Hostname: page-component-7479d7b7d-k7p5g Total loading time: 0 Render date: 2024-07-15T03:39:10.891Z Has data issue: false hasContentIssue false

Association of tea-drinking habits with the risk of non-Hodgkin lymphoma: a prospective cohort study among postmenopausal women

Published online by Cambridge University Press:  10 February 2022

Zikun Wang
Department of Epidemiology and Biostatistics, School of Public Health, Indiana University Bloomington, Bloomington, IN, USA
Rhonda Arthur
Department of Epidemiology and Population Health, Albert Einstein College of Medicine, Bronx, NY, USA
Aladdin H. Shadyab
Herbert Wertheim School of Public Health and Human Longevity Science, University of California, San Diego, La Jolla, CA, USA
Nazmus Saquib
Department of Research, College of Medicine, Sulaiman AlRajhi University, Al Bukayriyah, Kingdom of Saudi Arabia
Karen C. Johnson
Department of Preventive Medicine, University of Tennessee Health Science Center, Memphis, TN, USA
Linda G. Snetselaar
Department of Epidemiology, College of Public Health, University of Iowa, Iowa City, IA, USA
Lina Mu
Department of Epidemiology and Environmental Health, School of Public Health and Health Professions, University at Buffalo, Buffalo, NY, USA
Zhongxue Chen*
Department of Epidemiology and Biostatistics, School of Public Health, Indiana University Bloomington, Bloomington, IN, USA
Juhua Luo
Department of Epidemiology and Biostatistics, School of Public Health, Indiana University Bloomington, Bloomington, IN, USA
*Corresponding author: Zhongxue Chen, email
Rights & Permissions [Opens in a new window]


Although biological evidence suggests that tea consumption may protect against non-Hodgkin lymphoma (NHL), epidemiological evidence has been unclear. The aim of this study was to examine the association between tea-drinking habits and the risk of NHL in a large nationwide prospective cohort of postmenopausal US women. 68 854 women who were enrolled from 1993 through 1998 in the Women’s Health Initiative Observational Study and responded to year 3 annual follow-up questionnaire comprised the analytic cohort. Newly diagnosed NHL cases after the year 3 visit were confirmed by medical and pathology reports. Multivariable-adjusted Cox proportional hazards models were performed to assess the associations of tea-drinking habits (specifically, the amounts of caffeinated/herbal/decaffeinated tea intake) with the overall risk of NHL and three major subtypes (diffuse large B-cell lymphoma (n 195, 0·3 %), follicular lymphoma (n 128, 0·2 %) and chronic lymphocytic leukaemia/small lymphocytic lymphoma (n 51, 0·1 %)). Among 62 622 participants, a total of 663 (1·1 %) women developed NHL during a median follow-up of 16·51 (sd 6·20) years. Overall, different amounts of type-specific tea intake were not associated with the risk of NHL regardless of its histologic subtypes after adjustment for confounders. Our findings suggest that tea intake at the current consumption level does not influence the risk of NHL, regardless of its histologic types.

Research Article
Creative Commons
Creative Common License - CCCreative Common License - BY
This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (, which permits unrestricted re-use, distribution, and reproduction in any medium, provided the original work is properly cited.
© The Author(s), 2022. Published by Cambridge University Press on behalf of The Nutrition Society

Non-Hodgkin lymphoma (NHL) accounts for approximately 4·2 % of all cancer diagnoses, and about 720 000 people are living with NHL in the USA. Based on the estimate from the American Cancer Society, around 81 560 people are projected to be diagnosed with NHL in 2021(1). In addition, the incidence of NHL substantially increased in 1970 and then stabilised(Reference Groves, Linet and Travis2). However, the risks vary considerably in subtypes of NHL; in particular, the occurrence of diffuse large B-cell lymphoma remained stable, whereas the incidence of indolent B-cell lymphomas and T-cell NHL greatly rose from 1989 to 2007(Reference Van de Schans, Issa and Visser3).

Studies have revealed that NHL cases were more likely to be diagnosed at older ages(Reference Ries, Harkins and Krapcho4) among White people in the USA(Reference Miranda-Filho, Pineros and Znaor5) Moreover, chronic inflammation(Reference Ekstrom-Smedby6), including systemic lupus erythematosus (SLE) and rheumatoid arthritis (RA)(Reference Hu, Zhou and Wu7), is a known risk factor for NHL overall, except for certain subtypes of NHL. In addition, family history of any cancer is one of the established risk factors for NHL(Reference Ekstrom-Smedby6). Besides, studies indicate that lifestyle factors such as smoking(Reference Herrinton and Friedman8), higher alcohol consumption(Reference Chiu, Cerhan and Gapstur9) and lack of physical activity (PA)(Reference Hoffman-Goetz and Pedersen10) are adverse risk factors for certain NHL subtypes. However, the aetiology of most of the cases of NHL is largely unexplored(Reference Ekstrom-Smedby6). Thus, identifying modifiable risk factors for NHL is critical for NHL prevention.

Tea (Camellia sinensis; Theaceae) is one of the most popular beverages worldwide and is recognised as an antioxidant drink because of its wealth of bioactive compounds, leading to intense antioxidant activity in the body(Reference Butt, Imran and Sharif11). Traditionally, tea is defined by their formation process; for example, green tea is unfermented and black tea is completely fermented. At present, teas have been hypothesised to be associated with the risk of developing cancers, such as colorectal cancer, urinary tract cancer and NHL(Reference Chiu, Dave and Ward12Reference Parodi, Santi and Marani19). The beneficial impacts of tea are derived from diverse bioactive compounds, including flavonoids, l-theanine, theaflavins, caffeine, catechins, etc. Specifically, evidence from in vivo and in vitro studies(Reference Bertolini, Fusetti and Rabascio20Reference Kondo, Ohta and Igura22) has indicated that the aforementioned substances could exert anticancer effects protect against various types of haematologic and solid malignancies by inhibiting angiogenesis through the inhibition of cell proliferation and through mediating apoptosis of haematologic malignancies(Reference Naganuma, Kuriyama and Kakizaki16). Another major bioactive compound found in tea, caffeine (1,3,7-trimethylxanthine), has been recognised as a cancer-fighting property in humans since it can down-regulate the expression of many Wnt signalling proteins viz. cyclin D1 and apoptosis regulator Bcl-2(Reference Jabir, Islam and Tabrez23). In addition, some of these bioactive compounds in tea may act synergistically in anticancer action(Reference Bertolini, Fusetti and Rabascio20).

Thus far, epidemiological evidence on the association between tea intake and the risk of developing NHL is inconclusive(Reference Chiu, Dave and Ward12,Reference Zheng, Doyle and Kushi13,Reference Chang, Smedby and Zhang15Reference Parodi, Santi and Marani19,Reference Franceschi, Serraino and Carbone24) . A Japanese study showed that tea intake was statistically significantly related to a lower risk of NHL among male and female aged 40–79 years(Reference Naganuma, Kuriyama and Kakizaki16). However, a case–control study performed in the northeastern part of Italy(Reference Franceschi, Serraino and Carbone24) indicated that caffeinated tea but not decaffeinated or herbal tea was positively associated with a higher risk of NHL. Other observational studies showed no associations between tea intake and risk of NHL(Reference Chiu, Dave and Ward12,Reference Zheng, Doyle and Kushi13,Reference Chang, Smedby and Zhang15,Reference Ugai, Matsuo and Sawada17Reference Parodi, Santi and Marani19) . These inconsistencies may be due to small numbers of cases, different scales to assess tea intake, different types of tea consumption or different study populations. Besides, the evidence on examining associations between tea-drinking habits and NHL subtypes has also been insufficient. On the other hand, herbal tea, made from a wide variety of dried fruits, flowers, spices or herbal infusions rather than Camellia sinensis, contains many monomeric flavonoids(Reference Lallement and Bezanger25). Despite evidence indicates that a possible beneficial association has been observed between herbal tea and solid tumour risk(Reference Green, de Dauwe and Boyle26), the epidemiological evidence remains scarce for haematologic malignancies at present.

Existing evidence also suggests that the level of tea consumption and NHL risk or mortality may vary across ethnicity. Notably, tea consumption was about 8·5 times higher in Asian compared with White(27); whereas higher risk and higher mortality rate of NHL were found among White populations, compared with Asian populations(Reference Miranda-Filho, Pineros and Znaor5). The difference in the occurrence of malignant lymphomas between different racial groups may be partly explained by tea-drinking habits. Additionally, previous findings indicate that tea protects against cigarette smoke-induced apoptosis(Reference Banerjee, Maity and Mukherjee28). Moreover, tea consumption is correlated with age(Reference Schwarz, Bischof and Kunze29), and it is related to decreased risk of obesity(Reference Kao, Chang and Lee30). However, previous studies on NHL have not investigated the effect modification on tea intake by age, smoking status, alcohol consumption, BMI or racial group.

Given these mixed findings and limitations, we aimed to assess whether tea-drinking habits, such as the amount of type-specific tea intake, were associated with risks of NHL and its major subtypes among the Women’s Health Initiative Observational Study (WHI-OS)(31).


Study population

The Women’s Health Initiative (WHI) is a large, prospective study with long-term follow-up, which consists of three clinical trials (n 68 132) and an observational study (n 93 676). The study cohort comprised 161 808 postmenopausal women aged 50–79 who were recruited from forty clinical sites throughout the USA between 1993 and 1998. Information on the WHI design is described elsewhere(Reference Langer, White and Lewis32). All adjudicated primary cancers for observational study participants were updated through 1 March 2019. The WHI study protocol was reviewed and approved by the Fred Hutchinson Cancer Research Center (Fred Hutch) and the institutional review board at each participating institution.

Our follow-up baseline was year 3; specifically, participants were eligible for the current analyses if they responded to the year 3 follow-up questionnaire in the WHI-OS, which included information on the amount of type-specific tea intake (caffeinated/herbal/decaffeinated tea). Other information on covariates was collected from the year 3 follow-up questionnaire, except for education level, pack-years of smoking group, total energy intake, hypertension status and treatments, RA, SLE, racial or ethnic group, family history of any type of cancer and type 2 diabetes mellitus (T2DM), which were measured at WHI-OS baseline. In total, 82 568 postmenopausal women were initially included in the analysis. Moreover, we excluded women who reported any cancer history prior to annual visit year 3 (other than non-melanoma skin cancer, n 8819), women who died or were lost to follow-up or developed NHL before year 3 visit (i.e. women who reported NHL history at baseline WHI screening visit or who had incident NHL between WHI baseline and year 3, n 313) and women who had missing responses for primary exposures of interest, such as regular (caffeinated) tea intake (n 1651), herbal tea intake (n 2277) and decaffeinated tea intake(n 2865), the sample comprised 68 854 women. We further excluded 6232 women with missing data on the main covariate variables (racial and ethnic group (n 181), education levels (n 517), recreational PA (n 104), pack-years of smoking (n 2392), alcohol use (n 208), hormone therapy use for last 2 years (HRT, n 1341), hypertension status and treatments (n 1119), RA (n 470), SLE (n 815)). The final analytic cohort comprised 62 622 participants (Fig. 1).

Fig. 1. Algorithm of study participants for tea intake and the risk of non-Hodgkin lymphoma.

Exposure measures

Tea-drinking habits were recorded and collected from self-administered standardised observational study questionnaires at year 3. The information addressing habitual tea intake asked about: ‘How many cups of regular tea (not herbal, not decaf) did you drink per week in past 3 months, ‘How many cups of herbal tea did you drink per week in past 3 months’ and ‘How many cups of decaffeinated tea did you drink per week in past 3 months’. The information on the frequency of tea-drinking was collapsed for analyses to four levels (non-drinkers, < 1 cup per week, 1–6 cups per week or ≥ 7 cups per week) (1 cup = 0·237 l) because the number of participants in the latter levels was relatively sparse. In addition, tea-drinking statuses were merged into two levels (non-drinkers, tea drinkers) for chronic lymphocytic leukaemia/small lymphocytic lymphoma in herbal tea intake and decaffeinated tea intake due to sparse data.

Outcome ascertainment

The primary outcome of interest was the incidence of NHL. In general, new NHL cases, occurring after annual visit year 3, at which time our exposure variables (tea-drinking habits) were reported, was collected. Newly diagnosed NHL patients were first identified by self-reports through mail or telephone questionnaires and third-party reports and were further confirmed by centralised trained cancer adjudicators based on medical records and pathology reports from hospitals or laboratories(Reference Curb, McTiernan and Heckbert33). Women who self-reported a different diagnosis but were found upon review of medical pathology records to have NHL were considered as NHL cases. Classification and histology of tumours in the WHI were based on the Surveillance, Epidemiology, and End Results guidelines. The case morphology classification code used ICD-O-3 coding(Reference Curb, McTiernan and Heckbert33).

Three major subtypes of NHL were further examined in the current study, including diffuse large B-cell lymphoma (ICD-O-3 codes: 9678/3–9680/3 and 9684/3), follicular lymphoma (ICD-O-3 codes: 9690/3–9698/3) and chronic lymphocytic leukaemia/small lymphocytic lymphoma (ICD-O-3 codes: 9823/3 and 9670/3).

Follow-up duration was defined from the year 3 visit to the date of NHL diagnosis, the date of death or the last date of follow-up (1 March 2019), whichever occurred first.

Other covariates

Potential confounders(Reference Ries, Harkins and Krapcho4Reference Ekstrom-Smedby6,Reference Herrinton and Friedman8Reference Hoffman-Goetz and Pedersen10) in the models included socio-demographic variables: age group (50–59, 60–69 and 70–79+), racial or ethnic group (American Indian or Alaska Native, Asian or Pacific Islander, Black or African American, Hispanic White/Latino, White (not of Hispanic origin) and other), educational level (high school diploma and below, school after high school, college degree or higher)); lifestyle risk factors: BMI group (BMI, body weight in kilograms divided by squared body height in meters: < 25 kg/m2, 25 to < 30 kg/m2, 30 to < 35 kg/m2, ≥ 35 kg/m2), total energy intake (kcal/d), alcohol use (non-drinkers or past drinkers, ≤ 3 drinks per month, 1–4 drinks per week and ≥ 5 drinks per week), pack-years of smoking (non-smokers, < 5, 5–20, ≥ 20), recreational PA (total metabolic equivalents of energy expenditures per week (MET-hours/week)(Reference Hoffman-Goetz and Pedersen10), PA: < 5, 5–10, 10–20, 20–30 and > 30); and clinical risk factors/current medical conditions factors: hypertension status and treatment (no hypertension, untreated hypertension, treated hypertension), family history of any cancer (yes/no), hormone therapy use for last 2 years (HT: never use, oestrogen alone use, progesterone alone use, oestrogen or progesterone use, both use), RA (yes/no), SLE (yes/no) and T2DM (no diabetes, diabetes without treatment, diabetes with treatment).

Statistical analysis

Socio-demographic factors, lifestyle factors and medical conditions were examined by tea intake status. The frequencies of those selected categorical risk factors with percentage N (%) were presented. Mean and standard deviation of the selected continuous risk factors were calculated.

Incidence rates of newly diagnosed NHL and three subtypes of NHL were estimated using person-years as the denominator (number of cases/person-years × 10 000). Cox proportional hazards regression models were applied to evaluate associations between the amount of type-specific tea intake (caffeinated/herbal/decaffeinated tea) and the risk of NHL. Moreover, associations of tea-drinking habits with risks of the three major subtypes of NHL were further examined (diffuse large B-cell lymphoma: n 195, 0·3 %; follicular lymphoma: n 128, 0·2 % and chronic lymphocytic leukaemia/small lymphocytic lymphoma: n 51, 0·1 %). The rest of the NHL cases (n 289) were considered as a heterogeneous group and were not assessed in separate analyses. To avoid confounding, age-adjusted and multivariable-adjusted models were conducted. Hazard ratios and 95 % CI are presented. In the multivariable-adjusted models, potential confounders(Reference Ries, Harkins and Krapcho4Reference Ekstrom-Smedby6,Reference Herrinton and Friedman8Reference Hoffman-Goetz and Pedersen10) , including age group, racial or ethnic group, education, BMI group, alcohol use, pack-years of smoking group, PA, hypertension status and treatments, family history of any cancer, HT, RA, SLE and T2DM and treatments, were adjusted. Tests for trends were performed to determine whether or not a significant dose–response effect exists by modelling the ordinal variables as continuous variables. Interactions between tea intake (caffeinated/herbal/decaffeinated tea), selected categorical covariates (age group (50–59, 60–69 and 70–79+), BMI group (< 25 kg/m2, 25 to < 30 kg/m2, 30 to < 35 kg/m2, ≥ 35 kg/m2), pack-years of smoking group (non-smokers, < 5, 5–20, ≥ 20), racial or ethnic group (White (not of Hispanic origin)/others) and alcohol use (non-drinkers or past drinkers, ≤ 3 drinks per month, 1–4 drinks per week and ≥ 5 drinks per week)) and risk of NHL were examined in the multivariable-adjusted models. Additionally, mode imputation, a popular method that replaces missing data for categorical variables with mode values of the corresponding variables, was applied to reduce bias(Reference Batista and Monard34). Furthermore, we repeated the analyses for associations of tea-drinking habits (caffeinated/herbal/decaffeinated tea) with risks of NHL after excluding the first 5 years of follow-up to minimise possible reverse causality.

All analyses were conducted using SAS statistical software, version 9.4 (SAS Institute). P-values were two-tailed, and P-values ≤ 0·05 were considered statistically significant.


During a median 16·51-year follow-up (range, 0·01–21·22 years, (sd 6·20) years), 663 (1·1 %) were diagnosed with NHL among 62 622 participants.

Overall, frequent regular tea drinkers (≥ 7 cups/week) tended to be individuals who were Asian/Pacific Islanders (23·6 %), who had the highest education level (15·6 %), with the oldest age (70–70+) (16·5 %), who had the lowest BMI (< 25) (16·6 %) and never used hormone therapy at last 2 years (16·1 %) and who did not have T2DM (15·6 %), while non-tea drinkers were more likely to be non-alcohol drinkers or past drinkers (51·4 %), heavy smokers (50·5 %), with the highest physically active (≥ 30 METs-hours/week) (49·1 %) (Table 1). Similar distributions were obtained for herbal tea intake and decaffeinated tea intake, except for hypertension and treatment status (online Supplementary Tables S1S2); especially, individuals with treated hypertension were more commonly found in non-herbal tea drinkers (64·4 %), while women with untreated hypertension were commonly observed among non-decaffeinated tea consumers (76·9 %).

Table 1. General characteristics of study participants at year 3 according to regular tea consumed among the postmenopausal women, Women’s Health Initiative Observational Study

(Numbers and percentages, n 62 622)

The number of person-years by main exposures of interest and incidence rates is summarised in Tables 2 and 3 and online Supplementary Table S3. Besides, results from Cox proportional hazard regression models examining associations of caffeinated tea intake with risk of NHL overall are summarised in Table 2. There were no significant associations between caffeinated tea intake and occurrence of NHL, irrespective of histologic subtype. Similar associations were observed for decaffeinated tea intake (online Supplementary Table S3). Although the increased amount of herbal tea intake was non-significantly associated with the decreased risks of NHL (hazard ratios (95 % CI) for < 1 cup per week, 1–6 cups per week and ≥ 7 cups per week: 1·19 (0·96–1·46), 1·01 (0·82, 1·24), 0·74 (0·52, 1·05); P-value: 0·10) after controlling for potential confounders (Table 3), there were no significant dose–response associations among women who had tea-drinking habits in current analyses (Tables 2, online Supplementary Table S3, P-values for trend test > 0·05), and results were similarly null for NHL subtypes. Additionally, we searched for interactions between tea-drinking habits, selected categorical variables (age group, BMI group, pack-years of smoking group, racial or ethnic group and alcohol use) and risk of NHL, no interactions were found (online Supplementary Tables S4S8, P-values > 0·05). Furthermore, the results obtained by the mode imputation remained similar, and no significant changes were found between habitual tea intake and overall risks of NHL and its subtypes (P-value > 0·05) regardless of the amounts of tea intake and types of tea intake in the age-adjusted model and multivariable-adjusted models (data not shown). In order to minimise possible reverse causality, we performed additional analyses and excluded the first 5 years after being exposed to tea intake; however, the associations between tea-drinking habits and risk of NHL remained non-significant (data not shown).

Table 2. Associations of non-Hodgkin lymphoma (NHL) and major subtypes with regular tea intake among 62 622 postmenopausal women

(Hazard ratios and 95 % confidence intervals)

* Model adjusted for age group (50–59, 60–69 and 70–79+) only.

Model adjusted for all socio-demographic risk factors, lifestyle risk factors, clinical risk factors and current medical conditions.

P-value was estimated from Cox proportional hazard model across the total amount of regular tea consumed (categorical variable) at year 3.

§ P-value was estimated from linear trend test.

Table 3. Associations of non-Hodgkin lymphoma (NHL) and major subtypes with herbal tea intake among 62 622 postmenopausal women

(Hazard ratios and 95 % confidence intervals)

* Model adjusted for age group (50–59, 60–69 and 70–79+) only.

Model adjusted for all socio-demographic risk factors, lifestyle risk factors, clinical risk factors and current medical conditions.

P-value was estimated from Cox proportional hazard model across the total amount of herbal tea consumed (categorical variable) at year 3.

§ P-value was estimated from linear trend test.


Overall, in a large cohort of postmenopausal women with more than 20 years of follow-up, different amounts of type-specific tea intake (caffeinated/herbal/decaffeinated tea) were not associated with the risk of NHL irrespective of histology.

Thus far, findings on the associations between the different amounts of type-specific tea intake and the risk of NHL have been inconsistent(Reference Chiu, Dave and Ward12,Reference Zheng, Doyle and Kushi13,Reference Chang, Smedby and Zhang15Reference Parodi, Santi and Marani19) . The inconsistencies may be partly due to the certain tea types. For example, a beneficial association of heavy green tea intake on the risk of NHL was observed in one Japanese study(Reference Naganuma, Kuriyama and Kakizaki16). Green tea is especially rich in polyphenols (catechins and gallocatechin), and these compounds can induce apoptosis in lymphoma cells by inhibiting DNA replication(Reference Fang, Wang and Ai35,Reference Siddiqui, Adhami and Bharali36) . However, all analyses among Western countries, such as Sweden(Reference Chang, Smedby and Zhang15) and USA(Reference Chiu, Dave and Ward12,Reference Ward, Zahm and Weisenburger18) , where black tea-drinking is prevalent, there were no statistically significant associations between tea intake and risk of NHL. Existing evidence suggests that green teas have a higher concentration of antioxidants than black tea because they are the least processed tea with slight natural oxidation(Reference Lee, Lee and Lee37). Additionally, antioxidant properties in decaffeinated tea have been reduced during decaffeination(Reference Sang, Lambert and Ho38). Thus, the differences in types of tea intake are likely to account for inconsistent associations. However, we could not adjust for the specific formation process due to a lack of information on whether those teas are green tea or black tea in the current study. Besides, the study which was conducted in Japan indicated that participants who had heavy tea consumptions (≥ 5 cups/d) had a significantly decreased risk of NHL. In contrast, among this population, the overall tea intake level is relatively low compared with the Japanese study(Reference Naganuma, Kuriyama and Kakizaki16).

The inconsistencies in findings across studies might also be due to differences in exposure assessment; for example, the definition of tea intake (current v. non-current or ever v. never or cumulative lifetime intake) and differences in the distribution of NHL subtypes could also contribute to inconsistency. Moreover, lacking information about tea preparation methods might lead to inconsistencies. In particular, studies indicate that amounts of epigallocatechin gallate are almost equal in the first two cups brewed, but it can be largely declined in the third cup(Reference Wiseman, Mulder and Rietveld39). However, the relevant information was not asked for in the WHI(Reference Fang, Wang and Ai35,Reference Siddiqui, Adhami and Bharali36) .

Additionally, some clinical studies indicate that many drug interactions are caused by Camellia sinensis since its constituents can interact with the cytochrome P450 system(Reference Albassam and Markowitz40), which can be related to drug metabolism and occurrence of NHL. Although hypertension treatments, T2DM treatments and hormone therapy were adjusted in the multivariable-adjusted analyses, postmenopausal women may take many other medications, and we had limited power to examine each specific medication for the current study.

As mentioned previously, although herbal tea is called tea, it is quite different from traditional tea. This type of tea contains the major bioactive compounds known as monomeric flavonoids, such as anthocyanin, biochanin A, carotenoids, coumarins, curcumins, ellagic acid, plant sterols and terpenoids, which may protect against several chronic diseases(Reference Craig41). In accordance with the existing evidence, we observed that the risk of NHL decreased with an increase in the cups of herbal tea intake, albeit the inverse dose–response association was not statistically significant. Our findings were probably due to the fact that women who had habitual herbal tea intake might be more health conscious and would be more likely to choose a healthy lifestyle. This hypothesis was likely true and could be confirmed by examining the distribution of several health-related lifestyle factors(Reference Simard, Baecklund and Chang42), such as BMI, physical activities, smoking status and alcohol status according to herbal tea consumption (online Supplementary Table S9). Although previous case–control studies reported a significant inverse association between herbal tea intake and solid cancer risk(Reference Riza, Linos and Petralias43), further research is needed to evaluate the association between specific herbal teas and risks of NHL and certain subtypes of NHL.

The major strength of our study is the use of the prospective study design with a relatively large number of NHL cases ascertained over a long follow-up period, and only about 4 % of self-reported NHL cases did not get adjudicated because of not being able to get their medical records(44). This study had rich data on a wide range of potential confounders; besides, in order to avoid potential residual confounders, continuous variables, such as age, BMI and pack-years of smoking, had been used to replace categorical age, BMI and smoking status in multivariable-adjusted models, although results remained similar. In addition, important interactions between tea-drinking habits and covariates, such as smoking history, alcohol consumption, racial or ethnic group, BMI and age, were examined. To reduce bias, the imputation method was further conducted in this work, and categorical missing data were replaced by mode after applying standard complete-case analysis to the filled-in data, although no significant associations were found. Moreover, we further conducted lag analyses and excluded the first 5 years of follow-up after the habitual tea intake record to minimise reverse causality bias.

However, there are some limitations and pitfalls, which should be considered. First, our findings for tea intake should be interpreted with caution because information on tea intake was self-reported and could not be validated since it was not captured on the diet assessment tools. Thus, potential recall bias may be present in this study. However, observational epidemiological study of tea intake usually relies on self-reported exposure data since there is no other practical way of assessing tea intake at present. In addition, we only assessed tea information at year 3 of follow-up and did not examine changes in tea-drinking habits (caffeinated/decaffeinated tea were available), whereas the risk of NHL is expected to associate with cumulative tea intake over one’s lifetime rather than a 3-month window. As a control measure, we compared the tea intake at year 3 with year 6 and year 8 in the WHI-OS cohort. The patterns and distributions of tea intake remained similar between these time points. Therefore, we have some confidence that the tea-drinking habits among women in WHI-OS remained stable over time. Information on whether participants drank traditional tea products or tea beverages (i.e. iced tea, sweet tea, etc.) or certain types of tea (i.e. black, green or other types of non-herbal teas) was not available. Since men and younger women were not included in the current study, we could not examine sex differences, so our findings may not be generalised to other populations.

In conclusion, the results from our study indicate that habitual tea intake at the current consumption level was not associated with risk of NHL in the US postmenopausal women. More epidemiological studies, which include detailed assessments on the types of tea or tea preparation methods, are needed to further explore the associations between tea-drinking habits and the risk of NHL.


This work was supported by Women’s Health Initiative. The WHI program is funded by the National Heart, Lung, and Blood Institute, National Institutes of Health, US Department of Health and Human Services through contracts HHSN268201600018C, HHSN268201600001C, HHSN268201600002C, HHSN268201600003C and HHSN268201600004C. We thank the following WHI investigators for their expertise and help. Program Office: (National Heart, Lung, and Blood Institute, Bethesda, Maryland) Jacques Rossouw, Shari Ludlam, Joan McGowan, Leslie Ford, Nancy Geller. Clinical Coordinating Center: (Fred Hutchinson Cancer Research Center, Seattle, WA) Garnet L. Anderson, Ross Prentice, Charles Kooperberg, Lisa Johnson, Andrea LaCroix, Lesley Tinker, Marian Neuhouser, Susan Heckbert, Alex Reiner, Chongzhi Di, Xiaoling Song, Wayne Rosamond, Shirley Beresford, Chu Chen, Barbara Cochrane. Investigators and Academic Centers: (Brigham and Women’s Hospital, Harvard Medical School, Boston, MA) JoAnn E. Manson, Shari Bassuk, Howard Sesso, Lu Wang; (MedStar Health Research Institute, Washington, DC) Barbara V. Howard; (Stanford Prevention Research Center, Stanford, CA) Marcia Stefanick, Mark Hlatky, Marco Perez, Themistocles (Tim) Assimes and Jean Tang; (The Ohio State University, Columbus, OH) Rebecca Jackson, Randall Harris, Electra Paskett, W. Jerry Mysiw, Michael Blumenfeld; (University of Arizona, Tucson/Phoenix, AZ) Cynthia A. Thomson, Tamsen Bassford, Cheryl Ritenbaugh, Zhao Chen, Marcia Ko; (University at Buffalo, Buffalo, NY) Jean Wactawski-Wende, Michael LaMonte, Amy Millen, Heather Ochs-Balcom, Christopher Andrews; (University of Florida, Gainesville/Jacksonville, FL) Marian Limacher, Michael Perri, Andrew Kaunitz, R. Stan Williams, Yvonne Brinson; (University of Iowa, Iowa City/Davenport, IA) Jennifer Robinson, Robert Wallace, James Torner, Susan Johnson, Linda Snetselaar; (University of Nevada, Reno, NV) Robert Brunner, Sandra Daugherty1; (University of Pittsburgh, Pittsburgh, PA) Lewis Kuller, Jane Cauley, N. Carole Milas; (Wake Forest University School of Medicine, WinstonSalem, NC) Sally Shumaker, Stephen Rapp, Claudine Legault, Mark Espeland, Laura Coker, Michelle Naughton. Women’s Health Initiative Memory Study: (Wake Forest University School of Medicine, Winston-Salem, NC) Mark Espeland, Sally Shumaker, Stephen Rapp, Claudine Legault, Laura Coker, Michelle Naughton. All participants signed informed consent. The original protocol was approved by all forty institutional review boards (IRB) and was obtained from each of the participating WHI sites and from the Fred Hutchinson Cancer Research Center, which currently serves as the IRB of record for the WHI. Federalwide assurance number is 00001920, and the registration numbers for each of the IRBs are as follows: Fred Hutchinson Cancer Research Center IRB registration no. Com A (00000021), Com B (00000022), Com C (00005619), Com D (00009831).

There is no funding to disclose.

The authors’ responsibilities were as follows: Z. W., J. L. and Z. C. substantially designed the research; Z. K. W., J. H. L. and Z.X.C. acquired the data and performed the analyses; Z. W., J. L. and Z. C. participated in the interpretation of the results; Z. W., J. L. and Z. C. drafted the manuscript; all authors made critical revisions to the manuscript for intellectual content; J. L. and Z. C. supervised all aspects of the conduct of the study. All authors read and approved the final manuscript.

The authors declare that there is no conflict of interest.

Supplementary material

For supplementary material referred to in this article, please visit


Society AC (2021) Cancer Facts & Figures 2021. American Cancer Society. (accessed February 2021).Google Scholar
Groves, FD, Linet, MS, Travis, LB, etal. (2000) Cancer surveillance series: non-Hodgkin’s lymphoma incidence by histologic subtype in the United States from 1978 through 1995. J Natl Cancer Inst 92, 12401251.CrossRefGoogle ScholarPubMed
Van de Schans, SAM, Issa, DE, Visser, O, etal. (2012) Diverging trends in incidence and mortality, and improved survival of non-Hodgkin’s lymphoma, in the Netherlands, 1989–2007. Ann Oncol 23, 171182.CrossRefGoogle ScholarPubMed
Ries, LA, Harkins, D, Krapcho, M, etal. (2006) SEER Cancer Statistics Review, 1975–2003. Bethesda, MD: National Cancer Institute. (accessed November 2005).Google Scholar
Miranda-Filho, A, Pineros, M, Znaor, A, etal. (2019) Global patterns and trends in the incidence of non-Hodgkin lymphoma. Cancer Causes Control 30, 489499.CrossRefGoogle ScholarPubMed
Ekstrom-Smedby, K (2006) Epidemiology and etiology of non-Hodgkin lymphoma – a review. Acta Oncol 45, 258271.CrossRefGoogle ScholarPubMed
Hu, S, Zhou, D, Wu, Y, etal. (2019) Autoimmune disease-associated non-Hodgkin’s lymphoma-a large retrospective study from China. Ann Hematol 98, 445455.CrossRefGoogle ScholarPubMed
Herrinton, LJ & Friedman, GD (1998) Cigarette smoking and risk of non-Hodgkin’s lymphoma subtypes. Cancer Epidemiol Prev Biomarker 7, 2528.Google ScholarPubMed
Chiu, BC, Cerhan, JR, Gapstur, SM, etal. (1999) Alcohol consumption and non-Hodgkin lymphoma in a cohort of older women. Br J Cancer 80, 14761482.CrossRefGoogle Scholar
Hoffman-Goetz, L & Pedersen, BK (1994) Exercise and the immune system: a model of the stress response? Immunol Today 15, 382387.CrossRefGoogle Scholar
Butt, MS, Imran, A, Sharif, MK, etal. (2014) Black tea polyphenols: a mechanistic treatise. Crit Rev Food Sci Nutr 54, 10021011.CrossRefGoogle ScholarPubMed
Chiu, BC, Dave, BJ, Ward, MH, etal. (2008) Dietary factors and risk of t(14;18)-defined subgroups of non-Hodgkin lymphoma. Cancer Causes Control 19, 859867.CrossRefGoogle ScholarPubMed
Zheng, W, Doyle, TJ, Kushi, LH, etal. (1996) Tea consumption and cancer incidence in a prospective cohort study of postmenopausal women. Am J Epidemiol 144, 175182.CrossRefGoogle Scholar
Sun, CL, Yuan, JM, Koh, WP, etal. (2006) Green tea, black tea and colorectal cancer risk: a meta-analysis of epidemiologic studies. Carcinog 27, 13011309.CrossRefGoogle Scholar
Chang, ET, Smedby, KE, Zhang, SM, etal. (2005) Dietary factors and risk of non-Hodgkin lymphoma in men and women. Cancer Epidemiol Biomarkers Prev 14, 512520.CrossRefGoogle ScholarPubMed
Naganuma, T, Kuriyama, S, Kakizaki, M, etal. (2009) Green tea consumption and hematologic malignancies in Japan: the Ohsaki study. Am J Epidemiol 170, 730738.CrossRefGoogle ScholarPubMed
Ugai, T, Matsuo, K, Sawada, N, etal. (2017) Coffee and green tea consumption and subsequent risk of malignant lymphoma and multiple myeloma in Japan: the Japan Public Health Center-based Prospective Study. Cancer Epidemiol Biomarkers Prev 26, 13521356.CrossRefGoogle ScholarPubMed
Ward, MH, Zahm, SH, Weisenburger, DD, etal. (1994) Dietary factors and non-Hodgkin’s lymphoma in Nebraska (United States). Cancer Causes Control 5, 422432.CrossRefGoogle ScholarPubMed
Parodi, S, Santi, I, Marani, E, etal. (2016) Lifestyle factors and risk of leukemia and non-Hodgkin’s lymphoma: a case-control study. Cancer Causes Control 27, 367375.CrossRefGoogle ScholarPubMed
Bertolini, F, Fusetti, L, Rabascio, C, etal. (2000) Inhibition of angiogenesis and induction of endothelial and tumor cell apoptosis by green tea in animal models of human high-grade non-Hodgkin’s lymphoma. Leukemia 14, 14771482.CrossRefGoogle ScholarPubMed
Lee, YK, Bone, ND, Strege, AK, etal. (2004) VEGF receptor phosphorylation status and apoptosis is modulated by a green tea component, epigallocatechin-3-gallate (EGCG), in B-cell chronic lymphocytic leukemia. Blood 104, 788794.CrossRefGoogle Scholar
Kondo, T, Ohta, T, Igura, K, etal. (2002) Tea catechins inhibit angiogenesis in vitro, measured by human endothelial cell growth, migration and tube formation, through inhibition of VEGF receptor binding. Cancer Lett 180, 139144.CrossRefGoogle ScholarPubMed
Jabir, NR, Islam, MT, Tabrez, S, etal. (2018) An insight towards anticancer potential of major coffee constituents. BioFactors 44, 315326.CrossRefGoogle Scholar
Franceschi, S, Serraino, D, Carbone, A, etal. (1989) Dietary factors and non-Hodgkin’s lymphoma: a case-control study in the northeastern part of Italy. Nutr Cancer 12, 333341.CrossRefGoogle ScholarPubMed
Lallement, G & Bezanger, B (1970) Flavonoides de quelques labiate medicinales (rosmarin, menthe, sauge)(The flavonoid content of some medical plants from the labiatae family (rosemary, peppermint, sage)). Plant Med Phytother 4, 92107.Google Scholar
Green, CJ, de Dauwe, P, Boyle, T, etal. (2014) Tea, coffee, and milk consumption and colorectal cancer risk. J Epidemiol 24, 146153.CrossRefGoogle ScholarPubMed
Nations FAOU (2016) Food and Agriculture Organization of the United Nations Statistics Division. (accessed May 2021).Google Scholar
Banerjee, S, Maity, P, Mukherjee, S, etal. (2007) Black tea prevents cigarette smoke-induced apoptosis and lung damage. J Inflamm 4, 3.CrossRefGoogle ScholarPubMed
Schwarz, B, Bischof, HP & Kunze, M (1994) Coffee, tea, and lifestyle. Prev Med 23, 377384.CrossRefGoogle ScholarPubMed
Kao, YH, Chang, HH, Lee, MJ, etal. (2006) Tea, obesity, and diabetes. Mol Nutr Food Res 50, 188210.CrossRefGoogle ScholarPubMed
Study TWHI (1998) Design of the women’s health initiative clinical trial and observational study. Controlled Clin Trial 19, 61109.CrossRefGoogle Scholar
Langer, RD, White, E, Lewis, CE, etal. (2003) The women’s health initiative observational study: baseline characteristics of participants and reliability of baseline measures. Ann Epidemiol 13, S107121.CrossRefGoogle ScholarPubMed
Curb, JD, McTiernan, A, Heckbert, SR, etal. (2003) Outcomes ascertainment and adjudication methods in the women’s health initiative. Ann Epidemiol 13, S122128.CrossRefGoogle ScholarPubMed
Batista, GE & Monard, MC (2003) An analysis of four missing data treatment methods for supervised learning. Appl Artif Intell 17, 519533.CrossRefGoogle Scholar
Fang, MZ, Wang, Y, Ai, N, etal. (2003) Tea polyphenol (-)-epigallocatechin-3-gallate inhibits DNA methyltransferase and reactivates methylation-silenced genes in cancer cell lines. Cancer Res 63, 75637570.Google ScholarPubMed
Siddiqui, IA, Adhami, VM, Bharali, DJ, etal. (2009) Introducing nanochemoprevention as a novel approach for cancer control: proof of principle with green tea polyphenol epigallocatechin-3-gallate. Cancer Res 69, 17121716.CrossRefGoogle ScholarPubMed
Lee, KW, Lee, HJ & Lee, CY (2002) Antioxidant activity of black tea v. green tea. J Nutr 132, 785.CrossRefGoogle Scholar
Sang, S, Lambert, JD, Ho, CT, etal. (2011) The chemistry and biotransformation of tea constituents. Pharmacol Res 64, 8799.CrossRefGoogle ScholarPubMed
Wiseman, S, Mulder, T & Rietveld, A (2001) Tea flavonoids: bioavailability in vivo and effects on cell signaling pathways in vitro . Antioxidants Redox Signaling 3, 10091021.CrossRefGoogle ScholarPubMed
Albassam, AA & Markowitz, JS (2017) An appraisal of drug-drug interactions with green tea (Camellia sinensis). Planta Med 83, 496508.Google ScholarPubMed
Craig, WJ (1999) Health-promoting properties of common herbs. Am J Clin Nutr 70, 491S499S.CrossRefGoogle ScholarPubMed
Simard, JF, Baecklund, F, Chang, ET, etal. (2013) Lifestyle factors, autoimmune disease and family history in prognosis of non-hodgkin lymphoma overall and subtypes. Int J Cancer 132, 26592666.CrossRefGoogle ScholarPubMed
Riza, E, Linos, A, Petralias, A, etal. (2015) The effect of Greek herbal tea consumption on thyroid cancer: a case-control study. Eur J Public Health 25, 10011005.CrossRefGoogle ScholarPubMed
Women’s Health Initiative 2019 (2021) Annual Progress Report. Women’s Health Initiative 2019 Annual Progress Report. (accessed October 2021).Google Scholar
Figure 0

Fig. 1. Algorithm of study participants for tea intake and the risk of non-Hodgkin lymphoma.

Figure 1

Table 1. General characteristics of study participants at year 3 according to regular tea consumed among the postmenopausal women, Women’s Health Initiative Observational Study(Numbers and percentages, n 62 622)

Figure 2

Table 2. Associations of non-Hodgkin lymphoma (NHL) and major subtypes with regular tea intake among 62 622 postmenopausal women(Hazard ratios and 95 % confidence intervals)

Figure 3

Table 3. Associations of non-Hodgkin lymphoma (NHL) and major subtypes with herbal tea intake among 62 622 postmenopausal women(Hazard ratios and 95 % confidence intervals)

Supplementary material: File

Wang et al. supplementary material

Tables S1-S9

Download Wang et al. supplementary material(File)
File 81.8 KB