen WY, Tamimi RM. Alcohol consumption and risk of breast cancer by tumor receptor expression. Horm. Cancer. 2015;6:237–246. doi: 10.1007/s12672-015-0235-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Colditz GA, Stampfer MJ, Willett WC, Stason WB, Rosner B, Hennekens CH, et al. Reproducibility and validity of self-reported menopausal status in a prospective cohort study. Am. J. Epidemiol. 1987;126:319–325. doi: 10.1093/aje/126.2.319. [DOI] [PubMed] [Google Scholar]
- 32.Durrleman S, Simon R. Flexible regression models with cubic splines. Stat. Med. 1989;8:551–561. doi: 10.1002/sim.4780080504. [DOI] [PubMed] [Google Scholar]
- 33.Fine JP, Gray RJ. A proportional hazards model for the subdistribution of a competing risk. J. Am. Stat. Assoc. 1999;94:496–509. doi: 10.1080/01621459.1999.10474144. [DOI] [Google Scholar]
- 34.Guo, C. & So, Y. Cause-specific analysis of competing risks using the PHREG procedure. SAS Institute Inc. 2018. Paper No.: SAS2159-2018. Available from: https://www.sas.com/content/dam/SAS/support/en/sas-globalforum-proceedings/2018/2159-2018.pdf.
- 35.Wang J, Light K, Henderson M, O’Loughlin J, Mathieu ME, Paradis G, et al. Consumption of added sugars from liquid but not solid sources predicts impaired glucose homeostasis and insulin resistance among youth at risk of obesity. J. Nutr. 2014;144:81–86. doi: 10.3945/jn.113.182519. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The datasets used and analysed during this study are available from the corresponding author on reasonable request.