Live birth outcomes after adolescent and young adult breast cancer

Abstract

Reproductive outcomes are an important survivorship concern for women diagnosed with cancer as adolescents and young adults (AYAs). We examined the incidence of live birth and the prevalence of adverse birth outcomes according to tumor and treatment characteristics among AYAs with breast cancer. Women diagnosed with breast cancer at ages 15–39 during 2000–2013 were identified using the North Carolina Central Cancer Registry (n=4978). Cancer registry records were linked to state birth certificate files from 2000–2014 to identify births to women with and without a breast cancer history. The breast cancer cohort was followed until live birth, death, age 46, or December 31, 2014, whichever occurred first. For each birth to breast cancer survivors (n=338), we sampled 20 births to women without a recorded cancer diagnosis, with frequency matching on maternal age and year of delivery. The cumulative incidence of live births after breast cancer was 8% at 10 years. Births were less common among women treated with chemotherapy. Overall, the prevalence of preterm birth, low birth weight, small-for-gestational age (SGA), and Cesarean delivery did not differ substantially between births to women with and without breast cancer. However, births to women with ER-negative disease were more likely to be preterm (PR=1.84; 95% CI: 1.11–3.06). In this population-based study, <10% of AYA breast cancer survivors had a live birth within 10 years of their diagnosis. The increase in risk of preterm delivery among ER-negative survivors in our cohort warrants further investigation in larger studies.

Introduction

In the United States, breast cancer is the most common cancer diagnosed among adolescent and young adult women (AYA, ages 15–39 years),¹ accounting for over 12,000 new cases in 2015.² With recent improvements in breast cancer survival in young women,³ there is a growing number of young breast cancer survivors who face unique survivorship concerns related to the impact of a cancer diagnosis and treatment on reproductive outcomes. Increases in cancer survival coincide with continued trends toward a delayed age at first birth among U.S. women,⁴ and many women with breast cancer may not have completed childbearing by the time of their diagnosis. Though the importance of fertility counseling in younger cancer patients is increasingly recognized,⁵ there remains a paucity of population-based evidence on the incidence of live birth and birth outcomes among women with a history of breast cancer.

Pregnancy after breast cancer does not appear to reduce survival or increase risk of cancer recurrence.^6–8 However, physicians generally recommend that breast cancer patients wait at least 2 years before becoming pregnant. For women with hormone-responsive tumors receiving endocrine therapy, the recommended wait may be longer, as pregnancy is not recommended during the 5–10 year course of tamoxifen. These delays mean breast cancer survivors may be attempting to conceive at older ages when fertility is reduced and the risk of adverse birth outcomes is higher. Additionally, exposure to certain cancer treatments, particularly alkylating agent-based chemotherapies, may lead to decreased ovarian reserve,⁹ further threatening future fertility in women receiving these therapies. Although embryo and oocyte cryopreservation are accepted methods for fertility preservation in post-pubertal female cancer patients about to receive gonadotoxic therapies,⁵ these strategies remain underprescribed and underutilized.¹⁰

Recent studies from Canada¹¹ and Europe^{12, 13} have documented a decreased incidence of childbirth among breast cancer survivors relative to age-matched women in the general population. Other reports suggest that those who do give birth after breast cancer may have an increased risk of low birth weight and preterm deliveries compared to women without cancer.^14–18 However, variation in these outcomes according to factors such as hormone receptor status and endocrine therapy use has not been previously explored.

Using population-based data from North Carolina, the aims of this study were to 1) describe the incidence of live birth after an AYA breast cancer diagnosis, and 2) compare the prevalence of adverse birth outcomes between AYA breast cancer survivors and the general population. We also evaluated potential variation in these outcomes according to demographic, tumor, and treatment characteristics.

Materials and methods

We identified 4,978 women with a first primary diagnosis of breast cancer at ages 15–39 years between January 1, 2000 and December 31, 2013 in the North Carolina Central Cancer Registry (CCR). Breast cancer diagnoses were defined by the primary site codes C500–C509 and histology codes 8010–8589, using AYA-specific recodes of the International Classification of Diseases for Oncology (ICD-O-3) definitions.¹⁹ In addition to site and histology codes, information available from the CCR includes basic demographic variables, date of diagnosis, stage, primary treatments (surgery, radiation, chemotherapy, endocrine therapy) and tumor characteristics. Information on estrogen receptor (ER) status is available for women diagnosed in 2004 and later.

Using a probabilistic linkage strategy²⁰ in LinkPlus,²¹ data from the CCR were linked to North Carolina birth certificate files at the North Carolina State Center for Health Statistics to identify births to women with and without a breast cancer diagnosis. Linkage was based on maternal name, social security number, and date of birth; reliability estimates for these variables range from 96–98%. Information abstracted from birth certificates included maternal characteristics (race, parity, smoking during pregnancy, education, marital status), infant date of birth, gestational age, birthweight, mode of delivery, and plural birth. The study was approved by the institutional review board of the University of North Carolina and by the North Carolina State Center for Health Statistics.

To estimate the cumulative incidence of live birth after a breast cancer diagnosis, we included the first post-diagnosis birth to women with breast cancer where the total recorded gestational length, as indicated on the birth certificate, occurred after the woman’s date of breast cancer diagnosis. Thus births to women diagnosed during pregnancy were not considered events. For women who were diagnosed with breast cancer during pregnancy (N=91), we included the first additional live birth that was conceived after the breast cancer diagnosis. Multiple births (i.e. twins, triplets, etc.) were counted as a single event.

For analyses comparing birth outcomes between women with and without an AYA breast cancer diagnosis, we identified all singleton live births to women with breast cancer that had a gestational age of ≥20 weeks and a birthweight of ≥500 g and occurred after the date of the mother’s breast cancer diagnosis. Births to women diagnosed with breast cancer during pregnancy were excluded (N=91); however, subsequent births to these women that also occurred during the study period were included (N=6). For each included birth to women with breast cancer (N=338), we randomly sampled 20 births from 1,767,474 live births to women without a recorded cancer diagnosis, with frequency matching on year of delivery and maternal age. Outcomes included preterm birth (<37 weeks gestation), low birth weight (<2500 g), small-for-gestational-age (SGA), and Cesarean delivery. SGA was defined as a birthweight below the 10^th percentile of infants of the same sex and gestational age, according to the nomogram published by Oken et al.²²

Statistical analysis

Cox proportional hazards models were used to estimate cause-specific hazards ratios (HR) and 95% confidence intervals (CI) for first post-diagnosis live birth according to demographic, tumor, and treatment characteristics among women with breast cancer. Person-time began on the date of breast cancer diagnosis and ended on the date of first post-diagnosis live birth, death, age 46, or December 31, 2014, whichever occurred first. The assumption of proportional hazards was evaluated by visual inspection of log-log plots. This evaluation suggested evidence of non-proportional hazards in analyses according to ER status and endocrine therapy; thus these HR estimates should be interpreted as time-averaged summary measures. To estimate the cumulative incidence of live birth at 5 and 10 years after diagnosis, we used the Fine and Gray method²³ with death as a competing risk.

Multiple pregnancies per woman were included in analyses of birth outcomes. Prevalence ratios, comparing births to women with and without cancer, were estimated using Poisson regression models fitted with generalized estimating equation methods. Multivariable models were adjusted for maternal age (<30, 30–34, 35+), race (White, Black, other), education (high school or less, some college, Bachelor’s degree or higher), marital status (married, unmarried), smoking during pregnancy (any, none), and previous live births (0, 1, 2, 3+). Due to a change in the format of the birth certificate in 2010, all 2010 births were missing information on Cesarean delivery, mother’s education, and smoking during pregnancy. Therefore births occurring in 2010 were excluded from analyses of Cesarean deliveries, and all multivariable models of birth outcomes.

We used published general population birth rates for North Carolina in 2013²⁴ to compare observed birth rates among AYA breast cancer survivors to expected birth rates among women without cancer. Within each stratum of age and race/ethnicity, the number of expected births was calculated by multiplying the general population birth rate by the number of breast cancer survivors in our cohort who were alive at that age at the end of 2013. The standardized birth ratio (SBR) was then calculated as the ratio of observed to expected births, summed across strata of age and race/ethnicity. Confidence intervals were calculated using Fisher’s exact methods.²⁵

Results

The 4,978 AYA breast cancer survivors in our cohort had an average age at diagnosis of 35.0 years (SD=3.7) and contributed a total of 29,215 person-years of follow-up for first postdiagnosis live birth (median= 6.0 years/woman). A total of 293 women had at least one live birth after diagnosis, corresponding to an overall cumulative incidence of 5% and 8% at 5 and 10 years postdiagnosis, respectively (Table 1). In 2013, the age- and race-adjusted birth rate among AYA breast cancer survivors was significantly lower than that in the general population (SBR=0.43; 95% CI: 0.30–0.58). The reduction in rate was more pronounced for women with regional or distant disease (SBR=0.32; 95% CI: 0.17–0.54) than those with in situ or localized disease (SBR=0.53; 95% CI: 0.34–0.79).

Table 1.

Characteristics of AYA breast cancer survivors and estimated cumulative incidence and hazard ratios (HR) for postdiagnosis live births

	≥1 post- diagnosis birth		No post- diagnosis births		Cumulative incidence of birtha		Crude HR (95% CI)	Age-adjusted HR (95% CI)
	N	%	N	%	5-year	10-year
Total	293		4685		5%	8%	--	--
Race
White	200	68%	3179	68%	5%	9%	1	1
Black	81	28%	1296	28%	5%	8%	1.10 (0.85, 1.42)	1.03 (0.79, 1.33)
Other	12	4%	196	4%	6%	7%	1.05 (0.59, 1.89)	1.02 (0.57, 1.82)
Missing/unknown			14
Age at diagnosis
17–29	87	30%	392	8%	14%	25%	1.83 (1.40, 2.39)	1.83 (1.40, 2.39)
30–34	142	48%	1166	25%	9%	14%	1	1
35–39	64	22%	3127	67%	2%	2%	0.19 (0.14, 0.25)	0.19 (0.14, 0.25)
Summary stage
In situ	48	16%	485	10%	8%	12%	1.17 (0.85, 1.61)	1.64 (1.19, 2.28)
Localized	158	54%	1911	41%	6%	11%	1	1
Regional	79	27%	1947	42%	3%	5%	0.56 (0.43, 0.73)	0.54 (0.41, 0.70)
Distant	3	1%	269	6%	1%	2%	0.29 (0.09, 0.91)	0.27 (0.09, 0.84)
Unstaged	5	2%	71	2%
Missing/unknown	0		2
Treatment
Surgery only	72	26%	781	18%	8%	11%	1	1
Radiation, no chemotherapy	21	8%	359	8%	5%	8%	0.66 (0.41, 1.07)	0.77 (0.47, 1.25)
Chemotherapy, no radiation	94	34%	1556	35%	5%	9%	0.74 (0.55, 1.01)	0.62 (0.45, 0.84)
Radiation and chemotherapy	91	33%	1718	39%	4%	7%	0.63 (0.46, 0.86)	0.54 (0.40, 0.74)
Missing/unknown	15		271
ER statusb
Positive/borderline	104	60%	2012	66%	4%	10%	1	1
Negative	68	40%	1037	34%	5%	10%	1.35 (0.99, 1.83)	1.31 (0.97, 1.78)
Missing/unknown	22		303
Endocrine therapy (among ER+)b
No	59	61%	754	41%	7%	10%	1	1
Yes	38	39%	1076	59%	3%	11%	0.51 (0.34, 0.76)	0.47 (0.31, 0.71)
Missing/unknown	7		182
Time between diagnosis and first post-diagnosis birth (years)
Mean (SD)	3.8	(2.3)
0.8–2	71	24%
2–<3	69	24%
3–<5	80	27%
5–11.2	73	25%
Number of post-diagnosis births
1	237	81%
2	48	16%
3	6	2%
4	1	0%
5	1	0%
Previous live births (at time of first post-diagnosis birth)
0	114	39%
1	90	31%
2	50	17%
3+	39	13%

^aAccounting for death as a competing risk

^bAmong women diagnosed in 2004 and later

Women diagnosed at older ages were less likely to have a postdiagnosis birth than women diagnosed at younger ages. Compared to women aged 30–34 years at diagnosis, the HR was 1.83 (95% CI: 1.40–2.39) among those aged 17–29, and 0.19 (95% CI: 0.14–0.25) among those aged 35–39. Births were also less common among women with regional (HR=0.54; 95% CI: 0.41–0.70) or distant (HR=0.27; 95% CI: 0.09–0.84) stage disease compared to those with localized disease. In analyses of cancer treatment, women treated with chemotherapy were less likely to have a live birth. Compared to those treated with surgery only, the HRs for women who received chemotherapy were 0.54 (95% CI: 0.40–0.74) and 0.62 (95% CI: 0.45–0.84) for those with and without radiation, respectively (Table 1). In analyses restricted to women with localized or regional disease, corresponding HRs for chemotherapy with and without radiation were 0.70 (95% CI: 0.48–1.04) and 0.82 (95% CI: 0.56–1.21), respectively. Hazard ratios for live birth did not vary substantially by race.

Compared to ER-positive, women with ER-negative breast cancer were 1.31 times as likely to have a live birth (HR=1.31; 95% CI: 0.97–1.78); this was largely driven by the lower cumulative incidence of birth among ER-positive women over the first approximately 8 years of follow-up (Figure 1). However, at 10 years, the cumulative incidence of live birth was 10% in both groups. Among women with ER-positive tumors, those receiving endocrine therapy were less likely to have a live birth over most of the study period than women not receiving endocrine therapy (HR=0.47; 95% CI: 0.31–0.71), though the cumulative incidence among endocrine therapy users actually exceeded that in non-users by 10 years after diagnosis (11% vs 10%) (Figure 2).

Figure 1.

Cumulative incidence of post-diagnosis births according to estrogen receptor (ER) status among women diagnosed in 2004 and later.

Figure 2.

Cumulative incidence of post-diagnosis birth according to endocrine therapy among women with estrogen receptor-positive breast cancer diagnosed in 2004 and later.

Analyses of birth outcomes compared 338 singleton births to breast cancer survivors to 6,760 births to women without cancer. The average maternal age at delivery in both groups was 35 years (SD=4) (Table 2). Overall, the proportions of preterm birth, low birth weight, SGA, and Cesarean delivery, were similar for women with and without a breast cancer history (Table 3,4). In subgroup analyses, PRs were modestly elevated for women with invasive breast cancer relative to the noncancer cohort for both preterm birth (PR=1.29; 95% CI: 0.92–1.81) and low birth weight (PR=1.29; 95% CI: 0.89–1.87), though the prevalence of SGA did not differ substantially between groups (PR=1.08; 95% CI: 0.75–1.56) (Table 4). For women with in situ breast cancer, the small number of outcomes precluded meaningful analyses within this group.

Table 2.

Characteristics of singleton live births to breast cancer survivors and the comparison cohort

	Breast cancer survivors (n=338)		Comparison cohort (n=6760)
	N	%	N	%
Year of birth
2000–2003	22	7%	440	7%
2004–2006	63	19%	1260	19%
2007–2009	92	27%	1840	27%
2010–2012	81	24%	1620	24%
2013–2014	80	24%	1600	24%
Maternal age
Mean (SD)	35.1	(4.3)	35.1	(4.3)
<24	6	2%	120	2%
25–29	30	9%	600	9%
30–34	99	29%	1980	29%
35+	203	60%	4060	60%
Mother's race
White	227	67%	4796	71%
Black	94	28%	1179	17%
Other	17	5%	785	12%
Mother's marital status
Married	265	78%	5348	79%
Not married	73	22%	1411	21%
Missing	0		1
Mother's education
High school or less	75	24%	1750	28%
Some college	85	27%	1566	25%
Bachelor's degree or higher	158	50%	3030	48%
Missing	20		414
Maternal smoking during pregnancy
None	291	92%	5931	93%
Any	27	8%	417	7%
Missing	20		412
Previous live births
0	110	33%	1747	26%
1	118	35%	2204	33%
2	57	17%	1470	22%
3+	53	16%	1336	20%
Missing	0		3

Table 3.

Birth outcomes among AYA breast cancer survivors and the non-cancer comparison cohort

	Comparison cohort (n=6760)		Births to breast cancer survivors (N=338)		Births to ER+ breast cancer survivors (N=117)a		Births to ER− breast cancer survivors (N=71)a
	N	%	N	%	N	%	N	%
Preterm (<37 weeks)	622	9%	36	11%	12	10%	13	18%
Low birth weight (<2500 g)	474	7%	30	9%	4	3%	14	20%
Small for gestational age	601	9%	34	10%	8	7%	9	13%
Cesarean deliveryb	2318	36%	135	42%	42	40%	33	49%

Abbreviations: ER, estrogen receptor

^aAmong births to women diagnosed in 2004 and later

^bExcludes 2010 births (n=20 births to breast cancer survivors, 400 births to comparison cohort) due to missing information on Cesarean delivery and gestational hypertension)

Table 4.

Prevalence ratios (PR) and 95% confidence intervals (CI) comparing births to breast cancer survivors (n=338) to the comparison group (N=6760)

	Preterm (<37 weeks)		Low birth weight (<2500 g)		Small for gestational age		Cesarean deliveryb
	Age-adjusted	Multivariable a,b	Age-adjusted	Multivariable a,b	Age-adjusted	Multivariable a,b	Age-adjusted	Multivariable a,b
Overall	1.15 (0.83, 1.59)	1.10 (0.78, 1.54)	1.27 (0.89, 1.81)	1.11 (0.77, 1.61)	1.13 (0.80, 1.60)	1.02 (0.72, 1.45)	1.21 (1.05, 1.39)	1.14 (1.00, 1.31)
Disease extent
In situ	NC	NC	NC	NC	NC	NC	NC	NC
Invasive	1.39 (1.01, 1.92)	1.29 (0.92, 1.81)	1.53 (1.08, 2.18)	1.29 (0.89, 1.87)	1.25 (0.88, 1.78)	1.08 (0.75, 1.56)	1.20 (1.03, 1.40)	1.13 (0.97, 1.32)
ER statusc
Positive	1.04 (0.58, 1.88)	1.03 (0.54, 1.95)	0.49 (0.19, 1.30)	NC	0.76 (0.38, 1.50)	0.85 (0.44, 1.66)	1.14 (0.89, 1.46)	1.09 (0.85, 1.39)
Negative	1.97 (1.19, 3.26)	1.84 (1.11, 3.06)	2.80 (1.73, 4.52)	2.51 (1.53, 4.12)	1.43 (0.79, 2.58)	1.32 (0.72, 2.43)	1.38 (1.08, 1.77)	1.32 (1.03, 1.69)
Endocrine therapy
Yes	0.72 (0.31, 1.69)	NC	0.49 (0.16, 1.50)	NC	0.95 (0.43, 2.08)	NC	0.98 (0.71, 1.35)	1.15 (0.86, 1.54)
No	1.18 (0.83, 1.66)	1.11 (0.78, 1.58)	1.45 (1.01, 2.09)	1.23 (0.84, 1.81)	1.18 (0.82, 1.70)	1.04 (0.71, 1.51)	1.27 (1.10, 1.46)	1.19 (1.03, 1.38)
Chemotherapy
Yes	1.16 (0.78, 1.74)	1.12 (0.75, 1.69)	1.43 (0.95, 2.14)	1.26 (0.82, 1.93)	1.44 (1.00, 2.08)	1.32 (0.92, 1.91)	1.20 (1.01, 1.42)	1.13 (0.95, 1.34)
No	1.18 (0.71, 1.94)	1.12 (0.67, 1.85)	1.02 (0.54, 1.93)	0.88 (0.44, 1.74)	0.71 (0.34, 1.47)	0.57 (0.25, 1.28)	1.19 (0.96, 1.48)	1.13 (0.91, 1.40)
Time between diagnosis and birth, y
<5	1.03 (0.69, 1.55)	0.95 (0.62, 1.45)	1.05 (0.66, 1.66)	0.95 (0.59, 1.53)	0.96 (0.62, 1.49)	0.90 (0.59, 1.38)	1.19 (1.01, 1.40)	1.14 (0.97, 1.34)
≥5	1.40 (0.84, 2.33)	1.48 (0.91, 2.42)	1.73 (1.03, 2.90)	1.48 (0.83, 2.62)	1.50 (0.89, 2.52)	1.31 (0.75, 2.30)	1.18 (0.94, 1.49)	1.31 (1.05, 1.64)

Abbreviations: NC, not calculated due to small sample size; ER, estrogen receptor

^aMultivariable models adjust for maternal age, education, race, previous live births, marital status, and smoking during pregnancy

^bExcluding births in 2010

^cAmong births to women diagnosed in 2004 and later

ER-positive breast cancer was not associated with a higher prevalence of any of the four outcomes evaluated relative to the noncancer cohort. However, births to women with ER-negative disease were more likely to be preterm (PR=1.84; 95% CI: 1.11–3.06) and low birth weight (PR=2.51; 95% CI: 1.53–4.12), even after adjustment for maternal age, race, and other covariates. PRs did not differ greatly according to endocrine therapy or chemotherapy for most outcomes. However, a suggestive increase in SGA was identified among births to women treated with chemotherapy relative to the noncancer cohort (PR=1.32; 95% CI: 0.92–1.91). In general, PRs among births occurring 5 or more years after diagnosis were greater than among those occurring less than 5 years from diagnosis, but estimates were relatively imprecise.

Discussion

With trends toward delayed childbearing in the U.S., and increases in breast cancer survival in young women, there is an increasing number of reproductive-age breast cancer survivors who may wish to have children in the years following diagnosis and treatment. In this population-based study, we found that the birth rate among AYA breast cancer survivors was about 40% of that in the general population of North Carolina in 2013. Among AYAs with breast cancer, post-diagnosis live births were less common after a diagnosis of regional or distant stage disease and after treatment with chemotherapy. Compared to women with ER-negative tumors, those with ER-positive tumors had a lower cumulative incidence at 5 years, but not at 10 years, after diagnosis. After accounting for maternal age and other risk factors, we observed modest increases in preterm birth and low birth weight among births to women with invasive breast cancer relative to the noncancer comparison cohort, consistent with previous reports of women who conceived after a breast cancer diagnosis.^{17, 18}

Our overall estimates of cumulative incidence are within the range of those reported in previous population-based studies of breast cancer survivors in developed countries.^{11–13, 26, 27} However, our study reflects more recent diagnoses and treatment protocols and is, to date, the first to focus on AYAs and to evaluate characteristics such as ER status and endocrine therapy. In a study from Canada, the 10-year cumulative incidence of childbirth was 23% among breast cancer survivors diagnosed at ages 20–34 during 1992–1999.¹¹ In addition to the difference in age range, their analyses required all women to have survived at least 5 years without a recurrence. Similar studies from Northern Europe have reported postdiagnosis pregnancy proportions ranging from 1–6% among breast cancer patients diagnosed as early as the 1960s and 1970s and followed through the early 2000s.^{12, 13, 26, 27}

As expected, postdiagnosis live births in our analyses were significantly less common among women treated with chemotherapy and those with a more advanced stage of disease at diagnosis. We did not have information on the type or dose of chemotherapeutic agents. However, cyclophosphamide, an alkylating agent often used in breast cancer treatment, may cause ovarian toxicity,²⁸ potentially threatening future fertility. In addition to potential direct effects of chemotherapy, the time spent in active treatment may cause women to delay childbearing until older ages, when fertility is reduced. Though studies indicate that pregnancy after breast cancer does not adversely impact prognosis, psychological factors such as fear of recurrence and stress among women with more advanced disease and/or more intensive treatments may also contribute to the lower incidence of birth in these groups.

A recent meta-analysis reported that the probability of pregnancy after breast cancer among women with ER-positive tumors was about one-fourth that among women with ER-negative tumors.²⁹ Women with ER-positive tumors are often recommended to use 5 or more years of adjuvant endocrine therapy, and may be advised to postpone pregnancy until after therapy is completed. However, the POSITIVE study, an ongoing multicenter, single-arm trial, is currently investigating whether interruption of endocrine therapy to attempt pregnancy is associated with recurrence in women with hormone-responsive breast cancer.³⁰ In our cohort, women with ER-positive tumors were less likely to have given birth by 5 years post-diagnosis, but their cumulative incidence of childbirth matched that of women with ER-negative tumors at 10 years. This convergence likely reflects the pattern we observed among women with ER-positive tumors, in which the cumulative incidence rose more steeply between 5 and 10 years in endocrine therapy users compared to non-users.

Previous U.S.-based studies have suggested that low birth weight and preterm deliveries are more common among breast cancer survivors relative to the general population.^{14, 15, 18} However, among women who conceive after diagnosis, the magnitude of these associations appears to be modest. Recently, Hartnett et al. reported risk ratios (RR) of 1.3 (95% CI: 1.1–1.7) and 1.6 (95% CI: 1.3–2.0) for preterm birth and low birth weight, respectively, among postdiagnosis births to women diagnosed with invasive breast cancer at ages 20 to 45 in Georgia, Tennessee, and North Carolina relative to a matched comparison group. The corresponding RR for SGA was 1.2 (95% CI: 0.9–1.5).¹⁸ Likewise, in an older study from Denmark, PRs compared to women without cancer were 1.3 (95% CI: 0.7–2.2) for preterm birth and 1.2 (95% CI: 0.4–3.8) for low birth weight among women diagnosed with breast cancer before pregnancy.¹⁷ These estimates are similar to those of the current study for women with invasive breast cancer.

Our study adds information on variation in birth outcomes according to tumor and treatment characteristics. Chemotherapy was not strongly associated with either preterm birth or low birth weight among women, though the suggestive positive association with SGA in our data may warrant further investigation in larger studies. Exposure to chemotherapy may cause cardiovascular or pulmonary impairments, which could influence pregnancy outcomes through adverse effects on blood volume regulation.³¹ It is not immediately clear why low birth weight and preterm deliveries were more common among women in our cohort with ER negative tumors, even with adjustment for maternal age and race. However, these findings should be interpreted with caution, as they were based on a small number of outcomes and therefore are more subject to chance. It is possible that cancer recurrence and/or maternal characteristics such as comorbidities, socioeconomic status, and health behaviors could partially explain these associations. This information was not available in our data, but may be an important consideration for future studies of birth outcomes according to ER status.

Strengths of our study include the population-based design, and the 29,215 person-years of follow-up among almost 5000 AYA breast cancer survivors. Some limitations should be considered. We were unable to account for the potential impact of cancer recurrence on live birth incidence or birth outcomes, as this information is not captured in registry data. We also lacked detailed information on cancer therapies, such as chemotherapeutic agents and doses. Births to women who moved out of North Carolina during the study period would also not be captured in our data. However, census data suggest that only 7% of North Carolina women moved out of state during 2000–2010.³² Finally, our subgroup analyses of preterm birth and other birth outcomes were often limited by small sample sizes, and larger studies are needed to examine outcomes in groups defined by characteristics such as ER status.

Future reproductive outcomes are an important survivorship concern for many young women with a cancer diagnosis. Among AYA breast cancer survivors in our population-based study, the 10-year cumulative incidence of live birth was 8% overall, and was lower among women diagnosed in their 30s, those with regional or distant stage disease, and those treated with chemotherapy. Our findings reinforce the importance of fertility counseling and the use of accepted fertility preservation strategies for breast cancer patients who may want to have children after gonadotoxic cancer treatment. For women who do become pregnant after breast cancer, our overall findings suggest that the risk of adverse birth outcomes is not greatly elevated compared to women without cancer. However, the increased prevalence of preterm birth among women with ER-negative breast cancer may warrant further investigation in larger studies.

Novelty and Impact.

Reproductive outcomes are an important survivorship concern for adolescents and young adults with cancer. However, few studies have examined the incidence of live birth and birth outcomes among young women with a history of breast cancer. In this population-based study, <10% of breast cancer survivors had a live birth within 10 years of their diagnosis. Compared to women without cancer, preterm deliveries were significantly more common among breast cancer survivors with estrogen receptor-negative disease.

Abbreviations

AYA

adolescent and young adult

SGA

small-for-gestational age

CCR

Central Cancer Registry