Childbearing Recency and Modifiers of Premenopausal Breast Cancer Risk

First full-term pregnancy achieved before age 35 years is associated with a reduced lifetime risk of developing breast cancer in women (1). However, this protection does not occur immediately and women are initially at higher risk for breast cancer after giving birth (2), especially older first-time mothers (3). This transient increase in breast cancer lasts for at least 10 years for younger first-time mothers (<25 years) and up to 30 years for older first-time mothers (>30 years; refs. 4, 5). Excessive weight gain during pregnancy has been reported to increase breast cancer risk (6, 7), and the mechanism may be due to alterations in estrogens or other sex steroids, insulin, insulin-like growth factor-1, and adipokines (8-15). Conversely, breast-feeding has been shown to reduce breast cancer risk overall (16) but may contribute to the observed increase in breast cancer following pregnancy due to potential tumor-promoting effects of elevated prolactin levels seen in pregnancy and lactation (17). To examine if pregnancy weight gain or breast-feeding influences premenopausal breast cancer risk following childbirth, we examined data collected in a large breast cancer case-control study.

The case-control study has been described in detail in previous reports (18). In brief, cases were women with a first primary invasive breast cancer diagnosis identified from population-based cancer registries in Wisconsin, Massachusetts, and New Hampshire (1996-2001) according to protocols approved by institutional review boards at each site. Women eligible for the study resided in Wisconsin, Massachusetts (excluding metropolitan Boston), or New Hampshire, were ages 20 to 69 years at diagnosis, had a listed telephone number, and were verified by self-report to have a driver's license. Approximately 80% of eligible case women were successfully interviewed. Within each state, controls, frequency matched to cases within 5-year age strata, were randomly selected from lists of licensed drivers. To be eligible as a control, a woman must have had a listed telephone number and no personal history of breast cancer and be a licensed driver. Approximately 76% of eligible controls were successfully interviewed. A total of 3,464 premenopausal women (1,706 cases and 1,758 controls) ages <50 years at the reference date were eligible for the present analysis. For cases, the reference date was the date of diagnoses. For controls, the reference date corresponded to a date ∼1 year before the interview to reflect the average times between diagnosis and interview for the cases.

Information on reproductive history was ascertained in a structured telephone interview. Women were asked to report the dates, length, and outcome of each pregnancy. For each live birth, women were asked if the child was breast-fed and for how long. Women ages <50 years were also asked to report the number of pounds they had gained during the final or most recent live birth. The interview also elicited information on known and suspected breast cancer risk factors.

Unconditional logistic regression was used to estimate odds ratios (OR) and Wald 95% confidence intervals (95% CI) for recency of childbearing and interactions with weight gain and lactation in reference to the most recent birth. All models included terms for reference age (in 5-year categories), state of residence (Wisconsin, Massachusetts, or New Hampshire), total parity, and age at first birth. Models also included potential confounding breast cancer risk factors shown in Table 1 associated with breast cancer in these data including education, first-degree family history of breast cancer, recent prediagnosis body mass index (BMI), history of screening mammography in the 5 years before the reference age, and history of benign breast disease. Tests for linear trend were done by including ordinal variables in multivariate logistic regression models that also included terms for age, state of residence, and all other covariates. Among the 1,706 cases and 1,758 controls eligible for the analysis, data on total parity or age at first birth were missing for 2 controls. After excluding these women, a total of 1,706 cases and 1,756 controls were included in the present analysis.

The characteristics of the cases and controls included in the analysis are presented in Table 1. Compared with controls, cases had similar education, had lower parity, were older at the time of first birth, were more likely to have a family history of breast cancer, had a lower BMI, were less likely to have a mammogram in the previous 5 years, and were more likely to have benign breast disease.

The association among breast-feeding, weight gain during pregnancy, and recency with respect to the final or most recent birth and breast cancer risk is shown in Table 2. Recent childbearing was associated with a nonsignificant increased risk of breast cancer: when compared with nulliparous women, women that had given birth within 5 years of the reference date had a nonsignificant 35% increased risk of invasive breast cancer (OR, 1.35; 95% CI, 0.90-2.04), adjusting for age and all covariates, with no evidence of a trend with more proximal childbearing (P_trend = 0.14). Neither breast-feeding nor relative weight gain in pregnancy was significantly associated with breast cancer risk.

We examined whether breast-feeding or pregnancy weight gain associated with the last or most recent live birth modified the association between recent childbearing and breast cancer risk (Table 3). Among women who did not breast-feed, the risk of breast cancer was nonsignificantly elevated among women that recently gave birth (≤5 years; OR, 1.64; 95% CI, 0.97-2.78) when compared with nulliparous women. The excess risk associated with recent childbearing (≤5 years) was reduced in women that breast-fed for any duration during the last pregnancy (OR, 1.24; 95% CI, 0.80-1.91). There was no evidence for interaction between recency of childbirth (ordinal variable) and any breast-feeding among parous women (P_interaction = 0.30). For pregnancy weight gain, the elevated risk associated with recent childbirth appeared to be higher in women with a lower percent body weight gain relative to the last recorded body weight: the OR (95% CI) was 1.39 (0.80-2.42) for a weight gain of <24% and 1.21 (0.70-2.09) for a weight gain of ≥24% during the most recent pregnancy, adjusting for age and all covariates (P_interaction = 0.09). Results were similar for absolute weight gain (results not shown).

To our knowledge, this is the first study to consider whether pregnancy-associated factors have a modifying influence on the transient excess risk of breast cancer following childbirth. In line with previous studies (ORs, 1.1-1.3; refs. 2-4), we observed a modest increase in breast cancer risk associated with childbearing recency, although results in our analysis did not attain statistical significance. The transient increase in breast cancer after childbearing is often attributed to a promotional influence of pregnancy estrogens; however, as pregnancy-associated breast cancers are often estrogen receptor-negative (19), other mechanisms may be more plausible (20). Elevated levels of prolactin during pregnancy and breast-feeding may contribute to the observed increase in breast cancer following pregnancy (17). However, lactation has been found to reduce breast cancer risk overall (16) possibly by promoting terminal differentiation of breast tissue and/or suppressing ovulation and lowering breast estrogen levels (21, 22). In this analysis, we examined whether lactation has any net benefit or untoward influence on premenopausal breast cancer risk following childbirth and found, if anything, a slight attenuation in the risk associated with childbearing recency among women who breast-fed, although the test for interaction was nonsignificant. Greater weight gain during pregnancy may influence breast cancer risk secondary to alterations in estrogens or other sex steroids, insulin, insulin-like growth factor-1, and adipokines, but data have been conflicting (8-15). The present analysis suggests that pregnancy-associated breast cancer risk is not significantly altered by weight gain during the most recent live birth consistent with prior findings (23), although there was a suggestion that premenopausal breast cancer risk was lower among mothers who gained the most weight during pregnancy.

The strengths of this study include its large size and population-based design. Limitations included potential selection bias if cases and controls participated in the study differentially based on pregnancy-associated risk factors. Error in recall of lactation history and pregnancy weight gain would have attenuated all associations, although studies suggest that past weight is reported with reasonable validity (24, 25). In addition, the analysis was restricted to younger women (ages <50 years) and focused on the most recent live birth. Also, we did not attempt to specifically evaluate pregnancy-associated breast cancer, which is defined as cancer diagnosed during pregnancy or within 1 year of delivery. It is unclear whether the pathophysiology of these cancers is different from that of all other premenopausal breast cancers. Despite the large overall size of the study, relatively few women in the present analysis were of childbearing age; therefore, power in the study to evaluate recency main effects and interactions was less than optimal. In summary, we found no compelling evidence that pregnancy weight gain or breast-feeding following the final or most recent pregnancy modifies the association of childbearing recency with breast cancer risk in premenopausal women.

No potential conflicts of interest were disclosed.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank Drs. Henry Anderson, Patrick L. Remington, and Meir J. Stampfer, Laura Stephenson, and the staff of the Wisconsin Cancer Reporting System; Susan T. Gershman and the staff of the Massachusetts Tumor Registry; and John Hampton, Linda Haskins, Heidi Judge, Laura Mignone, and Shafika Abrahams-Gessel along with the study interviewers and programmers for assistance with data collection and the women who participated in this study and whose generosity made this research possible.