Reproductive Factors, Exogenous Hormone Use, and Risk of B-Cell Non-Hodgkin Lymphoma in a Cohort of Women From the European Prospective Investigation Into Cancer and Nutrition

Laura Costas; Leila Lujan-Barroso; Yolanda Benavente; Naomi E. Allen; Pilar Amiano; Eva Ardanaz; Caroline Besson; Heiner Boeing; Bas Bueno-de-Mesquita; Iris Cervenka; Renée T. Fortner; Agnès Fournier; Marc Gunter; Sophia Harlid; José María Huerta; Mats Jerkeman; Karin Jirström; Rudolf Kaaks; Anna Karakatsani; Kay-Tee Khaw; Anastasia Kotanidou; Eiliv Lund; Giovanna Masala; Amalia Mattiello; Beatrice Melin; Virginia Menéndez; Neil Murphy; Alexandra Nieters; Kim Overvad; Elio Riboli; Carlotta Sacerdote; Maria-Jose Sánchez; Julie A. Schmidt; Sabina Sieri; Anne Tjønneland; Antonia Trichopoulou; Rosario Tumino; Roel Vermeulen; Elisabete Weiderpass; Silvia de Sanjosé; Antonio Agudo; Delphine Casabonne


Am J Epidemiol. 2019;188(2):274-281. 

In This Article


In this analysis of women from a large prospective cohort study, we generally observed null associations with reproductive factors and exogenous hormone use and B-cell NHL, except for a moderately increased risk among women who at baseline reported having had a surgical menopause as compared with women who did not. Hysterectomy alone was not associated with B-cell NHL risk, while women with hysterectomy plus oophorectomy (especially bilateral oophorectomy) showed a higher risk of B-cell NHL.

The subject of the role of reproductive factors in lymphomagenesis has been controversial. Evidence for a role of hormonal factors in NHL etiology comes mainly from observational data summarized in a systematic review of the literature[10] and randomized data from the Women's Health Initiative Clinical Trial.[15] In summary, regarding observational data, 7 cohort studies,[16–25] 13 case-control studies,[26–39] and pooled analysis of 2 consortia of case-control studies[40–42] found associations between reproductive factors or hormone use and incident lymphoma. Several studies lacked detailed data on hormonal exposures and used heterogeneous definitions of lymphoma and hormonal exposures that hampered the performance of a meta-analysis.[10] The present analysis was based on detailed hormonal assessments, and it was the third largest individual study in terms of number of cases after 2 registry-based studies carried out in Sweden and Denmark, which assessed pregnancy variables and included 1,744 and 1,573 cases, respectively.[20,28]

Our finding of no association with parity is consistent with findings from the pooled analyses of the International Lymphoma Epidemiology Consortium (InterLymph), which included 3,816 cases and 5,151 controls from 18 studies.[40] We observed null results for oral contraception, in accordance with previous studies.[16,23,24] Postmenopausal hormone therapy has yielded contradictory findings. The protective role of hormone therapy observed in case-control studies[31–33,36,39,42] has not been replicated in cohort studies.[16,19,21,24,25] Additionally, in a systematic review of the literature, we concluded that the association between NHL and postmenopausal hormone therapy probably depends on the formulation used and oophorectomy status, which have been rarely assessed.[10] When these data were available, associations were derived from unopposed estrogen use, rather than use of estrogen + progestin, although there were still inconsistencies. However, in the present cohort analyses, we did not find associations between B-cell NHL and unopposed estrogen use or combined therapy. Randomized data can help in disentangling the role of hormone use in lymphoma risk and help to avoid biases commonly observed in observational studies.[43] In the Women's Health Initiative Clinical Trial, conjugated equine estrogens plus medroxyprogesterone acetate or conjugated equine estrogens alone were tested against placebo, and incidence rates of NHL were calculated by treatment group.[15] During the 13 years of follow-up, 27,229 women were randomized to treatments, and 383 incident NHL cases were identified. In that study, incidence of NHL was similar in the treatment and placebo groups. Together with our data and other cohort data, this suggests that hormone therapy is not associated with NHL.

We observed that women with surgical menopause, particularly women with hysterectomy and bilateral oophorectomy, showed a significantly higher risk of B-cell NHL. The ovaries secrete sex hormones, and several studies have suggested that women who undergo bilateral oophorectomy have reduced serum concentrations of androgens, rather than estrogens, compared with postmenopausal women with intact ovaries.[44,45] However, steroid metabolism is very complex, and in vitro data suggest that androgens could modulate NHL risk in either direction.[46,47] Epidemiologic literature on the role of oophorectomy in NHL risk or its subtypes is scarce. In the California Teachers Study cohort, Lu et al.[25] observed that women with bilateral oophorectomy had a significant 37% increased risk of NHL compared with women with natural menopause, in accordance with our results. In a case-control study evaluating risk factors for multiple myeloma, surgical menopause by hysterectomy with bilateral oophorectomy was statistically significantly associated with an 85% increased risk of multiple myeloma.[38] However, Lee et al.[31] found null results for NHL in a population-based case-control study, although women with oophorectomy and hysterectomy alone (without oophorectomy) were analyzed together and relevant misclassification of the surgical menopause may occurred. Considering our results and the 2 previous positive findings,[25,38] further assessment of bilateral oophorectomy is therefore warranted. However, the association was based on a small number of subjects, and the estimate for unilateral oophorectomy was below 1, which hampered the biological interpretation of this exposure. Therefore, this result could simply reflect type 1 error, given the large number of factors evaluated in the present analyses. Pooled analyses of cohort studies may provide the statistical power needed to corroborate or rule out these associations.

Our study was based on a large data set with a prospective design and detailed information on reproductive factors and exogenous hormone use, including formulations. In spite our relatively large sample size, associations for less common exposures among specific lymphoma subtypes relied on small numbers of cases. We had the ability to control for a variety of potential confounders, including education and body mass index, which biased previous studies of hormone therapy and disease because of confounding and a healthy user effect.[43] However, these variables may imprecisely measure complex factors such as socioeconomic factors or adiposity, and therefore residual confounding cannot completely be ruled out. A concern in this study is that menstrual and reproductive variables were based on self-reported data. However, the reliability of responses to questions on reproductive history, including self-reported oophorectomy and use of hormones, has been shown to be very high.[48] In addition, information on hormone therapy was not periodically updated; therefore, we could not evaluate incident use. Importantly, the reported association between B-cell NHL and surgical menopause may have been due to chance, because we performed multiple comparisons.

In conclusion, our prospective analysis does not support a strong role for reproductive factors or exogenous hormones in lymphomagenesis.