Keywords

1 Definition of Pregnancy-Associated Breast Cancer Risk Windows

Pregnancy-associated breast cancer (PABC) is commonly defined as a breast cancer during pregnancy and up to 1 year postpartum. Some authors only include the pregnancy window, while others also include the second year postpartum or up to 5 or 10 years postpartum in the PABC risk window. The different pregnancy and postpartum risk windows reflect short- and long-term exposures and effects of pregnancy on breast cancer detection and management, e.g. mammographic density, masking, potential detection delays, diagnostic workup, treatment and survival. For most outcomes it is of relevance to separate effects in the different risk windows during pregnancy (first, second, third trimesters) and after delivery (0–6 months, 6–12 months, etc.). The pregnancy window is of particular interest for long-term follow-up of children exposed to cancer treatment in utero.

From an epidemiological point of view, it is important to realize that pregnancy-association is an exposure, and that PABC is a subgroup of young patients diagnosed with breast cancer, namely those cases diagnosed while exposed to pregnancy and lactation.

2 Incidence of Pregnancy-Associated Breast Cancer

Breast cancer, together with malignant melanoma and cervical cancer, are the most common malignancies diagnosed in pregnant or recently pregnant women [1,2,3,4,5,6,7,8]. Around 4% of women with breast cancer under age 45 are diagnosed during pregnancy or within the first year postpartum [2, 4, 9].

The incidence rate of PABC has been reported from several population-based studies (Table 9.1). There is a large variation in incidence across different populations and calendar periods. Reported estimates of PABC incidence rates range from 17.5 to 39.9 per 100,000 births; rates range from 3.0 to 7.7 during pregnancy , and from 13.8 to 32.2 during the first year postpartum [1, 3,4,5,6,7].

Table 9.1 Population-based studies estimating incidence of pregnancy-associated breast cancer
Full size table

Hence, the risk pattern before and after delivery is strikingly different, with a lower incidence during pregnancy and an increasing incidence after delivery. This risk pattern was assessed in detail by Andersson et al. [1] who reported the relative risks of PABC as: 1st trimester: 0.05 (95% CI 0.02–0.11), 2nd trimester: 0.26 (0.18–0.36), 3rd trimester: 0.72 (0.59–0.87), 0–6 months postpartum: 0.59 (0.51–0.69), 6–12 months postpartum: 1.12 (1.01–1.24) and 12–24 months postpartum: 1.10 (1.03–1.18), compared to an age- and year-matched control population. This risk pattern could reflect a biologically lower risk during pregnancy , diagnostic delays or a healthy mother effect (reverse causation) during pregnancy [10].

Several studies have shown increasing incidence rates of PABC over calendar time [2,3,4,5,6]. The increasing incidence can to a large extent be explained by increasing maternal age and the ongoing trend of postponement of childbearing to ages where breast cancer is more common [2, 4, 5]. Other than age, risk factors of PABC are largely unknown.

There are several challenges when estimating PABC incidence rates. First, the estimates depend on which denominator has been utilized for calculating the incidence rate. Most commonly, PABC incidence is expressed as number of PABC cases per 100,000 deliveries or births, but also pregnancies (including elective and spontaneous abortions), live births, and person-time at risk have been used as denominator. The denominator should ideally capture the population at risk of PABC , namely pregnant (or recently pregnant) women. The total number of pregnant women may be difficult to ascertain in a population even with a birth registry, due to high rates of spontaneous abortion in the first trimester. Hence, births (deliveries) are a more stable measure of the pregnant population. Number of deliveries is also a good estimate of number of women at risk in the postpartum period, and using the same denominator before and after delivery makes the rates comparable.

Second, the incidence estimates may differ between studies due to the inclusion or exclusion of abortions (spontaneous or induced) and stillbirths in both the numerator and denominator of the incidence rate. Such hampered case ascertainment would lead to underestimation of PABC rates during pregnancy , in particular in early trimesters. Eibye et al. [4] reported that 81% of patients diagnosed with PABC in first trimester underwent elective abortion. However, in recent years the use of therapeutic abortion is likely to have decreased as more aggressive treatments are given during pregnancy .

Third, to obtain unbiased estimates of PABC incidence, population-based individual level data on both pregnancies and breast cancer is required. In many countries, population data is available in birth registries and in cancer registries. However, in order to classify a breast cancer as PABC these two registry databases must be linkable on an individual level, which may not be administratively possible in all countries.

3 Age-Specific Incidence of Preganancy-Associated Breast Cancer and Maternal Age

Similar to the overall breast cancer incidence, the age-specific incidence rates of PABC increases over a woman’s reproductive period and is highest above age 40 [2,3,4, 6]. However, the PABC incidence also depends on the age at childbirth (i.e. age at exposure ). In most populations, the mean age at childbirth is below 30 years. Because the age distribution of PABC is the overlap between the age distributions of pregnancy (exposure ) and of breast cancer (outcome), the absolute numbers of PABC are therefore highest in ages 30–34 years [2]. In this age group, pregnancy and breast cancer is most likely to co-occur. So, although the PABC incidence rate increases with age, the absolute numbers of PABC continuously decrease to zero at menopause after which women are no longer exposed to pregnancy and PABC does not occur.

In studies of PABC, women with PABC are often compared to other premenopausal women with breast cancer, often denoted “non-PABC”. Non-PABC is usually defined as a breast cancer diagnosed in a nulliparous woman or in women more than 1 year after the latest childbirth. Women with non-PABC are thus similar to “premenopausal breast cancer” and will have an increasing age-distribution, since breast cancer becomes more common at higher ages. Hence, since the age of PABC women is shifted towards younger ages, while the age of non-PABC women is shifted towards higher ages, age at diagnosis is a very strong confounder in comparisons between PABC and non-PABC. Any comparison between PABC and non-PABC must therefore be thoroughly adjusted for differences in age at diagnosis to avoid age confounding. This can either be achieved via fine matching between PABC cases and non-PABC controls (e.g. 1-year age categories are often required) or via adjustments in the statistical analysis. Residual age confounding is a problem that many studies of PABC may have overlooked when using too broad age categories in the adjustment.

4 Prognosis of Pregnancy-Associated Breast Cancer

Several studies of varying sizes have assessed prognosis following breast cancer during pregnancy and lactation with somewhat conflicting results [8, 9, 11,12,13,14,15,16,17,18,19,20]. Two meta-analyses including both hospital-based and population-based studies found a worse prognosis in women with PABC compared to non-PABC, but the association was stronger in the postpartum period and weaker in women diagnosed during pregnancy [12, 16]. Women with PABC are more often diagnosed with advanced stage tumors and hormone receptor negative disease [8, 9, 11]. After adjustments for tumor stage and biology, the survival is similar between PABC and non-PABC, indicating that the worse prognosis reported in some studies can to a large extent be explained by adverse tumor characteristics [8, 9, 11].

Differences between studies include study setting (hospital-based, population-based), country, ages at diagnosis, calendar periods, treatments, postpartum windows, length of follow-up and study size. Population-based studies of prognosis following PABC often lack detailed information on clinical factors and treatment data, while institution-based materials often include those variables at high quality, but are at risk of selection bias. The poorer survival in cases diagnosed postpartum may partly be explained by delayed diagnosis , suboptimal treatment and lack of adjustment for several important clinical factors (see also Chap. 11).